k0 instrumental neutron activation analysis (k0-INAA) requires high-purity Germanium (HPGe) gamma detectors, with their absolute full-energy peak efficiency (FEPE) known for as large an energy range as possible. The standard procedure for determining the FEPE is by using reference point gamma sources with known absolute activities, which yield the FEPE at a few discrete energies. Afterwards, polynomials are fitted to different energy regions of the efficiency data, with care taken to minimize the polynomial’s discontinuities at region boundaries. Due to the choice in number and degree of polynomials and their energy regions, this is a subjective approximation, inaccurate in regions where few data points are available, especially outside the region between the smallest and largest calibration energy.

We improve this by simulating a simplified HPGe detector and gamma source, using the Geant4 simulation toolkit. It performs a full physical Monte-Carlo simulation of the passage of particles through matter, yielding a simulated energy-efficiency function. This function has the desired general shape, but is inaccurate due to simplifications in the geometrical and physical models used for the simulation. We correct this by fitting the simulated efficiency curve to experimental data, analogously to how we fitted polynomials in the standard calibration procedure, but with fewer degrees of freedom. This gives a more accurate and less subjective model of efficiency than possible from a simple polynomial fit. Because we use a reasonably accurate physical model for the prediction, we also have greater confidence in the determined efficiency outside the extremal calibration energies.

Our method thus extends the energy range of HPGe calibration, and enables a better trade-off between the number of calibration sources used, and the accuracy of the determined FEPE. The latter is particularly useful for energy regions where few or no such sources are even available.