School of Physics and Astronomy

Helen Duncan


Research Student

Telephone: 020 7882 6568
Room Number: G. O. Jones Building, Room 214


Research Interests:

My research interests

My work focuses on modelling local order in organic ferroic materials using total neutron scattering and the reverse Monte Carlo (RMC)  method. Three materials have been investigated using this technique, two metal-organic frameworks, dimethylammonium manganese formate (DMMnF) and potassium imidazolium hexacyanoferrate (PIH) and the high-temperature ferroelectric triglycine sulfate (TGS).

The RMC method is used to produce atomic configurations of materials which are consistent with observed data. In this case the data are Bragg profiles, total neutron scattering data (S(Q)) and the associated pair distribution functions (PDFs). Atoms in the configuration are moved at random by a random amount, and the model's Bragg profile, S(Q) and PDF are calculated. If the atomic move improves agreement between the model and the observed data, that move is accepted and another atom is moved. If the atomic move doesn't improve agreement between model and observation, that move may still be accepted based on a probability distribution, this ensures the configurations do not get stuck in a local minimum. Atoms are moved at random and the agreement re-calculated until no more improvement to fit can be made. It is these conveged atomistic models which are then analysed.

In DMMnF, manganese2+ linked by formate bridges form a cage-like structure inside of which the dimethylammonium (DMA) cation rotates about the C-C axis. Upon cooling DMMnF displays first electric, and then magnetic ordering. Strain analysis reveals the framework contracts around the DMA moiety during this process. RMC modelling has shown that in both the electrically disordered and ordered states, the DMA cation forms two short hydrogen bonds (N...O) with the framework which are not observed in the average structure. The shortened bonds result from deformation of the framework to accommodate the cations. As the material cools further the shorter of the two bonds becomes less favourable, possibly due to the increasing distortion on the framework.

DMA cation surrounded by the formate framework, and the separation between the nitrogen of the DMA cation and oxygen of the framework as the material cools

Left: DMA cation surrounded by the formate framework. Right: The partial pair distribution functions - derived from RMC refinement - for the nitrogen of the DMA cation and the oxygen of the formate framework (gNO(r)) as a function of temperature. The dashed lines show the expected pair separations from the average structure in the high and low temperature phases. We can see that the N...O bonding scheme is more complicated than the average structure implies.

The second material Potassium Imidazolium Hexacyanoferrate (PIH) displays a first-order phase transition with accompanying change in crystal symmetry at 158 K, however there are also indications of a second order transition at 187 K. There is no change in crystal symmetry at the 187 K transition, however strain analysis, based on the average structure supports the permittivity and heat capacity measurements which suggested the transition.

Potassium Imidazolium Hexacyanoferrate, an example of a double perovskite metal-organic framework

The perovskite-like structure of PIH. Potassium and iron alternate on the B site forming with KN6 and FeC6 octahedra. The five-membered imidazolium cation is free to rotate within the cavity.

The crystal symmetry for all PIH structures above 158 K requires that the disordered 5-membered imidazolium ring be modelled as an ordered 6-membered ring. Even when the occupancy and scattering lengths are corrected for, the best Rietveld fit is still far from ideal, and disorder had to be introduced into the initial configurations before any RMC refinement could be carried out.

From left to right, the Bragg profile for PIH at 165 K when the disordered imidazolium is modelled as 1) an ordered benzene ring 2) disordered planar imidazolium 3) disordered imidazolium

From left to right, the calculated Bragg profile fit for PIH when the disordered imidazolium are modelled as 1) Ordered planar benzene. 2) Disordered planar imidazolium. 3) Disordered imidazolium accompanied by framework distortion

Analysis of configurations in the high- intermediate- and low-temperature phase found that during the intermediate temperature phase, the framework stiffens around the imidazole ring. This stiffening is seen by narrowing of the Fe-C-N and K-N-C angle distribution in the intermediate temperature phase..

PIH Fe-C-N bond angle distribution for the three phases

Distribution of the Fe-C-N bond angles as extracted from the configurations (data points on the left) and with their corresponding Gaussian fit on the right. In the intermediate temperature phase (T = 180 K and 165 K) there is a significant narrowing of the distribution, indicating a stiffening of the framework.

The final material is Triglycine Sulfate (TGS). TGS is one of the first discovered ferroelectrics. TGS differs from the two MOFs in that its ordered phase spears at room temperature (TC ≈ 49°C/342K). Due to its early discovery, it has been well studied, however the mechanism for the dielectric phase transition is not fully understood. In the high temperature phase the NH3 head of glycine molecule GI wags between two positions, and an intermediate hydrogen between glycine GII and GIII moves between both, with an average position midway between. As TGS undergoes the phase transition the NH3 heads preferentially orient in one direction, and the intermediate hydrogen will preferentially move towards either GII or GIII. The purpose of this study was to look for correlations between the orientation of the GI head with hydrogen bond formation in the nearby GII and GIII.

Triglycine Sulfate TGS. Glycine I shown in red, GII in blue and GIII in green. The inserts show the motion of the intermediate deuterium and the orientations of the GI head

Unit cell of  TGS. Glycine I (GI) is shown in red with the disorder of the NH3 head indicated by the oval inset. GII and GIII are shown in blue and green respectively. The intermediate hydrogen (square inset) preferentially forms a short bond with GII in the low temperature phase.

Atomistic configurations of TGS were generated above and below the phase transition, and the relative orientations of the GI head and the closeness of the intermediate deuterium to either GII or GIII calculated. Each configuration was approximately 50 Å x 50 Å x 50 Å which produced chains of 8 GI GII/GIII pairs which were investigated. Despite collecting data to Tc + 70 K, all configurations showed the GI head overwhelmingly pointing in one direction, indicating that even in the high temperature phase TGS forms domains of aligned GI heads at least the size of the RMC simulation box.


Published papers

Local structure of the metal–organic perovskite dimethylammonium manganese(II) formate
Duncan HD, Dove MT, Keen DA, Phillips AE
Dalton Trans., 2016,45, 4380-4391
DOI: 10.1039/C5DT03687A