A novel method to assess the incident angle and the LET of protons using a compact single-layer Timepix detector

Particle therapy can largely benefit from the detailed and wide-range spectrometric and directional characterization of energetic charged particles provided by compact Timepix detectors. Among several physical quantities that can be derived, the assessment of the linear energy transfer (LET) which is based on the deposited energy and particle's track length remains challenging. Due to the detector's pixel pitch, sensor thickness and charge sharing effect, an accurate estimation of the particle's incident angle and hence the track length, has been limited to particles with incident angles greater than 20 with respect to the normal of the sensor layer. This is critical for clinical beams which are highly directional, and measurements with radiation detectors are generally performed with sensitive volumes orthogonally placed with respect to the beam direction. In this work, we present a novel method in which we exploit the morphological cluster parameters to derive the proton's incident angle, thus enabling a precise directional reconstruction over the full field-of-view 2 pi (solid angle), and within 2 from the reference angles for Timepix detectors with 300 and 500 mu m thick Si sensors. As a consequence, the calculation of the track length was also improved, resulting in a more precise LET estimation. The experimental LET spectra and the frequency-averaged LET (LETF) were compared against Monte Carlo simulations using TOPAS for a wide range of proton energies (12 MeV-200 MeV) and incident angles (0-85 & DEG;). An agreement within 12% was found between measured and simulated LETF. A comparison with LET values based on the PSTAR database also showed an agreement within 10%. We demonstrated the feasibility of a precise LET calculation and directional response with an improved angular resolution down to normal incidence using a single-layer Timepix detector, while avoiding the use of a stacked telescope array.