Abstract
The electron paramagnetic resonance (EPR) alanine dosimetry system is based on EPR measurements of radicals formed in alanine by ionizing radiation. The system has been studied to determine its energy dependence for photons in the 10–30 MV region relative to those of 60Co and to find out if the system would be suitable for dosimetry comparisons. The irradiations were carried out at the National Research Council, Ottawa, Canada and the doses ranged from 8 to 54 Gy. The EPR measurements were performed at the University of Oslo, Norway.
The ratio of the slope of the alanine reading versus dose-to-water curve for a certain linac photon beam quality and the corresponding slope for a reference 60Co γ-radiation gives an experimental measure of the relative dose-to-water response of the EPR alanine dosimetry system. For calculating the linear regression coefficients of these alanine reading versus dose curves, the method of weighted least squares was used. This method is assumed to produce more accurate regression coefficients when applied to EPR dosimetry than the common method of standard least squares. The overall uncertainty on the ratio of slopes was between 0.5 and 0.6% for all three linac energies.
The relative response for all the linac beams compared to cobalt was less than unity: by about 0.5% for the 20 and 30 MV points but by more than 1% for the 10 MV point. The given standard uncertainties negate concluding that there is any significant internal variation in the measured response as a function of beam quality between the three linac energies. Thus, we calculated the average dose response for all three energies and found that the alanine response is 0.8% (±0.5%) lower for high energy x-rays than for 60Co γ-rays. This result indicates a small energy dependence in the alanine response for the high-energy photons relative to 60Co which may be significant. This result is specific to our dosimetry system (alanine with 20% polyethylene binder pressed into a particular shape) including its waterproofing sleeve of PMMA (2 mm thick); however, we expect that this result may apply to other similar detectors.