Response of synthetic diamond detectors in proton, carbon, and oxygen ion beams

Transcription

1 Response of synthetic diamond detectors in proton, carbon, and oxygen ion beams Severine Rossomme a) Center of Molecular Imaging, Radiotherapy and Oncology, Institut de Recherche Experimentale et Clinique, Universite catholique de Louvain, Brussels 1200, Belgium Marco Marinelli and Gianluca Verona-Rinati INFN-Dipartimento di Ingegneria Industriale, Universita di Roma Tor Vergata, Roma 00173, Italy Francesco Romano and Pablo Antonio Giuseppe Cirrone Laboratori Nazionali del Sud, Istituto Nazionale di Fisica Nucleare, Catania Sicily, Italy Andrzej Kacperek National Eye Proton Therapy Centre, Clatterbridge Cancer Centre, Wirral CH63 4JY, UK Stefaan Vynckier Center of Molecular Imaging, Radiotherapy and Oncology, Institut de Recherche Experimentale et Clinique, Universite catholique de Louvain, Brussels 1200, Belgium Department of Radiotherapy and Oncology, Cliniques Universitaires Saint-Luc, Brussels 1200, Belgium Hugo Palmans National Physical Laboratory, Acoustics and Ionising Radiation Division Teddington TW11 0LW, UK EBG MedAustron GmbH, Wiener Neustadt 2700, Austria (Received 1 March 2017; revised 29 June 2017; accepted for publication 4 July 2017; published 18 August 2017) Purpose: In this work, the LET-dependence of the response of synthetic diamond detectors is investigated in different particle beams. Method: Measurements were performed in three nonmodulated particle beams (proton, carbon, and oxygen). The response of five synthetic diamond detectors was compared to the response of a Markus or an Advanced Markus ionization. The synthetic diamond detectors were used with their axis parallel to the beam axis and without any bias voltage. A high bias voltage was applied to the ionization s, to minimize ion recombination, for which no correction is applied (+300 V and +400 V were applied to the Markus and Advanced Markus ionization s respectively). Results: The ratio between the normalized response of the synthetic diamond detectors and the normalized response of the ionization shows an under-response of the synthetic diamond detectors in carbon and oxygen ion beams. No under-response of the synthetic diamond detectors is observed in protons. For each beam, combining results obtained for the five synthetic diamond detectors and considering the uncertainties, a linear fit of the ratio between the normalized response of the synthetic diamond detectors and the normalized response of the ionization is determined. The response of the synthetic diamond detectors can be described as a function of LET as ( 6.22E E-3) LET + ( ) in proton beam, ( 2.51E E-4) LET + ( ) in carbon ion beam and ( 2.77E E-4) LET + ( ) in oxygen ion beam. Combining results obtained in carbon and oxygen ion beams, a LET dependence of about 0.026% (0.013%) per kev/lm is estimated. Conclusions: Due to the high LET value, a LET dependence of the response of the synthetic diamond detector was observed in the case of carbon and oxygen beams. The effect was found to be negligible in proton beams, due to the low LET value. The under-response of the synthetic diamond detector may result from the recombination of electron/hole in the thin synthetic diamond layer, due to the high LET-values. More investigations are required to confirm this assumption American Association of Physicists in Medicine [ Key words: LET-dependence, particle beams, relative dosimetry, synthetic diamond detector 1. INTRODUCTION Many detectors are used for clinical dosimetry applications, e.g., ionization s, radiochromic films or diodes. The behavior and performances of dosimetry detectors depend on their physical characteristics and properties, such as linearity, dose rate dependence, energy dependence, or spatial resolution. In particle therapy, another quantity can influence the response of the detector: the energy transferred per unit length of the track, i.e., the linear energy transfer (LET). This paper addresses the LET dependence of the response of PTW microdiamond detectors in particle beams Med. Phys. 44 (10), October /2017/44(10)/5445/ American Association of Physicists in Medicine 5445

2 5446 Rossomme et al.: LET-dependence of synthetic diamond detectors 5446 Details of the operational principle of this detector, which is based on a Schottky diode, can be found in papers by Almaviva et al. 1 and Di Venanzio et al. 2 Characteristics of the PTW detector can be found in PTW manuals. 3,4 In photon, electron and proton beams, this detector is described as suitable for small field dosimetry Also in protons, most dosimetry studies conclude that the response of the PTW microdiamond detector exhibits no LET dependence nor quenching effect (Mandapaka et al., 11 Marinelli et al., 12 Yuichi et al., 13 Rossomme et al., 14 Akino et al. 13 ). In, 15 which is based on an intercomparison of four microdiamond detectors in a clinical 138 MeV proton beam, authors reported a nonreproducibility between devices in terms of stability, sensitivity, and LET dependence. The response of the PTW microdiamond detector has been little investigated in carbon ion beams. A study by Marinelli et al. 16 indicates that the response of this type of detector exhibits no LET dependence, although this study was performed in highenergy clinical beams, where the energy spread in the Bragg peak is large and small quenching effects would not be very apparent. In a second study, Rossomme et al. 17 observed a 20% relative under-response of a PTW microdiamond detector compared to that of a Markus ionization, in the Bragg peak region of a 62 MeV/n carbon ion beam. Relating the data to the variation in the LET of the carbon ion beam, Rossomme et al. concluded that the response of the microdiamond detector used in the study is LET dependent compared to that of a Markus ionization. In this paper, we investigate the LET-dependence of the response of five different PTW microdiamond detectors, in three particle beams. The response of the detectors is studied as a function of beam LET-values, and by comparison with ionization response. SN , SN , and SN (sensitive volume: circular, radius = 1.1 mm, thickness 1 lm; water-equivalent window thickness = 1.0 mm) 3 and two different ionization s were used. Except the detector with the serial number , other detectors (with the serial numbers , , , and ) are noncommercial microdiamond detectors because of a slightly lower sensitivity. A Markus plane-parallel ionization with serial number SN 862 (sensitive volume: radius 2.65 mm, depth 2 mm; entrance foil = 0.03 mm polyethylene CH 2 ; protection cap = 0.87 mm PMMA and 0.4 mm air, guard ring width < 0.2 mm) was used at CCC and an Advanced Markus plane-parallel ionization with serial number SN 0068 (sensitive volume: radius 2.5 mm, depth 1 mm; entrance foil = 0.03 mm polyethylene CH 2 ; protection cap = 0.87 mm PMMA and 0.4 mm air, guard ring width = 2 mm) at INFN-LNS. 3 Measurements made with the microdiamond detectors were performed with their axis parallel to the beam axis. As recommended, the microdiamond detectors were used without any bias voltage. Figure 1 shows the schematic experimental set-up for the measurements using the ionization or the microdiamond detector. The rated maximum high voltages were applied to the ionization s (+300 V at CCC and +400 V at INFN- 2. METHOD The experimental work in this study was performed at Clatterbridge Cancer Centre (CCC) in United Kingdom and Istituto Nazionale di Fisica Nucleare-Laboratori Nazionali del Sud (INFN-LNS) in Sicily. At CCC, a clinical 60 MeV nonmodulated proton beam, produced by a Scanditronix MC62PF cyclotron, was used. The measurements at INFN- LNS were performed in a 62 MeV/n nonmodulated carbon ion beam and a 62 MeV/n nonmodulated oxygen ion beam, produced by a K800 superconducting cyclotron, using the beam line named 0 degree beam line. The configurations of the beam lines at CCC and INFN-LNS are similar. After being scattered by a tantalum foil (at LNS), the beam exits the vacuum tube through a Kapton window. In the nozzle, a 25 mm diameter brass aperture was used as a final collimator. All of these beams have a very sharp distal fall-off of less than 1 mm, which would indicate a good energy resolution. The experimental procedure consisted of measuring the response of each detector as a function of the depth in water. To this end, a computer controlled motorized water phantom was used in all measurements. Five PTW microdiamond detectors with serial number SN , SN , FIG. 1. Schematic set-up for the experimental determination using ionization (upper panel) and microdiamond detectors (lower panel). [Color figure can be viewed at wileyonlinelibrary.com]

3 5447 Rossomme et al.: LET-dependence of synthetic diamond detectors 5447 LNS), to minimize ion recombination. 18,19 Therefore, no ion recombination correction factor was applied to the response of the ionization s. Furthermore, the experimental data used in the analysis were restricted to a depth range from the surface up to the depth proximal to the Bragg peak, where the response of the ionization is 80% of the peak. In this region, recombination is not expected to be substantial nor substantially LET dependent, 19 justifying not to apply recombination corrections. Beyond this depth, a strong increase in ion recombination is expected, 19 requiring the determination of ion recombination corrections which was not feasible with the beam time available. As reported by Rossomme et al., 19 due to the increase in the LET in the Bragg peak region, initial recombination increases in that region. Consequently, this will induce an underestimation of the response of the ionization in the Bragg peak region. As no temperature and pressure variation was observed during the depth dose acquisition and as only relative dosimetry measurements were performed, no temperature and pressure correction was applied. The integral of the response of each detector was normalized to unity over a depth range of few millimeters, to minimize the effect of the instabilities of the nonclinical beams. The normalization region is chosen in the entrance dose region, between 0.18 cm and 0.4 cm for the carbon ion beam and 0.18 cm and 0.3 cm for the oxygen ion beam (0.18 cm corresponds to the first water depth where we measured a signal with the detector). For the proton beam, due to a technical positioning problem with the Markus when it was too close to the phantom window, the shallowest points are excluded of the analysis. For this reason, the normalization region was chosen between 1 cm and 1.28 cm. Based on previous experimental work for protons published in, 14 as long as the normalization region is in the entrance dose region, its position will not influence the results due to the low LET-variation. For each beam, the depth-ionization curves obtained with all detectors were shifted so as to have the same values of the continuous slowing down approximation (CSDA) range in order to account for small positioning errors and the small variation in the nominal thickness of the entrance window of each detector. The CSDA range was to this end experimentally determined as the depth on the distal edge of the Bragg peak where the dose drops to 80% of the maximum. The final results presented in this paper are the ratio of the normalized response of raw data obtained with the microdiamond detectors and a polynomial fit of the normalized response of the ionization, as a function of LET-values. These LETvalues were determined using Monte Carlo simulations, based on the Geant4 code, as explained in RESULTS Figures 2, 3 and 4 show the ratio between the normalized responses of the microdiamond detector and of the ionization, as a function of LET in proton, carbon, and oxygen ion beams respectively. Uncertainty bars shown take into account the reproducibility of the measurement (uncertainty Ratio microdiamond/ Linear interpolation: y = 6.22E 4*LET FIG. 2. The ratio between the normalized response of microdiamond detectors and the normalized response of a Markus ionization in a 60 MeV proton beam. [Color figure can be viewed at wileyonlinelibrary.com] Ratio microdiamond/ Linear fit : y = 2.51E 4*LET FIG. 3. The ratio between the normalized response of microdiamond detectors and the normalized response of an Advanced Markus ionization in a 62 MeV/n carbon ion beam. [Color figure can be viewed at wileyonlinelibrary.com] of type A), the fit of the ionization data (uncertainty of type B) and the range determination (uncertainty of type B). The uncertainty due to the Monte Carlo evaluation of the LET was not included in the analyses, because it will not contribute significantly to the total uncertainty given the much higher experimental uncertainty. The uncertainty contribution due to the range uncertainty was estimated as the variation in the ratio presented in Figs. 2, 3 and 4 due to shifting the curve obtained with the ionization toward lower and higher depths. The shift was chosen as 0.1% of the water range value, which corresponds to 31 lm for the proton beam, 9 lm for the carbon ion beam and 7 lm for the oxygen ion beam. The profiles of carbon ion beam and oxygen

4 5448 Rossomme et al.: LET-dependence of synthetic diamond detectors 5448 Ratio microdiamond/ ion beam used during the experimental session were not perfectly uniform. Due to the difference between the diameter of the Advanced Markus ionization and the diameter of the microdiamond detectors, a correction should be applied to the ratio presented in Figs. 3 and 4. The variation in the beam profiles was not investigated in detail during the experimental session but it was assumed to be dominated by the divergence of the beam and thus it can be assumed that no correction needs to be applied for the small nonhomogeneity of the beam. Figure 2 shows the results obtained in a 60 MeV proton beam. For the five microdiamond detectors, a linear fit of the ratio between the normalized response of the microdiamond diamond and the normalized response of the ionization was determined. The average of these five linear fits, given by the following equation R microdiamond Linear fit : y = 2.77E 4*LET FIG. 4. The ratio between the normalized response of microdiamond detectors and the normalized response of an Advanced Markus ionization in a 62 MeV/n oxygen ion beam. [Color figure can be viewed at wileyonlinelibrary.com] ¼ð 6: : ÞLET þð0:985 0:005Þ is represented by the continuous line in Fig. 2. Taking into account the uncertainty of both coefficients of the linear fit, dashed lines represent the standard deviation of the linear fit. In the 60 MeV proton beam, the LET-value is relatively low. Its value increases slowly in the entrance dose region, from 1.5 to 2 kev/lm. LET-value increases upto ~20 kev/lm, at the depth corresponding to the range 80%. As expected, results confirm the LET-independent response observed in previous experimental work Considering the uncertainties of the results, a LET dependence of the microdiamond detectors is not measurable and we conclude that there is agreement between the normalized response of the microdiamond detectors and the normalized response of Markus ionization. Figures 3 and 4 show the results obtained in a 62 MeV/n carbon ion beam and a 62 MeV/n oxygen ion beam respectively. Unlike the results obtained in Fig. 2, we observe an under-response of the microdiamond detectors compared to the response of the Advanced Markus ionization, when the LET-value increases. This observation highlights the dependence of the response of the microdiamond detector on LET. Taking into account the uncertainties, no significant difference in the LET dependence of the response of the five microdiamond detectors is observed. With carbon ions (Fig. 3), the microdiamond detectors SN , SN , and SN have a significant LET-dependence response, when the LET value is higher than 100 kev/lm. At 346 kev/lm, the under-response of the microdiamond detector SN equals 13 (4) %. Results obtained with the microdiamond detector SN confirm those obtained using the same detector, in the same beam, and published in. 17 The microdiamond detectors SN and SN show a much smaller LET-dependence that can be observed above 300 kev/lm. The continuous line in Fig. 3 represent the average of the five fits of the ratio between the normalized response of the microdiamond diamond and the normalized response of the ionization, which is given by R microdiamind ¼ð 2: : ÞLET þð1:008 0:005Þ: The linear fit indicates a LET-dependence of 0.025% (0.012%) per kev/lm. Comparing the behavior of the results obtained in the 62 MeV/n carbon ion beam and the 62 MeV/n oxygen ion beam up to an LET of 350 kev/lm, results are similar. In oxygen ion beam, for high LET-values, the response of all microdiamond detectors is LET-dependent. At 530 kev/lm, we observe an underestimation between 12% (for the microdiamond detector SN ) and 15% (for the microdiamond detector SN ). Similar to the analysis of the results obtained in the carbon ion beam, an average of experimental data was determined. The average of the five linear fits of the ratio between the normalized response of the microdiamond detector and the normalized response of the ionization is given by R microdiamond ¼ð 2: : ÞLET þð1:034 0:012Þ; which indicates a LET-dependence of 0.028% (0.006%) per kev/lm. This study reveals a LET-dependence of the response of the microdiamond detectors. Taking into account the uncertainties, contrary with the results presented by Marsolat et al., 9 all microdiamond detectors used in this work have a similar behavior as function of LET. Combining results obtained in carbon and oxygen ion beams (Figs. 3 and 4), a LET dependence of about 0.026% (0.013%) per kev/lm can be estimated. A negligible dependence is thus to be expected with protons due to the much lower LET-values. This result is

5 5449 Rossomme et al.: LET-dependence of synthetic diamond detectors 5449 confirmed by the experimental data obtained in protons (Fig. 2), in which no LET dependence is observed. These conclusions are in agreement with previous experimental work obtained with a single microdiamond detector (SN ). 14,16 As for the microdiamond detector under-response in carbon and oxygen beams, the authors believe that it could be explained in terms of certain recombination effects in the thin synthetic diamond layer due to the LET value. Electrons and holes in the synthetic diamond layer might recombine before being collected, causing a decrease in the collection efficiency of the detector. The substantial uncertainty of the LET dependence of the PTW microdiamond detector determined in this study results from a combination of two potential effects: detector-to-detector variability (as, for example, also observed by Marsolat et al. 9 for protons and Ralston et al. 10 for small photon fields) and instabilities of the non clinical beams in which the investigations were performed. 4. CONCLUSIONS The objective of this study was to evaluate the LET dependence of the PTW microdiamond detector, in three nonmodulated particle beams (proton, carbon, and oxygen). The response of five microdiamond detectors was studied by comparison with the response of a Markus ionization (proton beam) or an Advanced Markus ionization (carbon and oxygen ion beams). Results were presented as the ratio between the normalized response of the microdiamond detector and the normalized response of the ionization, as a function of the LET-values. No correction was applied to the response of the ionization or the response of the microdiamond detector. To minimize the under-response of the ionization s due to absence of ion recombination correction, they were used with a high bias voltage. Taking into account the uncertainties, the behavior of all microdiamond detectors is similar as function of LET. By combining results obtained in carbon and oxygen ion beams, an average LET dependence of the PTW microdiamond detector of about 0.026% ( 0.013%) per kev/lm is estimated. A negligible dependence is thus to be expected with protons due to the lower LET-values, which is demonstrated by the described results. ACKNOWLEDGMENTS Severine Rossomme was funded by Fond National de la Recherche Scientifique ( Televie Grant). The authors thank LNS Scientific Committee for providing the opportunity to perform the experimental session. The study performed at INFN LNS has received funding from the European Union HORIZON2020 research and innovation programme under Grant Agreement n ENSAR2. CONFLICT OF INTEREST Two of the authors, Marco Marinelli and Gianluca Verona Rinati, signed a contract with PTW-Freiburg involving financial interests deriving from the microdiamond dosimeter commercialization. a) Author to whom correspondence should be addressed. Electronic mail: severine.rossomme@uclouvain.be. REFERENCES 1. Almaviva S, Marinelli M, Milani E, et al. Chemical vapor deposition diamond based multilayered radiation detector: physical analyzis of detection properties. J Appl Phys. 2010;107: Di Venanzio C, Marinelli M, Milani E, et al. Characterization of a synthetic single crystal diamond Schottky diode for radiotherapy electron beam dosimetry. Med Phys. 2013;40: Detectors Ionizing radiation, Including Codes of Practice, manual, PTW MicroDiamond synthetic diamond detector for high-precision dosimetry: nearly as good as water, manual, PTW, Editor Azangwe G, Grochowska P, Georg D, et al. Detector to detector corrections: a comprehensive experimental study of detector specific correction factors for beam output measurements for small radiotherapy beams. Med Phys. 2014;41: Bagal P, Di Venanzio C, Falco M, et al. Radiotherapy electron beams collimated by small tubular applicators: characterization by silicon and diamond diodes. Phys Med Biol. 2013;58: Ciancaglioni I, Marinelli M, Milani E, et al. Dosimetric characterization of a synthetic single crystal diamond detector in clinical radiation therapy small photon beams. Med Phys. 2012;39: Lechner W, Palmans H, S olkner L, Grochowska P, Georg D. Detector comparison for small field output factor measurements in flattening filter free photon beams. Radiother Oncol. 2013;109: Marsolat F, Tromson D, Tranchant N, et al. A new single crystal diamond dosimeter for small beam: comparison with different commercial active detectors. Phys Med Biol. 2013;58: Ralston A, Tyler M, Liu P, McKenzie D, Suchowerska N. Over-response of synthetic microdiamond detectors in small radiation fields. Phys Med Biol. 2014;59: Mandapaka AK, Ghebremedhin A, Patyal B, et al. Evaluation of the dosimetric properties of a synthetic single crystal diamond detector in high energy clinical proton beams. Med Phys. 2013;40: Marinelli M, Pompili F, Prestopino G, et al. Dosimetric characterization of a synthetic single crystal diamond detector in a clinical 62 MeV ocular therapy proton beam. Nucl Instr Meth Phys Res A. 2014;767: Akino Y, Gautam A, Coutinho L, W urfel J, Das I. Characterization of a new commercial single crystal diamond detector for photon- and proton-beam dosimetry. J Radiat Res. 2015;56: Rossomme S, Denis JM, Souris K, et al. LET dependence of the response of a PTW microdiamond detector in a 62 MeV proton beam. Phys Med. 2016;32: Marsolat F, De Marzi L, Patriarca A, et al. Dosimetric characteristics of four PTW microdiamond detectors in high-energy proton beams. Phys Med Biol. 2016;61: Marinelli M, Prestopino G, Verona C, et al. Dosimetric characterization of a microdiamond detector in clinical scanned carbon ion beams. Med Phys. 2015;42: Rossomme S, Hopfgartner J, Vynckier S, Palmans H. Under response of a PTW microdiamond detector in the Bragg peak of a 62 MeV/n carbon ion beam. Phys Med Biol. 2016;61: Palmans H, Thomas R, Kacperek A. Ion recombination correction in the Clatterbridge Center of Oncology clinical proton beam. Phys Med Biol. 2006;51: Rossomme S, Hopfgartner J, Lee N D, et al. Ion recombination correction in carbon ion beams. Med Phys. 2016;43: Romano F, Cirrone GA, Cuttone G, et al. A Monte Carlo study for the calculation of the average linear energy transfer (LET) distributions for a clinical proton beam line and a radiobiological carbon ion beam line. Phys Med Biol. 2014;59: