A benchmark is an integral experiment which consists of a measurement of a charged-particle spectrum from a well characterized uniform thick target followed by a standard direct simulation using microscopic cross-section data in order to validate the data. This is an extension of the definition taken from nuclear reactor physics where microscopic neutron data are validated by comparison of calculated integral reactor characteristics such as neutron flux with results of direct measurements. The idea of benchmarking is to verify the validity of the data for the purpose they are intended for through their use in a typical application.
A feature of the cross-section validation against a benchmark spectrum (if it exists) has been developed in SigmaCalc. The spectrum simulation is performed in SigmaCalc taking straggling (Bohr's theory) on the inward and outward pathes into account separately in order to accurately reproduce effects of the cross-section fine structure if any. The simulated spectrum is calculated as a convolution of the cross-section with a spreading function using SRIM-2003 stopping power. Due to secondary effects such as multiple and plural scattering the comparison between simulated and measured spectra has significant importance mainly in a relatively narrow region near the high energy edge of the spectrum. Consequently in order to obtain significant results the cross-section energy range should extend to the energy exceeding the benchmark spectrum energy.