Themed issue: Modelling of materials

Materials chemistry is crucial to solving the many challenges that face society, from clean energy and emissions reduction, through supply of potable water, to improved health. As the complexity of materials increases and precise, engineered control of dimensions down to the nanoscale is required, computer modelling has a vital role to play as a complement to experiment (Kendrick et al.: DOI: 10.1039/c0jm02039g). Simulation and modelling not only provide valuable insights into phenomena, but also are increasingly capable of being predictive.

While the relentless advance of computer power has been a major factor in the dramatic expansion of materials modelling, the creation and evolution of techniques has been equally as significant. This themed issue of Journal of Materials Chemistry provides a flavour of the exciting current developments in materials modelling. The new research described covers a wide range of length scales from nano to meso; covers state-of-the-art modelling methods—from quantum descriptions of nanotubes (Jiao et al.: DOI: 10.1039/c0jm01416h) and oxides (Tanaka et al.: DOI: 10.1039/c0jm01932a), to atomistic descriptions of thin films (Yoneya et al.: DOI: 10.1039/c0jm01577f) and coarse-grained representations of soft matter at the micrometer length scale and beyond (Bhattacharya and Balaczs: DOI: 10.1039/c0jm01682a). At the atomistic level of theory, there have been recent significant advances in quantum mechanical techniques, through increasing availability of accurate hybrid functionals and post-Hartree–Fock methods, as well as an increasing drive towards linear scaling that allows more complex systems to be tackled. Force field approaches have also become more sophisticated, with the advent of reactive models that incorporate bond-order contributions seamlessly alongside dynamically fluctuating charges and van der Waals interactions (Joshi et al.: DOI: 10.1039/c0jm01556c).

The last decade or so has seen computational materials research move from consideration of relatively ideal systems, such as bulk properties, to embrace the full complexity of real materials. Here the examination of extended defects and interfaces is critical. Surfaces represent the ultimate defect that controls many of the chemical properties. Arguably, one of the most important materials at present is titania, given its applications to solar energy, water splitting and photocatalytic degradation of waste products. Here an understanding of the interaction between titania and water is vital, and theoretical models are making an important contribution (Sun et al.: DOI: 10.1039/c0jm01491e). Photovoltaic devices typically comprise many component layers of materials, including a transparent conducting film, such as indium tin oxide (ITO), and so quantum mechanical methods provide the opportunity to understand the surface properties and the influence of dopants in these systems (Walsh and Catlow: DOI: 10.1039/c0jm01816c). Both first principles and effective Hamiltonian methods can be powerful techniques in general for understanding optoelectronic devices, including organic photovoltaics, as demonstrated in a study of the potential application of organometallics in this field (Jacko et al.: DOI: 10.1039/c0jm01786h).

Aside from exposed surfaces, other extended defects can exert considerable influence on a material’s performance. Screw dislocations are responsible for accelerating the growth of materials, but also act as a location for the segregation of point defects and impurities, due to the influence of the strain field. They can also accelerate the movement of defects through pipe diffusion along the screw dislocation (Zhang et al.: DOI: 10.1039/c0jm01550d). Complex microstructures, such as intergranular films (Jiang and Garofalini: DOI: 10.1039/c0jm01555e) and polycrystalline materials are now also amenable to computer simulation by molecular dynamics. The strategy of simulated crystallisation, in which an assembly of nanoparticles are agglomerated and annealed, has been shown to generate realistic microstructure (Sayle et al.: DOI: 10.1039/c0jm01580f). More challenging again is the case of a heterogeneous interface between two materials, especially when there is considerable lattice mismatch strain. Theoretical methods can begin to establish the principles underlying the strain release in such heterofilms (Mohn et al.: DOI: 10.1039/c0jm01864c). Simulation is not restricted to interfaces between crystalline materials either. Using molecular dynamics, combined with electronic structure calculations sampled from the resulting trajectory, the properties of water layers trapped between surfactants in a Newton Black film can be characterised in a manner not possible by classical techniques alone (Bresme and Artacho: DOI: 10.1039/c0jm01572e).

One of the areas where computational techniques are at the cutting edge, and often being predictive ahead of experiment, is that of nanoscale materials. Predictions can now be made as to how to modify the electronic properties of carbon sheets, such as graphene and graphane, if the hydrogen content is controlled during transformation (Barnard and Snook: DOI: 10.1039/c0jm01436b). In addition to their electronic properties, the mechanical properties of carbon nanotubes are also unique, with their high strength parallel to the tube axis. However, the weak nature of interactions between tubes frustrates their application in this field. Here multiscale modelling can be used to explore possible bio-inspired cross-linking strategies via non-covalent interactions and identify optimised composite materials (Bratzel et al.: DOI: 10.1039/c0jm01877e).

The same “multiscale” strategy often applies to polymeric and dendrimeric materials, where information gained from atomistic simulation can be used to inform modelling at longer time and length scales (Lukyanov et al.: DOI: 10.1039/c0jm01654c, Huissmann et al.: DOI: 10.1039/c0jm01584a). At still longer length scales, the predictive power of modern computational methods is also apparent. Here, coarse-grained simulation methods can now be used as tools to help “engineer” the desired properties of new polymer–particle nanocomposite materials (Posocco et al.: DOI 10.1039/c0jm01561j); and to control the desired phase behaviour of new liquid crystalline materials (Peroukidis et al.: DOI: 10.1039/01692f, Kuriabova et al.: DOI: 10.1039/c0jm02355h). At length scales beyond the molecular level, hybrid lattice–Boltzmann (LB) methods are currently extending coarse-grained modelling to allow material scientists to look at phenomena that only emerge on still longer length scales. For example, when colloid particles are immersed in a chiral nematic material, hybrid LB methods show that both particle size and cholesteric pitch can be used to control defect configurations and hence influence the interactions between colloidal particles (Lintuvuori et al. DOI: 10.1039/c0jm01824d).

Given the need to rapidly, and reversibly, adsorb molecules for both hydrogen storage and removal of pollutants from exhaust gases, both topics are being extensively explored using computational methods. Here there is the possibility to explore candidate materials using theory and thereby provide guidance to experimental efforts. Although carbon nanostructures alone may not bind hydrogen strongly enough for practical usage for hydrogen storage, metal decorated systems have attracted considerable attention. Bias accelerated first principles dynamics can be used to explore the hydrogen reactivity of metals, such as palladium, supported on a carbon substrate (Mushrif et al.: DOI: 10.1039/c0jm01559h). Even the gas adsorption properties of carbon analogues, such as aluminium nitride single walled nanotubes, can be predicted (Jiao et al.: DOI: 10.1039/c0jm01416h). As an alternative to nanostructured clusters, a more immediate technology for gas adsorption and separation is offered by metal organic frameworks (MOFs). These materials, inspired by the nanoporous characteristics of zeolites, can be tailored to exploit the great variety of functionality in molecular cross-linking and metal coordination to generate new frameworks for specific technological goals. Again they are readily amenable to characterisation via simulation (Liu and Zhong: DOI: 10.1039/c0jm01045f).

In addition to synthetic nanostructures, there are also naturally occurring materials of this form. A fascinating case is that of the mineral imogolite, that consists of nanotubes created by the rolling of aluminium hydroxide layers, such as those found in the mineral gibbsite, but linked on the inside by silanol groups. The structure and energetics of this complex 1-D system have now been examined by accurate hybrid density functional theory for both the armchair and zig-zag configurations (Demichelis et al.: DOI: 10.1039/c0jm00771d).

Although Moore's Law may be hard to sustain in future, as transistor densities require components that encroach on the regime in which quantum confinement effects fundamentally alter the behaviour of semiconductors, there is every prospect that computer power will continue to grow rapidly for the foreseeable future through the adoption of new architectures, such as Graphical Processor Units, and an increased focus on massively parallel algorithms. Combined with continued innovation in the underlying techniques, the ability of computer modelling to address realistic materials with all their inherent complexities will continue to advance for the foreseeable future.

Julian D. Gale, Curtin University, Australia

Mark Wilson, University of Durham, UK

---