Crystallization is a complex process of fundamental importance to materials science, climate science, pharmaceutical production, food science, and beyond. State-of-the-art simulation work has pursued a molecular level understanding, by using powerful new simulation methods such as metadynamics, forward flux sampling and transition path sampling (Bolhuis, 2015, Peters 2015). These methods have been used with accurate force fields for studies of nucleation and growth processes of key systems including water (Lupi 2014), and calcium carbonate (Wallace 2013).

In experiment the state of the art is to combine quantitative induction time measurements with rapidly developing technologies such as microfluidics (Iledefonso 2013), electron microscopy (Nielsen 2014), and confocal microscopy. For example, confocal microscopy allows colloidal experimentalists to study systems at the single particle level. They can study designed particles crystallising into tunable architectures (Henzie et al 2012). But despite these advances, many important challenges remain.

In terms of quantitative experimental data and the maturity of our understanding, the state of the art in ice crystallisation is ahead of that in other fields (Sear 2014). For example, in many other crystallisation fields there is often debate as to whether nucleation occurs in the bulk (homogeneous nucleation), or occurs in contact with impurity. But it is now well established that at temperatures above approximately -35 C, ice only nucleates due to impurities (heterogeneous nucleation) (Hoffmann 2013, Sanz 2013).

However, for most crystallising substances, theory and simulation even struggle to predict trends in nucleation and growth rates, as variables such as composition and temperature are changed. A particularly problem is the important case of molecules that crystallize into multiple crystal lattices, called polymorphs. While methods to predict the existence and stability of polymorphs haveimproved, we have essentially no ability to predict the kinetics of polymorph selection (Little 2015). It remains difficult to disentangle the roles of nucleation and growth kinetics in determining the polymorph.

Another challenge comes from the countless surfaces and impurities that complicate nucleation and growth, in industrial crystallization and in the natural environment. Here the challenge is to provide insight into the most potent sites for nucleation, and into the components of the solution that most strongly impact growth. Here there are recent efforts aimed at practical quantitative models based on extreme value statistics (Sear 2014), and an increasing number of simulation studies of nucleation at a surface (Lupi 2014).

A workshop is the perfect environment to bring computer simulators and experimentalists together to work towards a consensus on the key problems in crystallisation, and how we can solve these problems. Crystallisation is important in many scientific and engineering fields but this diversity can dilute effort on solving key fundamental crystallisation problems. Our workshop will bring together leading scientists studying a wide range of systems. Scientists studying ice nucleation and those studying drug molecules or ions rarely meet to exchange ideas, and simulators and experimentalists often attend different meetings. The workshop will bring all these researchers together.

We believe that collaboration between computer simulation and experiment is often too weak. Present-day computer simulations, although commonly using validated potentials, often ignore features known to be key in experiment, such as surfaces for the heterogeneous nucleation of crystals, and the role of defects in crystal growth. Experiments on crystallisation are too often qualitative and lack time-resolved information. This workshop will provide guidance as to what kind of quantities need to be measured in both experiments and simulations.

The central idea of this workshop is to bring computer simulators and experimentalists together to build teams. The workshop will achieve this by focusing on 4 targeted topics: T1) Crystallisation of mixtures, T2) Crystallisation of molecules from solution, T3) Crystallisation in ionic systems, T4) Ice.

T1) Crystallisation of mixtures. There are crystallisation challenges that are unique to mixtures. For example, compositions that vary in space and time, hinders the formation of crystals. Also, almost all good glass formers are mixtures. We want to discuss the fundamental challenges here, as well as key application challenges such as how to optimise the crystallising mixtures used to make organic photovoltaic devices.

T2) Crystallisation of molecules from solution. There is now quantitative experimental and simulation data on these important systems, which include most drugs. However, the gap between experiment and simulation is still large. Polymorph control is key in the pharmaceutical industry, but we have essentially zero predictive ability here. We hope to stimulate a combined experiment and simulation attack on this problem.

T3) Crystallisation in ionic systems. Here, one system, calcium carbonate, has been the subject of experiment and simulation collaborations, and so we already know a great deal, but the behaviour is clearly complex. There is considerable evidence for multi-step crystallisation processes in this class of systems, but this evidence is mainly qualitative. We aim to catalyse the combination of quantitative experiments, and simulations, that we think are needed to understand these complex systems.

T4) Ice. The crystallisation of water is the most studied system in experiment, and a system of great importance. Models of ice formation are components of cloud models, and poor models of cloud formation are currently holding back progress in climate modelling. Over the last five years, reliable computer simulation results have become available, but collaboration between simulators and experimentalists is still weak. We will strengthen it.