Open Access

In the present work we applied the Optically read out PArticle track Chamber, OPAC, for the measurement of radial dose distributions, d(r), around tracks of heavy ions passing through the gas-filled sensitive volume of the chamber. The measured data were compared with d(r) functions derived from data calculated with the Monte Carlo particle transport code, TRAX – which is used for the heavy ion therapy planning at GSI. To measure this quantity we have used here an optically read out time projection chamber (OPAC) with a parallel-drift field and one or several electron and light amplification stages. The two dimensional projection of the three dimensional ionization pattern caused by the ionizing particle passing through the chamber is captured by an image intensified CCD camera. The work is motivated by the role the radial dose distribution plays in the estimation of the relative biological effectiveness (RBE) of heavy ions, e.g. in radiation therapy and in radiation protection. The most successful model for high-dose irradiation with ions (applicable e.g. for heavy ion therapy) is found to be the local effect model (LEM). The present work intends to deliver measured data for one of the basic physical parameters which serve as input for the application of the local effect model: the radial dose distribution, d(r). The first goal of our measurement program was the measurement of d(r) distributions around carbon ions of different energies from 400 MeV/u down to the Bragg peak regions. We found an excellent agreement between the measured and simulated distributions at all carbon energies for the r–range in which the measurements deliver useful results. The lower limit of this range is about 100 nm and the upper limit is 6000 nm at a resolution of down to 33 nm - if scaled to water density. Despite the simplifications in the TRAX code (e.g. binary encounter theory for the emission ionization electrons), the discrepancies between the simulated and measured d(r) distributions are found to be lower than the measurement uncertainties at most measured carbon ion energies in almost the whole observed r-range. Hence, within the limitations of our measurements we can conclude that the precision of TRAX is sufficient to simulate the d(r) distributions around carbon ions to serve as input parameter for therapy planning. However, this conclusion is only valid for larger radial distances (r >100 nm). For smaller radial distances the measured data are dominated by the diffusion. Apart from carbon ion tracks, tracks of very heavy ions (40Ar, 84Kr and 238U) were also measured with OPAC. The simulated d(r) values were typically slightly or significantly higher than the measured data in the 100 nm < r < 5000 nm region. The experience has shown: the heavier or the faster the ion, the higher the discrepancies. On the one hand, we found a surprisingly good agreement between measurements and simulations if the ions had energies of around 50 MeV/u (i.e. relatively low energy). On the other hand, at higher energies, simulated data underestimate the measured ones by up to a factor of two in the region of 100 nm < r < 1000 nm for 84Kr (E = 650 MeV/u) or in the region of 100 nm < r < 6000 nm for 238U (E = 1 GeV/u). A possible reason for these discrepancies is that the BEA model, used in TRAX for the production ionization electrons, is not adequate for very heavy projectiles. The energy values of the very heavy ions were selected with the aim of comparing the track structures - and namely the d(r) distributions - of ions with largely different atomic mass but similar LET values. From the Z-dependency of the stopping power we know that for heavier ions a higher specific ion energy (expressed in MeV/u) is required to provide the same LET. For example the common LET of 315 keV/micro-m was achieved at largely different specific energy levels of 4,4 MeV/u for 12C, 65 MeV/u for 40Ar and 650 MeV/u for 84Kr ions. The difference in the track structures was expected mainly due to the different ion velocities and thus e.g. different ranges of d-electrons. This expectation could be confirmed by the measurements. The reason why - in line with the simulations - no strong differences could be observed in the d(r) distributions of the argon and krypton ions is the relatively small difference in the velocities of the both ion types in conjunction with the limited range in r, where the data can be compared. In contrary, the d(r) function of the carbon ion shows a qualitatively different behavior than the heavier ions inside the observable radius-range - in agreement with the simulations.