Li Jinwei ^{1, 2}, Long Dengchuang ¹, Sun Yinuo ¹, Wu Lin ¹, Tang Min ¹

(1. Research Center of Ancient Ceramic, Jingdezhen Ceramic University, Jingdezhen 333001, Jiangxi, China; 2. Jiangxi Ceramic Heritage Conservation and Imperial Kiln Research Collaborative Innovation Center, Jingdezhen 333001, Jiangxi, China)

Extended abstract:

[Background and purposes] Thermoluminescence (TL) pre-dose dating, relying on the 110 ℃ TL peak of quartz, has become a core technique for absolute age determination of high-fired ancient porcelains. Beta irradiation is indispensable in pre-dose measurement, while 90Sr and 14C are two typical β sources with distinct half-lives, energy spectra and dose rates. However, systematic comparisons of their response discrepancies and influencing mechanisms in porcelain TL pre-dose testing remain insufficient. In this study, two well-dated Jingdezhen porcelain samples (Song Dynasty saggar and Yuan Dynasty egg-white porcelain) were selected to compare TL pre-dose results under 90Sr and 14C irradiation systems, in order to reveal the causes of data variations, verify the feasibility of 14C source and provide methodological references for archaeological dating and authenticity identification of ancient porcelains.

[Methods] Specimens with particle size of 125–200 μm were prepared from porcelain bodies. Energy-dispersive X-ray fluorescence (ED-XRF) was employed to determine the main chemical compositions of the sample bodies. TL pre-dose measurements were conducted in the two systems using 90Sr (99 mGy·s−1) and 14C (4 mGy·s−1) as β sources, with test dose of 198 mGy, calibration dose of 5940 mGy, thermal activation at 650 ℃ and heating rate 5 ℃·s-1. Paleodoses were calculated by using the pre-dose saturation index extinction method and ages were converted with an average annual dose rate of (4.9±0.85) mGy·a−1. Paleodose values, age results, data dispersion and TL response characteristics were quantitatively compared and statistically analyzed by using Welch t-test and Levene's test.

[Results] Both 90Sr and 14C β-irradiation systems yielded calculable and reliable pre-dose signals and both samples exhibited distinct 110 ℃ TL peak pre-dose responses. XRF results showed that the SiO₂ content was 73.03 wt.% for the Song Dynasty sample (SD) and 70.12 wt.% for the Yuan Dynasty sample (YD), indicating obvious differences in matrix composition and residual quartz distribution. For the SD sample, the average paleodose was (3.62±0.30) Gy in the 90Sr system and (3.57±0.59) Gy in the 14C system, corresponding to ages of (737.9±141.9) a and (728.4±174.5) a B.P., respectively. For the YD sample, the average paleodose was (3.31±0.22) Gy in the 90Sr system and (2.86±0.60) Gy in the 14C system, with ages of (675.1±125.6) a and (582.7±158.0) a B.P., respectively. Age values from both systems were generally consistent within error ranges. The 14C system showed slightly larger data dispersion and marginally younger ages, mainly owing to its lower dose rate and longer irradiation duration, which intensified the time-dependent decay of the short-life 110 ℃ TL peak of quartz. Differences in body compositions, such as SiO₂, Na₂O and Al₂O₃ contents, further affected response stability in the 14C system. Despite the variations, 14C source results fully satisfied the requirements of authenticity identification. Compared with 90Sr (half-life ~29 a), 14C (half-life ~5730 a) offered superior long-term stability, lower maintenance costs and weaker radiation safety demands.

[Conclusions] Both 90Sr and 14C β sources are applicable for TL pre-dose dating of ancient porcelains, yielding age results consistent within error ranges. The disparities between the two systems originate from dose rate, irradiation time and body compositions, which collectively regulate the decay behavior of the quartz 110 ℃ TL peak. The 14C system provides a low-cost, stable and user-friendly alternative for TL pre-dose measurement, with high practicability in archaeological dating and authenticity verification. For high-precision age determination, further studies on the decay kinetics of the 110 ℃ TL peak and corresponding time-decay corrections are recommended to improve measurement accuracy and reliability.

Key words: ancient porcelain; thermoluminescence dating; pre-dose method; 90Sr source; 14C source