Radiation Measurements 33 (2001) 883–892www.elsevier.com/locate/radmeas

Limits to depletion of blue-green light stimulated luminescencein feldspars: implications for quartz dating

Mayank Jaina ; ∗, Ashok. K. SinghvibaDepartment of Geology, Delhi University, Delhi 110 007, India

bPhysical Research Laboratory, Navarangpura, Ahmedabad 380 009, India

Received 14 October 2000; received in revised form 15 March 2001; accepted 22 March 2001

Abstract

Feldspar contaminants in quartz aliquots, either as micro-inclusions or as remnant grains (due to inadequate etching) cana3ect the accuracy and precision of paleodose estimates based on blue-green light stimulated luminescence (BGSL). Suchcontamination could also alter the shape of the BGSL stimulation curve of otherwise pure quartz. In this study, the functionalrelationship between the infra-red stimulated luminescence (IRSL) and BGSL of feldspars, (1) at di3erent preheats, and(2) with IR bleaching at di3erent stimulation temperatures and durations, is examined. The results suggest two trap populationsparticipate in the feldspar BGSL process. These are: (1) Type (A) trap populations that can be stimulated by both the infra-redand the blue-green light at 125

◦C and, (2) Type (B) trap populations that respond only to blue-green-light stimulation at

125◦C. However, infra-red stimulation at elevated temperature (220

◦C) (ETIR) permits depletions of charges in Type (A)

and Type (B) to the extent that the feldspar BGSL can be reduced by up to 97% in 5 min.These results o3er prospects for (1) improved precision in paleodose estimates based on quartz; (2) BGSL dating of quartz

in a polyminerallic >ne grain samples; (3) age estimates based on both quartz and feldspars from the same aliquots, and(4) dating based on feldspar micro-inclusions. c© 2001 Elsevier Science Ltd. All rights reserved.

1. Introduction

Multiple aliquot optical dating of quartz mineral sepa-rates from sediments is often beset with large experimentalerrors on account of signi>cant variance in the lumines-cence output from identical aliquots (henceforth termed scat-ter; Jain et al., 1999). Such a scatter in conjunction withnon-linear >tting algorithms result in large errors on palaeo-dose, thereby limiting the use of luminescence methods inhigh precision dating (Felix and Singhvi, 1997). Lumines-cence dating of quartz is attractive because of its resistanceto weathering and the absence of athermal fading. Scatter

∗ Corresponding author. Present address: RisH National Labora-tory, NUK-125, P.O. Box 49, DK-4000 Roskilde, Denmark. Tel.:+45-4677-4909; fax: +45-4677-4959.

E-mail address: mayank.jain@risoe.dk (M. Jain).

in the blue-green stimulated luminescence (BGSL) outputof identically dosed quartz aliquot can arise from (1) pres-ence of feldspar micro-inclusions (Stokes, 1994) or feldspargrain remnants due to inadequate etching, (2) heterogeneousbleaching (e.g. Clarke, 1996; Huntley and Berger, 1995;Olley et al., 1998; Duller and Murray, 2000), and (3) het-erogeneity in dosimetry (Murray and Roberts, 1997). Scat-ter in regeneration growth curves may suggest an inherentvariability in the sample behaviour for example due to phasecontamination (e.g. Stokes, 1994) or di3erent dose-responseof each aliquot. Studies on quartz mineral separates from ae-olian and alluvial sediments from the Thar desert indicatedthat two signi>cant causes of scatter were: (1) contamina-tion by BGSL from feldspar micro-inclusions, and (ii) theinadequacy of the natural-normalization procedure (Jain andSinghvi, 1999).

An assessment of the magnitude of BGSL contributionfrom feldspars and its removal is important for quartz BGSL

1350-4487/01/$ - see front matter c© 2001 Elsevier Science Ltd. All rights reserved.PII: S1350 -4487(01)00104 -4

884 M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892

dating for the following reasons:

(1) The functional form of BGSL decay curves forfeldspars and quartz are di3erent (e.g. Banerjee et al.,2001).

(2) The e3ective meanlives of quartz and feldspar BGSL aredi3erent. Feldspars occasionally su3er from anomalousfading (Wintle, 1973) and=or long term fading (Mej-dahl, 1988). The dominant OSL trap in quartz may havea lifetime of 108 years (Murray and Wintle, 1999). Thisdi3erence in stability makes it desirable to isolate thetwo signals.

(3) The dose rates for quartz and feldspars are di3erent be-cause of di3erences in internal potassium content andalpha eIciencies (Aitken, 1985). This leads to di3erentpalaeodoses in feldspars and quartz separates from thesame sample.

(4) The BGSL growth curves of quartz and feldspars aredi3erent. Li (1991) demonstrated that signi>cant errorsin paleodose will accrue if signals with di3erent growthcharacteristics are treated as a single signal.

(5) The luminescence sensitivity of quartz and feldsparsare substantially di3erent. Given the fact that lumines-cence sensitivity of feldspars is invariably 1–2 ordersof magnitude higher than quartz, and the bulk OSL ina quartz aliquot can originate with only ∼5% quartzgrains (Duller et al., 2000), even minor (and randomlydistributed) feldspar contamination could be signi>cant.

The lower thermal stability of feldspar BGSL compared tothat of quartz implies that preheating may result in dispro-portionate relative contributions from feldspars and quartzin the preheated OSL of natural and irradiated samplesas compared to the natural normalization signal (Stokes,1994). Increase in scatter in luminescence output of identi-cal aliquots with increasing laboratory irradiation and a de-crease in scatter with the use of ‘dose’ normalization as seene.g. by Galloway and Hong (1996), and Perkins and Rhodes(1994), could thus be partly explained by the presence offaded feldspar contaminants.

Feldspar contamination in quartz grains is routinelydetected using infra-red stimulated luminescence (IRSL)from a very brief stimulation (short shine) of natural andregenerated aliquots. This is based on the premise thatquartz has a negligible response to infra-red stimulation atroom temperature (Spooner and Questiaux, 1989; Short andHuntley, 1992). However, in view of substantially higherluminescence sensitivity, even a low IRSL yield may implysubstantial contribution to the total BGSL from feldspars ina supposedly ‘quartz only’ sample. Further, as the measure-ment of BGSL is invariably made at 125

◦C, the contamina-

tion of feldspar BGSL is expected to be enhanced by thermalassistance (Duller et al., 1995; Duller and BHtter-Jensen,1993; Galloway, 1994).

Galloway (1994) and Duller and BHtter-Jensen (1993)demonstrated that the majority of traps which providefeldspar BGSL can be emptied by prolonged infra-red

(IR) stimulation. However, feldspar BGSL bleaches at aslower rate than quartz and an upper limit of ∼90% toBGSL depletion using a room temperature IR stimulationhas been suggested (Galloway, 1994). Removal of feldsparBGSL by IR bleaching has also been examined by Liritziset al., (1997) and Perkins and Rhodes (1994). While Liritziset al. (1997) demonstrated that prolonged room temperatureIR stimulation did not a3ect quartz BGSL, more recently,Bailey (1998) reported signi>cant BGSL depletion in hisquartz sample at high temperature (220

◦C) IR stimulation.

Banerjee et al. (2001) have recently studied the e3ect of IRstimulation at 20

◦C; 60

◦C; 125

◦C and 200

◦C on the blue

light stimulated luminescence from >ne grains. They reportabout 92% OSL depletion at 125

◦C IR stimulation and

∼98% depletion at 200◦C IR stimulation in 3600 s. Their

data after 200◦C IR stimulation could be >tted to three

exponentials with relative contributions of 57%, 38% and4% with the slow component coming either from quartz orfeldspar. Banerjee et al. (2001) used IR laser emitting at500 mW=cm2 and blue light (470± 30 nm) for stimulation.

Physical separation of quartz and feldspars to give mineralspeci>c BGSL is desirable but it is not always practicallyachievable in cases where feldspar occurs as micro-inclusion(Fragoulis and Stoebe, 1990; Fragoulis and Redhead, 1991)or in polyminerallic >ne-grain aliquots where etching mayfail to remove the remnant feldspar grains. In such cases,optical separation of quartz and feldspar BGSL is important.A major potential advantage of such a procedure is thatit may allow individual measurements on both quartz andfeldspar phases in the same aliquot. With a view to achievingthis objective, the present study examines: (1) relationshipbetween BGSL and IRSL in feldspars, and (2) the extent ofBGSL depletion in feldspars as a function of IR bleachingtemperature and its duration.

2. Methodology

All BGSL and IRSL measurements were made on105–150 �m size feldspar fractions extracted from (i)K-feldspar mineral specimen (OMS), and (ii) natural buriedsediment about 27 ka old (SH4). OMS was an orthoclasederived from a pegmatite vein (kindly supplied by the Ge-ological Survey of India). Gross checks on its purity weremade using thin sections, X-ray di3ractometry and scan-ning electron microscopy. The lump was gently crushedand sieved for preparing aliquots. Feldspars from SH4 wereextracted following a sequence of pretreatment with 1NHCl and 30% H2O2, sieving, Frantz magnetic separationand sodium-polytungstate heavy liquid separation of themineral fractions (�=2:58 g=cm3). Heavy liquid separationwas repeated twice with prolonged ultrasonic treatmentto ensure an optimum feldspar extraction. In view of thelikelihood of minor quartz contamination, a known amountof SH4 sample was etched with 40% HF for 60 min. Neg-ligible (∼1% by weight) residue remained which gave

M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892 885

Fig. 1. Flow chart showing experimental protocol.

BGSL only indicating that sediment extract sample wasdominantly feldspar.

Luminescence measurements were carried out in a RisHTL-DA 15 reader. Blue-green stimulation was achievedusing a tungsten–halogen lamp >ltered through GG-420and SWP 550 interference >lters. IR stimulation usedTEMT-484 diode array emitting at 880± 80 nm. Detectionoptics was deliberately kept identical to that used for quartzBGSL dating and comprised BG39 + 2 × U340 >lters.Laboratory beta irradiations used 40 mCi 90Sr=90Y betaplaque source (dose rate = 0:13 Gy=s). Nominal estimatesof the stimulation Pux of infrared and green-light are 40and 15 mW=cm2, respectively. Brief (0:3 s) stimulations byblue-green light and infra-red were made at 125

◦C.

3. Measurements

The experimental protocols are summarized in Fig. 1. Forboth OMS and SH4, depletion of 125

◦C feldspar BGSL after

IR stimulation at 25◦C; 50

◦C; 100

◦C; 150

◦C and 220

◦C

for di3erent durations was examined.Aliquots of OMS were bleached for 6 h in daylight prior

to receiving a 30 Gy beta dose. Feldspar extracts SH4 wereused with their natural dose. The samples were given briefstimulations (0:3 s) by infra-red and blue green lights at125

◦C with 10 s pause at the elevated temperature prior

to optical readout. Subsequently, a preheat of 220◦C for 1

and 5 min was given on separate sets and again 0:3 s IRSLand BGSL short shines were conducted which provided ameasure of the initial signal. All the calculations of signaldepletion after IR bleaching were made with respect to thissignal. Two preheat times (60 and 300 s) were used to assessthe e3ect of heating alone on the depletion of IRSL andBGSL. Following this, measurements were made on twoseparate sets in the manner detailed below (Fig. 1).

1. Single aliquot–multiple cycles: Individual aliquots werebleached by IR at a given temperature for 100 s, fol-lowed by IRSL and BGSL short-shines. This cycle was

886 M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892

repeated until a cumulative duration of 800 s IR bleachwas reached. The same measurement cycle was repli-cated for at least three aliquots.

2. Multiple aliquots–single cycle: Each aliquot experienceda single IR bleaching event. A set of aliquots (typi-cally four) was IR stimulated at a given temperaturefor 100, 300, 500 and 700 s, respectively. This proce-dure was repeated on other sets for di3erent tempera-tures (50

◦C; 100

◦C; 150

◦C and 220

◦C). Brief IRSL and

BGSL (0:3 s) were made after IR bleaching.

The multiple aliquot–single cycle experiment was carriedout to make sure that the signal depletion in single aliquot–multiple cycle was not an artifact of sensitivity change orrecuperation e3ects (Duller and BHtter-Jensen, 1993).

To estimate the depletion of IRSL and BGSL due to ther-mal erosion alone during high temperature IR bleaching, andto assess thermal contribution to short-shine measurements,experiments similar to single aliquot–multiple cycle proto-col were carried out on separate set of aliquots under twoconditions, i.e.,

(a) 0:3 s stimulation at 125◦C followed by isothermal heat-

ing for 100 s at di3erent temperatures; and(b) IR bleaching at various temperatures followed by

brief readouts at 125◦C in the absence of any optical

stimulation.

Experiment (a) ascertained the depletion of IRSL andBGSL due to temperature alone and experiment (b) assessedthermal contribution to brief (0:3 s) OSL measurements at125

◦C.

4. Results and discussion

4.1. IRSL vs. BGSL in feldspars

Brief (0:3 s) IRSL and BGSL intensities of di3erentaliquots of OMS, before and after 300 s preheat are plottedin Fig. 2. Since all the aliquots are derived from subsets ofidentically treated feldspar specimen, the IRSL and BGSLyields from di3erent aliquots are expected to co-vary asa function of the inter-aliquot di3erences in mass and=orsensitivity. The coupled IRSL, BGSL values for all thealiquots should, therefore, fall on a regression line throughthe origin and with a slope = BGSL=IRSL ratio of anyaliquot. The fact that regression lines have good correla-tion coeIcients (R2 = 0:89–0.96 for di3erent sets) and areconsistent with the origin at 99% con>dence level (Fig. 2)suggests that the above premise largely holds.

The intercepts and slopes of the regression line >tted toIRSL and BGSL data, and also the actual range of depletionin IRSL and BGSL, due to preheat are given in Table 1.

Signi>cant observations from Table 1 and Fig. 2 are:

(1) BGSL from this feldspar is more than 3 times IRSLunder the experimental conditions; and

Fig. 2. Relationship between brief BGSL and IRSL (0:3 s) mea-sured prior to (a) and after preheat (b) in sample OMS on same setof aliquots. Con>dence levels are de>ned at 99% intervals. Thereis an increase in slope of regression line after the preheat at 220

◦C

for 300 s. Both the >gures are drawn to same scale.

(2) There is an increase in slope (i.e. in BGSL=IRSL ratio)after the preheat, and this change is comparatively largerfor the less stringent preheat.

The results suggest two trap populations viz. (a) Type-Atraps that are less stable and respond to both IR- andBG-stimulations; and (b) more stable Type-B traps thatrespond to blue-green-light stimulation but are relativelyless sensitive to IR. Trap populations that remain una3ectedby IR bleaching have been observed in past (e.g. Li andAitken, 1989; Duller and BHtter-Jensen, 1993; Spooner,1993; Galloway, 1994). The increase in the BGSL=IRSLratio after preheat (Fig. 2) indicates that Type B traps havea greater thermal stability.

4.2. Rate of IRSL and BGSL depletion

Fig. 3 presents the relationship between residualBGSL and IRSL after IR exposure at di3erent bleaching

M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892 887

Table 1Intercept and the slopes of the regression line >tted to IRSL plotted against BGSL values. These are obtained from brief (0:3 s) stimulationsbefore preheat (BPH) and after preheat (APH) of 220

◦C for 60 and 300 s, respectivelya

Preheat Slope Slope Intercept Intercept Reduction- Reduction-time (BPH) (APH) (BPH) (APH) BGSL(%) IRSL(%)

60 s 2:1± 0:2 3:1± 0:3 75000± 45000 25000± 25000 49:0± 0:4 63:3± 0:9300 s 2:5± 0:08 3:0± 0:09 40000± 25000 9600± 9400 61:3± 0:5 66:6± 0:6

aThe values of percentage depletion in IRSL and BGSL are directly calculated from the ratios of OSL values after preheat to that beforepreheat and the standard error (SE) on at least 24 aliquots is quoted.

Fig. 3. Residual BGSL and IRSL (as percentage) after IR-bleaching at a given temperature. Each point on a curve corresponds to brief(0:3 s) IRSL and BGSL measurements following an IR bleaching event. (a)(i) Measurements made on K-feldspar mineral specimen (OMS);successive measurements along the curve are made in a single aliquot manner. (ii) Experimental conditions same as in (i) but performed onfeldspar sediment extract. (b) Measurements made on sample OMS; successive measurements along a curve are made in a multiple aliquotmanner.

temperatures. Figs. 3a(i) and (ii) show the results using thesingle aliquot–multiple cycle procedure (method 1, Fig. 1)for OMS and SH4, respectively. Here cyclic IR bleaching(at a given temperature) followed by short-shine IRSL andBGSL was carried out on a single aliquot (see Section 3for details). Fig. 3b shows results from sample OMS usingthe multiple aliquot–single cycle procedure (method 2, Fig.1; Section 3). In this procedure each aliquot experienced

a single IR bleaching event followed by short-shine IRSLand BGSL measurements.

The relative decrease in IRSL and BGSL depends on stim-ulation temperature during IR bleaching. For the >rst 100 s,a relative decrease in the slope from 30

◦C to 100

◦C (stim-

ulation temperature) indicates that IRSL depletes at a fasterrate with increase in temperature compared to BGSL (Fig.3a(i)). At 150

◦C there is no further decrease in the slope and

888 M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892

Table 2Remaining BGSL and IRSL percentages after IR bleaching for di3erent times at 30

◦C; 150

◦C and 220

◦C on feldspar sediment extracts.

Measurements are made in multiple aliquot manner (method 2, Fig. 2)

Time (s) 30◦C 150

◦C 220

◦C

IRSL (%) BGSL (%) IRSL (%) BGSL (%) IRSL (%) BGSL (%)

100 34:8± 1:3 42:2± 1:4 3:7± 0:2 21:4± 0:9 1:6± 0:8 10:7± 0:9300 23:6± 0:7 34:6± 1:5 0:8± 0:01 14:9± 0:8 0:3± 0:01 5:2± 0:6700 16:8± 0:5 30:2± 1:3 0:3± 0:02 11:5± 0:6 0:2± 0:02 3:9± 0:5

it overlaps with the 100◦C line. At 220

◦C, however, there

is an increase in the slope indicating a relative decrease inBGSL=IRSL ratio as compared to 100

◦C and 150

◦C stim-

ulation cycles. Sediment extract SH4 (Fig. 3a(ii)) and theOMS (Figs. 3a(i), 3b) exhibit similar trends but di3er some-what in the absolute values.

The most rapid BGSL and IRSL depletion occurs duringthe >rst 100 s of IR stimulation. Larger depletion of IRSLin comparison to BGSL occurs with increasing stimulationtemperatures (up to 100

◦C) and this possibly rePects in-

creasing thermal assistance to type A traps. BGSL has beenshown earlier to have smaller temperature dependence thanIRSL in feldspars (Galloway, 1994). Relative rates of IRSLand BGSL depletion are identical in the temperature interval100

◦C and 150

◦C, although there is greater absolute evic-

tion in the case of 150◦C IR bleaching. However, an abrupt

increase in the relative rate of BGSL depletion for the 220◦C

IR bleaching indicates that Type B traps are accessed moree3ectively perhaps due to greater thermal assistance and=orthermal erosion of Type B traps at 220

◦C in comparison to

150◦C (we are sure that there is no quartz contamination

in the sample OMS; Section 2). Relative roles of thermalerosion, and thermal assistance are discussed below.

Table 2 gives the results of residual IRSL and BGSLas percentages of the initial value for di3erent IR bleach-ing temperatures and times. Table 3 shows the percentageIRSL and BGSL depletion due to isothermal annealing cy-cles alone. It should be noted that each percentage value inTable 3 refers to BGSL and IRSL lost during isothermalheating with respect to BGSL and IRSL measured imme-diately prior to that particular annealing cycle. Comparisonof Tables 2 and 3 indicates that the >rst 100 s of 220

◦C IR

bleaching causes e3ectively∼27% more depletion in BGSL(and∼29% more depletion in IRSL) than could be achievedby a combination of room temperature IR bleaching fol-lowed by 220

◦C isothermal heating (or vice versa).

The di3erences in IRSL and BGSL depletion between>rst 100 s of 150

◦C- and 220

◦C-IR bleaching cycles also re-

quire a closer scrutiny. At 220◦C approximately 9:1± 0:8%

higher depletion in BGSL occurs due to the e3ect of temper-ature alone (Table 3; di3erence between columns 3 and 5).Due allowance for such a ∼9% thermal depletion increasesresidual BGSL at 220

◦C from∼11% (Table 2, column 7) to

Table 3Percentage depletion of BGSL and IRSL due to isothermal heatingat 150

◦C and 220

◦C in seven 100 s cycles. Measurements are made

in single aliquot mannera

Time At 150◦C At 220

◦C

(s)IRSL (%) BGSL (%) IRSL (%) BGSL (%)

100 4:2± 0:07 3:4± 0:0 11:1± 0:2 12:5± 0:8100 4:8± 0:2 3:6± 0:07 8:0± 0:3 9:5± 0:8300 3:5± 0:07 2:4± 0:09 6:4± 0:8 8:4± 0:7400 1:9± 0:09 4:7± 0:09 7:9± 0:4 6:2± 0:6500 3:2± 0:08 3:4± 0:08 7:2± 0:09 8:2± 0:8600 3:9± 0:09 2:8± 0:07 4:4± 0:3 7:8± 0:8700 5:3± 0:4 0:9± 0:02 4:7± 0:4 1:6± 0:09

aEach quoted percentage value is calculated with respect to thepreceding OSL value, measured prior to isothermal heating cycle.i.e. [%OSLi = 100 − (100 × OSL(i)=OSL(i−1))], where OSL(i)refers to the OSL measured after ith isothermal annealing cycleand OSL(i−1) refers to OSL measured prior to ith annealing cycle.

∼12% still implying a di3erence of ∼9% in the remainingBGSL between 150

◦C and 220

◦C IR bleaching. Possible

causes include thermal assistance or a complex multi-stepeviction of the Type B traps. Since IRSL depletes to a lowvalue (∼4%) during the >rst 150

◦C IR bleaching cycle, a

combination of thermal assistance and thermal erosion at220

◦C causes a greater eviction of Type B traps manifested

as a sudden decrease in the BGSL=IRSL ratio (Fig. 3).Analysis of the later parts of the curves in Fig. 3 shows that

the BGSL depletion continues approximately linearly andmore e3ectively at 220

◦C as compared to lower temperature

IR bleaching cycles. Since thermal depletion in the latercycles never exceeds 8% even at 220

◦C (Table 3), thermal

assistance should carry out this eviction.It is possible that single aliquot–multiple cycles mea-

surements (method 1, Fig. 1; Fig. 3a) may have a contri-bution from recuperation e3ects (Duller and BHtter-Jensen,1993), thermal depletion and=or sensitivity changes due torepeated higher temperature IR bleaching cycles. However,the 100

◦C; 150

◦C and 220

◦C IR bleaching cycles in Figs.

3b and 3a suggest no signi>cant di3erence in the percentageBGSL depletion by either single aliquot or multiple aliquotexperimental protocols.

M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892 889

Fig. 4. Percentage decrease in feldspar BGSL as a function of IRstimulation temperature and time. Each point on a curve representsthe BGSL signal remaining after preceding IR bleaching event ata given temperature expressed as percentage of the initial value.(a) Measurements made on K-feldspar mineral (OMS) sample;successive measurements along the curve are made in a singlealiquot manner. (b) Experimental conditions same as in (a) butperformed on feldspar sediment extract.

4.3. IR stimulation temperature, duration and BGSLdepletion

The e3ect of infra-red stimulation at di3erent tempera-tures on the BGSL of feldspar can be seen in Figs. 4a and busing the single aliquot approach on OMS and SH4, respec-tively (method 1, Fig. 1). Maximum BGSL depletion occursin the initial 100 s followed by decay of a slower compo-nent. With 220

◦C IR bleaching, BGSL in OMS decreases

by ∼97% after 300 s and about 99% after 800 s. In com-parison IR bleaching at 30

◦C decreases BGSL by ∼84%

in 300 s and by ∼91% in 800 s (Fig. 4a). BGSL depletionbetween the 2nd and 8th cycle is relatively slower duringIR bleaching at room temperature (29–9%) than at 220

◦C

(7–1%). Thermal depletion during high temperature bleach-ing cycles and the thermal contribution to BGSL measure-ments at 125

◦C (Fig. 1) are shown in Fig. 5. Thermal

contribution is calculated to be ¡ 0:001% in all the eightcycles. Slower depletion of BGSL at room temperature may

Fig. 5. Thermal contribution to a brief (0:3 s) luminescence mea-surement at 125

◦C (in the absence of IR or BG illumination), after

150◦C (>lled triangles) and 220

◦C (open squares) IR bleaching

cycles. The BGSL remaining after isothermal annealing is shownas >lled squares (150

◦C stimulations) and open triangles (220

◦C

stimulations), as a function of anneal time. This provides an esti-mate of the BGSL loss arising from thermal erosion.

be partly due to the presence of 150◦C PTTL peak after IR

stimulation. This peak has been shown to be BG sensitive(Duller and BHtter-Jensen, 1993).

In case of SH4 (Fig. 4b), BGSL depletion at the end of300 s is∼95% and∼66% for 220

◦C and 30

◦C IR bleaching,

respectively. This result using a feldspar sediment extracthas a direct relevance for sediment dating; it implies thatroom temperature infra-red cleaning for 300 s can still leaveup to∼34% residual BGSL. When considered together withthe inherently higher luminescence sensitivity of feldspars,this implies that even minor contamination of feldspars ina quartz samples could mask the quartz BGSL. In contrast,IR stimulation at 220

◦C for 300 s e3ectively removes up to

∼95% of the feldspar BGSL signal in sample SH4. It shouldbe noted that the estimates on BGSL depletion using thisfeldspar sediment extract provides only a lower limit to thefeldspar BGSL depletion, as, the (unlikely) possibility ofsome residual quartz contamination in the feldspar sedimentextract must be acknowledged.

Our results on K-feldspar extracts extend those of earlierstudies by Duller and BHtter-Jensen (1993), and Galloway(1994) using a detection window de>ned by BG 39 + 7-59>lters as compared to our detection window which usedBG-39+U-340 >lters. Duller and BHtter-Jensen (1993) re-ported BGSL depletion of about 65%, 80% and 87% after100, 1000 and 6000 s of room temperature IR stimulations,respectively, leading to the conclusion that a part of BGSLsignal was inaccessible by IR bleaching. Similarly, Gal-loway (1994) reported ∼10% residual BGSL signal evenafter IRSL depletion by more than 98%. For our sampleOMS, the present studies suggest that ¿ 96% reduction inIRSL at the end of 800 s room temperature IR bleaching

890 M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892

leads to ∼91% reduction in BGSL (Fig. 4a). Since in ourstudy, the brief (0:3 s) stimulations by IR and BG lights aremade at 125

◦C, whereas IR bleaching is done at room tem-

perature, it may be that the number of traps sampled duringthe brief stimulation is greater than that which is potentiallyaccessible during IR bleaching. However, this should leadto slower depletion of BGSL in comparison to the earlierresults (Duller and Botter-Jensen, 1994; Galloway, 1994),which is contrary to our observations. These di3erences inthe rates of BGSL depletion may imply a relationship be-tween BGSL depletion and the UV or blue detection spec-trum. This needs further investigations.

Summarizing, the present studies suggests that elevatedtemperature IR (ETIR) cleaning at 220

◦C can access all or

nearly all BGSL traps (Type B) in our feldspar samples.ETIR may have many applications especially in situationswhere it is diIcult to separate quartz and feldspar phasescompletely by chemical and=or physical treatments, and=orif readouts on separate phases are desired in the same aliquot.For example:

(a) In routine quartz dating, ETIR cleaning at 220◦C could

serve as a preheat for quartz and at the same time reducethe BGSL signal from feldspar contaminants, if any, to asmall value. A comparison of quartz OSL dose-responseafter (a) routine preheating of 220

◦C for 300 s after

room temperature IR cleaning (Figs. 6a and (b)) afterETIR cleaning (Fig. 6b), indicates that there is a sig-ni>cant reduction in scatter when the latter approach isused. This, however, assumes that quartz BGSL is notin itself a3ected by high temperature IR stimulation. Re-cently, Bailey (1998) reported a maximum of ∼12%quartz BGSL (at 20

◦C) depletion due to 220

◦C IR stim-

ulation for 300 s. Banerjee et al. (2001) reported 50%depletion after 60 s of IR stimulation at 200

◦C. It is,

however, diIcult to extrapolate the rate of BGSL de-pletion from Banerjee et al. (2001) to the present resultsas they used IR laser (∼500 mW=cm2). On the otherhand, Bailey (1998) used similar experimental condi-tions as ours. Unfortunately, there are little experimen-tal data available to determine if this e3ect is commonand in any case a loss of ∼12% of the quartz signal isprobably not very worrying. Further work is needed todetermine if this degree of signal loss is typical.

The fast component has been shown to comprise∼99% (e.g. Wintle and Murray, 1988) to up to ∼62%of the total BGSL (e.g. Banerjee et. al, 2001). Thisimplies that the depletion of quartz OSL due to ETIRbleaching by about ∼12% in 300 s (Bailey, 1998) maynot a3ect the paleodose calculation if a reasonablecounting statistics and a non-dose-dependent BGSLdepletion are assumed. On the contrary, near completedepletion of the feldspar OSL by ETIR in its >nal ef-fect would signi>cantly increase the S=N ratio in caseswhere feldspar contamination is signi>cant. Further, wespeculate that the e3ect of IR stimulation at tempera-

Fig. 6. Multiple aliquot quartz additive-dose growth curves afternatural normalization. (a) After 60 s room temperature IR cleaningfollowing preheat; and (b) after elevated temperature IR (ETIR)cleaning as a substitute for the preheat.

tures below 200◦C on the quartz OSL is likely to be

weaker, hence any IR cleaning procedure using stim-ulation temperatures below or at 200

◦C may be worth

closer scrutiny.(b) Dating of 5ne grained quartz: A combination of etching

(if feldspars to quartz ratio is high) and high temperatureIR cleaning could be used to obtain a clean BGSL signalfrom quartz for dating the >ne-grained sediments. Thiswould imply a more e3ective use of the easy bleachabil-ity of a quartz fraction in a polyminerallic sample froma complex depositional environment (Godfrey-Smith,1988).

(c) Multiple dates on polyminerallic samples: In mixedpolyminerallic samples quartz based BGSL can beobtained after the feldspar age. A sequence such as pre-heat followed by IRSL for 100 s (for feldspar growthcurve) + ETIR (as preheat for quartz and for remov-ing feldspar BGSL); again followed by 125

◦C BGSL

for 100 s (for quartz growth curve) can be used. Inpractice, the illumination and=or preheat timings wouldneed to be optimized according to the sample-speci>c

M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892 891

behavior with due care for the detection optics. Sucha protocol allows the calculation of a paleodose onboth quartz and feldspars from the same set of aliquots.It would also be interesting to obtain the palaeodoseon known mixtures of quartz and feldspar samples(Prasad, 2000) by this method and ascertain the levelof feldspar contamination, at which the feldspar BGSLbecome dominant even after high temperature IRcleaning.

(d) Feldspar micro-inclusion dating: Similarity in IRSL andBGSL depletion rates with IR bleaching at 100

◦C and

150◦C (Fig. 3) indicates that the rate of BGSL and IRSL

depletion, and hence the physical processes of trap evic-tion are similar at these two temperatures. As discussedearlier, thermal assistance of type A traps is impor-tant up to 150

◦C, and detrimental e3ects of temperature

occur only beyond 150◦C. Hence for these feldspars,

dating can in principle be carried out up to 150◦C stim-

ulation temperatures. Higher IR stimulation temperaturecan be used to increase the S=N ratio thereby allowingfeldspar micro-inclusion dating even at very low levelof feldspar contamination. Thus elevated temperature IRstimulations have de>nite advantages over room tem-perature stimulations (see also, Huntley et al., 1993).Interestingly, a paleodose plateau in feldspar has beenobserved using IR stimulation temperatures upto 110

◦C

(Wiggenhorn, 1994); for higher temperatures there wasa decrease in palaeodose due to supralinear growth usinga preheat of 220

◦C for 200 s. By using IR stimulation

at high temperature, it should also be possible to com-pletely eliminate the Type A traps and use Type B trapsfor paleodose evaluation. This would help to determinewhether these also show athermal fading or age underestimation e3ects as seen for e.g. in wind transportedloess samples.

5. Conclusions

The present study allows the following conclusions:

1. The relationship between IRSL and BGSL inK-feldspar as observed using UV detection spectra(BG-39+U340) suggests that depletion of both IRSLand BGSL varies directly as a function of preheatand stimulation temperature during IR bleaching. Asystematical increase in the BGSL=IRSL ratio occurswhen the temperature of IR bleaching varies from30

◦C to 100

◦C. Rates of BGSL and IRSL depletion

during 150◦C and 100

◦C IR bleaching are identical,

suggesting similar processes of trap eviction at thesetemperatures. However, IR bleaching at 220

◦C causes

a signi>cant decrease in the BGSL=IRSL ratio. Theseresults indicate two distinct trap populations; (1) TypeA, which is responsive to both infra-red and green lightstimulations at 125

◦C and Type B, which responds

only to blue-green-light stimulation at 125◦C. Type B

traps are thermally more stable than type A traps, andcan be probed by high temperature IR stimulation.From room temperature to 150

◦C, thermal assistance

of Type A traps is dominant, leading to a relative in-crease in BGSL=IRSL ratio. At 220

◦C IR stimulation,

thermal assistance or a complex multi-step evictionof Type B traps leads to an abrupt relative decreasein BGSL=IRSL ratio. Thermal erosion alone can notexplain the observed di3erences in BGSL and IRSLdepletions.

2. Infra-red bleaching at a stimulation temperature of220

◦C allows probing of deeper Type B BGSL traps,

which were hitherto inaccessible by room tempera-ture IR stimulations. This can result in depletion ofthe BGSL signal by about 95–97% in 300 s. Twodi3erent experimental protocols using single and mul-tiple aliquots, respectively, indicate that any possiblesensitivity changes or recuperation e3ects, due to re-peated high temperature IR bleaching cycles, do notsigni>cantly a3ect the percentage of BGSL depletion.High temperature IR cleaning could provide e3ectiveremoval of Type A traps and test if Type B traps alsoshow any fading characteristics.

3. Elevated temperature IR (ETIR) cleaning, usinginfra-red stimulation at 220

◦C for 5 min can be used

as a substitute for the standard preheat in quartz datingtechniques. However, its e3ect on the BGSL depletionin quartz requires closer scrutiny and this will be re-ported elsewhere. We anticipate that its e3ect shouldbe equivalent to a relatively more stringent preheat.

4. The fact that there is∼95% and∼97% depletion in theBGSL of sediment mineral specimens, respectively,after 300 s of IR bleaching at 220

◦C shows a potential

for (1) sampling pristine quartz BGSL; (2) BGSL dat-ing of >ne grained quartz; (3) simultaneous estimationof paleodose on both quartz and feldspars in polymin-erallic aliquots and (4) feldspar micro-inclusiondating. Evaluation of these proposals needs futureinvestigations.

5. The results of absolute BGSL and IRSL depletionsdepend on the stimulation Puxes, hence, the exactmagnitude of photon Puxes and depletions for a givenstimulation duration would depend on speci>c experi-mental con>gurations. Nevertheless, overall inferencesof this study would remain unaltered.

Acknowledgements

MJ acknowledges University Grants Commission (UGC)and Physical Research laboratory (Ahmedabad) for exper-imental and other logistical facilities. AKS thanks DSTfor support provided through coordinated Program No.ESS=CA=A3-08=92. Authors thank Praveen Mishra andNavin Juyal for their valuable inputs. Dr. A.S. Murray,

892 M. Jain, A.K. Singhvi / Radiation Measurements 33 (2001) 883–892

Prof. R. Galloway, and Prof. S.K.Tandon are thanked for acritical review of the manuscript.

References

Aitken, M.J., 1985. Thermoluminescence Dating. Academic Press,London.

Bailey, R.M., 1998. Depletion of quartz OSL signal using lowphoton energy stimulation. Ancient TL 16 (2), 33–36.

Banerjee, D., Murray, A.S., BHtter-Jensen, L., Lang, A., 2001.Equivalent dose measurement using a single aliquot ofpolymineral >ne grains. Radiat. Meas. 33, 73–94.

Clarke, M.L., 1996. IRSL dating of sands: bleaching characteristicsat deposition inferred from use of single aliquots. Radiat. Meas.26, 611–620.

Duller, G.A.T., BHtter-Jensen, L., 1993. Luminescence fromPotassium feldspars stimulated by infra-red and green-light.Radiat. Prot. Dosim. 47, 683–688.

Duller, G.A.T., BHtter-Jensen, L., Poolton, N.R.J., 1995.Stimulation of mineral speci>c luminescence from multi-mineralsamples. Radiat. Meas. 24, 87–93.

Duller, G.A.T., BHtter-Jensen, L., Murray, A.S., 2000. Opticaldating of single sand-sized grains of quartz: sources ofvariability. Radiat. Meas. 32, 453–457.

Duller, G.A.T., Murray, A.S., 2000. Luminescence dating ofsediments using individual sediment grains. Geologos 5, 87–106.

Felix, C., Singhvi, A.K., 1997. Study of non-linear lumine-scence-dose growth curves for estimation of paleodose inluminescence dating: results of Monte-Carlo simulations. Radiat.Meas. 27, 599–607.

Fragoulis, D.V., Redhead, M.L., 1991. Feldspar inclusions and theanomalous fading and enhancement of thermoluminescence inquartz grains. Nucl. Tracks Radiat. Meas. 18, 291–296.

Fragoulis, D.V., Stoebe, T.G., 1990. Relationship of anomalousfading to feldspar inclusion in quartz. Radiat. Prot. Dosim. 34,65–68.

Galloway, R.B., 1994. Comparison of green and infrared stimulatedluminescence of feldspars. Radiat. Meas. 23, 617–620.

Galloway, R.B., Hong, D.G., 1996. Concerning the normalisationof additive dose optically stimulated luminescence data fromquartz. Ancient TL 14, 1–5.

Godfrey-Smith, D.I.G., 1988. Optical dating studies of quartz andfeldspar sediment extracts. Quat. Sci. Rev. 7, 373–380.

Huntley, D.J., Hutton, J.T., Prescott, J.R., 1993. Optical datingusing inclusions within quartz. Geology 21, 1087–1090.

Huntley, D.J., Berger, G.W., 1995. Scatter in luminescence datafor optical dating—some models. Ancient TL 30, 5–9.

Jain, M., Singhvi, A.K., 1999. Scatter and normalisation in quartzOSL: evaluation of current procedures and development of anew protocol. Ninth International Conference on Luminescenceand Electrospin Resonance, Rome, Book of Abstracts, p. 83.

Jain, M., Tandon, S.K., Bhatt, S.C., Singhvi, A.K, Mishra, S., 1999.Alluvial and aeolian sequences along the river Luni, Barmerdistrict: physical stratigraphy and feasibility of luminescencechronology methods. In: Radhakrishna, B.P., Merh, S.S. (Eds.),Vedic Saraswati, Geological Society of India Memoir, Vol. 42,pp. 273–295.

Li, S.-H., Aitken, M.J., 1989. How far can we go back withthe optical dating of the K-feldspars. Long and Short RangeLimits in Luminescence Dating. The Research Laboratoryfor Archaeology and the History of Art, Oxford University,Occasional publication, Vol. 9, pp. 41–46.

Li, S.-H., 1991. A cautionary note: apparent sensitivity changesresulting from curve >tting. Ancient TL 9, 12–13.

Liritzis, I., Galloway, R.B., Hong, D.G., 1997. Single aliquot datingof ceramics by green-light stimulation of quartz. Nucl. Instrum.Meas. Phys. Res. B 132, 457–467.

Mejdahl, V., 1988. Long term stability of the TL signal in alkalifeldspars. Quat. Sci. Rev. 7, 357–360.

Murray, A.S., Roberts, R.G., 1997. determining the burial time ofsingle grains of quartz using optically stimulated luminescence.Earth Planet. Sci. Lett. 152, 163–180.

Murray, A.S., Wintle, A.G., 1999. Isothermal decay of opticallystimulated luminescence in quartz. Radiat. Meas. 30, 119–125.

Prasad, S., 2000. HF treatment for the isolation of >ne grain quartzfor luminescence dating. Ancient TL 18, 15–17.

Perkins, N.K., Rhodes, E.J., 1994. Optical dating of Puvialsediments from Tattershall, UK. Quat. Geochr. 13, 517–520.

Olley, J., Caitcheon, G., Murray, A., 1998. The distribution ofdose as determined by optically stimulated luminescence insmall aliquots of Puvial quartz: implications for dating youngsediments. Quat. Geochr. 17, 1033–1040.

Short, M.A., Huntley, D.J., 1992. Infra-red stimulation of quartz.Ancient TL 10, 19–21.

Spooner, N.A., 1993. The validity of optical dating based onfeldspar. Unpublished D.Phil Thesis, Oxford University, Oxford.

Spooner, N.A., Questiaux, D.G., 1989. Optical dating—Achenheimbeyond the Eemian using green and infra-red stimulations. Longand Short Range Limits in Luminescence Dating. Occassionalpublication, Vol. 9, The Research Laboratory for Archaeologyand the History of Art, Oxford.

Stokes, S., 1994. Optical dating of selected Late Quaternary aeoliansediments from Western United States. Ph.D thesis, Universityof Oxford, Oxford.

Wiggenhorn, H., 1994. IRSL dating of K-feldspar at elevatedtemperature and infrared bleaching of TL. Radiat. Meas. 23,387–391.

Wintle, A.G., 1973. Anomalous fading of thermoluminescence inmineral samples. Nature 245, 143–144.

Wintle, A.G., Murray, A.S., 1988. Towards the development of apreheat procedure for OSL dating of quartz. Radiat. Meas. 29,81–94.