Simultaneous Double Dose Measurements Using TLD-100H

Abstract

Thermoluminescent dosimeters (TLD) and optically stimulated luminescent dosimeters (OSLD) are practical, accurate, and precise tools for point dosimetry in medical physics applications. The objective of this study is to investigate the luminescence properties—both OSL and TL—of lithium fluoride (LiF) doped with magnesium (Mg), copper (Cu), and phosphorous (P) (LiF: Mg, Cu, P), commercially known as TLD-100H. The goal is to devise a methodological approach for dose measurement that allows for obtaining two independently measured dose values at each irradiation point, thereby improving accuracy and precision. The luminescence properties of TLD-100H were studied using a beta irradiation source (90Sr/90Y) integrated into the TL/OSL DA-15 automated Risø reader. This study identified the ideal experimental conditions for optimal dose evaluation and used them for dosimeter calibration across doses ranging from 0.5 to 4.0 Gy. The results demonstrated that, under optimal measurement parameters, the OSL and residual thermoluminescence (ResTL) signals—correlated to two trap systems within the dosimeter—exhibited high reproducibility, stability over multiple cycles, and high precision and accuracy (≤2%). Specifically, the OSL response showed good linear behavior across the investigated dose range, while the ResTL signal exhibited linear behavior between 0.5 and 2 Gy and sublinear behavior for doses >2 Gy.

1. Introduction

In 1978, Nakajima et al. [1] first proposed a highly sensitive and tissue-equivalent thermoluminescent material by doping LiF crystals with Mg, Cu, and P impurities. Commercially known as TLD-100H, this material has been extensively studied by various researchers [2,3,4,5,6,7,8,9,10,11]. According to Bilski et al. [9], the optimal dopant concentrations for desirable dosimetric properties—such as stability and reproducibility—are as follows: Mg ≈ 0.2 mol%, Cu = 0.02–0.05 mol%, and P = 1.0–3.0 mol%. Compared to its predecessor, LiF: Mg, Ti (TLD-100), TLD-100H offers several advantages including higher sensitivity, an improved signal-to-noise ratio, an ideal tissue-equivalence response to low-energy photons, and a simple glow-curve structure [2,6,11,12,13,14]. In general, efficient dosimetric phosphors might exhibit both thermoluminescence (TL) and optically stimulated luminescence (OSL) properties. OSL offers advantages over TL, including high sensitivity, precise light delivery, fast readout times, and simpler automation. Additionally, TLDs exhibit a supralinearity behavior in the initial dose range. The ideal OSL material should satisfy several, sometimes conflicting, characteristics. It should have deep thermally stable traps for long-term storage of dosimetric information without significant fading. Simultaneously, these traps should be optically accessible when using light sources with wavelengths well-separated from the emission bands of the recombination centers. Investigations into OSL properties of several known thermoluminescent materials, including LiF (Mg, Ti), Li₂B₄O₇(Cu), CaSO₄ ™, CaF₂ (Mn), and Al₂O₃^©, have demonstrated their usefulness for clinical dosimetric measurements [15,16,17,18,19,20,21,22]. This work explores the possibility of simultaneously utilizing signals from optical stimulation with a blue diode (OSL) and residual signals from thermal stimulation (ResTL) using TLD-100H. The OSL and TL characterization includes an investigation of the ideal sample temperature during OSL readout, post-annealing conditions and pre-heating treatments, and the fading properties of the phosphor. With this information, we studied the stability of sensitivity and constructed well-defined calibration curves for dose determination using optimal measurement parameters.

2. Materials and Methods

TLD-100H rectangular chips were used for all the experiments performed.

Table 1. Specifications of LiF: Mg, Cu, P [23,24].

Property	Values
Density, ρ (g/cm3)	2.5
Effective atomic number, Zeff	8.31
Temperature of the main TL peak (°C)	≈210
Emission Spectra (nm)	≈370
Sensitivity	1 pGy–10 Gy
Sensitivity relative to LiF	30
Depth × Width × Height (mm)	3.16 × 3.16 × 0.89

shows rough estimate values of the specifications of TLD-100H.

For the TL and ResTL signals—used to determine kinetic parameters, experimental glow peaks, and the luminescence contributions of different peaks—a glow-curve deconvolution (GCD) with the first order kinetics equation [25] was employed for analysis:

$$I\left(T\right)=I_m\exp \left[1+\frac{E}{kT} \frac{T-T_m}{T_m}-\frac{T^2}{{T^2}_m}\times \exp \left[\frac{E}{kT} \frac{T-T_m}{T_m}\right]\left(1-\Delta \right)-{\Delta }_m\right]$$

(1)

where $I$ is the glow-peak intensity, $I_m$ is the intensity at the peak maximum, $T_m$ (K) is the temperature at the peak maximum intensity, $T$ is the stimulation temperature during readout, $E$ (eV) is the activation energy, $k$ (eV K⁻¹) is the Boltzmann constant, with $\Delta =2kT/E,{\Delta }_m=2k{T}_m/E$.

The Risø TL/OSL reader system model TL/OSL DA-15 from DTU Nutech University (Frederiksborgvej 399 DK 4000 Roskilde, Denmark) was used for luminescence measurements. The emitted luminescence was measured by a light detection system equipped with an EMI 9235QA photomultiplier tube and a detection filter (Hoya U340 optical filter, 260–390 nm region). In the reader, the samples were irradiated using a calibrated beta source based on the radionuclide 90Sr/90Y with a dose rate of 0.069 Gy/s at the sample position. The samples were stimulated with blue LEDs (470 ± 30 nm) using the continuous wave (CW) stimulation method (∼30 mWcm⁻² at 90% power) for OSL or an ohmic heating plate, thermocouple for TL. All measurements in the Risø system were conducted in a nitrogen atmosphere.

Additionally, in most experiments, following the OSL measurement, the ResTL was subsequently measured. For the OSL and ResTL calibration curves, it was necessary to conduct studies to determine the optimal experimental parameters, including pre-heating temperature conditions, temperature values during optical stimulation, and fading evaluation.

2.1. Effect of Temperature OSL Readout on ResTL

To assess the impact of temperature during the OSL readout on ResTL and to determine the appropriate reading temperature, three dosimeters were employed. These dosimeters were set at different temperatures during the OSL readout, ranging from 40 °C to 80 °C in 10 °C increments (see

Table 2. Parameters for studying the effect of temperature during the OSL readout on residual thermoluminescence (ResTL).

Step	Treatment	Observed
1	Annealing (Heating @ 250 °C for 60 s)
2	Annealing (Heating @ 250 °C for 60 s)
3	$\beta$-irradiation (2 Gy)
4	CW-OSL with Blue LEDs at X °C (40.00 s, 90% power)	OSLx
5	Heating@250 °C, 1 °C/s	ResTLx

Steps 1–5 were repeated for different X temperatures: 40 °C, 50 °C, 60 °C, 70 °C, 80 °C, 90 °C.

).

For the optically stimulated signal, the estimated background (BG) from the last 3.2 s was typically subtracted from the initial 0.8 s to calculate the OSL signal (OSLin) [26]. The readout background evaluation was crucial because it could have included various contributions, such as counts intrinsic to the photomultiplier tube, Raman-scattered photons, residual scattered light from the stimulation source, and OSL associated with less light-sensitive electron traps that remain unfilled after n measurement points [27].

2.2. Pre-Heat Temperature in OSL and ResTL

To determine the appropriate pre-heating temperature and assess its impact on the OSL, TL, and ResTL curves, different pre-heating temperatures (150 °C for 10 s, 170 °C for 10 s, and 190 °C for 10 s) were applied to three TLDs. The OSL signal was obtained from blue light stimulation, the ResTL from temperature stimulation following blue light stimulation, and the TL from temperature stimulation without light stimulation (

Table 3. Parameters for the study of the pre-heating temperature in OSL and ResTL.

Step	Treatment	Observed
1	Annealing (Heating @ 250 °C for 60 s)
2	Annealing (Heating @ 250 °C for 60 s)
3	$\beta$-irradiation (2 Gy)
4	Pre-heat (X °C for 10 s)
5	Heating@250 °C, 1 °C/s	TLx
6	Annealing (Heating @ 250 °C for 60 s)
7	Annealing (Heating @ 250 °C for 60 s)
8	$\beta$-irradiation (2 Gy)
9	Pre-heat (X °C for 10 s)
10	CW-OSL with Blue LEDs at 40 °C (40.00 s, 90% power)	OSLx
11	Heating@250 °C, 1 °C/s	ResTLx

Steps 1–10 were repeated for different X pre-heating temperatures: 150 °C, 170 °C and 190 °C.

).

For the annealing procedures, two heating cycles at 250 °C for 60 s, through a thermocouple mounted in the Risø TL/OSL reader system, were used.

To assess the sensitivity changes resulting from annealing, three TLDs were consecutively irradiated and read (for OSL and ResTL signals) using the correct parameters determined from previous tests (Section 2.1 and Section 2.2).

2.3. Fading and Calibration Curves

To evaluate fading phenomena, ten dosimeters were exposed to a beta source (2 Gy) after the annealing procedure. The luminescence measurements (OSL and ResTL) were conducted at different time intervals between the irradiation and readout (0 h, 2 h, 4 h, 12 h, 20 h, 26 h) to assess the region of signal stability.

The OSL and ResTL calibration curves were determined using five TLDs for each dose point. The doses administered were: 0.5 Gy, 1.0 Gy, 1.5 Gy, 2.0 Gy, 3.0 Gy, 4.0 Gy.

3. Results and Discussion

Figure 1 presents an example of the experimental TL data and the luminescence contributions of the four individual peaks obtained through the deconvolution procedures using Equation (1). The fitting procedures were performed using in-house developed software.

Table 4. Fit parameters of the four peaks obtained by the deconvolution procedure.

Peak	${\mathit{T}}_{\mathit{m}}$ (°C)	$\mathit{E}$ (eV)
1	74 ± 2	0.93 ± 0.05
2	120 ± 3	1.22 ± 0.06
3	174 ± 3	1.35 ± 0.07
4	222 ± 4	1.97 ± 0.07

displays the fit parameter values obtained through the deconvolution procedures. The quality of the fit was assessed using the figure of merit (FOM) proposed by Balian and Eddy [28]. For each fit, an FOM value of approximately 2% was observed, which typically indicates a satisfactory fit.

3.1. Effect of Temperature OSL Readout on ResTL

In Figure 2a, we observe that the ResTL signal exhibits thermal quenching in the dosimetric peak as the OSL readout temperature increases. Additionally, the ResTL glow curve shows a depletion of TL peak 2 at higher OSL readout temperatures, with a nearly complete removal of TL peak 2 at 80 °C readouts. This observation aligns with the fact that higher OSL readout temperatures deplete the low-temperature peaks. Because the initial OSL readout temperature is 40 °C, TL peak 1 is completely depleted and not visible in any of the glow curves.

The observed increase in the OSL signal with an increasing readout temperature in Figure 2b is partly due to the thermal de-trapping of the phototransferred charges (PTTL) and partly to the reduced probability of re-trapping during readout. However, as the readout temperature increases, the OSL decay curve deviates from the usual stretched-exponential behavior. This deviation is interpreted as resulting from the increased contribution of phototransferred charges at higher temperatures, specifically the thermal release of charge from shallow traps. Notably, the decay rate does not appear to increase or decrease significantly, suggesting that higher temperatures do not significantly affect the decay rate.

To gain further insight into the behavior of the ResTL peaks with increasing OSL readout temperatures, we conducted a glow-curve deconvolution (GCD) procedure. This allowed us to extract the luminescence contributions from TL peaks 2, 3, and 4 and observe their evolutions as the OSL readout temperature increased. TL peak 1, having a negligible contribution, was excluded from the GCD procedure.

Figure 3 illustrates the normalized ResTL intensity obtained from the three TL peaks as they vary with the increasing OSL readout temperature. Notably, ResTL peak 2 exhibits a significant decrease as the OSL readout temperature rises. This finding confirms our previous observations that the OSL signal is strongly linked to this particular TL peak. Additionally, the consistent stability of TL peak 4 underscores its reliability for dosimetric purposes. The OSL readout temperature between 40 and 60 °C significantly affects the depletion in peaks 2 and 3, while the re-trapping effects remain negligible. As the OSL readout temperature increases, in addition to the emptying effects of the traps, the recombination effects become more pronounced as the number of de-trapped electrons increases.

As shown in Figure 4, a direct comparison of the ResTL measured during the OSL readout at room temperature (RT) without pre-heating conditions and the TL glow curves confirms that the OSL signal of LiF: Mg, Cu, P is strongly correlated with TL peak 2 and partly correlated with TL peak 1. These findings align with observations from previous studies [29,30].

The comparison between the ResTL and TL signals reveals that the effect is negligible on the other TL peaks. However, the correlation between the OSL signal and the low-temperature TL peaks suggests the possibility of high OSL fading. This fading was measured at more than 60% (for integral continuous-wave OSL) and nearly 90% (for maximum continuous-wave OSL amplitude) signal loss after 24 h of storage in darkness [31]. Implementing a high pre-heat, pre-OSL readout could help minimize the effects of this significant fading.

3.2. Pre-Heat Temperature in OSL and ResTL

In Figure 5a, we observe two distinctive features in the ResTL signal as the pre-heating temperature increases. The primary feature is the reduction in intensity of the main dosimetric peak 4. Additionally, the ResTL glow curve shows a depletion of TL peak 3 with an increasing pre-heating temperature. This observation aligns with the expected behavior, as pre-heating is assumed to anneal the low-temperature peaks. Notably, the same behavior is also observed in the TL signal after different pre-heating temperatures.

Figure 5b illustrates how the increasing pre-heating temperatures affect the luminescence of TL peaks 3 and 4, as previously mentioned. TL peak 3’s contribution at a pre-heat of 190 °C is approximately 5–10% of the contribution at a pre-heat of 150 °C, approaching background levels. Meanwhile, TL peak 4 exhibits a slight decrease in value from a pre-heat of 150 °C to 170 °C and a significant loss of luminescence at 190 °C.

In Figure 6a, the OSL decay curve is shown for different pre-heating temperatures. Higher pre-heating temperatures result in relatively lower intensities in the OSL decay curve.

Figure 6b illustrates how the normalized OSL signal and its components change with increasing pre-heating temperatures. The plot reveals that the normalized OSL signal decays linearly with rising pre-heating temperatures, similar to the initial signal and the background. This observation confirms the decay curves seen in Figure 6a, where higher pre-heating temperatures lead to decreased OSL intensity.

An interesting aspect highlighted by Figure 6b is the linear decrease with the pre-heating temperature among three components: the initial signal, the background, and the OSL signal. However, this decrease is not uniform; at higher pre-heating temperatures, the background decreases more rapidly than the other two. This finding suggests that at a pre-heating temperature of 190 °C (located at the lower end of the plot), there is less contribution from the background. Consequently, this translates to greater accuracy and a better OSL signal.

Thus, from the perspective of OSL as well, a pre-heating temperature of 190 °C is preferred for dosimetric purposes. To test the effect of pre-heating at 190 °C for 10 s with heating at 40 °C during the optical stimulation, and to evaluate changes in sensitivity due to different irradiation cycles and readings, three dosimeters were irradiated (2 Gy). Subsequently, the OSL and ResTL signals were recorded for eight consecutive cycles. The results obtained are shown in Figure 7, with both the OSL and ResTL signals normalized to the value obtained during the first cycle of irradiation and the readout.

We previously established that the OSL signal of LiF: Mg, Cu, P is strongly correlated with TL peak 2 (located around 123 °C) and partially correlated with TL peak 1 (around 77 °C). However, both peaks are erased after a pre-heat of 190 °C or a TL readout of 250 °C, as shown in Figure 5. However, a distinct OSL signal persists. This signal arises from thermally stable, optically active traps that remain unaffected by either the 190 °C pre-heat or the 250 °C TL readout.

3.3. Fading and Calibration Curves

Figure 8 displays the OSL and ResTL signals, normalized to the signal value measured at time 0 h, after various time intervals between the irradiation and readout.

The OSL signal exhibited approximately 20% fading within the first 4 h. After this initial period of substantial instability, the signal gradually stabilized. Interestingly, the fading appeared to follow a logarithmic trend, with a significant loss immediately after irradiation, which then diminished over time. For the ResTL signal, a 6% loss was observed during the first 4 h, after which no significant fading occurred. Notably, the TL signal of TLD-100H exhibited a 2% fading over 3 months [23,31,32].

Figure 9 illustrates the OSL and ResTL signals in relation to the absorbed dose. The red line represents a polynomial fit describing the ResTL signal response, while the black line corresponds to a linear fit of the OSL signal response. For detailed fit parameters, refer to

Table 5. The fit parameters for the OSL and ResTL responses.

OSL	Linear fit: y = a + bx	a = 3.24 × 102 ± 6.40 × 101	R2 = 0.999
		b = 3.48 × 103 ± 2.70 × 101
ResTL	Pol fit: y = a + bx + cx2	a = 6.08 × 106 ± 1.98 × 106	R2 = 0.999
		b = 1.48 × 108 ± 2.12 × 106
		c = −1.26 × 106 ± 4.60 × 105

.

4. Conclusions

The results presented show the good accuracy of TLD-100H in the 0.5–4.0 Gy dose range. The proposed methods allow us to have two dosimetric responses for the same measurement point obtained with two different stimulation methods: linear with optical stimulation and polynomial with thermal stimulation (after OSL). These responses, corresponding to two different trap systems in the dosimeter, allow for greater accuracy in the whole range investigated thanks to the linear response of the TLDs with optical stimulation. Further extensive study is needed to understand how the TLD responds to other radiation sources, such as MV photon beams.