The paper stems from collaborative work between Christ Church’s Dr Clare Rees-Zimmerman and Professor Dirk Aarts, as well as other academics from the Complutense University of Madrid and Durham University. It explores the behaviour of microscopic particles in suspensions and concerns colloids – microscopic particles measuring between 1 nanometre (1×10−9m) and 1 micrometre (1×10−6m) in diameter that are dispersed in fluids. 

Many everyday formulations consist of colloidal dispersions, from personal care products and pharmaceuticals to fertilisers and paints. Crucial to the function of each formulation is the nature and behaviour of the particles in the dispersion. We can secure a better grasp of the functions of formulations, and determine how the functions might be enhanced, by improving our understanding of how the colloidal particles in them interact. 

As Dr Rees-Zimmerman explains, ‘particles typically strongly repel each other at very short distances (they cannot overlap!) and have no interaction when they are far enough apart.’ The paper explores the exact form of particle interactions as a function of the distance between them. According to Dr Rees-Zimmerman, ‘In industry, such interactions are typically approximated, but it would be much better to be able to measure them: this would further our understanding of the formulations and enable us to improve them.’

The three groups of researchers (in Oxford, Madrid and Durham) each investigated a different potential method for measuring interactions between colloidal particles, each group testing their method on the same data of simulated or experimental particle coordinates. The measurement of interactions is not straightforward. The three teams relied on ‘inverse methods’ for their measurements – that is, they solved the problem by going backwards from the particle coordinates to the underlying interactions that led to those particle positions. In each case, they began with simulated data, where they knew what interaction was input, and saw whether they could get the same interaction out. Dr Rees-Zimmerman then carried out experiments, imaging particles under an optical microscope in the lab: ‘The comparison of the experimental analyses was particularly interesting, as we did not know a priori what the correct answer would be!’

The group in Durham tested a traditional method for measuring particle–particle interactions, serving as a benchmark to compare newer tools against. This involves an initial guess for the interaction, which is iteratively updated through intensive simulations. The Madrid group used a recent machine learning method – again computationally intensive, yet highly adaptable. 

The Christ Church team implemented another recent method: one that again works by updating an initial guess, but in a computationally simpler way. Whilst saving on computing power is better for the environment, there are many other considerations that may recommend one approach to measurement over another: the researchers also compared the methods on their required data inputs and accuracy, proposing in which scenario each method should be used.

The broader objective of the researchers is to better describe and control the behaviour of soft matter both for gaining theoretical understanding and for industrial applications. Tools such as the inverse methods provide greater insight into the behaviour of colloids, and the authors of the JCP paper are now turning their attention from the simpler dispersions considered in their initial study to more complex colloidal systems such as bacteria and mixtures.