Quantifying signal quality in scanning transmission X-ray microscopy

While the general effects of experimental conditions such as photon flux and sample thickness on the quality of data acquired by scanning transmission X-ray microscopy (STXM) are widely known at a basic level, the specific details are rarely discussed. This leaves the community open to forming misconceptions that can lead to poor decisions in the design and execution of STXM measurements. A formal treatment of the uncertainty and distortions of transmission signals (due to dark counts, higher-order photons and poor spatial or spectral resolution) is presented here to provide a rational basis for the pursuit of maximizing data quality in STXM experiments. While we find an optimum sample optical density of 2.2 in ideal conditions, the distortions considered tend to have a stronger effect for thicker samples and so similar to 1 optical density at the analytical energy is recommended, or perhaps even thinner if significant distortion effects are expected (e.g. lots of higher-order light is present in the instrument). (Note that X-ray absorption calculations based on simple elemental composition do not include near-edge resonances and so cannot accurately represent the spectral resonances typically employed for contrast in STXM.) Further, we present a method for objectively assessing the merits of higher-order suppression in terms of its impact on the quality of transmission measurements that should be useful for the design of synchrotron beamlines.