Dry Fallout

Ice melting and Erosion Lacustrine Sediments

Earth's Crust

FIGURE 3.21 Schematic diagram of the 2l0Pb global cycle, illustrating the source, transport, deposition, and post-depositional, redistribution processes (from Preiss et al., 1996).

amount of ionizing radiation that the sample has been exposed to over time, from surrounding sediments. The decay of radioisotopes in the surrounding matrix produces free electrons in the mineral grains, which become trapped at defects in the crystal lattice; the longer the mineral has been exposed to radiation, the higher will be its trapped electron population and the greater the resulting luminescence signal. Luminescence is thus a measure of the accumulated dose of ionizing radiation (expressed in units of Grays, Gy) which is a function of sample exposure age. Age is determined by exposing subsets of the sample to known doses of radiation and measuring the resulting luminescence signal. The amount of radiation needed to produce the same luminescence signal as that from the original sample is called the equivalent dose (ED) (sometimes known as the paleodose; Duller, 1996). If the amount of radiation a sample is exposed to each year (the dose rate) is known (by direct measurement of the radioactive properties of the surrounding matrix) the sample age can be calculated as:

The key constraint is that the sample to be dated must either have no residual luminescence signal from the period before the event being dated, or any residual signal must be quantifiable. Because luminescence is released by either heating or optical bleaching of the sample, the event to be dated must have involved one of these processes that effectively set the "luminescence clock" to zero. Thus, TL dating has been widely used in archeology to date pottery or baked clay samples as well as baked flints from fire hearths attributable to early man (Wintle and Aitken, 1977). And TL has also been used to date sediment baked by contact with molten lava (Huxtable et al., 1978) and inclusions within lava (Gillot et al., 1979). By controlled heating experiments in the laboratory, measurements of TL can indicate the amount of time that has elapsed since the sample was last heated. The same principle applies to wind-blown and fluvially transported sediments. Exposure to sunlight can also dislodge electrons, resetting the luminescence clock. As shown in Fig. 3.22, the TL signal is reduced to an unreachable residual signal, but the OSL signal is reduced to zero on exposure to sunlight. The TL or OSL signal at any time after deposition and burial will thus be a measure of the time that has elapsed since the grains were transported to their depositional site (Huntley et al., 1985; Duller, 1996). Luminescence dating is thus extremely useful in studies of mainly inorganic sediments, such as loess and other aeolian deposits (Wintle, 1993; Zoller et al., 1994) as well as water-lain sediments (Balescu and Lamothe, 1994). However, if zeroing of the sample luminescence is incomplete prior to burial, the age of the deposit will be overestimated. This is a common problem in fluvial or glacio-fluvial sediments where exposure to sunlight may be limited (in duration and/or wavelength) by turbid water (Forman et al., 1994).

The useful timescale for TL dating depends on the radiation dose to which the sample has been exposed, and the capacity of the sample to continue to accumulate electrons, before becoming saturated. Samples with high quantities of potassium feldspars, for example, are potentially useful for dating older deposits, because such

Equivalent dose (ED) Dose rate

-level of "geological" TL signal signal reduced by light exposure level of residual "unbleachable" TL

time

(assuming constant flux of ionizing radiation)

level of natural TL signal signal readout in laboratory by heating to 450°C

-level of "geological" OSL signal

TL signal signal zeroed by light exposu

FIGURE 3.22 Schematic illustration of changes in thermoluminescence (TL) and optically stimulated luminescence (OSL) signals with time (from Rendell, 1995).

time level of natural OSL signal signal readout in laboratory by exposure to green or infrared light

(assuming constant flux of ionizing radiation)

FIGURE 3.22 Schematic illustration of changes in thermoluminescence (TL) and optically stimulated luminescence (OSL) signals with time (from Rendell, 1995).

minerals are less easily saturated. Most analysts are reluctant to place much faith in TL age estimates of more than 200,000 yr, but ages of up to 800 ka B.P. have been reported from areas where dose rates are extremely low and saturation levels are potentially high; nevertheless, such dates are controversial (Berger et al., 1992). Accuracy of luminescence dates may approach ±10% of the sample age, though comparison with 40Ar/39Ar dates suggests that TL dates on older samples are commonly 5-15% too young, perhaps due to saturation of the electron traps (see Section 3.2.4.2).

3.2.4.1 Thermoluminescence (TL) Dating

The thermoluminescence of a sample is a function of age. The older the sample, the greater will be the TL intensity. This is assessed by means of a glow curve, a plot of TL intensity vs temperature as the sample is heated (Fig. 3.23a). The TL emission at lower temperatures is not a reliable age indicator; such emissions correspond to shallow traps in the sample where electrons are not stable. The precise temperature necessary to dislodge the deeper "stable" electrons will depend on individual sample characteristics and is assessed by finding the point at which the ratio of natural TL

TL glow curves

TL growth curves

-natural

laboratory radiation — residual black body

TL glow curves

-natural

laboratory radiation — residual black body

200 300 temperature (°C)

Additiye dose growth curve at 300°C

200 300 temperature (°C)

TL growth curves

Additiye dose growth curve at 300°C

laboratory radiation dose

OSL decay curves OSL growth curves

OSL decay curves OSL growth curves

FIGURE 3.23 Schematic representation of Equivalent Dose (ED) determination by the additive dose method forTL and OSL. The OSL is measured after the removal of an unstable component of the signal by preheating or long-term storage of the samples (from Rendell, 1995).

to artificially induced TL becomes approximately constant. Generally this is in the range 300-450 °C, so TL intensity at these temperatures is used for age assessment.

To determine the age of a sample it is necessary first to know how much TL results from a given radiation dose, as not all materials produce the same amount of TL from a given radiation dose. In one approach (the additive dose method) subsets of the sample are exposed to a known quantity of radiation. Then TL is measured in each sample; the initial paleodose can then be determined by extrapolation to a laboratory-determined residual level (Fig. 3.23b). Another approach (the regenerative method) involves first bleaching subsamples of all their TL, then exposing them to different levels of radiation, followed by measurement of the TL emitted, corresponding to different levels of radiation exposure (Fig. 3.24). The TL in the original sample is then measured and compared with the artificially radiated samples to assess the corresponding paleodose (ED) as accurately as possible. Sample age can then be calculated simply by dividing the paleodose by the measured dose rate at the sample site. The dose rate is assessed by measuring the quantity of radioactive uranium, thorium, and potassium in the sample itself, and in the surrounding matrix. Alternatively, radiation-sensitive phosphors (such as calcium fluoride) may be buried at the sample site for a year or more to measure directly the environmental radiation dosage. A number of procedures have been devised to estimate the long-term dose rate and reduce the inherent uncertainty (Aitken, 1985).

3.2.4.2 Problems of Thermoluminescence Dating

In the age equation given in the preceding section, it is assumed that there is a linear relationship between radiation dose and the resulting TL. It is known, however, that this is not always the case at extremely low radiation dose levels, or at extremely high dose levels. The former problem (supralinearity) is most significant for relatively young samples (<5000 yr old) as the rate at which a sample acquires TL is reduced at relatively low radiation dose levels (or perhaps is nonexistent until a certain radiation threshold is exceeded). At the other end of the scale, very long exposures to radiation (high doses) may result in saturation of the available electron traps so that further exposure will not appreciably increase the sample TL. When this is likely to occur depends on sample age and mineral composition, but, in general, very old ages indicated by TL dating (greater than several hundred thousand years) are likely to be only minimum estimates.

A further difficulty in assessing the relationship between TL and radiation dose occurs when irradiated samples "lose" TL after very short periods of time, perhaps only a few weeks. This phenomenon is called anomalous fading (Wintle, 1973) and is common among certain minerals, particularly feldspars of volcanic origin. Unless corrected for, anomalous fading will result in underestimation of a sample age, but it can be identified relatively easily by storing irradiated samples in the dark and re-measuring TL periodically over a period of several months.

Perhaps the most significant problems in TL dating stem from variations in the environmental dose rate. Of particular importance is the mean water content of the sample and surrounding matrix during sample emplacement. Water greatly attenuates radiation, so a saturated sample will receive considerably less radiation in a given time period than a similar sample in a dry site; as a result the TL intensity will be much lower, giving an incorrect age indication. If the water content of the site can be assessed, this problem can be taken into account in the age calculation. Better still, if a radiation-sensitive phosphor can be placed in the environmental setting of the sample for a period of time, the effect of groundwater on the radiation dose may be assessed directly. However, there is always the problem of knowing how groundwater content has varied in the past and this uncertainty is a major barrier to more accurate TL dating. Groundwater may also leach away radioactive decay products, so long-term changes in groundwater content may further complicate the TL-dose relationship. Finally, it should be noted that one of the decay products of uranium is an inert gas (radon-222) which has a half-life of 3.8 days, long enough for it to escape from the sample site and effectively terminate the decay series (96% of the U-series y-radiation energy is post-radon). Fortunately, laboratory studies have shown that relatively few soils exhibit significant radon loss either in the laboratory or in situ.

3.2.4.3 Optical and Infrared Stimulated Luminescence (OSL and IRSL) Dating

Although the TL method has been widely used for dating of sediments since its first application to deep sea sediments (Wintle and Huntley, 1979) it was always known that it would be more appropriate to use light as the stimulation mechanism rather than heat. This would allow measurement of the light-sensitive luminescence signal, as distinct from the light-insensitive signal that remains in the sample at deposition (Fig. 3.22). With TL dating, a relatively large light-insensitive TL signal found in modern samples prevents the dating of sediments less than 1000-2000 yr old. This problem was addressed by Huntley et al. (1985), who reported on the first samples dated by measurement of an optically stimulated luminescence (OSL) signal. The OSL signal was shown to be zero for modern sediments, thus opening up the possibility of dating sediments as young as a few decades in age. Subsequently, Godfrey-Smith et al. (1988) measured the OSL (stimulated with green light from a laser) from a number of quartz and feldspar samples extracted from sediments. They showed that not only was there a negligible residual OSL signal after a prolonged sunlight bleach, but the initial rate of luminescence signal loss was several orders of magnitude faster than that for the TL signal from the same samples.

The OSL signal is derived from electrons displaced from electron traps in the crystal by phptons, such as those from a 514-nm argon laser (Fig. 3.25b). An electron thus released is able to recombine at a luminescence center, a process that results in the emission of a photon with a wavelength characteristic of the center. The ED is obtained by extrapolation as for TL (Fig. 3.24b). It is thus necessary to be able to observe this luminescence while totally rejecting the light from the stimulating light source. Fortunately, optical filters can be found which pass light from the violet and near ultraviolet emission, characteristic of quartz and feldspars, but reject the green laser light (Fig. 3.25b). Hiitt et al. (1988) discovered that they were able to stimulate trapped electrons from feldspars (but not quartz) using near infrared radiation. This resulted in infrared stimulated luminescence (IRSL) signals

Laboratory Irradiation (Gy)

FIGURE 3.24 Examples of (a) the additive dose and (b) the regenerative methods of estimating paleo-dose in thermoluminescence dating. In the additive dose method, the samples are exposed to increasing levels of radiation, and the associated TL is measured. Extrapolation back to zero reveals the paleodose of the sample before it was exposed to further radiation exposure. In the regenerative method, subsamples are first bleached then exposed to radiation and theTL is measured for each exposure level.The radiation exposure corresponding toTL measured in the original sample is then readily obtained (Duller, 1996).

Laboratory Irradiation (Gy)

FIGURE 3.24 Examples of (a) the additive dose and (b) the regenerative methods of estimating paleo-dose in thermoluminescence dating. In the additive dose method, the samples are exposed to increasing levels of radiation, and the associated TL is measured. Extrapolation back to zero reveals the paleodose of the sample before it was exposed to further radiation exposure. In the regenerative method, subsamples are first bleached then exposed to radiation and theTL is measured for each exposure level.The radiation exposure corresponding toTL measured in the original sample is then readily obtained (Duller, 1996).

that could be observed in a wider wavelength range (blue and green) because an optical filter could be chosen to reject the stimulation wavelength region, around 850-880 nm (Fig. 3.25c).

The principal advantage of OSL and IRSL is that the laser-stimulated electron traps are very sensitive to light, so that a relatively brief exposure to sunlight (on the order of a few tens of seconds to a few minutes) is likely to have reduced the sample luminescence to near zero. Such brief exposures would certainly not reduce

(a) Thermoluminescence (TL) quartz and feldspar

Glow curve

Glow curve

0 100 200 300 400 500 Temperature (°C)

0 100 200 300 400 500 Temperature (°C)

(b) Optically stimulated luminescence (OSL) quartz and feldspar

OSL counts

Laser exposure

Laser exposure

OSL counts

green

514 nm

A

laser line ---1---1-1

Wavelenght (nm)

Wavelenght (nm)

(c) Infrared stimulated luminescence (IRSL) feldspars

FIGURE 3.25 (a-c) Schematic representation of the principles of luminescence dating techniques (from Wintle etal., 1993).

FIGURE 3.25 (a-c) Schematic representation of the principles of luminescence dating techniques (from Wintle etal., 1993).

to zero the luminescence signal measured in TL studies; thus, OSL and IRSL open up the prospect of dating a wider variety of material, particularly quite young aeolian sands and silts (loess) as well as fluvial and lacustrine sediments where exposure to sunlight may have been brief (Wintle, 1993). For example, recent studies have dated aeolian sands deposited only a few decades to a few centuries ago, using IRSL techniques (Wintle et al., 1994; Clarke et al., 1996).

In TL studies, all of the luminescence in a sample is reduced to zero by heating, so repeat measurements are not possible. However, in OSL and IRSL dating, very brief exposures to the stimulation light source yield measurable signals that do not significantly deplete the potential luminescence in the sample, so multiple measurements can be made on the same sample (Duller, 1995). Indeed, multiple analyses on individual grains are now routinely possible (for grains with a high sensitivity), providing a set of results that helps to confirm whether the sample as a whole was adequately zeroed, thereby enhancing confidence that the event in question has been accurately dated (Lamothe et al., 1994). However, it must be borne in mind that many of the problems facing TL dating described in the foregoing (saturation of traps, anomalous fading, and especially estimation of the dose rate) still apply to OSL and IRSL dating.

3.2.5 Fission-track Dating

As already discussed in Section 3.2.3, uranium isotopes decay slowly through a complex decay series, ultimately resulting in stable atoms of lead. In addition to this slow decay, via the emission of a and (3 particles, uranium atoms also undergo spontaneous fission, in which the nucleus splits into two fragments. The amount of energy released in this process is large, causing the two nuclear fragments to be ejected into the surrounding material. The resulting damage paths are called fission tracks, generally 10-20 (xm in length. The number of fission tracks is simply a function of the uranium content of the sample and time (Naeser and Naeser, 1988). Rates of spontaneous fission are very low (for 238U, 10~16 a1) but if there is enough uranium in a rock sample a statistically significant number of tracks may occur over periods useful for paleoclimatic research (Fleischer, 1975).

The value of fission-track counting as a dating technique stems from the fact that certain crystalline or glassy materials may lose their fission-track records when heated, through the process of annealing. Thus, igneous rocks and adjacent metamorphosed sediments contain fission tracks produced since the rock last cooled down. Similarly, archeological sites may yield rocks that were heated in a fire hearth, thereby annealing the samples and resetting the "geological" record of fission tracks to zero. In this respect, the environmental requirements of the sample are similar to those necessary for 40K/40Ar dating. Because different minerals anneal at different temperatures, careful selection is necessary; minerals with a low annealing temperature threshold, such as apatite, will be the most sensitive indicator of past thermal effects (Faul and Wagner, 1971).

Fission tracks can be counted under an optical microscope after polishing the sample and etching the surface with a suitable solvent; the damaged areas are preferentially attacked by solvents, revealing the fission tracks quite clearly (Fleischer and Hart, 1972). After these have been counted, the sample is heated to remove the "fossil" fission tracks and then irradiated by a slow neutron beam, which produces a new set of fission tracks as a result of the fission of 235U. The number of induced fission tracks is proportional to the uranium content and this enables the 238U content of the sample to be calculated. Sample age is then obtained from a knowledge of the spontaneous fission rate of 238U. For a much more detailed discussion of the technique, and the problems of calibrating fission-track dates, see Hurford and Green (1982).

Fission-track dating may be undertaken on a wide variety of minerals in different rock types, though it has been most commonly carried out on apatite, micas, sphene and zircons in volcanic ashes, basalts, granites, tuffs, and carbonatites. It has also been widely used in dating amorphous (glassy) materials such as obsidian and is therefore useful in tephro-chronological studies (see Section 4.2.3; Westgate and Naeser, 1995). Its useful age range is large, from 103 to 10s years, but error margins are very difficult to assess and are rarely given. Microvariations in crystal uranium content may lead to large variations in the fission track count on different sections of the same sample (Fleming, 1976). This potential source of error may be reduced by repeated measurements, but for samples that are old and/or contain little uranium, the labor involved in counting precludes such checks being made. Tracks may also "fade" under the influence of mechanical deformation or in particular chemical environments and thus lead to underestimation of age (Fleischer, 1975). Fission-track dating is rarely used in paleoclimatology but its use in archeology and in tephrochronology is relatively common (Meyer et al., 1991). In most cases where fission-track dating is used, 40Ar/39Ar dating would be preferable, but may not always be feasible.

0 0

Post a comment