Evaluation of a simple graphite calorimeter for ultra-high dose-rate electron beam dosimetry

Objective.This study evaluates the performance of a simple secondary-standard graphite calorimeter (SSCal) for absorbed-dose measurements in ultra-high dose-rate (UHDR) electron beams, motivated by the need for reliable, traceable dosimetry for FLASH radiotherapy.Approach. The SSCal was calibrated against the primary absorbed-dose standard at the National Physical Laboratory (NPL) and characterised in conventional 6-10 MeV electron beams. Monte Carlo (MC) simulations were used to derive detector-specific perturbation and field-size correction factors, including conversion from absorbed dose to graphite to absorbed dose to water. The SSCal was then tested in a 9 MeV UHDR beam (0.06-5.78 Gy pulse^-1) at the Centro Pisano Flash Radiotherapy facility. Absorbed dose was determined using two independent approaches: (1) application of the calibration coefficient derived experimentally at NPL, and (2) first-principles calorimetry. Alanine dosimetry provided an independent traceability route for cross-validation.Main results. Calibration coefficients obtained at NPL showed minimal dependence on field size or beam quality. MC simulations indicated that perturbations were dominated by internal air gaps, with increasing magnitude at small field sizes. Corresponding beam quality correction factors enabled accurate application of the NPL calibration to the UHDR beam. Across all dose-per-pulse (DPP) values, the two dose determination methods agreed within 1%. The SSCal exhibited excellent linearity with pulse number and stable response for three pulses or more. A systematic ΔT/MU trend was observed only for single- or double-pulse exposures, for which transient perturbations in the beam current transformer signal provide a plausible explanation. In addition, or alternatively, the observed reduction in detector response may be influenced by non-linear heat-transfer effects within the calorimeter, potentially associated with cumulative heating and delayed thermistor equilibration. Agreement between SSCal and alanine was within ±1% for all but the highest DPP (5.71 Gy pulse^-1), where a 2.6% deviation was observed under low-pulse conditions, which is consistent with the ΔT/MU trend.Significance. The results demonstrate that the SSCal provides stable, linear, and largely dose-rate-independent absorbed-dose measurements in UHDR electron beams when appropriate MC-derived corrections are applied, while indicating that a potential deviation from ideal response cannot be excluded at the highest investigated DPP values (>5 Gy pulse^-1). These findings support suitability of SSCal as a reference-quality dosimeter for FLASH research and development.