This study deals with the dating by thermoluminescence (TL) of quartz from six volcanic formations of the Saint Lucia Island (Lesser Antilles Arc)
Subsequently, Berger (1992) showed that TL signals from 4 to 11 ?m grains of glass separated from tephra beds can be used to accurately date volcanic deposits spanning the Holocene to middle Pleistocene age range. Attempts have been made to date Quaternary volcanic glint desktop events in the Chaine-des-Puys by using the red thermoluminescence signal from quartz in pyroclastic ash layers, and sediments heated by lava flows (Pilleyre et al., 1992). With regard to anomalous fading, additional work is required to determine whether storage at elevated temperatures can assist in removal of the fading component of the luminescence signals.
The results of experiments which characterise the optically stimulated luminescence (OSL) signals of an ash sample (BI07-TL-05) from Barren Island are presented. The infrared stimulated luminescence signal decreases to 5% of its initial value when preheated at 150 °C for 10 s, suggesting that the infrared stimulated luminescence signal associated with the 290–390 nm emission in this sample arises from a single trap evicted by heating to 150 °C. The post-IR blue stimulated luminescence emission has greater thermal stability and arises from traps which are emptied by heating to temperatures between 120 °C and 240 °C. Dose recovery experiments demonstrate that a laboratory dose can be reliably determined to within 5% for the post-IR blue stimulated luminescence signal. However, the fading rate for the post-IR blue stimulation is high, and the g-value is estimated to be (9.6 ± 3.5)% per logarithmic decade for BI07-TL-05.
Red thermoluminescence (RTL) of natural quartz grains offers many desirable properties for quaternary chronology and archaeological dating, although RTL measurements suffer from high thermal background due to black-body radiation on heating. To reduce the thermal background to as low a level as possible, a silver sample disc covered with a biotite plate with a sample hole was used in combination with a light guide, cluster heater, optical filters, and photomultiplier tube cooling to — 20 ? C in the present system. As a result, the thermal background decreased from 2 ? 10 4 to 1000 cps in the temperature range 350 – 380 ? C , resulting in a detection limit of approximately 100 cps, corresponding to the RTL signal from a single quartz grain ( 250 – 500 ? m ) irradiated with 4.0 Gy. In addition, application of lower heating rates retarded the thermal quenching effect, resulting in high RTL signals, which are preferable for young or insensitive quartz samples. Using RTL measurements with the single quartz grain method under optimal RTL conditions, comparison of equivalent doses from artificially irradiated single quartz grains to the known dose was within the 20% measurement error. Based on equivalent dose determinations for single quartz grains, large irregularities on non-etched quartz surfaces might be very detrimental to the TL detection process. This result confirms that surface etching treatment is required to achieve reliable dating with high counting efficiency.
Aberrant thermoluminescence dates obtained from primary volcanic quartz
Quartz microcrystals up to one millimetre in size were extracted from dacites and pumice flows and prepared in a way similar to the well-known inclusion technique. The TL properties of these quartz were used to estimate apparent palaeodoses using the multi-aliquot protocol. The quartz TL was studied in three different spectral domains: red, green and ultraviolet/blue. The calculated annual dose-rates yielded a set of 18 age-estimates. For some samples complementary dates were obtained using high temperature TL (HTTL) of plagioclase feldspars. These latter dates combined with previously determined radiocarbon and unspiked K–Ar dates were used to explore the validity of ages computed from the TL of quartz. Individual values for quartz appear to be scattered and do not match ages deduced from 14 C, unspiked K–Ar or HTTL on plagioclase dates. These results indicate that when conventional TL methodologies derived from the inclusion method are applied to volcanic quartz major dating problems are to be expected.