Approaches to modelling thermal diffuse scattering in triglycine sulphate
Science post alert!
Coming up in Journal of Applied Crystallography soon will be our latest opus; ‘Approaches to modelling thermal diffuse scattering in triglycine sulphate, (NH2CH2COOH)3.H2SO4‘ by Jessica Hudspeth, D. J. Goossens and T. R. Welberry. It is the latest in a line of papers extending back into the mists of time, exploring the short-range order in crystalline materials. Usually crystallography delivers the average structure, as described by a unit cell. This is like a little box or prism which contains atoms in a particular arrangement. These cells are then stacked up like bricks to make the macroscopic crystal. In a real crystal, not every cell is identical. We have been working on using the diffuse scattering, that is the broad distribution of scattered X-rays (or neutrons) as well as the Bragg reflections, to explore the structure. The Bragg reflections give you the average building block, the unit cell, and the diffuse scattering gives you some information on the local variations from one cell to the next.
Triglycine sulphate, or TGS, is a ferroelectric material that has a phase transition at about 50oC; below that it is ferroelectric, above it is not. Below this critical temperate, TC,one of the glycine molecules points in a particular direction (some atoms get displaced), and neighbouring ones point the same way, and the net result is an electrical polarisation of the crystal, or at least of a small domain within the crystal. We have explored this in some previous work (link here). When this transition is incomplete, the system consists of small linear domains of glycines with like displacements. This gives rise to small, plate-like regions of scattering in the diffraction pattern. And here’s a picture:
So these little bands of scattering do not exist far above TC where there is no ferroelectric order, or far below TC, where the ferroelectric ordering is fully three dimensional, but only at temperatures close to TC.
This most recent work explores the modelling of the diffuse scattering in TGS using several different methods, in particular different ways of setting up the forces between molecules as they interact and buffet each other back and forth in the crystal. We find that a parameterisation method proposed in some earlier work, link here, proves the most effective at modelling the diffuse scattering. The full reference for this work will be retroposted as it comes to hand.