Do we need to measure quasielastic neutron scattering, or is there a better way to measure the underlying dynamics?

General ILL Seminar organised by College 1
Wednesday, 4 September 2019 at 11:00
Seminar room (no 163), ILL 4

Antonio Benedetto & Don Kearley

University College Dublin - Ireland

Diffusive dynamics of species (atoms, molecules or molecular fragments) within a sample are usually characterised via quasielastic neutron scattering (QENS), which is the envelope of scattering in the energy-spectrum around the elastic peak. Paradoxically, the main interest in this type of dynamics is in the relaxation processes that occur time-domain, and over the years methods sophisticated methods for extracting this information from the QENS energy-domain have evolved. So why not measure in the time domain directly, as neutron spin-echo (NSE) does? NSE measures the signal at the elastic-scattering condition, and over the past decade others [ref.1 and references therein] have noted that even without spin-echo there appears to be a decrease in elastic scattering as the instrumental resolution is increased, which resembles the relaxation, I(t). Whilst this has even been equated to I(t), it turns out that this apparent similarity is totally misleading.

Almost by accident, we realised that under the correct instrumental conditions, the elasticscattering technique actually measures the running-integral of the relaxation function, I(t), and the derivative of this provides the desired relaxation function directly [ref. 2]. The crucial point is remarkably simple, and entails moving away from the conventional matching of primary and secondary resolutions. In many ways this result seems too good to be true: it avoids Fourier transforms and inherently matches the resolution to that required for the measurement being made. We compare the results for a numerical simulation “under the same conditions” for QENS and the elastic scattering spectroscopy method in the figure below. Clearly, they are similar, but QENS requires fitting and/or Fourier transform to obtain I(t). We will show how this method works and welcome discussion of its potential advantages, limitations and validation.

1. Doster, W. Nakagawa, H. & Appavou, J. Chem. Phys. 139, 045105 (2013).
2. Benedetto, A. & Kearley G. J. Nature Scientific Reports (in press)

Sandra Cabeza
College 1 Secretary

External visitors may ask for a site access to dubouloz(at)ill.eu