Apatite composition: tracing petrogenetic processes in Transhimalayan granitoids
Transcript of Apatite composition: tracing petrogenetic processes in Transhimalayan granitoids
Apatite Composition Tracing PetrogeneticProcesses inTranshimalayan Granitoids
MEI-FEI CHU123 KUO-LUNG WANG2WILLIAM L GRIFFIN3SUN-LIN CHUNG1 SUZANNE Y OrsquoREILLY3 NORMAN J PEARSON3
AND YOSHIYUKI IIZUKA2
1DEPARTMENT OF GEOSCIENCES NATIONAL TAIWAN UNIVERSITY TAIPEI 10699 TAIWAN2INSTITUTE OF EARTH SCIENCES ACADEMIA SINICA TAIPEI 11529 TAIWAN3ARC NATIONAL KEY CENTRE FOR GEOCHEMICAL EVOLUTION AND METALLOGENY OF CONTINENTS (GEMOC)
DEPARTMENT OF EARTH AND PLANETARY SCIENCES MACQUARIE UNIVERSITY SYDNEY NSW 2109 AUSTRALIA
RECEIVED JULY 9 2008 ACCEPTEDJULY 28 2009
Apatites crystallized from different types of igneous rocks show sig-
nificant variations in the abundances of some minor and trace ele-
ments In this study electron probe microanalysis and laser ablation
inductively coupled plasma mass spectrometry were used to determine
the concentrations of 25 minor and trace elements in apatite separated
from three principal rock types of theTranshimalayan igneous pluto-
nic suite S-type granites the I-type Gangdese batholith and post-
collisional adakites F Mn Sr and rare earth elements (REE) in
apatite vary systematically with the composition of the host magma
and thus have high potential as petrogenetic tracers More specifi-
cally the F and Mn contents of apatite can be used as an indicator
of magma aluminosity or differentiation index Combined with Sr
and REE data which show significant variations in apatite from
different rock types these elements are useful for constructing lsquodis-
crimination diagramsrsquo This study also reveals that apatite has the
capacity to retain geochemical information about the host magma
through the course of magmatic evolution Systematic variations of
Sr and REE in apatite with bulk-rock aluminosity are the results of
partition competition with pre-existing and coexisting major and
accessory minerals in silicate melts and thus are useful for more
detailed investigations of petrogenetic processes such as fractional
crystallization and magma mixing which is signaled by inconsistent
Eu anomalies Sr abundances and REE patterns relative to bulk-
rock compositions
KEY WORDS apatite apatite minor elements apatite trace elements
igneous petrogenesis
I NTRODUCTIONThough tiny and sparse accessory minerals that concen-trate various geologically significant trace elements canprovide critical constraints on the details of igneous pro-cesses in magma chambers The rapid recent progress inmicro-analytical techniques (eg laser ablation inductivelycoupled plasma mass specrometry LA-ICP-MS) allowstheir in situ geochemical analysis and further applicationin igneous petrogenesis This type of microanalysis canprovide information that is not accessible through conven-tional bulk-rock analysis For example zircon a commonaccessory mineral has been intensively used to studymagma evolution the assembly of magma chambers andcrustal growth history by in situ analysis of hafnium oroxygen isotopes sometimes combined with U^Th^Pb agedeterminations and trace element patterns (eg Schalaquo reret al 1997 Griffin et al 2000 2002 Wilde et al 2001Valley 2003 Belousova et al 2006 Hawkesworth ampKemp 2006 Kemp et al 2007)Apatite although less intensively studied may be the
next candidate for such expanded application In terms ofits common occurrence stability during magma evolutionand chemical diversity apatite is comparable with zirconand even better in some respects Apatite is an early crys-tallizing and long-lasting phase that reaches saturationduring the evolution of a range of silicate melts (Hoskinet al 2000) In situ Sr-isotope analyses (Bizzarro et al
Corresponding author Telephone thorn61 2 9850 6125 Fax thorn61 2 98506904 E-mail mchuelsmqeduau
The Author 2009 Published by Oxford University Press Allrights reserved For Permissions please e-mail journalspermissionsoxfordjournalsorg
JOURNALOFPETROLOGY VOLUME 00 NUMBER 0 PAGES1^27 2009 doi101093petrologyegp054
Journal of Petrology Advance Access published September 3 2009
2003) and U^Th^Pb dating (Sano et al 1999 Willigerset al 2002) of apatite have shown preliminary successMore specifically the minor- and trace-element patternsof apatites vary with their host-rock type particularly thealuminosity (Bea 1996 Sha amp Chappell 1999 Belousovaet al 2001 2002 Hsieh et al 2008) and thus detrital apatitehas potential as a provenance indicator in sedimentaryrocksThere are two major sources of data on minor- and
trace-element abundances in apatite in plutonic rocks Shaamp Chappell (1999) reported minor- and trace-elementcompositions of apatite from lsquoclassicalrsquo I- and S-type grani-toids (SiO2457wt ) from the Lachlan Fold Belt easternAustralia and interpreted the data in terms of fractionalcrystallization redox conditions ionic substitution andmineral competition Belousova et al (2002) classified
apatite compositions in terms of their host-rock chemistryincluding the variable silica contents of granitoids Tounderstand the relationships between magmatic processesand the geochemical characteristics of apatite apatitesfrom a range of Transhimalayan plutonic rocks have beenanalysed in this study (Fig 1) The geochemical composi-tion of apatite from post-collisional adakites (Chung et al2003) is reported for the first time and apatites from theI-type Gangdese magmatic suite with SiO2 contents ran-ging from 52 to 74wt (Table 1) are interpreted interms of magma evolution during fractional crystalliza-tion Integration of these data comparing them with thegeochemical compositions of their host-rocks and the Hf-isotope composition of zircon in the same host-rocksallows an evaluation of the usefulness of apatite geochemis-try in studies of igneous processes
Fig 1 Simplified geological map showing sample localities and the distribution of major magmatic suites in the Transhimalayan domain ofsouthernTibet (after Chung et al 2003) BNS Bangong^Nujiang suture YTSYarlung^Tsangpo suture
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
2
Table1
Sum
marymajor-elementtrace-elementU^PbageNdisotopiccomposition
andmineralogicaldataforthestudiedTranshimalayan
samples
Sam
ple
Noof
mount(s)
Lat
(8N)
Long
(8E)
SiO
2
(wt)
Fe 2O3
(wt)
MnO
(wt)
CaO
(wt)
Na 2O
K2O
ASI
Sr
(ppm)
Th
(ppm)
Age
(Ma)
e Nd(T)
Majormineral
phases
Accessory
mineral
phases
Oligo-M
iocenepost-collisional
adakites
ST107B
A016
2927
9189
575
651
010
512
128
090
527
423
310
27
quartzthorn
alkalifeldsparthorn
apatitethorn
zirconthorn
ST107A
A015
2927
9189
653
328
005
367
111
097
708
381
303
22
amphibolethorn
biotitethorn
titanite
allanite
ET025C
A010
2969
9175
673
114
003
185
055
098
360
192
150
40
plagioclasethorn
opaq
ueminerals
T060B
A004
2952
9004
660
298
005
294
108
100
732
387
151
31
quartzthorn
alkalifeldsparthorn
apatitethorn
zirconthorn
ET023
A016
2961
9160
653
352
003
353
142
102
1048
88
170
03
biotitethorn
plagioclase
monazite
T016
A010
2957
9458
635
338
004
398
242
112
914
76
262
31
amphibolethorn
opaq
ueminerals
titanite
allanite
Gan
gdesebatholith(I-typ
e)
T044E
A001
A015
2949
8908
521
1125
017
777
251
076
633
68
483
thorn41
plagioclasethorn
amphibolethorn
apatitethorn
zircon
ST152A
A015
2933
9069
497
1093
018
101
133
078
646
02
527
thorn46
alkalifeldspar
pyroxene
titanite
epidote
T036C
A010
2932
9031
566
788
019
572
080
080
622
131
102
thorn31
biotite
quartzthorn
opaq
ue
rutile
T153
A010
3013
8541
534
887
015
643
074
081
985
260
499
14
minerals
ST141A
A013
2940
8909
506
950
016
813
155
083
613
37
905
thorn39
ST147A
A013
2940
9018
539
819
015
771
232
087
608
32
506
thorn42
ST129A
A013
2939
8963
577
754
015
616
153
091
551
81
941
thorn43
plagioclasethorn
amphibolethorn
apatitethorn
zircon
ST150A
A016
2929
9038
582
768
014
619
128
091
730
74
biotitethorn
alkalifeldspar
titanite
epidote
ET021E
A013
2969
9163
551
687
011
713
217
092
417
30
640
quartzthorn
opaq
ueminerals
monazite
pyroxene
ST143A
A014
2931
8981
576
771
013
664
161
092
493
90
848
thorn47
ST144A
A014
2932
8994
593
703
012
577
160
095
442
50
852
thorn48
ET026I
A001
A007
2948
9087
665
398
007
351
093
100
366
150
464
quartzthorn
alkalifeldsparthorn
apatitethorn
zircon
ST104A
A014
2927
9181
656
399
007
425
119
104
336
85
601
thorn23
plagioclasethorn
biotite
epidote
titanite
T148A
A016
3080
8512
707
288
005
303
122
107
267
132
amphibole
monazite
ST146A
A015
2935
9019
747
155
005
203
104
112
344
75
102thorn49
ST134Ay
A014
2952
8962
724
209
008
214
176
114
458
54
1881thorn59
T150
A015
3065
8513
747
114
005
093
077
118
714
294
[adakitic
Gan
gdese]
T027
A010
2900
9332
665
287
009
383
224
105
622
35
827
thorn22
plagioclasethorn
alkalifeldsparthorn
apatite
zirconep
idote
T024
A001
A016
2914
9375
695
222
007
354
255
113
738
24
804
thorn23
quartzthorn
biotitethorn
opaq
ue
muscovitetitanite
minerals
allanite
rutile
Northmag
matic
belt(S-typ
egranites)
T006C
A001
A007
A012
2999
9304
673
430
006
469
122
107
315
190
141
90
quartzthorn
orthoclasethorn
biotitethorn
apatitethorn
zircon
T138A
A009
3138
8670
686
288
006
152
060
121
252
416
129
plagioclase
muscovitethorn
allanite
monazite
T172A
A008
A012
3092
9258
725
217
003
118
046
123
121
367
opaq
ueminerals
titanite
xenotime
T170A
A008
3106
9243
733
230
004
142
066
123
116
206
92
rutile
T125A
A009
A012
3140
9001
723
196
005
179
062
129
212
142
121
T048C
A003
3011
8916
733
083
003
069
081
130
2970
143
153
Wen
etal(2008a2008b)
yChuet
al(2006)
ASIAluminium
SaturationIndexmolecu
larAl 2O3(Na 2Othorn
K2Othorn
CaO
)ratio
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
3
GEOLOGICAL BACKGROUNDAND SAMPLESThe continental collision between India and Asia startingin the early Cenozoic resulted in the rise of theTibetan pla-teau and the Himalayas (Molnar amp Tapponnier 1975)Tibet is a tectonic collage of three east^west-trendingGondwana-derived continental fragments from north tosouth these are the Songpan^Ganze Qiangtang andLhasa terranes (Fig 1) sequentially accreted to Asiaduring the Paleozoic to early Cenozoic (Alle gre et al1984) The Lhasa terrane the so-called Transhimalayabounds the southernmost part of Asia and contacts withIndia along the Yarlung^Tsangpo suture zone which isthe relic of the Neo-Tethys Ocean that existed before theIndia^Asia collision (Yin amp Harrison 2000)The Lhasa terrane consists primarily of Paleozoic to
Mesozoic sedimentary rocks associated with Mesozoic^Cenozoic igneous rocks (Pan et al 2004) The latter maybe categorized into three principal rock types accordingto their geochemistry andor occurrence (Fig 1) (1) S-typegranites in the northern magmatic belt (2) I-type grani-toids of the Gangdese batholith in the south (3) post-collisional adakites that occur as small plugs or sills cuttingacross the Gangdese batholith (Table 1) For this study apa-tite was separated from representative samples of each ofthese three suites
S-type granites the northernmagmatic beltThe northern magmatic belt is located in the northernpart of the Lhasa terrane (Fig 1) and is largely composedof Early Cretaceous (c 150^120 Ma) peraluminous orS-type granitoids (Xu et al 1985 Harris et al 1990 Panet al 2004) These rocks have evolved compositions with67^73wt SiO2 and negative eNd(T) values (Table 1)indicating derivation largely from the remelting of oldercrust Their main mineral constituents are quartz thorn alkalifeldspar thorn plagioclase thorn biotite thorn opaque minerals asso-ciated with accessory phases including apatite thorn zircon allanite monazite titanite xenotime (Table 1)Although zircon appears to be a common accessory min-eral most of the zircon in these rocks is inherited (seeChu et al 2006)There has been a long-term debate about the petrogen-
esis of the northern magmatic belt (see Kapp et al 2005)this has been attributed to (1) crustal anatexis during thecontinental collision of the Lhasa^Qiangtang terranes(eg Xu et al 1985) (2) high-temperature crustal meltingrelated to asthenospheric upwelling after the Lhasa^Qiangtang terrane collision (eg Harris et al 1990) (3)low-angle northward subduction of the Neo-Tethyan oce-anic slab (eg Coulon et al 1986) Our recent work(Chu et al 2006) which first identified the existence of
Jurassic-age Gangdese magmatism lends supports to thethird tectonic interpretation
I-type granitoids the Gangdese batholithsThe Gangdese batholith belt extends through most of thesouthern Lhasa terrane Its occurrence has been widelyattributed to northward subduction of Neo-Tethyan oce-anic lithosphere beneath the Lhasa terrane Magmatismtook place from Early Jurassic (Chu et al 2006) toPaleogene times (Wen et al 2008b and references therein)and resulted in both the voluminous Gangdese batholithand the associated Linzizong volcanic succession (Alle greet al 1984 Yin amp Harrison 2000 Lee et al 2007 2009Wen et al 2008b Fig 1)The Gangdese batholith consists dominantly of calc-
alkaline diorite typical of I-type granitoid compositions(Debon et al 1986 Wen et al 2008a 2008b Ji et al 2009Table 1) However the Gangdese rocks actually show awide range of compositions from gabbro to graniteSamples analyzed in this study for example have SiO2
contents varying from 50 to 75wt coupled with aneNd(T) range of ^14 to thorn59 (Wen 2007 Wen et al2008a) there is no correlation between silica content andNd isotopic composition With increasing whole-rocksilica content the major mineral constituents change fromplagioclase amphibole pyroxene biotite alkalifeldspar thorn opaque minerals to quartz thorn alkali feldspar thornbiotite thorn plagioclase amphibole thorn opaque mineralsApatite zircon titanite and epidote occur as commonaccessory phases (Table 1)Wen et al (2008a) reported the existence of a granodior-
ite body with adakitic geochemical characteristics heretermed lsquoadakitic Gangdesersquo (Table 1) which intruded at80 Ma in the southeastern part of the Gangdese batho-lith Its petrogenesis was attributed to a stage of flat sub-duction of the Neo-Tethyan slab
Post-collisional adakitesAdakites that were emplaced during Oligo-Miocene time(c 30^10 Ma) in the southern Lhasa terrane usually occuras small plugs or sills intruding the Gangdese batholith(see Chung et al 2005) These lsquocollision-typersquo adakitesshow overall geochemical characteristics similar to thoseof lsquonormalrsquo adakites formed in modern subduction zonesthat is the rocks are characterized by relative depletionsin heavy rare earth elements (HREE) and Y enrichmentin Sr and thus elevated SrY Garnet which stronglyretains HREE could have been a residual aluminum-richphase in the sources of the adakites which therefore havebeen interpreted as products of partial melting in a colli-sion-thickened mafic lower crust beneath southern Tibet(Chung et al 2003)Most of the adakites are of intermediate in composition
with 57^66wt SiO2 and eNd(T) of ^40 to ^03(Table 1) They are composed of quartz thorn alkali feldspar
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
thorn biotite thorn plagioclase amphibole thorn opaque mineralstogether with apatite thorn zircon titanite allanite monazite as accessory mineral phases (Table 1)
ANALYTICAL METHODSStandard heavy-liquid and magnetic methods were used toobtain pure apatite separates from 5 kg rock samplesApatite grains with no visible inclusions were hand pickedunder a stereo-microscope and then mounted in epoxydiscs Most of the Transhimalayan apatites in this studyare euhedral to subhedral granular and occasionally hex-agonal columnar in shape Back-scattered electron (BSE)images were taken of some of the apatites (eg Gangdesebatholith samples ET021E ST146A ST147A T036C andT153 S-type granites T006C and T172A) no internalzoning patterns were observed
Electron probe microanalysis (EPMA)Major- and minor-element contents of apatite were deter-mined by electron microprobe at the Institute of EarthSciences (IES) Academia Sinica Taiwan or at GEMOCMacquarie University Australia Up to four spot analyseswere performed on each crystal depending on the grainsize Apatite grains mounted in targets A001 to A006(Table 1) were analysed using a JEOL JXA-8900R electronmicroprobe using a wavelength-dispersive (WDS) methodthat employed TAP PET and LIF crystals with 2 mm spa-tial resolution15 kV beam conditions and 10 nA beam cur-rent For analysis of apatites in the remaining targets atGEMOC we used the methods described by Belousovaet al (2002) an electron beam of 10 mm diameter with anaccelerating voltage of 15 kV and a beam current of 20nA Analytical precision for most elements is better than1 but for F Cl and SO3 precision is around 5
Laser ablation-inductively coupled plasmamass spectrometry (LA-ICPMS)The trace-element contents of apatites were measured inthe LA-ICP-(Q)MS laboratory in the Department ofGeosciences National Taiwan University using an Agilent7500s ICP-MS system and a New Wave LUV213 lasersystem (Jackson 2001) Analytical methods which mainlyfollow those of Norman et al (1996 1998) involved time-resolved analysis (TRA) using a fast peak-hoppingsequence Each spot analysis consisted of c 60 s backgroundacquisition (gas blank) followed by c 60 s sample ablationused a 30 mm diameter laser beamTwenty-seven isotopeswere analysed in this study including 43Ca as the internalstandard (IS) using the concentrations obtained byEPMA Both 86Sr (isotopic abundance 986) and 88Sr(isotopic abundance 8258) were analysed to assessthe effect of linear calibration for variable elementcontents
Samples were measured in short runs bracketed by anal-yses of the external standard (ES) NIST SRM 610 withreference values taken from Norman et al (1996) Aftereach analysis data reduction was carried out usingVersion 40 of the lsquoreal-time on-linersquo GLITTER software(Griffin et al 2008) which allows the user to select a datarange free of effects produced by ablating inclusions Asthere is no matrix-matched and homogeneous apatite-likestandard available NIST SRM 610 NIST SRM 614 andUSGS international standards including BCR-2(G) TB-1(a basaltic glass) and GSE-1 (a doped rock-glass) wereanalysed as secondary standards during the analyticalruns to test the accuracy and precision of the LA-ICP-MSmethod The REE concentrations of apatites are generallybetween those of NIST SRM 610 and BCR-2(G)One hundred analyses of NIST SRM 610 show that the
minimum limit of detection (LOD) of this method formost trace elements in igneous rocks is around the ppmlevel commonly not more than 10 ppm For REE mini-mum LODs are substantially below the ppm level andmostly not more than 2 ppm Comparison of our resultswith those from the literature or the compiled values forstandard materials show that they are mutually consistentwith an accuracy better than 5 relative (ElectronicAppendix Table 1 available for downloading at httpwwwpetrologyoxfordjournalsorg) This agreementimplies that matrix effects are not significant during ourmeasurements when the doped synthetic glass is used asthe external standard for natural sample measurementsAccordingly this method can be applied to the trace ele-ment analysis of apatite For further comparison JC4 anapatite EPMA reference material was analysed as anunknown sample and these results are listed in ElectronicAppendixTable 1Precision for elements with mass480 is inversely corre-
lated with element abundance from 4^5 for 450 ppm(NIST SRM 610) to 7^14 for 08 ppm (NIST SRM614 Electronic AppendixTable 1) Regardless of concentra-tions or matrix the precision for Zn Rb Sn Sb Cs andPb (10^20) is significantly poorer than for the otheranalytes in any sample This lower precision may be dueto the fractionation of these elements relative to the IS(see Gulaquo nther et al 1999) or to heterogeneity in the refer-ence glasses on the scale of the LA-ICP-MS spatial resolu-tion (see Eggins amp Shelley 2002)
MINOR ELEMENTS ANALYTICALRESULTS AND DI SCUSSIONIn the following discussion the aluminum saturation index[ASI calculated as molecular Al2O3(Na2O thorn K2O thornCaO)] of the host-rocks is used to illustrate the range ofelemental variations in the Transhimalayan apatites Forthe studied samples ASI increases in general with the
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
5
silica content of the host-rock regardless of rock type(Table 1) If host-rock silica contents were used instead itwould not make any difference to our discussion or conclu-sions although the scatter of apatite compositions relativeto this index of magma evolution would become greater(see Fig 2 and Electronic Appendix Fig 1)In the subsequent discussion plutons with ASI 09 (57wt SiO2) except adakites are termed granites andthose with ASI 509 generally equivalent to 557wt SiO2 are termed lsquomafic rocksrsquoAccording to this classifica-tion the characteristics of minor element distributions inapatites (Electronic Appendix Table 2) from differenttypes of Transhimalayan igneous intrusions are summar-ized below
Fluorine and chlorineMost of the apatites are fluorapatite typical of igneousapatite with F contents41wt and FCl41 (ElectronicAppendixTable 2) The abundance of F increases and thatof Cl decreases with increasing ASI (Fig 2) Howeverapatites in some S-type granites (eg T1048 and T138)have Cl contents below the detection limits of EPMAthese analyses may have much higher FCl ratios thanthose shown calculated from the actual analyses (seeElectronic Appendix Table 2) Accordingly although Fand Cl abundance andor the FCl ratios of apatite havethe potential to monitor magma evolution the F contentis recommended as the better indicatorThe range of F concentrations analysed in one apatite
grain is up to 033 and the range within one rocksample generally is from 05 to 15 For a rock samplewith a specific ASI value the F variation in its apatites isup to 18
Manganese and ironMn variations in a single apatite separate can be up to01wt Whole-rock ASI values exert a control on theMn content of apatite in all types of Transhimalayan intru-sion (Fig 3) independent of the host-rock Mn abundance(Table 1) In metaluminous (ASI51) host rocks such asmost adakites (I-type) mafic rocks and some I-type gran-ites apatite consistently has 502 MnO In peralumi-nous rocks with ASI 11 both the MnO concentrationsof apatite and the proportions of apatite grains with402 MnO rise with the host-rock ASI values In rockswith 1 ASI511 the MnO abundances of apatite aretransitional Therefore the Mn content of apatite can beused as an indicator of the ASI of magmatic rocksIron concentrations in apatite from all Transhimalayan
plutonic rocks are mostly lower than 02 FeO except insome highly evolved S-type granites (eg T170A up to14 Electronic Appendix Table 2) The host-rock totaliron contents (Table 1) appear to have little effect on theFeO contents in apatite
SulfurSulfur (expressed as SO3wt ) in apatites from peralumi-nous rocks falls with increasing ASI of the host rocks tothe extent that more than half of the data are below theEPMA detection limit (001^0001wt SO3 ElectronicAppendix Table 2) In metaluminous Transhimalayan plu-tons the S contents of apatite cover a wide range (Fig 4)and are irregularly correlated with variation in host-rockASIAs for Mn and Fe S abundances in apatite may be
related to the redox condition of the host magma In an
Fig 2 F contents of apatites from different rock types apatite F vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
2003) and U^Th^Pb dating (Sano et al 1999 Willigerset al 2002) of apatite have shown preliminary successMore specifically the minor- and trace-element patternsof apatites vary with their host-rock type particularly thealuminosity (Bea 1996 Sha amp Chappell 1999 Belousovaet al 2001 2002 Hsieh et al 2008) and thus detrital apatitehas potential as a provenance indicator in sedimentaryrocksThere are two major sources of data on minor- and
trace-element abundances in apatite in plutonic rocks Shaamp Chappell (1999) reported minor- and trace-elementcompositions of apatite from lsquoclassicalrsquo I- and S-type grani-toids (SiO2457wt ) from the Lachlan Fold Belt easternAustralia and interpreted the data in terms of fractionalcrystallization redox conditions ionic substitution andmineral competition Belousova et al (2002) classified
apatite compositions in terms of their host-rock chemistryincluding the variable silica contents of granitoids Tounderstand the relationships between magmatic processesand the geochemical characteristics of apatite apatitesfrom a range of Transhimalayan plutonic rocks have beenanalysed in this study (Fig 1) The geochemical composi-tion of apatite from post-collisional adakites (Chung et al2003) is reported for the first time and apatites from theI-type Gangdese magmatic suite with SiO2 contents ran-ging from 52 to 74wt (Table 1) are interpreted interms of magma evolution during fractional crystalliza-tion Integration of these data comparing them with thegeochemical compositions of their host-rocks and the Hf-isotope composition of zircon in the same host-rocksallows an evaluation of the usefulness of apatite geochemis-try in studies of igneous processes
Fig 1 Simplified geological map showing sample localities and the distribution of major magmatic suites in the Transhimalayan domain ofsouthernTibet (after Chung et al 2003) BNS Bangong^Nujiang suture YTSYarlung^Tsangpo suture
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
2
Table1
Sum
marymajor-elementtrace-elementU^PbageNdisotopiccomposition
andmineralogicaldataforthestudiedTranshimalayan
samples
Sam
ple
Noof
mount(s)
Lat
(8N)
Long
(8E)
SiO
2
(wt)
Fe 2O3
(wt)
MnO
(wt)
CaO
(wt)
Na 2O
K2O
ASI
Sr
(ppm)
Th
(ppm)
Age
(Ma)
e Nd(T)
Majormineral
phases
Accessory
mineral
phases
Oligo-M
iocenepost-collisional
adakites
ST107B
A016
2927
9189
575
651
010
512
128
090
527
423
310
27
quartzthorn
alkalifeldsparthorn
apatitethorn
zirconthorn
ST107A
A015
2927
9189
653
328
005
367
111
097
708
381
303
22
amphibolethorn
biotitethorn
titanite
allanite
ET025C
A010
2969
9175
673
114
003
185
055
098
360
192
150
40
plagioclasethorn
opaq
ueminerals
T060B
A004
2952
9004
660
298
005
294
108
100
732
387
151
31
quartzthorn
alkalifeldsparthorn
apatitethorn
zirconthorn
ET023
A016
2961
9160
653
352
003
353
142
102
1048
88
170
03
biotitethorn
plagioclase
monazite
T016
A010
2957
9458
635
338
004
398
242
112
914
76
262
31
amphibolethorn
opaq
ueminerals
titanite
allanite
Gan
gdesebatholith(I-typ
e)
T044E
A001
A015
2949
8908
521
1125
017
777
251
076
633
68
483
thorn41
plagioclasethorn
amphibolethorn
apatitethorn
zircon
ST152A
A015
2933
9069
497
1093
018
101
133
078
646
02
527
thorn46
alkalifeldspar
pyroxene
titanite
epidote
T036C
A010
2932
9031
566
788
019
572
080
080
622
131
102
thorn31
biotite
quartzthorn
opaq
ue
rutile
T153
A010
3013
8541
534
887
015
643
074
081
985
260
499
14
minerals
ST141A
A013
2940
8909
506
950
016
813
155
083
613
37
905
thorn39
ST147A
A013
2940
9018
539
819
015
771
232
087
608
32
506
thorn42
ST129A
A013
2939
8963
577
754
015
616
153
091
551
81
941
thorn43
plagioclasethorn
amphibolethorn
apatitethorn
zircon
ST150A
A016
2929
9038
582
768
014
619
128
091
730
74
biotitethorn
alkalifeldspar
titanite
epidote
ET021E
A013
2969
9163
551
687
011
713
217
092
417
30
640
quartzthorn
opaq
ueminerals
monazite
pyroxene
ST143A
A014
2931
8981
576
771
013
664
161
092
493
90
848
thorn47
ST144A
A014
2932
8994
593
703
012
577
160
095
442
50
852
thorn48
ET026I
A001
A007
2948
9087
665
398
007
351
093
100
366
150
464
quartzthorn
alkalifeldsparthorn
apatitethorn
zircon
ST104A
A014
2927
9181
656
399
007
425
119
104
336
85
601
thorn23
plagioclasethorn
biotite
epidote
titanite
T148A
A016
3080
8512
707
288
005
303
122
107
267
132
amphibole
monazite
ST146A
A015
2935
9019
747
155
005
203
104
112
344
75
102thorn49
ST134Ay
A014
2952
8962
724
209
008
214
176
114
458
54
1881thorn59
T150
A015
3065
8513
747
114
005
093
077
118
714
294
[adakitic
Gan
gdese]
T027
A010
2900
9332
665
287
009
383
224
105
622
35
827
thorn22
plagioclasethorn
alkalifeldsparthorn
apatite
zirconep
idote
T024
A001
A016
2914
9375
695
222
007
354
255
113
738
24
804
thorn23
quartzthorn
biotitethorn
opaq
ue
muscovitetitanite
minerals
allanite
rutile
Northmag
matic
belt(S-typ
egranites)
T006C
A001
A007
A012
2999
9304
673
430
006
469
122
107
315
190
141
90
quartzthorn
orthoclasethorn
biotitethorn
apatitethorn
zircon
T138A
A009
3138
8670
686
288
006
152
060
121
252
416
129
plagioclase
muscovitethorn
allanite
monazite
T172A
A008
A012
3092
9258
725
217
003
118
046
123
121
367
opaq
ueminerals
titanite
xenotime
T170A
A008
3106
9243
733
230
004
142
066
123
116
206
92
rutile
T125A
A009
A012
3140
9001
723
196
005
179
062
129
212
142
121
T048C
A003
3011
8916
733
083
003
069
081
130
2970
143
153
Wen
etal(2008a2008b)
yChuet
al(2006)
ASIAluminium
SaturationIndexmolecu
larAl 2O3(Na 2Othorn
K2Othorn
CaO
)ratio
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
3
GEOLOGICAL BACKGROUNDAND SAMPLESThe continental collision between India and Asia startingin the early Cenozoic resulted in the rise of theTibetan pla-teau and the Himalayas (Molnar amp Tapponnier 1975)Tibet is a tectonic collage of three east^west-trendingGondwana-derived continental fragments from north tosouth these are the Songpan^Ganze Qiangtang andLhasa terranes (Fig 1) sequentially accreted to Asiaduring the Paleozoic to early Cenozoic (Alle gre et al1984) The Lhasa terrane the so-called Transhimalayabounds the southernmost part of Asia and contacts withIndia along the Yarlung^Tsangpo suture zone which isthe relic of the Neo-Tethys Ocean that existed before theIndia^Asia collision (Yin amp Harrison 2000)The Lhasa terrane consists primarily of Paleozoic to
Mesozoic sedimentary rocks associated with Mesozoic^Cenozoic igneous rocks (Pan et al 2004) The latter maybe categorized into three principal rock types accordingto their geochemistry andor occurrence (Fig 1) (1) S-typegranites in the northern magmatic belt (2) I-type grani-toids of the Gangdese batholith in the south (3) post-collisional adakites that occur as small plugs or sills cuttingacross the Gangdese batholith (Table 1) For this study apa-tite was separated from representative samples of each ofthese three suites
S-type granites the northernmagmatic beltThe northern magmatic belt is located in the northernpart of the Lhasa terrane (Fig 1) and is largely composedof Early Cretaceous (c 150^120 Ma) peraluminous orS-type granitoids (Xu et al 1985 Harris et al 1990 Panet al 2004) These rocks have evolved compositions with67^73wt SiO2 and negative eNd(T) values (Table 1)indicating derivation largely from the remelting of oldercrust Their main mineral constituents are quartz thorn alkalifeldspar thorn plagioclase thorn biotite thorn opaque minerals asso-ciated with accessory phases including apatite thorn zircon allanite monazite titanite xenotime (Table 1)Although zircon appears to be a common accessory min-eral most of the zircon in these rocks is inherited (seeChu et al 2006)There has been a long-term debate about the petrogen-
esis of the northern magmatic belt (see Kapp et al 2005)this has been attributed to (1) crustal anatexis during thecontinental collision of the Lhasa^Qiangtang terranes(eg Xu et al 1985) (2) high-temperature crustal meltingrelated to asthenospheric upwelling after the Lhasa^Qiangtang terrane collision (eg Harris et al 1990) (3)low-angle northward subduction of the Neo-Tethyan oce-anic slab (eg Coulon et al 1986) Our recent work(Chu et al 2006) which first identified the existence of
Jurassic-age Gangdese magmatism lends supports to thethird tectonic interpretation
I-type granitoids the Gangdese batholithsThe Gangdese batholith belt extends through most of thesouthern Lhasa terrane Its occurrence has been widelyattributed to northward subduction of Neo-Tethyan oce-anic lithosphere beneath the Lhasa terrane Magmatismtook place from Early Jurassic (Chu et al 2006) toPaleogene times (Wen et al 2008b and references therein)and resulted in both the voluminous Gangdese batholithand the associated Linzizong volcanic succession (Alle greet al 1984 Yin amp Harrison 2000 Lee et al 2007 2009Wen et al 2008b Fig 1)The Gangdese batholith consists dominantly of calc-
alkaline diorite typical of I-type granitoid compositions(Debon et al 1986 Wen et al 2008a 2008b Ji et al 2009Table 1) However the Gangdese rocks actually show awide range of compositions from gabbro to graniteSamples analyzed in this study for example have SiO2
contents varying from 50 to 75wt coupled with aneNd(T) range of ^14 to thorn59 (Wen 2007 Wen et al2008a) there is no correlation between silica content andNd isotopic composition With increasing whole-rocksilica content the major mineral constituents change fromplagioclase amphibole pyroxene biotite alkalifeldspar thorn opaque minerals to quartz thorn alkali feldspar thornbiotite thorn plagioclase amphibole thorn opaque mineralsApatite zircon titanite and epidote occur as commonaccessory phases (Table 1)Wen et al (2008a) reported the existence of a granodior-
ite body with adakitic geochemical characteristics heretermed lsquoadakitic Gangdesersquo (Table 1) which intruded at80 Ma in the southeastern part of the Gangdese batho-lith Its petrogenesis was attributed to a stage of flat sub-duction of the Neo-Tethyan slab
Post-collisional adakitesAdakites that were emplaced during Oligo-Miocene time(c 30^10 Ma) in the southern Lhasa terrane usually occuras small plugs or sills intruding the Gangdese batholith(see Chung et al 2005) These lsquocollision-typersquo adakitesshow overall geochemical characteristics similar to thoseof lsquonormalrsquo adakites formed in modern subduction zonesthat is the rocks are characterized by relative depletionsin heavy rare earth elements (HREE) and Y enrichmentin Sr and thus elevated SrY Garnet which stronglyretains HREE could have been a residual aluminum-richphase in the sources of the adakites which therefore havebeen interpreted as products of partial melting in a colli-sion-thickened mafic lower crust beneath southern Tibet(Chung et al 2003)Most of the adakites are of intermediate in composition
with 57^66wt SiO2 and eNd(T) of ^40 to ^03(Table 1) They are composed of quartz thorn alkali feldspar
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
thorn biotite thorn plagioclase amphibole thorn opaque mineralstogether with apatite thorn zircon titanite allanite monazite as accessory mineral phases (Table 1)
ANALYTICAL METHODSStandard heavy-liquid and magnetic methods were used toobtain pure apatite separates from 5 kg rock samplesApatite grains with no visible inclusions were hand pickedunder a stereo-microscope and then mounted in epoxydiscs Most of the Transhimalayan apatites in this studyare euhedral to subhedral granular and occasionally hex-agonal columnar in shape Back-scattered electron (BSE)images were taken of some of the apatites (eg Gangdesebatholith samples ET021E ST146A ST147A T036C andT153 S-type granites T006C and T172A) no internalzoning patterns were observed
Electron probe microanalysis (EPMA)Major- and minor-element contents of apatite were deter-mined by electron microprobe at the Institute of EarthSciences (IES) Academia Sinica Taiwan or at GEMOCMacquarie University Australia Up to four spot analyseswere performed on each crystal depending on the grainsize Apatite grains mounted in targets A001 to A006(Table 1) were analysed using a JEOL JXA-8900R electronmicroprobe using a wavelength-dispersive (WDS) methodthat employed TAP PET and LIF crystals with 2 mm spa-tial resolution15 kV beam conditions and 10 nA beam cur-rent For analysis of apatites in the remaining targets atGEMOC we used the methods described by Belousovaet al (2002) an electron beam of 10 mm diameter with anaccelerating voltage of 15 kV and a beam current of 20nA Analytical precision for most elements is better than1 but for F Cl and SO3 precision is around 5
Laser ablation-inductively coupled plasmamass spectrometry (LA-ICPMS)The trace-element contents of apatites were measured inthe LA-ICP-(Q)MS laboratory in the Department ofGeosciences National Taiwan University using an Agilent7500s ICP-MS system and a New Wave LUV213 lasersystem (Jackson 2001) Analytical methods which mainlyfollow those of Norman et al (1996 1998) involved time-resolved analysis (TRA) using a fast peak-hoppingsequence Each spot analysis consisted of c 60 s backgroundacquisition (gas blank) followed by c 60 s sample ablationused a 30 mm diameter laser beamTwenty-seven isotopeswere analysed in this study including 43Ca as the internalstandard (IS) using the concentrations obtained byEPMA Both 86Sr (isotopic abundance 986) and 88Sr(isotopic abundance 8258) were analysed to assessthe effect of linear calibration for variable elementcontents
Samples were measured in short runs bracketed by anal-yses of the external standard (ES) NIST SRM 610 withreference values taken from Norman et al (1996) Aftereach analysis data reduction was carried out usingVersion 40 of the lsquoreal-time on-linersquo GLITTER software(Griffin et al 2008) which allows the user to select a datarange free of effects produced by ablating inclusions Asthere is no matrix-matched and homogeneous apatite-likestandard available NIST SRM 610 NIST SRM 614 andUSGS international standards including BCR-2(G) TB-1(a basaltic glass) and GSE-1 (a doped rock-glass) wereanalysed as secondary standards during the analyticalruns to test the accuracy and precision of the LA-ICP-MSmethod The REE concentrations of apatites are generallybetween those of NIST SRM 610 and BCR-2(G)One hundred analyses of NIST SRM 610 show that the
minimum limit of detection (LOD) of this method formost trace elements in igneous rocks is around the ppmlevel commonly not more than 10 ppm For REE mini-mum LODs are substantially below the ppm level andmostly not more than 2 ppm Comparison of our resultswith those from the literature or the compiled values forstandard materials show that they are mutually consistentwith an accuracy better than 5 relative (ElectronicAppendix Table 1 available for downloading at httpwwwpetrologyoxfordjournalsorg) This agreementimplies that matrix effects are not significant during ourmeasurements when the doped synthetic glass is used asthe external standard for natural sample measurementsAccordingly this method can be applied to the trace ele-ment analysis of apatite For further comparison JC4 anapatite EPMA reference material was analysed as anunknown sample and these results are listed in ElectronicAppendixTable 1Precision for elements with mass480 is inversely corre-
lated with element abundance from 4^5 for 450 ppm(NIST SRM 610) to 7^14 for 08 ppm (NIST SRM614 Electronic AppendixTable 1) Regardless of concentra-tions or matrix the precision for Zn Rb Sn Sb Cs andPb (10^20) is significantly poorer than for the otheranalytes in any sample This lower precision may be dueto the fractionation of these elements relative to the IS(see Gulaquo nther et al 1999) or to heterogeneity in the refer-ence glasses on the scale of the LA-ICP-MS spatial resolu-tion (see Eggins amp Shelley 2002)
MINOR ELEMENTS ANALYTICALRESULTS AND DI SCUSSIONIn the following discussion the aluminum saturation index[ASI calculated as molecular Al2O3(Na2O thorn K2O thornCaO)] of the host-rocks is used to illustrate the range ofelemental variations in the Transhimalayan apatites Forthe studied samples ASI increases in general with the
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
5
silica content of the host-rock regardless of rock type(Table 1) If host-rock silica contents were used instead itwould not make any difference to our discussion or conclu-sions although the scatter of apatite compositions relativeto this index of magma evolution would become greater(see Fig 2 and Electronic Appendix Fig 1)In the subsequent discussion plutons with ASI 09 (57wt SiO2) except adakites are termed granites andthose with ASI 509 generally equivalent to 557wt SiO2 are termed lsquomafic rocksrsquoAccording to this classifica-tion the characteristics of minor element distributions inapatites (Electronic Appendix Table 2) from differenttypes of Transhimalayan igneous intrusions are summar-ized below
Fluorine and chlorineMost of the apatites are fluorapatite typical of igneousapatite with F contents41wt and FCl41 (ElectronicAppendixTable 2) The abundance of F increases and thatof Cl decreases with increasing ASI (Fig 2) Howeverapatites in some S-type granites (eg T1048 and T138)have Cl contents below the detection limits of EPMAthese analyses may have much higher FCl ratios thanthose shown calculated from the actual analyses (seeElectronic Appendix Table 2) Accordingly although Fand Cl abundance andor the FCl ratios of apatite havethe potential to monitor magma evolution the F contentis recommended as the better indicatorThe range of F concentrations analysed in one apatite
grain is up to 033 and the range within one rocksample generally is from 05 to 15 For a rock samplewith a specific ASI value the F variation in its apatites isup to 18
Manganese and ironMn variations in a single apatite separate can be up to01wt Whole-rock ASI values exert a control on theMn content of apatite in all types of Transhimalayan intru-sion (Fig 3) independent of the host-rock Mn abundance(Table 1) In metaluminous (ASI51) host rocks such asmost adakites (I-type) mafic rocks and some I-type gran-ites apatite consistently has 502 MnO In peralumi-nous rocks with ASI 11 both the MnO concentrationsof apatite and the proportions of apatite grains with402 MnO rise with the host-rock ASI values In rockswith 1 ASI511 the MnO abundances of apatite aretransitional Therefore the Mn content of apatite can beused as an indicator of the ASI of magmatic rocksIron concentrations in apatite from all Transhimalayan
plutonic rocks are mostly lower than 02 FeO except insome highly evolved S-type granites (eg T170A up to14 Electronic Appendix Table 2) The host-rock totaliron contents (Table 1) appear to have little effect on theFeO contents in apatite
SulfurSulfur (expressed as SO3wt ) in apatites from peralumi-nous rocks falls with increasing ASI of the host rocks tothe extent that more than half of the data are below theEPMA detection limit (001^0001wt SO3 ElectronicAppendix Table 2) In metaluminous Transhimalayan plu-tons the S contents of apatite cover a wide range (Fig 4)and are irregularly correlated with variation in host-rockASIAs for Mn and Fe S abundances in apatite may be
related to the redox condition of the host magma In an
Fig 2 F contents of apatites from different rock types apatite F vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Table1
Sum
marymajor-elementtrace-elementU^PbageNdisotopiccomposition
andmineralogicaldataforthestudiedTranshimalayan
samples
Sam
ple
Noof
mount(s)
Lat
(8N)
Long
(8E)
SiO
2
(wt)
Fe 2O3
(wt)
MnO
(wt)
CaO
(wt)
Na 2O
K2O
ASI
Sr
(ppm)
Th
(ppm)
Age
(Ma)
e Nd(T)
Majormineral
phases
Accessory
mineral
phases
Oligo-M
iocenepost-collisional
adakites
ST107B
A016
2927
9189
575
651
010
512
128
090
527
423
310
27
quartzthorn
alkalifeldsparthorn
apatitethorn
zirconthorn
ST107A
A015
2927
9189
653
328
005
367
111
097
708
381
303
22
amphibolethorn
biotitethorn
titanite
allanite
ET025C
A010
2969
9175
673
114
003
185
055
098
360
192
150
40
plagioclasethorn
opaq
ueminerals
T060B
A004
2952
9004
660
298
005
294
108
100
732
387
151
31
quartzthorn
alkalifeldsparthorn
apatitethorn
zirconthorn
ET023
A016
2961
9160
653
352
003
353
142
102
1048
88
170
03
biotitethorn
plagioclase
monazite
T016
A010
2957
9458
635
338
004
398
242
112
914
76
262
31
amphibolethorn
opaq
ueminerals
titanite
allanite
Gan
gdesebatholith(I-typ
e)
T044E
A001
A015
2949
8908
521
1125
017
777
251
076
633
68
483
thorn41
plagioclasethorn
amphibolethorn
apatitethorn
zircon
ST152A
A015
2933
9069
497
1093
018
101
133
078
646
02
527
thorn46
alkalifeldspar
pyroxene
titanite
epidote
T036C
A010
2932
9031
566
788
019
572
080
080
622
131
102
thorn31
biotite
quartzthorn
opaq
ue
rutile
T153
A010
3013
8541
534
887
015
643
074
081
985
260
499
14
minerals
ST141A
A013
2940
8909
506
950
016
813
155
083
613
37
905
thorn39
ST147A
A013
2940
9018
539
819
015
771
232
087
608
32
506
thorn42
ST129A
A013
2939
8963
577
754
015
616
153
091
551
81
941
thorn43
plagioclasethorn
amphibolethorn
apatitethorn
zircon
ST150A
A016
2929
9038
582
768
014
619
128
091
730
74
biotitethorn
alkalifeldspar
titanite
epidote
ET021E
A013
2969
9163
551
687
011
713
217
092
417
30
640
quartzthorn
opaq
ueminerals
monazite
pyroxene
ST143A
A014
2931
8981
576
771
013
664
161
092
493
90
848
thorn47
ST144A
A014
2932
8994
593
703
012
577
160
095
442
50
852
thorn48
ET026I
A001
A007
2948
9087
665
398
007
351
093
100
366
150
464
quartzthorn
alkalifeldsparthorn
apatitethorn
zircon
ST104A
A014
2927
9181
656
399
007
425
119
104
336
85
601
thorn23
plagioclasethorn
biotite
epidote
titanite
T148A
A016
3080
8512
707
288
005
303
122
107
267
132
amphibole
monazite
ST146A
A015
2935
9019
747
155
005
203
104
112
344
75
102thorn49
ST134Ay
A014
2952
8962
724
209
008
214
176
114
458
54
1881thorn59
T150
A015
3065
8513
747
114
005
093
077
118
714
294
[adakitic
Gan
gdese]
T027
A010
2900
9332
665
287
009
383
224
105
622
35
827
thorn22
plagioclasethorn
alkalifeldsparthorn
apatite
zirconep
idote
T024
A001
A016
2914
9375
695
222
007
354
255
113
738
24
804
thorn23
quartzthorn
biotitethorn
opaq
ue
muscovitetitanite
minerals
allanite
rutile
Northmag
matic
belt(S-typ
egranites)
T006C
A001
A007
A012
2999
9304
673
430
006
469
122
107
315
190
141
90
quartzthorn
orthoclasethorn
biotitethorn
apatitethorn
zircon
T138A
A009
3138
8670
686
288
006
152
060
121
252
416
129
plagioclase
muscovitethorn
allanite
monazite
T172A
A008
A012
3092
9258
725
217
003
118
046
123
121
367
opaq
ueminerals
titanite
xenotime
T170A
A008
3106
9243
733
230
004
142
066
123
116
206
92
rutile
T125A
A009
A012
3140
9001
723
196
005
179
062
129
212
142
121
T048C
A003
3011
8916
733
083
003
069
081
130
2970
143
153
Wen
etal(2008a2008b)
yChuet
al(2006)
ASIAluminium
SaturationIndexmolecu
larAl 2O3(Na 2Othorn
K2Othorn
CaO
)ratio
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
3
GEOLOGICAL BACKGROUNDAND SAMPLESThe continental collision between India and Asia startingin the early Cenozoic resulted in the rise of theTibetan pla-teau and the Himalayas (Molnar amp Tapponnier 1975)Tibet is a tectonic collage of three east^west-trendingGondwana-derived continental fragments from north tosouth these are the Songpan^Ganze Qiangtang andLhasa terranes (Fig 1) sequentially accreted to Asiaduring the Paleozoic to early Cenozoic (Alle gre et al1984) The Lhasa terrane the so-called Transhimalayabounds the southernmost part of Asia and contacts withIndia along the Yarlung^Tsangpo suture zone which isthe relic of the Neo-Tethys Ocean that existed before theIndia^Asia collision (Yin amp Harrison 2000)The Lhasa terrane consists primarily of Paleozoic to
Mesozoic sedimentary rocks associated with Mesozoic^Cenozoic igneous rocks (Pan et al 2004) The latter maybe categorized into three principal rock types accordingto their geochemistry andor occurrence (Fig 1) (1) S-typegranites in the northern magmatic belt (2) I-type grani-toids of the Gangdese batholith in the south (3) post-collisional adakites that occur as small plugs or sills cuttingacross the Gangdese batholith (Table 1) For this study apa-tite was separated from representative samples of each ofthese three suites
S-type granites the northernmagmatic beltThe northern magmatic belt is located in the northernpart of the Lhasa terrane (Fig 1) and is largely composedof Early Cretaceous (c 150^120 Ma) peraluminous orS-type granitoids (Xu et al 1985 Harris et al 1990 Panet al 2004) These rocks have evolved compositions with67^73wt SiO2 and negative eNd(T) values (Table 1)indicating derivation largely from the remelting of oldercrust Their main mineral constituents are quartz thorn alkalifeldspar thorn plagioclase thorn biotite thorn opaque minerals asso-ciated with accessory phases including apatite thorn zircon allanite monazite titanite xenotime (Table 1)Although zircon appears to be a common accessory min-eral most of the zircon in these rocks is inherited (seeChu et al 2006)There has been a long-term debate about the petrogen-
esis of the northern magmatic belt (see Kapp et al 2005)this has been attributed to (1) crustal anatexis during thecontinental collision of the Lhasa^Qiangtang terranes(eg Xu et al 1985) (2) high-temperature crustal meltingrelated to asthenospheric upwelling after the Lhasa^Qiangtang terrane collision (eg Harris et al 1990) (3)low-angle northward subduction of the Neo-Tethyan oce-anic slab (eg Coulon et al 1986) Our recent work(Chu et al 2006) which first identified the existence of
Jurassic-age Gangdese magmatism lends supports to thethird tectonic interpretation
I-type granitoids the Gangdese batholithsThe Gangdese batholith belt extends through most of thesouthern Lhasa terrane Its occurrence has been widelyattributed to northward subduction of Neo-Tethyan oce-anic lithosphere beneath the Lhasa terrane Magmatismtook place from Early Jurassic (Chu et al 2006) toPaleogene times (Wen et al 2008b and references therein)and resulted in both the voluminous Gangdese batholithand the associated Linzizong volcanic succession (Alle greet al 1984 Yin amp Harrison 2000 Lee et al 2007 2009Wen et al 2008b Fig 1)The Gangdese batholith consists dominantly of calc-
alkaline diorite typical of I-type granitoid compositions(Debon et al 1986 Wen et al 2008a 2008b Ji et al 2009Table 1) However the Gangdese rocks actually show awide range of compositions from gabbro to graniteSamples analyzed in this study for example have SiO2
contents varying from 50 to 75wt coupled with aneNd(T) range of ^14 to thorn59 (Wen 2007 Wen et al2008a) there is no correlation between silica content andNd isotopic composition With increasing whole-rocksilica content the major mineral constituents change fromplagioclase amphibole pyroxene biotite alkalifeldspar thorn opaque minerals to quartz thorn alkali feldspar thornbiotite thorn plagioclase amphibole thorn opaque mineralsApatite zircon titanite and epidote occur as commonaccessory phases (Table 1)Wen et al (2008a) reported the existence of a granodior-
ite body with adakitic geochemical characteristics heretermed lsquoadakitic Gangdesersquo (Table 1) which intruded at80 Ma in the southeastern part of the Gangdese batho-lith Its petrogenesis was attributed to a stage of flat sub-duction of the Neo-Tethyan slab
Post-collisional adakitesAdakites that were emplaced during Oligo-Miocene time(c 30^10 Ma) in the southern Lhasa terrane usually occuras small plugs or sills intruding the Gangdese batholith(see Chung et al 2005) These lsquocollision-typersquo adakitesshow overall geochemical characteristics similar to thoseof lsquonormalrsquo adakites formed in modern subduction zonesthat is the rocks are characterized by relative depletionsin heavy rare earth elements (HREE) and Y enrichmentin Sr and thus elevated SrY Garnet which stronglyretains HREE could have been a residual aluminum-richphase in the sources of the adakites which therefore havebeen interpreted as products of partial melting in a colli-sion-thickened mafic lower crust beneath southern Tibet(Chung et al 2003)Most of the adakites are of intermediate in composition
with 57^66wt SiO2 and eNd(T) of ^40 to ^03(Table 1) They are composed of quartz thorn alkali feldspar
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
thorn biotite thorn plagioclase amphibole thorn opaque mineralstogether with apatite thorn zircon titanite allanite monazite as accessory mineral phases (Table 1)
ANALYTICAL METHODSStandard heavy-liquid and magnetic methods were used toobtain pure apatite separates from 5 kg rock samplesApatite grains with no visible inclusions were hand pickedunder a stereo-microscope and then mounted in epoxydiscs Most of the Transhimalayan apatites in this studyare euhedral to subhedral granular and occasionally hex-agonal columnar in shape Back-scattered electron (BSE)images were taken of some of the apatites (eg Gangdesebatholith samples ET021E ST146A ST147A T036C andT153 S-type granites T006C and T172A) no internalzoning patterns were observed
Electron probe microanalysis (EPMA)Major- and minor-element contents of apatite were deter-mined by electron microprobe at the Institute of EarthSciences (IES) Academia Sinica Taiwan or at GEMOCMacquarie University Australia Up to four spot analyseswere performed on each crystal depending on the grainsize Apatite grains mounted in targets A001 to A006(Table 1) were analysed using a JEOL JXA-8900R electronmicroprobe using a wavelength-dispersive (WDS) methodthat employed TAP PET and LIF crystals with 2 mm spa-tial resolution15 kV beam conditions and 10 nA beam cur-rent For analysis of apatites in the remaining targets atGEMOC we used the methods described by Belousovaet al (2002) an electron beam of 10 mm diameter with anaccelerating voltage of 15 kV and a beam current of 20nA Analytical precision for most elements is better than1 but for F Cl and SO3 precision is around 5
Laser ablation-inductively coupled plasmamass spectrometry (LA-ICPMS)The trace-element contents of apatites were measured inthe LA-ICP-(Q)MS laboratory in the Department ofGeosciences National Taiwan University using an Agilent7500s ICP-MS system and a New Wave LUV213 lasersystem (Jackson 2001) Analytical methods which mainlyfollow those of Norman et al (1996 1998) involved time-resolved analysis (TRA) using a fast peak-hoppingsequence Each spot analysis consisted of c 60 s backgroundacquisition (gas blank) followed by c 60 s sample ablationused a 30 mm diameter laser beamTwenty-seven isotopeswere analysed in this study including 43Ca as the internalstandard (IS) using the concentrations obtained byEPMA Both 86Sr (isotopic abundance 986) and 88Sr(isotopic abundance 8258) were analysed to assessthe effect of linear calibration for variable elementcontents
Samples were measured in short runs bracketed by anal-yses of the external standard (ES) NIST SRM 610 withreference values taken from Norman et al (1996) Aftereach analysis data reduction was carried out usingVersion 40 of the lsquoreal-time on-linersquo GLITTER software(Griffin et al 2008) which allows the user to select a datarange free of effects produced by ablating inclusions Asthere is no matrix-matched and homogeneous apatite-likestandard available NIST SRM 610 NIST SRM 614 andUSGS international standards including BCR-2(G) TB-1(a basaltic glass) and GSE-1 (a doped rock-glass) wereanalysed as secondary standards during the analyticalruns to test the accuracy and precision of the LA-ICP-MSmethod The REE concentrations of apatites are generallybetween those of NIST SRM 610 and BCR-2(G)One hundred analyses of NIST SRM 610 show that the
minimum limit of detection (LOD) of this method formost trace elements in igneous rocks is around the ppmlevel commonly not more than 10 ppm For REE mini-mum LODs are substantially below the ppm level andmostly not more than 2 ppm Comparison of our resultswith those from the literature or the compiled values forstandard materials show that they are mutually consistentwith an accuracy better than 5 relative (ElectronicAppendix Table 1 available for downloading at httpwwwpetrologyoxfordjournalsorg) This agreementimplies that matrix effects are not significant during ourmeasurements when the doped synthetic glass is used asthe external standard for natural sample measurementsAccordingly this method can be applied to the trace ele-ment analysis of apatite For further comparison JC4 anapatite EPMA reference material was analysed as anunknown sample and these results are listed in ElectronicAppendixTable 1Precision for elements with mass480 is inversely corre-
lated with element abundance from 4^5 for 450 ppm(NIST SRM 610) to 7^14 for 08 ppm (NIST SRM614 Electronic AppendixTable 1) Regardless of concentra-tions or matrix the precision for Zn Rb Sn Sb Cs andPb (10^20) is significantly poorer than for the otheranalytes in any sample This lower precision may be dueto the fractionation of these elements relative to the IS(see Gulaquo nther et al 1999) or to heterogeneity in the refer-ence glasses on the scale of the LA-ICP-MS spatial resolu-tion (see Eggins amp Shelley 2002)
MINOR ELEMENTS ANALYTICALRESULTS AND DI SCUSSIONIn the following discussion the aluminum saturation index[ASI calculated as molecular Al2O3(Na2O thorn K2O thornCaO)] of the host-rocks is used to illustrate the range ofelemental variations in the Transhimalayan apatites Forthe studied samples ASI increases in general with the
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
5
silica content of the host-rock regardless of rock type(Table 1) If host-rock silica contents were used instead itwould not make any difference to our discussion or conclu-sions although the scatter of apatite compositions relativeto this index of magma evolution would become greater(see Fig 2 and Electronic Appendix Fig 1)In the subsequent discussion plutons with ASI 09 (57wt SiO2) except adakites are termed granites andthose with ASI 509 generally equivalent to 557wt SiO2 are termed lsquomafic rocksrsquoAccording to this classifica-tion the characteristics of minor element distributions inapatites (Electronic Appendix Table 2) from differenttypes of Transhimalayan igneous intrusions are summar-ized below
Fluorine and chlorineMost of the apatites are fluorapatite typical of igneousapatite with F contents41wt and FCl41 (ElectronicAppendixTable 2) The abundance of F increases and thatof Cl decreases with increasing ASI (Fig 2) Howeverapatites in some S-type granites (eg T1048 and T138)have Cl contents below the detection limits of EPMAthese analyses may have much higher FCl ratios thanthose shown calculated from the actual analyses (seeElectronic Appendix Table 2) Accordingly although Fand Cl abundance andor the FCl ratios of apatite havethe potential to monitor magma evolution the F contentis recommended as the better indicatorThe range of F concentrations analysed in one apatite
grain is up to 033 and the range within one rocksample generally is from 05 to 15 For a rock samplewith a specific ASI value the F variation in its apatites isup to 18
Manganese and ironMn variations in a single apatite separate can be up to01wt Whole-rock ASI values exert a control on theMn content of apatite in all types of Transhimalayan intru-sion (Fig 3) independent of the host-rock Mn abundance(Table 1) In metaluminous (ASI51) host rocks such asmost adakites (I-type) mafic rocks and some I-type gran-ites apatite consistently has 502 MnO In peralumi-nous rocks with ASI 11 both the MnO concentrationsof apatite and the proportions of apatite grains with402 MnO rise with the host-rock ASI values In rockswith 1 ASI511 the MnO abundances of apatite aretransitional Therefore the Mn content of apatite can beused as an indicator of the ASI of magmatic rocksIron concentrations in apatite from all Transhimalayan
plutonic rocks are mostly lower than 02 FeO except insome highly evolved S-type granites (eg T170A up to14 Electronic Appendix Table 2) The host-rock totaliron contents (Table 1) appear to have little effect on theFeO contents in apatite
SulfurSulfur (expressed as SO3wt ) in apatites from peralumi-nous rocks falls with increasing ASI of the host rocks tothe extent that more than half of the data are below theEPMA detection limit (001^0001wt SO3 ElectronicAppendix Table 2) In metaluminous Transhimalayan plu-tons the S contents of apatite cover a wide range (Fig 4)and are irregularly correlated with variation in host-rockASIAs for Mn and Fe S abundances in apatite may be
related to the redox condition of the host magma In an
Fig 2 F contents of apatites from different rock types apatite F vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
GEOLOGICAL BACKGROUNDAND SAMPLESThe continental collision between India and Asia startingin the early Cenozoic resulted in the rise of theTibetan pla-teau and the Himalayas (Molnar amp Tapponnier 1975)Tibet is a tectonic collage of three east^west-trendingGondwana-derived continental fragments from north tosouth these are the Songpan^Ganze Qiangtang andLhasa terranes (Fig 1) sequentially accreted to Asiaduring the Paleozoic to early Cenozoic (Alle gre et al1984) The Lhasa terrane the so-called Transhimalayabounds the southernmost part of Asia and contacts withIndia along the Yarlung^Tsangpo suture zone which isthe relic of the Neo-Tethys Ocean that existed before theIndia^Asia collision (Yin amp Harrison 2000)The Lhasa terrane consists primarily of Paleozoic to
Mesozoic sedimentary rocks associated with Mesozoic^Cenozoic igneous rocks (Pan et al 2004) The latter maybe categorized into three principal rock types accordingto their geochemistry andor occurrence (Fig 1) (1) S-typegranites in the northern magmatic belt (2) I-type grani-toids of the Gangdese batholith in the south (3) post-collisional adakites that occur as small plugs or sills cuttingacross the Gangdese batholith (Table 1) For this study apa-tite was separated from representative samples of each ofthese three suites
S-type granites the northernmagmatic beltThe northern magmatic belt is located in the northernpart of the Lhasa terrane (Fig 1) and is largely composedof Early Cretaceous (c 150^120 Ma) peraluminous orS-type granitoids (Xu et al 1985 Harris et al 1990 Panet al 2004) These rocks have evolved compositions with67^73wt SiO2 and negative eNd(T) values (Table 1)indicating derivation largely from the remelting of oldercrust Their main mineral constituents are quartz thorn alkalifeldspar thorn plagioclase thorn biotite thorn opaque minerals asso-ciated with accessory phases including apatite thorn zircon allanite monazite titanite xenotime (Table 1)Although zircon appears to be a common accessory min-eral most of the zircon in these rocks is inherited (seeChu et al 2006)There has been a long-term debate about the petrogen-
esis of the northern magmatic belt (see Kapp et al 2005)this has been attributed to (1) crustal anatexis during thecontinental collision of the Lhasa^Qiangtang terranes(eg Xu et al 1985) (2) high-temperature crustal meltingrelated to asthenospheric upwelling after the Lhasa^Qiangtang terrane collision (eg Harris et al 1990) (3)low-angle northward subduction of the Neo-Tethyan oce-anic slab (eg Coulon et al 1986) Our recent work(Chu et al 2006) which first identified the existence of
Jurassic-age Gangdese magmatism lends supports to thethird tectonic interpretation
I-type granitoids the Gangdese batholithsThe Gangdese batholith belt extends through most of thesouthern Lhasa terrane Its occurrence has been widelyattributed to northward subduction of Neo-Tethyan oce-anic lithosphere beneath the Lhasa terrane Magmatismtook place from Early Jurassic (Chu et al 2006) toPaleogene times (Wen et al 2008b and references therein)and resulted in both the voluminous Gangdese batholithand the associated Linzizong volcanic succession (Alle greet al 1984 Yin amp Harrison 2000 Lee et al 2007 2009Wen et al 2008b Fig 1)The Gangdese batholith consists dominantly of calc-
alkaline diorite typical of I-type granitoid compositions(Debon et al 1986 Wen et al 2008a 2008b Ji et al 2009Table 1) However the Gangdese rocks actually show awide range of compositions from gabbro to graniteSamples analyzed in this study for example have SiO2
contents varying from 50 to 75wt coupled with aneNd(T) range of ^14 to thorn59 (Wen 2007 Wen et al2008a) there is no correlation between silica content andNd isotopic composition With increasing whole-rocksilica content the major mineral constituents change fromplagioclase amphibole pyroxene biotite alkalifeldspar thorn opaque minerals to quartz thorn alkali feldspar thornbiotite thorn plagioclase amphibole thorn opaque mineralsApatite zircon titanite and epidote occur as commonaccessory phases (Table 1)Wen et al (2008a) reported the existence of a granodior-
ite body with adakitic geochemical characteristics heretermed lsquoadakitic Gangdesersquo (Table 1) which intruded at80 Ma in the southeastern part of the Gangdese batho-lith Its petrogenesis was attributed to a stage of flat sub-duction of the Neo-Tethyan slab
Post-collisional adakitesAdakites that were emplaced during Oligo-Miocene time(c 30^10 Ma) in the southern Lhasa terrane usually occuras small plugs or sills intruding the Gangdese batholith(see Chung et al 2005) These lsquocollision-typersquo adakitesshow overall geochemical characteristics similar to thoseof lsquonormalrsquo adakites formed in modern subduction zonesthat is the rocks are characterized by relative depletionsin heavy rare earth elements (HREE) and Y enrichmentin Sr and thus elevated SrY Garnet which stronglyretains HREE could have been a residual aluminum-richphase in the sources of the adakites which therefore havebeen interpreted as products of partial melting in a colli-sion-thickened mafic lower crust beneath southern Tibet(Chung et al 2003)Most of the adakites are of intermediate in composition
with 57^66wt SiO2 and eNd(T) of ^40 to ^03(Table 1) They are composed of quartz thorn alkali feldspar
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
4
thorn biotite thorn plagioclase amphibole thorn opaque mineralstogether with apatite thorn zircon titanite allanite monazite as accessory mineral phases (Table 1)
ANALYTICAL METHODSStandard heavy-liquid and magnetic methods were used toobtain pure apatite separates from 5 kg rock samplesApatite grains with no visible inclusions were hand pickedunder a stereo-microscope and then mounted in epoxydiscs Most of the Transhimalayan apatites in this studyare euhedral to subhedral granular and occasionally hex-agonal columnar in shape Back-scattered electron (BSE)images were taken of some of the apatites (eg Gangdesebatholith samples ET021E ST146A ST147A T036C andT153 S-type granites T006C and T172A) no internalzoning patterns were observed
Electron probe microanalysis (EPMA)Major- and minor-element contents of apatite were deter-mined by electron microprobe at the Institute of EarthSciences (IES) Academia Sinica Taiwan or at GEMOCMacquarie University Australia Up to four spot analyseswere performed on each crystal depending on the grainsize Apatite grains mounted in targets A001 to A006(Table 1) were analysed using a JEOL JXA-8900R electronmicroprobe using a wavelength-dispersive (WDS) methodthat employed TAP PET and LIF crystals with 2 mm spa-tial resolution15 kV beam conditions and 10 nA beam cur-rent For analysis of apatites in the remaining targets atGEMOC we used the methods described by Belousovaet al (2002) an electron beam of 10 mm diameter with anaccelerating voltage of 15 kV and a beam current of 20nA Analytical precision for most elements is better than1 but for F Cl and SO3 precision is around 5
Laser ablation-inductively coupled plasmamass spectrometry (LA-ICPMS)The trace-element contents of apatites were measured inthe LA-ICP-(Q)MS laboratory in the Department ofGeosciences National Taiwan University using an Agilent7500s ICP-MS system and a New Wave LUV213 lasersystem (Jackson 2001) Analytical methods which mainlyfollow those of Norman et al (1996 1998) involved time-resolved analysis (TRA) using a fast peak-hoppingsequence Each spot analysis consisted of c 60 s backgroundacquisition (gas blank) followed by c 60 s sample ablationused a 30 mm diameter laser beamTwenty-seven isotopeswere analysed in this study including 43Ca as the internalstandard (IS) using the concentrations obtained byEPMA Both 86Sr (isotopic abundance 986) and 88Sr(isotopic abundance 8258) were analysed to assessthe effect of linear calibration for variable elementcontents
Samples were measured in short runs bracketed by anal-yses of the external standard (ES) NIST SRM 610 withreference values taken from Norman et al (1996) Aftereach analysis data reduction was carried out usingVersion 40 of the lsquoreal-time on-linersquo GLITTER software(Griffin et al 2008) which allows the user to select a datarange free of effects produced by ablating inclusions Asthere is no matrix-matched and homogeneous apatite-likestandard available NIST SRM 610 NIST SRM 614 andUSGS international standards including BCR-2(G) TB-1(a basaltic glass) and GSE-1 (a doped rock-glass) wereanalysed as secondary standards during the analyticalruns to test the accuracy and precision of the LA-ICP-MSmethod The REE concentrations of apatites are generallybetween those of NIST SRM 610 and BCR-2(G)One hundred analyses of NIST SRM 610 show that the
minimum limit of detection (LOD) of this method formost trace elements in igneous rocks is around the ppmlevel commonly not more than 10 ppm For REE mini-mum LODs are substantially below the ppm level andmostly not more than 2 ppm Comparison of our resultswith those from the literature or the compiled values forstandard materials show that they are mutually consistentwith an accuracy better than 5 relative (ElectronicAppendix Table 1 available for downloading at httpwwwpetrologyoxfordjournalsorg) This agreementimplies that matrix effects are not significant during ourmeasurements when the doped synthetic glass is used asthe external standard for natural sample measurementsAccordingly this method can be applied to the trace ele-ment analysis of apatite For further comparison JC4 anapatite EPMA reference material was analysed as anunknown sample and these results are listed in ElectronicAppendixTable 1Precision for elements with mass480 is inversely corre-
lated with element abundance from 4^5 for 450 ppm(NIST SRM 610) to 7^14 for 08 ppm (NIST SRM614 Electronic AppendixTable 1) Regardless of concentra-tions or matrix the precision for Zn Rb Sn Sb Cs andPb (10^20) is significantly poorer than for the otheranalytes in any sample This lower precision may be dueto the fractionation of these elements relative to the IS(see Gulaquo nther et al 1999) or to heterogeneity in the refer-ence glasses on the scale of the LA-ICP-MS spatial resolu-tion (see Eggins amp Shelley 2002)
MINOR ELEMENTS ANALYTICALRESULTS AND DI SCUSSIONIn the following discussion the aluminum saturation index[ASI calculated as molecular Al2O3(Na2O thorn K2O thornCaO)] of the host-rocks is used to illustrate the range ofelemental variations in the Transhimalayan apatites Forthe studied samples ASI increases in general with the
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
5
silica content of the host-rock regardless of rock type(Table 1) If host-rock silica contents were used instead itwould not make any difference to our discussion or conclu-sions although the scatter of apatite compositions relativeto this index of magma evolution would become greater(see Fig 2 and Electronic Appendix Fig 1)In the subsequent discussion plutons with ASI 09 (57wt SiO2) except adakites are termed granites andthose with ASI 509 generally equivalent to 557wt SiO2 are termed lsquomafic rocksrsquoAccording to this classifica-tion the characteristics of minor element distributions inapatites (Electronic Appendix Table 2) from differenttypes of Transhimalayan igneous intrusions are summar-ized below
Fluorine and chlorineMost of the apatites are fluorapatite typical of igneousapatite with F contents41wt and FCl41 (ElectronicAppendixTable 2) The abundance of F increases and thatof Cl decreases with increasing ASI (Fig 2) Howeverapatites in some S-type granites (eg T1048 and T138)have Cl contents below the detection limits of EPMAthese analyses may have much higher FCl ratios thanthose shown calculated from the actual analyses (seeElectronic Appendix Table 2) Accordingly although Fand Cl abundance andor the FCl ratios of apatite havethe potential to monitor magma evolution the F contentis recommended as the better indicatorThe range of F concentrations analysed in one apatite
grain is up to 033 and the range within one rocksample generally is from 05 to 15 For a rock samplewith a specific ASI value the F variation in its apatites isup to 18
Manganese and ironMn variations in a single apatite separate can be up to01wt Whole-rock ASI values exert a control on theMn content of apatite in all types of Transhimalayan intru-sion (Fig 3) independent of the host-rock Mn abundance(Table 1) In metaluminous (ASI51) host rocks such asmost adakites (I-type) mafic rocks and some I-type gran-ites apatite consistently has 502 MnO In peralumi-nous rocks with ASI 11 both the MnO concentrationsof apatite and the proportions of apatite grains with402 MnO rise with the host-rock ASI values In rockswith 1 ASI511 the MnO abundances of apatite aretransitional Therefore the Mn content of apatite can beused as an indicator of the ASI of magmatic rocksIron concentrations in apatite from all Transhimalayan
plutonic rocks are mostly lower than 02 FeO except insome highly evolved S-type granites (eg T170A up to14 Electronic Appendix Table 2) The host-rock totaliron contents (Table 1) appear to have little effect on theFeO contents in apatite
SulfurSulfur (expressed as SO3wt ) in apatites from peralumi-nous rocks falls with increasing ASI of the host rocks tothe extent that more than half of the data are below theEPMA detection limit (001^0001wt SO3 ElectronicAppendix Table 2) In metaluminous Transhimalayan plu-tons the S contents of apatite cover a wide range (Fig 4)and are irregularly correlated with variation in host-rockASIAs for Mn and Fe S abundances in apatite may be
related to the redox condition of the host magma In an
Fig 2 F contents of apatites from different rock types apatite F vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
thorn biotite thorn plagioclase amphibole thorn opaque mineralstogether with apatite thorn zircon titanite allanite monazite as accessory mineral phases (Table 1)
ANALYTICAL METHODSStandard heavy-liquid and magnetic methods were used toobtain pure apatite separates from 5 kg rock samplesApatite grains with no visible inclusions were hand pickedunder a stereo-microscope and then mounted in epoxydiscs Most of the Transhimalayan apatites in this studyare euhedral to subhedral granular and occasionally hex-agonal columnar in shape Back-scattered electron (BSE)images were taken of some of the apatites (eg Gangdesebatholith samples ET021E ST146A ST147A T036C andT153 S-type granites T006C and T172A) no internalzoning patterns were observed
Electron probe microanalysis (EPMA)Major- and minor-element contents of apatite were deter-mined by electron microprobe at the Institute of EarthSciences (IES) Academia Sinica Taiwan or at GEMOCMacquarie University Australia Up to four spot analyseswere performed on each crystal depending on the grainsize Apatite grains mounted in targets A001 to A006(Table 1) were analysed using a JEOL JXA-8900R electronmicroprobe using a wavelength-dispersive (WDS) methodthat employed TAP PET and LIF crystals with 2 mm spa-tial resolution15 kV beam conditions and 10 nA beam cur-rent For analysis of apatites in the remaining targets atGEMOC we used the methods described by Belousovaet al (2002) an electron beam of 10 mm diameter with anaccelerating voltage of 15 kV and a beam current of 20nA Analytical precision for most elements is better than1 but for F Cl and SO3 precision is around 5
Laser ablation-inductively coupled plasmamass spectrometry (LA-ICPMS)The trace-element contents of apatites were measured inthe LA-ICP-(Q)MS laboratory in the Department ofGeosciences National Taiwan University using an Agilent7500s ICP-MS system and a New Wave LUV213 lasersystem (Jackson 2001) Analytical methods which mainlyfollow those of Norman et al (1996 1998) involved time-resolved analysis (TRA) using a fast peak-hoppingsequence Each spot analysis consisted of c 60 s backgroundacquisition (gas blank) followed by c 60 s sample ablationused a 30 mm diameter laser beamTwenty-seven isotopeswere analysed in this study including 43Ca as the internalstandard (IS) using the concentrations obtained byEPMA Both 86Sr (isotopic abundance 986) and 88Sr(isotopic abundance 8258) were analysed to assessthe effect of linear calibration for variable elementcontents
Samples were measured in short runs bracketed by anal-yses of the external standard (ES) NIST SRM 610 withreference values taken from Norman et al (1996) Aftereach analysis data reduction was carried out usingVersion 40 of the lsquoreal-time on-linersquo GLITTER software(Griffin et al 2008) which allows the user to select a datarange free of effects produced by ablating inclusions Asthere is no matrix-matched and homogeneous apatite-likestandard available NIST SRM 610 NIST SRM 614 andUSGS international standards including BCR-2(G) TB-1(a basaltic glass) and GSE-1 (a doped rock-glass) wereanalysed as secondary standards during the analyticalruns to test the accuracy and precision of the LA-ICP-MSmethod The REE concentrations of apatites are generallybetween those of NIST SRM 610 and BCR-2(G)One hundred analyses of NIST SRM 610 show that the
minimum limit of detection (LOD) of this method formost trace elements in igneous rocks is around the ppmlevel commonly not more than 10 ppm For REE mini-mum LODs are substantially below the ppm level andmostly not more than 2 ppm Comparison of our resultswith those from the literature or the compiled values forstandard materials show that they are mutually consistentwith an accuracy better than 5 relative (ElectronicAppendix Table 1 available for downloading at httpwwwpetrologyoxfordjournalsorg) This agreementimplies that matrix effects are not significant during ourmeasurements when the doped synthetic glass is used asthe external standard for natural sample measurementsAccordingly this method can be applied to the trace ele-ment analysis of apatite For further comparison JC4 anapatite EPMA reference material was analysed as anunknown sample and these results are listed in ElectronicAppendixTable 1Precision for elements with mass480 is inversely corre-
lated with element abundance from 4^5 for 450 ppm(NIST SRM 610) to 7^14 for 08 ppm (NIST SRM614 Electronic AppendixTable 1) Regardless of concentra-tions or matrix the precision for Zn Rb Sn Sb Cs andPb (10^20) is significantly poorer than for the otheranalytes in any sample This lower precision may be dueto the fractionation of these elements relative to the IS(see Gulaquo nther et al 1999) or to heterogeneity in the refer-ence glasses on the scale of the LA-ICP-MS spatial resolu-tion (see Eggins amp Shelley 2002)
MINOR ELEMENTS ANALYTICALRESULTS AND DI SCUSSIONIn the following discussion the aluminum saturation index[ASI calculated as molecular Al2O3(Na2O thorn K2O thornCaO)] of the host-rocks is used to illustrate the range ofelemental variations in the Transhimalayan apatites Forthe studied samples ASI increases in general with the
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
5
silica content of the host-rock regardless of rock type(Table 1) If host-rock silica contents were used instead itwould not make any difference to our discussion or conclu-sions although the scatter of apatite compositions relativeto this index of magma evolution would become greater(see Fig 2 and Electronic Appendix Fig 1)In the subsequent discussion plutons with ASI 09 (57wt SiO2) except adakites are termed granites andthose with ASI 509 generally equivalent to 557wt SiO2 are termed lsquomafic rocksrsquoAccording to this classifica-tion the characteristics of minor element distributions inapatites (Electronic Appendix Table 2) from differenttypes of Transhimalayan igneous intrusions are summar-ized below
Fluorine and chlorineMost of the apatites are fluorapatite typical of igneousapatite with F contents41wt and FCl41 (ElectronicAppendixTable 2) The abundance of F increases and thatof Cl decreases with increasing ASI (Fig 2) Howeverapatites in some S-type granites (eg T1048 and T138)have Cl contents below the detection limits of EPMAthese analyses may have much higher FCl ratios thanthose shown calculated from the actual analyses (seeElectronic Appendix Table 2) Accordingly although Fand Cl abundance andor the FCl ratios of apatite havethe potential to monitor magma evolution the F contentis recommended as the better indicatorThe range of F concentrations analysed in one apatite
grain is up to 033 and the range within one rocksample generally is from 05 to 15 For a rock samplewith a specific ASI value the F variation in its apatites isup to 18
Manganese and ironMn variations in a single apatite separate can be up to01wt Whole-rock ASI values exert a control on theMn content of apatite in all types of Transhimalayan intru-sion (Fig 3) independent of the host-rock Mn abundance(Table 1) In metaluminous (ASI51) host rocks such asmost adakites (I-type) mafic rocks and some I-type gran-ites apatite consistently has 502 MnO In peralumi-nous rocks with ASI 11 both the MnO concentrationsof apatite and the proportions of apatite grains with402 MnO rise with the host-rock ASI values In rockswith 1 ASI511 the MnO abundances of apatite aretransitional Therefore the Mn content of apatite can beused as an indicator of the ASI of magmatic rocksIron concentrations in apatite from all Transhimalayan
plutonic rocks are mostly lower than 02 FeO except insome highly evolved S-type granites (eg T170A up to14 Electronic Appendix Table 2) The host-rock totaliron contents (Table 1) appear to have little effect on theFeO contents in apatite
SulfurSulfur (expressed as SO3wt ) in apatites from peralumi-nous rocks falls with increasing ASI of the host rocks tothe extent that more than half of the data are below theEPMA detection limit (001^0001wt SO3 ElectronicAppendix Table 2) In metaluminous Transhimalayan plu-tons the S contents of apatite cover a wide range (Fig 4)and are irregularly correlated with variation in host-rockASIAs for Mn and Fe S abundances in apatite may be
related to the redox condition of the host magma In an
Fig 2 F contents of apatites from different rock types apatite F vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
silica content of the host-rock regardless of rock type(Table 1) If host-rock silica contents were used instead itwould not make any difference to our discussion or conclu-sions although the scatter of apatite compositions relativeto this index of magma evolution would become greater(see Fig 2 and Electronic Appendix Fig 1)In the subsequent discussion plutons with ASI 09 (57wt SiO2) except adakites are termed granites andthose with ASI 509 generally equivalent to 557wt SiO2 are termed lsquomafic rocksrsquoAccording to this classifica-tion the characteristics of minor element distributions inapatites (Electronic Appendix Table 2) from differenttypes of Transhimalayan igneous intrusions are summar-ized below
Fluorine and chlorineMost of the apatites are fluorapatite typical of igneousapatite with F contents41wt and FCl41 (ElectronicAppendixTable 2) The abundance of F increases and thatof Cl decreases with increasing ASI (Fig 2) Howeverapatites in some S-type granites (eg T1048 and T138)have Cl contents below the detection limits of EPMAthese analyses may have much higher FCl ratios thanthose shown calculated from the actual analyses (seeElectronic Appendix Table 2) Accordingly although Fand Cl abundance andor the FCl ratios of apatite havethe potential to monitor magma evolution the F contentis recommended as the better indicatorThe range of F concentrations analysed in one apatite
grain is up to 033 and the range within one rocksample generally is from 05 to 15 For a rock samplewith a specific ASI value the F variation in its apatites isup to 18
Manganese and ironMn variations in a single apatite separate can be up to01wt Whole-rock ASI values exert a control on theMn content of apatite in all types of Transhimalayan intru-sion (Fig 3) independent of the host-rock Mn abundance(Table 1) In metaluminous (ASI51) host rocks such asmost adakites (I-type) mafic rocks and some I-type gran-ites apatite consistently has 502 MnO In peralumi-nous rocks with ASI 11 both the MnO concentrationsof apatite and the proportions of apatite grains with402 MnO rise with the host-rock ASI values In rockswith 1 ASI511 the MnO abundances of apatite aretransitional Therefore the Mn content of apatite can beused as an indicator of the ASI of magmatic rocksIron concentrations in apatite from all Transhimalayan
plutonic rocks are mostly lower than 02 FeO except insome highly evolved S-type granites (eg T170A up to14 Electronic Appendix Table 2) The host-rock totaliron contents (Table 1) appear to have little effect on theFeO contents in apatite
SulfurSulfur (expressed as SO3wt ) in apatites from peralumi-nous rocks falls with increasing ASI of the host rocks tothe extent that more than half of the data are below theEPMA detection limit (001^0001wt SO3 ElectronicAppendix Table 2) In metaluminous Transhimalayan plu-tons the S contents of apatite cover a wide range (Fig 4)and are irregularly correlated with variation in host-rockASIAs for Mn and Fe S abundances in apatite may be
related to the redox condition of the host magma In an
Fig 2 F contents of apatites from different rock types apatite F vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
6
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Table 2 Summary of the geochemical characteristics of apatite in different types and aluminosities of Transhimalayan
granitoids
Metaluminous Peraluminous
ASI 509 09ndash10 10ndash11 4 11
SiO2 (wt ) 557 57ndash65 465 465
Geochemical abundances of apatite
F () 10ndash30 13ndash33 18ndash33 20ndash36
mostly 15ndash25 mostly 18ndash27 mostly 20ndash30 mostly 23ndash33
MnO () bdlndash019 0023ndash017 003ndash028 ( 075) 004ndash335
Sr (ppm) 341ndash747 266ndash490 69ndash472 ( 1850) 32ndash202
adakite adakite adakite
309ndash353 326ndash441 601ndash645
adakitic Gangdesey adakitic Gangdesez
306ndash341 ( 732) 351ndash553
HREE
Yb (ppm) 45ndash745 21ndash148 27ndash252 88ndash543
adakite adakite adakite
42ndash13 56ndash14 29ndash44
adakitic Gangdesey adakitic Gangdesez
23ndash68 49ndash151
Y (ppm) 91ndash1078 236ndash1788 451ndash3129 1056ndash5817
adakite adakite adakite
51ndash200 96ndash200 435ndash629
adakitic Gangdesey adakitic Gangdesez
314ndash1113 496ndash1779
HREE thorn Y 157ndash1728 389ndash2912 705ndash4903 1838ndash9507
(ppm) adakite adakite adakite
103ndash406 156ndash356 788ndash1117
adakitic Gangdesey adakitic Gangdesez
536ndash1950 719ndash2709
REE pattern steep right-inclined right-inclined with (transition from) flat with strong
with or without Eu(ndash) right-inclined to Eu(ndash) and Nd(ndash)
Eu(ndash) flat with Eu(ndash)
adakite adakite adakite
steep right-inclined Steep right-inclined gentle right-inclined
with Eu(ndash) with Eu(ndash) with Eu(ndash)
adakitic Gangdesey adakitic Gangdesez
right-inclined with left-inclined
Eu(ndash)
(LaNd)N 01ndash62 13ndash37 04ndash23 04ndash12
adakite adakite adakite
28ndash87 23ndash51 037ndash063
adakitic Gangdesey adakitic Gangdesez
111ndash177 018ndash091
(LaYb)N 02ndash118 50ndash338 04ndash312 038ndash22
adakite adakite adakite
79ndash371 37ndash124 27ndash53
adakitic Gangdesey adakitic Gangdesez
98ndash36 006ndash049
Eu anomaly 010ndash105 014ndash080 004ndash052 001ndash042
adakite adakite adakite
015ndash059 043ndash068 032ndash036
adakitic Gangdesey adakitic Gangdesez
033ndash053 071ndash165
Nd anomaly 097ndash119 088ndash110 090ndash113 073ndash100
adakite adakite adakite
101ndash111 082ndash122 110ndash115
adakitic Gangdesey adakitic Gangdesez
102ndash110 087ndash125
Eu(ndash) negative Eu anomaly in REE pattern Nd(ndash) negative Nd anomaly in REE pattern bdl below detection limitData from sample T016yData from sample T027zData from sample T024
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
7
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
oxidizing magma more S occurs as SO42 and thus can
easily replace PO43 in apatite (see Sha amp Chappell
1999) However no correlation between any two of SO3MnO and FeO is observed in theTranshimalayan apatites
SodiumApatites from different Transhimalayan rocks generallyhave similar Na contents (mostly Na2O502) howeverNa increases slightly with magma fractionation in the(peraluminous) S-type granites The variations in thesodium contents of apatite are not as clear and systematic
as those reported from the granites of the Lachlan FoldBelt by Sha amp Chappell (1999)
TRACE ELEMENTS RESULTSAND DISCUSS IONStrontiumSr concentrations in apatite grains separated from a singlerock are variable but mainly confined to the range of 025 SrapatiteSrhost-rock 1 (Fig 5a Electronic Appendix
Fig 3 MnO contents of apatites from different rock types apatite MnO vs host-rock ASI
Fig 4 SO3 contents of apatites vs host-rock ASI for different rock types
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
8
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Table 3) The Sr content of apatites from S-type graniteshave a restricted range with variations less than 100 ppm(mostly550 ppm) In apatites from I-type intrusions andadakites the range of Sr contents may be up to 330 ppm(mostly 20^150 ppm) However a few analyses extend tomore than 400 ppm (T027 and ET025C) and these Srvalues are higher than those of the host-rocksThe Sr contents of apatites from theTranshimalayan plu-
tonic rocks are generally correlated with the Sr in thehost-rock (Fig 5a) as observed in other studies (eg Shaamp Chappell 1999) For example adakites and (I-type)Gangdese adakitic rocks commonly have higher Sr concen-trations than other Transhimalayan rocks with the sameSiO2 contents or ASI a feature ascribed to the rarity ofresidual plagioclase in their sources (Fig 5b Chung et al2003) The Sr contents of apatites from these rocks are
comparable with those from (I-type) mafic rocks theunderplated equivalents of these mafic rocks have beensuggested as a likely source of adakitic magmas (Chu2006 Wen et al 2008a) However apatites from I-typegranite (ET026I) have high Sr abundances similar tothose in adakitic rocks (Fig 5b) and have much higher Srcontents than their host-rocks All of these examples arediscussed in more detail belowBecause Sr abundances major element concentrations
and the levels of some other trace elements in igneousrocks of a single suite vary with the degree of fractionation(eg whole-rock SiO2) and ASI the Sr contents of apatitesmight be used as a monitor of whole-rock compositionBelousova et al (2001) showed that the Sr contents of apa-tites from granites of four different (super-) suites in theMt Isa Inlier Australia exhibit a systematic variation
Fig 5 Variation of Sr content of apatite as a function of (a) Sr and (b) SiO2 content of the host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
9
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
with whole-rock SiO2 Al2O3 FeO K2O and RbSrHowever in the I-type Gangdese batholith it is difficult tosee such correlations for example between the Sr contentsof apatite and the K2O content of their host-rocks Thisprobably is because in the Gangdese belt the K2O con-tents of the intrusive rocks show no correlation with theirSr and SiO2 contents (Table 1) Thus the potential of Sr inapatite to predict whole-rock composition essentiallydepends on how well whole-rock SiO2 correlates withother element abundances This application of apatite Srcontents is feasible in principle but must be independentlydefined for each suite
Thorium and uraniumThe behavior of Th and U in apatite is independent of thecomposition and type of the host-rocks (Fig 6) In the
metaluminous granites and adakites the Th contents ofapatites vary mainly from 20 to 200 ppm similar to therange in the (metaluminous) mafic rocks In the peralumi-nous rocks apatite can have lowerTh concentrations com-monly 2^50 ppm The depletion of Th in these apatitescan be attributed to the crystallization of monazite (CeLa Th Nd Y)PO4 before andor concurrently with apa-tite as monazite is one of the main accessory phases in per-aluminous magmas but not in metaluminous ones Thereis no distinguishable difference in U content among apa-tites in adakites I-type rocks and S-type granites with dif-ferent aluminosities most values are between 10 and100 ppm The small range of ThU (05^6) of apatites inmetaluminous granitoids reflects these consistent Th andU contents (Fig 6b) In peraluminous granites apatiteThU ratios span a wider range from 001 to 20 because
Fig 6 Variation of (a) Th and (b) ThU in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
10
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
of the variable Th contents These observations for thegranites are inconsistent with those made in the LachlanFold Belt (Sha amp Chappell 1999)
Rare earth elements (REE)REE concentrations
Awhole-rock ASI between 10 and 11 also marks a bound-ary in the behavior of the REE in apatite (Fig 7)Regardless of rock type apatites from metaluminousTranshimalayan rocks have lower contents of the HREE(Gd to Lu plus Y 100^2000 ppm) than apatites from pera-luminous granites (400^10 000 ppm) although bothgroups of apatite have similar light REE (LREE) contents(La to Eu mostly 1000^10 000 ppm)The around 10-times enrichment of HREE in apatites
from peraluminous rocks may reflect (1) relatively fewerHREE-rich accessory minerals competing for the HREEin the evolved melts when apatite crystallized or (2) a pro-nounced increase in partition coefficients (D) for theHREE relative to LREE The measured REE partitioncoefficients of apatite in granites are highly variable(Electronic Appendix Table 4) and possibility (2) cannotbe properly evaluated before there is a better understand-ing of these HREE valuesWith regard to the first possibil-ity xenotime zircon and amphibole are known toconcentrate the HREE (Bea1996) Xenotime is not a ubiq-uitous phase in Transhimalayan peraluminous magmasZircon can be found in both peraluminous and metalumi-nousTranshimalayan rocks However zircons in peralumi-nous S-type granites are mainly inherited and the lessercompetition from zircon could be the reason for thehigh HREE in apatite In the I-type peraluminous gran-ites apatite can capture more of the HREE budgetbecause of the lesser crystallization of amphibole which iscommon in metaluminous magmas but not in peralumi-nous onesIn apatite from Transhimalayan intrusions the differ-
ences in the abundances of single REE show similartrends to the total contents of LREE and HREE(including Y) In this study apatite in general contains1000^20 000 ppm REE (including Y) The total LREEabundances of Transhimalayan apatites are essentially con-stant over a range of whole-rock aluminosity The some-what higher HREE contents of apatites in peraluminousgranites lead to somewhat higher total REE contents
REE patterns
Transhimalayan apatites show significant variations inchondrite-normalized REE patterns (Figs 8^11) betweendifferent rock types In a single apatite the within-grainvariation in absolute REE abundances can be large (egLa up to 1000 ppm Electronic Appendix Table 3) butthere is no significant difference in the shape of the REEpatterns
Apatites from the metaluminous rocks including ada-kites are characterized by LREE-enriched patterns withlittle or no Eu depletion (Figs 8a^c and 9a^k) Theirenrichment in LREE relative to HREE and Eu in generalfalls with increasing ASI accompanying the fractionationof the host magma (Figs 12 and 13a) In the I-typeGangdese magmatic suite the apatites with the most mark-edly LREE-enriched patterns and the least Eu depletionare all from the least fractionated Gangdese mafic rocks(Fig 9a c^e) except for one special case granite ET026I(Fig 9l) However the most pronounced enrichment ofLREE relative to HREE can be observed in apatites fromadakites (ASI409) although these commonly have a sig-nificant negative Eu anomaly (Fig 8a^c) the relativeLREE enrichments reflect the low HREE contents of themagma from which the apatites crystallized The composi-tions of the source rocks of the magmas are thus an impor-tant control on the apatite REE patternsIn contrast to apatites in metaluminous rocks the REE
patterns of those in the peraluminous granites especiallythose with ASI 411 (Figs 8f 9o^q 10b and 11b^f) areessentially flat or upwardly convex in linear-scaleplots and show Eu and Nd depletion they are similar tothe patterns of apatite in granites from the LachlanFold Belt and south China (Sha amp Chappell 1999 Hsiehet al 2008) The relatively low LREE abundances ofthese apatites probably reflect the crystallization of acces-sory phases including allanite (Electronic AppendixTable 5a) and particularly monazite (ElectronicAppendixTable 5b)Apatite in the only peraluminous (post-collisional) ada-
kite (T016) is mildly LREE-depleted (Fig 8f) In factthese patterns show lower LREE and higher HREE thanthose of apatite in metaluminous (post-collisional) ada-kites which is consistent with the difference between apa-tites in metaluminous and peraluminous lsquocommonrsquogranitoids Variations of apatite HREE in adakites andadakitic Gangdese rocks essentially follow the trenddefined by I- and S-type plutonic rocks but their abun-dances are much lower owing to the HREE depletion inthe magmas (Fig 7) Such LREE-enriched patterns repre-sent lsquocommonrsquo flat REE patterns modified by the HREE-depleted magma compositionApatites in peraluminous I-type Gangdese suite rocks
and adakites both with ASI of 1^11 show variable REEpatterns (Figs 8d and e 9m and n 10a and 11a) which canbe LREE-enriched flat andor transitional sometimeswith moderate Nd and Eu depletion like those of T148A(Fig 9n) and ST146A (Fig 9o) However those in theS-type granites show coherent flat patterns with pro-nounced negative Nd and Eu anomalies (Fig 11a)LREE-depleted patterns are also observed in
Transhimalayan apatites Some apatites in the titanite-bearing I-type mafic rock ST147A show La^Nd depletion
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
11
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Fig 7 Variation of (a) total REE andY (b) total LREE and (c) total HREE andYcontents in apatites vs host-rock ASI
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
12
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
patterns with a Eu negative anomaly (Fig 9f) The concen-tration of La can be down by a factor of 10 The three apa-tites with the most depleted LREE patterns have slightlylower Sr contents in comparison with the other apatites
(around 15^90 ppm difference) The crystallization oftitanite can cause these compositional variations witha strong depletion in LREE and mild depletion in Srin coexisting apatite A similar pattern is also observed
Fig 8 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks Transhimalayan adakites (lightgrey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference Chondrite REE values fromTaylor amp McLennan (1985)
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
13
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
in apatites in another titanite-bearing I-type mafic rockET021E (Fig 9i)LREE-depleted patterns with depletion extending as
far as Gd are observed in apatites from strongly
fractionated granites (eg T024 Fig 10b) Such patternshave been reported from apatite in Norwegian granite peg-matites and other granitoids and may be more commonin highly evolved granites including pegmatites
Fig 9 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan I-type Gangdeserocks (light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
14
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
(Belousova et al 2002) The much stronger depletion ofthese apatites in LREE results from competition with theabundant LREE-enriched minerals in T024 includingallanite epidote and feldspar (mainly K-feldspar) (Wenet al 2008a) Unfortunately in this study we have notobserved any examples of transitions between flat and
LREE-depleted patterns Whether there are evolutionaryrelationships between them remains uncertainA negative Nd anomaly in REE patterns is typical of
apatites with flat REE patterns from granites of theLachlan Fold Belt (Sha amp Chappell 1999) and theNanling Mountains south China (Hsieh et al 2008) and
Fig 9 Continued
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
15
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
is also seen in apatites from the peraluminousTranshimalayan granites Apatites from T148A andST146A show a transition between LREE-enriched andflat REE patterns Nd depletion is not present in the flatpatterns The Nd anomaly is particularly conspicuous inapatites crystallized from S-type and evolved I-typemagmas (Figs 9p and q and 11)
Mainly following the crystallization of feldspar andother minerals in I-type and S-type intrusions the Euanomaly in apatite in general becomes more negative asthe aluminosity of the host-rock increases (Fig 13a) andmore specifically as its LREE enrichment decreases(Fig 13b) However there are two main exceptions T024and ST146 In the former because apatite has a
Fig 9 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
16
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
comparable absolute Eu abundance its strong LREEdepletion reduces the apparent negative Eu lsquoanomalyrsquo(Fig 10b) The positive Eu anomaly of apatite in ST146provides specific petrogenetic information (discussedbelow) Moreover the common relationships between Eudepletion and LREE enrichment and host-rock aluminos-ity are not observed in the adakites (Fig 13)
DISCUSSIONNd depletion in apatite of peraluminousmagmasSha amp Chappell (1999) showed that Nd depletion inapatite from peraluminous intrusions does not reflect the
bulk-rock composition and argued that it reflects the crys-tallization of monazite Monazite is a common accessorymineral in peraluminous but not metaluminous magmasand its REE partition coefficients peak at Nd (Yurimotoet al 1990 Electronic Appendix Table 5b) However thisinterpretation may be inadequateAs in the granites of the Lachlan Fold Belt (Sha amp
Chappell 1999) there seems to be little Nd depletion inthe host-rock REE patterns of the peraluminousTranshimalayan rocks (Figs 8^11) although their apatitesall have negative Nd anomalies However because of apa-titersquos high partition coefficient for REE it has the abilityto lsquoamplifyrsquo a subtle negative Nd anomaly present in itshost-rock and make it visible when plotted on a
Fig 10 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocks I-type adakitic Gangdese rocks(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
17
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
logarithmic scale There is a simple method to checkwhether the Nd depletion in apatite can be inherited fromits host magma or is due to competition with monaziteThe possibility of control by the host magma compositionis negated if after normalization to the bulk-rock the
REE patterns of apatites still show significant Nd deple-tion However if the bulk-rock-normalized apatite REEpatterns show only a small Nd depletion or none thealternative explanation is confirmed that is that the Nddepletion in the apatite REE patterns originates from the
Fig 11 Chondrite-normalized REE patterns of apatites (black lines with open symbols) and their host-rocksTranshimalayan S-type granites(light grey lines with filled symbols) ASI and silica content of the bulk-rock are given in parentheses for reference
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
18
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
geochemistry of the host magma The latter case is clearlydemonstrated by our data (Fig 14)Monazite strongly favors Th over U and its crystalliza-
tion will result in a marked decrease in Th contents andThU in the residual magma This should produce a posi-tive correlation between the depletion of Nd in the REEpatterns and theThU of apatite However this correlationis not obvious in apatites from (peraluminous) S-typegranites The main accessory minerals in peraluminousmagmas are apatite thorn zircon allanite monazite titanite xenotime Of these zircon is the only commonliquidus phase in magmas that has a much higher prefer-ence for U than for Th (Electronic Appendix Table 5d)Fractionation of zircon could compensate for the variationof ThU in the magma produced by monazite crystalliza-tion and thus cause the observed lack of correlationbetween ThU and Nd anomalies in apatites HoweverU^Pb dating shows that zircon in the Transhimalayan S-type granites is mostly inherited and thus that zirconrarely crystallized from the S-type magmas (see Chuet al 2006) Xenotime could be another potential candi-date However it is not a ubiquitous mineral phase andthe poor understanding of its mineral^melt partition coef-ficients and the large ranges of Th and U in xenotime(Folaquo rster 1998) make it difficult to assess the effects of itscrystallizationThese inconsistencies might be explained if the apatites
in the S-type granites like their zircons are inheritedhowever the origin of the negative Nd anomalies is stilldifficult to explain For fractionation of a mineral to pro-duce a negative Nd anomaly in apatite its Nd partitioncoefficient would need to be distinctly greater than thosefor Pr and Sm However partition coefficients of REE in
monazite and other minerals typically vary as a smoothfunction of their continuously varying ionic radius Thusthe strong fractionation of monazite with the highest par-tition coefficients around Nd can result in a trough-likeREE pattern around Nd similar to examples shown byYurimoto et al (1990) but not a distinct kink at Nd asobserved in the REE patterns of apatiteIndeed apatites with Nd depletion in chondrite-
normalized REE patterns also tend to show a broad nega-tive anomaly around Ho and Er (eg Figs 9p and q and11) both are essential features of the so-called lsquotetradeffectrsquo (see Jahn et al 2001) The tetrad effect reflects non-CHARAC (CHarge-And-RAdius-Controlled) geochemis-try specifically in highly evolved magmas (Bau 1996)where the behavior of high field strength elements(HFSE) including the REE is controlled by both electronconfiguration and complexing ligand types and cannotbe modeled by partition coefficients that vary as smoothfunctions of ionic radius This phenomenon originatesfrom the electronic repulsion attributed to the formationof various ligands between HFSE and volatiles such asH2O Li B F andor Cl which are enriched in themagma As a result of the tetrad effect the bulk-rock andits mineral phases including apatite and monazite (Irber1999) have REE patterns comprising four convex subpat-terns in the subdivisions of La to Nd Pm to Gd Gd toHo and Er to Lu The tetrad effect and other non-CHARAC features such as non-chondritic ratios of YHoand ZrHf are mostly recognized in highly differentiatedgranites with strong hydrothermal interaction such assome of the Transhimalayan S-type and highly evolvedI-type granites
Fig 12 Correlation between chondrite-normalized LaYb ratios in apatites and ASI of their host-rocks
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
19
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Effects of major minerals in maficmagmasCrystallization of major minerals can affect both the REEabundances and patterns of apatite this effect is particu-larly pronounced in some mafic rocks of the I-typeGangdese magmatic suite for example T044E (Fig 9a)T036C (Fig 9c) and ST147A (Fig 9f) Apatites in theserocks show large variations in REEWe suggest that apatitecrystallized earlier in these mafic magmas tends to show(1) the most marked LREE enrichment with high (LaNd)N (eg 44) (2) the least Eu depletion and (3) Sr
contents similar to those of the bulk-rock Such stronglyLREE-enriched apatite always shows no (or weak) nega-tive Eu anomaly This corresponds to the least amount offractionation of feldspars from the host melt and accord-ingly the least magmatic differentiationThe crystallization of LREE-rich andor Eu- and Sr-rich
minerals such as titanite and plagioclase before or togetherwith apatite can significantly reduce the LREE enrich-ment in apatite and produce both negative anomalies inEu and lower Sr abundances Titanite (CaTiSiO5) is onecandidate because (1) its REE partition coefficients are
Fig 13 Apatite Eu anomalies (EuEu) vs (a) host-rock ASI and (b) apatite chondrite-normalized LaYb ratios EuN frac14 (12)SmN thorn (12)GdN
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
20
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
similar to those of apatite (Electronic AppendixTable 5c)and it concentrates LREE and (2) it can precipitate fromoxidized mafic magmas although it usually crystallizeslater than apatite in the Lachlan Fold Belt granites (Shaamp Chappell1999) Once crystallization of titanite becomesimportant in the magma the LREE enrichment of apa-tites crystallized later or synchronously becomes smallerIn metaluminous Gangdese rocks as the slopes of the
LREE patterns of the apatites become less steep their neg-ative Eu anomalies generally deepen especially in themore mafic rocks (Fig 15) This effect can be produced bycrystallization of plagioclase concurrent with or beforeapatite In mafic magmas only the crystallization of plagi-oclase can deplete Eu by selectively extracting Eu2thorn (seeTable 1) and plagioclase is the major rock-forming mineralin these rocks that prefers the LREE to the HREE (egDunn amp Sen 1994 DLa frac14 0075^018 DCe frac14 0062^014DEu frac14 048^079 DGd frac14 0016^0067 and DYb frac14 0004^0014 for plagioclase in basaltic to andesitic arc magmas)although its partition coefficients (D) for the LREE are
less than unity Amphibole and pyroxene can also fraction-ate the REE but these are not major phasesWe assume that only apatite and plagioclase compete for
the REE in the mafic magmas as apatite is the onlycommon early crystallized accessory phase that can con-centrate the REE (Hoskin et al 2000) In the literatureapatite partition coefficients for basic and intermediatemagmas vary widely (eg DLafrac14 25^282 ElectronicAppendix Table 4) The ability of apatite to concentrateLa for example can be 15^380 times that of plagioclaseNevertheless the modal abundance of plagioclase typicallyis a few hundred times that of apatite in mafic intrusiverocksMass-balance considerations suggest that crystallization
of both accessory minerals and rock-forming minerals cancause subtle changes in apatite REE patterns within asingle magma chamber Plagioclase also preferentially con-centrates Sr Although apatite has a Sr partition coefficientof more than unity (11 to 8 Electronic AppendixTable 4) comparable with that of feldspar the Sr content
Fig 14 Host rock-normalized REE patterns of apatites (black lines with open symbols) and chondrite-normalized REE patterns of their host-rocks (light grey lines with filled symbols) that is some examples of S-type granites and evolved I-type peraluminous granites
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
21
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
of apatite is mostly lower than that of its host magma aftercompetition with abundant plagioclase (Fig 5a)Apatites that crystallized earlier than plagioclase
have lower HREE contents (down to510 times the con-tents of the bulk-rock) and upward-concave REE patternsBecause the highest partition coefficients for apatite arearound Sm Gd or Tb in mafic magmas (ElectronicAppendixTable 4) this can be attributed to the fractiona-tion of pyroxene and amphibole which are HREE-enriched In summary apatite in these mafic magmasstarted to crystallize later than pyroxene andor amphi-bole but together with or a little earlier than plagioclaseThis demonstrates that the REE and Sr contents of apatitecan be used as tracers for magmatic processes and apatitehas the ability to retain information on the geochemicalevolution of the host magma
Indicators of petrogenesisThe Sr contents REE patterns and Eu anomalies of apa-tite are related to the geochemical composition of its hostmagma and thus they have high potential as indicators ofmagma mixing andor compositional heterogeneity in themagma source region The peculiar behavior of these indi-cators in apatites from ET025C (a metaluminous adakiteFig 8c) ET026I (an isaluminous I-type granite Fig 9l)ST146A (a peraluminous I-type granite Fig 9o) andT027 (an I-type peraluminous adakitic granite Fig 10a)provides examples These apatites share some of thefollowing characteristics (1) a positive Eu anomaly incon-sistent with that of the host-rock (2) higher Sr abundances
than their host-rocks or widely variable Sr abundances(3) REE patterns inconsistent with the host-rockaluminosityApatite REE patterns commonly exhibit negative (or
no) Eu anomalies reflecting both the redox state of thehost melt and particularly the fractionation of feldspar(alkali felspar and plagioclase) and of epidote to somedegree (Bea 1996) Positive Eu anomalies in apatitescannot be produced by fractionation processes but reflecta magma source with a positive Eu anomaly In some apa-tites in sample ST146A for example the Eu peaks in theREE patterns show that these apatites crystallized in amagma with a positive Eu anomaly but the bulk magmaof ST146 shows mild Eu depletion (Fig 9o) The inconsis-tent positive Eu anomalies in the apatites strongly suggestthat at least two magmas or magma sources were involvedOne from which most of the apatites with negative Euanomalies crystallized was peraluminous with a composi-tion more like the bulk-rock of ST146 The other has anASI less than 11 with a positive Eu anomaly suggestingthat the source of this magma essentially consisted of feld-spar cumulatesThe Sr contents of apatites are analogues of the magma
from which they crystallized Commonly apatite containsless Sr than its host-rock and shows limited ranges in Srconcentration (less than 200 ppm Fig 5a) Some of theapatite grains in T027 ET026I and ET025C have muchhigher Sr contents than the values expected from theirwhole-rock ASI or Sr contents and they also show arange in Sr abundance of4400 ppm (Fig 5a) During theformation of these three samples magmas with much
Fig 15 Correlation between Eu anomaly and chondrite-normalized LaNd ratios representing the degree of LREE enrichment of apatites inI-type Gangdese mafic rocks and ET026I an isaluminous I-type granite
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
22
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
higher Sr contents than the final bulk-rock must haveexisted in the magma chamber or the magma sourceBecause the Sr abundances of magmas decrease with frac-tional crystallization such magmas were probably moremafic than the final bulk-rock composition (Fig 5b)Values of (SrapatiteSrbulk-rock) 41 are therefore suggestedas an indicator that a more mafic magma was involved inmagma genesis The observed correlations between apatiteSr contents and some host-rock major and trace elements(eg Fig 5b) make it possible to identify the geochemistryof the possible end-member magmaZircons inT027 and ET026I were analysed for their Hf
isotopic compositions the results display pronounced vari-ation in and correlation between trace-element abun-dances and ratios [eg 176Yb177Hf (001^007)] and Hfisotope compositions which vary over 10 epsilon-units(Chu 2006 see Electronic Appendix Fig 2) These data
suggest that at least two magmas with different isotopicand trace-element signatures mixed during the formationof samples T027 and ET026IAlthough sample ET026I has a whole-rock ASI frac1410 its
apatite has an inconsistent steep LREE-enriched REEpattern with no Eu anomaly similar to those in maficrocks such as T044E ST141A and T036C (Fig 9a c ande) These apatites also have higher Sr concentrations thantheir host-rock The geochemical characteristics of theET026I apatites therefore suggest that they originated in amuch less evolved magma Either magma mixing orsource heterogeneity can be responsible for these inconsis-tent apatite REE patternsApatite in ET025C a metaluminous adakite would be
expected to have a strongly LREE-enriched REE patternbased on the general trend of REE variation vs ASIHowever ET025C contains some apatite grains with flat
Fig 16 Minor and trace element discrimination plots for apatite from Transhimalayan intrusive rocks with variable aluminosityCorresponding aluminosity of the granitoids is shown by numbers
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
23
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
REE patterns (Fig 8c) and no transitional patterns areobserved These apatites may be (1) inherited from themagma source (2) crystallized from an evolved adakiticmagma after differentiation or (3) derived through wall-rock contamination by the (peraluminous) crust of theLhasa terrane If a peraluminous adakitic magma hadevolved through magma mixing or fractional crystalliza-tion the newly forming apatites should have characteris-tics similar to those inT016 with mild LREE enrichmentIn addition although the source rocks of the adakitemagma may contain apatite with flat REE patterns (Beaamp Montero 1999) high Sr contents in apatite are alsoexpected This is not the case in ET025C however asmall degree of wall-rock contamination which would notsignificantly modify the bulk-magma REE pattern canreasonably explain the presence of two types of apatiteREE patterns
Provenance discriminationIn the Transhimalayan intrusive rocks apatite geochemis-try is more strongly controlled by whole-rock aluminosity(ASI) than by silica content Systematic variations in theminor- and trace-element composition of apatite relativeto host-rock ASI are observed especially in F Mn Sr andREE contents and chondrite-normalized REE patternsincluding the LREEHREE ratio Eu anomaly and Ndanomaly (Table 2) The behaviour of Fand Mn in apatitesis associated with magmatic differentiation or aluminosityand independent of rock type whereas that of Sr andREE varies with both aluminosity and rock typeMn Sr and REE were suggested as discriminants based
on the statistical analysis approach of Belousova et al(2002) although adakites were not included in their data-base The abundances of these elements and F inTranshimalayan apatites show variations comparable with
Fig 16 Continued
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
24
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
those observed by Sha amp Chappell (1999) These studiesalso demonstrate the potential of the abundances of FMn Sr and REE as provenance indicators in detrital apa-tite After rejecting those apatite data that are not repre-sentative of their host-rock composition the geochemicaldata for apatite in Transhimalayan rocks have been usedto generate discrimination plots distinguishing apatitesfrom I-type mafic rocks I-type granites and S-type gran-ites (Fig 16) Where apatite compositions in these rocktypes plot far from the recommended fields in Fig 16magma mixing or source heterogeneity may be involvedand this emphasizes the need to evaluate the data by popu-lation single grains may give misleading resultsApatites in peraluminous I-type and S-type rocks par-
ticularly those with ASI411 have similar geochemistryin terms of major minor and trace elements and typicallyflat REE patterns This makes it difficult to distinguishbetween peraluminous I-type and S-type rocks using apa-tite geochemistry If in situ Sr or Nd isotope compositionsof apatites can be measured better constraints on the pet-rogenesis might be obtained
CONCLUDING REMARKSThe aluminosity (ASI) of the host magma is a critical con-trol on the geochemical behavior of apatite inTranshimalayan plutonic rocks This is consistent with theobservations of Bea (1996) Sha amp Chappell (1999) andHsieh et al (2008) Abundances of trace elements inTranshimalayan apatites vary not only with the aluminos-ity of the host magma but also more specifically withrock type The composition of apatite shows significantvariation between metaluminous and peraluminousmagmas and between rock types particularly in terms ofF Mn Sr and REE contents demonstrating the potentialof apatite geochemistry as a petrogenetic tracerThe behavior of minor elements including Fand Mn in
Transhimalayan apatites is independent of the hostmagma type that is I-type intrusions S-type granitesand adakites F contents in apatite increase with host-rockaluminosity Mn contents rise with increasing aluminosityof the peraluminous host-rock Accordingly F and Mncontents of apatites carry specific information aboutmagma aluminosity or fractionationSr concentrations in apatites mimic those of the host
magmas and in general are lower than those of hostmagma with restricted ranges of 5200 ppm Apatite Srcontents can be used to discriminate between host-rocktypes the Sr content of apatite in adakites is substantiallyhigher for a given host-rock aluminosity or a given Fabun-dance in the apatite itself and the value is comparablewith that in I-type mafic rocks Significantly larger rangesin Sr content and apatite Sr contents higher than theirhost-rock may indicate that these apatites crystallized in
magmas of different trace-element composition from thebulk-rock and signal the effects of magma mixingThe REE patterns of apatite show potential in petroge-
netic applications The patterns normally vary fromstrongly LREE-enriched to flat reflecting the crystalliza-tion of competing rock-forming and accessory mineralsproducing depletion in Eu and even Nd They also can bemodified by specific magma compositions for instancethe HREE-depleted characteristics of apatites from ada-kites When apatite shows REE patterns andor positiveEu anomalies that are inconsistent with the host-rock alu-minosity and Eu depletion this implies the effects ofmagma mixing crustal contamination or source heteroge-neity in the generation of the magma
ACKNOWLEDGEMENTSThis study benefited from financial support by theNational Science Council Taiwan We thank J Q Ji(School of Earth and Space Sciences Peking University)H-Y Lee (National Taiwan University) Q Qian(Institute of Geology and Geophysics Chinese Academyof Sciences) D-RWen (National Taiwan University) andQ Zhang (Institute of Geology and Geophysics ChineseAcademy of Sciences) for help with fieldwork FernandoBea Julian A Pearce and two anonymous reviewers arethanked for their positive and constructive commentsAnalytical data were obtained at GEMOC using instru-mentation funded by ARC LIEF and DEST SystemicInfrastructure Grants and Macquarie University Thisis contribution 609 from the ARC National Key Centrefor the Geochemical Evolution and Metallogeny ofContinents (wwwesmqeduauGEMOC)
SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online
REFERENCESAlle gre C J CourtillotV Tapponnier P et al (1984) Structure and
evolution of the Himalaya^Tibet orogenic belt Nature 307 17^22Bau M (1996) Controls on the fractionation of isovalent trace ele-
ments in magmatic and aqueous systems evidence fromYHo ZrHf and lanthanide tetrad effect Contributions to Mineralogy and
Petrology 123 323^333Bea F (1996) Residence of REEYTh and U in granites and crustal
ptotoliths implications for the chemistry of crustal melts Journalof Petrology 37 521^552
Bea F amp Montero P (1999) Behavior of accessory phases and redis-tribution of Zr REEYTh and U during metamorphism and par-tial melting of metapelites in the lower crust an example from theKinzigite Formation of Ivrea^Verbano NW Italy Geochimica et
Cosmochimica Acta 63 1133^1153Belousova E A Walters S Griffin W L amp OrsquoReilly S Y (2001)
Trace-element signatures of apatites in granitoids from the Mt Isa
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
25
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Inlier northwestern Queensland AustralianJournal of Earth Sciences48 603^619
Belousova E A GriffinW L OrsquoReilly S Y amp Fisher N I (2002)Apatites as an indicator mineral for mineral exploration trace-ele-ment compositions and their relationship to host rock typeJournal of Geochemical Exploration 76 45^69
Belousova E A GriffinW L amp OrsquoReilly S Y (2006) Zircon mor-phology trace element signatures and Hf-isotope composition as atool for petrogenetic modeling examples from Eastern Australiangranitoids Journal of Petrology 47 329^353
Bizzarro M Simonetti A Stevenson R K amp Kurszlaukis S(2003) In situ 87Sr86Sr investigation of igneous apatites and carbo-nates using laser-ablation MC-ICP-MS Geochimica et Cosmochimica
Acta 67 289^302Chu M-F (2006) Application of ICP-MS to the study of
Transhimalayan petrogenesis PhD Thesis National TaiwanUniversityTaipei
Chu M-F Chung S-L Song B Liu D OrsquoReilly S YPearson N J Ji J amp Wen D-J (2006) Zircon U^Pb and Hf iso-tope constraints on the Mesozoic tectonics and crustal evolution ofsouthernTibet Geology 34 745^748
Chung S-L Liu D Ji J Chu M-F Lee H-YWen D-J Lo C-H Lee T-Y Qian Q amp Zhang Q (2003) Adakites from conti-nental collision zone melting of thickened lower crust beneathsouthernTibet Geology 31 1021^1024
Chung S-L Chu M-F Zhang Y Xie Y Lo C-H Lee T-YLan C-Y Li X Zhang Q amp Wang Y (2005) Tibetan tectonicevolution inferred from spatial and temporal variations in post-col-lisional magmatism Earth-Science Reviews 68 173^196
Coulon C Maluski H Bollinger C amp Wang S (1986) Mesozoicand Cenozoic volcanic rocks from central and southern Tibet39Ar^40Ar dating petrological characteristics and geodynamicalsignificance Earth and Planetary Science Letters 79 281^302
Debon F Le Fort P Sheppard S M F amp Sonet J (1986) The fourplutonic belts of the trans-Himalaya a chemical mineralogicalisotopic and chronological synthesis along a Tibet^Nepal sectionJournal of Petrology 27 219^250
Dunn T amp Sen C (1994) Mineralmatrix partition coefficients fororthopyroxene plagioclase and olivine in basaltic to andesitic sys-tems a combined analytical and experimental study Geochimica etCosmochimica Acta 58 717^733
Eggins S M amp Shelley J M G (2002) Compositional heterogeneityin NIST SRM 610-617 Glasses Geostandards Newsletter 26 269^286
Folaquo rster H-J (1998)The chemical composition of REE^Y^Th^U-richaccessory minerals in peraluminous granites of the Erzgebirge^Fichtelgebirge region Germany Part II Xenotime American
Mineralogist 83 1302^1315Griffin W L Pearson N J Belousova E Jackson S E van
Achterbergh E OrsquoReilly SY amp Shee S R (2000)The Hf isotopecomposition of cratonic mantle LAM-MC-ICPMS analysis ofzircon megacrysts in kimberlites Geochimica et Cosmochimica Acta
64 133^147GriffinW LWang X Jackson S E Pearson N J OrsquoReilly SY
Xu X amp Zhou X (2002) Zircon chemistry and magma mixingSE China in-situ analysis of Hf isotopes Tonglu and Pingtanigneous complexes Lithos 61 237^269
Griffin W L Powell W J Pearson N J amp OrsquoReilly S Y (2008)GLITTER Data reduction software for laser ablation ICP-MS(appendix) In Sylvester P (ed) Laser Ablation-ICP-MS in the
Earth Sciences Mineralogical Association of Canada (MAC) Short Course
Series 40 308^311Gulaquo nther D Jackson S E amp Longerich H P (1999) Laser ablation
and arcspark solid sample introduction into inductively coupled
plasma mass spectrometers Spectrochimica Acta Part B Atomic
Spectroscopy 54 381^409Harris N B W Inger S amp Xu R (1990) Cretaceous plutonism in
Central Tibet An example of post-collision magmatism Journal ofVolcanology and Geothermal Research 44 21^32
Hawkesworth C J amp Kemp A I S (2006) Using hafnium andoxygen isotopes in zircons to unravel the record of crustal evolu-tion Chemical Geology 266 144^162
Hoskin PW O Kinny P DWyborn D amp Chappell BW (2000)Identifying accessory mineral saturation during differentiation ingranitoids magmas an integrated approach Journal of Petrology 411365^1396
Hsieh P-S Chen C-H Yang H-J amp Lee C-Y (2008)Petrogenesis of the Nanling Mountains granites from SouthChina constraints from systematic apatite geochemistry andwhole-rock geochemical and Sr^Nd isotope compositions Journalof Asian Earth Sciences 33 428^451
Irber W (1999) The lanthanide tetrad effect and its correlation withKRb EuEu SrEu YHo and ZrHf of evolving peraluminousgranite suites Geochimica et Cosmochimica Acta 63 489^508
Jackson S E (2001) The application of NdYAG lasers in LA-ICP-MS In Sylvester P (ed) Principles and Applications of Laser Ablation
ICP-Mass Spectrometry in the Earth Sciences Mineralogical Association of
Canada (MAC) Short Course Series 29 29^45Jahn B M Wu F Capdevila R Martineau F Wang Y amp
Zhao Z (2001) Highly evolved juvenile granites with tetrad REEpatterns the Woduhe and Baerzhe granites from the GreatXingrsquoan (Khingan) Mountains in NE China Lithos 59 171^198
JiW-QWu F-Y Chung S-L Li J-X amp Liu C-Z (2009) ZirconU^Pb geochronology and Hf isotopic constraints on petrogenesis ofthe Gangdese batholith southernTibet Chemical Geology 262 229^245
Kapp P Yin A Harrison T M amp Ding L (2005) Cretaceous^Tertiary shortening basin development and volcanism in centralTibet Geological Society of America Bulletin 117 865^878
Kemp A I S Hawkesworth C J Foster G L Paterson B AWoodhead J D Hergt J M Gray C M amp Whitehouse M J(2007) Magmatic and crustal differentiation history of graniticrocks from Hf^O isotopes in zircon Science 315 980^983
Lee H-Y Chung S-L Wang Y Zhu D C Yang J H Song BLiu D amp Wu FY (2007) Age petrogenesis and geological signif-icance of the Linzizong volcanic successions in the Linzhou basinsouthern Tibet evidence form zircon U^Pb and Hf isotopes (inChinese with English abstract) Acta Petrologica Sinica 23 493^500
Lee H-Y Chung S-L Lo C-H Ji J Lee T-Y Qian Q ampZhang Qi (2009) Eocene Neotethyan slab breakoff in southernTibet inferred from the Linzizong volcanic record Tectonophysicsdoi101016jtecto200902031
Molnar P amp Tapponnier P (1975) Cenozoic tectonics of Asia effectsof a continental collision features of recent continental tectonicsin Asia can be interpreted as results of the India^Eurasia collisionScience 189 419^426
Norman M D Pearson N J Sharma A amp Griffin W L (1996)Quantitative analysis of trace elements in geological materials bylaser ablation ICPMS instrumental operating conditions and cali-bration values of NISTglasses Geostandards Newsletter 20 247^261
Norman M D Griffin W D Pearson N J Garciac M O ampOrsquoReilly S Y (1998) Quantitative analysis of trace element abun-dances in glasses and minerals A comparison of laser ablationinductively coupled plasma mass spectrometry solution inductivelycoupled plasma mass spectrometry proton microprobe and elec-tron microprobe data Journal of Analytical Atomic Spectrometry 13477^482
JOURNAL OF PETROLOGY VOLUME 00 NUMBER 0 MONTH 2009
26
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27
Pan G Ding J Yao D amp Wang L (2004) Guide book of 11500000geologic map of the Qinghai^Xizang (Tibet) plateau and adjacent areasChengdu China Chengdu Cartographic Publishing House
SanoY OyamaTTerada K amp Hidaka H (1999) Ion microprobeU^Pb dating of apatite Chemical Geology 153 249^258
Schalaquo rer U Corfu F amp Demaiffe D (1997) U^Pb and Lu^Hf iso-topes in baddeleyite and zircon megacrysts from the Mbuji-Mayikimberlite constraints on the subcontinental mantle Chemical
Geology 143 1^16Sha L-K amp Chappell B W (1999) Apatite chemical composition
determined by electron microprobe and laser-ablation inductivelycoupled plasma mass spectrometry as a probe into granite petro-genesis Geochimica et Cosmochimica Acta 63 3861^3881
Taylor S R amp McLennan S M (1985) The Continental Crust its
Composition and Evolution Oxford Blackwell Scientific PublicationsValley J W (2003) Oxygen isotopes in zircon In Hanchar J H amp
Hoskin P W O (eds) Zircon Mineralogical Society of America
Reviews in Mineralogy and Geochemistry 53 343^385Wen D-R (2007) The Gangdese batholith southernTibet ages geo-
chemical characteristics and petrogenesis PhD Thesis NationalTaiwan UniversityTaipei
Wen D-R Chung S-L Song B IizukaY Yang H-J Ji J LiuD amp Gallet S (2008a) Late Cretaceous Gangdese intrusions of
adakitic geochemical characteristics SE Tibet petrogenesis andtectonic implications Lithos doi101016jlithos200802005
Wen D-R Liu D Chung S-L Chu M-F Ji J Zhang QSong B Lee T-Y Yeh M-W amp Lo C-H (2008b) ZirconSHRIMP U^Pb ages of the Gangdese Batholith and implicationsfor Neotethyan subduction in southernTibet Chemical Geology 252191^201
Wilde S A Valley J W Peck W H amp Graham C M (2001)Evidence from detrital zircons for the existence of continentalcrust and oceans on the Earth 44 Gyr ago Nature 409 175^178
Willigers B J A Baker J A Krogstad E J amp Peate DW (2002)Precise and accurate in situ Pb^Pb dating of apatite monaziteand sphene by laser ablation multiple-collector ICP-MSGeochimica et Cosmochimica Acta 66 1051^1066
Xu R Scharer U amp Alle gre C J (1985) Magmatism and meta-morphism in the Lhasa Block (Tibet) a geochronological studyJournal of Geology 93 41^57
Yin A amp Harrison T M (2000) Geologic evolution of theHimalayan^Tibetan orogen Annual Review of Earth and Planetary
Sciences 28 211^280Yurimoto H Duke E F Papike J J amp Shearer C K (1990) Are
discontinuous chondrite-normalized REE patterns in pegmatiticgranite systems the results of monazite fractionation Geochimica etCosmochimica Acta 54 2141^2145
CHU et al APATITE IN TRANSHIMALAYAN GRANITOIDS
27