Chapter 3-10: Thermoluminescence

Thermoluminescence (TL) dating is a technique that is based on the analysis of light release when heating crystalline material. TL-dating is used in mineralogy

and geology, but is also increasingly being applied for dating of anthropological and archaeological samples.

chlorophane mineral

cool state heated state

Applicability of TL datingTL dating is mainly applicable for material with mineral

or crystalline structure or with spurious crystalline contents. It is only usable of insulating material not for metallic artifacts.

In archaeology TL is mainly used for pottery analysis.

In anthropology the main use of TL is the dating of flint stone as early tool material for mankind.

The glow curve

first heating

second heating

Typical phenomenon of thermoluminescence, when sample isheated above 200 oC, light emission in blue range is observedup to 400 oC. At higher temperature, material emits a red glow.For second heating no blue light emission is observed, only thered glow curve remains at the higher temperature range.

Origin of thermoluminescenceThermoluminescence originates from the temperatureinduced release of energy, stored in the lattice structure of the crystal following long-term internal and external exposure to nuclear radiation from natural sources.

TL accumulates in the material with time depending on radiationexposure. TL is released by heating. The dating clock starts with the initial firing of the material, whenoriginally accumulated TLis being driven out.

Age determinationthe amount of accumulated TL is proportional to the age of the sample and inverse proportionalto the nuclear radiation exposure of the sample.

dose)(TL/unit Dose) (AnnualTL icalArchaeologAge

⋅=

Amount of energy E deposited by radiation exposure into sample of mass m within one year. It dependson the content of radioactive nucleiin the sample material and on the exposure to external radioactivity.

tmE

tD

⋅=Annual Dose:

reminder from Chapter 1-2:

Dose: DEm

= Amount of energy E deposited by radiation into body of mass m.

1 Gray (Gy) = 1 Joule/kg (Energy/mass)1 rad = 1 centigray = 10 milligrays ( 1 rad = 1cGy = 10 mGy ) Nominal background radiation absorbed dose 100 mrad/year = 1 mGy/yr.

TL/unit dose is material specific. It describes the probability for a material to develop thermoluminescence (TL) under radiation exposure. This has to be determined by independent analysis typically by exposure to well defined radioactive sources.For metals: TL/unit dose =0

Example 1

temperature oC

21 000 TL-cts/s

TL

[103

cts/

s]

52 000 TL-cts/(s·0.1Gy)The TL curve shows 21000 TL-counts/s at an average annual dose of 1 mGy/y

Measurement of TL withan calibrated source gives52000 TL-cts/s after an exposure to 0.1Gy radiation.TL/dose= 520000 TL-cts/(s·Gy)

y4038Age]Gy[s 520000[Gy/y]101

][s 21000Age 11-3

-1

=⋅⋅⋅

= −−

Physical Mechanismradiation creates defects in lattice structure of ionic crystal,e.g. CaCO3 with Ca+ and CO3

- ions constructing the lattice.Typically defects arecharacterized by theabsence of negative or a positive ion in its lattice location;e.g. negative ion vacancy acts as atrap for electrons.Intra-lattice positionmay serve as trapfor electrons. Alsoelemental impurities can cause traps (Ag+).

solid state structureThe energy configuration of electrons within a lattice can be described by the electron band model; bound electrons (attached to an ion) are in the valence band; free electrons – floating between ions – are in the conduction band. A good insulator has no electrons in the conduction band and has a filled valence band.

Energy

conduction band energy absorption causes excitation of electron from valence to conduction band.De-excitation by immediateemission of photon energy.

band gap (energeticallynot allowed for electrons

valence band

Thermoluminescence

free movingelectrons

trappedelectrons

bound electrons

excitation or de-excitation to intermediate (trap) level

time-scale for TL clock

decayλτ 1

=Life time τ of theintermediate trap configuration is:

λdecay is the probability for the decay transition between the electron in its intermediate state and the valence band.

While the decay probability between the conductive and the valence band is >1012 s-1 (τ<1 ps), the probability for the

decay of the intermediate state can be up to 10-12 s-1, that means the life time τ<106 y providing a long time clock.

There is also a range of intermediate life time configurations.

experimental set-up

heater with temperature control; blue filter for red light glow absorptionand photomultiplier with scintillatorfor TL photon measurement.

TL emission light

the red component from the natural heat induced light emission must be separated by using special filters between sample and photomultiplier detection system.

the plateau test

two portions of the sample are TL tested, one after additional exposure to radioactive source (b radiation) and one without additional exposure.

The ratio of the singleTL curves shouldprovide a plateauin the TL range.

short-lived TL components

The plateau method

The plateau method – initial developed as an independent test –provides a more reliable method since with this relative approachthe systematic uncertainties in the absolute dose rates cancel out.

dose annualpaleodoseAge =

The paleodose is the natural dose received by sample since its last heating (production); it can be determined from theplateau ratio R between the paleodose and the paleodose

plus the source related dose from the sample irradiation. Theannual dose is material dependent and must be measured.

Paleodose Determination

)R(N

N

)R(NNR

N)NN(R

NNNR

1

1

1 −=

−=⋅

=+⋅

+=

−β

β

β

β

If R=0.5 (R-1=2) the sample received during the source irradiation the same dose as naturally over its life time. If R=0.33 (R-1=3) the sample received during the source irradiation the twice the dose as naturally over its life time.

Example 2

Nβ=10 Gy

Gy.N;GyN

N.)./(

NN;.R

β 868 :Paleodose10with

886014701

470

==

⋅=−

== ββ

Annual DoseFor most pottery samples the TL is produced in roughly

equal proportions by the nuclear radiation from potassium40K (β, γ emitter), thorium (α-emitter), and uranium (α-emitter)

plus its additional β and γ radiation which is emitted along the decay chains. Minor contributions yield from rubidium

contamination (β-emission) and cosmic radiation bombardment.

cr' DDDDD +++= γβα :dose Annual

α radiation saturates the TL traps due to its high ionization probability, therefore only 10-15% of the α leads to TL dose.

15010 ..kDkD' −≈⋅= αααα

Dose origin in sample

The α and the β particles are completely stopped in pottery material. The α, β dose Dα, Dβis therefore completely internal. The γ radiation has a very long range and Dγ therefore is mostly from the surrounding material. Cosmic ray dose Dcr is external.

Sample preparation

The mineral grains (high TL) is separated from the clay powder by chemical techniques to increase the light emission efficiency.

Due to the short range of α and β particles the sample material must be converted into small grain powder (range size ~4-8µm).

mean α-range

fine grain method

α dose from U,Th

γ-dose~100%

β-dose from 40K

grain size [in mm]

dose

in %

coarse grain method

fine and coarse grain method• the fine grain method seeks to maximize the α-dose in sample powder of3-10 µm grain size!

• the coarse grain method is based only on β and γ induced luminescence using a coarse powder with more than 500 µm grain size !

0.5mm

Annual doses for typical potteryfine grain technique

Example 3 (fine grain method)Annual dose for ‘typical’ pottery and soil conditions can be estimated (neglecting Cosmic Ray Dose):

[Gy/y]10185[Gy/y]102411058110362

3

333

−

−−−

⋅=

⋅+⋅+⋅=

++⋅=

.D)...(D

)DDDk(D γβα

Comparison with the paleodose yields an age of:

[y]1710[Gy/y]10185

[Gy]868dose annual

paleodoseAge 3 =⋅

== −..

Example 4 (coarse grain method)For coarse grain size the α-dose is negligible and the β dose has dropped to 90-95% of its small size value.

Gy/y10241

Gy/y10421Gy/y10581903

33

−

−−

⋅=+

⋅=⋅⋅=⋅=

.DD

...DkD

CR

'

γ

βββ

considerably smaller dose values than with fine grain methodthe TL paleodose from a fraction of the same sample is 4.6Gy.

[y]1704[Gy/y]1072

[Gy]64Age

kpaleodoseAge

3 =⋅

=

++⋅=

−..

DDD CRγββ

Coarse grain technique TL datingDating of mineral fractions from Viking pottery fragments

Uncertainties & ProblemsIf sample has been exposed to light for considerable timebleaching may occur, de-excitation of TL traps by photon

interaction. TL drops, suggesting a lower paleodose & age.

a) natural TL curveb) after 1h light exposurec) after 24 h light exposure

TL

≈ 0.72·TL

≈ 0.15·TL

since age is proportionalto paleodose, significant

errors can occur!

bleaching time dependenceSince the light exposure causes photon induced de-excitation ofintra-band states the TL drops exponentially with exposure time t.

constant material0

=⋅= ⋅−

κ

κ teTL)t(TL

Quartz is more slowly bleached than Feldspar, if κ and light exposure time t is known, corrections can be applied.

bleached piece of potteryunbleached case:

[y]1710[Gy/y]10185

[Gy]868dose annual

paleodoseAge 3 =⋅

== −..

1 h bleached case paleodose is reduced to 72%:

[y]1231[Gy/y]10185

[Gy]386dose annual

paleodoseAge 3 =⋅

== −..

sample appears considerably younger!

water uptake and water damagePottery within a meter or so from ground level are typically wet! Water uptake has direct impact on annual dose since it acts as absorber material for the radiation. Wet pottery haslower annual dose suggesting a higher age of pottery sample.

wf.k

k

wf.k

k

wf.k

k

γβα,

drywet

drywet

drywet

⋅⋅+=

⋅⋅+=

⋅⋅+=

1411

2511

511

:is exposureand,for rates dosein Reduction

γγ

ββ

αα

water absorbs α-radiation 50%more efficiently than dry clay, water is 25% more efficient inabsorbing β-radiation but only 14% more efficient for γ-radiation.

f: fraction of water retained in soil material(0.95 in UK, 0.6 in Italy)w: saturation of water uptake; obtained byweighting wet and dry sample (w=0.05-0.3)

wet piece of potteryin soaked soil, f=0.95, soaked piece of pottery, w=0.3.If sample has a paleodose of 8.86 Gy like a dry sample:

[Gy/y]10763[Gy/y]10241760105817401036270

3

333

−

−−−

⋅=

⋅⋅+⋅⋅+⋅⋅=

⋅+⋅+⋅=

.D)......(D

)DkDkDk(D wetwetwetγγββαα

about 73% reduction in annual dose

[y]2356[Gy/y]10763

[Gy]868dose annual

paleodoseAge 3 =⋅

== −..