Abstract
Medical linear accelerators (linacs) operated above 8 MV photon energy have their output contaminated with neutrons resulting from the photon interactions with the materials of the linac head components. Photoneutron contamination in the photon output was investigated on the Varian Clinac 2100C linear accelerator operating at 15 MV photon beam energy and a dose rate of 4 Gy/min using experimental and Monte Carlo (MCNP) simulations. In order to ensure that the output of the MCNP code was valid, an 241Am-Be isotopic irradiation facility was simulated to study the scattered and transmitted neutron fluxes emanating from a rectangular phantom placed in 'short' and 'long' width orientations. The results of the scattered and transmitted neutron fluxes were compared with those obtained by measurements using the Microspect-2 Neutron Probe (N-Probe) and a new neutron detector, the LGB detector, based on a scintillator containing Li, Gd, and B. Results show a reasonable agreement between measurements and MCNP calculations for both transmitted and scattered neutron flux. Good and accurate knowledge of all the relevant nuclear parameters involved and reliable as well as reproducible experimental conditions are required in neutron flux measurements using foils. A Monte Carlo based Fortran 90 program, COLDET, was developed to calculate the solid angle subtended by both 'bare' and collimated y-ray detector to point, disc and cylindrical sources. Results show good agreement with theory and those obtained by others, however, some differences arise when the finite dimensions of the source and detector are taken into account. Superheated drop detectors (SDDs) were employed in the photoneutron dose measurements due to their insensitivity to high energy and intensity photons in contrast to activation foils. The high and low neutron sensitivity SDDs (Apfel Enterprises Inc., U.S.A), recommended for out of beam and in-beam measurements were used, respectively. Measurements were carried out for both in air and in a water phantom in the patient plane at 100 cm source-to-surface (detector) distance (SSD) in order to investigate the variation of neutron dose equivalent with field size in and outside the irradiation beam and also in the maze of the linac bunker and the control room. The neutron dose equivalent on the beam axis increased gradually as field size was varied from 5x5 cm to 40x40 cm for in-air measurements and was independent of field size outside the irradiated field at distances greater than 20 cm. The neutron dose equivalent for 10x10 cm2 and 40x40 cm2 field sizes was found to be 1.57+/-0.10 mSvGy-1 and 1.74 +/- 0.09 mSvGy-1, respectively. The neutron dose equivalent in the maze for all the field sizes was much higher at gantry angles 0 and 180. The neutron dose equivalent on the beam axis for the 10x10 cm2 field size at a depth of 1 cm in water was 1.42 +/- 0.09 mSvGy-1 for the in-phantom measurements and 1.81+/-0.08 mSvGy-1 for the 5x5 cm2 field size for the same depth but was independent of field size at depth >5 cm in water. MCNP simulation of the 15 MV photon energy Varian Clinac 2100C head was carried out to investigate the photoneutron contamination in its output for the purpose of comparison with experiment. Though the precise linac information about the treatment head was not made available to us due to manufacturer's proprietary rights and commercial secrecy, there was good agreement between simulation and experiment for both in-air and in-phantom to within 15% and 20%, respectively. The fractional neutron dose equivalent to radiosensitive organs of the patient during high-energy photon treatment was determined using the tissue equivalent phantom BOMAB compatible with MCNP. In a design study, MCNP simulation of a linac bunker was undertaken to determine the effect on the total neutron flux and dose at the exit of the maze in terms of treatment room modification and in cladding the maze with neutron absorbing materials. The neutron spectrum of an isotropic 252Cf source was used for the purposes of simulation to represent the unfiltered neutron component of the linear accelerator beam and was placed at the SSD. The modification of the treatment room reduced the total neutron flux and dose by approximately 40 and 45%, respectively, whereas the addition of neutron absorbing materials resulted in further reduction of approximately 90%.