Exploring the application of blue and red thermoluminescence for dating volcanic glasses

Tephras are significant markers in the stratigraphic record and play a key role in establishing paleoenvironmental and paleoclimate histories worldwide. Despite burgeoning research focused on tephra characterization and correlation techniques, there are still few techniques that allow for the direct dating of tephra, particularly below the lower age limit of K/Ar and Ar/Ar dating methods.

Radiation Measurements 153 (2022) 106731

Available online 26 February 2022

1350-4487/© 2022 Elsevier Ltd All rights reserved

Exploring the application of blue and red thermoluminescence for dating

volcanic glasses

aDivision of Earth and Ecosystem Sciences, Desert Research Institute, 2215 Raggio Parkway, Reno, NV, 89512, USA

bIllinois State Geological Survey, Prairie Research Institute, University of Illinois at Urbana-Champaign, Champaign, IL, 61820, USA

A R T I C L E I N F O

Keywords:

Luminescence

Tephrochronology

Thermoluminescence

Volcanic glass

TL

A B S T R A C T Tephras are significant markers in the stratigraphic record and play a key role in establishing paleoenvir-onmental and paleoclimate histories worldwide Despite burgeoning research focused on tephra characterization and correlation techniques, there are still few techniques that allow for the direct dating of tephra, particularly below the lower age limit of K/Ar and Ar/Ar dating methods In this study, we test different thermoluminescence (TL) dating approaches on the 4–11 μm volcanic glass constituents of three different independently different tephras By comparing against independent age control, we demonstrate the utility of both blue (320–450 nm) and red (587–651 nm) TL emissions for dating volcanic glasses using single aliquot regenerative (SAR) dose techniques We find that both blue and red TL emissions from the volcanic glass shards are dim but reproducible and show no evidence for significant sensitivity changes occurring between the natural TL and the first test dose during the SAR protocol Fading tests on the blue TL signal show that g-values range from 1.6 ± 1.0 to 2.9 ± 1.1%/decade and are statistically indistinguishable with zero at 2σ for the red TL Bleaching experiments show that both blue and red TL signals are sensitive to light exposure, with sensitivity corrected signals declining by

~40% over a 2-h period For all three tephras, both the fading-corrected blue and red SAR-TL ages are consistent with age expectation These successful results demonstrate the effectiveness of TL techniques for determining the eruption ages of tephra deposits in primary position between ~1 and at least 30 ka

1 Introduction

Tephrochronology is a geochronologic dating technique based on

geochemically identifying and correlating horizons of tephra erupted

from volcanic eruptions (Lowe and Alloway, 2015) In recent decades,

the application of tephrochronology has increased considerably in direct

response to advances in geochemical techniques used to ‘fingerprint’

individual tephras and correlate them with specific eruption events

(Lowe, 2011) Because tephras are deposited virtually instantaneously

(in geologic terms) and over long distances, they effectively represent

isochronic marker horizons within stratigraphic sequences (Lowe,

2011) If the age of a tephra layer is known, its occurrence in

strati-graphic sequences offers an opportunity to constrain the age of

brack-eting stratigraphic units and link spatially separated stratigraphic

sequences into an integrated chronological framework

Accurate tephrochronology requires that tephra deposits (a) have a

known age, (b) originate from a (geologically) short lived volcanic

eruption, and (c) can be fingerprinted and geochemically discriminated

from other eruption events (Lowe, 2011) A major obstacle in the application of tephrochronology is accurate age determination The age

of tephras is most commonly established either directly by radiometric methods (commonly K/Ar or Ar/Ar, e.g., Van den Bogaard, 1995), or fission track dating of primary mineral constituents for older deposits (≳100 ka, e.g., Seward, 1974) or indirectly by radiocarbon dating of

associated organic material for younger tephras (<50 ka, e.g., Benson

et al., 1997) For older tephras lacking in K-rich components, ages can also be indirectly defined by correlating paleomagnetic secular varia-tions with GISP2 ice core records (e.g., Benson et al., 2003), but these techniques rely on several untested assumptions about intra-hemispheric relationships between marine and continental records

Techniques that can accurately date tephra between the upper limits

of radiocarbon dating (~50 ka) and lower limits of K/Ar and Ar/Ar dating (~100 ka) are limited, but luminescence dating methods hold considerable promise Previous luminescence work on volcanic mate-rials has primarily focused on dating either (1) volcanic phenocrysts

* Corresponding author

E-mail address: krodrigues@nevada.unr.edu (K Rodrigues)

Contents lists available at ScienceDirect Radiation Measurements journal homepage: www.elsevier.com/locate/radmeas

https://doi.org/10.1016/j.radmeas.2022.106731

Received 1 December 2021; Received in revised form 15 February 2022; Accepted 23 February 2022

(Tsukamoto et al., 2007) or (2) xenoliths that have had their

lumines-cence signal reset upon interaction with lava flows (Schmidt et al., 2017)

or via phreatomagmatic explosions (Rufer et al., 2012) Fattahi and

Stokes (2003) and B¨osken and Schmidt (2020) summarize many

luminescence-based efforts for determining the ages of various volcanic

products Several studies have observed that the major limitation in

applying luminescence dating techniques to volcanic materials has been

anomalous fading, which can severely affect the accuracy of the

technique (e.g., Wintle, 1973; Tsukamoto and Duller, 2008) Addition-ally, the aforementioned approaches, which depend on the presence of phenocrysts or xenoliths, are not applicable for dating volcanic products lacking macroscopic mineral grains, particularly from deposits far from the eruptive source

With respect to the volcanic glass constituents of tephra, previous works have used blue thermoluminescence (TL) to produce age results with varying degrees of success between 0.5 ka to 400 ka (Berger and Huntley, 1983; Berger, 1985, 1987, 1991; Berger and Davis, 1992) TL dating of volcanic glass was not continued because of problems associ-ated with low signal intensities, anomalous fading, and poor inter-aliquot reproducibility resulting in large analytical errors up to 20–25% (e.g., Berger, 1991) Moreover, several TL ages had an apparent discrepancy with the established tephra ages and were considered inaccurate Since that time, the regional tephrochronology has been revised and are now in agreement with Berger’s (1991) age results with one exception: the Dibekulewe tephra (Redwine, 2013)

Beyond the work carried out by Berger, there have been few attempts

to apply luminescence dating to glass Several attempts to apply TL dating to archaeological glasses have provided unsatisfactory results (Sanderson et al., 1983), though some recent attempts on glass tiles (tesserae) have been more promising (Chiavari et al., 2001; Galli et al.,

2006, 2011) apparently owing to the relatively high degree of crystal-linity within the amorphous silica and/or to the presence of individual minerals dispersed within the glass Other amorphous natural materials including flint (amorphous or microcrystalline SiO2) have demonstrated

a long history of TL dating success (e.g., G¨oksu et al., 1974)

New technology and advances in technique development provide an opportunity to revisit the applicability of luminescence dating to vol-canic glass A successful volvol-canic glass TL dating approach would enable direct dating of tephras and allow for age determination of their asso-ciated eruption events In this paper, we expand on the work of Berger and others for TL dating of volcanic glass by refining previously suc-cessful techniques (blue TL), and testing the application of a new dating technique focused on red TL emission that has previously been observed for volcanic glasses with andesitic or dacitic composition (Kanemaki

et al., 1991)

2 Study area and sample characteristics

Tephra samples were collected from the Lahontan basin of Nevada—a sub-basin of the Great Basin that hosted a series of large lakes throughout the Quaternary (Fig 1; Russell, 1885) Three well-documented tephra outcrops were selected for sampling based on the following criteria: (1) thick (ideally 10+ cm), well preserved tephra beds with minimal visible evidence for reworking, thus reducing the possibility for post-depositional signal resetting, and (2) independently

Fig 1 Location map of the Lahontan basin showing the sample locations of the

Wono, Trego Hot Springs, and Turupah Flat tephras described in this study

Details about the sample location, and independent age control for each of the

tephra beds are outlined in Table 1 The thin grey line denotes the late

Pleis-tocene highstand of Lake Lahontan approximately 12.7 ka T = Lake Tahoe, P =

Pyramid Lake, W = Walker Lake

Table 1

Information about tephra samples collected for luminescence age determination

Turupah Flat Salt Wells (39 ◦ 23 ′ 15.36 ′′ N, 118 ◦ 38 ′ 3.48 ′′ W) 0.4 0.6–2.0 a Radiocarbon dating Wood (1977)

Davis (1978) Miller (1985) Sieh and Bursik (1986) Adams (2003)

Trego Hot

Springs Squaw Creek, southern amphitheater (40

◦ 47 ′ 16.56 ′′ N,

119 ◦ 30 ′ 30.88 ′′ W) 25.0 23.4 ±0.5 Radiocarbon, thermoluminescence dating Davis (1983) Berger (1991)

Benson et al (1997)

Wono North of Bunejug Mountains (39 ◦ 22 ′ 10.20 ′′ N,

118 ◦ 37 ′ 53.04 ′′ W) 3.5 27.3 ±0.3 Radiocarbon dating, paleomagnetic correlation to GISP2 Davis (1983) Benson et al

(1997) Zic et al (2002)

aAdams (2003) provides the geographically closest radiocarbon age of 650–920 cal years BP at Salt Wells

Radiation Measurements 153 (2022) 106731

dated with at least one type of dating technique, with ages that are

widely accepted in the regional tephrochronology The three selected

tephras were sourced from different eruptions along the Cascade Range

spanning ~1 to ~30 ka (Table 1) The tephras in this study have all been

identified as rhyolitic (>75% SiO2) on the basis of major and minor

elemental analysis of volcanic glass shards by electron microprobe

(Benson et al., 1997; Kuehn and Negrini, 2010; Bursik et al., 2014;

Pouget et al., 2014)

2.1 Turupah Flat tephra

The Turupah Flat tephra bed is comprised of a series of

geochemi-cally similar tephras dated between 0.6 and 2.0 ka years ago (Table 1,

Wood, 1977; Davis, 1978; Miller, 1985; Sieh and Bursik, 1986; Adams,

2003) The Turupah Flat tephra sampled in this study is exposed at Salt

Wells on the landward (south) side of a beach barrier that borders a

small playa in northern Nevada The beach barrier—originally mapped

as the Fallon 1 lake shoreline by Morrison (1964) and since referred to as

the Salt Wells beach barrier (Adams, 2003)—has been interpreted to

represent a lake stand that covered most of the Carson Sink (a remnant

of Lake Lahontan) during the late Holocene The Turupah Flat tephra is

thought to have been deposited through overwash processes on the

backside of the beach barrier at the time of this highstand (Davis, 1978;

Adams, 2003) The only local age constraint of the Turupah Flat tephra

at Salt Wells is based on a single radiocarbon age (650–920 cal years BP) determined from a charcoal sample collected immediately below the tephra layer within ~20 m of the site sampled in this study At the site selected for sample collection, the Turupah Flat tephra is ~5 cm in thickness and situated 0.4 m below the ground surface The tephra layer has an abrupt lower contact and grades upward into parallel laminae of silt and tephra throughout the upper 1 cm (Fig 2A) Adams (2003) documents the depositional environment, geomorphology, stratigraphic succession, and relevant geochronology at this site

2.2 Trego Hot Springs tephra

The Trego Hot Springs tephra is widespread throughout the Lahon-tan basin and has been independently dated to 23.4 ka (Table 1, Davis,

1983; Berger, 1991, Benson et al., 1997) For this study, the Trego Hot Springs tephra was sampled at a locality on Squaw Creek (the southern amphitheater of Davis, 1983), a site that has been studied thoroughly over the last 40 years (Davis, 1983; Benson et al., 1997; Adams, 2010) Here, Squaw Creek has exposed parts of a delta belonging to the Sehoo Formation of Morrison (1964) An ~12–15 cm thick exposure of the Trego Hot Springs tephra is situated near the base of the delta and grades upward into an ~30 cm thick siltier unit with prominent reworked

Fig 2 Pictures of sample sites showing tephra beds and their bounding sedimentary units (A) The Turupah Flat tephra at Salt Wells The arrow in the picture is 10

cm in length for scale (B) The Trego Hot Springs tephra at Squaw Creek (the southern amphitheater of Davis, 1983) Hammer is ~60 cm in length for scale (C) The Wono tephra near the Bunejug Mountains The arrow in the picture is 10 cm in length for scale

K Rodrigues et al

tephra (Fig 2B)

2.3 Wono tephra

Approximately 2 km southeast of the Turupah Flat tephra sampled at

Salt Wells is a prominent exposure of the Wono tephra, located at the

northern end of the Bunejug Mountains The Wono tephra has a well-

established age of 27.3 ka (Table 1, Davis, 1983; Benson et al., 1997;

Zic et al., 2002) At the site selected for sample collection, the Wono

tephra is interbedded with coarse beach gravels and has been

inter-preted to represent deposition in backset beds on the margin of a former

lake (Adams, 2010) At this location, the Wono tephra forms an ~15 cm

thick bed with abrupt upper and lower contacts (Fig 2C)

3 Methods

3.1 Sample collection and preparation

At each of the three sites, luminescence samples were taken directly

from the tephra bed by hammering steel tubes (for Wono and Trego Hot

Springs tephras: ~20 cm L x ~5 cm D, for Turupah Flat tephra: ~10 cm L

x 3 cm D) into freshly cleaned vertical sections The ends of tubes were

then wrapped to avoid light exposure during transport To account for

the heterogeneous gamma dose rate environments at each of the

lumi-nescence sample sites, sediment samples were collected from the upper

and lower bounding layers for dose rate assessment A third ‘average’

dose rate sample was collected from sediment within a 30 cm radius

surrounding the luminescence sample tube This third dose rate sample

was used exclusively to check for equilibrium conditions in the 238U

decay chain

Luminescence sample preparation was conducted at the DRI

Lumi-nescence Laboratory (DRILL) Preparation of the tephra for TL

mea-surement adapted the methodology from Berger (1991) Following this

protocol, the tephra samples were chemically treated to remove

car-bonates and organic material (10% HCl and 30% H2O2, respectively),

and the fine grained (4–11 μm) fraction was separated from the bulk

sample by extraction from suspension at appropriate settling velocity

times according to Stokes’ Law Volcanic glass was isolated from bulk

tephra using a solution of lithium heteropolytungstate in methanol

prepared with a specific gravity of 2.45 g/cm3 and centrifuged at 3000

rpm for 10 min The heavy liquid separation protocol was carried out a

minimum of two times with the float The effectiveness of the separation

technique was evaluated by visual inspection under both petrographic

and scanning electron microscope (SEM) at the University of Nevada,

Reno Microbeam Laboratory Panchromatic SEM-cathodoluminescence

(SEM-CL) was also applied to individual glass shards in effort to

char-acterize them further

3.2 Luminescence measurements

TL measurements on each sample were performed using two

different methods, each utilizing a different range of emission spectra for

measurement: blue TL and red TL Blue TL measurements on separated

volcanic glass were conducted using a Risø TL/OSL-DA-20 reader

housed at the DRILL with an integrated 90Sr/90Y β-source delivering a

dose rate of 0.10 Gy/s Blue TL signals were detected with an EMI

9235QA photomultiplier tube (PMT) after passing through a blue filter

pack (4 mm of Corning 7–59 in combination with 2 mm Schott BG 39)

Red TL measurements were conducted on a Lexsyg Smart TL/OSL reader

housed at the Illinois State Geologic Survey OSL Dating Lab with

a90Sr/90Y β-source delivering a dose rate of 0.074 Gy/s Red TL emission

was detected through a thermoelectrically cooled H7421-40

Hama-matsu PMT after passing through a combination of Chroma ET 620/60

and Schott KG 3 filters The beta sources for both systems were

cali-brated with 4–11 μm quartz prepared by Risø

All TL measurements were conducted in an N2 atmosphere and using

a heating rate of 2 ◦C/s The background TL was recorded in a second measurement on the same aliquot immediately after signal readout and subtracted channel-wise to obtain net signals TL measurements were made on 9.8 mm diameter stainless steel discs mounted with ~1 mg of sample by settling in acetone

3.3 Initial testing and measurement protocols

To define the thermally stable part of the TL glow curve and deter-mine an appropriate preheat temperature for experimentation, a plateau test was carried out on three aliquots of each sample by using the ratio of natural TL to the TL after laboratory bleaching (herein defined as heating to 450 ◦C to remove the signal) and subsequently administering

a β-dose (Aitken, 1985)

Standard SAR procedures were carried out following the methods of Murray and Wintle (2000) and incorporated a preheat of 200 ◦C for 10 min to eliminate the thermally unstable part of the TL glow curve prior

to TL readout, 4–5 regeneration doses bracketing the natural dose including a recycled dose and a zero dose, in addition to sensitivity correction with a 50 Gy test dose (Table 2) SAR-TL dose recovery tests were carried out for both blue and red TL to further test the reliability of the SAR protocol and determine the spread of the recovered doses Blue

TL dose recovery tests (Murray and Wintle, 2003) were carried out on 10 aliquots of each sample that were first bleached and then β-irradiated with a known dose approximating the natural equivalent dose (natural

De) Red TL dose recovery tests followed the same protocol but for only 5 aliquots of Trego Hot Springs

All previously published TL dating work on volcanic glass (Berger and Huntley, 1983; Berger, 1985, 1987, 1991; Berger and Davis, 1992) has employed a multiple aliquot additive dose (MAAD) approach using the methodology described in Aitken (1985) However, MAAD ap-proaches require significantly more prepared material and yield values

of De that are based on extrapolation and yield lower age precision relative to interpolative approaches like the single aliquot regenerative protocol (SAR, Murray and Wintle, 2000) In order to test whether a SAR approach would be appropriate for both blue and red TL measurements

on volcanic glass, a single-aliquot regeneration and added dose (SARA, Mejdahl and Bøtter-Jensen, 1997) approach was applied to assess for potential sensitivity changes occurring between natural and regenera-tive TL readout Owing to sample availability, the SAR-SARA experi-ments were only carried out for Trego Hot Springs The SAR-SARA experiments included four to five groups of aliquots which were given a different additive prior to the determination of De by our SAR-TL pro-tocol (Table 2) These SAR-TL measured doses were then plotted against the known added doses, and the value of De was obtained by a linear extrapolation of the data to the dose axis at the intercept A line with a slope of 1 was considered to reflect insignificant sensitivity change be-tween measurement of the natural and first regenerative dose The ‘Luminescence’ package (Kreutzer, 2021) for R was used to calculate De values using the SAR-TL protocol The appropriate TL in-tegral used for De calculation was determined by identifying a stable region showing a plateau in De values (Aitken, 1985) All results of TL

Table 2

The SAR-TL protocol applied to the samples in this study

1 Give dose, Di a

2 Preheat (200 ◦ C for 10 min)

3 TL measurement to 450 ◦ C at 2 ◦ C/s Lx

4 Background TL measurement to 450 ◦ C

5 Administer test dose (50 Gy)

6 Preheat (200 ◦ C for 10 min)

7 TL measurement to 450 ◦ C at 2 ◦ C/s Tx

8 Background TL measurement to 450 ◦ C

9 Return to step 1

a For measurement of the natural signal, i = 0

Radiation Measurements 153 (2022) 106731

measurements were required to pass the following criteria for further

analysis: >40 ◦C plateau range, <10% test dose error, <10% recycling

ratio error, <10% recuperation, and a signal > 3σ above background

Owing to the small spread in values of De, the central age model (CAM,

Galbraith et al., 1999) was used to calculate a weighted mean value of

De

To determine anomalous fading rates (g-values, Aitken, 1985) of the

TL signal in both the blue and red emission windows, β-irradiations with

doses similar to the natural De were administered to a series of aliquots

following laboratory bleaching Fading tests incorporated pauses of

varying duration (up to 40 h) after irradiation and prior to measurement

of the test dose corrected signal (Lx/Tx) following the approach of

Auclair et al (2003) For Lx/Tx measurements, TL signals were

inte-grated from ~260 to 300 ◦C for blue TL and ~230–270 ◦C for red TL

Anomalous fading tests for the blue TL signal were performed on five

aliquots of each of the samples Anomalous fading tests for the red TL

signal were only carried out on five aliquots of Trego Hot Springs

The optical bleachability of both the blue and red TL signal was

evaluated using aliquots that had been laboratory bleached, β-irradiated

with a dose similar to the natural De, and subsequently exposed to

natural sunny day light conditions for intervals of time ranging from 0 to

8.3 h Owing to sample availability, light exposure experiments were

only carried out on one sample: Trego Hot Springs Three aliquots were

measured for every light exposure interval The sensitivity corrected

signals (Lx/Tx) remaining after light exposure were measured using a

test dose of 50 Gy Like the fading tests described above, TL signals used

for Lx/Tx measurements were integrated from ~260 to 300 ◦C for blue

TL and ~230–270 ◦C for red TL

3.4 Dose rate measurement and age calculation

Samples for luminescence dating should ideally be collected from a

sedimentary deposit that is homogenous within a radius of at least ~30

cm (Aitken, 1998), but where sampling distal tephras are concerned, this

condition is rarely met Because of their short penetration distances,

alpha and beta dose rate contributions are likely best estimated using

subsamples of the original dating sample However, distal tephras are typically thin relative to the distance traveled by gamma rays, and contributions from both bounding sedimentary units and the tephra bed itself must be considered In this study, external dose rates were esti-mated from both (1) untreated sub-samples collected from the lumi-nescence sample tube for alpha, beta, and a fraction of the gamma dose rate, and (2) from representative samples collected from the bounding sediments above and below the luminescence dating sample which contribute to the modeled gamma dose rate

Dose rates were determined by measuring the elemental concentra-tions of U, Th, and K via ICP-MS/AES (from ~20 g of material) at the ALS facility in Reno, Nevada and then converted to dose rates based on the factors of Adamiec and Aitken (1998) Alpha and beta dose attenuation factors were based on Brennan et al (1991) and Gu´erin et al (2012), respectively The α-efficiency (a-value, Aitken and Bowman, 1975) of each sample was determined on three aliquots of each sample Following laboratory bleaching, the aliquots were α-irradiated with a known dose and then subsequently measured using the SAR-TL protocol outlined in Table 2 with β-irradiations for regenerative and test dose steps Irradi-ations were performed using a calibrated six-seater vacuum alpha irra-diation facility using 241Am alpha sources delivering ~0.14/μm2 min housed at the University of Washington Luminescence Lab A-values were only measured for the blue TL emission of each sample The frac-tional gamma dose rate contribution to the luminescence sample from both the tephra bed itself and the bounding sediment layers at each site were calculated as a function of their thickness following Aitken (1985, Appendix H) The cosmic dose rate contribution was calculated after Prescott and Hutton (1994) Internal dose rates were determined from the elemental concentrations of U, Th, and K from the isolated glass fraction of each sample (~5 g of material) Alpha, beta, and gamma dose rates were adjusted for estimated moisture content of the sample during the burial history (1 ± 5%)

To test for U-series disequilibrium in our samples, a representative portion of each sample (~20 g) was encapsulated in thin disk-shaped containers and sealed with two layers of epoxy gel A minimum wait-ing time of 21 days after sealwait-ing was observed to restore the radioactive

Fig 3 Images of Wono bulk tephra pre- (A) and post-volcanic glass separation (B) (C) and (D) are images of the same glass shard displaying a series of unidentifiable

elongated or acicular microlites under SEM (C) and as observed under cathodoluminescence (CL) Note the dim visible range luminescence emitted from these microlites under CL

K Rodrigues et al

equilibrium of 222Rn daughter products The specific activities (Bq/kg)

were measured with a broad-energy high-purity germanium detector in

a planar configuration, shielded by 15 cm of lead

4 Results and discussion

4.1 Volcanic glass characterization and assessment of the separation

protocol

Bulk tephra samples collected for this study were largely dominated

by the glass component, with minor (≲10%) contributions of quartz and

feldspar (either primary or detrital), lithics, and pumice fragments

Several glass shards featured small unidentifiable dark inclusions, as

observed under a standard optical microscope (Fig 3A) Petrographic and SEM examination of the separated volcanic glass component showed no observable non-glass grains (Fig 3B) Moreover, we did not observe the typical 110 ◦C UV TL peak that would likely be emitted in the presence of detrital quartz (Supplementary Fig 1) Taken together, these results suggest that the TL emission of our volcanic glass samples likely arises from the glass shards (and potentially the inclusions within them) as opposed to some other detrital component that was not adequately removed during separation

Approximately 10–15% of the 4–11 μm glass shards imaged dis-played prominent nano-to micro-scale crystalline inclusions with tabular, acicular, or skeletal crystal habits (Fig 3C) It is well docu-mented that volcanic glass shards frequently feature microlites and nanolites with compositions that can vary depending on the magmatic source (Sharp et al., 1996; Raymond, 2002) This contrasts with the observations made by Berger (1991) who reported that the fine-grained glass component was free from inclusions on the basis of petrographic examination Under panchromatic SEM-CL, several of the observable microlites appeared dimly luminescent in the visible spectrum in contrast to the glass matrix (Fig 3D) These microlites were too small to reveal chemical variations by energy dispersive X-ray spectroscopy elemental mapping or spot chemical analyses

4.2 Blue TL (320 nm–450 nm)

Blue TL signals arising from each of the measured volcanic glass samples were dim with peaks centered near ~300 ◦C (Fig 4A), consis-tent with the findings of Berger (1983, 1985, 1991, 1987) and Berger and Davis (1992) In those previous studies, a low temperature, long preheat (e.g., 50 ◦C over 8 days) was employed to improve the separa-tion of thermally unstable and stable components of the glow curve In our samples, we find that thermal instability is a prominent feature of the glow curves below ~200 ◦C and opt for the application of a 200 ◦C preheat for 10 min prior to TL measurement An example of background subtracted TL measured after preheating from a natural and artificially dosed aliquot of Trego Hot Springs is displayed in Fig 4A A zone of thermal stability indicated by a plateau in the ratio of natural and irradiated blue TL was prevalent from ~250 to 325 ◦C across all sam-ples When no initial preheat was performed, blue TL glow curves exhibited a broad peak spanning ~150–400 ◦C (Supplementary Fig 2), suggesting the presence of several overlapping (first order) peaks Blue TL dating of volcanic glass, like many TL-dating applications on heated materials, has routinely employed a multiple aliquot protocol However, an additive protocol requires several aliquots and therefore a significant amount of prepared material to carry out SAR protocols (Murray and Wintle, 2000) are preferable because they provide more precise estimates of De (and therefore ages) Berger and Huntley (1983) reported significant age underestimation using regeneration methods for dating volcanic glass, ostensibly owing to sensitivity changes be-tween natural and regeneration dose measurements Our results from SAR-SARA experiments on Trego Hot Springs yield a slope of 0.9, sug-gesting that sensitivity changes between the natural and first regener-ated dose are negligible while employing our SAR-TL protocol (Fig 3C)

We expect that the other samples studied here would behave similarly given their otherwise comparable luminescence characteristics Dose recovery tests performed on each sample indicate that the SAR protocol reproduces laboratory doses with reasonable accuracy (mean dose re-covery ratio of 0.9 ± 0.1, n = 22) Blue SAR-TL dose response curves (Fig 3B) were fit with a single saturating exponential

4.3 Red TL (587 nm–651 nm)

The red TL emission from our volcanic glass samples were dimmer than their blue TL counterparts and centered near ~250 ◦C, consistent with the findings of Kanemaki et al (1991) who similarly describe high temperature red TL emission at ~620 nm Our TL glow curves featured

Fig 4 Various features of the blue TL signal from volcanic glass (A) Shows

natural and regenerated glow curves for an aliquot of prepared volcanic glass

from Trego Hot Springs Red dots indicate the channel by channel ratio of the

natural to regenerated dose (B) The sensitivity corrected dose response of a

single aliquot of Trego Hot Springs with test dose response (inset) (C) Results of

the SAR-SARA protocol on Trego Hot Springs

Radiation Measurements 153 (2022) 106731

plateaus that extended from ~220 to 275 ◦C (Fig 5A) At temperatures

above ~325 ◦C a high thermal background was dominant, making it

difficult to extract a meaningful net signal Manually shifting the

back-ground TL by a single data channel (i.e 2 ◦C) improved black body

subtraction at high temperatures but had no appreciable impact on the

plateau region where the TL signal was integrated for calculation of De

(Supplementary Fig 3) We therefore carried out standard

channel-by-channel subtraction for all red TL glow curves

Results of the SAR-SARA protocol on Trego Hot Springs using the red

TL signal are displayed in Fig 5B Like the blue TL emission, a slope of

1.0 confirms that there is negligible sensitivity change between

mea-surement of the natural and first test dose (Fig 5C) The mean dose

recovery ratio carried out using the SAR-TL protocol for Trego Hot Springs is 1.0 ± 0.1 Red SAR-TL dose response curves were fit with a single saturating exponential (Fig 5B)

4.4 Light exposure tests

In principle, the TL signal from volcanic glasses should begin to accumulate shortly after cooling and solidification Provided that tephras are subsequently deposited within days to a few years (Lowe,

2011), the TL age should closely represent the timing of eruption However, if the electron trap storing the signal is optically bleachable, post-depositional reworking may reset or conflate the age of interest Though the samples collected for this study were considered primary with no indication of post-depositional reworking, we carried out a se-ries of tests to determine the sensitivity of both the blue and red TL signals of volcanic glass to light exposure

Both red and blue TL signals appear sensitive to light exposure and display a logarithmic decline in signal intensity over ~8 h of sunlight exposure with an average signal reduction of ~40% over a 2-h period (Fig 6) Within the reported analytical uncertainties, there is no appreciable difference in the bleaching rates between blue and red TL signals These results suggest that both blue and red TL dating of vol-canic glass can only accurately determine the timing of an eruption event if sampling is conducted on primary ashfall, or where reworking has occurred over short post-depositional time scales

4.5 Equivalent dose determinations

Results of the blue and red SAR-TL and SAR-SARA De measurements are compiled in Table 3 Values of De obtained using the SAR and SAR- SARA protocols for Trego Hot Springs are indistinguishable from each other within 1σ error uncertainty, further demonstrating that the SAR protocol accurately measures radiation doses stored in our volcanic glass samples Values of De generally display low spread with overdispersion values under 10% (Fig 7, Table 3) The observation of low inter-aliquot scatter suggests that the TL signal is dominated by an average lumi-nescence signal derived from many emitters at the aliquot level, but further work is needed to determine whether this is associated with discrete inclusions embedded within the glass shards, the glass matrix

Fig 5 Various features of volcanic glass red TL signals displayed for Trego Hot

Springs (A) shows natural and regenerated glow curves for an aliquot of

pre-pared volcanic glass Red dots indicate the channel-by-channel ratio of the

natural to regenerated dose (B) The sensitivity corrected dose response of a

single aliquot with test dose response (inset) (C) Results of the SAR-

SARA protocol

Fig 6 Blue and red TL signal loss as a function of bleaching time for Trego Hot

Springs (THS) Data was normalized to initial measurements made on three artificially irradiated, unbleached aliquots Bleaching was carried out using natural sunny daylight conditions for both blue and red TL Each data point reflects the average of measurements on three aliquots with errors that repre-sent the standard deviation

K Rodrigues et al

Table 3

A summary of red and blue TL equivalent doses, dose rates, and calculated ages Preferred ages are in bold

Sample Method Aliquots Accepted

(measured) De (Gy) Overdispersion (%) Total dose rate (Gy/ka) Average g-value (%/decade) Measured age (ka) Final age (ka) Expected age (ka) Turupah Flat Blue SAR-TL 11 (20) 12.0 ±

1.1 0 ± 0 13.1 ± 1.3 2.9 ± 1.1 0.9 ± 0.1 1.1 ±0.2 a 0.6–2.0 Red SAR-TL 9 (16) 14.4 ±

Trego Hot

Springs Blue SAR-TL 13 (20) 119.1 ±3.7 1 ± 5 5.5 ± 0.5 2.4 ± 1.0 21.7 ± 2.0 26.3 ±3.0 a 23.4 ± 0.5

Red SAR-TL 8 (15) 133.0 ±

Blue SAR-

Red SAR-

Wono Blue SAR-TL 20 (35) 135.4 ±

4.1 8 ± 3 5.9 ± 0.6 1.6 ± 1.0 22.9 ± 2.5 25.9 ±2.5 a 27.3 ± 0.3 Red SAR-TL 9 (13) 155.0 ±

aAges corrected for anomalous fading

Fig 7 Equivalent dose values displayed as kernel density estimate (KDE) plots RTL = red TL, BTL = blue TL

Radiation Measurements 153 (2022) 106731

itself, or both

4.6 Fading tests

Consistent with the findings of Berger (1985, 1987) and Berger and

Davis (1992) we find that the blue TL emission from our volcanic glass

samples exhibits anomalous fading with average g-values ranging from

1.6 ± 1.0 to 2.9 ± 1.1%/decade (tc =2 days) (Table 3, Fig 8) G-values

listed in Table 2 represent an average from 5 aliquots per sample These

average fading rates were used to correct the measured blue TL De values

following the method of Huntley and Lamothe (2001) The red TL

emission from Trego Hot Springs displays negligible fading over a period

of ~40 h (− 0.4 ± 1.2%/decade, tc =2 days) (Table 3, Fig 8) These

results are consistent with the work of Tsukamoto et al (2007) which

demonstrated that the blue emission from volcanic quartz fades, while

the red does not

4.7 Dose rate determination

Dose rates for the volcanic glass samples were typically dominated

by the external alpha and beta dose rates (together making up ~75–80%

of the total dose rate), with minor contributions from the gamma, cos-mic, and internal dose rate components (Table 4) With the exception of Trego Hot Springs, we find no evidence for disequilibrium in either the U

or Th decay chains that may contribute to inaccurate age estimation (see Supplementary Table 3) For Trego Hot Springs, the 226Ra/238U and

210Pb/226Ra activity ratios (both 1.2 ± 0.1, Supplementary Table 3) indicate minor U-series disequilibrium, but any resulting dose rate un-certainty is likely minimized given the high concentration of K in the sample

The blue TL derived a-values for these samples ranged from 0.18 to 0.21, in agreement with Berger (1991) who report a range from 0.08 to 0.36 Due to technical considerations, we were not able to carry out alpha efficiency measurements on the red TL signal This parameter might differ between TL emissions (e.g., for flint, Richter and Krbet-schek, 2006) A larger uncertainty (~33%) was incorporated on the a-value for red TL age calculation to account for the variability that may present between the different TL emissions Further work should be carried out to test for the differences in alpha efficiency between the TL emissions for volcanic glasses

4.8 Blue and red TL ages

Final age calculations are reported in Table 3 and incorporate beta and alpha dose rates determined from a sub-sample of the luminescence dating sample, and a gamma dose rate modeled from representative samples collected from the bounding sediments above and below the luminescence dating sample SAR-TL ages are fully consistent with in-dependent age control for both the fading-corrected blue TL and red TL signals Our SAR protocol applied regeneration doses across a limited

range (<400 Gy) which did not permit confident determination of the

saturation dose and thus the upper dating limit for these materials Previous applications of red TL on volcanic quartz have reported a significantly higher dose saturation level relative to conventional UV or blue TL approaches (e.g., Pilleyre et al., 1992; Miallier et al., 1994; Fattahi and Stokes, 2003) Additional work should be carried out to characterize saturation limits of both the blue and red TL emission from volcanic glass

Importantly, relative to red TL, blue TL ages are less precise (relative errors of ~14% vs 10%) owing to the additional errors that are intro-duced on the final age after fading correction The fading correction method proposed by Huntley and Lamothe (2001) is also limited to

young (<50 ka) samples that exhibit relatively low fading rates Despite

yielding statistically equivalent ages, we propose that red TL be used over blue TL particularly when dating older volcanic glass samples and/or if a higher level of precision is required on the age

Fig 8 Results of blue and red TL fading experiments for a series of Trego Hot

Springs (THS) aliquots Data is normalized to the first measurement Dashed

lines represent a logarithmic trend fit to the data G-values are normalized to

2 days

Table 4

Elemental concentrations and dose rates used for age calculation

Sample U a

(ppm) Th

a (ppm) K

a (%) a-value External Alpha Dose Rate (Gy/ka) External Beta Dose Rate (Gy/

ka)

External Gamma Dose Rate (Gy/ka) b Cosmic Dose

Rate (Gy/ka) Internal Dose Rate (Gy/ka) c Total Dose

Rate (Gy/ ka) d Turupah

Flat 6.3 ±0.6 19.0 ±1.9 3.7 ± 0.4 0.2 ±0.05 6.1 ± 1.2 4.3 ± 0.4 1.7 ± 0.2 0.3 ± 0.03 0.8 ± 0.3 13.1 ± 1.3 Trego Hot

Springs 2.3 ±0.2 5.3 ±0.5 2.7 ± 0.3 0.2 ±0.05 1.6 ± 0.4 2.6 ± 0.3 1.1 ± 0.1 0.03 ± 0.003 0.2 ± 0.1 5.5 ± 0.5

0.2 6.0 ±0.6 2.9 ± 0.3 0.2 ±0.05 1.6 ± 0.4 2.8 ± 0.4 1.1 ± 0.1 0.2 ± 0.02 0.3 ± 0.1 5.9 ± 0.6

aElemental concentrations provided here were determined from a subsample of the luminescence dating sample and used to determine external alpha and beta contributions

b Estimates of external gamma dose rate were determined on the basis of elemental concentrations of U, Th, and K from representative samples collected from both the bounding sediments above and below the luminescence dating sample and from the luminescence dating sample itself See Supplementary Table 1 for the elemental concentrations used in these determinations

cElemental concentrations used for internal dose rate calculation can be found in Supplementary Table 2

dDose rates assume a moisture content of 1 ± 5% for all samples

K Rodrigues et al

5 Conclusions

This study has demonstrated the potential for SAR-TL dating of

volcanic glasses Ages were measured from two distinct luminescence

color centers: blue and red Fading corrected blue TL and red TL ages are

in good agreement with independent age control for the three samples

that we studied Aside from imposing additional uncertainties on the

calculated age, the fading correction methods that we utilize for the blue

TL ages are also limited to young samples with low fading rates We

recommend that red TL be used over blue TL methods, particularly when

dating older volcanic glass samples and/or if a higher level of precision

is required on the age Our results show that both blue and red TL dating

of volcanic glass can accurately determine the timing of an eruption

event provided that the tephra bed is in primary depositional context or

has only experienced reworking over short post-depositional time scales

We have demonstrated that TL dating of volcanic glass offers an

accurate and direct dating approach that has the potential to fill the gap

between the upper range of radiocarbon dating and the lower range of

K/Ar and Ar/Ar dating techniques The results of this study are

encouraging and emphasize the need for continued research Future

work should investigate the saturation limits for both blue and red TL to

better define the effective dating range for these approaches At present,

the TL observations that we report cannot provide conclusive evidence

for the origin of the signal, but further work should seek to establish and

isolate a mineralogical source for both the blue and red TL emission

arising from volcanic glasses Additional efforts should also be made to

extend testing to compositionally different volcanic glasses

Declaration of competing interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence

the work reported in this paper

Acknowledgments

This work was funded through an NSF grant (EAR 2026019)

awar-ded to AKZ, a UNR Graduate Student Association Research Grant

awarded to KR and other in-house support from the Desert Research

Institute and Illinois State Geological Survey The authors thank Jim

Feathers for gracious use of the alpha irradiation facilities at the

Uni-versity of Washington KR thanks Ken Adams at the Desert Research

Institute for field assistance

Appendix A Supplementary data

Supplementary data to this article can be found online at https://doi

org/10.1016/j.radmeas.2022.106731

References

Adamiec, G., Aitken, M.J., 1998 Dose-rate conversion factors : update Anc TL 16,

37–50

Adams, K.D., 2010 Lake levels and sedimentary environments during deposition of the

Trego Hot Springs and Wono tephras in the Lake Lahontan basin, Nevada, USA

Quat Res 73, 118–129

Adams, K.D., 2003 Age and paleoclimatic significance of late Holocene lakes in the

Carson Sink, NV, USA Quat Res 60, 294–306

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

Aitken, M.J., 1998 Introduction to Optical Dating: the Dating of Quaternary Sediments

by the Use of Photon-Stimulated Luminescence Oxford University Press, Oxford,

p 267

Aitken, M.J., 1985 Thermoluminescence Dating Academic Press, London, p 359

Aitken, M.J., Bowman, S.G.E., 1975 Thermoluminescent dating: assessment of alpha

particle contribution Archaeometry 17, 132–138

Auclair, M., Lamothe, M., Huot, S., 2003 Measurement of anomalous fading for feldspar

IRSL using SAR Radiat Meas 37, 487–492

Benson, L., Liddicoat, J., Smoot, J., Sarna-Wojcicki, A., Negrini, R., Lund, S., 2003 Age of

the Mono Lake excursion and associated tephra Quat Sci Rev 22, 135–140

Benson, L.V., Smoot, J.P., Kashgarian, M., Sarna-Wojcicki, A., Burdett, J.W., 1997 Radiocarbon ages and environments of deposition of the Wono and Trego Hot Springs tephra layers in the Pyramid Lake Subbasin, Nevada Quat Res 47, 251–260

Berger, G.W., 1991 The use of glass for dating volcanic ash by thermoluminescence

J Geophys Res Solid Earth 96, 19705–19720

Berger, G.W., 1987 Thermoluminescence dating of the Pleistocene Old Crow tephra and adjacent loess, near Fairbanks, Alaska Can J Earth Sci 24, 1975–1984

Berger, G.W., 1985 Thermoluminescence dating of volcanic ash J Volcanol Geoth Res

25, 333–347

Berger, G.W., Davis, J.O., 1992 Dating volcanic ash by thermoluminescence: test and application Quat Int 13–14, 127–130

Berger, G.W., Huntley, D.J., 1983 Dating volcanic ash by thermoluminescence PACT 9, 581–592

B¨osken, J.J., Schmidt, C., 2020 Direct and indirect luminescence dating of tephra: a review J Quat Sci 35, 39–53

Brennan, B.J., Lyons, R.G., Phillips, S.W., 1991 Attenuation of alpha particle track dose for spherical grains Int J Radiat Appl Instrum Part Nucl Tracks Radiat Meas 18, 249–253

Bursik, M., Sieh, K., Meltzner, A., 2014 Deposits of the most recent eruption in the Southern Mono Craters, California: description, interpretation and implications for regional marker tephras J Volcanol Geoth Res 275, 114–131

Chiavari, C., Martini, M., Sibilia, E., Vandini, M., 2001 Thermoluminescence (TL) characterisation and dating feasibility of ancient glass mosaic Quat Sci Rev 20, 967–972

Davis, J.O.J., 1978 Quaternary tephrochronology of the Lake Lahontan area, Nevada and California Nev Archeol Surv Res Pap 7, 137

Davis, J.O., 1983 Level of Lake Lahontan during deposition of the Trego Hot Springs tephra about 23,400 years ago Quat Res 19, 312–324

Fattahi, M., Stokes, S., 2003 Dating volcanic and related sediments by luminescence methods: a review Earth Sci Rev 62, 229–264

Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M., 1999 Optical dating of single and multiple grains of quartz from Jinmium Rock Shelter, northern Australia: Part I, experimental design and statistical models Archaeometry 41, 339–364

Galli, A., Martini, M., Sibilia, E., Vandini, M., Villa, I., 2011 Dating ancient mosaic glasses by luminescence: the case study of San Pietro in Vaticano Eur Phys J Plus

126, 121

Galli, A., Poldi, G., Martini, M., Sibilia, E., Montanari, C., Panzeri, L., 2006 Study of blue colour in ancient mosaic tesserae by means of thermoluminescence and reflectance measurements Appl Phys A 83, 675–679

G¨oksu, H., Fremlin, J., Irwin, H., Fryxell, R., 1974 Age determination of burned flint by a thermoluminescent method Science 183, 651–654

Gu´erin, G., Mercier, N., Nathan, R., Adamiec, G., Lefrais, Y., 2012 On the use of the infinite matrix assumption and associated concepts: a critical review Radiat Meas

47, 778–785

Huntley, D.J., Lamothe, M., 2001 Ubiquity of anomalous fading in K-feldspars and the measurement and correction for it in optical dating Can J Earth Sci 38, 1093–1106

Kanemaki, M., Ninagawa, K., Yamamoto, I., Nakagawa, M., Wada, T., Yamashita, Y., Endo, K., 1991 Red thermoluminescence of volcanic glass fractions from tephras Int J Radiat Appl Instrum Part Nucl Tracks Radiat Meas 18, 81–88 Kreutzer, S., 2021 analyse_SAR.TL: Analyse SAR TL measurements Function version 0.3.0 In: Kreutzer, S., Burow, C., Dietze, M., Fuchs, M.C., Schmidt, C., Fischer, M., Friedrich, J., Mercier, N., Philippe, A., Riedesel, S., Autzen, M., Mittelstrass, D., Gray, H.J (Eds.), Luminescence: Comprehensive Luminescence Dating Data Analysis R package version 0.9.14, 2021 https://CRAN.R-project.or

Kuehn, S.C., Negrini, R.M., 2010 A 250 k.y record of Cascade arc pyroclastic volcanism from late Pleistocene lacustrine sediments near Summer Lake, Oregon, USA Geosphere 6, 397–429

Lowe, D.J., 2011 Tephrochronology and its application: a review Quat Geochronol 6, 107–153

Lowe, D.J., Alloway, B.V., 2015 Tephrochronology In: Rink, W.J., Thompson, J.M (Eds.), Encyclopaedia of Scientific Dating Methods Springer, Dordrecht,

pp 783–799

Mejdahl, V., Bøtter-Jensen, L., 1997 Experience with the SARA OSL method Radiat Meas 27, 291–294

Miallier, D., Fạn, J., Sanzelle, S., Pilleyre, Th, Montret, M., Soumana, S., Falgu`eres, C.,

1994 Attempts at dating pumice deposits around 580 ka by use of red TL and ESR of xenolithic quartz inclusions Radiat Meas 23, 399–404

Miller, C.D., 1985 Holocene eruptions at the Inyo volcanic chain, California: implications for possible eruptions in Long Valley caldera Geology 13, 14–17

Morrison, R.B., 1964 Lake Lahontan: Geology of Southern Carson Desert, Nevada U.S Government Printing Office, p 156

Murray, A.S., Wintle, A.G., 2003 The single aliquot regenerative dose protocol: potential for improvements in reliability Radiat Meas 37, 377–381

Murray, A.S., Wintle, A.G., 2000 Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol Radiat Meas 32, 57–73

Pilleyre, Th, Montret, M., Fain, J., Miallier, D., Sanzelle, S., 1992 Attempts at dating ancient volcanoes using the red TL of quartz Quat Sci Rev 11, 13–17

Pouget, S., Bursik, M., Cort´es, J.A., Hayward, C., 2014 Use of principal component analysis for identification of Rockland and Trego Hot Springs tephras in the Hat Creek Graben, northeastern California, USA Quat Res 81, 125–137