Figure 1 shows the variation of TL intensity against temperature observed during the heating of a sample of calcite. This graph, which is referred to as a glow curve, is a unique product of the TL measurement. As seen in this example of calcite TL, the glow curve is formed from a number of peaks having different heights, widths and shapes. The TL belonging to each peak is generated by a particular structural element within the crystal, by a mechanism which is described below. The term TL signal is used to refer to the luminescence produced by a given structure. The positions and shapes of the TL peaks are related to the characteristics of these structures, which are in turn typical of the crystal containing them.
The colour dimension is not unique to TL, but is shared by all types of luminescence. In all such processes, the emission spectrum conveys a certain amount of information about the mechanism that is generating the light output. In the case of TL, however, the extension of this information into the temperature dimension greatly increases the amount of data which is available for practical application. Thus, the extra dimension in TL opens new possibilities of analysis.
In order to understand the value of the information contained in the TL observation, it is helpful to understand something about the process that produces the light emission. The crystal is rendered capable of emitting TL by exposure to ionising radiation, which redistributes electrical charges within the material. Some of charged particles become confined at trap sites, while others are localised at luminescence centres. During the TL measurement, light is produced in a two stage process, which is illustrated in figure 3. In the first stage, the charges held at trap sites are released by the action of heat and become mobile. In the second stage, these charges are attracted to oppositely charged particles at the luminescence centres, combine with them, and release energy in the form of light. At the end of the measurement, the crystal contains fewer centres with unpaired charges, and thus returns to a situation similar to the one it was in prior to radiation exposure.
The glow curve owes its significance to the fact that the temperature of the crystal defines the amount of thermal energy which is available for initiating the first stage of the TL process. It requires more thermal energy to release charge from a deep trap than from a shallow trap, so that the TL associated with a deeper type of trap appears at a higher temperature. Thus, the temperature scan which constitutes the TL glow curve represents a scan through the various types of trap present in the crystal.
By contrast, the colour of the light emitted is determined, in the second stage of the TL process, by the type of luminescence centre where the charges combine. Since the form of the glow curve is related to the charge traps, and the emission spectrum is determined by the luminescence centres, the TL measurement provides information about both stages of the luminescence process simultaneously.
The uses to which the additional glow curve information of the TL measurement can be put fall under three headings:
Luminescence dating applications treat materials as radiation dosimeters, capable of giving a measurement of the total radiation dose that they have received since some significant event in the past. This accumulated radiation dose is called the palaeodose, and its effect on the material is to induce in it a capacity to luminesce. The present-day measured intensity of the luminescence is referred to as its natural intensity.
Regardless of the type of luminescence employed, the evaluation of the palaeodose is based essentially on the relationship between the observed natural intensity and the rate at which the intensity increases with additional radiation dose. In general, the measured light intensity is the sum of a number of components from different luminescence processes, each of which contributes its own signal. In order to avoid errors in the evaluation of the palaeodose, it is necessary that the signal composition of the natural intensity is similar to that of the intensity increase. This is particularly important in respect of sediments older than 100 ka, because in these samples it is possible for the natural intensity to be dominated by one signal, while the luminescence increase is due mainly to another signal. Since these two signals are the products of independent processes, palaeodoses evaluated under these circumstances can have no meaning.
The use of glow curve information has made an important contribution to understanding the composite nature of sediment TL, and to discovering the properties of the individual signals that contribute to it. The basic technique is one in which the same TL sample is repeatedly measured under different experimental conditions. This method yields a series of glow curves in which the contributing TL signals are observed in varying ratios of intensity. Figure 4 shows a series of natural TL glow curves, labelled "A+B", obtained from one sediment measured with four different combinations of wavelength selection and sample preparation. By extrapolating the variations of these glow curves, they can be decomposed into the shapes of the two TL signals, labelled A and B, from which they are all formed.
From such observations, it has been found that sediment TL is dominated by two fundamental signals whose glow curves and emission spectra overlap. The lower temperature signal, Signal A, has an emission spectrum which extends into ultra-violet wavelengths, while Signal B emits most strongly at a temperature of 360°C and at blue wavelengths. Only Signal A has the necessary properties which make it suitable for dating sediments. These properties include a rapid and almost complete removal of the TL (bleaching) on exposure to light, and a high saturation dose. The routine method for obtaining an adequately pure sample of this TL signal makes use of both temperature and wavelength dimensions of the TL emissions. Firstly, an optical filter which transmits ultra-violet rays while absorbing blue light is placed in front of the TL detector. Secondly, the glow curve is sampled at a low temperature where Signal A is maximised relative to Signal B. This ability to work in two dimensions ensures an effective discrimination between the two TL signals.
The section above dealt with the selection of the TL signal most suited for the dating of sediments. The glow curve of this signal rises sharply on its low temperature side, then peaks and falls away less steeply. The interesting feature of this glow curve is that its shape is dependent on the age of the sediment and on the past ambient temperature of its location. For old sediments from hot regions, the rising edge occurs at a higher temperature compared to the onset of TL from young sediments in cold climates. This is illustrated in figure 5, which compares the shapes of natural TL glow curves obtained from three sediments collected from Arctic, temperate and warm climatic regions, respectively. This effect is a consequence of the slow decay of low temperature TL when it is exposed to ambient levels of thermal energy. Thus, in a sediment exposed to a high ambient temperature for a long period, a greater quantity of its low temperature TL is lost.
In order for the sediment to act as a reliable radiation dosimeter, the measured TL intensity must be stable against the kind of decay described above. Using the glow curve information, it is a simple matter to select the TL which is stable and reject that which is unstable. Over the long burial period of the sediment, the low temperature TL decays while the high temperature TL survives to produce the natural glow curve. By contrast, when the sediment is exposed to ionising radiation in the laboratory and its TL measured soon after, there is insufficient time between irradiation and TL measurement for signal loss to occur. By comparing the shape of the natural glow curve with that of the artificially regenerated glow curve, it is easy to identify the temperature below which natural TL has decayed, and above which it is stable. This is illustrated in the following section, which explains in greater detail how the comparison of natural and regenerated glow curves is used in TL dating.
When sedimentary deposits are measured by luminescence techniques, the event that is dated is the exposure of the sediment grains to daylight at the time of deposition. This exposure to daylight greatly reduces the luminescence intensity that can potentially be yielded by the grains, but does not entirely remove it. The intensity of the remainig, or residual, luminescence is the starting level from which the signal regenerates under the action of natural radiation in the burial environment. The residual level must therefore be subtracted from the observed natural intensity in order to determine the amount of luminescence that has been induced by the palaeodose.
In the regeneration method of palaeodose evaluation, the processes that the sediment has undergone during and since its deposition are simulated in the laboratory. The changing intensity of the sediment's luminescence is observed at various stages of the simulation and compared with the intensity of its natural luminescence. The laboratory simulation begins with an exposure to light which is intended to replicate the effect of the original daylight exposure at the time of deposition. Various artificial radiation doses are then administered to regenerate luminescence as it would have appeared in the sediment at different times during its burial. The palaeodose is assumed to be equal to the amount of artificial radiation which regenerates the same intensity of luminescence as found in the natural material.
The main uncertainty in this method is the quantity of light which needs to be given to the sediment in order to simulate the original exposure at deposition. If this quantity is too great, the simulated residual luminescence will be less than the actual residual at the time of deposition, and a higher artificial radiation dose will be required to regenerate the natural intensity. Conversely, if the laboratory light exposure is too low, the measured palaeodose will also be too low. This uncertainty in the palaeodose evaluation is minimised by selecting a luminescence signal which is rapidly bleached to a low residual intensity. However, when young sediments are dated, the natural luminescence inevitably contains a relatively high proportion of residual signal. In these cases, it is desirable to use a mehod which can detect the initial condition of the sediment's luminescence immediately after deposition.
As emphasised above, the TL glow curve from sediments is composed almost entirely from two overlapping signals whose properties are widely different. Signal A is selected as the dating signal because it is easily bleached. The higher temperature signal, Signal B, is comparatively difficult to bleach and its residual intensity is much greater as a fraction of its natural TL intensity. The fact that the TL glow curve is formed from two signals possessing such different bleaching properties is very fortunate, because it provides a means of detecting the original intensity of residual TL.
Essentially, the method involves the detailed comparison of the shapes of the natural and regenerated TL glow curves over a wide temperature range. If the regeneration method is repeated using varying degrees of laboratory light exposure, it is observed that an exact matching of the glow curve shapes occurs only for one value of the exposure. It seems plausible that this unique artificial bleach treatment corresponds to the actual bleaching at deposition.
The condition of shape matching of natural and regenerated TL glow curves is illustrated in figure 6. The continuous lines show the variation of the measured palaeodose against temperature for a sample of sediment in two separate experiments. These experiments differed only in the degree of the laboratory bleaching exposure. The behaviour of the two curves encapsulates a large amount of information. The steep rise of the curves from zero shows the effect of the thermal instability of the low temperature TL, which was explained in the previous section. The temperature at which the curves level out, shown by the vertical line, is the temperature above which the natural TL has remained stable.
The lower curve, which falls away above this temerature, shows that the laboratory bleaching used in this measurement is probably less than the light exposure to the sediment at the time of its deposition. Likewise, the rising upper curve indicates a probable over-bleaching in the laboratory. The divergence of the two curves at higher temperatures is a consequence of the increasing proportion of Signal B in the glow curves, and the large intensity of residual TL associated with this signal. The discontinuous line in figure 6 shows the palaeodose values that would have been obtained if the laboratory bleaching had been set to the critical amount of light exposure where the natural and regenerated glow curve shapes exactly match each other. The vertical line gives the optimum temperature for TL selection, being the temperature where, firstly, the natural TL is stable, and secondly, the palaeodose changes least as a function of laboratory bleaching. This last condition is directly related to the fact that the proportion of the easily bleached TL of Signal A is greatest at low temperatures.
The glow curve is unique to the TL measurement procedure, and provides the only means of access to information in the temperature dimension. When combined with the emision spectrum it creates a data set of great value. The applications to which this information may be put are especially relevant to the requirements of sediment dating.
Glow curve information may be used to detect and separate different TL processes which contribute to the total light intensity, and allows the properties of the component signals to be determined. This knowledge, in turn, informs the sampling strategy for optimising the TL signal most suited for dating purposes. Additionally, the glow curve provides the means, firstly, to distinguish between thermally stable and unstable luminescences, and secondly, to detect the likely condition of sedimentary TL immediately after deposition.