Speaker
Description
Neutron and x-ray scattering experiments traditionally rely upon
histogrammed data sets, which are analysed using least-squares curve
fitting of multiple probability distribution components to quantify
separately the various scientific contributions of interest. The main
advantage to this approach is the relative ease of deployment due to
its intuitive nature. Despite the great popularity of the method,
there are known drawbacks relative to alternative methods, such as
systematic errors, biases, and instability in some scenarios that are
common in neutron scattering. Improvements over the base methods
include dynamic optimisation of histogram bin width and the
application of modern numerical optimisation methods that can be less
unstable when pushed to the edge of the performance envelope.
In this new study, we demonstrate analysis of neutron scattering data
entirely on an event-by-event basis, without resorting to any kind of
numerical integration, histogramming, or least squares fitting. This
method is demonstrated first using synthetic events for a standard
distribution, to establish the method in a controlled environment;
then in synthetic small angle scattering application, with more
realistic features; and ultimately deployed on event data measured by
the ARCS spectrometer at the Spallation Neutron Source, TN, USA. This
wide range of tests also shows the broad applicability of this method.
The benefits of this approach are revealed: orders of magnitude
greater efficiency (i.e. fewer data points required for the same
parameter accuracy). The efficiency gain can be viewed as either a
significant increase in scientific output, or equivalently reaching
the finest possible time resolution in kinetic studies. Considering
the cost of neutron scattering beam time, it may also result in
significant operational cost savings. The main drawbacks are an
increase in computation time, and perhaps a less intuitive analysis
method for some users.
