October 13, 2000 ver. 2
From: Shannon Mahan
To: Prospective Thermoluminescence (TL) and Optically Stimulated Luminescence (OSL) users
Thank you for your interest in using the U.S. Geological Survey TL/OSL lab for sediment dating. In this information “booklet” there are several sections which you will find useful. Section I covers general TL and OSL methodology and assumptions, Section II covers dose rate (DR) calculations and how the elements are measured quantitatively, while Section III guides the reader through lab preparation of a sample and subsequent measurement of equivalent dose (DE).
Section IV reviews sampling in the field for TL and OSL and other material that may be dated using luminescence. Section V introduces quick and general facts about the technique range, explanations for differences in TL and OSL techniques, reasons why different luminescent techniques are applied to different depositional sample types and other labs to contact in the U.S. Some labs, other than the USGS lab, may be more helpful for archeological applications, may respond with a cheaper price or offer faster response time depending on the need for a particular analysis.
Section I: General TL and OSL methodology (and assumptions)
If a sample of sediment is heated rapidly to 500°C, there is weak but measurable emission of light. This light is known as thermoluminescence (TL) and is based on time-dependent accumulation of radiation damage in minerals. Optical Stimulated Luminescence (OSL) is measured by shining a beam of light onto mineral grains and measuring the resulting luminescence back. A low level of ionizing radiation, which we measure from 40K, 238U, 235U, 232Th and daughter products, 87Rb and cosmic rays is omnipresent in nature. Almost exclusively, luminescence from quartz and feldspar grains is used in dating.
The interaction between this radiation and the atoms of minerals results in gradually increasing radiation damage. The intensity of the radiation damage in crystal lattices is a measure of the Equivalent Dose (DE) which the mineral has received since formation or last “resetting” by exposure to sunlight or heat over 300°C. The mineral is used as a natural dosimeter. DE is measured in Gray (Gy) or absorbed radiation energy per unit mass. Once one has “read” the DE by means of a TL or OSL measurement, the Dose Rate (DR) is obtained via measurement of K, U, Th, Rb and cosmic rays as dose per unit time or Gy/ka. The equation for obtaining an age is:
Age (ka) = DE (Gy)/DR (Gy/ka)
To evoke the emission of TL and OSL, additional thermal or optical stimulation, respectively, has to be supplied to the crystal. Luminescence is created by the ionizing radiation freeing electrons which wander through the crystal lattice until they encounter a trapped hole or recombine and become trapped at electron traps which are lattice defects with negative charge deficits. During luminescence measurement, “traps” are emptied and luminescence centers destroyed. The longer the crystal has been exposed to ionizing radiation, the more “traps” can be filled, resulting in an increased luminescence signal, with a practical time limit of 800 ka (Berger, G.W., 1994, Thermoluminescence Dating of Sediments Older than 100 ka, Quaternary Geochronology, Quaternary Science Reviews, 13, 445-456).
Key assumptions are:
1). Materials have uniform and definable dose rates.
2). Moisture content of the sample and its environment can be determined.
3). Depth, altitude and intensity of cosmic rays on site can be calculated or are known.
4). The radiation-induced signal has to be thermally or optically reset by the event to be dated. The rate and completeness of “resetting” can be reliably obtained.
5). The TL or OSL must have been stable during the time span in question. Any spurious “fading” of TL or OSL can be measured and compensated for in age calculations.
6). The TL and OSL growth characteristics have to follow a mathematical function.
The principal minerals used in OSL are quartz and K feldspar, although use of volcanic glass, some forms of calcite and zircon have been used. The principle minerals used in TL are polymineralic or separation of the quartz or feldspar. M.J. Aitken (Thermoluminescence Dating, 1985, Academic Press, 359 pages and An Introduction to Optical Dating, 1998, Oxford University Press, 267 pages) presents a very able discussion of the complete TL and OSL techniques.
Section II: Dose Rates and how they are obtained
The rate at which trapped electrons are accumulated is proportional to the energy absorbed during burial. Several components are needed for an accurate DR; 1). measurement of K, U, Th and Rb 2). calculation of moisture content in field at time of collection and saturation potential of the sample sediment and 3). cosmic ray component calculation.
DR’s are taken from the combination of material around each site to be dated. These include neutron activation (INAA), atomic absorption, X-ray fluorescence (XRF), flame photometry (K only) and inductively coupled plasma mass spectrometry (ICP-MS). The fundamental disadvantage of these methods is they do not account for U and Th being in disequilibrium.
Equilibrium can be checked by using alpha spectrometry or high resolution gamma spectrometry since these methods measure the activities of several individual radionuclides in the decay chains. Usually the facilities required for these techniques are expensive and measurement time can be long.
Another routine approach that minimizes possible error due to disequilibrium is thick source alpha counting (TSAC) which determines the Th and U and the K as above. TSAC can also be used to count only for alpha contribution and the beta contribution can be determined by a beta TLD or particle counter (high sensitivity TL dosimetry phosphor).
The method in use at the USGS lab is gamma ray spectrometry, due to an inherited collection of NaI crystals from radioactive chemistry labs. High-resolution gamma-spectrometry is carried out on a 600-g sample (the ideal weight since less material means higher counting errors); admittedly a lot of sample, but the crystals date from the early 1980’s. However, this amount of sample more adequately represents the variations and mix that the collected sample might have absorbed during deposition, rather than relying on the one or two gram quantity normally counted in other techniques. Apart from elemental quantification, this technique enables the checking of radioactive equilibrium (see assumption #1). The samples are measured on four (4) different crystals for ten (10) hours per crystal.
Moisture and radon migrations are not factors because the bulk sample has been dried and sealed for a month and radon allowed to equilibrate before counting. We collect gamma spectra and then fitted to standard spectra of K, U, and Th using the least square criterion. Elements measured in standard materials provide quality assurance as they run alongside the unknown samples. Comparisons between the USGS luminescence laboratory and published NBS (now NIST) and USGS Open-Files literature values for these standards show excellent agreement. Most of the scatter can be attributed to counting statistics.
Occasionally, the lab will compare the K concentration of samples using X-ray fluorescence (XRF). General agreement, within 1 sigma, provides another check on secular equilibrium. If the lab suspects the sample to be high in Rb, this element will be analyzed on the XRF as well.
It is most desirable to measure the gamma dose-rate on-site. This is so that if there is any doubt about uniformity of radioactivity within the 30-cm sphere of influence of the surrounding sample, the readings will show such variations, even if a laboratory high-resolution gamma spectrometer is available to count after collection. The USGS lab keeps a portable NaI crystal for this purpose and will loan it out if the USGS scientist is unable to collect the samples. The sample site should be counted for an hour or more, to provide high resolution and account for present field moisture or large stones. The lab crystals are unable to reproduce moisture conditions or stone placement, since the sample is usually dried and sieved before measurement.
The annual DR to the sample is computed from the concentrations of K, U and Th by the method described in several literature sources, Aitken, 1985 (Thermoluminescence Dating, M.J. Aitken, 1985, Academic Press, London, p. 10-12 and p. 282-288). The lab initially assumes secular equilibrium in the U decay series, unless comparisons in randomly picked samples show otherwise. The measurements must now include alpha, beta, and cosmic radiation and factor in the reduced effect of alpha radiation relative to gamma and beta radiation. Annual radiation doses in Gy/ka taken from Aitken, 1985 are adapted as shown in Table 1.
Table 1 |
|||
|
Alpha |
Beta |
Gamma |
|
|
1% Rb |
----------- |
4.00 |
-------- |
|
1% K |
----------- |
0.83 |
0.24 |
|
1 ppm U |
2.78 |
0.15 |
0.11 |
|
1 ppm Th |
0.74 |
0.03 |
0.05 |
A reasonable estimate is made of the moisture content through geologic time with the understanding that this estimate carries a large uncertainty. Ages are calculated using field moisture percentage, unless unusual conditions prevailed at the time of sample collection, i.e. sustained rain over a period of time, drought or human disturbance. Water attenuation corrections for each type of radiation are made using moisture correction factors taken from Aitken, 1985. The actual equations used can be found in USGS Open File Report 94-249 p.18-21.
In comparison with silicates, water has a significantly higher mass absorption coefficient for alpha, beta and gamma rays, but negligible radioactivity; hence it more or less attenuates the DR and can significantly change the radiation a sample may have absorbed. Therefore, the lab also calculates ages based on the sample being fully saturated. In the final report there will be listed a “halfway” value; that is the sample moisture content may have had dry and wet frequencies during deposition of unknown duration, but somewhere between conditions at time of collection and a fully saturated value. There will be three ages listed for each sample for each technique, each age corresponding to a different water moisture value. It is the client’s prerogative to quote which age best fits their carefully researched scenario.
The lab then calculates the dose rates for element/radiation combination using the annual radiation doses in Table 1, the corrections for attenuation of water, and the alpha k-effective value (0.10 ± 0.03). (Equations can be seen in USGS Open File Report 94-249 p. 18). The dose rate for the various element/radiation combination are combined to give the dose rates for the radioelements as follows:
Table 2 |
|||
|
Alpha |
Beta |
Gamma |
|
|
Total DR Rb |
----------- |
DR Rb |
-------- |
|
Total DR K |
----------- |
DR K |
DR K |
|
Total DR U |
DR U |
DR U |
DR U |
|
Total DR Th |
DR Th |
DR Th |
DR Th |
The final component of cosmic ray value is added now to the dose rate calculations. The value, 0.291 Gy/ka, is the cosmic ray dose at sea level and latitude 38° S and is taken from Prescott and Hutton, 1988. This value is valid for latitude greater than 40°, but must be corrected for the sample elevation above sea level and depth within the sediment. Aitken, 1985 p. 298 presents a graph of the elevation factor vs. elevation for different latitudes. The low elevation portion of the curve (about 3,000 m or 9,000 ft.) is approximately linear with slopes shown in Table 3.
Table 3 |
|
|
Latitude |
Slope |
|
>40° |
8.2-5/foot |
|
25° |
6.1-5/foot |
|
0° |
4.0-5/foot |
Values for the depth factor are taken from Prescott and Hutton, 1988 and may be read from USGS Open File Report 94-249 p. 19. Finally, the total DR of the sample is computed using DR of cosmic ray, DR of Rb, DR of U, DR of Th and DR of K.
Errors are calculated in a separate program to 2 sigma, using standard formulas from Taylor, 1982 (An Introduction to Error Analyses, University Science Books, 1982, pp. 148-150).
Section III: Lab prep and equivalent dose (DE) analyses
A full explanation of USGS lab procedure may be read in USGS Open File Report 94-249. The USGS lab uses three grain sizes: a polymineralic 4-11 µ fine silt for both TL and OSL dating, 90-125 µ quartz for OSL (Blue Light-Optically Stimulated Luminescence) and 180-125 µ or 90-125 µ (depending on quantity and difficulty in calculating DR) k-feldspar for IRSL-OSL (Infrared Stimulated Luminescence). Before sieving for desired size fractions, each sample is treated with 4N HCl to remove carbonate, 30-35% H2O2 to destroy organic material and is dispersed in Na pyrophosphate solution. Wet sieving and Stoke’s settling in de-ionized water are employed to achieve desired size fractions. Heavy liquid separation is used on the larger grains to separate quartz from feldspar by using lithium polytungstate (LST). Short sample etches in HF for the quartz fraction may be used to clean surfaces.
The silt size portion of the sample (4-11 µ) are then plated on 1 cm size aluminum disks from a grain suspension in methanol, while the larger fraction (90 µ and larger) are poured loosely into cups. All procedures are performed under reduced lighting conditions to minimize artificial bleaching of the samples. The intensity of the sodium vapor lights in the laboratory, which emit a single wavelength (589 nm), are adjusted to provide enough requisite sensitivity for the human eye but not bleach the samples.
The silt size sample disks are normalized for sample homogeneity by a 5-second exposure to infrared illumination and detection of the resulting luminescence by a photomultiplier tube covered with Schott BG-39 and Kopp 7-59 filters, as well as a Pyrex window. The TL signal is reduced by about 2% by this treatment and the IRSL signal by about 1%. The larger size fraction are measured using a single aliquot technique and thus require no normalization.
Samples are measured for TL and OSL separately, which requires the use of many disks or cups. TL is obtained by heating the disk to 500°C and the electronically intensified light emission is recorded as a function of the heating temperature. After each measurement the heating is repeated to record and subtract background. Most of the signal is due to TL from quartz and feldspars. The TL instrument records the signal as photons/5°C intervals and plots the resulting glowcurve as TL signal vs. temperature. The glows are actually a composite of TL peaks from different traps. It is crucial to conduct TL measurements under the purest nitrogen (or argon) atmosphere to prevent unwanted chemiluminescence, which is energetically fed from exothermic chemical reactions.
Silt size, polymineralic IRSL is obtained by exposing the disk to a 30mA current to infrared LED’s, 1 second dwell time per channel for 100 channels, 100 second total time exposure, sample temperature held to 30°C and a background count taken before and after a set of samples is exposed under the photomultiplier tube. (It is not necessary to obtain IRSL under nitrogen atmosphere). The coarser grains measured on the Riso System are done so under a steady light from either blue or infrared diodes. The blue light emission is given in the 400-550 nm range (centered at 470 ± 30) and infrared emission is from the 800-900 nm range (centered on 880 ± 20 nm).
Optical filters are inserted between the sample and photomultiplier tube to permit the recording of specific spectral regions. For optical filters the USGS lab uses Schott BG-39 (at the upper level) and Kopp 7-59 (at the lower level) for both TL and IRSL, on the Daybreak system analyses. On the Riso system analyses a U-340 detection window helps cut off spillover from blue light emission at the 350-400 nm range and BG-39 filters out 300-700 nm (helps to gain sensitivity). For Riso measurement of feldspars U-340 and RG-715 are the common filters.
The fine-grain technique requires the determination of the a-value and thus the use of an alpha source as well as a beta source. These sources are used to artificially irradiate sets of disks so that a curve will result from increased TL growth induced by the laboratory “aging”. This curve is then extrapolated back to its intersection with a residual baseline. The baseline is defined by the TL or IRSL signal from a natural disk. The disks are exposed to 8 to 16 hours of natural sunlight for TL or 5 to 10 minutes for OSL. The alpha source in use at the USGS lab is 241Am at 0.5-mCi strength and two beta sources of 90Sr at 100 mCi and 200 mCi respectively.
There are many methods a lab can use to generated DE. The methods in use by the USGS lab are the “total bleach” with a preheat of 124°C (62 hours) and a “total bleach” with a preheat of 140°C (6 hours) for TL. “Additive dose” of 124°C (62 hours) and “additive” dose of 140°C (6 hours) for IRSL, and single aliquot analysis for both blue light OSL and IRSL large grains at 220°C (5 minutes), fading tests at all preheats and a sunlight sensitivity test. Not all experiments are performed for every sample.
The traps of some minerals, particularly feldspars are afflicted by the malign phenomenon of anomalous fading. Such fading is anomalous in that observed stability is much less than predicted from kinetic considerations. The test for anomalous fading establishes the effectiveness of the preheat treatment for removing the less stable TL and OSL components generated by beta or alpha irradiation.
In the “total bleach” method, irradiating sets of natural disks for various lengths of time generates a growth curve. The longest of these times is chosen to produce about 8x the best estimate of the DE for a particular sample. The resulting TL signals are plotted against the radiation dose. The intersection of this curve with a residual defines DE. IRSL “additive dose” technique is also approached in the same manner. Glowcurves illustrating this process are included in a sample report.
In order to test if the signal of natural TL has been stable over the age range in interest, the plateau test is applied. In the plateau test the DE or the TL age is calculated in dependence of the glow curve temperature. The start of the plateau value is indicative of the thermal stability of the analyzed TL signal. The plateau can be thermally “washed” through preheating to obtain better development of plateaus. This is why the client might note differently defined plateaus for samples at different preheat temperatures. The USGS lab uses two distinct preheats in an effort to fully test plateau placement. Experience and judgement can play a large role in choosing the DE for each sample. OSL DE does not have a plateau property; thus a TL run on the sample is usually done simply to judge the potential stability of luminescence within the sample, even if it is believed that the TL has been incompletely reset due to attenuated light exposure.
The TL and silt-size IRSL data are run and reduced using the Daybreak 1100SI TL Application software (registered trade name). This software computes a growth curve for the data points generated by passing a vertical line (single temperature) through the family of glowcurves. A least squares fit is made either to a line passing through the points or, for the saturating exponential, a line passing through a logarithmic transformation of the data. An iterative algorithm is employed for the latter fit. The natural (unirradiated) disks are given double weight. The Daybreak software estimates errors for DE using Rendell’s equation #1 for the regression of TL signal on dose (Rendell, 1985).
OSL data is run and reduced using a Riso TL/OSL-DA-15A/B (registered trade name) system. Riso system software was specially developed for the determination of the archeological and geological ages from quartz and feldspar in collaboration with Dr. Rainer Grun (Australian National University, Canberra, Australia). The special application software for the single aliquot regeneration (SAR) analysis were developed by G.A.T. Duller (University College of Wales, Aberystwth, UK) and Dr. Andrew Murray (Nordic Laboratory of Luminescence dating at Riso, Roskilde, Denmark). Complete analysis and procedures followed in the USGS lab are taken from a 2000 paper of Andrew Murray and Ann Wintle (Luminescence dating of quartz using an Improved Single-aliquot Regenerative-dose protocol, Radiation Measurements, 32 (2000), p. 57-73).
Various combinations of these parameters furnish many variants of the TL and OSL technique of dating. In consequence the dating of a particular sample by several laboratories will not necessarily yield directly comparable ages---although we feel they should at least have overlapping age ranges. It is important to present this experimental procedure together with the age data. The results of the measurements can not be judged without this important information. Only then can ages be evaluated with a realistic quoted error.
Section IV: How to collect samples for TL and OSL dating
In general, try to avoid sediments that show any post-depositional disturbances such as root penetration, krotavinia (bioturbation), carbonates, ground-water leaching, certain soil formations or large stones that may hamper sampling. If the only exposure that is available is one in which cracks, roots or krotovina may be encountered at depth, or the sediment is so hard that it will be difficult to remove a tube driven in at depth, then take a cube block sample at the site. Mark the sample as to the side exposed to the surface and wrap it in aluminum foil, a black bag (several layers if being shipped) and tape tightly to prevent disintegration.
Ideally, a 0.5 m (about a one-foot hole) must be augured into the soil sample above or below the boundary of the layer so that the luminescence and bulk samples are taken or counted on a homogenous layer. If this is not possible, take the luminescence sample and dig sideways for adequate bulk sample sediment. Sampling tubes of either 0.15” wall PVC (white or black) or steel should be driven into the back of the hole to collect the luminescence sample. PVC tubes often benefit from a sharpened or beveled edge on the back of the tube to be driven into the sediment. Tubes are driven in by sledgehammer, however, a hard metal plate should be used against the end being hammered to prevent shattered or collapse.
If the sample is being collected during the daylight hours, a black or opaque cloth MUST be used to shield the sample collection and the immediate site from light upon removal of the sample. If the sample is collected at dusk or night, a red-filter light can be employed at a minimum for collector’s visual aid. The lab will need about 50 g of sediment if the sample is fine-grained, about 75 to 100g if the sample is sand-size for adequate replicate analyses.
The sample tube should then be carefully pried or dug out from the hole by pushing a knife or chisel alongside the tube and levering it sideways. Cover the sample hole with a black cloth (this may require the aid of others) while freeing the tube. Once the tube is free, DO NOT remove it from the hole or black cloth until the open end is closed with a black cap or duct tape. Both ends of the tube can then be sealed with black electrical tape if the caps look loose or if they might pop off during sample shipment. Place the tubes in a black polyethylene black bag (the kind used in photography supply stores and not black trash bags). Write the sample name on both tube and bag.
A second sample must be collected from the back or sides of the same hole. This sample, referred to as the “bulk sample”, does not need protection from light. The bulk sample should be collected in a double Ziploc 1-quartz size bag to inhibit the loss of moisture, so that a gross moisture content can be measured in the lab. This sample can also be used to obtain a DR if field analysis is not possible. In that case, at least 600-g of sample is needed for the gamma ray counting.
Photographs or trench logs showing the locations of samples should be included if possible. Usually one photograph is required to locate the sample hole in stratigraphic context and one close-up to show texture of the sediment sample. The augured holes often show up well in distant photos of the site. Location of the site using GPS systems is appreciated and depth of sample below the original surface of the ground should be noted, as these are needed for a calculation of cosmic ray component of the DR.
Other materials that can be collected to date archeological finds using luminescence:
1). Artificial Glass: There have been few successful attempts, caused principally by glass’ non-crystalline state. TL and OSL can only used if glass has not been heated above “glass transition” temperature (? °C). (Mueller, P. and Schvoerer, M., 1993, Archaeometry 35: pp. 299-304, Factors affecting the viability of TL dating of Glass).
2). Burned Flint and Stones: Well suited for TL, especially in dating Middle to Lower Paleolithic periods. Zeroing requires 450°C and this assumption can be tested for. Oldest age obtained is 453 ± 39 ka. (Richter, D.G., 1997, Dissertation for University Tubingen, in German, no translation).
Potboilers (heated stones) have also been successfully dated. (Huxtable, J., Aitken, M. J., Hedges, J. W., Renfrew, A. C., 1976, Archaeometry 18: pp. 5-17, Dating a settlement pattern by TL: the burnt mounds of the Orkneys).
3). Ceramics and Burnt Clayware: Important concepts of TL methods were established on ceramic shards, bricks and kilns during the 1960’s and 1970’s. The usual method is to date both polymineralic fine-grains (4 to 11 µ) and quartz coarse-grains (100-200 µ), thus combining TL ages. Typical error is 6-10%. (Wagner, G.A. and Lorenz, I. B., 1997, another untranslated German paper).
Medieval kiln structures or the burned surface lining (containing quartz) have been successfully dated (Wagner, G. and Wagner, I., 1994, another untranslated German paper).
Authenticity dating of ceramic objects uses both TL and OSL. For TL, the lab needs 200 mg of powder and only really looks for baseline signal, indicating recent age or a natural TL glowcurve, indicating some elapsed time greater than 500 years. (Aitken, M. J., 1985, Thermoluminescence Dating, Academic Press, 359 p.). For OSL, 100 mg or less is needed and higher precision of measurement gives 1-2% error if the single-aliquot method is used. (Mejdahl, V. and Botter-Jensen, L., 1997, Radiation Measurements 27: pp. 291-294, Experience with the SARA OSL Method).
4). Slags: Attempts have been made to date archeometallurgic slag by TL, but the glass phase is problematic. Since slags consist essentially of silicates, heated to high degrees, in theory even OSL should work. Uncertainties about DR, however, render errors at 20%. (Elitzsch, C., Pernicka, E., and Wagner, G. A., 1983, PACT 9: p. 271- 286, TL Dating of Archeometallurgical Slags).
5). Vitrified Forts or Earth Mounds: Prehistoric walls consisting of vitrified quartz and feldspar bearing rocks are of widespread occurrence in western Europe and TL ages have been obtained from Scotland. (Sanderson, D. C. W., Placido, F. and Tate, J. O., 1988, Nuclear Tracks and Radiation Measurements 14: pp. 307-316, Scottish vitrified forts: TL results for six study Sites).
In northeastern Louisiana Paleo-Indian earthworks have been dated, presuming exposure to sunlight during construction or occupation of the mound. Dating was by means of OSL using 90-120 µ quartz. Insufficient bleaching proved to be a problem, even with OSL, but minimum ages were obtained (Feathers, J. K., 1997, Quaternary Geochronology 16: pp. 333-340, Luminescence dating of Early Mounds in Northeast Louisiana).
6). Wasp Nests: Using OSL to date quartz grains imbedded in petrified mud-wasps nests is a new way to give dates to rock petroglyphs and paintings. Nest thickness need to be at least 5mm and in some cases, an experimental method of single-aliquot analyses (due to lack of material) gave excellent results. Ages up to 16.4 ka were obtained with about 10% error. (Roberts, R., et. al., 1997, Nature: Vol. 387, pp. 696-699, Luminescence Dating of Rock Art and Past Environments using Mud-wasps Nests in Northern Australia).
Other materials that can be collected to date geological sediment using luminescence:
1). Carbonates: TL has interesting potential in dating calcitic sinter, and an extensive series of stalagmite samples were taken from a French cave which yielded agreement with U-series and U-Th ages. TL age underestimation was found for samples whose dose rate is dominated by the internal alpha and beta components, however. (Debenham, N. C. and Aitken, M. J., 1984, Archaeometry 26: pp. 155-170, TL dating of Stalagmitic Calcite).
Another form of TL has been proposed which dates “dirty” pedogenic carbonates. The method uses changes in the dose rate to quartz grains when the sediment becomes carbonated. (Singhvi, A., et. al., 1996, Earth and Planetary Science Letters 139: pp. 321-322, A Luminescence method for dating “dirty” pedogenic carbonates for paleoenvironmental Reconstruction).
Because the samples of mollusk shells suffer structural alteration at usual TL glow-curve temperatures only sporadic studies exist for these. General consensus seems to be that it works in a limited number of cases, if it can be proven that the shells exhibit a glow-curve in the 240°C red spectral region or are of the Pectinidae or Ostreidae family. (Ninagawa, K., et. al., 1994, Quaternary Science Reviews 13: pp. 589-593, TL dating of calcite shell Crassostrea gigas (Thunberg) in the Ostreidea Family).
A recent literature research reveals no OSL studies being done on carbonates, possibly because ESR use is more prevalent.
2). Colluvial and Alluvial Silts: Most recent studies have found that different degrees of bleaching depend on depositional conditions and that all too often only distal wash facies or the upper part of the colluvial wedge are sufficiently bleached and suitable for TL dating. (Forman, S. L., et. al., 1991, Journal of Geophysical Research 96 B1: pp. 595-605, TL Dating of Fault-scarp Derived Colluvium: Deciphering the Timing of Paleoearthquakes on the Weber segment of the Wasatch Fault Zone, North Central, Utah). Careful application of the partial bleach method however, can give ages for some alluvium.
Combinations of TL and IRSL have been done with excellent agreement in many arid climates, notably Israel (Porat, N., et. al., 1996, Quaternary Research 46: pp. 107-117, Late Quaternary earthquake chronology from luminescence dating of colluvial and alluvial deposits of the Arava Valley, Israel). Typical minerals used in IRSL are K-feldspars of coarse-grained size, although with judicious application the polymineralic fine-grained portion is also acceptable. The emerging opinion seems to be that modern colluvial sediments in arid regions are bleached better and far more uniformly than modern alluvium, thus better suited to luminescence age control and reflect past sediment conditions.
3). Dune Sand: Along with loess, considered routine for both TL and OSL. It has an advantage over loess in that monomineralic, coarse-grained fractions can be separated. Any feldspar found in these sands has unusually high TL sensitivity (important for dating Holocene sands). Many studies have been done with respect to sea levels over time, interglacial sea maximas (responsible for dune formation), rock shelter occupational periods and human migration patterns. (Roberts, R. G., et. al., 1994, Quaternary Science Reviews 13; pp. 575-583, The Human Colonisation of Australia: OSL dates of 53,000 and 60,000 years Bracket Human arrival at Deaf Adder Gorge, Northern Territory).
On the other hand, OSL has been extremely useful for dating dunes too young to be assessable by TL. Ages as young as 100 years have been obtained with the use of quartz, sanidine and potassium feldspars (Stokes, S., et. al., 1997, Nature pp. 154-158, Multiple Episodes of Aridity in Southern Africa since the last Interglacial Period).
4). Fluvial Systems: Due to its better “bleachability”, feldspar is generally preferable to quartz for TL dating in fluvial systems. However, if sands have been transported in shallow, clear rivers, even quartz grains may be fully bleached. More systematic studies on the suitability of fluvial, glaciofluvial, limnic and coastal marine sands are needed. Light conditions are crucial in each case, since the water column tends to absorb light, particurlary short-wave spectral regions.
Most sands of fluvial origin can be dated, provided some other technique (such as U-series or Electron Spin resonance) is used as a check. Recent cases also combine TL with PPTL (phototransfer thermoluminescence) and results are encouraging. (Murray, A., S., 1996, Geochimica Cosmochima Acta 60: pp. 565-576, Developments in OSL and PTTL dating of young sediments: Application to a 20000-year sequence of Floods). Green light luminescence on coarse grains is encouraging and flood deposits using OSL have also been successfully dated.
5). Fulgurite: Basically, no one has ever tried it and reported on it in the literature. In theory, if lightening hits a sandy surface the heat effect is sufficient to cause fritting and melting of quartz grains, therefore TL signal would be reset.
6). Glacial Deposits (including glaciofluvial): Not much is known about the suitability of glacio-fluvial sediments for TL dating on a general basis. The degree of bleaching strongly depends on type and distance of transport; thus every site must be individually assessed. Most sites just don’t possess enough “reset” sediments, some showing a bleaching equivalent of only 10 minutes (on a sunny day)!! TL is a bust on most tills and debris flows, although several studies have been done on lakebed cores of mainly glacigenic origin with much success, using infrared stimulation (IRSL) on the fine-grained portion. (Hardy, F. and Lamothe, M., 1997, Quaternary Science reviews 16: pp. 417-426, Quaternary Basin Analysis using IRSL on Borehole Cores and Cuttings).
Single-aliquot analysis may be necessary because glacio-fluvial sediments contain such a mix of bleached and unbleached grains. IRSL ages have been obtained on coarse-grains (100-200 µ) with limited success. (Hutt, G. and Junger, H., 1992, Quaternary Science Reviews 11: pp. 161-163, Optical and TL dating of Glaciofluvial Sediments). OSL holds the most promise to date any glacial deposit, but even OSL is not “reset” in most of glacial terrain.
7). Impactites: This potential of the TL method was tried on the Arizona Meteor Crater and it was determined that 680°C and a shock-wave intensity of 10Gpa was enough to reset the Coconino Ss. quartz. An age of 49 ± 3 ka was determined by TL. (Sutton, S. R., 1985, Journal of Geophysical Research 90: pp. 3690-3700, TL measurements on Shock-metamorphosed Sandstone and Dolomite from Meteor Crater, Arizona).
8). Loess: Now considered to be “cookbook” for TL dating, particularly the fine-grained portion of quartz and feldspar. For particularly sandy loess coarse-grained fractions are also used. Most reliable ages are under 120 ka (last glacial-interglacial), although ages up to 800 ka have been reported. (Berger, G., W., et. al., 1992, Geology 20: pp. 403-406, Dating Loess up to 800 ka by TL). If the site has a low dose rate (i.e. 4Gy/ka) ages of 100 to 200 ka can be obtained routinely and reliably. (Waters, M. R., Forman, S. F., and Pierson, J. M., 1997, Science 275: pp. 1281-1284, Diring Yuriakh: a Lower Paleolithic site in central Siberia).
If the TL signal has been reset in loess, OSL is expected to work just as well. In fact, there are several problems, if the fine-grained portion is used. Ages past 120 ka are frequently underestimated using green light OSL and Infrared. It is now suggested that medium grains (43 to 54 µ) and coarse grains are used, especially for infrared. (Li, S.H., and Wintle, A. G., 1992, Quaternary Science Reviews 11: pp. 133-137, A global view of the stability of luminescence signals from Loess).
9). Marine sands: Sands are commonly able to be dated by TL due to the fact that they are well exposed to light by repeated displacement in shallow coastal regions. Both quartz and feldspars are used to date littoral deposits and high sea levels. (Mauz, B., et. al., 1997, Palaeogeogr Palaeoclimatol Palaeoecol 128: pp. 269-285, Middle to upper Pleistocene morpho-structural evolution of the NW coast of Sicily: TL dating and palaeontological-stratigraphicalevaluations of littoral sediments). Other marine facies such as terraces, sublittoral, and intertidal sediments all show age overestimations with TL.
Very recent studies in sublittoral sediments indicates that Red Light OSL holds promise for dating modern and Holocene sediments, but that IRSL is the more robust geochronometer. IRSL still shows considerable variability in solar resetting, which erodes precision for ages <50 ka. The fine-grained fraction of the sediment is used most frequently as it receives the greatest light exposure during deposition. (Forman, S. L., 1998, Arctic and Alpine Research, in press, Infrared and Red Stimulated Luminescence Dating of late Quaternary Near Shore Sediments from Spitsbergen, Svalbard).
Coarse-grain K-feldspar grains using OSL has been used to date tsunami deposits (mainly in tidal-flat and tidal-channel sands) that had probably been reworked by currents, waves and burrowing organisms prior to the tsunami (thus increasing light exposure). (Huntley, D. J. and Clague, J. J., 1996, Quaternary Research 46: pp. 127-140, Optical Dating of Tsunami-laid Sands).
10). Pseudotachylite and Fault Breccia: The glass phase of the pseudotachylite offers a potential for TL dating. Linear growth is usually observed and high temperatures associated with the formation of friction melts during the intense fault movements are usually enough to reset the TL signal in the quartz and feldspar relics. Both coarse- and fine-grained fractions of gouge sample have been used. (Singhvi, A., 1994, Quaternary Science reviews 13: pp. 595-600, Luminescence Studies on Neotectonic Events in South-Central Kumaun Himalaya-a Feasibility Study).
11). Soil Horizons: A horizon-This soil horizon forms first but is often eroded away. If present or formed in an arid to semi-arid environment, the top of this horizon is acceptable for TL and IRSL dating unless contaminated by younger clay and silt filtering down from above. The presence of clay films is evidence of such contamination. The mineral grains are repeatedly exposed to sunlight through bioturbation and therefore, TL age corresponds to the time of the formation of the soil. These soils often have humified organic matter mixed in with mineral material. (Buol, S. W., et. al., 1997, Soil Genesis and Classification, 527 p., Iowa State University Press).
B horizon-Forms after the A horizon and is less suitable. Dominant features include evidence of removal of carbonates; residual concentrations of sesquioxides or silicate clays; coatings of sesquioxides that give a darker or stronger or redder color to the horizon; illuvial concentration of silicate clay, iron aluminum, carbonates, gypsum or humus. If it is high in clay, it can swell and shrink and thus, is often contaminated with younger silt filtering down via roots and cracks. The radiation history is less certain because water moving through this horizon creates oxidizing or reducing conditions. Weathered horizons, such as B soils, should be avoided as they often exhibit pedogenic accumulations of clay, silt, carbonate or silica.
C horizon-This horizon usually provides the most reliable TL and IRSL ages. Those samples collected at the bottom of the horizon date the start of deposition. This layer is altered very little by soil forming process. Also used as a designation for coprogenous earth, diatomaceous earth, saprolite and other sediments that are not hard enough to qualify as rock. (Millard, H. T. and Maat, P. B., 1994, USGS Open File Report 94-249, TL Dating Procedures in Use at the U.S. Geological Survey, Denver, Colorado).
12). Tephra (Volcanic Ash): The fine-grained glass phase of volcanic ash bears considerable potential for TL dating. Distal ash deposits are workable as well, due to their fine-grained nature and ages up to 400 ka have been obtained. Ages as young as a few 100 years old may be dated this way as well. The grain size used is 4-11 µ and special attention should be paid to preheating techniques and TL growth (Berger, G. W., 1991, Journal for Geophysical Research 96: pp. 19705-19720, The use of Glass for Dating Volcanic Ash by TL). The IRSL (infrared) procedure (880 nm wavelength), the green light OSL (514 nm wavelength) and red light OSL (633 nm wavelength) techniques were examined, but only the IRSL was seen as a favorable signal for Mazama Ash. (Berger, G. W. and Huntley D. J., 1994, Quaternary Science Reviews 13: pp. 509-511, Tests for OSL from Tephra Glass).
13). Volcanites: Direct dating (using TL) of lava streams causes difficulties owing to the rarity of quartz and problems with anomalous fading in the feldspars. Tholeiitic basalts, in particular, exhibit unfavorable TL behavior. Using plagioclase for TL analyses has been successfully done with an accuracy of 10%, (Guerin, G. and Valladas, G., 1980, Nature Vol. 286: pp. 697-699, TL Dating of Volcanic Plagioclases) but the lab work is very time consuming and takes a year or more to get results back. Baked argillaceous shales in volcanic slag have also been shown to be suitable candidates for TL, with careful selection of sampling sites (Zoeller, L., 1989, untranslated German paper).
Section V: Explanation for differences in TL and OSL techniques and other closing comments
Advantages of TL over OSL:
1). Stability of luminescence can be judged by “plateau test”
2). Less sample preparation and analysis time required
3). Wider age range published in literature
4). More reliable older ages past 200 ka (published or otherwise)
Advantages of OSL over TL:
1). Deals with easily bleached component, requires less “resetting” time (5 minutes vs. 8 hours)
2). No brute heating of sample needed, can be measured easily
3). Single grain or aliquot analyses readily available (great for fluvial, glacial and lacustrine sediments)
4). Thoroughly bleached grains can be picked out for analysis
General outside agency pricing: $1000/sample (US$)
USGS in-house pricing: one payperiod of time ($3500) allows for four samples to be processed
Optional: If we like the project, feel like we are contributing to a worthwhile endeavor and get published somewhere, prices negotiable
Services offered: TL on 4-11 micron (µ) silts, IR-OSL (infrared stimulated luminescence) on 4-11 µ polymineralic or feldspar grains or 90-125 µ feldspar grains and BLUE-OSL (blue light optically stimulated luminescence) on 90-125 µ quartz grains
Methods offered: Additive, multi-aliquot; partial bleach; multi-aliquot for TL
Additive, multi-aliquot; single aliquot for IRSL
Single aliquot for BLUE-OSL
Anomalous fading tests (every sample) and sunlight sensitivity tests (one sample per site)
Dose Rate: Field measurement by gamma spectrometry (K, U, Th and cosmic ray) or lab gamma spectrometry with NaI crystals (K, U and Th). Optional measurements of K by XRF; of U, TH with ICP-MS; or K, U and Th by INAA.
Sample turnaround time: Six (6) samples take three months to date (from time of sample receipt in lab). No ages are released until we are satisfied about the quality of the data. Sample turnaround times may be slower, but never faster. Remember that these samples have accumulated for thousands of years, it takes a few months to reveal their secrets.
Range of TL: 1,000-800,000 years with the 800,000 years heavily dependent on Dose Rate. Most samples have a practical application of 1,000 to 500,000 years.
Range of OSL: 500- 150,00 years (some practitioners report success at 100 years, but this does not seem feasible for most studies)
USGS lab: In operation since 1992. Latest equipment from Riso Labs offering BLUE-light OSL capability. Specializes in all geological applications, with published results mostly in the southwest US. Contact: Shannon Mahan, Box 25046, MS 974, Denver Federal Center, Denver, CO 80225, smahan@usgs.gov
Other USA Labs and contact personnel:
1). Glenn Berger (Earth and Ecosystems Sciences, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512-1095, gwberger@dri.edu) specializes in all geological applications
2). James Feathers (University of Washington, Dept. of Anthropology, Box 353100, Seattle, WA 98195-3100, jimf@u.washington.edu) specializes in archeological applications
3). Steve Forman/James Pierson (Dept. of Earth and Environmental Sciences, University of Illinois, 845 W. Taylor St., Chicago, IL 60607-7059, slf@uic.edu) specializes in all geological applications
4). Ed Haskell (Center for Applied Dosimetry, University of Utah, 729 Arapeen Dr., N 205, Salt Lake City, UT 84108-1218, e.haskell@m.cc.utah.edu) specializes in dose rate problems and theoretical applications, not necessarily sample ages
5). Steve McKeever (Oklahoma State University, Dept. of Physics, 145 Physical Science Bdlg., Stillwater, OK 74078, u1759aa@okstate.edu) specializes in theory and physics of luminescence along with solid-state TL, some students doing geological applications
6). Joel Spencer and Lewis Owen (University of California, Riverside, Dept. of Earth Sciences-036, 1432 Geology, Riverside, CA 92521-0423, joel.spencer@ucr.edu) specializes in glacial applications to geological sediment
7). Ron Goble and Mike Blum (University of Nebraska, Lincoln, NE, mblume@unl.edu and rgoble2@unl.edu) specializes in Nebraska Sand Hill and other Mid-West features or geological dating projects linked to the University.
8). Dr. Cathy Wilson is in charge of setting up a lab at Los Alamos National Labs in New Mexico for the Environmental Sciences Group. Announcement sent out 11/13/00 for Post Doctoral researcher with hands-on experience of OSL, but might consider other outstanding experience in experimental research. E-mail cjw@lanl.gov.