The first ESR dating attempt of optically bleached quartz grains extracted from Quaternary sediment was published more than 30 years ago (Yokoyama
Overview of the different groups working on ESR dating of quartz over the last decades (non-exhaustive list). The description of the experimental setup is based on the information provided in the respective publications cited in the last column.
Museum National d’Histoire Naturelle, Paris, France | Bruker EMX (107 K) | Voinchet |
University of Cologne, Cologne, Germany | Bruker ESP 300 (115 K) | Beerten |
Centro Nacional de Investigacion sobre la Evolucion Humana (CENIEH), Burgos, Spain | Bruker EMXmicro 6/1 (90 K) | Duval |
McMaster University, Hamilton, Canada | JEOL FA-100 (77 K) | Burdette |
China Earthquake Administration, Beijing, China | Bruker ER041XG (77 K) | Liu |
Okayama University of Science, Okayama, Japan | JEOL PX-2300 (81–84 K) | Toyoda |
Consequently, the present work aims at contributing to fill this knowledge gap. Taking advantage of the wide range of pieces of equipment available at the National Centre of Human Evolution (CENIEH), we measured a couple of quartz samples using different combinations of ESR spectrometers and resonators. We evaluated their performance, sensitivity and then compared the resulting ESR measurements and DE estimates.
We selected two quartz samples (BIZ1201, BIZ1202) from the Early Pleistocene site of Bizat Ruhama, (Northern Negev, Israel). Sample preparation was performed following Duval
Quartz extracts from each natural sample were divided into 14 multiple grains aliquots. Twelve of these aliquots (D1 to D12) were irradiated using a calibrated 137Cs Gammacell-1000 gamma source (dose rate = 7.16 Gy/min for SiO2) to the following doses: 98, 196, 393, 835, 1376, 2261, 3932, 6390, 9830, 16711, 24575 and 39320 Gy. The non-bleachable residual ESR signals of the Aluminum centre were obtained after exposing one aliquot of each natural sample in a SOL2 (Dr Hönle) solar light simulator for about 1460 h.
ESR measurements were performed in the ESR dating laboratory at CENIEH with two different ESR spectrometers (Bruker EMXmicro 6/1 and Bruker Elexsys E500) and three different resonators (two standard ER4102ST rectangular resonators and one cylindrical high sensitivity SHQE resonator). Full description of the three experimental setups employed in the present study can be found in
Description of Experimental Setups #1, #2 and #3.
Resonator model | ER4102ST(1): standard cavity | ER4102ST(2): standard cavity | SHQE: high sensitivity cavity |
Low temperature system | ER4141VT Digital Temperature control system | ER4131VTM Digital Temperature (stabilized via PID) | ER4131VTM Digital Temperature (stabilized via PID) |
Chiller | Thermo Scientific NESLAB ThermoFlex 3500 (water-cooled) | Riedel PC-100-02 (water-to-air cooled) | Riedel PC-100-02 (water-to-air cooled) |
Acquisition software | Bruker Win EPR | Bruker Xepr | Bruker Xepr |
In order to obtain comparable data, the Al and Ti centres were measured with the three experimental setups using similar acquisition parameters (
Acquisition parameters employed for the three experimental setups during the ESR measurements of Al and Ti centres.
ESR spectrometer | EMX 6/1 micro | Elexsys E500 | EMX 6/1 micro | Elexsys E500 |
Resonator model | ER4102ST(1) | ER4102ST(2) SHQE | ER4102ST(1) | ER4102ST(2) SHQE |
Microwave power (mW) | 10 | 10 | 5 | 5 |
Sweep width (mT) | 18 | 18 | 18 | 18 |
HF modulation (kHz) | 100 | 100 | 100 | 100 |
Modulation amplitude (mT) | 0.1 | 0.1 | 0.1 | 0.1 |
Number of points | 1024 | 1024 | 1024 | 1024 |
Conversion time (ms) | 40 | 40 | 60 | 60 |
Sweep time (s) | 40.96 | 40.96 | 61.44 | 61.44 |
Number of scans | 1 | 1 | 2–4 | 2–4 ST resonator 1–2 HSQE |
Receiver Gain | 3169.8 | 60 dB | 25178.5 | 81 dB |
Temperature (K) | 90 | 94 | 90 | 94 |
In order to ensure similar resonance conditions in the resonator for all the aliquots of a given sample (
Each sample aliquot was measured three times after ∼ 120º rotation in the resonator in order to take into account the uncertainty on the angular dependence of the ESR signal due to sample heterogeneity. Then, measurements were repeated three times over distinct days to evaluate the reproducibility of the DE values. The procedure was carried out for both the Al and Ti centres of the two samples and with the three different experimental setups.
The intensity of the ESR signal of the Al centre was extracted from peak-to-peak amplitude measurements between the top of the first peak (g = 2.0185) and the bottom of the 16th peak (g = 1.9928) (Toyoda and Falguères, 2003). Following the conclusions from Duval and Guilarte (2015), the ESR intensity of the Ti centre was measured by considering the peak-to-baseline amplitude around around g = 1.913–1.915 (Option D), as shown in
For a given aliquot, each ESR intensity was corrected by the corresponding receiver gain value, the number of scans, aliquot mass and a temperature correction factor (see Section 3 – Evaluating the performance of the different experimental setups – Influence of the temperature on the ESR signal of the Al centre in quartz). Then, a mean ESR intensity and the associated standard deviation was derived from the three angular measurements of a given time (day 1, 2 and 3).
ESR dose evaluation for both Al and Ti centres was performed following the standard ESR dating procedure developed at CENIEH (e.g., Duval
In ESR dosimetry/dating, it is crucial to ensure constant experimental conditions, in order to eliminate, or at least minimize, the variations of the ESR signal that are induced by the system, and to make sure that only the variations derived from the sample itself are recorded. The main external factors that could impact measurement repeatability are: (1) the vertical position of the sample within the resonator and (2) the stability of the measurement temperature over time.
Because the resonator has a non-uniform volume sensitivity distribution (Barr
The procedure was repeated several times with different quartz grains samples. The signal intensities distribution of all the resonators used in this work: two Standard rectangular (ER4102ST) and SHQE resonators are represented in
Results show a bell-shaped pattern for the three resonators, with a maximum signal intensity achieved at the cavity centre. The standard resonators (used in setup #1 and #2) show a very similar behaviour, with an intensity decrease of 4–5% for the first 2 mm around the resonator centre, and then up to 60–70% for a distance of 8 mm. Interestingly, the HSQE profile shows less curvature, with an intensity decrease of only 1.2–1.8% and 20–24% for a distance of 2 and 8 mm from the resonator centre. This indicates that a small variation of the sample vertical position within ±1 mm around the resonator centre will have much less impact on the resulting ESR intensity (1.0–1.5% for the standard resonators
To sum up, these results imply that (i) the sample centre should match the cavity centre in order to obtain the highest ESR intensity, and (ii) the sample should be systematically placed at the same position within the resonator to achieve repeatable measurements. This could be achieved by measuring the distance (d, see
Following our standard measurement procedures (e.g. Duval, 2012; Duval and Guilarte-Moreno, 2012 and Duval and Guilarte, 2015), around 1 hour for the Al centre and 2–3 hours for the Ti centre are usually needed to complete the measurement of one sample formed by 12–14 aliquots. It is therefore crucial to make sure that the experimental setup is sufficiently stable in time to allow highly repeatable measurements at low temperature (<100 K) over several hours.
Consequently, the stability of setup #2 (see
Initial test showed that setup #2 could reach a minimum temperature of 90–91 K, but with questionable stability. Consequently, stability tests were performed at temperatures between 92 and 95 K for 1–5 hours (
Repeated long measurements with both setups #1 and #2 do not show any systematic trend or significant drift in the ESR signal intensity with time. However, setup #1 offers comparatively higher performance than setup #2: it allows not only to reach lower temperatures (85–90 K) but also displays slightly higher stability over time. Based on our stability tests, 90 K was the temperature chosen to perform measurements with setup #1 and 94 K for those carried out with setup #2 and #3. These temperatures are low enough to obtain the spectra of Al and Ti centres with enough resolution and intensity and to ensure that the temperature systems are stable for several hours.
In addition, for both setups, it is advisable to wait for a minimum period of 20–30 minutes once the systems reach the required temperature before starting the ESR measurements, in order to avoid variations in the temperature and gas flow originated from the stabilization of the systems.
It is known that when measuring quartz grains (Al and Ti centres), the lower the temperature of the measurements, the better the resolution of the signal and the higher the signal intensity. However, the relation between the signal intensity and the temperature may be setup dependent. Consequently, we performed 2D scan measurements by acquiring 1 scan every 1 K from 92 to 130 K. This procedure was carried out with setup #2 and was repeated with two different quartz samples (some examples of the ESR spectra are shown in Supplementary information
The resulting fitted curve was then compared to the data previously obtained by Duval and Guilarte-Moreno (2012) for setup #1 (
To conclude, when using a variable temperature unit, it is not uncommon to observe some slight temperature variations of up to 0.3 K among the measurements of the different aliquots of a given sample. For a given measurement, the temperature of the resonator is saved together with all the other acquisition parameters. Consequently, given the dependence of the ESR intensity on the temperature, it is recommended to systematically correct the corresponding ESR intensities in order to eliminate the small bias caused by variations of temperature during the measurement. For setup #1, a variation of about 0.2 K in the temperature range of 90 K may induce a systematic error of about 2% in the ESR intensity (Duval and Guilarte-Moreno, 2012).
The results reported in the next sections were obtained considering these preliminary studies. Each ESR tube was placed in the centre of the resonator to ensure that all the aliquots of a sample were measured under identical conditions, and all the ESR intensities were corrected by the corresponding temperature factor.
We evaluated and compared the sensitivity of two of the most common types of resonators: the standard resonator ER4102ST and the high sensitivity cylindrical resonator (HSQE).
These data show the interest of using a high sensitivity resonator, such as HSQE, for quantitative ESR determination, especially when working with paramagnetic centres that generally have low signal intensities (
The Al and Ti centre DRCs of samples BIZ1201 and BIZ1202 were measured with the three different experimental setups and following the conditions detailed in Materials and Methods. The main objective was to gain knowledge about the precision of the ESR measurements using the three different experimental setups and to study its impact on the DE values.
Each aliquot of a given sample was measured at three different rotations in the resonator, resulting in the calculation of a mean ESR intensity, an associated standard deviation and a coefficient of variation (CV). Therefore, for each day of measurement, a mean CV value can be derived from the individual CV obtained for each aliquot of a given sample (see column 5 for Al centre, and column 7 for Ti centre in
Variability of the ESR intensities of samples BIZ1201, BIZ1202 (Al and Ti centres) derived from experimental setups #1, #2 and #3.
Setup #1 | 1 | 13 | 0.82 | 1.67 | 3.80 | 2.59 | |
2 | 13 | 1.05 | 3.63 | ||||
3 | 13 | 1.04 | 3.62 | ||||
Setup #2 | 1 | 13 | 0.79 | 2.02 | 2.57 | 4.96 | |
2 | 13 | 1.25 | 3.32 | ||||
3 | 13 | 1.76 | 3.57 | ||||
Setup #3 | 1 | 13 | 1.31 | 1.28 | 3.82 | 4.15 | |
2 | 13 | 1.58 | 2.66 | ||||
3 | 13 | 1.54 | 4.08 | ||||
Setup #1 | 1 | 13 | 1.37 | 1.21 | 4.08 | 3.78 | |
2 | 13 | 1.47 | 3.65 | ||||
3 | 13 | 1.01 | 2.59 | ||||
Setup #2 | 1 | 13 | 1.07 | 1.84 | 2.79 | 3.06 | |
2 | 13 | 1.45 | 2.84 | ||||
3 | 13 | 1.19 | 2.39 | ||||
Setup #3 | 1 | 13 | 2.29 | 1.81 | 3.00 | 5.00 | |
2 | 13 | 1.63 | 2.54 | ||||
3 | 13 | 1.47 | 2.92 |
Considering the Al centre in sample BIZ1201, the CV values obtained for each day of measurement range from 0.82 to 1.05% (mean value 0.97 ± 0.13%) for setup #1, from 0.79 to 1.76% (mean value 1.27 ± 0.48%) for setup #2 and from 1.31 to 1.58% (mean value 1.48 ± 0.15%) for setup #3 (
Regarding the Ti centre, the ESR intensity variabilities are higher than those of the Al centre for both samples and with the three experimental setups. This observation was expected (Duval and Guilarte, 2015), as this is simply the consequence of the much lower S/N response for the Ti centre in comparison with the Al centre. The ESR intensity variability of the Ti centre measurements is similar for the three setups. Hence, for the Ti centre, the CVs range from 2.39 to 4.08% (
The measurement of each complete sample was repeated over three different days with the three setups, in order to consider the reproducibility of the measurements. Then, a mean CV value was calculated by averaging the CV values achieved for the different aliquots of a sample, considering the three days (
For each sample, BIZ1201 and BIZ1202, the final ESR intensities (
Final DRCs derived from the three experimental setups are displayed in
For samples BIZ1201 and BIZ1202, most of the ESR intensities achieved with the three different setups are within error (
In summary, data show that the three different experimental setups provide comparable intensities for both Al and Ti centres, which suggests that any of the setups could be used for quantitative ESR measurements.
For each sample and setup, three DRCs were obtained: each one derived from the measurements performed on a different day (time 1, 2 and 3). A fitting process and a DE evaluation were carried out for each DRC as described in Section 2 – DE evaluation.
Results indicate that the reproducibility of the DE values is similar for the three different experimental setups. Actually, the DE variability over repeated measurements is ranging from 3.7% (BIZ1202) to 10.7% (BIZ1201) with setup #1, from 1.1% (BIZ1201) to 7.5% (BIZ1202) for setup #2, and from 6.2% (BIZ1201) to 6.8% (BIZ1202) for the third one. Although the DE reproducibility achieved with the setups show a large difference from one sample to another, except for setup #3 (both samples around 6%), the limited number of samples used in this study does not allow to draw any definitive conclusion. Moreover, the magnitude of the variability in the DE values is actually in the range of previous studies (typically between 3 and 10%, e.g., Duval
In addition, for each sample and each setup, a final DE value was obtained from the final DRC (Supplementary information
Finally, calculated adjusted r2 values could provide an indication of the goodness-of-fit achieved by the EXP+LIN function fitted through the experimental data points. For all the samples, overall good fittings (adjusted r2 values > 0.98) were obtained when using the three different setups (
Data show that for the three different setups, all the DEs derived from the measurements performed on different days are consistent within 1-sigma error (
Final DRCs obtained by averaging the three repeated ESR measurements (Supplementary information,
Overall, data show no appreciable difference in the calculated DEs for a given paramagnetic centre. Therefore, results suggest that any of the three experimental setups could independently be used for dose estimation of optically bleached quartz grains.
Finally, it is interesting to notice that setup #3 offers a significant reduction of the measurement time for the Ti centre, thanks to the higher S/N ratio achieved with the high sensitivity resonator in comparison with a standard one (see Section 3 – Comparison of the resonator sensitivities). For instance, for the Ti centre measurements of the present study, not only was the acquisition time halved (see
This intra-laboratory comparison provides a further step in the standardization of the ESR method applied to optically bleached grains. It gives some insights about the potential impact of different experimental setups on measurement precision and dose evaluation of quartz grains. The main conclusions and recommendations of this study are:
Firstly, measurements should be performed under carefully controlled experimental conditions in order to ensure repeatable measurements. Among the factors to take into account, one could mention the measurement temperature (
Secondly, our results show the interest of using a HSQE resonator for an improved measurement sensitivity. This resonator provides S/N ratios about 3–4 times higher than a standard resonator. Therefore, it is of special interest for measuring samples showing ESR signals with low intensity, as it could happen with the Ti centre signal in quartz samples from many site localities; or in cases where only a small quantity of sample is available.
On the other hand, the three experimental setups yield comparable results for both Al and Ti centres in terms of intensities, measurement precision and DE estimates. Dose values are all consistent within 1 sigma errors for both Al and Ti centres, suggesting that all the setups could be independently used for dose estimation of optically bleached quartz grains.
As a next step, a future inter-laboratory study would be of special interest in order to evaluate a possible bias on dose estimates due to the use of different analytical procedures. So far, this has never been really checked, although it appears to be essential in order to achieve a minimum standardization of the analytical procedure and data reporting (e.g. Duval