Category Archives: Luminescence dating

Speed dating?

UntitledLuminescence dating provides an estimate of the time that has elapsed since a sediment was last exposed to daylight. This makes it an extremely valuable technique for dating landforms deposited by the wind, rivers and glaciers. Put in slightly more technical terms it uses a light-sensitive signal that has built up in sand grains from exposure to background radiation (see this acceisble quick fact sheet and this excellent set of guidelines written by Geoff Duller). It has an applicable age range from 10 to 500,000 years, depending on the type of sediment and the amount/rate of background radiation it has been exposed to.

However, as any luminescence dating practitioner is only too aware, the dating process is time consuming and complicated process, involving lots of time in a subdued-light laboratory, a reasonable volume of chemicals (to refine your sample down solely to quartz or to feldspar-dominated fractions) and some pretty expensive and weighty bits of analytical kit. For this reason a number of researchers have been developing ways to assess the depositional age of materials more rapidly. This rapid assessment of age might be particularly useful in the early stages of working out the approximate age of new sites, or if the  research project is focussed on understanding landscape dynamics during a particular time period, such as the late Holocene or the Last Glacial Maximum. To make the dating process more speddy we can either:

1) reduce sample preparation time, or   2) reduce sample analysis time.

In a little more detail, methods include:

  • Luminescence ‘profiling’ in which initial measurements of luminescence behaviour are taken on samples undergoing no, or minimal, chemical pre-treatment (minimising sample preperation time) and this guides the longer, more laborous process of full dating (e.g. Sanderson et al., 2003; 2007; Burbidge et al., 2007).
  • Streamlined chemical pre-treatment for range finder ages (e.g. Roberts et al., 2009; Durcan et al., 2010) again to speed up sample preparation to make initial estimates of ages to guide subsequent full dating procedures.
  • Standardised growth curves, which is an approach to speed up the measurement side of the process rather than sample preparation. It uses fully pre-treated samples but takes fewer analytical measurements for any new sample collected, basing interpolations on an established standardised curve from previous sample (e.g. Roberts and Duller, 2004; Telfer et al., 2008; Yang et al., 2011). This requires some initial effort to build the training curve first, but once established for a region means short amounts of analysis time needed.
  •  Portable luminescence readers (e.g. Sanderson and Murphy, 2010; Kinnard et al., 2011; Stang et al., 2012 and others). These amazing bits of kit measure sediment as it is found in the field (so no sample preparation) and take minute-long measurements of the light emission intensity (rather than many hours to days in the laboratory).

A version of a portable luminescence reader made by SUERC.

So this brings us up to the rationale of what we were doing in this study… Using one of these portable luminescene readers (photo) to make rapid age assessment – our speed dating. It both reduces sample preparation time (we put bulk material as collected in the field in the petri dish and close the door) and reduces sample analysis time (2 blocks of 1 minute, rather than days). However the measurements are instantly transferable to sample age. Therefore, we wanted to see how far we could get toward turning the intentisty of the light signal measured in the portable reader into a rough estimate of actual sample burial age…

So. I chose samples which I had slavishly previously dated using both full sample preparation (to quartz in this case) and full established measurement protocols on which to also measure using the portable luminescence reader to make the comparison. We chose to take a very simple approach to the comparison – that of  making a linear regression of the portable reader signal intensity against established sample burial age (using full dating). The samples chosen span three broad groups of ages (modern, very late Holocene and last interglacial). The relationship between portable reader signal and age was then defined and tested using a second group of samples that had not been used in the regression.

Regression of portable luminescence reader signal (the post-IR BLSL signal) against sample age (both on log axis), which can be used to predict unknown sample ages from their portable reader signals.


The result? The regression is very good suggesting that sample burial age drives the portable reader signal more than other potential factors (which include different mixtures of sediment compostion). This may be because the sites we have chosen, all inside the Namib Sand Sea have a common sedimentary origin.

A word of caution. We are not advocating that this should be done instead of full optically stimulated luminescence dating, with full sample preparation and full analytical protocols (time-intensive as it is, including lots of time in the dark laboratory). Rather, this research has helped to establish the utility of the portable reader to make rapid age estimates in the field, which help us in the field to develop guided sampling strategies when we are addressing particular research questions. For example, if the question was ‘how dyanmic was dune accumulation and migration over the past 5,000 years’ we could rapidly make measurements on small amounts of material in the field to target those parts of the 34,000 km^2 of the Namib Sand Sea that were about 5,000 years old or less. Then we would need only to sample those areas and depths into the sand dunes. This means less sampling in the field, much less material to pack up and ship them home for full analsysis. This is good because sediment samples are heavy and expensive to transport, and we would have also wasted a lots of laboratory time and chemicals preparing the samples and analytical time measuring those samples not relevant to the time frame we are interested in. Why collect, sieve, treat and measure more sand that you need to?

See this link for our full paper in Quaternary Geochonology and

a condensed poster version  for a simpler version of the story.

On a related note, what is the future potential for these bits of kit? It’s not overblown to say they could be out of this world, with a number of groups of luminescence scientists working at developing an instrument that could be used on space missions of the future to our red planet neighbour and beyond…

For example, see PhD project of Maizakiah Abdullah and this presentation by Yukihar and Mckeever