Elsevier

Quaternary Geochronology

Volume 39, April 2017, Pages 105-111
Quaternary Geochronology

Research paper
Optimization of laboratory illumination in optical dating

https://doi.org/10.1016/j.quageo.2017.02.010Get rights and content

Abstract

As part of the development of new laboratory lighting, we present a methodological approach applicable to the characterization of any light source intended for illumination in optical dating laboratories. We derive optical absorption cross-sections for quartz and feldspar from published data and compare these with the human eye response. The optimum peak wavelength giving the best clarity for non-dark adapted vision for the least trapped charge loss lies within the wavelength range 590–630 nm; we argue that it is unnecessary to consider dark-adapted vision. The predicted relative decay rates of quartz optically stimulated luminescence (OSL) and feldspar infrared stimulated luminescence (IRSL) signals by an incandescent light bulb and a compact fluorescent lamp (CFL) through an ILFORD 902 filter are first derived. These predicted decay rates are then compared with those of three relevant light-emitting diodes (LEDs); this comparison demonstrates the significant advantage of the LED sources over the filtered light sources, because essentially all of the reduction of both OSL and IRSL signals by the LEDs occurs at wavelengths to which the human eye is most sensitive. We conclude that exposure of quartz and feldspar extracts from various samples to the light from an LED with emission peak at 594 nm results in a ≤1% OSL or IRSL signal loss for a 48-h exposure at a power density of ∼0.2 μW.cm−2.

Introduction

A prerequisite for the reliable measurement of a light-sensitive trapped charge population is that the original population remains unaffected during any treatment prior to measurement (see section 3.2.2 in Aitken, 1998). Ambient lighting in optical dating laboratories is a significant source of concern, and should be kept as low as possible. On the other hand, laboratory pretreatment of samples before measurement often involves significant personal hazard (e.g. hydrofluoric acid) and working several hours per day under dim light condition; there is thus an obligation to demonstrate that the intensity of ambient lighting is not unnecessarily low.

Optimization of ambient lighting in luminescence dating laboratories has been the subject of several studies since the early days of thermoluminescence (TL) dating. Almost all of these are published in Ancient TL. The first studies focused on blocking wavelengths <550 nm by conversion of white fluorescent tubes to safelights using UV-green absorbing filters (Sutton and Zimmerman, 1978, Jensen and Barbetti, 1979, Spooner and Prescott, 1986, Spooner et al., 1988). Filtered white fluorescent tubes were chosen over red bulbs and red fluorescent tubes because they provided better clarity of vision (Sutton and Zimmerman, 1978). However, the batch-to-batch variation in the transmittance of colour filters as pointed out by Smith (1988) and the general convenience of lighting without any additional filtering prompted further investigation into unfiltered coloured light sources (Galloway and Napier, 1991, Galloway, 1991) even after the advent of optically stimulated luminescence (OSL) dating of quartz which makes use of traps that are more sensitive to light than those used in TL dating (Huntley et al., 1985).

With the introduction of infrared stimulated luminescence (IRSL) dating of feldspars (Hütt et al., 1988) and the recognition of an IR absorption resonance at ∼850 nm as, at least partially, responsible for age underestimation (Aitken, 1994), later studies aimed at reducing the IR component of laboratory lights as well as blocking wavelengths <550 nm. For instance, Lamothe (1995) used a combination of red filters with a bandpass glass filter to limit the emission spectrum of white fluorescent tubes to 600–650 nm, while Spooner et al. (2000) developed a light module consisted of a low-pressure sodium vapor lamp emitting predominantly at 589 nm screened with five layers of yellow plastic filters to reject the high energy photons; this light module was later adopted by others (e.g. Mauz et al., 2002). Spooner et al. (2000) rightly argued that yellow-orange wavelengths are not considerably more effective at luminescence signal reduction (bleaching) than orange-red wavelengths; since the operator's eye is more sensitive to yellow-orange, these wavelengths can be used at a lower intensity. This argument was also used by Huntley and Baril (2002) who obtained their IR-suppressed yellow-orange laboratory light by placing orange filters in front of compact fluorescent light (CFL) bulbs (Huntley and Baril, 2002).

Given the narrow emission band and ready power control of light emitting diodes (LEDs) compared to incandescent light bulbs and CFLs, Berger and Kratt (2008) tested the suitability of a green and a red LED for laboratory illumination. They deduced that exposure of quartz to either of these LEDs at an intensity of ∼0.7 μW.cm−2 will have no significant effect on the OSL signal for up to few hours, while the safe exposure time for feldspar IRSL signal was found to be 30–40 min. and 15–20 min. for the green and red LED, respectively. Although such exposure times are rather short for most routine laboratory procedures, no guidance on a practical combination of light intensity and exposure time was provided (Berger and Kratt, 2008). Moreover, these green and red wavelengths are distant from the yellow-orange light discussed by previous workers (e.g. Spooner et al., 2000, Huntley and Baril, 2002). Consequently, despite their well-known unreliable characteristics (Smith, 1988), colour filters have remained popular components of laboratory safelights in optical dating; our own laboratory has, for example, consistently used ILFORD 902 glass filters (developed for photographic processing), but these are no longer available.

As existing laboratories undergo renovation and new laboratories are established, there is a constant demand for new laboratory light sources; this demand is also stimulated by laboratory intercomparisons and the dating of standard samples (Murray et al., 2015). Unfortunately, there are no standard methods for determining ambient light conditions, and as a consequence there is, in our view, a tendency to adopt the principle of ‘the darker and redder the better’. However, such an approach can lead to unsafe laboratory practices, and so the development of a methodological approach to the optimization of ambient illumination is of crucial importance.

The choice of the optimum light source for illumination in an optical dating laboratory hinges on the answers to two questions: first of all, what is the most appropriate wavelength for exposure, if the optical absorption cross-sections of quartz and feldspar, and the light response of the human eye are taken into account? Secondly, given some minimum exposure time required for sample preparation in the laboratory, what is the maximum power density to which grains of quartz and feldspar can be exposed at this wavelength before any loss of the trapped charge of interest becomes unacceptable? To address these questions, we first derive the quartz OSL and feldspar IRSL optical absorption cross-sections from published data and divide these into the known response of the human eye to determine the wavelength region providing the best clarity of vision for the least trapped charge loss – not surprisingly this lies within the yellow-orange region. The emission spectra of the widely-used ILFORD 902 safelights and three relevant commercially-available LEDs are measured to determine the relative effect of the peak emission compared to the short and long wavelength tails on the reduction of the quartz and feldspar signals. The most suitable light source is then tested extensively on samples of various geological origins to determine the maximum acceptable power density for a given exposure time.

Section snippets

Instrumentation and measurement conditions

Spectral measurements were carried out using an Ocean Optics MAYA2000-Pro spectrometer with a spectral range of 200–1100 nm. All spectra were measured over a total integration time of 150 s and corrected for detector non-linearity, electrical dark current and stray light. The power densities were measured with a Thorlabs PM200 handheld optical power meter console equipped with a Thorlabs S120VC photodiode power sensor with an aperture diameter of 9.5 mm. The luminescence measurements were

Methodology

Finding an ideal laboratory light source that provides reasonable illumination but no depletion of trapped charge populations is challenging. Quartz OSL and feldspar IRSL signals are derived from trapped charges that can be excited by photons over a wide range of wavelengths, from the visible to the near infrared (e.g. Hütt et al., 1988, Bailiff and Poolton, 1991, Ditlefsen and Huntley, 1994, Bøtter-Jensen et al., 1994, Spooner, 1994a, Spooner, 1994b). Furthermore, although the human eye is

Results

We now investigate the relative bleaching rates of different light sources on quartz OSL and feldspar IRSL signals. This is carried out by multiplying the measured emission spectra of each light source by the photoionization cross-sections of the two signals (see Fig. 1) to derive their relative bleaching rates over a wide wavelength range (300–1100 nm). The resulting bleaching spectra are then compared with the desired peaks identified above (see Fig. 2) to evaluate the importance to bleaching

Bleaching tests

Having identified the most appropriate wavelength and light source, the next step is to determine the bleaching response of quartz OSL and feldspar IRSL signals to this wavelength. This will allow the quantification of any of the three parameters of signal loss, power density and exposure time, by setting the desired values of the other two. To this end, 48 aliquots each of Risø calibration quartz (Hansen et al., 2015) and K-rich feldspar from a loess sample (sample D38139) were prepared. Seven

Discussion

The overall higher bleaching rate of feldspar IRSL than quartz OSL at 594 nm makes feldspar the determining factor for setting the upper limit to power density and the exposure duration in the laboratory; this applies to almost all the samples studied here, with the exception of sample 981009, for which the two signals seem to be reduced by the same amount (∼30%) after 24 h of exposure. The higher than expected apparent bleaching rate of this OSL signal may arise from feldspar contamination

Conclusion

A methodological approach to the characterization of light sources suitable for illumination in optical dating laboratories has been presented. Based on published data, we argued that the best compromise between minimum bleaching and maximum visibility is achieved at a wavelength range from ∼590 to 630 nm. Comparison of the predicted relative bleaching rates of quartz OSL and feldspar IRSL signals by an incandescent light bulb and a CFL through an ILFORD 902 filter with those of three different

Acknowledgments

Louise Maria Helsted, Lars Peter Pirtzel and Myungho Kook are gratefully acknowledged for their assistance with sample preparation and technical support with the electronics. J.-P.B. receives funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme ERC-2014-StG 639904 – RELOS.

References (40)

Cited by (24)

  • Constraining the age of Quaternary megafloods in the Altai Mountains (Russia) using luminescence

    2022, Quaternary Geochronology
    Citation Excerpt :

    A total of 24 samples for luminescence dating were collected from the three pits. Sample preparation took place in Moscow under dim orange light (Sohbati et al., 2017). Samples were processed using standard procedures (Murray et al., 2021): wet sieving to obtain the 180–250 μm sand fraction, followed by treatment with 10% HCl, H2O2, HF, and HCl.

View all citing articles on Scopus
View full text