British Journal of Radiology (2005) 78, 543-547
© 2005 British Institute of Radiology
doi: 10.1259/bjr/73133162
The low energy X-ray response of the LiF:Mg:Cu:P thermoluminescent dosemeter: a comparison with LiF:Mg:Ti
C R Edwards, PhD, MIPEM1,
P J Mountford, PhD, FIPEM1,
S Green, PhD, FIPEM2,
J E Palethorpe, HND, IIPEM2 and
A J Moloney, BSc, MIPEM1
1 Medical Physics Directorate, University Hospital of North Staffordshire, Princes Road, Hartshill, Stoke-on-Trent, Staffordshire ST4 7LN and 2 RRPPS, University Hospital Birmingham, Edgbaston, Birmingham B15 2TH, UK
 |
Abstract
|
|---|
LiF:Mg:Cu:P thermoluminescent dosemeters (TLD) can be used for the same X-ray dosimetry applications as LiF:Mg:Ti, with each type having the disadvantage of a response dependent on energy, particularly at low energies. Measurements were made of the response per unit air kerma of LiF:Mg:Cu:P and LiF:Mg:Ti to nine quasi-monoenergetic X-ray beams with mean energies from 12 keV to 208 keV. Each measurement was normalized to the value produced by 6 MV X-rays. LiF:Mg:Cu:P was found to under-respond to a majority of these radiations whereas LiF:Mg:Ti over-responded to a majority. Their smallest relative measured response was produced by the lowest energy beam, and the maximum measured relative response of 1.15±0.07 and 1.21±0.07 for LiF:Mg:Cu:P and LiF:Mg:Ti, respectively, occurred at 33 keV. Energy response coefficients were derived from these measurements to estimate the error introduced by using either type of TLD to measure the dose from an X-ray spectrum different to that used for its absolute response calibration. It was calculated that if the response of either type of TLD was calibrated at 100 kVp, then an error of no more than ±2% would be introduced into measurements of tube output at potentials of 50–130 kVp. LiF:Mg:Cu:P was found to introduce a larger error (up to 30%) into the measurement of body exit dose than LiF:Mg:Ti at tube potentials of 40–150 kVp, if its absolute response was calibrated using the corresponding body entrance beam. The method should allow this type of error to be estimated in other dosimetry applications for either type of TLD.
|
Introduction
|
|---|
The LiF:Mg:Cu:P thermoluminescent dosemeter (TLD) can be used as an alternative to LiF:Mg:Ti for several diagnostic radiology and radiotherapy dosimetry applications. Both types of TLD have the disadvantage of a non-uniform response to low energy X-rays [1, 2] which may be significant in applications where the energy spectrum differs from that used to calibrate the absolute response of the TLD. However, there are limitations to the methods that have been used to investigate the low X-ray energy response of LiF:Mg:Cu:P such as a limited range of energies (10.0–99.6 keV) [2, 3], and the use of heteroenergetic sources [1] which will yield a less representative variation of the response than that produced by monoenergetic sources [4].
The purpose of this study was to measure the relative response of LiF:Mg:Cu:P to low energy quasi-monoenergetic X-ray sources over a wider range of energies than had been derived already using monoenergetic sources [2], and to compare this response with that of LiF:Mg:Ti. Although the measurement technique had already been used to derive the low X-ray energy response of LiF:Mg:Ti TLDs [5], these measurements were repeated in order to provide a direct comparison with the LiF:Mg:Cu:P measurements. The results were then used to derive energy response coefficients which allowed a comparison to be made between the potential errors introduced by each type of TLD when used for two dosimetry applications where the X-ray energy spectrum differed from that used to calibrate the absolute TLD response.
|
Material and methods
|
|---|
Energy response measurements
Fifty LiF:Mg:Cu:P chips (0.45 cm x 0.45 cm x 0.06 cm) and 50 LiF:Mg:Ti TLDs cylindrical rods (0.1 cm diameter, 0.5 cm length) were annealed five times by the procedure recommended by the manufacturer (QADOS Ltd., Camberley, UK) to reduce the risk of sensitivity variations. Each chip and rod was then positioned in a Perspex phantom at the depth of maximum dose produced by a Varian 2100C linear accelerator (Varian Oncology Systems, CA, USA) with a 6 MV 10 cm x 10 cm field. A dose of 100 cGy (absorbed dose to water) was delivered to the TLD and its thermoluminescent response was recorded. This process was repeated three times before 30 TLDs of each type, whose response did not vary by more than ±2.5%, were selected for the energy response measurements. Each TLD was then encased individually in black polyvinyl chloride.
Three TLDs of each type were attached to a 2 cm thick Styrofoam block to minimize backscatter (attenuation <0.5%) [5] and were irradiated in the central 0.003 m2 of a 0.025 m2 circular field by one of the quasi-monoenergetic radiations generated by a filtered Pantak HF-320 X-ray unit (Pantak, Inc., CT, USA) with a stabilized output (Table 1
). More detailed information regarding the characteristic of this X-ray unit has been reported by Green et al [6]. The dose received by each TLD ranged from 24 cGy to 58 cGy depending upon the dose rate at each energy and the irradiation time available. The 6 MV calibration reference was converted from absorbed dose-to-water to air kerma following the method outlined in an earlier study by Edwards et al [5]. The response per unit air kerma of a TLD at each X-ray energy relative to that produced by 6 MV X-rays was derived from the mean value of the three measured responses for each TLD type using the method described in the previous study [5].
View this table:
[in this window]
[in a new window]
|
Table 1. Characteristics of the quasi-monoenergetic X-ray sources [6] and the relative response per unit air kerma to quasi-monoenergetic X-rays of LiF:Mg:Cu:P and LiF:Mg:Ti thermoluminescent dosemeters (TLD) normalized to that produced by 6 MV X-rays
|
|
Energy response coefficients
Energy response coefficients were derived to estimate the potential error in the two following diagnostic X-ray dosimetry applications: first, for dose output measurements over the typical range of beam qualities produced by a diagnostic X-ray unit (Table 2), but where the absolute response of the TLD had been calibrated just at one beam quality; and second, for body exit dose measurements (e.g. in high dose fluoroscopy examinations) where the TLD absolute response had been calibrated using the corresponding entrance beam. The energy response coefficient 
d for each TLD type d was calculated from Equation 1 [7, 8]:
where rd(Ei) was the variation with X-ray energy Ei of the relative response per unit air kerma for each TLD type d, obtained by fitting a curve to the results of the above measurements using the method of cubic splines. The variation with X-ray energy Ei of the fluence-to-air kerma conversion factor 
(Ei) was obtained from the electronic version of IPEM Report 78 [9]. In the first application, 
x(Ei) was the spectral variation of the fluence of X-rays with energy Ei for a beam quality x, and it was assumed that the absolute TLD response had been calibrated with a 100 kVp X-ray spectrum cal(Ei). In the second application, x(Ei) was the body exit X-ray energy spectrum for each of the above range of beam qualities incident on a body thickness consisting of 16 cm of tissue and 2 cm of bone, and it was assumed that for each beam quality, the absolute TLD response had been calibrated with the incident X-ray spectrum cal(Ei). Each X-ray energy spectrum and its mean energy were calculated using the IPEM Spectrum Processor for a tungsten target and a 3 mm thick aluminium filter [9, 10]. As defined by Equation 1
, an energy response coefficient of more or less than unity will be produced by a TLD over or under responding, respectively, to an X-ray spectrum compared with its response to the calibration spectrum.
View this table:
[in this window]
[in a new window]
|
Table 2. Characteristic properties of the X-ray sources and body exit X-ray beams and the energy response coefficients  d calculated for two dosimetry applications
|
|
|
Results
|
|---|
The variation in relative response per unit air kerma of each type of TLD with X-ray energy showed the same overall shape, with the smallest value at the lowest X-ray energy measured of 12 keV (Table 1 and Figure 1). Above this energy, their relative response reached a maximum measured value of 1.15±0.07 and 1.21±0.07 for LiF:Mg:Cu:P and LiF:Mg:Ti, respectively, at an energy of 33 keV. LiF:Mg:Cu:P showed the smallest value of minimum measured relative response (0.39±0.02), and the largest overall change in relative response at energies above its maximum measured value (1.15–0.76).
View larger version (17K):
[in this window]
[in a new window]
|
Figure 1. The relative response per unit air kerma to low energy quasi-monoenergetic X-rays of LiF:Mg:Cu:P and LiF:Mg:Ti thermoluminescent dosemeters normalized to that produced by 6 MV X-rays (error bars are ±1 standard deviation).
|
|
The calculated values of the energy response coefficient are given in Table 2. If the response of both types of TLD was calibrated at 100 kVp, the relative response of LiF:Mg:Ti to X-rays generated at a tube potential of 40 kVp to 150 kVp did not vary by more than ±2%. However, the response of LiF:Mg:Cu:P remained within the range of ±2% over a slightly narrower tube potential range of 50–130 kVp and outside this range, it increased by more than 2% above the calibration value at 100 kVp. The ratio of the response of each type of TLD to a body exit X-ray beam spectrum compared with its calibration response to the entrance beam spectrum decreased continuously with tube potential over the range 40–150 kVp, passing through a value of unity for LiF:Mg:Cu:P and LiF:Mg:Ti at a tube potential of 60 kVp and 70 kVp, respectively. The response of a LiF:Mg:Ti TLD to a body exit beam was 3% above and 6% below its entrance beam calibration response at a tube potential of 40 kVp and 150 kVp, respectively, whereas the response of LiF:Mg:Cu:P to the exit beam was about 30% above and 13% below its entrance beam calibration response at a tube potential of 40 kVp and 150 kVp, respectively.
|
Discussion
|
|---|
The average fractional uncertainty in the relative response per unit air kerma for each LiF:Mg:Ti and LiF:Mg:Cu:P TLD was approximately ±6% (1 standard deviation (SD)). Systematic errors produced the greatest contribution to these overall uncertainties, and these errors arose from calibration of the NE2250 Secondary Standard Ionisation Chamber (±2%) used to determine the air kerma output from the Pantak HF-320 X-ray unit, the reproducibility of the response of the TLDs (±2.5%), and from the variation in the X-ray unit's output owing to fluctuations in beam flatness, current and timing (±2%) [5]. However, variation of these uncertainties with energy will have had little effect on the variation of each detector's response with energy.
The LiF:Mg:Ti TLD produced a peak sensitivity of 1.2 at 30 keV when its stopping power ratio had been compared with water [11–13]. The synchrotron radiation studies by Kron et al [2, 3] found a peak sensitivity of 1.47 at an incident energy of 26.8 keV. In the previous quasi-monoenergetic study of the response of LiF:Mg:Ti, a maximum measured relative sensitivity of 1.31±0.09 was produced for an incident energy of 33 keV [5], which overlapped within one standard deviation of the value of 1.21±0.07 found in this later study at that energy (Figure 1). A paired t-test, used to compare these two sets of results showed that there was no statistically significant difference in the mean response (p>0.5), despite the fact that this type of test does not consider the experimental uncertainties. Since the time of the first study, these TLD rods had been in continual routine use and handling of the rods may have resulted in surface damage. This is one factor that could have affected their thermoluminescent properties.
The study by Kron et al [2] using synchrotron radiation to produce monoenergetic X-rays from 10.0 keV to 99.6 keV suggested a peak sensitivity of 1.05 when LiF:Mg:Cu:P was irradiated at an energy of 24 keV, compared with its response when irradiated by 6 MV photons. Cai [1] found a peak sensitivity of 1.13 for this type of TLD when irradiated by a heteroenergetic source whose mean energy was 39 keV, compared with its response when irradiated by a 60Co source. The maximum measured relative response for LiF:Mg:Cu:P in this study of 1.15±0.07 occurred at 33 keV (Figure 1 and Table 2), which was the source energy nearest to that of the maximum response found by Cai [1]. At energies above the peak sensitivities, both of these previously published studies gave a detector under-response, with a minimum sensitivity of 0.8 at 122 keV found by Cai [1] and approximately 0.75 at 165 keV found by Kron et al [2]. These under-responses were consistent with the minimum measured value found here of 0.76 at 118 keV. Within their statistical uncertainty, the results have demonstrated a uniform response for LiF:Mg:Cu:P above the maximum monoenergetic energy of 99.6 keV used by Kron et al [2].
LiF:Mg:Cu:P had a smaller value of maximum measured relative response to the quasi-monoenergetic radiations than LiF:Mg:Ti, and in general it tended to under-respond over the range 12–208 keV, whereas LiF:Mg:Ti tended to over-respond across this range. However, when either type of TLD is used to measure the dose from an X-ray source with a spectrum different from that used to calibrate the TLD response, the error introduced by these effects will depend on the difference between the shape of the two spectra over this range of X-ray energies, as well as on the variation in the TLD response with energy. The energy response coefficients derived in this study have been used as a simple and convenient method of estimating and comparing these potential errors for each type of TLD in two dosimetry applications for a typical range of beam qualities.
In the first application, the results indicated that for both types of TLD, there was only a small difference between their response to X-rays generated at 50–130 kVp and their response to 100 kVp radiation. Hence there would be little error introduced if the response of either type of TLD was calibrated at 100 kVp and then used for dose output measurements at other values of tube potential within that range. However, LiF:Mg:Ti would produce a smaller error than LiF:Mg:Cu:P at tube potentials above and below this range.
In the second application, the results indicated that the response of LiF:Mg:Ti to the exit beam transmitted by an 18 cm thick body (16 cm tissue, 2 cm bone) would vary from its calibration response to the entrance beam by much smaller differences than that produced by LiF:Mg:Cu:P at every value of tube potential listed in Table 2. Calibration of the response of LiF:Mg:Ti with a body entrance beam over 40–120 kVp was found to introduce an error into the corresponding body exit dose measurement which was similar in magnitude to the variation in the response to repeat irradiations (±2.5%), but less than the average fractional uncertainty in its measured relative response per unit kerma (±6%) to the quasi-monoenergetic radiations. The incident calibration spectrum always had a lower mean energy than the body exit spectrum in the second application (Table 2). Tissue backscatter was omitted from the calculation of the incident spectrum by the IPEM Spectrum Processor [10], but the effect of its inclusion would have been to reduce the mean energy of the incident spectrum and therefore to increase further the difference between the mean energy of the incident spectrum and the body exit spectrum. The actual effect of tissue backscatter on the energy response coefficient could be explored by a Monte Carlo calculation, and included in a study of the differences between the TLD response at a depth in a tissue equivalent phantom and the response to the incident calibration spectrum. In this case, its effect would be to reduce the mean spectral energy at both positions.
Finally, the results of this study suggest that a similar comparison should be made between the response of LiF:Mg:Cu:P and LiF:Mg:Ti when used for tissue surface in vivo dosimetry outside the main treatment field of a linear accelerator X-ray beam. The normal practice would be to calibrate the TLD absolute response with the central axis beam. However, it has been shown that the near surface X-ray spectrum outside the main field has a large peak over the energy range 25–500 keV where both types of TLD have a non-uniform response, but which is absent in the central axis spectrum [8].
|
Conclusions
|
|---|
The measurements have confirmed that both types of TLD have a non-uniform relative response to low energy X-rays, with LiF:Mg:Cu:P tending to under-respond to quasi-monoenergetic radiations from 12 keV to 208 keV, whereas LiF:Mg:Ti tended to over-respond. Derivation of an energy response coefficient based on these measurements will allow the potential errors to be estimated in dosimetric applications where the X-ray energy spectrum differs from that used to calibrate the absolute TLD response. This method could be used for a preliminary investigation before a series of experimental measurements, or to assist with the choice of the most suitable TLD for a particular application. However in the latter case, the decision should also be based on the other physical characteristics of the TLD types.
|
Acknowledgments
|
|---|
The authors would like to express their gratitude to Mr Simon Evans for use of the equipment in the Medical Physics Department at the Derby Royal Infirmary to measure the thermoluminescence of the LiF:Mg:Cu:P chips, and to the Radiology Physics staff at the University Hospital of North Staffordshire for their advice.
Received for publication July 9, 2004.
Revision received November 25, 2004.
Accepted for publication January 13, 2005.
|
References
|
|---|
- Cai G. The influence of the grain size of LiF:Mg:Cu:P thermoluminescent material on photon energy response and dose response. J Phys D Appl Phys 1995;28:1259–62.[CrossRef]
- Kron T, Duggan L, Smith T, Rosenfeld A, Butson M, Kaplan G, et al. Dose response of various radiation detectors to synchrotron radiation. Phys Med Biol 1998;43:3235–60.[CrossRef][Medline]
- Kron T, Smith A, Hyodo K. Synchrotron radiation in the study of the variation of dose response in thermoluminescence dosimeters with radiation energy. Australas Phys Eng Sci Med 1996;19:225–36.[Medline]
- Thompson IMG. International Standard Reference Radiations and Their Application to the Type Testing of Dosimetric Apparatus (Proc. IAEA Symposium on National and International Standardization of Radiation Dosimetry). Vienna: IAEA, 1977:343–65.
- Edwards CR, Green S, Palethorpe JE, Mountford PJ. The response of a MOSFET, p-type semi-conductor and LiF TLD to quasi-monoenergetic X-rays. Phys Med Biol 1997;42:2383–91.[Medline]
- Green S, Palethorpe JE, Peach DE, Bradley DA. Development of a calibration facility for test instrumentation in diagnostic radiology. Radiat Prot Dosim 1996;67:41–7.[Abstract]
- Yin Z, Hugtenburg RP, Beddoe AH. Response of silicon diode dosimeters to scattered radiation from megavoltage photon beams. Radiat Prot Dosim 2002;101:415–8.[Abstract]
- Edwards CR, Mountford PJ. Near surface photon energy spectra outside a 6 MV field edge. Phys Med Biol 2004;49:N293–N301.[Medline]
- Cranley K, Gilmore BJ, Fogarty GW, Desponds L. IPEM Report 78 - Catalogue of diagnostic X-ray spectra and other data. York: The Institute of Physics and Engineering in Medicine, 1997.
- Reilly AJ, Sutton D. IPEM Report 78 – Spectrum processor. York: The Institute of Physics and Engineering in Medicine, 1997.
- McKinlay AF. Thermoluminescence dosimetry. Bristol: Adam Hilger Press, 1981.
- Mobit PN, Mayles P, Nahum AE. The quality dependence of LiF TLD in megavoltage photon beams: Monte Carlo simulation and experiments. Phys Med Biol 1996;41:387–98.[CrossRef][Medline]
- Williams JR, Thwaites DI. Radiotherapy physics. New York: Oxford University Press Inc., 1993:231–8.
Top
Abstract
Introduction
