IV. AUTHENTICITY TESTING OF CERAMICS USING THE THERMOLUMINESCENCE METHOD S. J. Fleming Research Laboratory for Archaeology, Oxford, England Introduction Thermoluminescence (TL) is the light emitted by a mineral when heated to 5000C in addition to the inherent red-hot glow which is associated with such high temperature treatment. The TL emission represents a release of stored energy from the mineral. The accumulation of this stored energy occurs over the archaeological burial time in the case of the minerals responsible for the TL observed when pottery fabric is heated. The source of the energy storage is the trapping of electrons in the mineral's crystal lattice at defect sites and impurity ions, following the excitation of these electrons from lattice atoms by radiation from the sur- roundings of the pottery. This radiation environment consists of the trace levels of impurities of uranium, thorium and radioactive potassium (K 40) that are present in the pottery itself and the surrounding burial soil. Some radiation dosage also arises from the cosmic rays which continuously bombard this planet, but the contribution this involves is only approximately 3% (on average) of the total dosage experienced by the pottery minerals. If the mineral is heated, the trapped electrons are forcibly released to wander once more through the crystal lattice until encountering their host lattice source atom (or an identical one elsewhere in the crystal) where they de-excite in recombination with their parent with the emission of that light which we term thermoluminescence. Age Determination The practical observation of this TL is illustrated in the schematic diagram of Fig. 1. During the heating of the sample on a nichrome metal strip the light emitted is detected by a photomultiplier which converts the light signal to an electrical current that feeds the y-axis of a chart recorder. Meanwhile a thermocouple records the temperature of the heating strip and plots the increasing temperature during measurement on the x-axis of the recorder. The curve obtained of TL versus temperature is termed a Glow curve (Fig. 2). Common crystalline minerals will be mixed in the clay used for making pottery. During firing of the pottery (to as high as 10000C) the TL accrued from geological dosage will have been driven off. If a sample of such pottery had been given an immediate reheat it would exhibit no TL. In ancient times the pottery will have fallen into disuse and eventually -23- 24 have become buried. The thermoluminescence measured in present times repre- sents a record of the radiation dosage of the burial conditions. It is assumed that this radiation environment of burial does not change during the archaeo- logical life of the pottery i.e. the annual dose-rate is constant. The age of the pottery follows from: Age (years) = Total radiation dosage during burial Annual dose rate Dose-rate evaluation Firstly let us discuss some of the factors which control the evaluation of the denominator of this equation, and so limit the accuracy of the dating method. Precise dating has only become possible after closestudy of the various forms of radiation the pottery minerals experience. The alpha rad- iation which arises from uranium and thorium travels only 23 microns, on average, through pottery fabric. The beta radiation from these radio isotopes and from radioactive potassium travel approximately 1 mm, on average, in such fabric. These two forms of radiation are the source of internal dosage in the pottery. The surrounding soil supplies an external dosage component through the gamma radiation coming from the natural radioisotopes, effective from as much as 30 cms away from the buried ceramic. As mentioned earlier, cosmic rays also contribute some external dosage. In quantitative terms average values for the different dosage components have been estimated from radioactive analysis of around 60 pottery fragments, and various soils from many different archaeological sites: alpha radiation dose-rate 0.189 rads/year beta radiation dose-rate 0.211 rads/year gamma radiation dose-rate 0.067 rads/year cosmic ray dose-rate 0.014 rads/year Total dose-rate 0.481 rads/year Short range alpha radiation contributes 39% of the total dosage on average. However,a further non-uniformity is present in pottery - the clay matrix con- tains almost all the internal radioactivity of the pottery while the crystal- line minerals embedded in that matrix (such as quartz), are primarily respon- sible for the TL observed, but almost radioactivity-free (Fig. 3a). Grains of such quartz of radius larger than the alpha particles' range will have an internal region which has not been affected by this short-range radiation (see Fig. 3b). By gentle, controlled, crushing it is possible to extract the very small grains of quartz of dimensions around 1-5 microns which are fully affected by the alpha particles emitted by the clay matrix. Such control per- mits precise knowledge of the radiation geometry of the material used for dating to be retained. 25 Further, as the external dosage is around 17% the present estimated accuracy of dating of better than +8% standard deviation for each pottery piece studied from archaeological contexts has only been possible by accur- ate radioactive analysis of the burial soil as well as of the pottery fabric itself. Authenticity testing of ceramics Such high accuracy is neither possible nor necessary in the majority of applications of the TL method to authenticity studies. It is not necessary since we may only require to distinguish between a piece of Chinese T'ang ware between 1100 and 1350 years old (if authentic) and a modern reproduction of such ware which might have an age of up to 60 years corresponding to the initial period of popularity of such T'ang material in the western art markets around 1910. Nor is such high accuracy possible in authenticity studies for the following reasons. Firstly examine the form of sample usually used for this work. A 25 mg. powder sample obtained by drilling is treated as the practi- cal limit to the quantity of material.which can reasonably be removed from a figure without detracting from the artistic value of the piece. The drilling procedure is at once a contradiction of the approach used for precise dating using controlled crushing. Now the sample contains a whole spectrum of grain sizes of crystalline inclusions. The spectrum is expected to differ from that of the original pottery structure as the drill- ing damages and fragments larger crystals and enhances the concentration of finer grains in the powder collected. The 1-5 micron grains are extracted from the drilled powder (by suspension in acetone, utilizing gravity - vis- cosity discrimination of grain size) and a series of samples prepared for TL measurement by deposition of around 1 mg. of the powder onto individual aluminum discs. Some of these 1-5 micron grains may well have originally been part of the inner regions of a large quartz grain. As this region received no alpha radiation dosage in the original radiation geometry of the pottery it acts as a diluent to the archaeological dosage new evaluated by TL measure- ment of these disc-deposited fine grains. The true internal dose-rate suffered by the grains used for measurement must then be evaluated as: f a + S where f <1. Empirically using the data from the drilling of a limited number of pottery pieces of known age an effective value of f has been deduced of around 0.85. By continued study of pottery of various degrees of hardness it is anti- cipated that further information will be accumulated on this point. More obviously damaging to any attempt at precise dating of art ceramics 26 is the incomplete knowledge of the true level of environmental dosage for the ceramic's TL, since the piece has invariably been cleaned free of adhering representative burial soil by the time it begins its museum shelf-life, or appears in an auction sale-room. It is true that we have a useful average working value for the environ- mental dosage of around 0.081 rads/year. To determine the maximum possible age of the ceramic we must know the minimum possible level of the environmental dosage. This is the contribution of cosmic-rays alone, assuming burial of the ceramic in a non-radioactive medium. This amounts to 0.014 rads/year. The minimum possible age of the ceramic is controlled by the maximum possible level of the environmental dosage. The present level used is 0.170 rads/year corresponding to the highest level of activity determined from soil samples studied at the Oxford Laboratory. The possibility that the pottery spent its archaeological lifetime in the close vicinity of a uranium mine is presumed exclusive. Thus for a piece to be authentic it is essential that its minimium possible age determined by TL does not overlap the documented era of forgery of such ware. Further it is essential that the maximum age of a piece must not overlap the purported archaeological period of manufacture if that piece is to be declared an imitation. Clearly the greater the gap between the archaeological period and the period of forgery (or imitation) the more certain it becomes that a definit- ive decision about authenticity can be obtained by TL testing. Plate I illustrates the application of TL dating to authenticity test- ing of two significant art ceramics: (a) a Han Weasel, purported to date be- tween 206 B.C. and 220 A.D., and (b) an Amlash Zebu Bull purported to date circa 1000 B.C. (The Amlash culture originates from the mountains of N.W. Persia overlooking the Caspian Sea. The Zebu is the domestic bull of that area). The approach to dating the Han Weasel is initially to study the TL glow curves in order to establish in what temperature region the electron trapping, which is responsible for the light observed, is stable against thermal decay at ambient burial temperatures. This is achieved by comparison of the ordi- nates of the natural TL glow curve with those of the TL curve induced by lab- oratory radiation, at various temperatures. In the case of the Han Weasel this ordinate ratio becomes constant at and beyond 3500C. Subsequent dating analysis is only carried out on the portions of the curves above this temper- ature. By comparison with the additinal TL induced by 1000 rads of beta radiation 27 at 3750C it is deduced that the natural TL at that temperature was induced by 1100 rads. Radioactive analysis on this figure yielded that the inter- nal dose-rate from its clay fabric was 0.559 rads/year. Using the average value of 0.081 rads/year for the external dose-rate it is thus estimated that the figure experienced a dose-rate of 0.64 rads/year in antiquity. A date of 1720 years for the manufacture of the Weasel follows from these cal- culations. When the factors which determine the maximum and minimum age of the piece are introduced,date limits of 1330 to 2270 years are set, a result consistent with the originally suggested period of the Weasel's manufacture. The date obtained for the Amlash Zebu is not so satisfactory with re- spect to authenticity. The natural TL glow curve is scarcely distinguish- able above the background red-hot glow around 5000C. Yet 150 rads of labor- atory applied beta radiation induce substantial TL light levels in that temperature region. From these curves a limit of 6 rads is set on the archaeological dosage this piece has suffered since the last time it was fired. Radioactive analysis of the pottery clay yield a maximum age of 14 years. The Copenhagen. Amphora It would be a simple task to illustrate this type of analysis time after time for over 200 pieces,many of them with some interesting special feature of either scientific or artistic value. The intrusion of this physical method into authenticity problems in the art world is well served by the example of the Copenhagen 'Pontic' amphora (Plate 2a). This amphora together with some similar pieces appeared on the Swiss art market in the 1950's. Doubts about the group were first voiced by Professor Dohrn in 1966, when several errors in painting were noted in production of the various 'heroic' scenes depicted on the amphorae. Depicted in the upper register of the Copenhagen amphora we find a scene of the death of Hector in which it is possible to suggest confusion between left and right in the artist's efforts. The assailant in the left of the scene is satisfactory, with the breat cuirass depicted simply by two linked circles, the shield held in the left hand while the right hand hurls a spear. In contrast the soldier attacking Hector from behind has the shield apparently held in the right hand while the spear is thrust by the left hand. Oddly, however, the sword sheath is painted on the left side of the warrior creating appreciable problems should he wish to withdraw that sword still holding his shield to the right. Such confusion of left and right together with several other peculiar- ities of amphora decoration led Dohrn to conclude that the piece was a pro- duct of a modern imitatorill-versed in the techniques of ancient Greek art- ists. In response Hampe and Simon, in 1967, responded that these painting errors could well have stemmed from a similar lack of knowledge on the part of the plagiarists working in Etruria where this amphora is believed to have been made. Further these authors noted that even masters of Greek style on the mainland itself were not above such confusion of position, stressing the 28 example of Nikosthenes in the amphora depicting Dionysos and dancing maenads and satyrs (Plate 2b). Note the kantharos held in-the left hand, yet painted. as a right hand. TL dating was attempted upon a cube of 4mm. dimensions removed from the rim of the piece. Glow curves shown in Fig. 4 were obtained which suggested an archaeological dosage of 1365 rads, leading to a date for the amphora of between 420 B.C. and 2000 B.C. The authenticity of the piece was duly support- ed despite the sound artistic criticisms directed at it. It is hoped that this method will continue to yield valuable information which can be turned to advantage in the understanding of the cultural develop- ment of early civilizations. 29 I--- Pulse D; srimrki Ratemeter| . _ Photomultiplier .-,Quartz light guide 0 010, : L -Blue filters Cold Thermocouple Rcre junction _controlled heater Fig. 1. Schematic diagram of the thermoluminescent apparatus (details of the electronics of the system are given in: Aitken, M. J., Alldred, J. C., Thompson, J. (1968) - Proceedings of the Second International Confer- ence on Luminescence Dosimetry, Conf.-680920, 248. U.S.A.E.C. publication). 300 0 200 a, 100102030 00 Temperature 1?C) Fig. 2. A typical thermoluminescent glow curve from a sample of ancient pottery. -Q Ai Indino IU**svU!WnlOWJ4LI m ~~~0 m UP h. ~~~a 0 Ind;no jUvoSoU!WnloWJ4L 30 as N co 0 co 44 0 U) *1-4 Ui) Cu : 4- Co p 0 U) 0) Co Co a) P4 U) co Cu 0) Cu '44 a Plate 2a The Copenhagen Amphora .Plate 2b Amphora., by the Greek painter, Nikosthenes. Dionysos and dancing maenads and satyrs are depicted b 32 Light output CD p H. . . . . ~~. . . . . . . . . . . . OQ . . . . . . . . . . . . . . . . . . c-) C-)~~~~~~~~~~~~C o . . . . .. . . . .S (Cl~ ~ ~ 3 *0 ...... * clQ- . . - (Cl~~*v-@@ 0.... .......... 0 01 0 *** 0 OQO s0 -*1 I- 33 References a. Thermoluminescence Aitken, M. J., D. W. Zimmerman, S. J. Fleming, 1968: 'Thermoluminescent Dating of Ancient Pottery'. Nature, vol. 219, No. 5153, pp. 442-445. Aitken, M. J., 1969: 'Thermoluminescent Dosimetry of Environmental Radiation on Archaeological Sites'. Archaeometry, 11, pp. 109-114. Fleming, S. J., H. W. Moss, A. Joseph, 1970: 'Thermoluminescent Authenticity testing on some "Six Dynasties" figures'. Archaeometry, 12 (1), pp. 59- 70. Fleming, S. J., 1970: 'Thermoluminescence dating by the Inclusion Technique'. Archaeometry, 12 (2), p. 135. b. Art History Joseph, A. M., H. M. Moss, S. J. Fleming, 1970: 'Chinese Pottery Burial Objicts of the Sui and T'ang Dynasties'. (Hugh M. Moss Ltd.) Dohrn, T., 1966: Bonner Jahrbuch, 166, p. 113. Hampe, R., E. Simon, 1967: Jahrbuch des Rlomisch-Germanischen Zentralmuseums Mainz, 14, Jahrgang, p. 73.