A note on estimating the error when subtracting background counts from weak OSL signals

Abstract

The counting statistics of the instrumental background that underlies all OSL measurements was investigated and found not to follow a Poisson distribution. However, if the instrumental background is assumed to have a Poisson distribution, substantial underestimation of the error of the background-corrected OSL signal may occur, especially for samples with a weak OSL signal. An alternative method is suggested for estimating the error on the background-corrected OSL signal by making a separate measurement of the variance of the instrumental background.