Thermoluminescence (TL) Dating

David M. Price

Basic Principles

Thermoluminescence (TL) dating of sediments dependsupon the acquisition and long term stable storage ofTL energy by crystalline minerals contained within asedimentary unit. This energy is stored in the form oftrapped electrons and quartz sand is the most commonlyused mineral employed in the dating process. Prior tothe final depositional episode it is necessary that anypreviously acquired TL is removed by exposure to sunlight.After burial the TL begins to build up again at a ratedependent upon the radiation flux delivered by long-livedisotopes of uranium, thorium and potassium. The presenceof rubidium and cosmic radiation generally play a lesserbut contributory roll, and the total radiation dose deliveredto the TL phosphor is modified by the presence of water.The period since deposition is therefore measured bydetermining the total amount of stored TL energy, thepalaeodose (P), and the rate at which this energy isacquired, the annual radiation dose (ARD).

TL Age = Palaeodose (P) / (ARD)

Sampling techniques and strategies

TL samples may be collected in open ended or opaquePVC tubes approximately 12cm in length and 6cm in diameter.Whilst it advisable to protect the sample from directsunlight there is no need to sample at night and theorientation of the specimen is not important. The sampleis taken by introducing the tube into a freshly cleanedback surface; if this proves difficult a block may becut from the unit of interest. The specimen tube or blockshould then be wrapped in black plastic to prevent furtherexposure to light and to preserve the environmental moisturecontent. The exposed material at either end of the tubeis used in the determination of the annual radiationdose and the internal unexposed portion for the palaeodose determination.
 

The sample collected should be taken from the centre of a homogenous sphere of 30cm radius. In the absence of a field gamma spectrometer this is extremely important as the presence of rock or any other dissimilar material within this distance may have an effect upon the radiation dose received by the sample. This effect cannot be determined in the laboratory from the sample submitted. If a field gamma spectrometer is available measurements should be made in the field by placing the gamma probe into the hole from which the sample has been removed. This correction will subsequently be applied to the laboratory measured radiation dose value.

A sample of recently deposited similar material shouldalso be submitted in order that the TL starting point atthe time of deposition can be measured and a suitablecorrection applied. In the case of an aeolian depositthis may be collected as a surface peel on adhesive tapeor from the subsurface of the deposit. This correctionbecomes a little more difficult in the case of waterborne sediments but may be determined from a subsurfacesample of recently deposited similar material. In generalthe older the sediment the less significant the surfaceresidual correction becomes. The technique is best suitedto the age determination of sediments which have undergonelong transport distances and, in the case of water bornesediments, as suspended load under shallow water, lowenergy conditions. Aeolian sediments are more likelyto have undergone considerable solar exposure prior todeposition and therefore are more likely to have beeneffectively zeroed. Care must also be taken to ensurethat the sample is taken from an undisturbed site and has not been subject to re-exposure by bioturbation.

Laboratory sample processing

Upon arrival in the laboratory TL samples normally consistof two parts: the sample to be dated and a modern analoguesample for the surface residual correction. Both specimensare carefully sieved to separate the 90-125 micrometregrain size fraction, chemically cleansed in dilute HCl,etched in 40% w/w HF and finally subjected to heavy liquidseparation. The sample so prepared consists of betterthan 99% pure quartz grains.

The quartz from the specimen under investigation isdivided into two parts one of which is heavily bleachedunder a UV sunlamp. This exposure effectively removesalmost all of the previously acquired TL leaving onlywhat is termed as the unbleachable TL. Aliquots ofboth the bleached and the unbleached quartz are depositedonto a series of aluminium planchettes and a number ofthese are incrementally irradiated using a calibrated90Sr plaque source. Each planchette, complete with itssample aliquot, is heated to 500 degC at a controlledrate and in an oxygen free atmosphere. The light emitted(TL) is recorded and in this way it is possible to establisha TL growth curve which relates TL output and the absorbedradiation dose. With reference to this curve the measurednaturally accumulated sample TL may be converted to absorbedradiation units (Palaeodose P). The surface residualTL correction is determined from the modern analoguesample by means of a similar procedure and this correctionis applied to the palaeodose value. In the absence ofa suitable modern sample the laboratory induced unbleachableTL level is assumed which has the effect of maximisingthe resultant TL age determined. In the case of an oldersample this correction may only represent a small proportionof the total age.

The radiation dose received annually by the sample ismeasured by means of calibrated thick source alpha countingwhich determines the specific activity of the uraniumand thorium decay chains assuming that secular equilibriumexists. This process requires that the sample be crushedto an extremely fine grain size such that all of theshort range alpha particles may be detected. The crushedsample is placed in immediate contact with a scintillationscreen which is sealed in an alpha counting cell whichin turn is positioned on a photomultiplier tube assembly.Because certain of the daughters within the uranium andthorium decay chains are gaseous it is necessary to waita period of three weeks before introducing the cell intothe counter. This period allows the decay chains to bere-established. The amount of potassium present in thesample is determined by means of atomic emission spectroscopyand the rubidium content by X-ray fluorescence. Thus,assuming the cosmic contribution and applying a correctionfor the modifying effect of the sample moisture content,the radiation dose received upon an annual basis (ARD)may be computed and the depositional age of the sample determined from the equation shown.

Dateable material

TL dating as practised in the Wollongong laboratorymay be applied to aeolian, fluvial, coastal and, in somecases, marine sediments. The technique is also successfullyapplied to volcanic materials and heated fire hearthsamples and therefore may be directly applicable in certainarchaeological contexts. The age determination of potteryis also undertaken and the method may also be appliedto casting core material removed from bronze artefacts.Certain of these analytical procedures may make use ofthe polymineral 1 – 8 micrometre fine grain sample fraction rather than the 90 – 125 micrometre quartz grains.

Cost and turn around time

The full cost of a sedimentary age determination iscurrently $715.00 (inc. GST) although in certain circumstancesa reduced collaborative rate may be applicable. Thischarge includes the analysis of the modern analogue samplewhich may be applicable to more than one sediment. Acharge of $247.50 (inc. GST) is made for the authenticationof pottery and bronze artefacts the purpose of whichis to establish the originality of the piece rather thanto accurately determine the age. This pricing structurewill be in place commencing January 2004 until furthernotice.

The turn around time for authenticity work is generallyaround two weeks. Often a verbal result is availablewithin a week of submission of the ware. It is preferableto bring the item for test to the laboratory such thatit can be sampled and taken away on the same day. Sedimentsmay be sent to the laboratory but full field data mustalso be submitted. It is essential to work closely withthe laboratory to ensure a full understanding of theprocesses involved in the analysis and also to provideany special information that my have an effect upon thedating of the sample. TL dating is both a costly andtime consuming process and, as with most dating methods,good samples give good dates, poor samples seldom providereliable ages. By the nature of the procedure practisedTL dating of sediments cannot be achieved reliably inless than a month and because of the backlog analysiswill generally take longer. Small numbers of samplesare normally analysed rather more promptly than largerbatches which tend to reduce the flexibility of the laboratory. Every effort is made to meet all agreed realistic deadlines.

Australian quarantine regulations

Australia necessarily has quarantine regulations whichprohibit the entry of unauthorised soil/samples fromother countries. Before submitted such a sample to thelaboratory it is necessary to obtain an import permit.These may be obtained from Australian Quarantine andInspection Service at cost of AUD$60 and are valid overtwo years. Additionally the laboratory possesses a permitof approval to work upon imported samples which is renewedupon an annual basis. When sending such samples to thelaboratory it is essential that a copy of the importpermit is attached to the package and that it is clearlymarked with the laboratory approval number N0555. Itis also necessary to clearly label sample packages toprevent x-ray or exposure to light. The addition of thelaboratory phone number may also assist in the case of problems encountered during quarantine inspection.

Other applications and advantages

TL is an electron trap method of dating as opposed toa radiometric technique such as carbon dating. This hasthe advantage that the TL signal increases with timerather than decreasing. At a palaeodose level, dependentupon the physical properties of the quartz grains, asaturation point is reached beyond which there is nofurther increase in the TL with additional irradiation.This saturation level represents the point at which allof the electron traps are full. Once this palaeodoselevel has been reached a finite depositional age forthe sediment cannot be determined and it is only possibleto determine a minimum age value. This level frequentlyoccurs at around 250 Grays in the case of quartz butthis value may vary from one sample to another. An accumulatedpalaeodose of this level may be reached as soon as 50kaor, in exceptional cases, almost 1ma dependent upon theoperative radiation flux level.

Naturally occurring quartz grains contain chemicallyinbuilt impurities such as silver and manganese and itis from within these electron trap sites that TL is emitted.Pure quartz in fact does not emit a TL signal. Quartzis then often the preferred phosphor employed in theTL dating process because of the stability of the electronswithin the crystalline lattice ie long lifetime c109years. Feldspars may be subject to loss of TL signal over a period of time and must be treated with caution.

Because of the finite electron trap energy levels associatedwith each of the inbuilt impurities within the quartzcrystals, TL peaks are exhibited at defined temperaturesas the quartz crystals are heated and the trapped electronsreleased. Thus a TL spectra may be taken as being indicativeof the impurities contained within the quartz grainswhich in turn may represent the formation conditionsand history of the quartz grains. This property, in someinstances, may be utilised in the detection of changesin provenance of the quartz and the sediment in whichit is contained which can indicate a change in palaeowinddirection or an altered water flow regime.

Electrons stored at low energy electron trap levelsare more easily released and thus, upon heating, a naturallyaccumulated TL spectrum will not contain TL peaks inthese regions. When a quartz sample is irradiated inthe laboratory and fairly quickly heated these low energyTL peaks are immediately evident in the so-called secondglow curve. If the amplitude of this spectrum is comparedwith a natural TL glow curve over a range of temperaturesa plateau region will result over the region where thetrapped electrons are stored in a stable fashion, usuallybetween 200 and 500 degC.

If a sediment is only partially exposed during its transportphase, or perhaps, has been re-exposed by a process ofbioturbation electrons stored at less stable, lower energyelectron traps, may be released leaving only those trappedat the higher energy levels. If the TL spectrum displayedby such a sample is compared with that emitted by a recentlylaboratory induced spectrum then a stepped temperatureplateau characteristic may well result. This is sometimesevident in the case of tsunami deposited sediment wherethere is little time and solar exposure to enable thecomplete resetting of the previously acquired TL signal.A similar process may take place if a sediment is partiallyre-exposed perhaps as a result of bioturbation or other processes.

Recent literature references and handbooks

Extremely readable accounts of TL dating, processesand application may be found in Thermoluminescence Datingby M.J. Aitken, Academic Press, 1985 and Science-basedDating in Archaeology, Longman, 1990 by the same author.There is also a users newsletter available, Ancient TL,in which the recent developments are reported. This appearsthree or four times a year and is available from theLaboratoire de Physique Corpuschulaire, IN2P3-CNRS UniversiteBlaise-Pascal, 63177 Aubiere Cedex, France. The resultsof the application of TL in its many aspects are reportedin many international and Australian scientific journals,Australian Geographer, Australian Journal of Earth Sciences,Quaternary Research, Geomorphology, Marine Geology etc.An international meeting of TL and related electron capturepractitioners is held every three years and the proceedingsof these meetings are reported in Radiation Measurementsand Quaternary Science Reviews. The last such meetingwas held in Lake Tahoe on the Californian-Nevada borderduring 2002 and the proceedings of the 1999 meeting,held in Rome, appear in Rad. Meas. vol 32, No. 5-6 andQSR vol. 20, Nos. 5-9. These volumes provide the most up to date information regarding the state of the art.