Experimental comparison of cylindrical and plane parallel ionization chambers for reference dosimetry in continuous and pulsed scanned proton beams

Objective. In this experimental work we compared the determination of absorbed dose to water using four ionization chambers (ICs), a PTW-34045 Advanced Markus, a PTW-34001 Roos, an IBA-PPC05 and a PTW-30012 Farmer, irradiated under the same conditions in one continuous- and in two pulsed-scanned proton beams. Approach. The ICs were positioned at 2 cm depth in a water phantom in four square-field single-energy scanned-proton beams with nominal energies between 80 and 220 MeV and in the middle of 10 × 10 × 10 cm ³ dose cubes centered at 10 cm or 12.5 cm depth in water. The water-equivalent thickness (WET) of the entrance window and the effective point of measurement was considered when positioning the plane parallel (PP) ICs and the cylindrical ICs, respectively. To reduce uncertainties, all ICs were calibrated at the same primary standards laboratory. We used the beam quality (k _Q) correction factors for the ICs under investigation from IAEA TRS-398, the newly calculated Monte Carlo (MC) values and the anticipated IAEA TRS-398 updated recommendations. Main results. Dose differences among the four ICs ranged between 1.5% and 3.7% using both the TRS-398 and the newly recommended k _Q values. The spread among the chambers is reduced with the newly k _Q values. The largest differences were observed between the rest of the ICs and the IBA-PPC05 IC, obtaining lower dose with the IBA-PPC05. Significance. We provide experimental data comparing different types of chambers in different proton beam qualities. The observed dose differences between the ICs appear to be related to inconsistencies in the determination of the k _Q values. For PP ICs, MC studies account for the physical thickness of the entrance window rather than the WET. The additional energy loss that the wall material invokes is not negligible for the IBA-PPC05 and might partially explain the low k _Q values determined for this IC. To resolve this inconsistency and to benchmark MC values, k _Q values measured using calorimetry are needed.