Simple and Robust Algorithm for Unfolding Hard X-ray Spectra from Filter Stack Spectrometers*

The Filter Stack Spectrometer (FSS) is an instrument that measures x-ray spectra using alternating filters and image plates (IPs). The FSS has several attractive features: it is an easy-to-field, portable, and passive diagnostic capable of measuring hard (10s of keV to 10s of MeV) x-ray spectra for radiographic systems and HED physics experiments. However, unfolding x-ray spectra from the IP signals is an ill-posed problem requiring physics principles to constrain the unfold. Strong constraints can enforce a unique solution but can also strongly impose preconceptions onto the inferred x-ray spectrum.

Perturbative minimization (PM) has been shown to be an effective method for unfolding x-ray transmission data (without scatter), with the restriction that filter attenuation coefficients monotonously decrease with x-ray energy [1,2]. Here, we extend the PM method to unfold x-ray spectra from FSS data (which includes scatter) without the previous restriction and without a physics-informed initial guess. We apply the algorithm to unfold a set of synthetic spectra with known ground truths, as well as experimental spectra obtained at OMEGA EP and the Texas Petawatt Laser Facility.

References:

[1] Waggener et al., Med. Phys. 26, 1269-1279 (1999).

[2] Iwasaki et al., Rad. Phys. Chem. 67, 81-91 (2003).