A well-established high-resolution chronology from southern mid-latitude sites is needed for providing a consistent framework of the inter-hemispheric climate changes. Archives in New Zealand play an important role for palaeoclimate reconstruction in the Southern Hemisphere (Alloway
Despite the well-established chronology of the North Island, there are few dating studies on the South Island of New Zealand. It was shown that some of the Quaternary dating methods have partial or no utility for chronological studies on loess-paleosol deposits from South Island. The current situation is that in addition to the fact that radiocarbon dating has a limited age range, previous studies have shown that both organic material and charcoal that is interbedded within loess may be contaminated by younger mobile organic compounds (Goh
The OSL dating method has been successfully applied on quartz and feldspars extracted from various sediments on both the Northern and Southern Hemisphere during the last decades. The precision and the accuracy of the ages obtained have increased due to the development of the single-aliquot regenerative-dose (SAR) protocol on quartz (Murray and Wintle, 2000, 2003), as well as the versions of the protocol developed for feldspars, especially the post-infrared–infrared protocols (pIRIR), pIRIR225 (Buylaert, 2009) and pIRIR290 (Thiel, 2011). However, relatively few luminescence dating studies have been reported on sediments of various origins from South Island, New Zealand (e.g. Berger
The aim of this study is to further explore the luminescence properties of quartz extracted from a loess deposit from Canterbury Plains, in the South Island, and to assess the applicability of two pIRIR stimulated luminescence protocols on polymineral fine grains extracted from the same samples.
Apart from the Ross (2.4–2.6 Ma) and Porika (2.1–2.2 Ma) glaciations, New Zealand has undergone at least five glaciations during the Quaternary (Suggate, 1990). The most extensive deposits of loess are located across the foothills of the Southern Alps and in the lowlands of the Canterbury Plains (Yates
The study site (44.018870°S, 171.882054°E) is located in the southern Canterbury Plains, on the eastern side of central South Island (
For luminescence investigations, five samples were collected from the top section of the flank. The uppermost sample was discarded from the analysis due to its proximity to the modern disturbed surface. The uppermost three samples analysed (NZ 2, NZ 3 and NZ 4) were collected at depths of 30 cm, 50 cm and 70 cm, whereas the last sample (NZ 5) was collected from a depth of 140 cm. According to Berger
Luminescence samples were collected in stainless steel tubes and were prepared under subdued red-light laboratory conditions. The material from the end of each tube was removed and used for gamma spectrometry measurements. The material from the inner part of the tubes was used for 63–90 μm, 90–125 μm, 125–180 μm, 180–250 μm quartz and 4–11 μm polymineral grains’ extraction. Treatments with hydrochloric acid (10% concentration) and hydrogen peroxide (10% concentration followed by 30%) were employed for calcium carbonate and organic matter removal. Finer (<63 μm) and coarser (>63 μm) grains were separated through wet sieving. Polymineral fine grains (4–11 μm) were separated from the finer polymineral fraction obtained after Stoke's law, settling and centrifuging in distilled water (Frechen
All luminescence measurements were carried out using two Risø TL/OSL readers (model DA-20), equipped with a classic or automated detection and stimulation head (DASH) (Lapp
The natural luminescence signal of the polymineral fine grain aliquots used for bleaching experiments and residual dose measurements was erased by exposing the disks to window light under natural conditions in March (latitude 46° 47′ N).
Raman spectra have been recorded using a Renishaw InVia Reflex Confocal Raman system with a Leica microscope using the 100× (NA 0.9) long working-distance objective. For excitation, a Cobolt DPSS laser emitting at 532 nm was employed. Wavenumbers’ calibration has been achieved using the internal silicon standard. An edge filter was employed for removing the Rayleigh line and the spectra were recorded in the 50–1840 cm−1 range with 0.5 cm−1 resolution. The acquisition conditions were 1s exposure, 1 accumulation and 100% of the output laser power (≅200 mW). A Rencam CCD was employed for signal detection and data acquisition, and processing has been achieved using the WIRE 3.4 and Origin 9.2 software.
The single-aliquot regenerative-dose optically stimulated luminescence (SAR-OSL) protocol (Murray and Wintle 2000, 2003) was applied on coarse quartz grains. Full details on the protocol are given in
Equivalent doses measured on polymineral fine grains were carried out by applying two elevated-temperature infrared-stimulation methods based on the SAR procedure and the pIRIR225 (Buylaert
High-resolution gamma spectrometry was applied to determine the radionuclide specific activities using a well-type HPGe detector. For 226Ra-222Rn equilibrium to be reached, the samples were stored for 1 month before measurement. Annual dose rates were determined using the conversion factors tabulated by Guérin
We have checked the purity of the quartz mineral extracts and their crystallinity by conducting RAMAN spectroscopy on sample NZ 5 63–90 μm.
To check if the quartz extracted from the investigated samples is suitable for OSL dating, the SAR-OSL protocol labelled in
We have further tested the behaviour of sample NZ 3 63–90 μm quartz by performing a dose-recovery test. Three aliquots were bleached with blue diodes twice at room temperature for 100 s with a 10000 s pause in between the two stimulations. A dose of 100 Gy was then administered and measured as unknown, using the SAR protocol as described in Section 3.3.
Results of dose recovery tests on NZ 3 63–90
aliq 1 | 3789 | 2571± 684 | 6268 | 4472 ± 930 | 1.13 | 0.78 | 0.81 |
aliq 2 | 1422 | 3154 | 0.83 | 0.76 | 0.60 | ||
aliq 3 | 2502 | 3993 | 1.11 | 0.78 | 0.94 |
IR, infrared.
The failure of the quartz from New Zealand to accurately measure a given dose in the SAR protocol, as well as its poor OSL sensitivity, have been previously reported by Preusser
To test whether OSL properties of the samples investigated here can be improved by the application of different irradiation and stimulation steps, we have conducted a dose recovery test on sample NZ 3 63–90 μm under the same conditions as presented above, except that the whole procedure was carried out subsequent to the application of three different treatments:
the repetition of 5 bleach (blue diodes stimulation for 100 s at room temperature)/dose (100 Gy) cycles, irradiating the sample with a dose of 100 Gy followed by heating to 500°C five times and simply by repeating five times the heat treatment that consistied of a ramp heating to 500°C. The results of the dose recovery test performed following these treatments are presented in
Results of dose recovery tests after repeated cycles of bleaching and irradiation and heating and irradiation.
Experiment (i) Bleach/dose (100 Gy) × 5 | 1 | 2685 | 2858 ± 718 | 2869 | 3015 ± 565 | 1.06 | 1.23 | 0.91 |
2 | 1710 | 2119 | 1.14 | 0.89 | 0.66 | |||
3 | 4180 | 4058 | 0.82 | 0.57 | 1.08 | |||
Experiment (ii) Heat to 500°C/dose (100 Gy) × 5 | 1 | 34665 | 49960 ± 12765 | 41173 | 60220 ± 16152 | 1.00 | 0.97 | 0.94 |
2 | 39905 | 47148 | 0.98 | 1.01 | 0.95 | |||
3 | 75311 | 92339 | 1.00 | 0.99 | 0.92 | |||
Experiment (iii) Heat to 500°C× 5 | 1 | 10406 | 20287 ± 6244 | 11556 | 25794 ± 8760 | 0.98 | 0.99 | 0.97 |
2 | 18612 | 24073 | 1.02 | 0.98 | 0.92 | |||
3 | 31842 | 41754 | 0.98 | 1.01 | 0.94 |
IR, infrared.
Contrary to the findings of Preusser
Another important observation is that significant sensitisation of the 110°C TL signals was observed following annealing to 500°C (
The effect of repeating the annealing cycles on the OSL signal is shown in
The effect of varying the annealing temperature was tested, as well. Aliquots of sample NZ 3 63–90 μm quartz were bleached and subjected to the following measurement steps: irradiation with 100 Gy, preheat for 10 s at 220°C, OSL measurement (for 40 s at 125°C), anneal to Ta (ramp heating to a temperature between 225°C and 525°C), irradiation with 100 Gy, preheat for 10 s at 220°C, OSL measurement (for 40 s at 125°C). The OSL response after annealing can be compared to the OSL response before the application of the thermal treatment. This was carried out on three aliquots for each anneal temperature employed and the results of the average ratio between the post annealing and pre-annealing signals are presented in
An exponential increase is reported as a function of annealing temperature, significant sensitisation occurring for heating temperatures above ≅350°C. This sensitisation with heating is intriguing, given the fact that the quartz analysed should have an important metamorphic (sub-greenschist) component from the Eastern greywacke basement rocks. For this reason, the crystal should have been subjected to moderate heating 300–100 Ma. Poor OSL properties of metamorphic quartz were also reported by Guralnik
Sensitisation by heating, and not due to pre-dose activation, was previously reported to occur in partially fired (i.e. annealed only to moderate temperatures) quartz (Yang and McKeever, 1990; Bøtter-Jensen
The dependency presented is suggestive not only of the thermal activation curves presented by Yang and McKeever (1990) for their as-received Arkansas quartz, but also of the annealing temperature dependence of the C band (3.4 eV) for radioluminescence signals reported by Martini
Unlike OSL signals of quartz, the IRSL signals of polymineral fine grains displayed a satisfactory behaviour in the SAR protocol. Equivalent doses were obtained by interpolating the natural sensitivity corrected luminescence signal on the dose response curve constructed for each sample using both pIRIR protocols.
The dose response curves were best fitted by a sum of two saturating exponential functions. The values for the equivalent doses range from 64 ± 2 Gy to 92 ± 23 Gy in the case of pIRIR225 protocol and from 83 ± 3 Gy to 120 ± 6 Gy in the case of pIRIR290 protocol, respectively. The measured equivalent doses along with the results of recycling and recuperation tests for each sample are shown in
It is well known that the post-IR IRSL signals are difficult to bleach, as residual doses of a few grays have been reported even after prolonged exposure to daylight or light in a solar simulator (e.g Buylaert
To quantify residual values, four fresh aliquots from each sample were exposed to window for 30 days to remove the natural signal and quantify the residual level. Residual doses measured using the pIRIR225 protocol range from 3.3 ± 0.4 Gy to 3.7 ± 0.3 Gy, while the values obtained using the pIRIR290 protocol are from 4.1 ± 0.6 Gy to 13.6 ± 2.5 Gy. The values obtained on each sample are presented in
To check whether this is indeed the minimum residual level that can be achieved as function of exposure time, sets of five fresh aliquots of sample NZ5 were exposed to window light for different periods of time (from 0.5 h to 192 h). The results are shown in
Several studies have reported that the magnitude of the residual dose is dependent on the equivalent dose (e.g. Sohbati
We compare the data obtained for 1600 Gy (8.9 ± 1 Gy for pIRIR225 and 18.5 ± 1.7 Gy for pIRIR290) to the values obtained in the bleaching experiment for the natural signals (corresponding to a measured equivalent dose of ≅60–120 Gy), namely average values on all samples of 3.5 ± 0.2 Gy for pIRIR225 and 9.2 ± 2.0 Gy for pIRIR290, respectively. We conclude that (i) different doses accrued by the mineral grain before the bleaching event should not result in dramatically different residual doses and (ii) performing bleaching experiments on natural samples instead of using a modern analogue should not result in an offset of more than a few Grays. Bleaching corrections should not cause significant inaccuracies unless very young samples are dated. However, since a dose dependence is reported, and since it remains questionable whether the conditions in the laboratory fully reproduce the bleaching process in nature, it is advisable to measure a modern analogue for residual dose estimation, as well. Here, a modern sample was collected from a nearby site (latitude 44.014973° S, longitude 171.891569° E). Following the same methodology, an equivalent dose of 7.5 ± 0.5 Gy was measured using pIRIR225 protocol; on the other hand, using pIRIR290 protocol, a dose of 22.8 ± 1.5 Gy was obtained. Based on the results of the laboratory experiments presented above, we consider these residual doses as maximum values. We have chosen to perform residual corrections using both laboratory obtained values, as well as by the modern analogue approach (
Uncertainties were determined following Aitken and Alldred (1972). A 15% water content was assumed with a relative error of 25%. Errors quoted on equivalent doses and specific activities represent random uncertainties. The uncertainties quoted for the ages represent total (random and systematic) errors computed in quadrature. pRIR, post-infrared–infrared protocols.
To further check the reliability of the measurement protocol, a dose recovery test (Murray, 1996; Wallinga
Dose recovery ratios obtained for the pIRIR225 protocol range from 0.97 ± 0.02 (NZ 2) to 1.03 ± 0.01 (NZ 5), while for pIRIR290 they range from 1.02 ± 0.04 (NZ 2) to 1.09 ± 0.05 (NZ 3), suggesting that both pIRIR225 and pIRIR290 protocols can successfully recover the laboratory doses given.
The greatest disadvantage of using feldspars as luminescence dosimeters is the loss of the signal during burial time. This loss of the signal is known as anomalous fading and can be usually quantified in terms of fading rates (percentage of the signal lost per decade; Aitken, 1985). Thiel
On the other hand, in the case of pIRIR225 signals, it remains debatable whether the signals are affected by fading. Some studies reported g-values >1%/decade (Buylaert
Luminescence ages obtained based on pIRIR protocols—along with the dosimetry information—are presented in
Summary of the pIRIR225 and pIRIR290 ages.
30 | 57 ± 2 | 60 ± 4 | 635 ± 17 | 36 ± 1 | 39 ± 1 | 4.2 ± 0.1 | 14 ± 1 | 14 ± 1 | 15 ± 1 | 18 ± 2 | |
50 | 59 ± 2 | 63 ± 4 | 607 ± 17 | 27 ± 2 | 24 ± 1 | 3.4 ± 0.1 | 17 ± 1 | 18 ± 2 | 18 ± 2 | 24 ± 2 | |
70 | 76 ± 3 | 100 ± 7 | 599 ± 16 | 26 ± 1 | 27 ± 1 | 3.4 ± 0.1 | 25 ± 2(*) | 29 ± 3 | 26 ± 2(*) | 32 ± 3 | |
140 | 85 ± 3 | 91 ± 6 | 604 ± 16 | 35 ± 1 | 37 ± 2 | 4.0 ± 0.1 | 22 ± 2 | 25 ± 3 | 23 ± 2 | 28 ± 3 |
The residual dose estimated based on a modern analogue sample
The residual dose measured after laboratory bleaching
Indicates the age is corrected for fading
Only the ages obtained by residual dose correction using the modern analogue technique are represented as function of depth in
The pIRIR-ages increase down-section, except for an age reversal between the sample NZ 4 and NZ 5. Such age reversals have been previously reported by others in the Canterbury region (e.g. Berger
The difficulty of obtaining reliable luminescence chronologies for loess in New Zealand South Island is well known. In this context, the applicability of SAR-OSL protocol on quartz as well as the feldspar-SAR–based pIRIR225 and pIRIR290 protocols on polymineral fine grains was investigated on four loess samples collected from the Southern Canterbury Plains. Although the purity and crystallinity of quartz extracts was confirmed by RAMAN spectroscopy, OSL signals displayed low sensitivity and significant sensitivity-change during the measurement cycles. It was shown that sensitisation of the OSL signal as well as that of the 110°C TL peak could be achieved by annealing in the 300–500°C range, likely by the activation of luminescence centres. However, the same effect could not be achieved by repeated irradiation and bleaching cycles as previously suggested. While equivalent doses could not be securely obtained using quartz, the application of pIRIR225 and pIRIR290 methods were successful in achieving the following outcomes, namely: (i) dose recovery tests resulted in satisfactory results for both these protocols, results being consistent to unity at a 95% confidence level; (ii) fading rates of pIRIR225 signals were generally negligible, with measured values of less than 1%; (iii) although pIRIR290 signals are more difficult to bleach than pIRIR225 signals, constant residual values of ≅4 and ≅10 Gy were achieved after exposure under window light exposure for 48 h in the case of pIRIR225 and 96 h in the case of pIRIR290 signals, respectively; and (iv) a dependence of the residual on the previous given dose was observed; however, for a dose as large as 1600 Gy, the residuals obtained for the two methods are ≅9 and ≅19 Gy, respectively. Thus, the equivalent dose of the sample used in the bleaching experiment should not have a significant effect when the residual dose is determined by laboratory experiments. The ages obtained by the application of the two pIRIR protocols are generally in agreement, with values ranging from 14 ± 1 ka to 29 ± 3 ka, suggesting that loess from the investigated site was deposited during the last glacial maximum.