Determining the elapsed time since exposure to ionizing radiation is crucial in worker accident scenarios for accurate dose reconstruction and timely medical intervention. Detailed timestamps are essential for regulatory compliance, incident investigations, and safety improvements. Furthering previous studies [1,2], this work evaluated using the signal loss in luminescent materials over time. Discs made from GR-200A (LiF:Mg,Cu,P) [3], with higher sensitivity than TLD-100 (LiF:Mg,Ti), enabled the measurement of lower doses, important for environmental monitoring and personal dosimetry. The study focused on the decay of peaks II and III, with half-lives of approximately one day and 3-4 months, respectively, and the stable peak IV, used as a reference. Crystals were placed on a solid water support in a Plexiglas apparatus with a 2.22 MBq Ra-226 source. A thin layer of solid water served as a buildup thickness. Irradiated discs were read using a Harshaw model 4500 reader, followed by an annealing phase at 240 °C. The mass of each crystal was measured, and charge values were reported per unit mass of the phosphor. Glow curve decomposition was performed using GlowFit v1.3 software, which determines trap energy depth, temperature, and maximum intensity of each thermoluminescence peak. Initially, each crystal's response was characterized to identify homogeneous subgroups with the same response at the same absorbed dose, with a maximum deviation of 5%. Combinatorial analysis techniques compared crystals based on the fraction of total charge associated with each peak. A double irradiation cycle with different final doses (~1 mGy in 4 hours and ~4 mGy in 16 hours) excluded subgroups whose response homogeneity depended on the absorbed dose. Chauvenet's criterion identified and removed outliers. The homogeneous subgroups were irradiated again and read according to a predetermined timeline. The temporal variation of the ratios of peaks II to IV and III to IV generated fading curves, fitted by least squares to a single-term exponential model, y=a exp(bx) (b<0). The homogeneity requirements resulted in remarkable consistency in the fading curves. The maximum uncertainty in elapsed time estimates was ±1 day and ±1 week during the first week and the first two months after exposure, respectively. The minimum detectable dose (MDD) of the dosimeter, calculated as three times the standard deviation of the zero-dose reading, was a few μGy. To expand the study, the depth of energy levels responsible for thermoluminescent properties was statistically determined. Histograms showed the distribution of trap depth values calculated using GlowFit for each crystal [4]. Knowing the time elapsed between irradiation and readout, contributions from crystals with the same fading interval were separated. This revealed an apparent redistribution of electrons in the energy levels of peaks II and III and suggested a possible fine structure of the energy level for dosimetric peak IV. Further investigations are ongoing to validate these results, as adequate characterization would provide an additional method for determining elapsed time since exposure. Analyses of other TLD materials are also underway, as the half-life of the peaks constrains the ability to determine the elapsed time since exposure.