Phenomenological models are frequently used to analyze experimental signals in thermally and optically stimulated luminescence experiments. Typically, these models consist of systems of differential equations describing various electronic transitions. An alternative to the differential equation approach is the use of Monte Carlo (MC) methods, which also allow an estimation of the theoretical stochastic uncertainty of the intensity of the luminescence signal. By running and averaging several MC variants, these stochastic uncertainties are estimated in this paper for various luminescence models.
In the case of first-order kinetics processes, the MC results compare well with previously published analytical results for the coefficient of variation (CV) in stochastic linear pure death processes. By contrast, no analytical results are available for the more general one trap one recombination center model (OTOR), and MC is the only method available for estimating the stochastic uncertainties.
In this paper the CV coefficients are simulated for three commonly used experimental stimulation modes, namely thermally stimulated luminescence (TL), continuous-wave optically stimulated luminescence (CW-OSL) and linearly modulated OSL (LM-OSL). The results of the simulations show that CW-OSL signals have the smallest CV values among the three stimulation modes, and therefore these signals are least likely to exhibit stochastic variations. The stochastic uncertainties in these phenomenological models are discussed in the context of single grain luminescence experiments and nanodosimetric materials, in which one deals with small numbers of charge carriers.