The principles of Luminescence Dating
Luminescence dating is used to identify when a sample was last exposed to daylight or extreme heat by estimating the amount of ionising radiation absorbed since burial or firing. Free electrons, excited mainly by environmental alpha, beta and gamma radiation, become trapped within the crystalline defects of minerals such as quartz and feldspar, and continue to accumulate until exposure to daylight or sufficient heat evicts them, bleaching the sample of its signal, and effectively resetting the ‘luminescence clock’ to zero. Once evicted from the ‘traps’ the electrons are attracted are attracted back to the ‘holes’ created by ionisation and, on recombination at these centres emit energy in the form of photons. These are measured by a photomultiplier tube within a luminescence reader and are used to estimate the Equivalent Dose (De) of absorbed radiation required to create that signal, and is measured in Grays (1 Gy = 1J/kg). The measurement of radionuclides present within the surrounding sediment can be calculated either by in-situ gamma spectrometry or high-resolution gamma spectrometry in the laboratory to identify the annual Dose Rate (Dr) of ionising radiation, Gy kyr-1, after which the age of the sample can be calculated using the equation:
Estimated age = De/Dr
This equation very simply expresses the calculations necessary, but it is important to be aware of the factors influencing the two values used. Heterogeneous sediments and radioactive disequilibria will increase errors on Dr, while incomplete bleaching of the sample prior to burial, anomalous fading in feldspars, and the estimation of past sediment moisture content may all also add to increased errors. Despite this, through the careful choice and collection of samples, as well as stringent preparation and analytical methods, it is possible to produce ages with an accuracy of between 5 and 12%.
The dating of sediments using the luminescence signal generated by optical stimulation (OSL) offers an independent dating tool, and is used most often on the commonly occurring minerals of quartz and feldspar and, as such, has proved particularly useful in situations devoid of the organic component used in radiocarbon dating. The ‘depth’ of the traps within which the electrons are held effectively dictates the point at which a sample will become saturated with charge and so the limit of the dating range. Quartz has been used for dating to at least 200 ka, while the deeper traps of feldspar have produced dates as old as 1 ma. The use of fine-grain dating for samples such as pottery, loess, burnt flint and lacustrine sediments, and coarse-grain dating of aeolian, fluvial and glacial sediments is regularly undertaken. While thermoluminescence (TL, the generation of a luminescence signal generated by thermal stimulation) is still conducted on pottery and burnt flint samples, the bulk of luminescence dating now uses optical stimulation as this releases a signal that is far more readily zeroed than that re-set by heat.
Analysis of fully bleached samples is preferred as this ensures that associated errors are kept to a minimum. Despite this, procedures exist with which to identify and take account of partially bleached grains, as may be seen in fluvial, or more likely glacial sediments, where light exposure may have been attenuated by turbid or turbulent conditions.
Collection of Samples
It is important to observe certain conventions when collecting samples in order to reduce errors as much as possible. By taking samples from well-sorted sediment structures problems with heterogeneous dose rates may be avoided, and all grains are more likely to have undergone the same depositional history. Any areas of disturbance such as soil formation, groundwater leaching, bioturbation or slumping, should be avoided to remove the potential for post-depositional mixing of grains. As the exclusion of light exposure is vital, opaque steel or plastic tubes are driven into sediment taking care to discard any material that may have undergone exposure. All subsequent preparation is conducted under subdued orange light to avoid the early stimulation of any trapped charge. It is recommended that the assistance of a luminescence specialist is sought before samples are taken as mistakes at this stage may subsequentlz be responsible for incorrect age estimates later.
Luminesence measurement procedures
LEDs are now widely used for optical stimulation, with the photons emitted being measured by a photomultiplier tube (PMT) attached to a luminescence reader (see Figure 3). By placing optical filters in front of the PMT it is possible create detection windows to isolate particular bandwidths of emissions, and also to avoid the influence of backscattering of the stimulation source. The use of longer wavelength stimulation light to that of the emissions helps to ensure reliable evaluations and blue or green light is now most widely used for the stimulation of quartz. As feldspar can be measured using the longer wavelengths of infrared (IR) stimulation, a broader range of emissions are available for measurement (see Figure 4).
Measurement of De is now generally made using the Single Aliquot Regenerative (SAR) protocol (Murray and Wintle, 2000). Firstly, the natural signal (Ln) present in the sample is measured, followed by a series of rising known radiation doses from which a dose response curve can be constructed (Figure 4). These subsequent doses are designed to bracket the natural signal and so make it possible to interpolate from the curve the Equivalent Dose needed to generate Ln. It is recognised that the luminescence generated following laboratory irradiation contains a proportion of electrons that are considered thermally unstable over geological time and are not present when measuring the natural signal in a sample. To remove this element a preheat is administered, which in turn may increase the sensitivity of the sample to further irradiation. The introduction of a standard test dose (Tn) following the measurement of the natural and laboratory doses allows the calculation of a ratio with which to monitor, and correct for, any change in sensitivity. Further parameters are tested during and before this sequence to confirm that the protocol can successfully recover a signal.
The SAR protocol offers significant advantages over the earlier multiple aliquot approach which only produced one normalised natural value (Ln) from the use of 24-48 aliquots per sample, and was then used to extrapolate a De value from the dose response curve. The SAR approach provides multiple Des and also allows for the explicit testing of partial bleaching.
Calculation of the Dose Rate
The calculation of the dose rate introduces uncertainties due to substantial variability being observed in potassium, uranium and thorium which are the principal sources of environmental radiation. Although minor below altitudes of 1 km, an additional contribution from cosmic rays is also considered, and can be limited to the soft component fraction by only using sediments below depths of more than 0.5 m (Aitken, 1998). Increased moisture content and carbonate coating of grains can lead to an attenuation of the radiation reaching the sample, and so ultimately to an age overestimation if not identified. Chemical and radiometric assessment of the surrounding sediment, through in-situ gamma spectrometry or laboratory high-resolution gamma spectrometry, identifies the proportions of each element together with the contribution from the alpha, beta and gamma components. With alkali-feldspars, their internal potassium content contributes a significant proportion of the total radiation it experiences, and so considerably reduces the reliance on the external dose rate, variations in moisture content and overburden and so the associated uncertainties (Duller, 1997; Preusser, 2003). While the present-day moisture content can be measured it is that which has prevailed over the majority of the burial period that is pertinent, a maximum for which can be calculated by identifying the porosity of the sediment, with a 1% error in moisture content leading to an approximately equivalent error in age (Wallinga, 2002).
Aitken, M J 1998 An introduction to optical dating. Oxford University Press
Duller, G A T 1997 Behavioural studies of stimulated luminescence from feldspars, Radiation Measurements, 27(5/6), 663-694
Murray, A S and Wintle, A G 2000 Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol, Radiation Measurements, 32, 57-73
Preusser, F 2003 IRSL dating of K-rich feldspars using the SAR protocol: comparison with independent age control, Ancient TL, 21(1), 17-22
Wallinga, J 2002 Optically stimulated luminescence dating of fluvial deposits: a review, Boreas, 31, 303-322