-
Revista Mexicana de Física 23 (974) FA87-FAI04 FAS7
THERMOLUMINESCENCE MEASUREMENTS IN MESOAMERICAN•POTTERY WITII A
COMMERCIAL TL READER
A. Moreno y Moreno*. and Klaus Bccker
I/ea/lh Physics Division. Oak Ridge Naliona/ Laboralory
Oak Ridge. Tennessee, USA 37830
(Recibido: enero 10. 1974)
(Versión revisada: ocrubre 29. 1974)
ABSTRAer: The feasibiljty of an unmodified commercial TI~D
reader(lIarshaw 2000) for authenticity tests of ancient ceramics
was
demonstTated by TI.•measurements in seven diHerent typcs
ofprecolombian potsherds reom various paMS of Mexico as well as
modeen pottery, including faked artifacts. The techniques
for
carrying out such tests are described in detail, and the
possi.
bilities of dating, detection of irradiat,"d fakes, and long
.•term
integration of low environmental gamma radiation exposures
by
reading the TI.. stored in the interior of large quartz grains
Insamples of known age, are discussed.
Research sponsored by eh," U.S. Atomic Energy Commision under
contraet withUn ion Carbide Corporarion••Visiting scientisr £rom
Institute of Physics, Universidad Nacional Autónoma deMéxico.
-
FA88
INTRODUCTION
Moreno and Beeke1
Ancient ceramics, such as pottery, ccrracottas, bricks, and
tiles aceoot onIy of scientific imponance for Que underscanding of
rhe cultural andsocioeconomic history of roan, bul a150 frequently
of great artistic, andcons~quendy monetary valuc. A Greek vase, rhe
Metropolitan Museum' scontroversial erarer, recentIy changed hands
for une million dollars. Notsurprisingly, faking of such relics
became rachee widespread. Collectors,art dealers, and museum
directors estima te thar currently 50 to 75% of rhesupposedly oId
precolumbiao pots and s(atue s 00 rhe market are probablyrecent
fakes. Families, and perhaps whole villages are in the business
ofproducjng such fakes, employi~g aEtco rachee sophisticated
techniques. Forexample, use of the same ciay, che same molds, and
the sarne firing techniqueswhich were used in antiquity, and
careful "aging" by burying, etc., makes italmost impossible to
detect forgeries by visual or microscopic inspection,or chemical
and mineralogical anaIysis.
ObviousIy, a reIativeIy simple method for testing the
authenticity ofan art object is of great interest to archeologists,
art' historians, museums,and art dealers and collectors, Measuring
of the radiation-induced thermo.Iuminescence (TI.) signaI which has
been accumulated in quartz and otherconstituen[s of ceramics since
it was manufactured by firing (resuIting in ananneaIing ol all
previous TI... signals) provides such a method. Althoughflrst
suggested twenty year ago1 and studied for more than twelve years
byvarious groups, up to now only about ten laboratories in the
world were e-quipped to carry out such tests. The main reason was
that the pioneers in[his fieId repeatedIy pointed out in their
publications (for sorne good modernreviews, see re£. 2-5) that
highIy sophisticated and correspondingly expensiveTI... equipment
with a vacuum system, pIJoton counter, etc., would be requiredfor
this type ol studies, thus discouraging smaller laboratories and
less ex-perienced investigators. Commercially available
Thermoluminescence Dosi-metey (TLD) instruments have, however, been
considerably improved in recentyears (for a compilation, see rel.
6). It was therefore,one of the primarygoals of this study to
determine whether one ol [he better and more widelyused commercial
TLD readers, the I1arshaw 2000 instrument (several ol whichare
presently in use in Mexico City) could be employed for such
investigations.
The next steps in this investigation about which sorne
preliminary re-sults also are included in this report will
inelude:
l. To check the feasibility ol unrnodified or modified
comrnercial TIJDreaders for actual age determination within
reasonable limits oíaccuracy (t 5-10%). TL dating is gaining
rapidly in importance
-
Tbermoluminescence measurements ••• FA89
because of the well ...known difficulties in obtaining suitable
samples,and disagreemcnt in the results of other absolute dating
techniques,such as l"e, dendrochronology, obsidian hydration, and
fissiontrack dating.
2. Assuming that fakers already have, or may begin soon to
irradiatetheir products in order to simulate a greater TL age, we
intend toestablish simple and reliable methods to distinguish
between re ...cently irradiated and truly ancient samples.
3. It should be determined if there is a stable Thetmally
StimulatedExoelectron Emission (TSEE) signal in those samples,
whetherit can be used for dating, and how it compares with the TL
signals.
4. For the first time, we will al so attempt to use the TL
signal whichis produced by penetrating environmeotal radiatioo to
determineeovironmental radiation levels averaged over hundreds or
thousaodsof years, in order to compare them with current levels and
annualor seasonal fluctuations.
The principIe of this lalter method IS simple .. The age of a
samplecan be determined 00 the basis of the equation
age = TL signal/(av. dose-rate x sensitivity)
With a sample of known age, this equation can be reversed to
result 10
av. dose-rate = TL signal/(age x sensitivity)
We intend to use large (> 100,um) quartz grains which are
koown to containalmost no radioactivity and'can, therefore, be used
as a reliable detectorfor the long-range gamma radiation, after the
alpha radiadon effect at theirsurface has be en eliminated by
etcllÍng away the alpha. exposed surface layerwith hydrofluoric
acid, and the "internar" beta radiation dose has been sub
...tracted.
EXPERIMENT
In aH experiments, a standard lIarshaw 2000 rcader with a
factoryprovidcd slightly extended temperaturc range (500 instead of
400oC)* and a
•T-he heater as supplied with the standard instrument would
actually be capable ofreaching temperatures of almost IOOO"C; the
high-temperature modifications conslstof a few simple measures,
such as a change in the galn of the thermocouple ampli-fier,
repJacement of the temperature meter, and re-ealibration.
-
FA90 Moreno and Becker
plarinum-iridium "permanent'" planchct was employed. Unless
mentionedothcrwisc, a red illumination (reJ-painted light bulbs)
was used during samplepreparar ion, hecausc sorne of (he samples
sccrncd to respond [O the normalfluorescenl Iaboratory light.
Becausc small traces of oxygen are known (Oproduce a ¡arge spurious
signal, lhe rcadout c1uunhcrwas Oushed fOf no lessthan 30 seconds
with pure n¡trogen prior to rcadoul. Longer f1ushing did ootimprove
the results; neither did additional washing of the n¡trogen in
aconcentrated aqueous solution of KOII and pyrogallic aeid (v.'hich
is known toremovc traces of oxygen Croro gases) fol1owedby
desiccation of me gas streamwith calcium chloridc. Ir has beco
suggested7 ro mix the materials with aheat.resistant silicone oil
to minimize environmental effects. We found,however, that the oil
used for lhis purpose, begins to evaporate when heatedfor an
extended period ro 450-500° C, thus possibly contaminating dle
interiorof the TL reader.
The heating rate was ~lffiost linear and about 20°C/sec, the
heatingtime 30 sec., the photomultiplier vohage 750 V, the range
for recording theglow curve usually 10.8 or 10..9A and ((or the
heater set at 9.0) the maximumtemperature"" 450°C. In sorne
instances, samples h.ave also been read at alower maxlffium
temperature of about 400°C (heater set at 7.5). This stillreleases
most of the ancient and modern TL, but reduces dramatically
thebackground signal, and is, therefore, morc suitable for
dctcrmining the arcaunder ,he glnw peak.
As the temperature indicadon of the reader is not considered
very ac-curate, no ternperatures are given in the graphs ((or
orientation purposes: theleh end of the graphs corresponds ro "" 25
° C; the main stable peak of quartzis in the 300-350oC range';
background bccomes substantial a' 350'0 400°C;and the curves break
off al "" 450°C). The recording rate in a plotmaticX -y plouer was
0.5 cm/ sec. Switching to the operation mode of the readerto photon
counting would ha ve be en easy by adding a preamplifier and a
scaler,but it was not considered neccssary because none oC the
disturbiog "bursts"which have been reponed by others2 have been
observcd
Pottery samples wcre taken by eithcr breaking oc sawiog a picce
a fewhundced mg in weight off (rom a larger potshcrd oc figure,
cemoving the upper-most"" 1 mm which have been exposed to intense
light, and crushing thesample gently bctween aluminum foil with
pliers. This was followed by
•Although only a broad single pea~ is seen at the high heating
rate used in these ex-periments. it should be noted that it
actually consists of tWQ peaks at 325 aud375°C, with half-lives at
20°C of -3x 103 and.- -ix 107 years. respectively2 ..
-
FA91
sicving, lhe 107 lo 175¡.Lrn size fraetion being used foe mose
of [he expertAmcnts because ir contaios rhe large quauz graios
which are of inrcres( foccnvironmental measurements of long-range
gamma radiation. Most of rhe darkiron.rich particles in mis size
fraction ("'-10 to 60% of (he sample) were removedby gently shaking
the powdet on a glossy paper sheet between the poles ofa strong
perrnanent magneto
Washing with ..••10% IICI, followed by washing with distilled
waterand alcohol oc aceroue, and drying al 50-80° e inirially
completed rhe basiecleaning procedurca There remained, however,
apparently a layer of impuri-des on rhe partic1e surface which
absorbs pan of rhe TL I¡ght. This canbe removed by etching for 0.5
to 2 minutes with diluted (1-2%) IIF. Etching¡ncreases [he light
ourput in rhe large graios by up co a factor of (hree fOfboth the
ancient and the newly induced TL signals (Fig. 1). In small
grains,
XOUIC1250-950 BC107-1'75Jlm
10
>-•...iñz~•...~j•...~i=«j~
-
FA92 Moreno and Beeker
[he effect is even more pronounced (as an example, rhe effect of
0.5 mino 01etching in"" 1% UF al 20°C is given in Fig. 2 foc
ehe
-
Thermolumi"escence measu,ements ••.
20 I S
-
FA94 Moreno and Hecker'
fluorescent laborawry light al50 produces a low-temperature
peak, probablydue to a trap transíer process). Consequendy, an
overnight (l8-24h) post.irradiation annealing al 800 was adopted
fOf all further cxperimenrs, unl~,snoted otherwise. As can be seco
in rhe following graphs, (his annealingproceJur:l;' ¡ndeed produces
a glow-curve which is almase identical wirh rhe"natural" ancient
signal.
Unlike aH orhcr TI. dating laborarorics who use beta radiation
sourcesfOf low Linear Energy Transfcr (LEY) calibration
irradiations, we employeda 6OCo gamma radiation soucee (close-rate
10 R/sec) fOf rhe un¡form exposureof OUT samples. le was confirmed
in (he experiments rhar
a) [he inherent sensiriviry of rhe quartz graios varies froID
sample tosample; and
b) the supralinearity of the radiadon response does not remaln
thesame when an annealed sample is re" irradiated.
It can be seen in Fig. 4 that irradiadon (followed by the
standard overnightannealing at 800 C) in addition to the ancient
signal yields, by extrapolation,an estimate of the ancient
radiation dose. If the same sample is re" irradiated
ORNL-DWG 73-12066
RE-EXPOSED AFTERANNEALlNG AT 500°C
300020001000EXPOSURE (Ryl
o
GUAMUCHALBLANCO DIFERENCIADO1250-950 BC107-175pm
IEXPOSUREOF UNANNEALEDMATERIAL
5
6
O-1000
en><
-
Thermo/umit'Jescence measuremet'Jls ••• FA95
after annealing at '5000 C, its supralinearity becomes more
pronounced.Alpha irradiations of a few samples were carried out by
coveri~a
'39pu layer electroplatcd on a metal disc by 10!-,m Kimfol (a
polycarbonatefoil), on which (he grains are spread in a monolaycr
and ocassionally movedslightly for homogeneous exposure. The
inhcren( alpha activity of sampleswas measurcd by covering a ,hick
(> 3 mm) laycr of lhc powdered samplewith a 4.'52 cm2 (24 mm
diame(cr) ZnS:Ag scin(illa(or foil, enclosing i( with athin
polycarbonatc foil, and counting overnight witll a photomultiplier
in oneof the devices which arc useJ for roudne bioassay alpha
counting at ORNL.In the Xquic matcrial for example, a nct count of
0.144 cpm would amount,according to thc calibradon factors given in
reL 2, to an internal alpha radi-ation exposure of 1.17rad/year. In
a modero sample (a small ceramic dogmade by highland indians in
Chiapas) the count-rate was almost twice ashigh (0.26 nel cpm).
The internal exposure of the sample to beta radiadon was
detcrminedwi,h a sandwich of a ,hin (- 0.5 mm) layer aboul 5 cm' in
area of CaSO, :Dy,enclosed in a 0.05 mm (- 6 mg/cm') polye,hylenc
foil'. ldenlical pOlsherdscovered both rides of this detector. The
foil stops most of the alpha parti-eles, but permits the penetradon
of almost all the beta radiation. (Accordingto ref. 4, a thicker
layer of 30 rng/cm2 would be even beuer for this purposeand will be
used in future experiments). A referencc sample of freshly
an-nealed CaS04 :Dy was left in a plastic botde in order to
subtract the environ-mental gamma radiadon exposure. Alter 3.75
months of exposure, theCaS04 :Dy was read out as described
earlier8• '111einternal beta dose amountedro 95.6mR/ycar in rhc
Xquic potshcrd, and 58.1 mR/ycar in the modernChiapas dog (aftcr
subtraction of the external background radiadon of54.8
mR/year).
RESULTS AND DlSCUSSION
In Table 1, the samples which are discussed in this repon are
listed.They were selected in order to c()ver a wide range of times
and locations inMesoamerican archaeology, with reladvely
well-established ages. Theyoungest sample(a.Tenochritlan Black 00
Oraogc Azrcc 11I, 1350-1500 A.D.from Lomas de Sanra Cruz, Puebla)
has nor yet beco inc1uded in the study.In addition, numerous
samples from rhe colJecdon ol one of the authors (K. B.)and his
acquaintances were tcsted, wirh mosr samples giving a rarher
clear"oId" or "new" resulto A few amoiguous cases may be
expIainabIe with low
-
FA96 Moreno and Becker.
firing temperatures (insufficient annealing oí deep filled craps
in the quartz,followed by retrapping), or abnormally high
radioactivity in the clay. Re-fining oí old ceramics would, oí
course, al50 make them look new. No ev.kdence for irradiated fakes
has yet been found.
TAHLE 1
List oí Precolumbian Potsherds from Mexico Used in [his
Study
No.
2
3
4
5
6
7
Designation Phase Location Date
Guamuchal Cuadros Aquiles Serdán. 1250-950 B. C.Blanco
Diferenciado Chiapas
Xquic Cuadros Aquiles Scrdán, 1250-950 B.e.Chiapas
Michis Tuztlan Red Deos Aquiles Scrdán, 1350-1250
B.e.Chiapas
Thio Orange Teotihuacan La Ventilla, 350-500 A. D.11-11I
México
Burnished Brown Teotihuacan La Ventilla, 400-
-
Thermol ••",irJescence ",eas ••re",enlS ...
GUAIIUCHAlIIUNCO OlFERENC1AOO12~O-950lIe107-175,_
ORNL-OWG 13~
FA97
>-..;¡;z•••..~-'..•••>>:e-'r
15
10
5
O
NATURAL+600Rr
TEMPERA TURE
Fig.5. Guamuchal '-Blanco diferenciado" (Table 1, No. 1)
l07-175,um largegrain fraction, magnetically c1eaoed and etched for
1 mino in 2% UF;TL signal duriog f¡rst heating of an unirradiated
sample (natural TL),oC the background during the second reading oC
the same sample, andof a pteviously unheated sample whichhas beeo
exposed to 600R of6OCo gamma radiation for calibration purposes,
and stored overnight
at BOOe in order to anneal the low-temperature peaks. Each
curverepresents the average oC at least three individual readings
oC- 20 mg samples, each.
-
FA98
,
___ n_...--TUln..vl ll£OIWI-~!ICJ:)7-11"!ll""
Moreno and Becker
Fig.6. Michis Tuztlan Red (Table 1, No. 3), treated as described
in Fig. 5.
TlOfMJll,C,UlII-m--400-600"0lOT-I1!>I''''
..
o
Fig.7. Teotihuacan IJI-IV Burnished Brown (Table 1, No. 5).
treated asdescribed in Fig. S.
-
Thermolumjnescence measu,emenls ...
CHOLUL.I. POlYCHl¡OlIlE1100-1200 .1.0101-11~ }1m
FA99
JO
.~.-~~5
o
Fig.8. Cholula Polychrome (Table 1, No. 7) I treated as
described In Fig. 5.
TEQTltlJl(;,AN u-mTliIN OIW«;Em-m AO107 175}1m
"
~2••
o
Fig.9. Thio Oraoge from Teotihuacan 11-111(Table 1, No. 4),
ueated as lO Fig. 5.
-
FAIOO Moreno and Becker
10 TEOTIHUACAN lATE NBURNISHEO REO ON BUfF600-7'50
AOJ01-I75vm
oTEMP£RATUR£
Fig. 10. Burnished Red 00 Buff from late Teotihuacan IV (Table
1, No. 6),treated as in F ¡g. S.
TABLE 11
Doses and Dose.rates in Various Mesoamerican Potshercls
Sample No. • Method(2) Dose Estimated Date Dose- rateR Age (y)
(y) (mR/y)
1 a 38S 3070 3223-2993 125b 380
2 a 653 3070 3223-2993 2133 a 565 3270 3323-3223 175
b S80a 133 1470 400~00 120b 220
6 b 324 1300 600-750 2S0• See Table ICor sample description.(2)
a = peak height. b = peak area with maximum temperature .-
400.C.
-
The,moluminescence measu,ements ••• FAIOI
a. the etching procedure as used in dIese tests did not remove
mostof the alpha radiation signal which is accumulated primarily
inthe grain's surface ~extended etching and/or a higher UF
concen.tration would be required to remove most oE the
alpha.exposedsurface layer); and
b. the internal beta dose (mostly due to potassium) has to be
sub.tracted, which would, for example, reduce the high reading
oEsample No. 2 (Xquie) to a more realistie 117 mR/y.
lo Fig. 11, lhe results from a pieee of modelO 'Mexieao pOllery
(a IOYdog from Chiapas made aboul lhree years ago) are giveo.
Allhough lhere isa difference between the first and the second
heating of the sample, no signalcan be observed in the peak region.
In other cases of modern pottery, suchas a faked piece oE "ancient"
pouery wlJich was purchased Jast year in Quito,Ecuador, not even a
trace of a signal could be found in the unirradiated sample(Fig.
12). Other specimen which were assumed to be Eakes turned out,
onlhe olher haod, to be aulheotie. Pieees of Cholula Polyehrome
offered bychildren to tourists in the Cholula area were, for
example, identical in theirTL to the authentic potsherds.
10
>->-¡¡;z•••>-~~ 3•••>;::'"..J•••'"
O
CHIAPAS DOG 11l00ERN)107-173pm
TEIlPERATURE
ORM.-DWG 73-12063
Fig. 11. !L oE. a modern piece oE Mexican pottery, a toy dog
made by Indiansin ChIapas three years ago (sample preparadon as
described in Fig. S).
-
FA 102
10
>-..¡¡z~..;O;J 5..w>¡:oCJ~E
o
EDWXJl FAC£--01-115 ••-\4or
-
FAlO}
of lJ Of Th compounds ,,'ould, of course, be easil)~ detectable.
Srill, rhepossibilit}, of enoes canDor be completely cIcluded in
authenticit)O tests.Foe cIample, valuable old specimen are olteo
bcin,e repaiced .¡lh moderoceramlc. lf a sample is ralíen from a
rerJ.~r("d an"a. !his ;¡,ouM indic3U' amodero falce. Carerul
interpretaríon of Tl~aurhenticiry checks ¡5,
therefore,esscnlial.
SiOz (SEASANO.44-1061"'"A_",mAT5OO"C1
o-lIRAllIAlUl T-1RRAllUlTID,PREJ''PEAI ro I doyJlTllO'C
15
~•..;¡;z~ 10;l;..••..•••>;::cr..•••••• 5
o
\1••11111•,,,
~llACIlIlOtN7'
2 ••• 2'l1.HF
EltItilI7-30 •••5HF
Fig. 13. Glow cunes of 44-106p. quartz «rajos (sea s&Illd,
pl"r-oeaaealcd al 500.C).lIec esposure lo Zl9Pu alplla radiarioa.
(lf:ft) aocI6Deo£_. radiarioo •.• ithout aod aÍl:er elcbiDI .ith 2%
UF f. diffn~t ti.es &1.- 25.C.
Preliminary results on me thermally süna.1ating ~oelectrm
CIIIissioo(TSF.F.) 01
-
FAI04
ACKNOWLEDGMENTS
Moreno and Becker
We appreciatc discussions with numerous colleagues, su eh
as~t.J. Aitken, Oxford University; 11. Barker, British Museum; and
Kaplan.Arte Primitivo, New York. The samples, as well as valuable
advice, wereprovided by J. Li,vak King, Depattment of Arehaeology
of UNAM, andN. Castillo Tejero, Musco Nacional de Antropología,
México .
•REFERENCESI. F. Daniels, C.A. Boyd, and n. F. Saunders, Science
117(953) 343.2. M. J. Aithen and S. J. f1eming, "Thermolumincsccnce
Dosimctry In
Archacological nating" in F.IL Anix, ed., Tapies in
RadíationDosimetry, (Acaclemic Prcss, New York, 1972).
3. II.N. Michael and E.K. Ralph, Daling Techniques lar Ihe
Archae%gisl,(MIT Press, Cambridge, Mass., 1971l.
4. V. Mejdahl, Dosimetry Techniques in ThermoJuminescence
Dating.Riso-Rep. 261, Danish AEC, Riso, Roskilde (972).
5. J. W. Miche1s, Daling ,\/ elhods in Archae%gy, (Seminar
Press,New yo,k and London, (973).
6. K. Becker, Solid-Slale Dosimelry,(CRC Press, Cleveland, Ohio,
1973).7. M. C. lIan, University of Pennsylvania, personal
correspondence.
(Dr. lIan also provided samples of his silicone oil).8. K.
Ilecker, Nue!. lnstr. Me,h. 104 (972) 405.
RESUMEN
•For more complece bibliographies on TL dacing, see reference
liscs in re£. 2-5.