- Email: [email protected]
On isothermal heating as a method of separating closely collocated thermoluminescence peaks for kinetic analysis
Journal of Luminescence 155 (2014) 70–78
Contents lists available at ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
On isothermal heating as a method of separating closely collocated thermoluminescence peaks for kinetic analysis M.L. Chithambo n, P. Niyonzima Department of Physics and Electronics, Rhodes University, PO Box 94, Grahamstown 6140, South Africa
art ic l e i nf o
a b s t r a c t
Article history: Received 9 September 2013 Received in revised form 11 June 2014 Accepted 12 June 2014 Available online 20 June 2014
The experimental separation of overlapping peaks is a common practical concern in thermoluminescence experiments. In this regard, we have investigated the use of isothermal heating as a method for separating closely collocated peaks for kinetic analysis. The method has been developed and applied on a sample of synthetic quartz in which the thermoluminescence intensity is not reproducible in successive measurements on the same aliquot. Preparatory experiments performed to establish a suitable combination of temperature and heating time are described in detail. A comparative analysis of results from conventional thermal cleaning with those from isothermal cleaning, done as a means to assess the effectiveness of the latter, has shown isothermal heating to be a reliable method of separating closely collocated glow-peaks. The method is conveniently applied in combination with a new technique for kinetic analysis based on the use of the temperature-dependence of the area under an isothermal decaycurve. Isothermal heating was also applied to separate collocated peaks in europium-doped orthosilicate (Sr2SiO4:Eu2 þ ), a luminophor, and in carbon-doped aluminium oxide (α-Al2O3:C), a dosemeter. The set studied in Sr2SiO4:Eu2 þ , unlike in the quartz, consists of peaks where the intensity of the lowertemperature one is greater than subsequent ones at higher temperature. On the other hand, the example investigated in α-Al2O3:C is similar except that the dominant peak subsumes the lower intensity secondary peak to the extent that they appear as one. & 2014 Elsevier B.V. All rights reserved.
Keywords: Thermoluminescence Peak separation Phosphorescence Quartz Sr2SiO4:Eu2 þ α-Al2O3:C
1. Introduction Thermoluminescence (TL) is a well-documented method of measuring low light levels from a previously irradiated semiconductor or insulator under a controlled heating rate [1]. The temperature-dependence of the emission appears as a set of peaks collectively termed as a glow-curve with each peak corresponding to an electron-trapping defect. TL glow-curves in some cases consist of well-defined peaks, and in others, of complex sets of overlapping ones as can be appreciated by a cursory examination of any of the many TL papers in the literature. Characteristic features of the peaks may be understood in terms of kinetic parameters, for example the activation energy, defining the electron traps. These parameters aid the description of the mechanisms involved in luminescence emission in the material concerned [1]. However, analysis of a peak for kinetic parameters can usually be conveniently done if the peak is isolated. As such, the experimental separation of overlapping TL peaks is a common practical problem and is the concern of this report.
n
Corresponding author. Tel.: þ27 46 603 8450; fax: þ 27 46 603 8757. E-mail address: [email protected] (M.L. Chithambo).
http://dx.doi.org/10.1016/j.jlumin.2014.06.026 0022-2313/& 2014 Elsevier B.V. All rights reserved.
The conventional method of separating overlapping peaks involves heating to just beyond the maximum of the first peak in order to deplete charge responsible for the ‘satellite’ thereby leaving the next peak with a clean rising edge. The main requirement in this so-called thermal cleaning method is that the partial heating should substantially remove charge responsible for the lower temperature peak. It seems to be reasonable then to expect that any other procedure that also selectively removes charge from a particular peak should affect thermal cleaning. In view of this, this report is concerned with static or isothermal rather than dynamic heating as a method of thermal cleaning. Isothermal heating as a procedure for separating overlapping peaks was discussed by Pagonis and Shannon [2] who applied it on LiF:Mg,Ti. The method is based on the fact that when an irradiated sample is held at a constant temperature, some charge is lost from its electron traps. For this reason, isothermal heating should achieve the same purpose as the conventional thermal cleaning method. Pagonis and Shannon [2] observed the method to be of particular use in the case of first-order kinetics and added that it can be better exploited in combination with curve-fitting. One of the key requirements in the method of Pagonis and Shannon [2] is that the luminescence intensity from the same aliquot should be reproducible. To our knowledge, the technique has not been reported for materials in which this is not the case.
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
This report is concerned with isothermal heating as a method of thermal cleaning and has been developed and applied on synthetic quartz in which the sensitivity changes due to the predose effect. The quartz used here therefore exemplifies a material in which the TL intensity is not reproducible in successive measurements. Coincidentally, the glow-peaks on which the method was applied are closely collocated unlike in LiF (the test sample of Pagonis and Shannon [2]) where the peaks used are somewhat well separated. Preparatory experiments necessary before application of the isothermal cleaning method in this case have been described in detail. The effectiveness of the method was assessed by comparing kinetic parameters found from an isothermally cleaned peak with those determined from a peak thermally cleaned using the conventional thermal cleaning procedure. The method is most effective when used together with a new method of kinetic analysis [3] that uses the temperature-dependence of the area under an isothermal decay-curve. We emphasize here that results from isothermal and conventional thermal cleaning methods were compared only to provide some evidence of reliability of the isothermal cleaning method to be discussed. The comparison must not be misinterpreted as an attempt to find which method is better. The discussion is organized into three parts. The first one (Section 3.1) briefly presents results of kinetic analysis following conventional thermal cleaning. The second part (Section 3.2) describes experiments intended to establish a suitable combination of measurement temperature and heating time for use in isothermal heating. The third part (Section 3.3) then discusses results of kinetic analysis following isothermal heating. Taking advantage of phosphorescence emitted during isothermal cleaning, we applied (see Section 3.3.3) a new method of kinetic analysis [3] based on the temperature-dependence of the area under an isothermal decay curve. The utility of the isothermal cleaning method is further demonstrated when the technique is applied to isolate TL peaks in europium-doped orthosilicate (Sr2SiO4:Eu2 þ ), a luminophor (Section 3.3.4) and to extract a collocate of the main peak in carbon-doped aluminium oxide (α-Al2O3:C), a dosemeter (Section 3.3.5). The set of peaks studied in Sr2SiO4:Eu2 þ , unlike in the quartz, exemplifies a case where the intensity of the first peak at a lower temperature exceeds that of the neighbouring one at the higher temperature. On the other hand, the example studied in αAl2O3:C is similar except that the dominant component subsumes
71
the weaker-intensity one to the extent that the peaks appear as a single peak.
2. Experimental methods The synthetic quartz used in this work was ground into coarse grains (90–500 μm) from a block supplied by Sawyer Research Products (OH, USA). Experiments were conducted using a RISØ TL/ OSL-DA-20 Luminescence Reader. The luminescence was detected by an EMI 9235QB photomultiplier tube through a 7 mm Hoya U340 filter (transmission band 260–390 nm). Samples were irradiated in situ at room temperature using a 90Sr/90Y beta source at a dose rate of 0.10 Gy s 1. The thermoluminescence was measured in a nitrogen atmosphere to prevent spurious signals from air and to improve thermal contact between the sample holder and the heater planchet.
3. Results and discussion The synthetic quartz used in this study produces negligible or no thermoluminescence if it is heated after irradiation only. Thermoluminescence is only properly observed after the sample has been preheated to a temperature above 300 1C before any irradiation. Fig. 1 shows a glow-curve measured at 5 1C s 1 from a sample of synthetic quartz irradiated to 30 Gy. The glow-curve consists of at least 5 peaks (as clarified in the insets) at about 200 and 390 1C as well as the dominant peak near 100 1C which is overlapped at both ends by weaker intensity peaks at approximately 68 and 135 1C respectively. In illustrating the method of isothermal cleaning, this report discusses the kinetic analysis of the main peak and its lower-temperature ‘satellite’ in view of their location as part of the first three closely collocated peaks. The reason for choosing the two peaks studied will be explained later in the text. To aid visual clarity, the glow-curve otherwise measured to 500 1C is shown truncated at 280 1C in Fig. 1. For ease of reference, the main peak is labelled MP and its ‘satellites’ denoted S1 and S2. In this section and elsewhere in the text, collocation refers to the case where a significant component or all of a peak is concealed by another with which it overlaps. Examples of collocated peaks include high-temperature secondary peaks in γirradiated α-Al2O3:C [4], TL emission near 700 nm in Mn or Cr doped beryl [5], TL in X-ray irradiated CaF2:Dy [6] and many others in the literature. The main peak in α-Al2O3:C is a topical example and will be discussed in Section 3.3.5. 3.1. Kinetic analysis using the conventional thermal cleaning method
Fig. 1. A glow-curve measured from synthetic quartz at a heating rate of 5 1C s 1 following irradiation to 30 Gy. The insets show in (a) a glow-curve measured after irradiation but without any prior heating and in (b) the semi-logarithmic plot of intensity against temperature included to better show weaker intensity thermoluminescence peaks beyond 150 1C.
In these series of experiments, the main peak was cleaned of its lower temperature overlapping ‘satellite’ by applying the conventional thermal cleaning method described by McKeever [1]. Wherever necessary, a sample irradiated to 30 Gy was heated to 70 1C at 5 1C s 1 in order to remove peak S1 after which the complete glow-curve was measured by heating to 500 1C at the same rate. It should be emphasized that the only necessary and sufficient condition for using the thermal cleaning method is that the sample ought to be heated to a temperature just beyond the maximum of the peak immediately before the one to be analysed. Kinetic parameters associated with the main peak were determined, after conventional thermal cleaning using the initial-rise and various heating rate methods discussed elsewhere [1]. The values found are subsequently compared with those evaluated after an alternative thermal cleaning method, isothermal heating, to be described later.
72
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
3.1.1. Analysis using the initial-rise and variable heating rate methods The activation energy of the main peak using the initial-rise and variable heating rate methods was found, after conventional thermal cleaning, as 0.77 70.02 eV and 0.65 7 0.02 eV respectively. The value from the variable heating rate method is slightly less than that from the initial-rise method. Of interest is that Pagonis et al. [7] recently showed that the variable heating rate method can underestimate E values whereas the initial-rise method consistently produces correct values of the activation energy. 3.2. Tests preparatory to use of isothermal heating as a method of thermal cleaning The second series of kinetic analyses was carried out on the main peak cleaned of its lower-temperature overlap using isothermal heating. The purpose of this procedure was to achieve the same aim as conventional thermal cleaning, that is, to remove a preceding peak without necessarily affecting the adjacent one. Whereas the main protocol in dynamic thermal cleaning is simply to exceed the maximum temperature of the lower temperature ‘satellite’ peak, the eponymous isothermal cleaning method requires that a sample be held at a temperature on the rising edge of the ‘satellite’ peak for a certain time until its removal reveals a clean rising edge of the next peak. The aim of this section is to describe a set of measurements that were carried out in order to establish a suitable combination of measurement temperature and heating time for use in the isothermal cleaning method. The procedure used in this study is based in outline on the work of Pagonis and Shannon [2] who developed an isothermalheating based protocol for separating overlapping thermoluminescence peaks involving three main steps. The first step is to experimentally determine the number of overlapping peaks. Pagonis and Shannon [2] suggested the usage of either the E T stop or T m T stop method for this purpose. Once the number of peaks and their position are known, the sample is irradiated and heated isothermally at a temperature below the maximum of the first peak. If successful, this step produces the next of the overlapping peaks with a clean rising edge. Lastly, a complete first peak is digitally mustered by subtracting the isothermally cleaned peak from the original glow-curve. This procedure therefore produces two results: a complete first peak and the adjacent one with a clean rising edge. The peaks can then be analysed for their kinetic parameters using suitable methods. The procedure of Pagonis and Shannon [2] is explicit in expecting that the overlapping peaks in question are somewhat well separated or at least appear with distinguishable peak maxima. In applying this method to LiF:Mg,Ti; they found that isothermal heating at an appropriate temperature for a proper duration reduced the intensity of the lower temperature peak but not that of the succeeding one. They stated that the successful application of their method is based on the assumptions that (a) the peaks are sufficiently separated in activation energy E and frequency factor s, (b) the TL intensity is reproducible in successive measurements on the same aliquot, and (c) the position of the first peak is independent of the duration and temperature of isothermal heating. They concluded that their method is most appropriate for first-order TL peaks. The measurements reported here were made and developed on synthetic quartz where the overlapping peaks in question are not well separated but are, incidentally, closely collocated. The sample is also affected by the pre-dose effect and thus the TL intensity is not reproducible in successive measurements on the same aliquot. This means that our material exemplifies samples that do not
satisfy one of the criteria set out by Pagonis and Shannon [2], namely that the TL intensity ought to be reproducible in successive measurements on the same aliquot. As such, their procedures, which were devised from practical considerations on LiF:Mg,Ti, needed to be adapted for use in the quartz whose TL intensity, unlike in LiF: Mg,Ti, was not reproducible when re-measured. This section therefore describes a gamut of tests that were made preparatory to using isothermal heating as a method of thermal cleaning. The measurements discussed may be of interest to others planning to use isothermal heating for thermal cleaning in materials that do not satisfy the criteria of Pagonis and Shannon [2]. We describe experiments on how to choose the temperature for and duration of isothermal heating. We examine the influence of temperature, time and their combination on overlapping TL peaks. We consider the stability of peak positions and the influence of sample re-use, inevitable in TL procedures, on TL features of the peaks. The method has been applied on the closely collocated first three TL peaks apparent in the glow-curve of synthetic quartz as evident in Fig. 1. The various tests are described with respect to the main peak, the only well-defined of the three peaks. It should be noted that since MP is the peak to be analysed, the isothermal heating was done on peak S1. On the other hand, if peak S2 were the peak to be analysed, then peaks S1 and MP would have had to have been removed in this way. Thus the principle is the same for conventional thermal cleaning, namely that the peak preceding the one of interest is the one that needs to be removed. In the final part of the report on synthetic quartz (Section 3.3.3), we will show how the phosphorescence emitted during isothermal heating can be exploited for kinetic analysis using a new method reported elsewhere [3]. This will be discussed with respect to peak S1 but for completeness, results for peak MP will also be presented. The reason for using this new method rather than the whole glow-peak technique as in the method of Pagonis and Shannon [2] will then be explained. In the quartz discussed, the intensity of the first peak (S1) is less than that of the succeeding one (MP). Examples where this is not the case will also be looked at. 3.2.1. Establishing the number of component peaks using the T m –T stop method The glow-curve measured from a sample of synthetic quartz at 5 1C s 1 after irradiation to 30 Gy was shown in Fig. 1. The number of peaks overlapping the main one (peak MP) on its rising edge was confirmed to be only one at 68.8 71.1 1C and to be of firstorder based on the fact that the position of this peak did not shift by more than 2 1C with dose or with pre-heating from 30 to 78 1C. Similarly, the position of peak MP was stable at 101.1 7 1.2 1C. The corresponding values of the activation energy, calculated using the initial-rise method, are shown in an E–T stop graph in Fig. 2. There is a step-like advance from the set E¼ 0.50 7 0.03 eV corresponding to the first peak (S1) to E¼ 0.78 70.01 eV for peak MP. 3.2.2. Influence of Tstop on intensity of the main peak The influence of change in Tstop on the luminescence intensity of the main peak is shown in Fig. 3. The intensity is essentially independent of Tstop up to 56 1C beyond which it decreases: less rapidly up to 70 1C and with a more pronounced reduction thereafter. Two points immediately become apparent from the nature of Fig. 3. The first is that in the sample studied here, partial heating to a temperature Tstop is de facto thermal cleaning when T stop 4 68 1C, the position of peak S1. Thus conventional thermal cleaning affects
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
Fig. 2. The activation energy corresponding to a temperature a sample irradiated at 30 Gy was partially heated to at 5 1C s 1.
Fig. 3. Dependence of the luminescence intensity of the main peak on Tstop.
the peak of interest (peak MP in this study). The second fact is that if a temperature for isothermal heating is to cause a decrease in the intensity of the ‘satellite’ peak but not that of the peak of interest (i.e. to decrease intensity of peak S1 rather than that of peak MP), the heating temperature needs to be selected in the ‘plateau’ region say up to 56 1C in this example. 3.2.3. Reproducibility of the position of the main peak The position of the main peak was measured 50 times consecutively at 5 1C s 1 and determined to be stable at 100 1C. This feature is particularly important because isothermal cleaning involves several cycles of measurement on the same sample and the unstated assumption in such tests is that the peak position is constant. 3.2.4. Influence of repetitive measurement on intensity of the main peak The influence of repetitive measurement on the intensity of the main peak is shown in Fig. 4. The change in intensity is shown for five sets of measurement in which the same sample was measured up to ten times in each set. The same sample (and not different samples of different mass) was used in the five sets. In the first one, the thermoluminescence intensity increases to a steady value when the sample is re-measured. In the second set, the intensity
73
Fig. 4. The influence of repetitive measurement on the intensity of the main peak and in the inset, the thermal activation profile for synthetic quartz. In the latter, each data point corresponds to a different sample of similar size.
decreases with re-measurement (solid circles). In subsequent measurements (sets 3 through 5), the intensity becomes independent of re-use. The aim of Fig. 4 is to simply show that the thermoluminescence intensity in this material is not necessarily reproducible in successive identical measurements on the same aliquot and this fact has to be taken into account when applying the isothermal cleaning method. That changes in luminescence intensity when the sample is reused (as evident in Fig. 4) may be caused by sensitivity variations were investigated by testing for the pre-dose effect. In this context, it is known that in order to observe sensitization due to irradiation, a sample has to be heated after irradiation. However, in order to discover if there is sensitization at all and if so, to find the temperature at which sensitization is most efficient, it is necessary to construct a thermal activation (TAC) curve by heating different irradiated samples to various temperatures before applying a small test dose to determine the sensitivity [8,9]. The thermal activation curve (TAC) obtained is shown in the inset of Fig. 4. The sensitivity starts to increase from about 300 1C with more efficient activation seen from 350 1C. This behaviour suggests that each time a sample is heated after irradiation, it undergoes sensitization. Studies of the synthetic quartz of this report using positron annihilation [10] showed that annealing up to 1000 1C did not produce new defects or change the concentration of chargetrapping defects. However, annealing did cause a change in lifetimes from a lower value τL to a higher one τH . In a behaviour that is diametric to that in natural quartz [11,12], increasing the annealing temperature enhances recombination in the principal recombination centre LH. On this basis, the concentration of holes in the principal recombination centre LH before a sample has been annealed is negligible and this would explain the lack of response when the sample is heated after irradiation only (Fig. 1, inset (a)). The transfer of holes to the principal recombination centre becomes more efficient when the sample is heated beyond the 300 1C activation temperature as evident by the increase of the resultant signal (Fig. 4, inset (a)). Heating beyond the activation temperature after irradiation increases the amount of thermoluminescence per unit dose due to the accumulation of holes in the principal recombination centre from ionization, as well as from the non-radiative and other recombination centres. It should be understood that the only purpose of Fig. 4 is to show that the intensity from a particular TL peak is not reproducible for measurements made from the same aliquot. Expansive discussions to account for the detailed changes observed, apart
74
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
thermal cleaning will affect the next peak in the series if the peaks are closely collocated as in the example discussed. In the last set of experiments, the position of both the main peak (MP) and its overlap (S1) were found to be independent of the duration of isothermal heating at 40 1C. That the position of the first peak is not affected by the duration of isothermal heating is one of the requirements for successful separation of overlapping peaks in the method of Pagonis and Shannon [2].
3.3. Kinetic analysis following isothermal heating
Fig. 5. Influence of the duration of isothermal heating at 40 1C on the TL intensity of the main peak. The measurements are consecutive (run 1; solid circles) through run 5 (solid inverted triangles). The inset shows measurements on a previously used sample.
from the notes above, are not essential for this discussion and have thus been omitted.
3.2.5. Influence of isothermal heating on peak properties The series of experiments described in this section were carried out to determine the suitable time t to be used for isothermal heating in combination with a temperature T selected from Fig. 3. We first report results of experiments on the influence of isothermal heating on peak intensity. This is then followed by a brief account on the influence of isothermal heating on peak position. The influence of isothermal heating (intended to remove peak S1) on the position and the intensity of the main peak (MP) was studied in a sample held at 40 1C (chosen on the basis of Fig. 3) for various duration. Two sets of measurements were made, one on a fresh sample and the second on a previously used one. The samples were each irradiated to 20 Gy. Fig. 5 shows the effect of duration of isothermal heating at 40 1C on the intensity of the main peak for measurements made consecutively five times on the previously unused sample. In the first measurement (solid circles) the intensity of the main peak was not affected when the sample was held at 40 1C for up to 100 s but decreased when the holding time was increased further. This decrease is evidence of the onset of charge depletion from the main peak. When the measurements were repeated, the profile of the first measurement was not reproduced but rather the intensity was now stable only up to a holding time of 40 s. In further measurements, the intensity of the main peak decreased despite the holding time. This latter behaviour was also observed when similar measurements were carried out on the previously used sample (inset). There are two salient points from Fig. 5. The first is that on the basis of the first set of measurements (solid circles), it would be prudent to select a holding time anywhere between 5 and 100 s. However, the actual combination of time and temperature should be one that leads to removal of the ‘satellite peak’ (S1) with minimal effect on the next peak (MP) in the series. It should be noted that although this discussion is with respect to peaks S1 and MP, the same procedure would be valid if peak MP were to be removed to obtain peak S2. In general, the same principle would apply for a pair of closely collocated glow-peaks. The second point is that Fig. 5, as did Fig. 3, shows that it is inevitable that thermal cleaning whether by isothermal heating or by conventional
Kinetic parameters associated with the main peak were determined after isothermal cleaning at 40 1C for 100 s using the initial rise- and variable heating rate methods. The combination of 40 1C and 100 s was chosen on the basis of experiments described in the previous section. The kinetic parameters found were then compared with those found from the same peak after conventional thermal cleaning as described in Section 3.1. The comparison was done to look for consistency in values and in that way used as a means to assess the effectiveness of isothermal heating as a method for thermal cleaning. The comparison was not done to determine which method is better.
3.3.1. Kinetics study using the initial-rise method: influence of measurement temperature on E The main peak obtained after isothermal cleaning at six different temperatures from 40 up to 65 1C was in each case analysed for the activation energy using the initial-rise method. The values of the activation energy corresponding to the various holding temperatures are shown in Fig. 6. The apparent activation energy decreases as the temperature at which the sample is held before measurement of a glow-curve is increased. Considering this result together with Fig. 3, it seems to be reasonable to conclude that the only valid range of values for E in Fig. 6 is those corresponding to temperatures below 55 1C. Increasing the holding temperature otherwise (Fig. 6) or the holding time much beyond 100 s (Fig. 3) serves to not only remove peak S1 as desired but also leads to the unwanted after-effect of significantly depleting peak MP. Although isothermal cleaning is conceptually straightforward, this result emphasizes the need for a valid combination of time and temperature when isothermal heating is used as a tool for thermal cleaning. The change in E as shown in Fig. 6 may be attributed to significant depletion of charge from the main peak as the holding temperature increases such that the
Fig. 6. Dependence of the apparent activation energy on the temperature of isothermal cleaning.
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
Fig. 7. Procedures involved in separation of peaks S1 and MP. The glow-curve measured following isothermal cleaning (solid circles) is subtracted from the one measured without this step (open circles) to obtain data for peak S1 (inset).
isothermal heating. The difference between the two peaks then reveals the lower-temperature lying peak S1 (inset of Fig. 7). In the method of Pagonis and Shannon [2], the digitally mustered peak is analysed by curve-fitting. This procedure may however be unsuitable for some cases of closely collocated peaks as in this example since mustering will inevitably produce a wellresolved but poor intensity glow-peak as illustrated in the inset of Fig. 7. We emphasize that Fig. 7 is only used to illustrate a problem that can arise, namely poor statistics, when a peak is digitally recovered from some cases of closely collocated peaks as in this example. To address this limitation, we applied a new method of kinetic analysis [3] that uses the temperature-dependence of the area under an isothermal decay-curve. Rather than using a digitally mustered peak, the method instead exploits the phosphorescence emitted during isothermal heating. The method, described in detail elsewhere [3], is not to be confused with the technique of Furetta et al. [13] where the activation energy is calculated from the ratio of areas corresponding to isothermal decay-curves measured at two different temperatures. The method should also be distinguished from that of Mariani et al. [14] which uses a semi-logarithmic plot of inverse lifetime against inverse temperature. The method using the temperaturedependence of the area under an isothermal decay-curve was found to be suitable for closely collocated peaks here and is outlined below. Starting from the fact that the concentration of trapped electrons is proportional to the area under an isothermal decaycurve [13], it can be shown that the segment of area measured between t and t þΔt depends on measurement temperature T as Φ ¼ κ expð E=kTÞ;
Fig. 8. A plot of ln Φ against 1/kT for peak S1 and in the inset, for peak MP.
underlying assumption of the initial-rise method, that the change in trapped charge is negligible, becomes void.
3.3.2. Analysis using the variable heating rate method The variable heating rate method applied on a sample held at 40 1C for 100 s before measurement of a glow-curve at a specific heating rate gave E ¼ 0:77 7 0:02 eV.
3.3.3. Analysis using the temperature-dependence of the area under an isothermal decay curve The main aim of the method of Pagonis and Shannon [2] is to separate overlapping peaks, obtain their complete shape and then evaluate their kinetic parameters using the whole glow-peak procedure and possibly other methods. In their method, the complete shape of the ‘satellite’ peak at a relatively lower temperature was obtained by digitally subtracting isothermally cleaned data from the initial glow-peak. The same procedure was used in this study to digitally muster data for peak S1 except that the isothermally cleaned data were corrected for difference in peak intensity with the original glow-curve. The steps involved are illustrated in Fig. 7. The plot shows the glow-curves measured with and without isothermal heating. The intensity of the TL following isothermal heating was normalized to that before
75
ð1Þ
where κ is a constant [3]. The activation energy E can then readily be found from the slope in a plot of ln Φ against 1/kT. Fig. 8 shows an example of a semi-logarithmic plot of Φ against 1/kT for isothermal decay-curves measured for 10 s between 30 and 50 1C immediately after irradiation from which E ¼0.53 7 0.01 eV. This value compares favourably with E¼ 0.507 0.03 eV from the initial-rise method (Fig. 2). The same procedure applied on isotherms measured after thermal cleaning to 64 1C, that is, for the main peak gave E ¼0.67 70.02 eV (Fig. 8, inset) comparable to the result from the variable heating rate method (Sections 3.1.1 and 3.3.2). The consistency in values shows that when the temperature and the duration of isothermal heating are judiciously chosen, isothermal cleaning in combination with the method based on peak areas of isothermal decay-curves is appropriate to address the separation and analysis of closely collocated peaks. It should be recalled that the combination of heating duration and temperature was determined in the context of preparatory experiments on a material in which the TL intensity is not necessarily reproducible in successive measurements. The relevance of the method based on the use of the temperature-dependence of the area under an isothermal decaycurve is that by sampling only a portion of an isothermal decaycurve, the possibility of the signal being a multiplex of emissions from the collocated peaks can at most be avoided or at least minimized. Concerning useability and applicability of our method, partially overlapping peaks as is common would rather be separated using the normal thermal cleaning method [1] or the method of Pagonis and Shannon [2]. Although an alternative, the method of isothermal cleaning we have developed would be rather ill-suited and not really offer any advantage for this purpose because of the additional preparatory experiments needed to establish a suitable combination of heating time and temperature. The discussion thus far has been about a pair of peaks where the intensity of the first peak is less than that of the second at a
76
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
Fig. 9. A glow-curve of Sr2SiO4:Eu2 þ measured at 1 1C s 1 from a sample irradiated to 10 Gy. The decrease of signal immediately before the rising edge of the main peak is due to decay of phosphorescence. The background is shown for comparison.
higher temperature. We now discuss a case where the opposite is true and another similar example but where the peaks in question appear to merge.
3.3.4. Application of isothermal cleaning to TL peaks in Sr2SiO4:Eu2 þ This section presents the use of isothermal heating to separate TL peaks in europium-doped strontium orthosilicate (Sr2SiO4: Eu2 þ ), a material with properties that make it suitable for use in a range of eclectic applications including as a phosphor in white light emitting diodes [15]. Fig. 9 shows a glow-curve measured from a sample of Sr2SiO4: Eu2 þ at 1 1C s 1 after irradiation to 10 Gy. The glow-curve shows the main peak at 52 1C collocated with an as yet undetermined number of lower intensity components on its higher temperature side. The main peak is affected by phosphorescence on its rising edge. This case, unlike the quartz studied thus far, exemplifies a set of collocated peaks where the intensity of the lower-temperature one exceeds that of the higher temperature ‘satellite(s)’. Here, as with conventional thermal cleaning, the next peak at a higher temperature can be found by removal of the lower-temperature one by isothermal heating. In general, the optimum combination of heating temperature and time should be one that least affects the intensity of the adjacent subsidiary peak and can be determined using systematic experiments similar to the ones described in Section 3.2. In this work, an irradiated sample of Sr2SiO4:Eu2 þ was heated to a temperature corresponding to a position on the rising edge of the main peak and held at this temperature for a certain time. When the complete glow-curve was then re-measured but without further irradiation, the next peak in the series emerged. It was observed that every subsequent peak found after isothermal cleaning showed visual evidence of collocation. Isothermal heating using a suitable combination of heating time and duration was then done on every such new peak until isothermal cleaning produced no more peaks. Fig. 10 shows seven subsidiary peaks recovered after isothermal heating at various temperatures for various times (see inset (a)). The subsidiaries, at 88, 108, 120, 138, 160, 176 and 196 1C, are of successively decreasing intensity such that when measured simultaneously (Fig. 9), their overlap creates what appears to be a collocated component of the main peak on its higher-temperature side. Each of the subsidiary peaks, except the last one, is indeed also collocated with others as can be deduced from Fig. 10.
Fig. 10. The seven subsidiary peaks recovered after isothermal heating at various temperatures for various times overlain with the original glow-curve (inset (a), the dotted line is only a visual guide). The subsidiaries are of successively decreasing intensity but the complete shape of a particular peak can be digitally recovered by subtracting the isothermally cleaned peak from the original glow-curve as illustrated for the main peak only in inset(b).
Fig. 11. A glow-curve measured at 0.4 1C s 1 from α-Al2O3:C after irradiation to 3 Gy.
The complete shape of a particular peak can be digitally mustered by subtracting the isothermally cleaned peak from the original glow-curve. This is illustrated for the main peak only in Fig. 10, inset (b). If necessary, the recovered peaks can then be analysed for kinetic parameters using various methods. For example, the activation energy of the digitally mustered peak shown in the inset (b) was estimated, using the option of the full-width at half-maximum in the peak-shape method [1], as 0.91 eV. The aim of this section was not to carry out a detailed study of the thermoluminescence and kinetic analysis of Sr2SiO4:Eu2 þ but only to demonstrate that isothermal heating can also be applied where the intensity of the first peak exceeds that of the succeeding one. A comprehensive presentation of kinetic parameters would necessitate a discussion of mechanisms of luminescence in Sr2SiO4:Eu2 þ which is outside the scope of this study but is addressed elsewhere e.g. [15,16].
3.3.5. Application of isothermal heating to the main TL peak of α-Al2O3:C Fig. 11 shows a glow-curve measured at 0.4 1C s 1 from α-Al2O3:C after irradiation to 3 Gy. The glow-curve shows a
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
number of secondary peaks (labelled I, III, IV and V) and the dominant peak at 152 1C (labelled II). As is apparent by inspection, peak II is possibly collocated with an as yet undefined component on its higher temperature end. We applied isothermal heating to isolate this higher-temperature collocate. The main peak in α-Al2O3:C has often been deduced to be composed of multiple closely spaced components based on various reasoning including kinetic analysis and mathematical modelling e.g. [17–21]. This peak therefore exemplifies a case where a peak (or peaks as the case may be) is completely subsumed by another to the extent that they appear as one. In a recent study, Chithambo et al. [22] found that when a particular sample irradiated to 3 Gy was partially heated at 0.4 1C s 1 to 200 1C to remove preceding peaks including the main one at 156 1C, reheating from ambient temperature without further irradiation revealed not only peak III at 264 1C but also a previously disguised but weaker intensity peak (labelled IIA) at 170 1C. There was no temperature at which peaks II and IIA appeared simultaneously. The aim of this section is to apply isothermal heating to the main TL peak of α-Al2O3:C in order to try and isolate the component peak IIA. Since in our experience [22], peak IIA could only be removed by preheating an irradiated sample to 200 1C under the experimental conditions mentioned, that is, way above the peak maximum of peak II, it was deduced that isothermal heating to remove peak II and isolate peak IIA also needed to be carried out at a temperature way above the peak maximum of peak II. Preparatory tests showed that isothermal heating at 180 1C for 130 s removed peak II to reveal peak IIA. Fig. 12, a glow-curve (open circles) measured after removal of peak II by isothermal heating shows not only peak III but also the previously disguised peak IIA, at 170 1C. This glowcurve (open circles) has been scaled up 50 times to aid visual clarity. The main peak partially recorded to 209 1C at 0.4 1C s 1 after a 3 Gy dose (in a separate measurement) is only included for comparison. Further investigations were made to establish combinations of heating time and temperature able to remove peak II to reveal peak IIA and the results are shown in Fig. 12 (inset (a)). As evident, the duration of isothermal heating required to isolate peak IIA decreases when the holding temperature is increased. In principle, the optimum combination of heating temperature and time should
77
be one that least affects the intensity of peak IIA and may depend on the dose the sample is exposed to. It should be noted that the effect of thermal quenching is irrelevant in this method here. Analysis of peak IIA for the activation energy using the whole glow-curve method described elsewhere [1] gave the best fit corresponding to an order of kinetics b¼1.1 (Fig. 12, inset (b)) and E ¼0.84 70.01 eV. This can be compared to say 0.85 70.04 eV or 0.92 7 0.08 eV from the initial-rise and variable heating rate methods or indeed to an average value of 0.85 70.13 eV from peak-shape method determined for the same peak following conventional thermal cleaning to 200 1C [22] as explained earlier. The same study [22] showed that peak IIA is subject to first-order kinetics in agreement with Fig. 12 (inset (b)) where b is effectively equal to 1. In contrast, previously reported values of E for peak II include E ¼1.00 70.02 eV [20], 1.33 70.01 eV [23], and 1.48 eV [24]. Thus peaks II and IIA can be qualitatively and quantitatively distinguished. The results presented should not be interpreted to mean that peak II has only two components nor that peak IIA can only emerge under the experimental conditions described here. The study simply shows that a secondary peak (peak IIA) was isolated at 170 1C using isothermal heating as a method of thermal cleaning as described.
4. Summary The need for experimentally separating overlapping TL peaks is understood. The particular case of using isothermal heating for this purpose was discussed by Pagonis and Shannon [2] who developed a method applicable to materials in which the TL intensity is reproducible in successive measurements. They applied their method in combination with curve-fitting to resolve and analyse TL peaks in LiF:Mg,Ti. We have reported a method for use in materials in which the TL intensity is not reproducible in successive measurements on the same aliquot and applied it on a sample of synthetic quartz where the peaks are, coincidentally, closely collocated. Preparatory experiments necessary before application of the method in this case have been described. The experiments are concerned with finding an appropriate combination of temperature and heatingduration. Kinetic analysis after isothermal cleaning as discussed produces results consistent with contemporary methods. In particular, the isothermal separation and the subsequent analysis of closely overlapping peaks are most effective in combination with a Table 1 The activation energy corresponding to dynamic thermal- and isothermal cleaning (data for Fig. 2 denotes average values). In principle, various issues may contribute to a systematic deviation from the true value in the calculated activation energy [1]. For example, for the peak-shape method the bias may be due to imprecision in determining temperatures at half-height; for the initial-rise method, significant depletion of trapped charge, which would violate the basic assumption of the method, would render the outcome void (for example see Section 3.3.1). Sample
Peak
Method
E (eV)
Reference
Var. heat. rate. Initial-rise Initial-rise
0.777 0.02 0.78 7 0.01 0.65 7 0.02 0.50 7 0.03 0.85 7 0.04
Section 3.1.1 Fig. 2 Section 3.1.1 Fig. 2 [21]
Var. heat. rate. Phosp., Area Phosp., Area Peak-shape Whole-glowpeak
0.777 0.02 0.677 0.02 0.53 7 0.01 0.91 0.84 7 0.01
Section 3.3.2 Fig. 8 Fig. 8 Section 3.3.4 Fig. 12
Conventional thermal cleaning Quartz MP Initial-rise
Fig. 12. A glow-curve measured at 0.4 1C s 1 from α-Al2O3:C after irradiation to 3 Gy and isothermal heating at 180 1C for 130 s highlighting the presence of peak IIA. Peak II, partially measured up to 209 1C afterwards after re-irradiation of the sample, is only shown for comparison. The data for the glow-curve after isothermal heating (open circles), including that of peak IIA, has been scaled up 50 times. The inset (a) shows the dependence of heating duration on measurement temperature whereas the inset (b) is the outcome of an application of the whole glow-curve method to determine the activation energy.
α-Al2O3:C
S1 IIA
Isothermal cleaning Quartz MP
SrSiO4:Eu2 þ α-Al2O3:C
S1 Mainpeak IIA
78
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
newly reported technique [3] that uses the temperaturedependence of the area under an isothermal decay-curve. Isothermal heating was also applied to separate collocated peaks in europium-doped orthosilicate (Sr2SiO4:Eu2 þ ), a luminophor, and in carbon-doped aluminium oxide (α-Al2O3:C), a dosemeter. The luminophor Sr2SiO4:Eu2 þ , unlike the quartz, has a set of peaks where the intensity of the lower-temperature one is greater than subsequent ones at higher temperature. On the other hand, the example investigated in α-Al2O3:C is similar except that the dominant peak subsumes the lower intensity secondary peak to the extent that they appear as a single peak. Values of the activation energy determined using various methods in this work are listed for reference in Table 1. Acknowledgements We gratefully acknowledge financial support from the National Research Foundation of South Africa, Rhodes University and the Government of Rwanda. It is also a pleasure to thank Odireleng M. Ntwaeaborwa of the University of the Free State for the sample of Sr2SiO4:Eu2 þ . References [1] S.W.S. McKeever, Thermoluminescence of Solids, Cambridge University Press, Cambridge, 1985. [2] V. Pagonis, C. Shannon, Radiat. Meas. 32 (2000) 805. [3] M.L. Chithambo, J. Lumin. 151 (2014) 235.
[4] M.L. Chithambo, S. Afr. J. Sci. 100 (2004) 524. [5] M.L. Chithambo, S.G. Raymond, T. Calderon, P.D. Townsend, J. Afr. Earth Sci. 20 (1995) 53. [6] P.D. Townsend, D.R. White, Radiat. Prot. Dosim. 65 (1996) 83. [7] V. Pagonis, L. Blohm, M. Brengle, G. Mayonado, P. Woglam, Radiat. Meas. 51–52 (2013) 40. [8] X.H. Yang, S.W.S. McKeever, Radiat. Prot. Dosim. 33 (1990) 27. [9] R. Chen, P.L. Lueng, Radiat. Prot. Dosim. 84 (1999) 43. [10] M.L. Chithambo, P. Sane, F. Tuomisto, Radiat. Meas. 46 (2011) 310. [11] M.L. Ogundare, F.O. Feathers, J. Radiat. Meas. 43 (2008) 1. [12] V. Pagonis, M.L. Chithambo, R. Chen, A. Chruścińska, M. Fasoli, S.H. Li, M. Martini, K. Ramseyer, J. Lumin. 145 (2014) 38. [13] C. Furetta, J. Marcazzó, M. Santiago, E. Caselli, Radiat. Eff. Defect Solids 162 (2007) 385. [14] D.F. Mariani, G. Gonzales, J. Vignolo, M. Jimenez de Castro, A. Ibarra, Phys. Status Solidi (a) 112 (1989) K53. [15] J.S. Kim, J.Y. Kang, P.E. Jeon, J.C. Choi, H.L. Park, T.W. Kim, Jpn. J. Appl. Phys. 43 (2004) 989. [16] M.A. Tshabalala1, H.C. Swart, O.M. Ntwaeaborwa, J. Vac. Sci. Technol., http://dx. doi.org/10.1116/1.4862752. [17] S.W.S. McKeever, M. Moscovitch, P.D. Townsend, Thermoluminescence Dosimetry Materials: Properties and Uses, Nuclear Technology Publishing, Kent, 1995. [18] A.N. Yazici, S. Solak, Z. Ozturk, M. Topaksu, Z. Yegingil, J. Phys. D: Appl. Phys 36 (2003) 181. [19] M. Zahedifar, L. Eshraghi, E. Sadeghi, Radiat. Meas. 47 (2012) 957. [20] F.O. Ogundare, S.A. Ogundele, M.L. Chithambo, M.K. Fasasi, J. Lumin. 139 (2013) 143. [21] G.I. Dallas, D. Afouxenidis, E.C. Stefanaki, N.F. Tsagas, G.S. Polymeris, N.C. Tsirliganis, G. Kitis, Phys. Status Solidi (a) 205 (2008) 1672. [22] M.L. Chithambo, C. Seneza, F.O. Ogundare, Radiat. Meas. 66 (2014) 21. [23] A.N. Nyirenda, Mechanisms of Luminescence in α-Al2O3:C: Investigations using Time-resolved Optical Stimulation and Thermoluminescence Techniques (unpublished M.Sc. thesis), 2012. [24] G. Kitis, Phys. Status Solidi (a) 191 (2002) 621.
Contents lists available at ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
On isothermal heating as a method of separating closely collocated thermoluminescence peaks for kinetic analysis M.L. Chithambo n, P. Niyonzima Department of Physics and Electronics, Rhodes University, PO Box 94, Grahamstown 6140, South Africa
art ic l e i nf o
a b s t r a c t
Article history: Received 9 September 2013 Received in revised form 11 June 2014 Accepted 12 June 2014 Available online 20 June 2014
The experimental separation of overlapping peaks is a common practical concern in thermoluminescence experiments. In this regard, we have investigated the use of isothermal heating as a method for separating closely collocated peaks for kinetic analysis. The method has been developed and applied on a sample of synthetic quartz in which the thermoluminescence intensity is not reproducible in successive measurements on the same aliquot. Preparatory experiments performed to establish a suitable combination of temperature and heating time are described in detail. A comparative analysis of results from conventional thermal cleaning with those from isothermal cleaning, done as a means to assess the effectiveness of the latter, has shown isothermal heating to be a reliable method of separating closely collocated glow-peaks. The method is conveniently applied in combination with a new technique for kinetic analysis based on the use of the temperature-dependence of the area under an isothermal decaycurve. Isothermal heating was also applied to separate collocated peaks in europium-doped orthosilicate (Sr2SiO4:Eu2 þ ), a luminophor, and in carbon-doped aluminium oxide (α-Al2O3:C), a dosemeter. The set studied in Sr2SiO4:Eu2 þ , unlike in the quartz, consists of peaks where the intensity of the lowertemperature one is greater than subsequent ones at higher temperature. On the other hand, the example investigated in α-Al2O3:C is similar except that the dominant peak subsumes the lower intensity secondary peak to the extent that they appear as one. & 2014 Elsevier B.V. All rights reserved.
Keywords: Thermoluminescence Peak separation Phosphorescence Quartz Sr2SiO4:Eu2 þ α-Al2O3:C
1. Introduction Thermoluminescence (TL) is a well-documented method of measuring low light levels from a previously irradiated semiconductor or insulator under a controlled heating rate [1]. The temperature-dependence of the emission appears as a set of peaks collectively termed as a glow-curve with each peak corresponding to an electron-trapping defect. TL glow-curves in some cases consist of well-defined peaks, and in others, of complex sets of overlapping ones as can be appreciated by a cursory examination of any of the many TL papers in the literature. Characteristic features of the peaks may be understood in terms of kinetic parameters, for example the activation energy, defining the electron traps. These parameters aid the description of the mechanisms involved in luminescence emission in the material concerned [1]. However, analysis of a peak for kinetic parameters can usually be conveniently done if the peak is isolated. As such, the experimental separation of overlapping TL peaks is a common practical problem and is the concern of this report.
n
Corresponding author. Tel.: þ27 46 603 8450; fax: þ 27 46 603 8757. E-mail address: [email protected] (M.L. Chithambo).
http://dx.doi.org/10.1016/j.jlumin.2014.06.026 0022-2313/& 2014 Elsevier B.V. All rights reserved.
The conventional method of separating overlapping peaks involves heating to just beyond the maximum of the first peak in order to deplete charge responsible for the ‘satellite’ thereby leaving the next peak with a clean rising edge. The main requirement in this so-called thermal cleaning method is that the partial heating should substantially remove charge responsible for the lower temperature peak. It seems to be reasonable then to expect that any other procedure that also selectively removes charge from a particular peak should affect thermal cleaning. In view of this, this report is concerned with static or isothermal rather than dynamic heating as a method of thermal cleaning. Isothermal heating as a procedure for separating overlapping peaks was discussed by Pagonis and Shannon [2] who applied it on LiF:Mg,Ti. The method is based on the fact that when an irradiated sample is held at a constant temperature, some charge is lost from its electron traps. For this reason, isothermal heating should achieve the same purpose as the conventional thermal cleaning method. Pagonis and Shannon [2] observed the method to be of particular use in the case of first-order kinetics and added that it can be better exploited in combination with curve-fitting. One of the key requirements in the method of Pagonis and Shannon [2] is that the luminescence intensity from the same aliquot should be reproducible. To our knowledge, the technique has not been reported for materials in which this is not the case.
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
This report is concerned with isothermal heating as a method of thermal cleaning and has been developed and applied on synthetic quartz in which the sensitivity changes due to the predose effect. The quartz used here therefore exemplifies a material in which the TL intensity is not reproducible in successive measurements. Coincidentally, the glow-peaks on which the method was applied are closely collocated unlike in LiF (the test sample of Pagonis and Shannon [2]) where the peaks used are somewhat well separated. Preparatory experiments necessary before application of the isothermal cleaning method in this case have been described in detail. The effectiveness of the method was assessed by comparing kinetic parameters found from an isothermally cleaned peak with those determined from a peak thermally cleaned using the conventional thermal cleaning procedure. The method is most effective when used together with a new method of kinetic analysis [3] that uses the temperature-dependence of the area under an isothermal decay-curve. We emphasize here that results from isothermal and conventional thermal cleaning methods were compared only to provide some evidence of reliability of the isothermal cleaning method to be discussed. The comparison must not be misinterpreted as an attempt to find which method is better. The discussion is organized into three parts. The first one (Section 3.1) briefly presents results of kinetic analysis following conventional thermal cleaning. The second part (Section 3.2) describes experiments intended to establish a suitable combination of measurement temperature and heating time for use in isothermal heating. The third part (Section 3.3) then discusses results of kinetic analysis following isothermal heating. Taking advantage of phosphorescence emitted during isothermal cleaning, we applied (see Section 3.3.3) a new method of kinetic analysis [3] based on the temperature-dependence of the area under an isothermal decay curve. The utility of the isothermal cleaning method is further demonstrated when the technique is applied to isolate TL peaks in europium-doped orthosilicate (Sr2SiO4:Eu2 þ ), a luminophor (Section 3.3.4) and to extract a collocate of the main peak in carbon-doped aluminium oxide (α-Al2O3:C), a dosemeter (Section 3.3.5). The set of peaks studied in Sr2SiO4:Eu2 þ , unlike in the quartz, exemplifies a case where the intensity of the first peak at a lower temperature exceeds that of the neighbouring one at the higher temperature. On the other hand, the example studied in αAl2O3:C is similar except that the dominant component subsumes
71
the weaker-intensity one to the extent that the peaks appear as a single peak.
2. Experimental methods The synthetic quartz used in this work was ground into coarse grains (90–500 μm) from a block supplied by Sawyer Research Products (OH, USA). Experiments were conducted using a RISØ TL/ OSL-DA-20 Luminescence Reader. The luminescence was detected by an EMI 9235QB photomultiplier tube through a 7 mm Hoya U340 filter (transmission band 260–390 nm). Samples were irradiated in situ at room temperature using a 90Sr/90Y beta source at a dose rate of 0.10 Gy s 1. The thermoluminescence was measured in a nitrogen atmosphere to prevent spurious signals from air and to improve thermal contact between the sample holder and the heater planchet.
3. Results and discussion The synthetic quartz used in this study produces negligible or no thermoluminescence if it is heated after irradiation only. Thermoluminescence is only properly observed after the sample has been preheated to a temperature above 300 1C before any irradiation. Fig. 1 shows a glow-curve measured at 5 1C s 1 from a sample of synthetic quartz irradiated to 30 Gy. The glow-curve consists of at least 5 peaks (as clarified in the insets) at about 200 and 390 1C as well as the dominant peak near 100 1C which is overlapped at both ends by weaker intensity peaks at approximately 68 and 135 1C respectively. In illustrating the method of isothermal cleaning, this report discusses the kinetic analysis of the main peak and its lower-temperature ‘satellite’ in view of their location as part of the first three closely collocated peaks. The reason for choosing the two peaks studied will be explained later in the text. To aid visual clarity, the glow-curve otherwise measured to 500 1C is shown truncated at 280 1C in Fig. 1. For ease of reference, the main peak is labelled MP and its ‘satellites’ denoted S1 and S2. In this section and elsewhere in the text, collocation refers to the case where a significant component or all of a peak is concealed by another with which it overlaps. Examples of collocated peaks include high-temperature secondary peaks in γirradiated α-Al2O3:C [4], TL emission near 700 nm in Mn or Cr doped beryl [5], TL in X-ray irradiated CaF2:Dy [6] and many others in the literature. The main peak in α-Al2O3:C is a topical example and will be discussed in Section 3.3.5. 3.1. Kinetic analysis using the conventional thermal cleaning method
Fig. 1. A glow-curve measured from synthetic quartz at a heating rate of 5 1C s 1 following irradiation to 30 Gy. The insets show in (a) a glow-curve measured after irradiation but without any prior heating and in (b) the semi-logarithmic plot of intensity against temperature included to better show weaker intensity thermoluminescence peaks beyond 150 1C.
In these series of experiments, the main peak was cleaned of its lower temperature overlapping ‘satellite’ by applying the conventional thermal cleaning method described by McKeever [1]. Wherever necessary, a sample irradiated to 30 Gy was heated to 70 1C at 5 1C s 1 in order to remove peak S1 after which the complete glow-curve was measured by heating to 500 1C at the same rate. It should be emphasized that the only necessary and sufficient condition for using the thermal cleaning method is that the sample ought to be heated to a temperature just beyond the maximum of the peak immediately before the one to be analysed. Kinetic parameters associated with the main peak were determined, after conventional thermal cleaning using the initial-rise and various heating rate methods discussed elsewhere [1]. The values found are subsequently compared with those evaluated after an alternative thermal cleaning method, isothermal heating, to be described later.
72
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
3.1.1. Analysis using the initial-rise and variable heating rate methods The activation energy of the main peak using the initial-rise and variable heating rate methods was found, after conventional thermal cleaning, as 0.77 70.02 eV and 0.65 7 0.02 eV respectively. The value from the variable heating rate method is slightly less than that from the initial-rise method. Of interest is that Pagonis et al. [7] recently showed that the variable heating rate method can underestimate E values whereas the initial-rise method consistently produces correct values of the activation energy. 3.2. Tests preparatory to use of isothermal heating as a method of thermal cleaning The second series of kinetic analyses was carried out on the main peak cleaned of its lower-temperature overlap using isothermal heating. The purpose of this procedure was to achieve the same aim as conventional thermal cleaning, that is, to remove a preceding peak without necessarily affecting the adjacent one. Whereas the main protocol in dynamic thermal cleaning is simply to exceed the maximum temperature of the lower temperature ‘satellite’ peak, the eponymous isothermal cleaning method requires that a sample be held at a temperature on the rising edge of the ‘satellite’ peak for a certain time until its removal reveals a clean rising edge of the next peak. The aim of this section is to describe a set of measurements that were carried out in order to establish a suitable combination of measurement temperature and heating time for use in the isothermal cleaning method. The procedure used in this study is based in outline on the work of Pagonis and Shannon [2] who developed an isothermalheating based protocol for separating overlapping thermoluminescence peaks involving three main steps. The first step is to experimentally determine the number of overlapping peaks. Pagonis and Shannon [2] suggested the usage of either the E T stop or T m T stop method for this purpose. Once the number of peaks and their position are known, the sample is irradiated and heated isothermally at a temperature below the maximum of the first peak. If successful, this step produces the next of the overlapping peaks with a clean rising edge. Lastly, a complete first peak is digitally mustered by subtracting the isothermally cleaned peak from the original glow-curve. This procedure therefore produces two results: a complete first peak and the adjacent one with a clean rising edge. The peaks can then be analysed for their kinetic parameters using suitable methods. The procedure of Pagonis and Shannon [2] is explicit in expecting that the overlapping peaks in question are somewhat well separated or at least appear with distinguishable peak maxima. In applying this method to LiF:Mg,Ti; they found that isothermal heating at an appropriate temperature for a proper duration reduced the intensity of the lower temperature peak but not that of the succeeding one. They stated that the successful application of their method is based on the assumptions that (a) the peaks are sufficiently separated in activation energy E and frequency factor s, (b) the TL intensity is reproducible in successive measurements on the same aliquot, and (c) the position of the first peak is independent of the duration and temperature of isothermal heating. They concluded that their method is most appropriate for first-order TL peaks. The measurements reported here were made and developed on synthetic quartz where the overlapping peaks in question are not well separated but are, incidentally, closely collocated. The sample is also affected by the pre-dose effect and thus the TL intensity is not reproducible in successive measurements on the same aliquot. This means that our material exemplifies samples that do not
satisfy one of the criteria set out by Pagonis and Shannon [2], namely that the TL intensity ought to be reproducible in successive measurements on the same aliquot. As such, their procedures, which were devised from practical considerations on LiF:Mg,Ti, needed to be adapted for use in the quartz whose TL intensity, unlike in LiF: Mg,Ti, was not reproducible when re-measured. This section therefore describes a gamut of tests that were made preparatory to using isothermal heating as a method of thermal cleaning. The measurements discussed may be of interest to others planning to use isothermal heating for thermal cleaning in materials that do not satisfy the criteria of Pagonis and Shannon [2]. We describe experiments on how to choose the temperature for and duration of isothermal heating. We examine the influence of temperature, time and their combination on overlapping TL peaks. We consider the stability of peak positions and the influence of sample re-use, inevitable in TL procedures, on TL features of the peaks. The method has been applied on the closely collocated first three TL peaks apparent in the glow-curve of synthetic quartz as evident in Fig. 1. The various tests are described with respect to the main peak, the only well-defined of the three peaks. It should be noted that since MP is the peak to be analysed, the isothermal heating was done on peak S1. On the other hand, if peak S2 were the peak to be analysed, then peaks S1 and MP would have had to have been removed in this way. Thus the principle is the same for conventional thermal cleaning, namely that the peak preceding the one of interest is the one that needs to be removed. In the final part of the report on synthetic quartz (Section 3.3.3), we will show how the phosphorescence emitted during isothermal heating can be exploited for kinetic analysis using a new method reported elsewhere [3]. This will be discussed with respect to peak S1 but for completeness, results for peak MP will also be presented. The reason for using this new method rather than the whole glow-peak technique as in the method of Pagonis and Shannon [2] will then be explained. In the quartz discussed, the intensity of the first peak (S1) is less than that of the succeeding one (MP). Examples where this is not the case will also be looked at. 3.2.1. Establishing the number of component peaks using the T m –T stop method The glow-curve measured from a sample of synthetic quartz at 5 1C s 1 after irradiation to 30 Gy was shown in Fig. 1. The number of peaks overlapping the main one (peak MP) on its rising edge was confirmed to be only one at 68.8 71.1 1C and to be of firstorder based on the fact that the position of this peak did not shift by more than 2 1C with dose or with pre-heating from 30 to 78 1C. Similarly, the position of peak MP was stable at 101.1 7 1.2 1C. The corresponding values of the activation energy, calculated using the initial-rise method, are shown in an E–T stop graph in Fig. 2. There is a step-like advance from the set E¼ 0.50 7 0.03 eV corresponding to the first peak (S1) to E¼ 0.78 70.01 eV for peak MP. 3.2.2. Influence of Tstop on intensity of the main peak The influence of change in Tstop on the luminescence intensity of the main peak is shown in Fig. 3. The intensity is essentially independent of Tstop up to 56 1C beyond which it decreases: less rapidly up to 70 1C and with a more pronounced reduction thereafter. Two points immediately become apparent from the nature of Fig. 3. The first is that in the sample studied here, partial heating to a temperature Tstop is de facto thermal cleaning when T stop 4 68 1C, the position of peak S1. Thus conventional thermal cleaning affects
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
Fig. 2. The activation energy corresponding to a temperature a sample irradiated at 30 Gy was partially heated to at 5 1C s 1.
Fig. 3. Dependence of the luminescence intensity of the main peak on Tstop.
the peak of interest (peak MP in this study). The second fact is that if a temperature for isothermal heating is to cause a decrease in the intensity of the ‘satellite’ peak but not that of the peak of interest (i.e. to decrease intensity of peak S1 rather than that of peak MP), the heating temperature needs to be selected in the ‘plateau’ region say up to 56 1C in this example. 3.2.3. Reproducibility of the position of the main peak The position of the main peak was measured 50 times consecutively at 5 1C s 1 and determined to be stable at 100 1C. This feature is particularly important because isothermal cleaning involves several cycles of measurement on the same sample and the unstated assumption in such tests is that the peak position is constant. 3.2.4. Influence of repetitive measurement on intensity of the main peak The influence of repetitive measurement on the intensity of the main peak is shown in Fig. 4. The change in intensity is shown for five sets of measurement in which the same sample was measured up to ten times in each set. The same sample (and not different samples of different mass) was used in the five sets. In the first one, the thermoluminescence intensity increases to a steady value when the sample is re-measured. In the second set, the intensity
73
Fig. 4. The influence of repetitive measurement on the intensity of the main peak and in the inset, the thermal activation profile for synthetic quartz. In the latter, each data point corresponds to a different sample of similar size.
decreases with re-measurement (solid circles). In subsequent measurements (sets 3 through 5), the intensity becomes independent of re-use. The aim of Fig. 4 is to simply show that the thermoluminescence intensity in this material is not necessarily reproducible in successive identical measurements on the same aliquot and this fact has to be taken into account when applying the isothermal cleaning method. That changes in luminescence intensity when the sample is reused (as evident in Fig. 4) may be caused by sensitivity variations were investigated by testing for the pre-dose effect. In this context, it is known that in order to observe sensitization due to irradiation, a sample has to be heated after irradiation. However, in order to discover if there is sensitization at all and if so, to find the temperature at which sensitization is most efficient, it is necessary to construct a thermal activation (TAC) curve by heating different irradiated samples to various temperatures before applying a small test dose to determine the sensitivity [8,9]. The thermal activation curve (TAC) obtained is shown in the inset of Fig. 4. The sensitivity starts to increase from about 300 1C with more efficient activation seen from 350 1C. This behaviour suggests that each time a sample is heated after irradiation, it undergoes sensitization. Studies of the synthetic quartz of this report using positron annihilation [10] showed that annealing up to 1000 1C did not produce new defects or change the concentration of chargetrapping defects. However, annealing did cause a change in lifetimes from a lower value τL to a higher one τH . In a behaviour that is diametric to that in natural quartz [11,12], increasing the annealing temperature enhances recombination in the principal recombination centre LH. On this basis, the concentration of holes in the principal recombination centre LH before a sample has been annealed is negligible and this would explain the lack of response when the sample is heated after irradiation only (Fig. 1, inset (a)). The transfer of holes to the principal recombination centre becomes more efficient when the sample is heated beyond the 300 1C activation temperature as evident by the increase of the resultant signal (Fig. 4, inset (a)). Heating beyond the activation temperature after irradiation increases the amount of thermoluminescence per unit dose due to the accumulation of holes in the principal recombination centre from ionization, as well as from the non-radiative and other recombination centres. It should be understood that the only purpose of Fig. 4 is to show that the intensity from a particular TL peak is not reproducible for measurements made from the same aliquot. Expansive discussions to account for the detailed changes observed, apart
74
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
thermal cleaning will affect the next peak in the series if the peaks are closely collocated as in the example discussed. In the last set of experiments, the position of both the main peak (MP) and its overlap (S1) were found to be independent of the duration of isothermal heating at 40 1C. That the position of the first peak is not affected by the duration of isothermal heating is one of the requirements for successful separation of overlapping peaks in the method of Pagonis and Shannon [2].
3.3. Kinetic analysis following isothermal heating
Fig. 5. Influence of the duration of isothermal heating at 40 1C on the TL intensity of the main peak. The measurements are consecutive (run 1; solid circles) through run 5 (solid inverted triangles). The inset shows measurements on a previously used sample.
from the notes above, are not essential for this discussion and have thus been omitted.
3.2.5. Influence of isothermal heating on peak properties The series of experiments described in this section were carried out to determine the suitable time t to be used for isothermal heating in combination with a temperature T selected from Fig. 3. We first report results of experiments on the influence of isothermal heating on peak intensity. This is then followed by a brief account on the influence of isothermal heating on peak position. The influence of isothermal heating (intended to remove peak S1) on the position and the intensity of the main peak (MP) was studied in a sample held at 40 1C (chosen on the basis of Fig. 3) for various duration. Two sets of measurements were made, one on a fresh sample and the second on a previously used one. The samples were each irradiated to 20 Gy. Fig. 5 shows the effect of duration of isothermal heating at 40 1C on the intensity of the main peak for measurements made consecutively five times on the previously unused sample. In the first measurement (solid circles) the intensity of the main peak was not affected when the sample was held at 40 1C for up to 100 s but decreased when the holding time was increased further. This decrease is evidence of the onset of charge depletion from the main peak. When the measurements were repeated, the profile of the first measurement was not reproduced but rather the intensity was now stable only up to a holding time of 40 s. In further measurements, the intensity of the main peak decreased despite the holding time. This latter behaviour was also observed when similar measurements were carried out on the previously used sample (inset). There are two salient points from Fig. 5. The first is that on the basis of the first set of measurements (solid circles), it would be prudent to select a holding time anywhere between 5 and 100 s. However, the actual combination of time and temperature should be one that leads to removal of the ‘satellite peak’ (S1) with minimal effect on the next peak (MP) in the series. It should be noted that although this discussion is with respect to peaks S1 and MP, the same procedure would be valid if peak MP were to be removed to obtain peak S2. In general, the same principle would apply for a pair of closely collocated glow-peaks. The second point is that Fig. 5, as did Fig. 3, shows that it is inevitable that thermal cleaning whether by isothermal heating or by conventional
Kinetic parameters associated with the main peak were determined after isothermal cleaning at 40 1C for 100 s using the initial rise- and variable heating rate methods. The combination of 40 1C and 100 s was chosen on the basis of experiments described in the previous section. The kinetic parameters found were then compared with those found from the same peak after conventional thermal cleaning as described in Section 3.1. The comparison was done to look for consistency in values and in that way used as a means to assess the effectiveness of isothermal heating as a method for thermal cleaning. The comparison was not done to determine which method is better.
3.3.1. Kinetics study using the initial-rise method: influence of measurement temperature on E The main peak obtained after isothermal cleaning at six different temperatures from 40 up to 65 1C was in each case analysed for the activation energy using the initial-rise method. The values of the activation energy corresponding to the various holding temperatures are shown in Fig. 6. The apparent activation energy decreases as the temperature at which the sample is held before measurement of a glow-curve is increased. Considering this result together with Fig. 3, it seems to be reasonable to conclude that the only valid range of values for E in Fig. 6 is those corresponding to temperatures below 55 1C. Increasing the holding temperature otherwise (Fig. 6) or the holding time much beyond 100 s (Fig. 3) serves to not only remove peak S1 as desired but also leads to the unwanted after-effect of significantly depleting peak MP. Although isothermal cleaning is conceptually straightforward, this result emphasizes the need for a valid combination of time and temperature when isothermal heating is used as a tool for thermal cleaning. The change in E as shown in Fig. 6 may be attributed to significant depletion of charge from the main peak as the holding temperature increases such that the
Fig. 6. Dependence of the apparent activation energy on the temperature of isothermal cleaning.
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
Fig. 7. Procedures involved in separation of peaks S1 and MP. The glow-curve measured following isothermal cleaning (solid circles) is subtracted from the one measured without this step (open circles) to obtain data for peak S1 (inset).
isothermal heating. The difference between the two peaks then reveals the lower-temperature lying peak S1 (inset of Fig. 7). In the method of Pagonis and Shannon [2], the digitally mustered peak is analysed by curve-fitting. This procedure may however be unsuitable for some cases of closely collocated peaks as in this example since mustering will inevitably produce a wellresolved but poor intensity glow-peak as illustrated in the inset of Fig. 7. We emphasize that Fig. 7 is only used to illustrate a problem that can arise, namely poor statistics, when a peak is digitally recovered from some cases of closely collocated peaks as in this example. To address this limitation, we applied a new method of kinetic analysis [3] that uses the temperature-dependence of the area under an isothermal decay-curve. Rather than using a digitally mustered peak, the method instead exploits the phosphorescence emitted during isothermal heating. The method, described in detail elsewhere [3], is not to be confused with the technique of Furetta et al. [13] where the activation energy is calculated from the ratio of areas corresponding to isothermal decay-curves measured at two different temperatures. The method should also be distinguished from that of Mariani et al. [14] which uses a semi-logarithmic plot of inverse lifetime against inverse temperature. The method using the temperaturedependence of the area under an isothermal decay-curve was found to be suitable for closely collocated peaks here and is outlined below. Starting from the fact that the concentration of trapped electrons is proportional to the area under an isothermal decaycurve [13], it can be shown that the segment of area measured between t and t þΔt depends on measurement temperature T as Φ ¼ κ expð E=kTÞ;
Fig. 8. A plot of ln Φ against 1/kT for peak S1 and in the inset, for peak MP.
underlying assumption of the initial-rise method, that the change in trapped charge is negligible, becomes void.
3.3.2. Analysis using the variable heating rate method The variable heating rate method applied on a sample held at 40 1C for 100 s before measurement of a glow-curve at a specific heating rate gave E ¼ 0:77 7 0:02 eV.
3.3.3. Analysis using the temperature-dependence of the area under an isothermal decay curve The main aim of the method of Pagonis and Shannon [2] is to separate overlapping peaks, obtain their complete shape and then evaluate their kinetic parameters using the whole glow-peak procedure and possibly other methods. In their method, the complete shape of the ‘satellite’ peak at a relatively lower temperature was obtained by digitally subtracting isothermally cleaned data from the initial glow-peak. The same procedure was used in this study to digitally muster data for peak S1 except that the isothermally cleaned data were corrected for difference in peak intensity with the original glow-curve. The steps involved are illustrated in Fig. 7. The plot shows the glow-curves measured with and without isothermal heating. The intensity of the TL following isothermal heating was normalized to that before
75
ð1Þ
where κ is a constant [3]. The activation energy E can then readily be found from the slope in a plot of ln Φ against 1/kT. Fig. 8 shows an example of a semi-logarithmic plot of Φ against 1/kT for isothermal decay-curves measured for 10 s between 30 and 50 1C immediately after irradiation from which E ¼0.53 7 0.01 eV. This value compares favourably with E¼ 0.507 0.03 eV from the initial-rise method (Fig. 2). The same procedure applied on isotherms measured after thermal cleaning to 64 1C, that is, for the main peak gave E ¼0.67 70.02 eV (Fig. 8, inset) comparable to the result from the variable heating rate method (Sections 3.1.1 and 3.3.2). The consistency in values shows that when the temperature and the duration of isothermal heating are judiciously chosen, isothermal cleaning in combination with the method based on peak areas of isothermal decay-curves is appropriate to address the separation and analysis of closely collocated peaks. It should be recalled that the combination of heating duration and temperature was determined in the context of preparatory experiments on a material in which the TL intensity is not necessarily reproducible in successive measurements. The relevance of the method based on the use of the temperature-dependence of the area under an isothermal decaycurve is that by sampling only a portion of an isothermal decaycurve, the possibility of the signal being a multiplex of emissions from the collocated peaks can at most be avoided or at least minimized. Concerning useability and applicability of our method, partially overlapping peaks as is common would rather be separated using the normal thermal cleaning method [1] or the method of Pagonis and Shannon [2]. Although an alternative, the method of isothermal cleaning we have developed would be rather ill-suited and not really offer any advantage for this purpose because of the additional preparatory experiments needed to establish a suitable combination of heating time and temperature. The discussion thus far has been about a pair of peaks where the intensity of the first peak is less than that of the second at a
76
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
Fig. 9. A glow-curve of Sr2SiO4:Eu2 þ measured at 1 1C s 1 from a sample irradiated to 10 Gy. The decrease of signal immediately before the rising edge of the main peak is due to decay of phosphorescence. The background is shown for comparison.
higher temperature. We now discuss a case where the opposite is true and another similar example but where the peaks in question appear to merge.
3.3.4. Application of isothermal cleaning to TL peaks in Sr2SiO4:Eu2 þ This section presents the use of isothermal heating to separate TL peaks in europium-doped strontium orthosilicate (Sr2SiO4: Eu2 þ ), a material with properties that make it suitable for use in a range of eclectic applications including as a phosphor in white light emitting diodes [15]. Fig. 9 shows a glow-curve measured from a sample of Sr2SiO4: Eu2 þ at 1 1C s 1 after irradiation to 10 Gy. The glow-curve shows the main peak at 52 1C collocated with an as yet undetermined number of lower intensity components on its higher temperature side. The main peak is affected by phosphorescence on its rising edge. This case, unlike the quartz studied thus far, exemplifies a set of collocated peaks where the intensity of the lower-temperature one exceeds that of the higher temperature ‘satellite(s)’. Here, as with conventional thermal cleaning, the next peak at a higher temperature can be found by removal of the lower-temperature one by isothermal heating. In general, the optimum combination of heating temperature and time should be one that least affects the intensity of the adjacent subsidiary peak and can be determined using systematic experiments similar to the ones described in Section 3.2. In this work, an irradiated sample of Sr2SiO4:Eu2 þ was heated to a temperature corresponding to a position on the rising edge of the main peak and held at this temperature for a certain time. When the complete glow-curve was then re-measured but without further irradiation, the next peak in the series emerged. It was observed that every subsequent peak found after isothermal cleaning showed visual evidence of collocation. Isothermal heating using a suitable combination of heating time and duration was then done on every such new peak until isothermal cleaning produced no more peaks. Fig. 10 shows seven subsidiary peaks recovered after isothermal heating at various temperatures for various times (see inset (a)). The subsidiaries, at 88, 108, 120, 138, 160, 176 and 196 1C, are of successively decreasing intensity such that when measured simultaneously (Fig. 9), their overlap creates what appears to be a collocated component of the main peak on its higher-temperature side. Each of the subsidiary peaks, except the last one, is indeed also collocated with others as can be deduced from Fig. 10.
Fig. 10. The seven subsidiary peaks recovered after isothermal heating at various temperatures for various times overlain with the original glow-curve (inset (a), the dotted line is only a visual guide). The subsidiaries are of successively decreasing intensity but the complete shape of a particular peak can be digitally recovered by subtracting the isothermally cleaned peak from the original glow-curve as illustrated for the main peak only in inset(b).
Fig. 11. A glow-curve measured at 0.4 1C s 1 from α-Al2O3:C after irradiation to 3 Gy.
The complete shape of a particular peak can be digitally mustered by subtracting the isothermally cleaned peak from the original glow-curve. This is illustrated for the main peak only in Fig. 10, inset (b). If necessary, the recovered peaks can then be analysed for kinetic parameters using various methods. For example, the activation energy of the digitally mustered peak shown in the inset (b) was estimated, using the option of the full-width at half-maximum in the peak-shape method [1], as 0.91 eV. The aim of this section was not to carry out a detailed study of the thermoluminescence and kinetic analysis of Sr2SiO4:Eu2 þ but only to demonstrate that isothermal heating can also be applied where the intensity of the first peak exceeds that of the succeeding one. A comprehensive presentation of kinetic parameters would necessitate a discussion of mechanisms of luminescence in Sr2SiO4:Eu2 þ which is outside the scope of this study but is addressed elsewhere e.g. [15,16].
3.3.5. Application of isothermal heating to the main TL peak of α-Al2O3:C Fig. 11 shows a glow-curve measured at 0.4 1C s 1 from α-Al2O3:C after irradiation to 3 Gy. The glow-curve shows a
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
number of secondary peaks (labelled I, III, IV and V) and the dominant peak at 152 1C (labelled II). As is apparent by inspection, peak II is possibly collocated with an as yet undefined component on its higher temperature end. We applied isothermal heating to isolate this higher-temperature collocate. The main peak in α-Al2O3:C has often been deduced to be composed of multiple closely spaced components based on various reasoning including kinetic analysis and mathematical modelling e.g. [17–21]. This peak therefore exemplifies a case where a peak (or peaks as the case may be) is completely subsumed by another to the extent that they appear as one. In a recent study, Chithambo et al. [22] found that when a particular sample irradiated to 3 Gy was partially heated at 0.4 1C s 1 to 200 1C to remove preceding peaks including the main one at 156 1C, reheating from ambient temperature without further irradiation revealed not only peak III at 264 1C but also a previously disguised but weaker intensity peak (labelled IIA) at 170 1C. There was no temperature at which peaks II and IIA appeared simultaneously. The aim of this section is to apply isothermal heating to the main TL peak of α-Al2O3:C in order to try and isolate the component peak IIA. Since in our experience [22], peak IIA could only be removed by preheating an irradiated sample to 200 1C under the experimental conditions mentioned, that is, way above the peak maximum of peak II, it was deduced that isothermal heating to remove peak II and isolate peak IIA also needed to be carried out at a temperature way above the peak maximum of peak II. Preparatory tests showed that isothermal heating at 180 1C for 130 s removed peak II to reveal peak IIA. Fig. 12, a glow-curve (open circles) measured after removal of peak II by isothermal heating shows not only peak III but also the previously disguised peak IIA, at 170 1C. This glowcurve (open circles) has been scaled up 50 times to aid visual clarity. The main peak partially recorded to 209 1C at 0.4 1C s 1 after a 3 Gy dose (in a separate measurement) is only included for comparison. Further investigations were made to establish combinations of heating time and temperature able to remove peak II to reveal peak IIA and the results are shown in Fig. 12 (inset (a)). As evident, the duration of isothermal heating required to isolate peak IIA decreases when the holding temperature is increased. In principle, the optimum combination of heating temperature and time should
77
be one that least affects the intensity of peak IIA and may depend on the dose the sample is exposed to. It should be noted that the effect of thermal quenching is irrelevant in this method here. Analysis of peak IIA for the activation energy using the whole glow-curve method described elsewhere [1] gave the best fit corresponding to an order of kinetics b¼1.1 (Fig. 12, inset (b)) and E ¼0.84 70.01 eV. This can be compared to say 0.85 70.04 eV or 0.92 7 0.08 eV from the initial-rise and variable heating rate methods or indeed to an average value of 0.85 70.13 eV from peak-shape method determined for the same peak following conventional thermal cleaning to 200 1C [22] as explained earlier. The same study [22] showed that peak IIA is subject to first-order kinetics in agreement with Fig. 12 (inset (b)) where b is effectively equal to 1. In contrast, previously reported values of E for peak II include E ¼1.00 70.02 eV [20], 1.33 70.01 eV [23], and 1.48 eV [24]. Thus peaks II and IIA can be qualitatively and quantitatively distinguished. The results presented should not be interpreted to mean that peak II has only two components nor that peak IIA can only emerge under the experimental conditions described here. The study simply shows that a secondary peak (peak IIA) was isolated at 170 1C using isothermal heating as a method of thermal cleaning as described.
4. Summary The need for experimentally separating overlapping TL peaks is understood. The particular case of using isothermal heating for this purpose was discussed by Pagonis and Shannon [2] who developed a method applicable to materials in which the TL intensity is reproducible in successive measurements. They applied their method in combination with curve-fitting to resolve and analyse TL peaks in LiF:Mg,Ti. We have reported a method for use in materials in which the TL intensity is not reproducible in successive measurements on the same aliquot and applied it on a sample of synthetic quartz where the peaks are, coincidentally, closely collocated. Preparatory experiments necessary before application of the method in this case have been described. The experiments are concerned with finding an appropriate combination of temperature and heatingduration. Kinetic analysis after isothermal cleaning as discussed produces results consistent with contemporary methods. In particular, the isothermal separation and the subsequent analysis of closely overlapping peaks are most effective in combination with a Table 1 The activation energy corresponding to dynamic thermal- and isothermal cleaning (data for Fig. 2 denotes average values). In principle, various issues may contribute to a systematic deviation from the true value in the calculated activation energy [1]. For example, for the peak-shape method the bias may be due to imprecision in determining temperatures at half-height; for the initial-rise method, significant depletion of trapped charge, which would violate the basic assumption of the method, would render the outcome void (for example see Section 3.3.1). Sample
Peak
Method
E (eV)
Reference
Var. heat. rate. Initial-rise Initial-rise
0.777 0.02 0.78 7 0.01 0.65 7 0.02 0.50 7 0.03 0.85 7 0.04
Section 3.1.1 Fig. 2 Section 3.1.1 Fig. 2 [21]
Var. heat. rate. Phosp., Area Phosp., Area Peak-shape Whole-glowpeak
0.777 0.02 0.677 0.02 0.53 7 0.01 0.91 0.84 7 0.01
Section 3.3.2 Fig. 8 Fig. 8 Section 3.3.4 Fig. 12
Conventional thermal cleaning Quartz MP Initial-rise
Fig. 12. A glow-curve measured at 0.4 1C s 1 from α-Al2O3:C after irradiation to 3 Gy and isothermal heating at 180 1C for 130 s highlighting the presence of peak IIA. Peak II, partially measured up to 209 1C afterwards after re-irradiation of the sample, is only shown for comparison. The data for the glow-curve after isothermal heating (open circles), including that of peak IIA, has been scaled up 50 times. The inset (a) shows the dependence of heating duration on measurement temperature whereas the inset (b) is the outcome of an application of the whole glow-curve method to determine the activation energy.
α-Al2O3:C
S1 IIA
Isothermal cleaning Quartz MP
SrSiO4:Eu2 þ α-Al2O3:C
S1 Mainpeak IIA
78
M.L. Chithambo, P. Niyonzima / Journal of Luminescence 155 (2014) 70–78
newly reported technique [3] that uses the temperaturedependence of the area under an isothermal decay-curve. Isothermal heating was also applied to separate collocated peaks in europium-doped orthosilicate (Sr2SiO4:Eu2 þ ), a luminophor, and in carbon-doped aluminium oxide (α-Al2O3:C), a dosemeter. The luminophor Sr2SiO4:Eu2 þ , unlike the quartz, has a set of peaks where the intensity of the lower-temperature one is greater than subsequent ones at higher temperature. On the other hand, the example investigated in α-Al2O3:C is similar except that the dominant peak subsumes the lower intensity secondary peak to the extent that they appear as a single peak. Values of the activation energy determined using various methods in this work are listed for reference in Table 1. Acknowledgements We gratefully acknowledge financial support from the National Research Foundation of South Africa, Rhodes University and the Government of Rwanda. It is also a pleasure to thank Odireleng M. Ntwaeaborwa of the University of the Free State for the sample of Sr2SiO4:Eu2 þ . References [1] S.W.S. McKeever, Thermoluminescence of Solids, Cambridge University Press, Cambridge, 1985. [2] V. Pagonis, C. Shannon, Radiat. Meas. 32 (2000) 805. [3] M.L. Chithambo, J. Lumin. 151 (2014) 235.
[4] M.L. Chithambo, S. Afr. J. Sci. 100 (2004) 524. [5] M.L. Chithambo, S.G. Raymond, T. Calderon, P.D. Townsend, J. Afr. Earth Sci. 20 (1995) 53. [6] P.D. Townsend, D.R. White, Radiat. Prot. Dosim. 65 (1996) 83. [7] V. Pagonis, L. Blohm, M. Brengle, G. Mayonado, P. Woglam, Radiat. Meas. 51–52 (2013) 40. [8] X.H. Yang, S.W.S. McKeever, Radiat. Prot. Dosim. 33 (1990) 27. [9] R. Chen, P.L. Lueng, Radiat. Prot. Dosim. 84 (1999) 43. [10] M.L. Chithambo, P. Sane, F. Tuomisto, Radiat. Meas. 46 (2011) 310. [11] M.L. Ogundare, F.O. Feathers, J. Radiat. Meas. 43 (2008) 1. [12] V. Pagonis, M.L. Chithambo, R. Chen, A. Chruścińska, M. Fasoli, S.H. Li, M. Martini, K. Ramseyer, J. Lumin. 145 (2014) 38. [13] C. Furetta, J. Marcazzó, M. Santiago, E. Caselli, Radiat. Eff. Defect Solids 162 (2007) 385. [14] D.F. Mariani, G. Gonzales, J. Vignolo, M. Jimenez de Castro, A. Ibarra, Phys. Status Solidi (a) 112 (1989) K53. [15] J.S. Kim, J.Y. Kang, P.E. Jeon, J.C. Choi, H.L. Park, T.W. Kim, Jpn. J. Appl. Phys. 43 (2004) 989. [16] M.A. Tshabalala1, H.C. Swart, O.M. Ntwaeaborwa, J. Vac. Sci. Technol., http://dx. doi.org/10.1116/1.4862752. [17] S.W.S. McKeever, M. Moscovitch, P.D. Townsend, Thermoluminescence Dosimetry Materials: Properties and Uses, Nuclear Technology Publishing, Kent, 1995. [18] A.N. Yazici, S. Solak, Z. Ozturk, M. Topaksu, Z. Yegingil, J. Phys. D: Appl. Phys 36 (2003) 181. [19] M. Zahedifar, L. Eshraghi, E. Sadeghi, Radiat. Meas. 47 (2012) 957. [20] F.O. Ogundare, S.A. Ogundele, M.L. Chithambo, M.K. Fasasi, J. Lumin. 139 (2013) 143. [21] G.I. Dallas, D. Afouxenidis, E.C. Stefanaki, N.F. Tsagas, G.S. Polymeris, N.C. Tsirliganis, G. Kitis, Phys. Status Solidi (a) 205 (2008) 1672. [22] M.L. Chithambo, C. Seneza, F.O. Ogundare, Radiat. Meas. 66 (2014) 21. [23] A.N. Nyirenda, Mechanisms of Luminescence in α-Al2O3:C: Investigations using Time-resolved Optical Stimulation and Thermoluminescence Techniques (unpublished M.Sc. thesis), 2012. [24] G. Kitis, Phys. Status Solidi (a) 191 (2002) 621.