Profile

ISIS attracts scientists from all over the world to perform neutron scattering experiments.  These experiments reveal information about the microscopic and atomic scale structure of their samples in ways often inaccessible by any other technique.

The scientists that I work with are world leaders in their areas of expertise.  However, neutron scattering is a field and expertise in its own right.  My goal is to write software that will help these scientists to analyze the results of their experiments without having to become experts in neutron scattering.

The particular program that I’m working on right now is called Dissolve, which is short for Disordered Solver. It’s called that because it’s used for studying disordered materials, which is anything where there isn’t a regular arrangement to the molecules.  Any liquid or gel, where the atoms can freely move around, is an example of a disordered material.

A scientist tells our software what molecules were in their sample and what kinds of forces are acting between those molecules.  The software then simulates a small sample of those molecules and calculates what kind of neutron scattering would be expected.  The software then compares the calculated scattering with the what the scientist measured.  The software then starts adjusts the forces to make the calculated scattering closer to the measured experiment.

Eventually, the calculated scattering will be very close to the scattering that the scientist measured and the software looks at the arrangement of molecules that produces this scattering pattern.  The fact that the pattern matches the experiment doesn’t mean that the atoms in the actual experiment.  First off, the actual experiment will have trillions of times more atoms than our software could ever handle.  Furthermore, there’s a near infinite number of ways to arrange the atoms that will produce the same scattering pattern.  However, the configuration of the molecules at the end of the simulation will be closer to the arrangement in the actual sample than what the scientist had when they started.  This is still useful for allowing the scientist to look at how the molecules are interacting, the shapes that they form, and what forces develop between them.

To help improve the final arrangement that the software produces, the scientists often perform multiple measurements on the same sample.  For example, then might do one sample scattering neutrons and another scattering x-rays.  Or they might make a new sample the replaces all of the hydrogen with deuterium (which is like hydrogen with an extra neutron attached to the nucleus).  Using deuterium in place of hydrogen changes almost none of the chemistry, but massively changes the neutron scattering pattern.  As I said earlier, there’s an infinite number of molecule arrangements that will match any one pattern.  However, most of those will not match for two different experiments.  Thus, the more measurements that they can perform, the closer the final configuration of molecules will be to the one that actually occurred in the experiment.