Radiation Protection Dosimetry Advance Access published March 25, 2015 Radiation Protection Dosimetry (2015), pp. 1–4
doi:10.1093/rpd/ncv056
PRELIMINARY TESTING OF GAN-BASED DOSIMETERS FOR ELECTRON BEAM RADIOTHERAPY A. Ismail1,3,*, R. Wang2, A. Chaikh3, P. Pittet2 and J. Balosso3 1 Protection and Safety Department, Atomic Energy Commission of Syria, P.B. 6091, Damascus, Syria 2 Institut des Nanotechnologies de Lyon, INL, CNRS UMR5270, Universite´ Claude Bernard Lyon1, Lyon F-69003, France 3 Service de Cance´rologie-Radiothe´rapie, CHU de Grenoble, CS 10217, Universite´ Joseph Fourier Grenoble1, Grenoble Cedex 9 F-38043, France
The response of an implantable in vivo dosimetric system based on gallium nitride radioluminescence was investigated for electron beam radiotherapy using ELEKTA SLi and VARIAN Clinac 2100 CD Linear Accelerators. A bi-channel method has been implemented for fibre background rejection. The percentage depth dose (PDD) profiles were measured in polymethyl methacrylate for 6, 12 and 18 MeV electron beams. The PDD results were in excellent agreement with those measured with reference to ionisation chambers.
INTRODUCTION In radiation therapy procedures, in vivo dosimetry of the entrance beam became recently a mandatory quality assurance process(1). This measure is widely performed by semiconductors placed on the skin of the entrance port of the beam. However, the size of such detectors is not suitable for electron beam assessment because of unacceptable dose distribution disturbance. Therefore, it would be interesting to have very small detectors avoiding this drawback. Actually, numerous types of dosimeters have been developed and investigated for monitoring dose deposited during highenergy electron beams radiotherapies(2, 3). Nevertheless, some radiation damage effects for high-energy electron beam irradiation are responsible for a gradual longterm sensitivity decay(3). Recently, a fibre-optic dosimetric (FOD) system based on the radioluminescence (RL) of gallium nitride (GaN) has been proposed(4) for implantable dosimetry in photon beam radiotherapy. This system shows a linear response over a wide dose range, no dose rate dependence and no angular dependence(5, 6). Moreover, GaN is a direct wide band gap semiconductor (Eg ¼ 3.4 eV) having a higher radiation hardness when compared with silicon, thanks to its low crystal lattice constant(7). This chapter aims to investigate the dosimetric properties of GaN-based dosimetric probe under radiotherapy electron beams of different energies, in order to evaluate its potential as a quality assurance tool in electron beam applications. MATERIALS AND METHODS The real-time GaN-based dosimeter has two main parts: an optical fibre probe and a photo detection module.
The dosimetric probe consists of a small-volume highly Si-doped GaN bulk (,0.1 mm3) used as radioluminescent transducer coupled to a silica/silica optical fibre for collecting and transmitting the RL output. The photo detection module consists of a flat field concave spherical grating (Horiba—Jobin-Yvon, France) and a 32-channel linear array multi-anode photomultiplier module (H7260M-04, Hamamatsu, Japan) for spectrophotometric signal analysis. The grating efficiency is 40 % at GaN RL wavelength. It covers a spectral range of 260–900 nm with a resolution of about 24 nm by channel. The 32-channel outputs of the module are connected to a PC for data acquisition, monitoring and processing via a CH-3160 PCI acquisition board (12-bit, 10 MSPS A/D, Acquitek, France). The bichannel method described by Pittet et al. (2013)(8) was adopted to reject the background contribution of the irradiated fibre segment. Since it was initially developed for photon beam irradiation, the performance of this method had to be specifically validated for electron beam irradiations as reported in the present study. Spectra of the photo-detected luminescence were acquired for a GaN probe and for a dummy probe without GaN transducer (background). The two detection channels had then been selected according to the GaN and background spectra: Channel 1 corresponds to the GaN RL emission peak, and Channel 2 is in the broadband spectrum of the background. It is noted that Channel 2 was chosen to be centred at a shorter wavelength than Channel 1 to prevent any direct or induced GaN RL contribution within Channel 2. Then, the relationship between the background contributions measured in the two spectral channels had been evaluated for square field sizes ranging from 3` 3 to 40` 40 cm2 and for 6 and 18 MeV electron beams. This relationship was then used
# The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
*Corresponding author: [email protected]
A. ISMAIL ET AL.
to estimate in real time the fibre background contribution in Channel 1 from its contribution measured in Channel 2. This contribution was subtracted from the signal detected in Channel 1 to determine the GaN RL contribution as explained in (8, 9). The GaN probe was held between polymethyl methacrylate (PMMA) plates (30 ` 30 ` 15 cm3) and placed at the reference depth (zref equal to 1.6, 3.9 and 5.6 cm for 6, 12 and 18 MeV beams, respectively) and centred in radiation fields ranging from 6` 6 to 25` 25 cm2, with a 100 cm source-to-surface distance (SSD). A 100 monitor units (MU) irradiation was performed with a constant dose rate of 300 MU min21. The short-term repeatability was evaluated by consecutively acquiring measurements for the same radiation dose, for all energies. Percentage depth dose (PDD) measurement was also carried out in a PMMA phantom. The SSD was kept at 100 cm, changing the depth of the dosimetric probe from 0 to 15 cm with PMMA plates. Reference measurements were performed with PTW plane parallel air-filled ionisation chamber. Three electron beam energies were used in the PDD measurement: 6, 12 and 18 MeV, while the field size has been set to 10 2 ` 10 cm for all three beam energies. To perform a quantitative dose comparison, the gamma (g) index was used(10). This tool combines a dose difference in percentage (%) criteria with a distance-to-agreement in millimetre (mm) criteria. In this study, the chosen criteria were 3 % and 3 mm. If g , 1, it represents the fulfilment of the criteria. RESULTS AND DISCUSSION Figure 1a shows the acquired luminescence spectra for GaN and dummy probes irradiated by a 6 MeV electron beam over 25` 25 cm2 fields. The GaN RL
emission peak is centred at 370 nm and is superimposed to a broadband spectrum of the background corresponding to luminescence emission from the irradiated fibre segment. It is noted that the irradiated fibre’s volume was two orders of magnitude larger than the GaN transducer volume. Thus, Channels 1 and 2 centred at 370 and 322 nm, respectively, have been set for bi-channel method implementation. Figure 1b shows, with the same detector, the bichannel measurements. The ratio of the two channel responses is independent of the irradiated segment volume of the fibre and shows no significant dependence on the electron beam energy. Thus, it confirms that background contribution in Channel 1 can be subtracted by using this ratio and measured value in Channel 2. Dosimetric measurements with fibre background rejection were performed for 3` 3, 6` 6, 10` 10 and 20` 20 cm2 square fields and for 6, 12 and 18 MeV electron beams. The short-term repeatability has been evaluated over 48 measurements performed for this test. Standard deviation, divided by the mean, remains lower than 0.7 % for each data series as shown in Table 1. Dosimetric measurements were consistent with planned doses (PD) (within +2 % for 12 and 18 MeV and +3 % for 6 MeV beams). It confirms that the fibre background rejection method was efficient for 10` 10 and 20` 20 cm2 fields. The measurements of PDD by GaN dosimetric system for 6, 12 and 18 MeV electron beams are presented in Figure 2, along with reference PDD measured by a parallel plane ionisation chamber. A good agreement for PDD measurements between GaN and ionisation chamber was observed as shown by Figure 2; however, the gamma index in the build-up region (zref 1.6 cm) for 6 MeV electron beam is out of the tolerance criteria (3 mm, 3 %). This case represents the sharpest
Page 2 of 4
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
Figure 1. (a) Emission spectra of the GaN and the dummy probes for a 6 MeV electron beam irradiation with 25` 25 cm2 square field and (b) relationship between Channel 1 and Channel 2 measurements irradiated by 6 and 18 MeV electron beams with square fields ranging from 3` 3 to 40` 40 cm2.
PRELIMINARY TESTING OF GAN DOSIMETERS FOR E-BEAMS Table 1. GaN dosimetric measurements in 6, 12 and 18 MeV electron beams, in two field sizes, compared with the PD. Energy Field size (cm2) 10` 10 20` 20
6 MeV GaN
PD
1.762+0.007 1.8 1.761+0.007 1.816
12 MeV Deviation (%) 22.097 23.022
GaN
PD
1.764+0.006 1.8 1.730+0.008 1.760
18 MeV Deviation (%) 22.021 21.706
GaN
PD
1.766+0.005 1.8 1.705+0.0089 1.733
Deviation (%) 21.896 21.594
dose gradient region where the reference chamber and the GaN dosimeter difference of size are certainly increasing the uncertainty of the measurements. CONCLUSION We have evaluated the GaN dosimetric response including PDD measurements and its repeatability for several electron beam energies and field sizes. A good agreement between GaN and reference ionisation chamber has been observed, confirming the robustness of this dosimeter. GaN is thus suitable for measuring PDD and could be used for in vivo dosimetry even as implantable probe, for instance, to check the dose delivered beyond heterogeneities as costal bone when chest wall is to be irradiated.
ACKNOWLEDGEMENTS Authors would like to acknowledge the French National Research Agency (ANR-11-TECS-018) for research funding. The first author would like to acknowledge the Atomic Energy Commission of Syria for its support during this research.
REFERENCES 1. Ismail, A., Giraud, J., Lu, G. N., Sihanath, R., Pittet, P., Galvan, J. M. and Balosso, J. Radiotherapy quality insurance by individualized in vivo dosimetry: state of the art. Cancer/Radiothe´rapie. 13, 182–189 (2009). 2. Mijnheer, B., Beddar, S., Izewska, J. and Reft, C. In vivo dosimetry in external beam radiotherapy. Med. Phys. 40, 070903 (2013).
Page 3 of 4
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
Figure 2. Comparison of GaN measured PDD with reference to PDD measured by parallel plane ionisation chamber (IC) in three beam energies at different depths.
A. ISMAIL ET AL. 7. Granta, J., Cunninghama, W., Bluea, A., O’Sheaa, V., Vaitkusb, J., Gaubasb, E. and Rahman, M. Wide bandgap semiconductor detectors for harsh radiation environments. Nucl. Instru. Meth. Phys. Res. A. 546, 213–217 (2005). 8. Pittet, P., Ismail, A., Ribouton, J., Wang, R., Galvan, J. M., Chaikh, A., Lu, G. N., Jalade, P., Giraud, J. Y. and Balosso, J. Fiber background rejection and crystal over-response compensation for GaN based in vivo dosimetry. Phys. Med. 29, 487– 492 (2013). 9. Wang, R., Pittet, P., Ribouton, J., Lu, G.-N., Chaikh, A. and Ahnesjo¨, A. Implementation and validation of a fluence pencil kernels model for GaN-based dosimetry in photon beam radiotherapy. Phys. Med. Biol. 58, 6701–6712 (2013). 10. Bakai, A., Albert, M. and Nusslin, F. A revision of the gamma-evaluation concept for the comparison of dose distributions. Phys. Med. Biol. 48, 3543– 3553 (2003).
Page 4 of 4
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
3. Dos Santos, T. C., Neves-Junior, W. F. P., Gonc¸alves, J. A. C., Haddad, C. M. K. and Bueno, C. C. Evaluation of rad-hard epitaxial silicon diode in radiotherapy electron beam dosimetry. Radiat. Meas. 46, 1662–1665 (2011). 4. Pittet, P., Lu, G. N., Galvan, J. M., Loisy, J. Y., Ismail, A., Giraud, J. Y. and Balosso, J. Implantable real-time dosimetric probe using GaN as scintillation material. Sens. Actuat. A-Phys. 151, 29– 34 (2009). 5. Chaikh, A., Balosso, J., Giraud, J.-Y., Wang, R., Pittet, P. and Lu, G.-N. Characterization of GaN dosimetry for 6MV photon beam in clinical conditions. Radiat. Meas. 71, 392–395 (2014). 6. Ismail, A., Pittet, P., Lu, G. N., Galvan, J. M., Giraud, J. Y. and Balosso, J. In vivo dosimetric system based on Gallium Nitride radioluminescence. Radiat. Meas. 46, 1960–1963 (2011).
doi:10.1093/rpd/ncv056
PRELIMINARY TESTING OF GAN-BASED DOSIMETERS FOR ELECTRON BEAM RADIOTHERAPY A. Ismail1,3,*, R. Wang2, A. Chaikh3, P. Pittet2 and J. Balosso3 1 Protection and Safety Department, Atomic Energy Commission of Syria, P.B. 6091, Damascus, Syria 2 Institut des Nanotechnologies de Lyon, INL, CNRS UMR5270, Universite´ Claude Bernard Lyon1, Lyon F-69003, France 3 Service de Cance´rologie-Radiothe´rapie, CHU de Grenoble, CS 10217, Universite´ Joseph Fourier Grenoble1, Grenoble Cedex 9 F-38043, France
The response of an implantable in vivo dosimetric system based on gallium nitride radioluminescence was investigated for electron beam radiotherapy using ELEKTA SLi and VARIAN Clinac 2100 CD Linear Accelerators. A bi-channel method has been implemented for fibre background rejection. The percentage depth dose (PDD) profiles were measured in polymethyl methacrylate for 6, 12 and 18 MeV electron beams. The PDD results were in excellent agreement with those measured with reference to ionisation chambers.
INTRODUCTION In radiation therapy procedures, in vivo dosimetry of the entrance beam became recently a mandatory quality assurance process(1). This measure is widely performed by semiconductors placed on the skin of the entrance port of the beam. However, the size of such detectors is not suitable for electron beam assessment because of unacceptable dose distribution disturbance. Therefore, it would be interesting to have very small detectors avoiding this drawback. Actually, numerous types of dosimeters have been developed and investigated for monitoring dose deposited during highenergy electron beams radiotherapies(2, 3). Nevertheless, some radiation damage effects for high-energy electron beam irradiation are responsible for a gradual longterm sensitivity decay(3). Recently, a fibre-optic dosimetric (FOD) system based on the radioluminescence (RL) of gallium nitride (GaN) has been proposed(4) for implantable dosimetry in photon beam radiotherapy. This system shows a linear response over a wide dose range, no dose rate dependence and no angular dependence(5, 6). Moreover, GaN is a direct wide band gap semiconductor (Eg ¼ 3.4 eV) having a higher radiation hardness when compared with silicon, thanks to its low crystal lattice constant(7). This chapter aims to investigate the dosimetric properties of GaN-based dosimetric probe under radiotherapy electron beams of different energies, in order to evaluate its potential as a quality assurance tool in electron beam applications. MATERIALS AND METHODS The real-time GaN-based dosimeter has two main parts: an optical fibre probe and a photo detection module.
The dosimetric probe consists of a small-volume highly Si-doped GaN bulk (,0.1 mm3) used as radioluminescent transducer coupled to a silica/silica optical fibre for collecting and transmitting the RL output. The photo detection module consists of a flat field concave spherical grating (Horiba—Jobin-Yvon, France) and a 32-channel linear array multi-anode photomultiplier module (H7260M-04, Hamamatsu, Japan) for spectrophotometric signal analysis. The grating efficiency is 40 % at GaN RL wavelength. It covers a spectral range of 260–900 nm with a resolution of about 24 nm by channel. The 32-channel outputs of the module are connected to a PC for data acquisition, monitoring and processing via a CH-3160 PCI acquisition board (12-bit, 10 MSPS A/D, Acquitek, France). The bichannel method described by Pittet et al. (2013)(8) was adopted to reject the background contribution of the irradiated fibre segment. Since it was initially developed for photon beam irradiation, the performance of this method had to be specifically validated for electron beam irradiations as reported in the present study. Spectra of the photo-detected luminescence were acquired for a GaN probe and for a dummy probe without GaN transducer (background). The two detection channels had then been selected according to the GaN and background spectra: Channel 1 corresponds to the GaN RL emission peak, and Channel 2 is in the broadband spectrum of the background. It is noted that Channel 2 was chosen to be centred at a shorter wavelength than Channel 1 to prevent any direct or induced GaN RL contribution within Channel 2. Then, the relationship between the background contributions measured in the two spectral channels had been evaluated for square field sizes ranging from 3` 3 to 40` 40 cm2 and for 6 and 18 MeV electron beams. This relationship was then used
# The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
*Corresponding author: [email protected]
A. ISMAIL ET AL.
to estimate in real time the fibre background contribution in Channel 1 from its contribution measured in Channel 2. This contribution was subtracted from the signal detected in Channel 1 to determine the GaN RL contribution as explained in (8, 9). The GaN probe was held between polymethyl methacrylate (PMMA) plates (30 ` 30 ` 15 cm3) and placed at the reference depth (zref equal to 1.6, 3.9 and 5.6 cm for 6, 12 and 18 MeV beams, respectively) and centred in radiation fields ranging from 6` 6 to 25` 25 cm2, with a 100 cm source-to-surface distance (SSD). A 100 monitor units (MU) irradiation was performed with a constant dose rate of 300 MU min21. The short-term repeatability was evaluated by consecutively acquiring measurements for the same radiation dose, for all energies. Percentage depth dose (PDD) measurement was also carried out in a PMMA phantom. The SSD was kept at 100 cm, changing the depth of the dosimetric probe from 0 to 15 cm with PMMA plates. Reference measurements were performed with PTW plane parallel air-filled ionisation chamber. Three electron beam energies were used in the PDD measurement: 6, 12 and 18 MeV, while the field size has been set to 10 2 ` 10 cm for all three beam energies. To perform a quantitative dose comparison, the gamma (g) index was used(10). This tool combines a dose difference in percentage (%) criteria with a distance-to-agreement in millimetre (mm) criteria. In this study, the chosen criteria were 3 % and 3 mm. If g , 1, it represents the fulfilment of the criteria. RESULTS AND DISCUSSION Figure 1a shows the acquired luminescence spectra for GaN and dummy probes irradiated by a 6 MeV electron beam over 25` 25 cm2 fields. The GaN RL
emission peak is centred at 370 nm and is superimposed to a broadband spectrum of the background corresponding to luminescence emission from the irradiated fibre segment. It is noted that the irradiated fibre’s volume was two orders of magnitude larger than the GaN transducer volume. Thus, Channels 1 and 2 centred at 370 and 322 nm, respectively, have been set for bi-channel method implementation. Figure 1b shows, with the same detector, the bichannel measurements. The ratio of the two channel responses is independent of the irradiated segment volume of the fibre and shows no significant dependence on the electron beam energy. Thus, it confirms that background contribution in Channel 1 can be subtracted by using this ratio and measured value in Channel 2. Dosimetric measurements with fibre background rejection were performed for 3` 3, 6` 6, 10` 10 and 20` 20 cm2 square fields and for 6, 12 and 18 MeV electron beams. The short-term repeatability has been evaluated over 48 measurements performed for this test. Standard deviation, divided by the mean, remains lower than 0.7 % for each data series as shown in Table 1. Dosimetric measurements were consistent with planned doses (PD) (within +2 % for 12 and 18 MeV and +3 % for 6 MeV beams). It confirms that the fibre background rejection method was efficient for 10` 10 and 20` 20 cm2 fields. The measurements of PDD by GaN dosimetric system for 6, 12 and 18 MeV electron beams are presented in Figure 2, along with reference PDD measured by a parallel plane ionisation chamber. A good agreement for PDD measurements between GaN and ionisation chamber was observed as shown by Figure 2; however, the gamma index in the build-up region (zref 1.6 cm) for 6 MeV electron beam is out of the tolerance criteria (3 mm, 3 %). This case represents the sharpest
Page 2 of 4
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
Figure 1. (a) Emission spectra of the GaN and the dummy probes for a 6 MeV electron beam irradiation with 25` 25 cm2 square field and (b) relationship between Channel 1 and Channel 2 measurements irradiated by 6 and 18 MeV electron beams with square fields ranging from 3` 3 to 40` 40 cm2.
PRELIMINARY TESTING OF GAN DOSIMETERS FOR E-BEAMS Table 1. GaN dosimetric measurements in 6, 12 and 18 MeV electron beams, in two field sizes, compared with the PD. Energy Field size (cm2) 10` 10 20` 20
6 MeV GaN
PD
1.762+0.007 1.8 1.761+0.007 1.816
12 MeV Deviation (%) 22.097 23.022
GaN
PD
1.764+0.006 1.8 1.730+0.008 1.760
18 MeV Deviation (%) 22.021 21.706
GaN
PD
1.766+0.005 1.8 1.705+0.0089 1.733
Deviation (%) 21.896 21.594
dose gradient region where the reference chamber and the GaN dosimeter difference of size are certainly increasing the uncertainty of the measurements. CONCLUSION We have evaluated the GaN dosimetric response including PDD measurements and its repeatability for several electron beam energies and field sizes. A good agreement between GaN and reference ionisation chamber has been observed, confirming the robustness of this dosimeter. GaN is thus suitable for measuring PDD and could be used for in vivo dosimetry even as implantable probe, for instance, to check the dose delivered beyond heterogeneities as costal bone when chest wall is to be irradiated.
ACKNOWLEDGEMENTS Authors would like to acknowledge the French National Research Agency (ANR-11-TECS-018) for research funding. The first author would like to acknowledge the Atomic Energy Commission of Syria for its support during this research.
REFERENCES 1. Ismail, A., Giraud, J., Lu, G. N., Sihanath, R., Pittet, P., Galvan, J. M. and Balosso, J. Radiotherapy quality insurance by individualized in vivo dosimetry: state of the art. Cancer/Radiothe´rapie. 13, 182–189 (2009). 2. Mijnheer, B., Beddar, S., Izewska, J. and Reft, C. In vivo dosimetry in external beam radiotherapy. Med. Phys. 40, 070903 (2013).
Page 3 of 4
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
Figure 2. Comparison of GaN measured PDD with reference to PDD measured by parallel plane ionisation chamber (IC) in three beam energies at different depths.
A. ISMAIL ET AL. 7. Granta, J., Cunninghama, W., Bluea, A., O’Sheaa, V., Vaitkusb, J., Gaubasb, E. and Rahman, M. Wide bandgap semiconductor detectors for harsh radiation environments. Nucl. Instru. Meth. Phys. Res. A. 546, 213–217 (2005). 8. Pittet, P., Ismail, A., Ribouton, J., Wang, R., Galvan, J. M., Chaikh, A., Lu, G. N., Jalade, P., Giraud, J. Y. and Balosso, J. Fiber background rejection and crystal over-response compensation for GaN based in vivo dosimetry. Phys. Med. 29, 487– 492 (2013). 9. Wang, R., Pittet, P., Ribouton, J., Lu, G.-N., Chaikh, A. and Ahnesjo¨, A. Implementation and validation of a fluence pencil kernels model for GaN-based dosimetry in photon beam radiotherapy. Phys. Med. Biol. 58, 6701–6712 (2013). 10. Bakai, A., Albert, M. and Nusslin, F. A revision of the gamma-evaluation concept for the comparison of dose distributions. Phys. Med. Biol. 48, 3543– 3553 (2003).
Page 4 of 4
Downloaded from http://rpd.oxfordjournals.org/ at North Dakota State University on July 14, 2015
3. Dos Santos, T. C., Neves-Junior, W. F. P., Gonc¸alves, J. A. C., Haddad, C. M. K. and Bueno, C. C. Evaluation of rad-hard epitaxial silicon diode in radiotherapy electron beam dosimetry. Radiat. Meas. 46, 1662–1665 (2011). 4. Pittet, P., Lu, G. N., Galvan, J. M., Loisy, J. Y., Ismail, A., Giraud, J. Y. and Balosso, J. Implantable real-time dosimetric probe using GaN as scintillation material. Sens. Actuat. A-Phys. 151, 29– 34 (2009). 5. Chaikh, A., Balosso, J., Giraud, J.-Y., Wang, R., Pittet, P. and Lu, G.-N. Characterization of GaN dosimetry for 6MV photon beam in clinical conditions. Radiat. Meas. 71, 392–395 (2014). 6. Ismail, A., Pittet, P., Lu, G. N., Galvan, J. M., Giraud, J. Y. and Balosso, J. In vivo dosimetric system based on Gallium Nitride radioluminescence. Radiat. Meas. 46, 1960–1963 (2011).
