Monte-Carlo simulation of TLD/OSLD based out-of-field fast neutron dose measurements during proton therapy Reinhard Hentschel, Bhaskar Mukherjee, Xiaoning Ding, Beate Timmermann West German Proton Therapy Centre Essen (WPE), Hufelandstrasse 55, D-45147 Essen, Germany 52 nd Annual Conference of the Particle Therapy Co-Operative Group June 2-8, 2013, Essen, Germany Introduction In proton therapy, it could be desirable to measure the out-of-field fast neutron dose at critical locations near the patient’s body. The stray neutron fluence is weak and attenuates with the square of the distance of the measuring site from the treatment volume. Hence, a very sensitive fast neutron dose measuring method is mandatory. At the same time, the space available for measurements near the patient is limited and doesn’t allow the application of big-sized moderators. In [1,2] a highly sensitive thermal neutron detector based on paired carbon doped aluminum oxide thermoluminescence dosimeters (TLD500) one of which is wrapped with natural Gadolinium is presented. This detector combined with a small moderator could be suited for the aforementioned measuring task. The working principle of such a novel clinical fast neutron dose monitor is verified by MCNPX simulation. Materials and Methods The device is based on a small Lucite moderator of just 5 cm side length for easy handling. In the simulation, a 30x30x30 cm³ polystyrene plate phantom is bombarded with a scanned SOBP proton pencil beam (range = 20 cm, modulation = 10 cm, field size = 10x10 cm²). Fast neutrons are produced inside the treatment volume of the phantom hitting also the moderator which is located near the phantom. The secondary thermal neutron flux produced inside the moderator by the impinging fast neutrons is estimated by a pair of α-Al 2 O 3 :C (TLD500) chips, which are evaluated offline after the treatment either by TL or OSL methods. The first chip is covered with natural Gadolinium foil, thereby converting the thermal neutrons to gammas via the (n,γ) reaction. The second chip is covered with a dummy material. Both chips have a distance of 2 cm from each other and may have a small lead shielding in between. Thermal neutrons from outside reaching the moderator cube are sup- pressed by a neutron capturing layer placed around the moderator. Layer materials examined are Cadmium (Cd), Gadolinium (Gd), 6 LiF, and nat LiF. An additional lead or iron sheath may be placed inbetween the thermal neutron capturing layer and the moderator to increase the amount of thermal neutrons produced in the moderator. The MC calculations performed in this study have been done with MCNPX 2.6.0 [3]. MCNPX is a general purpose Monte Carlo radiation transport code, it uses evaluated nuclear interaction cross-section libraries for energies from 1 keV to 150 MeV and the Bertini intranuclear cascade model to calculate hadronic cross-sections above 150MeV. Variance reduction techniques like splitting / Russian roulette have been used. Results Fig.1 shows calculated total and neutron dose distributions in phantom and moderator. Usually four moderators at different distances from the proton beam axis under 90° have been calculated in one run. The two Al 2 O 3 :C (TLD500) chips are assembled in the centre of the moderator in that way that they are at the same phantom depth. The simulations show that the difference of gamma doses in the TLD500 chips is correlated to the mean fast neutron dose delivered to the moderator material. In the case of uncovered moderator cubes the TLD500 signals are even higher, but the majority of thermal neutrons stem from the phantom thus falsifying the measurement. Therefore, a thermal neutron capturing layer becomes mandatory. For Cd, Gd, and 6 LiF 1 mm layer thickness is sufficient, for nat LiF the thickness must be 1 cm to ensure satisfactory suppression. nat LiF and Cd are preferred for cost reasons; the latter is easier to mount and is therefore favourised. However, all kinds of thermal neutron shielding work well in the simulation. Fig.1 – Total dose and neutron dose distributions in phantom and moderator. A lead sheath between thermal neutron shielding and moderator may improve the thermal neutron production in the moderator. Its optimum thickness has been found to be about 5 mm. Iron is less suited than lead for this purpose. Another advantage of this sheath is the weakening of gammas produced in the thermal neutron shielding. A displacement of the TLD500 chips from the moderator centre e.g. to the more distant moderator edge gives no benefit. Fig.2 shows an example how the fast neutron dose in the moderator material is correlated to the TLD500 photon dose difference for different distances d of the moderator centre from the proton beam axis. For other materials, doses can be calculated using neutron kerma factors. The small passive device could be used in clinical practice as neutron dose monitor at places outside and in the vicinity of the patient. Fig.2 – Correlation of mean moderator fast neutron dose with TLD500 photon dose difference. References [1] B. Mukherjee, R. Hentschel, J. Lambert, J. Farr, Detector and method for detecting neutrons. EP-10724702.51240, 26.05.2010 [2] B. Mukherjee, J. Lambert, R. Hentschel, J. Farr, A novel 3-He Free Neutron Area Surveillance Monitor for Proton Therapy Facilities. Particle Therapy Co-Operative Group (PTCOG 50) Meeting, Philadelphia, Pennsylvania, USA 8-14 May 2011 [3] Denise B. Pelowitz (Ed.), MCNPX User’s Manual, Version 2.6.0, LA-CP-07-1473, April 2008 0 1000 2000 3000 4000 5000 6000 7000 8000 0 50 100 150 200 250 300 350 400 PMMA neutron dose (nGy/nAs) Al 2 O 3 gamma dose difference (nGy/nAs) d = 35 cm d = 25 cm d = 20 cm d =50 cm