Temperature lags of luminescence measurements in a commercial luminescence reader

Nuclear Instruments and Methods in Physics Research B 359 (2015) 60–63

Contents lists available at ScienceDirect

Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb

Temperature lags of luminescence measurements in a commercial luminescence reader George Kitis a, Nafiye G. Kiyak b, George S. Polymeris c,⇑ a

Aristotle University of Thessaloniki, Nuclear Physics Laboratory, 54124 Thessaloniki, Greece ISIK University, Faculty of Science and Arts, Physics Department, Sile, 34980 Istanbul, Turkey c Ankara University, Institute of Nuclear Sciences, Besßevler, 06100 Ankara, Turkey b

a r t i c l e

i n f o

Article history: Received 17 December 2014 Received in revised form 9 July 2015 Accepted 10 July 2015

Keywords: Temperature lag TLD 100 Luminescence reader De-convolution

a b s t r a c t The temperature recorded in thermoluminescence and optically stimulated luminescence equipments is not the temperature of the sample but that of the heating element on which the thermocouple is attached. Depending upon the rate of heating, a temperature difference appears between the samples and the heating element, termed as temperature lag, which could have serious effects on the curve shapes and trapping parameters. In the present work the temperature lag effect is studied in a newly developed luminescence equipment measuring both thermoluminescence and optically stimulated luminescence. It is found that the temperature lag could be large for heating rates above 2 K/s and it is strongly dependent upon the sample holder. A simple approximation method is proposed in order to both predict as well as correct for temperature lag effects in luminescence measurements. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Thermoluminescence (TL) and Optically Stimulated Luminescence (OSL) research tools are widely applied in the fields of radiation dosimetry and archaeological dating. Currently developed equipments can record both TL and OSL glow curves [1]. The exact knowledge of sample’s temperature is very important, not only for TL but also for OSL measurements. The temperature is measured using a thermocouple, which however, records the temperature of the heating strip and not the one of the sample. The temperature difference between the heating element and the sample is known as temperature lag, whereas the temperature differences within the sample are known as thermal gradient. Betts and Townsend [2] had modeled in detail all these effects for the case of contact heating TL readers, while Betts et al. [3] have performed detailed measurements of the effects. Although the methodology used by [3] analyzed thoroughly the effects in every detail, it is technically very difficult and neither easily repeatable nor applicable. Piters and Bos [4], Pradhan [5] and Furetta et al. [6] had shown the serious effects of the temperature lag on TL glow-curves and on the trapping parameters evaluation. Kitis and Tuyn [7] have proposed a method-approximation in order to correct for temperature lag and thermal gradients, using

only TL measurements. The temperature lag correction is based on the following equation:

T mgj

E-mail address: [email protected] (G.S. Polymeris). http://dx.doi.org/10.1016/j.nimb.2015.07.041 0168-583X/Ó 2015 Elsevier B.V. All rights reserved.

! ð1Þ

where Tmgj and Tmgi are the peak maximum temperatures received at heating rates bj and bi, respectively and c a constant. This aforementioned equation was derived for all cases of kinetics, namely first, second and general, by considering the glow peaks obtained for two different rates of heating bi < bj with the glow peak due to the higher heating rate being shifted towards the higher temperature keeping its integral stable and reducing slightly its peak height. Consequently, Eq. (1) was derived by considering the intensities of the two glow peaks at the same fraction of their maximum intensity and applying the corresponding kinetic order model [7]. In order to apply this specific method, one has to go through three distinctive steps. Step 1: Evaluation of the constant c in Eq. (1), using two very slow heating rates, so that the temperature lags are negligible. For example, for heating rates b1/b2 = 0.5 giving peak maximum temperatures Tmg1 and Tmg2, the constant will be:

c¼ ⇑ Corresponding author.

b ¼ T mgi  c  ln i bj

T mg2  T mg1 0:693

ð2Þ

Step 2: From Eq. (1) and the evaluated constant c, the Tm at every high heating rate is estimated relative to the lower one.

61

G. Kitis et al. / Nuclear Instruments and Methods in Physics Research B 359 (2015) 60–63

Step 3: If the peak maximum temperature of a glow-peak received with a high heating rate is Tmg and Tm is the respective value derived from step 2, then the temperature lag will be:

ð3Þ

The aim of the present work is to assess temperature lag effects appearing in a widely used TL/OSL reader, which will be described below. The dosimeter selected for this study is LiF:Mg,Ti (TLD-100). The reason is that this material is a reference material widely accepted in TL radiation dosimetry, and therefore its glow-curve properties and its physical parameters have been extensively studied.

TL (a.u)

DT ¼ T mg  T m

15000

10000

0 300

350

400

The main interest in the present work is the behavior of all glow-peak maximum temperatures as a function of the heating rate, but not the glow-curve intensity and trapping parameters which were extensively studied [4,5,8–11]. All the experimental glow-curves of LiF:Mg,Ti obtained were fitted using a first order kinetics function [12]:

"

#   E T  Tm T2 E T  Tm IðTÞ ¼ Im  exp 1 þ   2  exp   ð1  DÞ  Dm kT T m kT T m Tm ð4Þ where D is 2kT/E, Dm the value of D at Tm, E (eV) the activation energy, Im the TL intensity at the peak maximum temperature Tm and k (eV/K) the Boltzmann constant. All curve fittings were performed using the MINUIT computer program [13], while the goodness of fit was tested using the Figure Of Merit (FOM) of Balian and Eddy [14] given by:

ð5Þ

where YExper is the experimental glow-curve, YFit is the fitted glow-curve and A is the area of the fitted glow-curve.

450

500

(K)

Fig. 1. Glow-curve of LiF:Mg,Ti de-convolved into its individual glow-peaks.

evaluate the constant c of Eq. (1) for each glow-peak. Once the constant c was evaluated then the corrected values of Tm from Eq. (1) were obtained. The results for glow-peaks 2, 3, 4 and 5 are shown in Fig. 2. The open circles, for all glow-peaks, represent the experimental values without any correction. The solid circles, for all glow-peaks, represent the corrections procedure of Eq. (1) described in introduction section. It can be easily observed that for heating rates up to 1 K/s there are not differences between un-corrected Tmg and corrected Tm values. However, for higher heating rates the difference between un-corrected and corrected Tm values, which defines the temperature lag, becomes obvious. The crosses in Fig. 2 represent the Tm values resulted from a simulation study as following. The activation energies and frequency factors of all glow-peaks are evaluated from the curve fitting analysis. By considering, as in above case, that at the lowest heating rate of 0.25 K/s the temperature lag is negligible, then the (E, s) values obtained (presented in Table 1) can be used to evaluate theoretically the peak maximum positions through the usual first order Randall-Wilkins equation [15]. In fact this simulation is the obvious method to evaluate temperature lag, provided that very accurate values of (E, s) are known. The results of the simulation, which are represented by the crosses, show the excellent agreement achieved between the correction method of Eq. (1) and those of the theoretical simulation. The glow-peak 5 is the most extensively studied glow-peak and many results concerning its trapping parameters are summarized in the GLOCANIN inter-comparison program [9,10]. The values of its trapping parameters evaluated in the present work as a function of the heating rate are listed in Table 2. The temperature lag

550 Tm ( K )

3.1. Method of analysis

A

3

Temperature

3. Experimental results and discussion

i

2

250

The Harshaw LiF:Mg,Ti chips of dimensions 3  3  0.9 cm were used in the present work. The measurements were performed using the RISØ TL/OSL reader (model TL/OSL-DA-15) equipped with a 0.1 Gy/s 90Sr/90Y b-ray source [1]. The reader is fitted with an EMI 9635QA PM Tube. All TL measurements were performed using a combination of a Pilkington HA-3 heat absorbing and a Corning 7–59 blue filter. Seven different heating rates were applied, ranging between 0.25 and 16 K/s, namely 0.25, 0.5, 1, 2, 4, 8 and 16 K/s. Maximum heating temperature was 350 °C in all cases except for the two latter cases where TLD-100 was heated up to 400 °C. The test dose attributed was 0.5 Gy. The study was performed on both sample holders available, namely stainless steel cups and discs, both purchased from RISØ laboratory.

X jY Exper  Y Fit j

4

5000

2. Experimental procedure

FOM ¼

5

20000

500 5

450

4 3

3.2. LiF chips in cup holders

400 2

A curve fitting example of LiF:Mg,Ti is presented in Fig. 1. The glow-peaks on which the interest is concentrated are the glow-peaks 2, 3, 4 and 5, already shown in the Figure. According to the de-convolution analysis, the experimental Tmg values as a function of the heating rate were obtained. The values of Tmg for the lower heating rates of 0.25 and 0.5 K/s were used in order to

350

0.1

1 Heating rate

10 ( K/s )

Fig. 2. Tm as a function of the heating rate for glow-peaks 2, 3, 4 and 5. Open circles represent the experimental values. Solid circles are the corrected values according to Eq. (1) while crosses correspond to the simulated values.

62

G. Kitis et al. / Nuclear Instruments and Methods in Physics Research B 359 (2015) 60–63

Table 1 E and s values of glow-peaks 2, 3, 4 and 5 of LiF:Mg,Ti for a heating rate of 0.25 K/s. E (eV)

s (s1)

2 3 4 5

0.936 1.242 1.677 2.094

7.4801010 2.4551013 1.5881017 2.8351020

560 540 Tm ( K )

Peak

520 500

3.3. LiF:Mg,Ti single grain in cup and disk holder A LiF:Mg,Ti chip was broken and two grains of dimensions 300–500 microns were selected. The first one was positioned at the center of a cup holder while the other one in the center of a disk holder. In both cases the grains were glued using silicon grease. A heating rate sequence was applied to both aliquots and the resulted glow-curves were analyzed as described above. The results are shown in Fig. 5. In the case of the cup holder the results are identical with those of chips given in Fig. 4. However, the situation is different for the case of the disk holder. According to Fig. 5 the temperature lag is greater and furthermore is not the same for all glow-peaks but increases as the peak maximum temperature

Table 2 Trapping parameters of the glow-peak 5 of LiF:Mg,Ti as a function of the heating rate. b (K/s)

E (eV)

s (s1)  1021

0.25 0.5 1 2 4 8 Mean

2.094 2.118 2.186 2.232 2.278 2.215 2.187 ± 0.07

0.284 0.490 2.173 5.097 8.267 0.758 2.81 ± 3

480 460

0.1

1 Heating rate

10 ( K/s )

Fig. 3. Case of glow-peak 5. Open circles are the experimental values. Solid circles show the corrected values according to Eq. (1) and (a) simulation with (E, s) from glow-curves received at heating rate 0.25 K/s, whereas (b–d) simulation with (E, s) from glow-curve obtained using heating rates 2, 4 and 8 K/s.

ΔT ( K )

40 30 20 10 Peak 2 Peak 3 Peak 4 Peak 5

0 0.1

1

10

Heating rate

( K/s )

Fig. 4. Temperature lag as a function of heating rate evaluated from the glow peaks 2, 3, 4 and 5.

100 ΔT ( K )

influences the values of activation energy [4,5,8,11], mainly, due to the temperature lag at the peak maximum. It is not compensated by the slightly increase of the full width, which is the difference of two temperature points of the glow-peak shifted more or less equally. Although the curve fitting analysis of the present work has discriminated the increase of activation energy as a function of the heating rate, all the values listed in Table 2 are within the values listed in the GLOCANIN inter-comparison program [10]. Their mean value has a standard deviation of less than 4%. On the other hand the error in the frequency factor is much higher. Namely, 1% error in the activation energy causes a 25–30% error in frequency factor. The high dependence of the simulation method on the values of trapping parameters (E, s) is shown in Fig. 3 for the case of glow-peak 5. The case (a) is as in Fig. 2. The cases (b), (c) and (d) are evaluated using the trapping parameters values from Table 2 corresponding to the heating rates of 2, 4 and 8 K/s, respectively. Here one can point out very high differences in the theoretical simulation coming from (E, s) having an error less than 3%. The temperature lag, DT, which is the difference between uncorrected Tmg values (open circles in Fig. 2) and the corrected ones (solid circles in Fig. 2) are shown in Fig. 4, as a function of the heating rate. As it becomes prominent, the temperature lag appears even at the low heating rate of 2 K/s and increases fast, reaching the values of 20 K at 8 K/s and almost 40 K at 16 K/s. An important observation is that the temperature lag is more or less the same at the entire temperature region of the glow-curve, since no appreciable differences were found at the peak maximum positions of glow-peak 2, 3, 4 and 5. This means that the temperature lag, more or less, causes a shift of the whole glow-curve towards higher temperatures.

d c b a

Peak 2 Peak 3 Peak 4 Peak 5

80 60 40

Disks

20 0 0.1

Cups

1 Heating rate

10 ( K/s )

Fig. 5. Temperature lag as a function of the heating rate for two different sample holders.

increases. Therefore, the type and obviously the material of the sample holder have a crucial role in the value of the temperature lag. 4. Conclusions The temperature lag in a TL/OSL reader has been evaluated using TL measurements only. It is found that the temperature lag effect is significant and it is ubiquitously present. Specifically, the following important features of the temperature lag were yielded:

G. Kitis et al. / Nuclear Instruments and Methods in Physics Research B 359 (2015) 60–63

1. It depends upon the sample holder. 2. Depending upon the sample holder it appears even at the very low heating rate of 0.25 K/s. 3. It is not necessarily the same across the glow-curve but it could be different at each position of the glow-curve. The above observations are important not only for thermoluminescence studies, especially, when trapping parameters are involved, but also for OSL measurements. The reason is that routine OSL measuring protocols involve various kinds of heating pretreatments (as for example pre-heating. thermal activation etc.). According to the results of the present work a preset pre-heat up to a medium temperature (i.e. 200 °C) in practice could be lower by more than 20 °C, whereas a thermal activation up to, for example, 500 °C in practice could be less than 450 °C. Furthermore, a distortion of sample holders and even of the heater, due to their extensive use, cannot be excluded. All these cases lead to even worse results. Another topic where the temperature lag can has important implications is the Isothermal Thermoluminescence (ITL). This is a very interest topic which must be investigated. It is suggested that any TL/OSL equipment must be tested for the presence of temperature lag, since this effect can be always present even if very low heating rates (not always practical) are used. The test can be very easily performed using any TL material

63

(not necessary a reference one) by applying the simple temperature lag correction method of Kitis and Tuyn [7] used in the present work. References [1] L. Bøtter-Jensen, E. Bulur, G.A.T. Duller, A.S. Murray, Radiat. Meas. 32 (2000) 523. [2] D.S. Betts, P.D. Townsend, J. Phys. D Appl. Phys. 26 (1993) 849. [3] D.S. Betts, A. Couturier, A.H. Khayrate, B.J. Luff, P.D. Townsend, J. Phys. D Appl. Phys. 26 (1993) 843. [4] T.M. Piters, A.J.J. Bos, J. Phys. D Appl. Phys. 27 (1994) 1747. [5] A.S. Pradhan, Radiat. Prot. Dosim. 65 (1996) 73. [6] C. Furetta, G. Kitis, J.H. Kuo, L. Vismara, P.S. Weng, J. Lumin. 75 (1997) 341. [7] G. Kitis, J.W.M. Tuyn, J. Phys. D Appl. Phys. 31 (1998) 2065. [8] A.J.J. Bos, R.N.M. Vijverberg, T.M. Piters, S.W.S. McKeever, J. Phys. D Appl. Phys. 25 (1992) 1249. [9] A.J.J. Bos, T.M. Piters, J.M. Gomez Ros, A. Delgado, Radiat. Prot. Dosim. 47 (1993) 473. [10] A.J.J. Bos, T.M. Piters, J.M. Gomez Ros, A. Delgado, Radiat. Prot. Dosim. 51 (1994) 257. [11] G. Kitis, J.W.N. Tuyn, Radiat. Prot. Dosim. 84 (1999) 371. [12] G. Kitis, J.M. Gomez Ros, J.W.N. Tuyn, J. Phys. D Appl. Phys. 31 (1999) 2636. [13] F. James, M. Roos, MINUIT, CERN program library entry D506, 1977, . [14] H.G. Balian, N.W. Eddy, Nucl. Instr. Meth. 145 (1977) 389. [15] R. Chen, S.W.S. McKeever, Theory of Thermoluminescence and Related Phenomena, World Scientific, Singapore, 1997. p. 29.