lable at ScienceDirect

Quaternary Geochronology 22 (2014) 57e64

Contents lists avai

Quaternary Geochronology

journal homepage: www.elsevier .com/locate/quageo

Research paper

High precision multi-collector 40Ar/39Ar dating of young basalts:Mount Rouse volcano (SE Australia) revisited

Erin L. Matchan*, David PhillipsSchool of Earth Sciences, The University of Melbourne, Parkville, VIC 3010, Australia

a r t i c l e i n f o

Article history:Received 31 July 2013Received in revised form6 February 2014Accepted 18 February 2014Available online 26 February 2014

Keywords:40Ar/39Ar geochronologyMulti-collector mass spectrometryARGUSVIBasaltGroundmassIncremental heating

* Corresponding author. Tel.: þ61 3 9035 9621.E-mail addresses: ematchan@unimelb.edu.au, erin.matc

http://dx.doi.org/10.1016/j.quageo.2014.02.0051871-1014/� 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

The acquisition of high precision geochronological data for young basalts, particularly low potassiumexamples <500 ka old, is a non-trivial task. This is due to low 40Ar radiogenic argon yields, high at-mospheric argon levels and, sometimes, extraneous argon contamination. In this study we present new,high precision argon isotopic data for a basalt flow from the Mount Rouse volcano of the Newer VolcanicProvince, southeast Australia. These data were generated using a new generation, multi-collector,ARGUSVI mass spectrometer. We show that the multi-collector system achieves an order of magni-tude improvement in analytical precision and a concomitant improvement in our ability to detect minorisotopic disturbances, compared to ‘conventional’ single-collector mass spectrometry. We report arevised eruption age of 284.4 � 1.8 ka (0.6%, 95% CI) for Mount Rouse, based on laser step-heating ex-periments of several groundmass aliquants (50e100 mg) from a single sample (180e250 mm size frac-tion). This result is considerably more precise than our previous estimate of 303 � 13 ka (4.3%, 2s),obtained using an older generation VG3600 mass spectrometer and larger (w200 mg) sample aliquants.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction

A precise chronology of young (Quaternary; <2.6 Ma) basalticeruption events is critical for assessing volcanic hazards, magmadynamics, the causes of magmatism, eruption mechanisms, andchrono- and magnetoestratigraphic correlations and calibrations.To date, high precision 40Ar/39Ar dating of young basalts has provenchallenging, due to relatively lowpotassium contents (K2O typically<3 wt %), low proportions of radiogenic 40Ar (often <5%) and highlevels of atmospheric argon contamination. These measurementshave typically required large sample sizes (up to 1 g) and 40Ar/39Arstep-heating analyses of multiple aliquants, to achieve ‘reasonable’precision levels ofw10% (e.g. Singer and Pringle, 1996; Singer et al.,2002; Guillou et al., 2004; Singer et al., 2004; Jicha and Singer,2006; Matchan and Phillips, 2011). Until now, the ultimate preci-sion of individual measurements has been limited by the availablenoble gas mass spectrometers, most of which are essentially single-collector instruments, based on 1980s designs (e.g. VG3600,MAP215-50).

The laser step-heating studies of Sharp and Renne (2005) andCalvert and Lanphere (2006) on Hawaiian alkali basalts

han@gmail.com (E.L. Matchan).

demonstrate current analytical precision levels for conventionalmass spectrometry. Groundmass samples (w150 mg) analysed bySharp and Renne (2005) yielded plateau ages of ca. 240 ka withassociated errors of 5e10% (95% CI). Ages calculated for individualsteps show precision levels upwards of w12% (95% CI). Similarly,Calvert and Lanphere (2006) obtained plateau ages with reporteduncertainties of 18e25% (2s) for 150e250 mg aliquants fromsomewhat younger basalt flows (ca. 140e200 ka). In this case,uncertainties from individual heating steps ranged upwards fromw35% (2s).

Although averaging results from multiple concordant analysesgives rise to impressively precise apparent ages, large numbers ofanalyses are typically required to obtain reasonable precision levels(w5%; 2s). For instance, Cassata et al. (2008) reported an appar-ently precise age of 32.1�5.4 ka (2s) for the Crater Hill basalt (NewZealand); however, this result was obtained by averaging >5040Ar/39Ar measurements, each with uncertainty levels in excess of50% (2s). As another example, Jicha and Singer (2006) analysed 60e100 ka basalts from Seguam Island (Alaska) and noted that mea-surements on three or four w200 mg aliquants, were necessary toobtain weighted mean ages with ‘acceptable’ precision levels of10e15% (2s).

The advent of new generation, high sensitivity, multi-collectormass spectrometers (e.g. ARGUSVI, HELIX) provides renewed

Penshurst

Port Fairy

Koroit

50

100

150

200

250

km0 10

MT NAPIER

MT ECCLES

MT ROUSE

McArthur

Princes Hwy

NVP21

Hawkesdale

38.0°

38.2°

142.0° 142.2°N

VICSA

Moy

neRi

ver

Fig. 1. Locality map, modified from Matchan and Phillips (2011). Basalt flows fromMount Rouse and neighbouring volcanoes Mount Napier and Mount Eccles are shaded.Sampling locality of basalt sample NVP21 indicated by triangle. Also shown are 50 melevation contours and towns (circles). Inset map shows distribution of the NewerVolcanic Province and location of main map.

E.L. Matchan, D. Phillips / Quaternary Geochronology 22 (2014) 57e6458

impetus to this research field and significantly enhances our abilityto date young (<500 ka) volcanic rocks with extremely highanalytical precision (<1%). The first (in the world) new generationARGUSVI mass spectrometer was recently installed at the Univer-sity of Melbourne. The ultra-high analytical precision achievablewith this instrument has been demonstrated on 40Ar/39Ar datingstandards by Phillips and Matchan (2013). To illustrate theconsiderable potential of this system in the study of young basalts,we compare ARGUSVI 40Ar/39Ar step-heating analyses of aw300 kabasalt sample from theMount Rouse volcano (Victoria, Australia) todata for the same sample acquired using a ‘standard’ VG3600 massspectrometer (Matchan and Phillips, 2011). Although absolute agesfrom both mass spectrometer systems are indistinguishable withinuncertainties, individual measurements on the ARGUSVI returnedprecision levels of <1%, compared to �20% for the VG3600. Theseresults are unprecedented and demonstrate the exceptional capa-bility of the new generation mass spectrometers to deliver highprecision age data. In turn, this capability enhances potential foreffective inter-calibration of all noble gas dating methods (e.g.40Ar/39Ar, cosmogenic 3He, 21Ne) and dating standards, and facili-tates recognition of volcanic and geomorphic events that were notresolvable previously.

2. Geology

The Pliocene-Holocene Newer Volcanic Province (NVP) ofsoutheastern Australia is an extensive, well-preserved, intra-platebasaltic lava field containing more than 400 eruptive centres.Available geochronological data indicate that NVP volcanism(dominantly tholeiitic to alkalic) initiated at ca. 4.5 Ma in theMelbourne region and Central Uplands sub-province, and thenmigrated throughout southern and western Victoria (WesternPlains sub-province) and South Australia (Mt Gambier sub-province). Peak activity, as revealed by KeAr geochronology, isconsidered to have occurred between 3.0 and 1.4 Ma, with subse-quent eruption episodes at ca. 1.0e0.8 Ma and 60e5 ka (see sum-mary by Matchan and Phillips, 2011).

Mount Rouse is a composite, alkali basalt volcano located in theWestern Plains sub-province of the NVP. The Rouse lava field coversan area >450 km2 (Elias, 1973), with the longest continuous flowextending approximately 60 km to the coast at Port Fairy (Fig. 1).The sample chosen for the present study is a finer crush fraction(180e250 mm) of the sample (NVP21) studied by Matchan andPhillips (2011). The sample was collected from a fresh, hol-ocrystalline late-stage lava flow (K2O ¼ 1.18 wt %; J. Boyce pers.comm.) intercalated with scoria and ash in the southernmost craterof the Mount Rouse complex.

3. Sample preparation and irradiation

Thin section petrography of sample NVP21 from Mt Rouse re-veals a fresh, well-crystallised olivine basalt with <15 vol% vesicles(Fig. 2). Approximately 1 kg of the sample was crushed to w2 cmchips using a jaw crusher. Individual chips were then screened foralteration and large vesicles, with acceptable chips crushedmanually using a steel piston crusher. Crushed fragments werewashed of dust and sieved to a 180e250 mm grain size (after Ozawaet al., 2006). Following purification of whole rock groundmassmaterial by magnetic separation, approximately 600 mg ofgroundmass material was hand-picked under a binocular micro-scope. Grains were treatedwith 5% HNO3 for 10min in an ultrasonicbath, followed by 2% HF for 1 min, followed by deionised water for10 min. The sample was then loaded into an aluminium foil packet,placed in a quartz tube (UM#54), along with the flux monitorstandard Alder Creek Rhyolite (ACR) sanidine. Can UM#54 was

irradiated for 30 min in the Cd-shielded CLICIT facility of the Ore-gon State University TRIGA reactor.

4. Analytical procedures

Individual aliquants of either 50 or 100mgwere loaded as singleor double grain layers into a custom-made circular copper sampleholder containing four parallel slots, measuring 6 mm wide, 25e30 mm long and 3 mm deep. The holder was covered with a ZnSglass disc and loaded into the sample chamber of a gas-handlingsystem connected to a new generation multicollector ARGUSVImass spectrometer. As described in detail by Phillips and Matchan(2013), the ARGUSVI detector array comprises 5 Faraday detectorsand a CDD (compact discrete dynode) electron multiplier, allowingsimultaneous collection of all 5 argon isotopes.

The extraction line and contained samples were baked for atleast 12 hours at w120 �C. Once acceptable (UHV) backgroundlevels were achieved, each aliquant was heated with a PhotonMachines Fusions 10.6 CO2 laser to remove the majority of loosely-bound excess argon; this was achieved using a 6 mm homogenisedbeam operated step-wise between 1 and 5% laser power (w0.23e1.2W). For step-heating experiments, aliquants (n¼ 5) were heatedincrementally using the 6 mm beam over an energy range of 1.3e

Fig. 2. Photomicrograph of sample NVP21. Phenocrysts of plagioclase (0.5e2 mm),augite and olivine are set in an ophitic groundmass containing iron oxide (magnetite)and apatite. Glass is not observed. Olivine shows minor alteration to iddingsite,otherwise sample appears unaltered.

E.L. Matchan, D. Phillips / Quaternary Geochronology 22 (2014) 57e64 59

2.9 W (5.5e14.0% laser power). Samples were exposed to 3 passesof the beam, equating to w4 min duration, with evolved gascontinually exposed to two SAES ZreAl getters (see Phillips andMatchan, 2013).

Argon isotopic results, reported in Table 1, are corrected forsystem blanks, mass discrimination, radioactive decay and reactor-induced interference reactions. Isotopic results that exclude theinterference corrections (as per the recommendation of Renneet al., 2009) are presented in the electronic appendix, along withblank correction values (Table A1). Line blanks weremeasured afterevery second or third sample analysis and were typically<1.5 fA for40Ar, compared to �100 fA for typical sample analyses. Massdiscrimination and detector bias were characterised via automatedanalyses of air pipette aliquots prior to the first analysis, assumingan atmospheric 40Ar/36Ar ratio of 298.56 � 0.31 (Lee et al., 2006).Due to the short irradiation time it was not feasible to include Ca/K/Cl salts/glasses in the same package. Therefore, correction factorsdetermined for K-glass and Ca-salts contained in another recentpackage irradiated in the CLICIT facility (UM#52) were used:(39Ar/37Ar)Ca ¼ (6.5075 � 0.0033) � 10�4; (36Ar/37Ar)Ca ¼(2.5680 � 0.0017) � 10�4; (40Ar/39Ar)K ¼ (5.93 � 0.16) � 10�4; and(38Ar/39Ar)K ¼ (1.2055 � 0.0007) � 10�2.

The J-value for sample NVP21 (0.00013175 � 0.000000084;0.06% 1s) was calculated by averaging the mean 40Ar*/39Ar ratiosfrom fusion analyses conducted on aliquots (3 grains) of bracketingACR sanidine standards (AC2 and AC3, see Table A2). 40Ar/39Ar ages

were calculated relative to an ACR sanidine age of 1.186 � 0.012 Ma(1s) (Turrin et al., 1994), using the decay constants of Steiger andJäger (1977) and the Lee et al. (2006) atmospheric argon compo-sition. For direct comparison with the results of Matchan andPhillips (2011), data were also processed using the Nier (1950) airratio (see Table A1).

Calculated uncertainties associated with mean and plateau agesinclude uncertainties in the J-value, but exclude errors associatedwith the age of the flux monitor and the decay constant. Unlessotherwise stated, uncertainties are reported at the 2s or 95% con-fidence level (see Ludwig, 2012). Plateau ages are defined asincluding >50% of the total 39Ar, from at least 3 contiguous steps,with 40Ar*/39Ar ratios within error of the mean at the 95% confi-dence level (e.g. McDougall and Harrison, 1999).

5. Results

ARGUSVI analytical results from laser step-heating experimentson five NVP21 aliquants (V21b,c,d,e,f) are summarised in Table 1.The use of the Lee et al. (2006) versus the Nier (1950) values for the(40Ar/36Ar)air ratio does not significantly affect the calculated ages(Table A1), with values changing by �2.3% (typically <1.5%). Thisvariation is within 1s error bounds for all but two heating steps(V21e Step 9 and V21f Step 5), where values are within 2s (seeTable A1). Step-heating spectra and isochron plots (Fig. 3) weregenerated using ISOPLOT/Ex v.3.75 (Ludwig, 2012). In the two caseswhere plateau age criteria were not met (V21d,f), ‘weighted meanages’ were calculated from the weighted mean of 40Ar*/39Ar valuescalculated for each step, with the associated error incorporating theuncertainty in the J-value. A summary of age results, including in-verse isochron results and total gas ages, is provided in Table 2.

Steps from all five aliquants show generally concordant, highlyprecise apparent ages ranging from 273.6 � 5.2 to 317.4 � 12.6 ka(2s). Three of the five aliquants yield analogous plateau ages of283.1 � 1.7 ka (0.6%, 2s; V21b (Fig. 3a)), 286.8 � 2.4 ka (0.8%, 2s;V21c (Fig. 3c)) and 284.7 � 2.2 ka (0.7%, 2s; V21e (Fig. 3g)). Inverseisochron analysis reveals an atmospheric 40Ar/36Ari composition of303.3� 5.6 (2s) for V21c (Fig. 3d), but slightly discordant 40Ar/36Arivalues of 291.3 � 4.5 (2s) and 304.7 � 4.5 (95% CI) for aliquantsV21b and V21e respectively (Table 2). If only the plateau-formingsteps are considered for these two aliquants, the 40Ar/36Ari valuesbecome indistinguishable from air at 292.0 � 8.3 (2s, V21b) and301.9 � 4.2 (2s, V21e), with corresponding inverse isochron ages of294 � 15 (2s, V21b) and 278.1 � 9.1 (2s, V21e).

Age spectra for the remaining two aliquants V21d and V21f(Fig. 3f,j) are slightlymore discordant and do not yield plateau ages.Inverse isochron regression for these aliquants reveals an atmo-spheric trapped component composition for V21d (295.2� 8.1 (95%CI)) and a slightly sub-atmospheric ratio for V21f (289.6 � 7.5 (95%CI); see Table 2). Weighted mean ages calculated from individual40Ar*/39Ar values measured for each step are 281.6 � 4.2 ka (1.5%,95% CI) and 290.2 � 7.8 ka (2.7%, 95% CI) for aliquants V21d andV21f, respectively. These values are indistinguishable from theplateau ages calculated from the other three aliquants, as areassociated total gas ages (Table 2).

6. Discussion

The five Mount Rouse aliquants yielded generally reproducibledegassing profiles and apparent ages, although some discrepanciesare apparent (i.e. for V21f and the high-T steps of V21e) that areamplified by the remarkable analytical precision of the ARGUSVIsystem. Implications for eruption age estimates are addressed first,followed by a discussion of possible causes for the minor discor-dances within and between age spectra.

Table 1ARGUSVI 40Ar/39Ar laser step-heating analytical resultsa.

Sample ID Stepno.

Laserpower (W)

40Ar � 1s 39Ar �1s 38Ar �1s 37Ar �1s 36Ar �1s 39Ar (�10�15 mol) Ca/K � %40Ar* 40Ar*/39Ar �1s Cum.% 39Ar Apparentage (ka)c

�1s

(fA) (fA)b (fA) (fA)b (fA)

Aliquant V21b 103 mgV21b_1 1 1.5 W 119.195 0.043 44.719 0.018 0.04103 0.00026 23.89 0.08 0.2184 0.0014 0.159 0.9347 0.0032 45.29 1.207 0.009 13.61 287.0 2.2V21b_2 2 2.0 W 324.867 0.097 106.877 0.046 0.12420 0.00030 58.31 0.11 0.6611 0.0016 0.379 0.9548 0.0018 39.24 1.193 0.005 46.13 283.5 1.1V21b_3 3 2.4 W 290.694 0.064 87.951 0.041 0.11666 0.00041 57.82 0.08 0.6210 0.0022 0.312 1.1505 0.0017 36.21 1.197 0.007 72.89 284.5 1.8V21b_4 4 2.6 W 196.182 0.047 54.127 0.048 0.08338 0.00029 47.44 0.12 0.4438 0.0015 0.192 1.5339 0.0040 32.45 1.176 0.008 89.36 279.6 2.0V21b_5 5 2.9 W 141.566 0.042 34.960 0.019 0.06332 0.00021 38.03 0.08 0.3370 0.0011 0.124 1.9036 0.0041 28.92 1.171 0.009 100.00 278.4 2.2

Total gas age: 284 � 13 (2s)Aliquant V21c 52 mgV21c_1 1 1.5 W 94.134 0.036 33.355 0.033 0.03388 0.00025 17.64 0.04 0.1803 0.0013 0.118 0.9255 0.0022 42.80 1.208 0.012 20.55 287.1 2.8V21c_2 2 2.0 W 184.545 0.068 57.745 0.025 0.07252 0.00032 35.08 0.04 0.3860 0.0017 0.205 1.0630 0.0013 37.55 1.200 0.009 56.14 285.2 2.1V21c_3 3 2.4 W 134.157 0.047 37.633 0.025 0.05631 0.00028 32.22 0.06 0.2997 0.0015 0.134 1.4984 0.0031 33.30 1.187 0.012 79.33 282.1 2.8V21c_4 4 2.6 W 83.482 0.020 20.712 0.024 0.03659 0.00018 25.59 0.08 0.1948 0.0009 0.074 2.1618 0.0075 30.35 1.223 0.014 92.09 290.7 3.2V21c_5 5 2.9 W 56.607 0.036 12.834 0.032 0.02571 0.00009 19.84 0.05 0.1369 0.0005 0.046 2.7048 0.0097 27.82 1.227 0.012 100.00 291.6 2.9

Total gas age: 286 � 16 (2s)Aliquant V21d 103 mgV21d_1 1 1.5 W 273.101 0.079 80.025 0.057 0.11220 0.00023 76.77 0.15 0.5972 0.0012 0.284 1.6787 0.0034 34.70 1.185 0.005 26.82 281.5 1.1V21d_2 2 2.0 W 423.238 0.148 98.680 0.066 0.19192 0.00052 173.62 0.10 1.0216 0.0028 0.350 3.0791 0.0028 27.93 1.198 0.009 59.89 284.8 2.0V21d_3 3 2.0 W 133.667 0.045 33.758 0.018 0.05848 0.00019 54.84 0.15 0.3113 0.0010 0.120 2.8430 0.0077 30.47 1.206 0.009 71.21 286.7 2.1V21d_4 4 2.4 W 189.686 0.068 38.485 0.037 0.09148 0.00026 103.64 0.17 0.4869 0.0014 0.137 4.7130 0.0090 23.36 1.151 0.011 84.11 273.6 2.6V21d_5 5 2.6 W 142.660 0.044 26.765 0.022 0.06975 0.00035 90.63 0.21 0.3712 0.0019 0.095 5.9256 0.0149 22.30 1.189 0.021 93.08 282.6 5.0V21d_6 6 2.9 W 118.629 0.075 20.656 0.007 0.05952 0.00016 87.39 0.18 0.3168 0.0008 0.073 7.4040 0.0153 20.27 1.164 0.013 100.00 276.7 3.0

Total gas age: 284 � 15 (2s)Aliquant V21e 92 mgV21e_1 1 1.3 W 67.421 0.043 22.357 0.021 0.02576 0.00014 15.41 0.13 0.1371 0.0008 0.079 1.2059 0.0103 39.28 1.185 0.010 10.10 281.6 2.5V21e_2 2 1.4 W 87.593 0.033 25.690 0.051 0.03568 0.00020 22.03 0.10 0.1899 0.0011 0.091 1.5009 0.0077 35.26 1.203 0.013 21.71 285.8 3.1V21e_3 3 1.5 W 94.040 0.040 25.592 0.033 0.03970 0.00014 26.19 0.15 0.2113 0.0008 0.091 1.7909 0.0108 32.91 1.209 0.009 33.28 287.4 2.2V21e_4 4 1.6 W 96.629 0.048 24.899 0.028 0.04234 0.00017 29.91 0.12 0.2254 0.0009 0.088 2.1019 0.0084 30.36 1.178 0.011 44.53 280.0 2.6V21e_5 5 1.7 W 91.598 0.052 22.267 0.028 0.04075 0.00022 32.41 0.14 0.2169 0.0012 0.079 2.5473 0.0113 29.30 1.206 0.016 54.59 286.5 3.7V21e_6 6 1.8 W 83.912 0.040 18.944 0.031 0.03888 0.00023 33.85 0.13 0.2070 0.0012 0.067 3.1274 0.0129 26.36 1.168 0.020 63.15 277.6 4.7V21e_7 7 2.0 W 103.961 0.037 20.628 0.021 0.04947 0.00019 51.54 0.13 0.2633 0.0010 0.073 4.3724 0.0121 24.38 1.229 0.015 72.48 292.1 3.6V21e_8 8 2.2 W 102.611 0.042 18.170 0.036 0.05073 0.00029 60.90 0.18 0.2701 0.0015 0.065 5.8653 0.0205 21.42 1.210 0.026 80.69 287.6 6.1V21e_9 9 2.4 W 95.103 0.041 15.554 0.032 0.04789 0.00009 62.07 0.19 0.2549 0.0005 0.055 6.9833 0.0257 19.97 1.221 0.010 87.72 290.2 2.4V21e_10 10 2.6 W 84.882 0.055 13.181 0.042 0.04294 0.00024 61.88 0.18 0.2285 0.0013 0.047 8.2153 0.0351 19.61 1.263 0.030 93.67 300.2 7.0V21e_11 11 2.9 W 102.831 0.043 14.003 0.036 0.05294 0.00023 81.73 0.18 0.2818 0.0012 0.050 10.2137 0.0344 18.19 1.336 0.026 100.00 317.4 6.3

Total gas age: 291 � 19 (2s)Aliquant V21f 54 mgV21f_1 1 1.4 W 71.856 0.042 20.007 0.022 0.02928 0.00023 20.41 0.12 0.1559 0.0012 0.071 1.7855 0.0103 35.23 1.265 0.018 16.83 300.8 4.3V21f_2 2 1.6 W 99.404 0.038 24.055 0.028 0.04336 0.00031 34.33 0.09 0.2308 0.0017 0.085 2.4977 0.0072 30.68 1.268 0.021 37.08 301.4 4.9V21f_3 3 1.8 W 93.153 0.031 20.929 0.032 0.04242 0.00021 36.73 0.18 0.2258 0.0011 0.074 3.0715 0.0156 27.63 1.230 0.016 54.69 292.3 3.8V21f_4 4 2.2 W 138.938 0.042 24.340 0.039 0.06969 0.00031 84.45 0.15 0.3709 0.0017 0.086 6.0719 0.0143 20.29 1.158 0.020 75.17 275.3 4.9V21f_5 5 2.6 W 108.615 0.090 17.248 0.042 0.05517 0.00006 82.43 0.18 0.2937 0.0003 0.061 8.3632 0.0277 19.27 1.214 0.008 89.68 288.5 2.0V21f_6 6 2.9 W 80.700 0.020 12.265 0.033 0.04161 0.00031 66.23 0.12 0.2215 0.0017 0.044 9.4491 0.0303 18.05 1.188 0.040 100.00 282.3 9.6

Total gas age: 294 � 22 (2s)

a Data are corrected for mass spectrometer backgrounds, discrimination, radioactive decay and interference corrections (see Table A.1 for values excluding the interference correction). Errors are one sigma uncertainties andexclude uncertainties in the J-value (propagating this error only has an effect in the third decimal place).

b Interference corrections: (39Ar/37Ar)Ca ¼ (6.5075 � 0.0033) � 10�4; (36Ar/37Ar)Ca ¼ (2.5680 � 0.0017) � 10�4; (40Ar/39Ar)K ¼ (5.93 � 0.16) � 10�4; and (38Ar/39Ar)K ¼ (1.2055 � 0.0007) � 10�2.c J-value is 0.00013175 � 0.000000084 (0.06%; 1s), based on an age of 1.186 � 0.012 Ma (1s) for ACR sanidine (Turrin et al., 1994).

E.L.Matchan,D

.Phillips/Quaternary

Geochronology

22(2014)

57e64

60

Fig. 3. 40Ar/39Ar age spectra and inverse isochron diagrams for aliquants of Mount Rouse sample NVP21 (‘V21’).

E.L. Matchan, D. Phillips / Quaternary Geochronology 22 (2014) 57e64 61

Fig. 3. (continued).

E.L. Matchan, D. Phillips / Quaternary Geochronology 22 (2014) 57e6462

6.1. Eruption age and comparison with previous geochronology

Minor scatter of 40Ar/36Ari values for individual aliquants aboutthe atmospheric value suggests that the true trapped compositionis likely atmospheric and that minor recoil effects and possiblyisotope mass fractionation have subtly affected the five aliquants.Due to the apparent minor disturbance to isotope ratios, total gasages and plateau ages are preferred over the relatively poorlyconstrained inverse isochron ages. Total gas ages for all five ali-quants are indistinguishable at the 1s-level (Table 2), and yield aweighted mean age of 285.0 � 7.1 (2.5%, 2s; MSWD ¼ 0.16,p ¼ 0.96). A weighted mean age of 284.4 � 1.8 ka (0.6%, 95% CI;MSWD ¼ 2.1, p ¼ 0.01) is calculated from the 40Ar*/39Ar ratios ofplateau-forming steps (n¼ 16) yielded by aliquants V21b, V21c and

Table 2Summary of 40Ar/39Ar results from step-heating experiments on groundmass aliquants o

Age spectrum Inverse i

Aliquantno.

Increments usedfor plateau/wmb

39Ar% MSWD p Age (ka) � 2sa 40Ar/36Ar

V21b 2.0e2.6 W (steps 2e4) 75.8 1.9 0.15 283.1 � 1.7 291.3 �292.0 �

V21c 1.5e2.9 W (all) 100 1.9 0.11 286.8 � 2.4 303.3 �V21d 1.5e2.9 W (all) 100 4.1 0.00 281.6 � 4.2

(95% CI)295.2 �(95% CI)290.4 �

V21e 1.3e2.9 W (steps 1e8) 80.7 1.9 0.06 284.7 � 2.2 304.7 �(95% CI)301.9 �

V21f 1.4e2.9 W (all) 100 4.4 0.00 290.2 � 7.8(95%CI)

289.5 �(95% CI)283.3 �

Weighted mean age, all aliquots (n ¼ 33, 1 rej.): Composit

3.9 0.00 284.3 � 1.7(95% CI)

299.7 �(95% CI)

Weighted mean age, plateau points from aliquots V21b,c,e (n¼ 16): Composit2.1 0.01 284.4 � 1.8 300.2 �

Weighted mean age, all points from aliquots V21b,c,e (n ¼ 21, 1rej.):

Composit

2.7 0.00 284.6 � 1.9(95% CI)

302.3 �(95% CI)

a Ages have been calculated using the decay constants of Steiger and Jäger (1977), assu0.00013175 � 0.06% (1s).

b In the cases of aliquants V21d and V21f, data did not satisfy the requirements for plaaliquants.

V21e (Table 2). The corresponding composite inverse isochron in-dicates an atmospheric 40Ar/36Ari value of 300.2 � 4.1 (95% CI;MSWD ¼ 1.9, p ¼ 0.02) and an age of 282.8 � 8.0 ka (95% CI).Assuming that an atmospheric composition implies minimal iso-topic disturbance for these heating increments, we propose thatthe weighted mean age of 284.4 � 1.8 ka (95% CI) best reflects theeruption age of Mount Rouse. This is consistent with the moreconservative weighted mean total gas age of 285.0 � 7.1 (2s).

This result is essentially identical to the weighted mean age of280 � 19 ka (6.8%, 2s) reported for the same sample (NVP21) byMatchan and Phillips (2011), but an order of magnitude moreprecise. In addition, the VG3600 result was generated by combiningdata from the plateau-forming steps of four NVP21 aliquantsweighing 160e200 mg each (compared to 50e100 mg aliquants

f sample NVP21.

sochron analysis Total gas age

i � 2s MSWD p n Age (ka) � 2s (ka) � 2s

4.5 0.65 0.59 5 (all) 294.9 � 7.3 283 � 138.5 1.8 0.18 3 (steps 2e4) 294 � 155.6 1.6 0.19 5 (all) 278 � 11 286 � 168.1 4.2 0.002 6 (all) 290 � 21 (95%CI) 282 � 15

4.3 2.5 0.06 5 (steps 2e6) 304 � 124.5 2.6 0.005 11 (all) 272 � 12 (95%CI) 288 � 18

4.2 1.8 0.09 8 (steps 1e8) 278.1 � 9.17.5 2.7 0.03 6 (all) 318 � 21 (95%CI) 291 � 22

6.7 0.70 0.5 4 (steps 1e4) 279.4 � 3.9e inverse isochron, all aliquots (n ¼ 33): Weighted mean

total gas age:3.1 4.6 0.00 33 283.6 � 7.1

(95%CI)285.0 � 7.1(2s; MSWD ¼ 0.16,p ¼ 0.96)

e inverse isochron, plateau points from aliquots V21b,c,e:4.1 1.9 0.02 16 282.8 � 8.0

(95% CI)e inverse isochron age, all points from aliquots V21b,c,e:

3.2 32.7 0.00 21 278.1 � 6.5(95% CI)

ming an age of 1.186 � 0.012 Ma (1s) for ACR sanidine (Turrin et al., 1994); J-value is

teau age calculation. Weighted mean (wm) ages were therefore calculated for these

Fig. 4. Comparison of age spectra calculated for individual aliquants of sample NVP21 using the VG3600 versus the ARGUSVI systems. The J-values calculated for NVP21 packetsirradiated in UM#54 (for ARGUSVI analyses) and UM#23 (for VG3600 analyses) are 0.00013175 � 0.06% (1s) and 0.000743 � 0.44% (1s) respectively.

E.L. Matchan, D. Phillips / Quaternary Geochronology 22 (2014) 57e64 63

used in the current study). As illustrated in Fig. 4, analytical errorsassociated with the VG3600 data are relatively large, with 1s un-certainties for individual apparent ages being 6e40%. In contrast,apparent ages calculated for the ARGUSVI data (aliquants of 50 or100 mg) have associated 2s errors of 0.8e6.8% (but typically �2%),reflecting the marked improvement in analytical precision (Table 1,Fig. 4).

The main contributing factors to enhanced precision on theARGUSVI are use of the low-noise, high sensitivity CDD detector, tomeasure 36Ar, and longer counting times due to multi-collection,accounting for w60e65% of precision improvement. Contribu-tions from other factors are more difficult to deconvolute, butinclude the following: (i) use of Faraday detectors with low-noise1 � 1012U resistors to measure 40Ar, 39Ar, 38Ar, 37Ar (w15e20%contribution); (ii) the volume difference between the ARGUSVI(w900 cc) and the VG3600 (w2600 cc), coupled with enhancedsource sensitivity (3.55 � 10�17 mol/mV on ARGUSVI Faraday de-tectors, versus 4.00 � 10�17 mol/mV on the VG3600 Daly multi-plier); this effect is, however, partly offset by the larger sample sizesused in the VG3600 experiments (�10% contribution); (iii)increased 40Ar* yield due to improved sample preparation pro-cedures applied to the ARGUSVI samples (�10% contribution); (iv)lower line blanks (<5% contribution), although blank corrections(see Table A1) have a minimal impact on the age results.

The Matchan and Phillips (2011) study involved analysis ofwhole-rock samples (including NVP21) from three distinct Rouselava flows. By pooling concordant data from 14 furnace step-heating experiments on all three samples, Matchan and Phillips(2011) calculated a weighted mean age of 303 � 13 ka (95% CI),which was interpreted as the time of Mt Rouse basalt eruption. Thisresult is slightly older than the ARGUSVI NVP21 result of284.0 � 1.4 ka (2s). The disparity between these ages warrantssome discussion. One possibility is that the Mt Rouse lavas repre-sent multiple flows that erupted at different times, and could not beresolved using the VG3600 mass spectrometer. However, mappingby Sutalo and Joyce (2004) and the lack of evidence of inter-flowerosion, suggests that emplacement of Mt Rouse pyroclastic ma-terial and the lava flows took place over a relatively short timescale,on the order of weeks to months. An alternative explanation is

that the older age of 303 � 13 ka (95% CI) reflects a contributionfrom undetected extraneous argon in individual samples and/oraliquants. Inverse isochron regressions were poorly constrained inthe Matchan and Phillips (2011) study, making it difficult to resolvethe composition of the trapped component and inverse isochronages. Therefore, although 40Ar/36Ari values were statisticallyindistinguishable from the atmospheric composition, it is possiblethat the uncertainties mask the presence of extraneous argon (e.g.extraneous argon in phenocrysts), resulting in older apparent ages.

6.2. Evaluation of minor discordance in the age spectra

As noted above, the unprecedented analytical precision ob-tained from the ARGUSVI mass spectrometer amplifies minor dis-crepancies that exist within and between the Mount Rouse basaltage spectra. These discordances may reflect: (1) extraneous argoncontamination (e.g. inherited argon from mantle xenocrysts); (2)minor 39Ar recoil loss/redistribution; and/or (3) fractionation ofargon isotopes during baking and step-heating.

The initial 40Ar/36Ar ratios determined from inverse isochronregressions reveal atmospheric or slightly sub-atmospheric trap-ped component compositions in all cases except the highest tem-perature steps for aliquant V21e. Aliquant V21ewas heated over thesame temperature interval as the other four aliquants, but insmaller increments (1.3e2.9 W over 11 steps, see Table 1). The firsteight steps of the V21e age spectrum (Fig. 3g) yield a plateau age of284.7 � 2.2 ((2s); Steps 1e8). The final two steps yield olderapparent ages of 300 � 14 (2s) and 317 � 13 ka (2s). With theexception of the final step of V21c (apparent age¼ 292� 6 ka (2s)),older apparent ages are not observed for the other aliquants.Therefore, the elevated apparent ages associated with the high-Tsteps of aliquants V21e and V21c may indicate release of inheri-ted argon from phenocrysts/mantle xenocrysts overlooked duringhandpicking (e.g. Cassata et al., 2008).

In recent years, attention has focussed on the role of recoil loss/redistribution of 39ArK in minerals and whole rock samples and theaffects on age spectra. Vacuum-encapsulation 40Ar/39Ar studieshave quantified significant recoil loss of 39Ar during irradiation ofminerals such as clays, micas and sanidine (e.g. Dong et al., 1995;

E.L. Matchan, D. Phillips / Quaternary Geochronology 22 (2014) 57e6464

Hall et al., 1997, 2000, 2013). Recoil effects in whole-rock samplesare likely to be more complex and vary from sample to sample.Recoil loss and redistribution of 39Ar and 37Ar during irradiationmay be significant in glassy, altered basalt samples (e.g. Kopperset al., 2000). Recoil of 39Ar from high potassium, low temperaturesites into low-potassium, high temperature sites is thought toresult in elevated apparent ages for low temperate steps, whereasrecoil of 37Ar leads to decreased apparent ages for high temperaturesteps as calcium-rich sites outgas (e.g. Koppers et al., 2000). Theseeffects are not immediately apparent in the Rouse spectra, althoughcould be obscured by competing processes.

Age spectra showing increasing apparent ages with increasingtemperature could be indicative of mass fractionation effects (e.g.,Trieloff et al., 2005; Phillips and Matchan, 2013). However, if massfractionationwere a factor, 36Ar would be preferentially removed atlow temperatures, leading to younger apparent ages for these steps,and older apparent ages in subsequent steps. This is not observed inthe Rouse age spectra (with the exception of V21e) and 36Ar/40Arratios consistently increase with temperature.

In conclusion, the most likely explanation for the minor dis-cordances observed within and between the Mt Rouse aliquants, isconsidered to be release of heterogeneously distributed inheritedargon from phenocrysts and/or mantle xenocrysts, possibly com-bined with minor recoil loss/redistribution effects.

7. Conclusions

� The plateau-forming steps from aliquants V21b, V21c and V21eyield a weighted mean age of 284.4 � 1.8 ka (0.6%, 95% CI;MSWD ¼ 2.1, p ¼ 0.01). This value is 7 times more precise thanthe older pooled age of 303�13 ka (95% CI) previously proposedfor the eruption of Mount Rouse by Matchan and Phillips (2011),and 10 times more precise than the age of 280 � 19 ka (2s)previously reported for this particular sample (NVP21).

� The observed improvement in analytical precision is due tovariety of factors, such as the smaller volume of the ARGUSVIsystem, enhanced source sensitivity, improved sample selectionprotocols and, most significantly, the low-noise detector array(particularly the high-sensitivity CDD for measurement of 36Ar).

� The current results indicate that the multi-collector ARGUSVImass spectrometer is ideally suited for high precision geochro-nology of young basalts. Aliquot sizes of w50 mg are adequatefor 5-step heating experiments, at least in cases where 40Ar*comprises �30% of total 40Ar, yielding apparent ages for indi-vidual steps with 1e2% 1s uncertainties; however, aliquot sizesof w100 mg improve uncertainty levels to �1% for the samestep-heating routine.

� Minor discordance observed between age spectra is attributedto a combination of inherited argon in phenocrysts/xenocrystsand minor argon recoil loss/redistribution. The unprecedentedanalytical precision of the ARGUSVI will permit more detailedinvestigation into these processes.

� Careful monitoring of neutron flux gradients across bulk sam-ples subjected to short irradiation times is necessary to ensurehigh precision results, while the accuracy of the ARGUSVI resultsis predicated on the availability of co-irradiated flux monitorsthat are homogeneous, with well-constrained ages (see dis-cussion by Phillips and Matchan, 2013).

Acknowledgements

This study is supported by Australian Research Council Discov-ery grant DP130100517 to D. Phillips. S. Szczepanski is thanked fortechnical assistance in the Melbourne 40Ar/39Ar laboratory. The

authors acknowledge the constructive and insightful commentsfrom P. Renne, A. Calvert and an anonymous reviewer.

Editorial handling by: P.R. Renne

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.quageo.2014.02.005.

References

Calvert, A.T., Lanphere, M.A., 2006. Argon geochronology of Kilauea’s early sub-marine history. J. Volcanol. Geotherm. Res. 151, 1e18.

Cassata, W.S., Singer, B.S., Cassidy, J., 2008. Laschamp and Mono Lake geomagneticexcursions recorded in New Zealand. Earth Planet. Sci. Lett. 268, 76e88.

Dong, H., Hall, C.M., Peacor, D.R., Halliday, A.N., 1995. Mechanisms of argon reten-tion in clays revealed by laser 40Are39Ar dating. Science 267, 355e359.

Elias, M., 1973. The Geology and Petrology of Mount Rouse, a Volcano in theWestern District of Victoria. Honours thesis. University of Melbourne, Mel-bourne, Australia.

Guillou, H., Singer, B.S., Laj, C., Kissel, C., Scaillet, S., Jicha, B.R., 2004. On the age ofthe Laschamp geomagnetic excursion. Earth Planet. Sci. Lett. 227, 331e343.

Hall, C.M., 2013. Direct measurement of recoil effects on 40Ar/39Ar standards. In:Special Publication e Geological Society of London, vol. 378. First publishedonline March 13, 2013, http://dx.doi.org/10.1144/SP378.7.

Hall, C.M., Higueras, P.L., Kesler, S.E., Lunar, R., Dong, H., Halliday, A.N., 1997. Datingof alteration episodes related to mercury mineralization in the Almadén dis-trict, Spain. Earth Planet. Sci. Lett. 148, 287e298.

Hall, C.M., Kesler, S.E., Simon, G., Fortuna, J., 2000. Overlapping Cretaceous andEocene alteration, Twin Creeks Carlin-type deposit, Nevada. Econ. Geol. 95,1739e1752.

Jicha, B.R., Singer, B.S., 2006. Volcanic history and magmatic evolution of SeguamIsland, Aleutian Island arc, Alaska. Geol. Soc. Am. Bull. 118, 805e822.

Koppers, A.A.P., Staudigel, H., Wijbrans, J.R., 2000. Dating crystalline groundmassseparates of altered Cretaceous seamount basalts by the 40Ar/39Ar incrementalheating technique. Chem. Geol. 166, 139e158.

Lee, J.-Y., Marti, K., Severinghaus, J.P., Kawamura, K., Yoo, H.-S., Lee, J.B., Kim, J.S.,2006. A redetermination of the isotopic abundances of atmospheric Ar. Geo-chim. Cosmochim. Acta 70, 4507e4512.

Ludwig, K.R., 2012. User’s Manual for Isoplot 3.75. A Geochronological Toolkit forMicrosoft Excel. Spec. Publ. No. 5. Berkeley Geochronology Center, Berkeley,California, p. 75.

Matchan, E., Phillips, D., 2011. New 40Ar/39Ar ages for selected young (<1 Ma) basaltflows of the Newer Volcanic Province, southeastern Australia. Quat. Geochronol.6, 356e368.

McDougall, I., Harrison, T.M., 1999. Geochronology and Theromochronology by the40Ar/39Ar Method, second ed. Oxford University Press, New York.

Nier, A.O., 1950. A redetermination of the relative abundances of the isotopes ofcarbon, nitrogen, oxygen, argon, and potassium. Physiol. Rev. 77, 789e793.

Ozawa, A., Tagami, T., Kamata, H., 2006. Argon isotopic composition of some Ha-waiian historical lavas. Chem. Geol. 226, 66e72.

Phillips, D., Matchan, E.L., 2013. Ultra-high precision 40Ar/39Ar ages for fish canyontuff and Alder Creek Rhyolite sanidine: new dating standards required? Geo-chim. Cosmochim. Acta 121, 229e239.

Renne, P.R., Deino, A.L., Hames, W.E., Heizler, M.T., Hemming, S.R., Hodges, K.V.,Koppers, A.A.P., Mark, D.,F., Morgan, L.E., Phillips, D., Singer, B.S., Turrin, B.D.,Villa, I.M., Villeneuve, M., Wijbrans, J.R., 2009. Data reporting norms for40Ar/39Ar geochronology. Quat. Geochronol. 4, 346e352.

Sharp, W.D., Renne, P.R., 2005. The 40Ar/39Ar dating of core recovered by the HawaiiScientific Drilling Project (phase 2), Hilo, Hawaii. Geochem. Geophys. Geosyst. 6.

Singer, B.S., Ackert Jr., R.P., Guillou, H., 2004. 40Ar/39Ar and KeAr chronology ofPleistocene glaciations in Patagonia. Geol. Soc. Am. Bull. 116, 434e450.

Singer, B.S., Pringle, M.S., 1996. Age and duration of the Matuyama-Brunhesgeomagnetic polarity reversal from 40Ar/39Ar incremental heating analyses oflavas. Earth Planet. Sci. Lett. 139, 47e61.

Singer, B.S., Relle, M.K., Hoffman, K.A., Battle, A., Laj, C., Guillou, H., Carracedo, J.C.,2002. Ar/Ar ages from transitionally magnetized lavas on La Palma, CanaryIslands, and the geomagnetic instability timescale. J. Geophys. Res. 107 (B11),20.

Steiger, R.H., Jäger, E., 1977. Subcommission on geochronology: convention on theuse of decay constants in geo- and cosmochronology. Earth Planet. Sci. Lett. 36,359e362.

Sutalo, F., Joyce, B., 2004. Long basaltic lava flows of the Mt. Rouse Volcano in theNewer Volcanic Province of southeastern Australia. Proc. Royal Soc. Vic. 116,37e49.

Trieloff, M., Falter, M., Buikin, A.I., Korochantseva, E.V., Jessberger, E.K., Altherr, R.,2005. Argon isotope fractionation induced by stepwise heating. Geochim.Cosmochim. Acta. 69, 1253e1264.

Turrin, B.D., Donnely-Nolan, J.M., Hearn Jr., B.C., 1994. 40Ar/39Ar ages from therhyolite of Alder Creek, California: age of the Cobb mountain normal-polaritysubchron revisited. Geology 22, 251e254.