Limitations of silicon diodes for clinical electron dosimetry

This work investigates the relevance of several factors affecting the response of silicon diode dosemeters in depth-dose scans of electron beams. These factors are electron energy, instantaneous dose rate, dose per pulse, photon/electron dose ratio and electron scattering angle (directional response). Data from the literature and our own experiments indicate that the impact of these factors may be up to +/-15%. Thus, the different factors would have to cancel out perfectly at all depths in order to produce true depth-dose curves. There are reports of good agreement between depth-doses measured with diodes and ionisation chambers. However, our measurements with a Scantronix electron field detector (EFD) diode and with a plane-parallel ionisation chamber show discrepancies both in the build-up and in the low-dose regions, with a ratio up to 1.4. Moreover, the absolute sensitivity of two diodes of the same EFD model was found to differ by a factor of 3, and this ratio was not constant but changed with depth between 5 and 15% in the low-dose regions of some clinical electron beams. Owing to these inhomogeneities among diodes even of the same model, corrections for each factor would have to be diode-specific and beam-specific. All these corrections would have to be determined using parallel plane chambers, as recommended by AAPM TG-25, which would be unrealistic in clinical practice. Our conclusion is that in general diodes are not reliable in the measurement of depth-dose curves of clinical electron beams.