Sediment dating by luminescence: a review

Abstract

In the present article we review applications of luminescence methods for the determination of the age of sediments for Quaternary Earth science. “Sediment” is taken to include any mineral particles transported by wind, water and/or ice and subsequently deposited. Methodology is not discussed per se but appears in context. The emphasis is on recent work and is illustrative rather than exhaustive. Applications in archaeology are mentioned only briefly. Successful dating requires that the luminescence signal has been zeroed by sunlight at the time the sediment was laid down. Dune sands and loess transported in full daylight usually satisfy this condition and have been dated by both traditional thermoluminescence and by optical methods. Optical dating is advantageous for less well-bleached sediments, such as are found in colder climates and those laid down by fluvial and glacial processes, although the prospects for dating the latter are shown to be poor. Among less common contexts are volcanic eruptions, earthquake faults and tsunami deposits. The dating of very young sediments, less than 2 ka, is demonstrated and the factors limiting the dating of very old sediments, 500 ka to 1 Ma, are discussed. Validation of any dating method by comparison with other methods is necessary; luminescence ages have been successfully compared with ages obtained by, for example, 1 4C, U–Th series and δ 1 8O isotope ratios. The most difficult problems encountered in dating quartz, feldspars and undifferentiated fine grain mixtures include: anomalous fading (which leads to underestimates of age), incomplete zeroing (which gives overestimates) and sample inhomogeneity. Methods devised to minimise these effects include preheating regimes, selection of observing wavelengths and of stimulating wavelengths for optical dating, the use of single aliquot and single grain measurements. The use of appropriate data analysis procedures is also important. Advice on current best practice on procedures for obtaining reliable ages is offered and, in all applications, there are suggestions as to where further research might be directed.