MOSFET dosimetry with high spatial resolution in intense synchrotron-generated x-ray microbeams E. A. Siegbahn, a E. Bräuer-Krisch, and A. Bravin European Synchrotron Radiation Facility (ESRF), 6 Rue Jules Horowitz, 38043 Grenoble, France H. Nettelbeck, M. L. F. Lerch, and A. B. Rosenfeld Center for Medical Radiation Physics, University of Wollongong, Wollongong, New South Wales 2522, Australia Received 21 July 2008; revised 26 January 2009; accepted for publication 27 January 2009; published 11 March 2009 Various dosimeters have been tested for assessing absorbed doses with microscopic spatial reso- lution in targets irradiated by high-flux, synchrotron-generated, low-energy 30–300 keV x-ray microbeams. A MOSFET detector has been used for this study since its radio sensitive element, which is extraordinarily narrow 1 m, suits the main applications of interest, microbeam radiation biology and microbeam radiation therapy MRT. In MRT, micrometer-wide, centimeter- high, and vertically oriented swaths of tissue are irradiated by arrays of rectangular x-ray micro- beams produced by a multislit collimator MSC. We used MOSFETs to measure the dose distri- bution, produced by arrays of x-ray microbeams shaped by two different MSCs, in a tissue- equivalent phantom. Doses were measured near the center of the arrays and maximum/minimum peak/valley dose ratios PVDRs were calculated to determine how variations in heights and in widths of the microbeams influenced this for the therapy, potentially important parameter. Monte Carlo MC simulations of the absorbed dose distribution in the phantom were also performed. The results show that when the heights of the irradiated swaths were below those applicable to clinical therapy 1 mm the MC simulations produce estimates of PVDRs that are up to a factor of 3 higher than the measured values. For arrays of higher microbeams i.e., 25 m  1 cm instead of 25  500 m 2 , this difference between measured and simulated PVDRs becomes less than 50%. Closer agreement was observed between the measured and simulated PVDRs for the Tecomet ® MSC current collimator design than for the Archer MSC. Sources of discrepancies between measured and simulated doses are discussed, of which the energy dependent response of the MOS- FET was shown to be among the most important. © 2009 American Association of Physicists in Medicine. DOI: 10.1118/1.3081934 Key words: x-ray, microbeam, radiation therapy, MOSFET, dosimetry, MRT I. INTRODUCTION The unique features of synchrotron radiation enable intense monochromatic or polychromatic a spectrum of energies x-ray beams to be used in biomedical experiments, some of which are aimed toward radiotherapy. 1 This study focuses on the dosimetry of microbeam radiation therapy MRT. 2,3 In MRT, an array of rectangular x-ray beams, “microplanar” beams, is used to irradiate selected biological targets. A typi- cal microbeam array of area of 1  1 cm 2 contains 50 beams; each of these has a typical size of 1 cm  25 m height  width, separated by a center-to-center ctc spac- ing of 200 m. The minimum dose in the central region between two microbeams is called the “valley dose.” The peak and valley dose ratios PVDRs in the irradiation field are believed to be of importance for the therapeutic effect of the treatment. Figure 1 presents a lateral dose profile in tis- sue, typical for MRT, extracted along a line in the phantom transverse the beam direction, where the peak and valley doses which need to be measured are indicated. In the pre- clinical in vivo MRT trials, these arrays irradiate the target with either unidirectional or cross firing techniques. 4 In ra- diotherapy, there are strict requirements on the precision with which the delivered dose must be known. 5 Even if the accu- racy requirements in standard hospital based radiation therapy 3%Ref. 5 could be relaxed somewhat for MRT, it would still be the physicist’s responsibility to ascertain the dose as accurately as possible before therapy is implemented. The x-ray source properties and the radiation interaction probabilities in materials of interest can be used to simulate the absorbed dose or it can be derived from measurement. Accurate knowledge of the absorbed dose in tissue in MRT pose unique challenges: the beam sizes used are small which means that the detecting element must be small, ideally of micron or submicron size; the x-ray energies used are in the interval where many of the commonly used detectors have an energy dependent response. Furthermore, the microbeam dose deposition should also be measured at distances many beam widths away from the direct irradiation field more than 1000 m away from the field border which requires a wide dynamic range in the dose detection. Since the evolution of MRT, the production of planar beams has been investigated with various collimator designs. In the early MRT trials, a single slit was used to collimate the 1128 1128 Med. Phys. 36 „4…, April 2009 0094-2405/2009/36„4…/1128/10/$25.00 © 2009 Am. Assoc. Phys. Med.