Doctor of Philosophy
Department of Engineering Physics - Faculty of Engineering
Carolan, Martin G, Semiconductor dosimetry of epithermal neutron beams for Boron neutron capture therapy, PhD thesis, Department of Engineering Physics, University of Wollongong, 2003. http://ro.uow.edu.au/theses/158
Boron neutron capture therapy (BNCT) is a binary targeted therapy that uses suitably designed pharmaceuticals to deliver (superscript 10)B to tumor cells. The region is then irradiated with neutrons and neutron capture by the (superscript 10)B nucleus leads to the emission of an alpha particle and lithium ion. These have very short ranges similar to the dimensions of a biological cell and therefore the technique could have potential for selective killing of tumour cells. In order to achieve adequate neutron fluxes at the site of the tumour epithermal (0.5 eV � 10s of keV) neutron beams are used. A review of the general details of BNCT is presented in this thesis. This thesis investigates the use of two semiconductor devices for measuring the neutron and gamma dose components involved in epithermal neutron beams used for BNCT. The silicon lattice in PIN diodes undergoes displacement damage when irradiated with neutrons. This leads to a change in the forward bias voltage of the diode that is proportional to the neutron dose received. To verify that the energy dependence of this effect follows the published silicon displacement damage KERMA (Kinetic Energy Released per Mass of Absorber) data measurements were performed using quasi-monoenergetic neutrons obtained from a Van de Graff accelerator (Ansto) in the energy range from 90 keV � 890 keV. These measurements were in agreement with the published data for silicon displacement damage KERMA. A sensitivity factor for the diodes was also derived from these measurements. The thermal neutron sensitivity of the PIN diodes was then determined using the TC-10 thermal neutron column on the Moata reactor at Ansto. The sensitivity results were in general agreement with the Van de Graff derived sensitivity factor. Since the silicon damage KERMA is not the same as the tissue KERMA function PIN diodes are not intrinsically tissue equivalent. A Monte Carlo (MNCP) ideal beam study was undertaken to see if for some limited energy range tissue dose could be parameterised in terms of silicon damage dose and foil activation. This was found to be approximately true for neutron energies from thermal to 100 keV. Coefficients are given that allow tissue dose to be determined on the basis of a single PIN diode and activation foil measurement in spectra where the maximum neutron energy is 100 keV or less. MOSFETS can be used as gamma radiation dosimeters by measuring the change in threshold voltage (simplistically understood as the potential applied to the gate to initiate current flow from source to drain electrodes) that occurs when they are exposed to radiation. The MOSFETs used in this study were characterised using a Varian 2100C medical linac beam and low energy x-rays from a superficial x-ray unit. The sensitivity of the MOSFETs was measured with different potentials applied to the gate during irradiation. Depth dose profiles in 6 MV x-ray beams were measured and found to be in good agreement with both ionisation chamber measurements and MCNP simulations. This good agreement was also obtained for the buildup region. Although the silicon oxide layer of the MOSFET is not intrinsically very sensitive to neutron irradiation the presence of encapsulating materials leads to the generation of secondary photons and electrons which lead to shifts in threshold voltage and therefore confound gamma ray measurements in mixed neutron / gamma fields. To determine the energy dependant neutron response function of the MOSFET a detailed MCNP simulation was used. A lithiated shield was also incorporated into this model. The calculated neutron response functions were used to correct for neutron contributions to MOSFET measurements in mixed fields. MOSFET thermal neutron responses were measured using a series of measurements with MOSFETS both with and without lithiated covers exposed in the the Moata thermal neutron column at Ansto. The measurements were repeated with various gate potentials. The gamma doses measured were consistent with gamma doses measured using paired ionisation chambers. The Petten HB11 facility is briefly described as are phantoms and MCNP models fabricated by S Wallace for an associated work. Measurements using PIN diodes and MOSFETs in phantoms exposed in the HB11 beam are described. Foil activation data is compared to MCNP calculations to validate the MCNP models used. This thesis presents a number of results that have been recalculated in more detail and with various parameters changed. In particular the effect of variations in phantom hydration have been incorporated as have response functions for MOSFET detectors and associated shields. PIN diode measurements in a Perspex cube phantom exposed in the HB11 beam show good agreement with MCNP calculated silicon displacement dose. Similarly good agreement is obtained for a cylindrical phantom filled with tissue equivalent gel when the hydrogen content of the gel in the original model is corrected for dehydration. Measurements in a more complex skull phantom show larger discrepancies between the experimental results and a MCNP simulation especially at depth. The discrepancies range from 25 � 300% in absolute terms but are only 2-3% of the maximum silicon dose. MOSFET measurements performed in the Perspex cube phantom using lithiated covers show excellent agreement with ionisation chamber measurements (also with lithiated covers). Measurements in a cylinder phantom and head phantom using lithium/perpex covered MOSFETs are compared with Monte Carlo calculations of induced gamma dose. In this case the measured gamma doses at approximately 2 cm depth appear to be too low. Further investigation involving a detailed MCNP simulation including the lithiated MOSFET covers in the model indicated that the covers suppress the thermal neutron flux at the measurement point and therefore the gamma dose is also reduced. Measurements at greater depths show a similar effect but to a lesser extent. The epithermal neutron beam at the Brookhaven Medical Research Reactor (BMRR) is described is described. PIN diode and MOSFET measurement results in a Perspex cube phantom are also presented. Reasonable agreement between calculated and measured PIN diode results is observed. MOSFET measurements show good agreement with the known percentage depth dose curve for the total gamma dose. However there is a discrepancy in the absolute magnitude of the measured gamma doses. It is proposed that this is also due to thermal neutron flux depression arising from the use of relatively thick lithiated neutron shields around the MOSFETS. In summary; It is demonstrated that PIN diodes could be useful for verifying treatment planning dose distributions in epithermal neutron beams. This includes the possibility of on line real-time measurements. They could also be used in conjunction with an activation foil to yield tissue equivalent dose measurements where the maximum neutron energy is less than 100 keV. Lithium shielded MOSFET measurements can be reconciled with calculated gamma dose distributions when the effect of flux depression is taken into account. However the perturbations introduced by the shield mean that the measured dose does not represent the dose at the measurement point in the absence of the shield. In order to use MOSFETs for gamma dosimetry in epithermal neutron beams different encapsulation is required to minimise neutron response and eliminate the need for lithiated covers. It is suggested that MCNP simulations of MOSFETs similar to the models in this thesis would provide an adequate tool for optimising the appropriate encapsulation.