Chronology of the Huxushan Paleolithic Site in South China: Inferred from Multiple Luminescence Dating Techniques

Seven sediment samples were collected from the stratigraphic profile of the south wall of the excavation pit for luminescence dating at the Luminescence Dating Laboratory of Peking University. The samples were taken by hammering stainless steel tubes (5 cm in diameter and 20 cm in length) horizontally into the vertical wall of freshly cleaned exposures. The tubes were then covered with aluminum foil and sealed with opaque tape.

Sample preparation for D_e determination was done under subdued red light in the dark room using the procedure described in Zhang and Zhou (2007). Carbonates in the samples were dissolved in 10% HCl acid, and organic matter was oxidized in 30% H₂O₂. The samples were then separated by wet sieving to attempt to obtain medium-grained (45–63 μm) fractions, but failed to obtain enough amount of the grains for D_e measurements. The poly-mineral fine (4–11 μm) grains were obtained by settling after Stokes’ Law. To obtain fine-grained quartz, some of the polymineral fine-grained fractions obtained were further purified with silica saturated fluorosilicic acid (H₂SiF₆) for three days followed by treatment with 10% HCl to remove any fluorides produced. The purity of the quartz extracts was checked by infrared stimulation followed by an OSL measurement. The IRSL signals detected were less than 0.1% of the OSL signal for all seven samples, indicating that feldspar contaminants were completely removed. The fine quartz and polymineral grains were respectively dispersed in acetone and then deposited onto 0.97 cm diameter aluminum discs for D_e measurements.

All luminescence measurements were performed in an automated Risø TL/OSL-20 luminescence reader. Irradiations were carried out using a ⁹⁰Sr/⁹⁰Y beta source equipped in the reader. Quartz grains were stimulated using blue LEDs (λ = 470 ± 30 nm), and the OSL signal was detected using an EMI 9235QA photomultiplier tube (PMT) with a 7.5 mm Hoya U-340 filter in front. Poly-mineral grains were stimulated using IR diodes (λ = 870 ± 40nm) with the IRSL signal collected through the same PMT with a Schott BG39/Corning 7-59 filter combination in front (Bøtter–Jensen et al., 2003).

Three procedures in Table 1 were employed to determine the equivalent dose of quartz extracts. In the SAR protocol (Murray and Wintle, 2000, 2003), regenerative beta doses include a zero-dose used for monitoring recuperation effects and a repeat of the first regeneration dose used for checking the reproducibility of the sensitivity correction (i.e. recycling ratio). Test doses were irradiated immediately after measuring natural or regenerative-dose OSL signals (L_i) to produce the OSL signal (T_i) for correcting sensitivity changes due to preheating/OSL measurements. D_e value is obtained by projecting the corrected natural OSL intensity (L_N/T_N) onto the dose-response curve (DRC) constructed using the corrected regenerative-dose OSL intensity (L_i/T_i). Preheat of 250°C for 10 s (determined from preheat plateau tests, see the following section) and a cut-heat of 160°C were applied. A 280°C blue LED stimulations for 40 s was conducted at the end of each round of the measurement to reduce recuperation. The SMAR protocol (Zhou and Shackleton, 2001; Lu et al., 2007) was also used for the quartz extracts under the same conditions as in the SAR protocol. Six regeneration doses were performed, and their corrected OSL signals were used to construct DRC. Three aliquots were used for each regeneration dose.

In order to check the upper age limit of the quartz OSL-SAR and OSL-SMAR methods for our samples, the samples were also measured using the TT-OSL procedure proposed by Wang et al. (2006) (Table 1). The TT-OSL can be separated into two parts, the recuperated OSL (Re-OSL) and basic transferred OSL (BT-OSL). It is shown that the Re-OSL signal continues to grow with increasing dose up to 20, 000 Gy, while this is not the case for the BT-OSL signal. Using the equation L_Re-OSL = [L_TT-OSL/T_TT-OSL] – [L_BT-OSL/T_BT-OSL], Wang et al. (2006) showed that the ages for loess using both the OSL signal and the Re-OSL signal were consistent from modern to ~ 130 ka, demonstrating that Re-OSL could be used for dating. Here six aliquots were measured to obtain the intensity of natural TT-OSL (L_TT-OSL) and basic transferred OSL (L_BT-OSL), respectively. Test doses were used immediately to obtain T_TT–OSL and T_BT–OSL signals in exactly the same way as in the SAR protocol. The Re-OSL D_e value was then obtained by projecting the corrected natural L_Re-OSL intensity to the DRC constructed using the corrected regenerated L_Re-OSL intensities. For quartz extracts, the initial 0.64 s signal minus the last 3.2 s integral as a background in the decay curves was used for dose calculation.

Two procedures listed in Table 2 were used to determine the equivalent doses of the polymineral fine grains. In the MET-pIRIR protocol proposed by Li and Li (2011, 2012a), the preheat of 320°C for 60 s was applied after both regenerative and test dose to clear great influence from residual signals from measurements (steps 7 and 14). The regenerative and test dose IRSL signals after preheat were measured for 100 s at stimulated temperatures of 50, 100, 150, 200, 250, 290°C, respectively. A 100 s IR bleaching at 325°C was applied at the end of each cycle to clear recuperation. The two-step pIRIR protocol was also employed to determine the equivalent doses of the fine-grained polymineral fractions (Thomsen et al., 2008; Thiel et al., 2011; Buylaert et al., 2012), in which the polymineral aliquots were preheated at 320°C for 60 s, and IR stimulation time is 200 s. The first IR stimulation temperature was 100°C (IRSL signal) determined by D_e plateau tests (see the following section for details) (Li and Li, 2012a), and the subsequent IR stimulation temperature was 275°C (post-IR IRSL signal, pIRIR_{(100, 275)}). The IR clean-out at the end of each SAR cycle was carried out at 325°C for 200 s. The first 2 s signal minus a background estimated from the last 50 s in the decay curves was used for dose calculation. It is noted that the experiments for anomalous fading and residual doses were not carried out (see the following section for details).

The annual dose estimation was based on U, Th and K concentrations. U and Th concentrations were determined by laser ablation inductively coupled plasma mass spectrometer (ICP–MS), and K concentrations were determined by a Wavelength Dispersive X-Ray Fluorescence spectrometer. Water contents (weight water/dry sediment weight) for all samples were measured in the laboratory, but the experiment results (13–14%) were probably lower than the natural values since the sediment section had been exposed to air for a long time. Therefore, we prefer to use a higher value of 23% obtained by similar samples from the Fengshuzui site (Zhang et al., 2018) close to the studied site. U, Th, K and water contents were assumed to have an uncertainty of 5% to cover analytical errors. The cosmic ray contribution for each sample was calculated as a function of depth, altitude, and geomagnetic latitude (Prescott and Hutton, 1994). The α-efficiency was taken as 0.04 ± 0.02 for quartz and 0.08 ± 0.02 for polymineral grains, respectively (Lai and Brückner, 2008). Finally, annual dose rates and ages were calculated using the DRAC calculator (v1.2) (Durcan et al., 2015).

Examples of quartz OSL decay curves and SAR- and SMAR-OSL DRCs for our studied samples are illustrated in Fig. 2. Note that the DRCs for all samples were fitted with a single exponential saturation (SE) or saturating exponential plus linear function (SEPL) using the Risø luminescence Analyst version 4.53 software. The shine-down curve indicates that the quartz natural OSL signal is reduced to 4.3% of its initial intensity after 0.96 s of blue simulation at the sample temperature of 125°C, demonstrating that the quartz grains are very bright, and the OSL signal is dominated by the fast component. The DRCs demonstrate that the OSL signals are not saturated at a dose up to about 1000 Gy, and have the characteristic saturation doses (D₀1 and D₀2) of 35 and 464 Gy, if the curve is fitted with a double saturating exponential function. The shapes of the SAR- and SMAR-OSL DRCs are identical when regeneration dose is less than about 400 Gy if errors are considered. The recycling ratios for all samples using the SAR protocol are within the range of 1.0 ± 0.1 and the recuperation less than 5%, suggesting that the sensitivity correction for these samples is successful and thermal transfer negligible.

Fig. 2

Preheat plateau tests were carried out to determine a suitable preheat temperature for quartz aliquots. The D_e values of aliquots from sample L3478 were measured using the SAR procedure in Table 1 with different pre-heat temperatures ranging from 180 to 280°C at 20°C increments. Three aliquots were measured at each preheat temperature. The results shown in Fig. 3a indicate that there is no dependency on D_e values on preheat temperature in the range of 200–280°C. The recycling ratios are within the range of 0.9 to 1.1, and the recuperation is less than 2% for all samples (Fig. 3b), indicating that the preheat temperatures of 200–280°C are suitable for D_e determination on the quartz aliquots.

Fig. 3

Dose recovery tests were also performed on the same sample as used in the preheat plateau tests to confirm the results of the preheat plateau tests. Prior to irradiating a laboratory dose, the natural OSL signals were removed by two 100 s blue light (blue LED) exposures at room temperature, separated by a 10 ks pause to allow any photo-transferred charge in the 110°C thermoluminescence trap to decay. The bleached aliquots were then irradiated with a given dose approximately equivalent to the D_e values of the sample, and the irradiated aliquots were measured using the SAR procedure in Table 1. Three aliquots were measured for each preheat temperature, and the results are shown in Fig. 4. It shows that the ratios of measured dose to the given dose (dose recovery ratio) are closest to unity at the preheat temperature of 250°C. Therefore, the preheat of 250°C for 10 s was used in the quartz OSL-SAR, OSL-SMAR and TT-OSL SAR procedures.

Fig. 4

Examples of the decay curves of the natural TT-OSL (Fig. 5a) and BT-OSL (Fig. 5b) signals and the regeneration dose TT-OSL (Fig. 5c) and BT-OSL (Fig. 5d) signals are shown in Fig. 5. It is evident that there was a significant rise in the intensities of L_TT-OSL with increasing doses, and this is not the case for L_BT-OSL. The corrected recuperated intensity, L_Re-OSL, was obtained by the equation (mentioned above). The corresponding DRCs and one of the L_Re-OSL signals are shown in Fig. 5e and 5f, respectively, for which the recycling ratios fall into the acceptance range of 0.9–1.1, and recuperation is 1.01 ± 0.03. The DRC in Fig. 5f was fitted with a single saturating exponential, the D_e value obtained is much less than the D₀ value of about 470 Gy for the curve. In order to determine the residual TT-OSL signal of the studied samples, the natural aliquots of sample L3481 were exposed to light from a solar simulator (SOL2, Hönle UVACUBE400) for 1 min, 9 h and 2 days, respectively, and then measured using the TT-OSL procedure in Table 1. Three aliquots were used for each bleaching time. The residual TT-OSL D_e values of 230.5 ± 10.1 Gy, 70.0 ± 5.7 Gy and 33.4 ± 1.6 Gy were obtained for the three bleaching times, respectively. The dose of 33.4 Gy was taken as a residual TT-OSL dose for subtraction for all samples.

Fig. 5

The dose rates and D_e values obtained using the above three procedures on fine-grained quartz are summarized and listed in Table 3. It shows that the calculated dose rates range from 4.3 to 3.5 Gy/ka, and the D_e values varying from about 269 to 460 Gy obtained using the three procedures. The ages were obtained by dividing the D_e values by the dose rates and listed in Table 3 and displayed in Fig. 6. The top sample (sample L3475) was dated to SAR-OSL 67 ± 4 ka, SMAR-OSL 66 ± 4 ka and TT-OSL 62 ± 4 ka, and the bottom sample (sample L3481) to SAR-OSL 133 ± 9 ka, SMAR-OSL 123 ± 9 ka and TT-OSL 127 ± 9 ka. The figure demonstrates that the SAR-OSL, SMAR-OSL and TT-OSL ages for each sample are consistent within errors except for sample L3478, and they are in stratigraphical order.

Fig. 6

All the samples show relatively dim IRSL, pIRIR and MET-pIRIR signals, and their natural intensities decrease as depth increases (Figs. 7–9). Note that the intensity of the natural IRSL signals measured at the sample temperature of 250°C are much larger than at lower temperatures (Fig. 7). Fig. 7 demonstrates that the MET-pIRIR signals are very dim and close to the noise background, and this is especially the case for the test dose (91.5 Gy) signals. The dim MET-pIRIR signals result in very large uncertainties in MET-pIRIR D_e values. In this case, the MET-pIRIR D_es of the samples were not further determined.

Fig. 7

Fig. 8

Fig. 9

In the two-step pIRIR procedure (Table 2), the effect of the first IR stimulation temperature on pIRIR D_e values was tested using the temperatures of 50, 100, 150 and 200°C, respectively, and the results shown in Fig. 10 indicate that the pIRIR D_e values are independent of the first stimulation temperatures between 50 and 150°C, and the recycling ratios are close to unity and the recuperations less than 5%. Therefore, here the sample temperature of 100°C was used for the IRSL signal measurements before measuring the pIRIR signals at 275°C. Two pIRIR_{(100, 275)} DRCs for the top sample (L3475) and the bottom sample (L3481) shown in Fig. 9 demonstrate that the two samples are very far from being naturally saturated as the D₀ for samples L3475 and 3481 are 348 Gy and 402 Gy, respectively, and the recycling ratios are scattered but still close to unity. The pIRIR_{(100, 275)} D_e values and ages obtained are summarized in Table 3 and the ages also shown in Fig. 6. The pIRIR_{(100, 275)} ages obtained range from 77 ± 7 to 174 ± 17 ka, and they are not in strict stratigraphical order. The pIRIR_{(100, 275)} ages of samples L3475, L3479, L3480 and L3481 are in agreement with their quartz OSL (SAR-, SMAR- and TTOSL) ages, if considering 2σ uncertainties. However, the pIRIR_(100,275) ages of samples L3476, L3477 and L3478 are much larger than their corresponding quartz OSL ages. It is noted that the analysis of residual pIRIR_{(100, 275)} dose and fading corrections have not been performed for these samples.

Fig. 10