The sky is the limit for computational chemistry

Citation

Warde, S. (2000), "The sky is the limit for computational chemistry", Aircraft Engineering and Aerospace Technology, Vol. 72 No. 3. https://doi.org/10.1108/aeat.2000.12772caf.002

The sky is the limit for computational chemistry

Stephen Warde

Keywords Computational methods, Coatings, Corrosion prevention, Aircraft, BAe Systems

Computational chemistry, a technology traditionally associated with drug discovery, is now widely used to design and characterise many other materials. It is finding applications far beyond the pharmaceutical industry. Researchers at BAe Systems (formerly British Aerospace) are applying computation to the design of environmentally safe alternatives to today's anti-corrosive aircraft coatings.

Computational chemistry is the application of computer software to improve the understanding of chemical systems. Typically, researchers construct a graphical model of a molecule on their computer screen and then use predictive methods to refine the model and to compute values for its properties. This helps them to understand the underlying chemistry of the system and to optimise its behaviour for relevant applications. Such methods originated in the pharmaceutical industry, where they assist the characterisation of proteins and the discovery of new drugs. Computational chemistry is now an integral component of every serious drug discovery effort. All the leading pharmaceutical companies make significant investments in the technology, and the market for drug discovery software and services is highly competitive and dynamic.

Computational materials science

Yet computational chemistry can impact problems far removed from drug discovery. Since the late 1980s, software companies and industrial researchers have been applying these techniques to more general materials design problems. Initial efforts focused on polymeric systems, where methods developed for protein modelling found obvious parallels, and on catalysis, where the relationships between molecular geometry and chemical behaviour are of particular interest. Today, this computational materials science addresses problems in areas including surface and solid state chemistry, crystallisation, formulation, and materials characterisation. The core market is the chemicals and petrochemicals industry. But, as the technology matures, it is finding applications in sectors including electronics, foods, cosmetics, metals and, as we shall see, aerospace.

The challenges facing computational materials science are those of diversity. First, there is the varied chemistry of the systems that must be studied. Computational chemistry was first developed in the pharmaceutical industry partly because the biological systems studied are almost all organic, consisting of a limited range of chemical elements. To tackle a wider range of materials problems, predictive methods must deal with elements from across the periodic table. Yet the approximations inherent in such methods often break down when they are applied to materials dissimilar from those for which they were designed. Second, there is the wide variety of material types. Computational chemistry began by looking at comparatively simple interactions between isolated molecules. But materials research encompasses amorphous materials, many different types of crystalline systems, semi-crystalline solids, liquids, complex fluids, and gases. In fact, most practical problems cover mixtures of several of these systems!

Computational materials science must also combine effectively with other tools. Many important properties - for example, the mechanical behaviour of metals and ceramics - are determined as much by microscopic or mesoscopic behaviour as by molecular behaviour. Molecular-level simulation complements engineering or process simulations, rather than replaces them. Interaction with experimental work is also particularly important, providing a reality-check that enables researchers to establish and validate useful models. Computation works best as one component in an integrated research program.

The key technologies

The core technologies in computational materials science are molecular (or atomistic) simulation and quantum mechanical (QM) methods. Molecular simulation takes the atom as its basic unit. In effect, it assumes that molecules or materials are simply made up of "balls", representing atoms, and "springs", representing the bonds between them. Mathematical relationships are then determined that represent the interaction of the balls and the behaviour of the springs and which result in behaviour that matches experimental observations. These relationships are termed "force fields". A force field determined for a particular system is often accurate in describing the properties of similar materials. The advantage of such simulation is that it is extremely rapid and can simulate very large systems. Today's most sophisticated methods account for dozens of types of chemical interaction, many of them extremely subtle. Entire research programs are dedicated to developing force fields for particular classes of material and to establishing and extending the range of their validity.

Quantum mechanical methods are more accurate, but also more computationally demanding. They are based on the fundamental equation of chemistry, the SchrÎdinger equation. Using the sub-atomic behaviour of electrons, this equation theoretically enables the accurate determination of molecular structure and of the energy of a chemical system, which in turn enables prediction of many properties. In fact, the SchrÎdinger equation cannot be solved exactly for anything larger than a hydrogen atom. QM simulation is the attempt to overcome this limitation using various approximations. A major contribution was the introduction, in 1964, of the density functional theory (DFT) - an approximation that showed that chemical properties could be accurately computed using calculations on only some of a molecule's constituent electrons. This dramatically reduced computation times. Walter Kohn and John Pople received the 1998 Nobel Prize for their work in this area. Today, scientists routinely perform quantum mechanical calculations on systems with dozens of atoms, and restrictions on system size and calculation speed are being rapidly overcome.

A third important method is the simulation of data from analytical instruments. Given a model of a material's structure it is possible to predict, for example, its X-ray powder diffraction pattern. Sometimes, as with UV or infrared spectra, predictions of analytical instrument output are based on analyses of QM or molecular simulations. Sometimes, as in the case of diffraction, equations exist allowing the prediction to be made directly from the structure. Comparison between the simulated and experimental patterns enables validation of the model. The model can even be adapted so as to achieve a match between simulation and experiment - many sophisticated computational routines are now available that automate this refinement process.

Instrument simulation provides an essential link between computational work and reality. This validation is a requirement if these methods are to be practically applied. And more and more real-world applications of materials simulation are being reported. On-going research at BAe Systems provides an excellent example of such work. BAe Systems researchers are using quantum mechanical calculations to predict results from analytical instruments, aiding their design of new aircraft coatings.

Improving aircraft coatings

Aircraft bodies are generally constructed out of an aluminium copper alloy. Although the copper strengthens the lightweight aluminium it also acts as a cathode, resulting in corrosion. Some aircraft manufacturers anodise their craft in order to prevent corrosion, but many prefer to use chromate conversion coatings. Chromate coatings have a lower electrical contact resistance than anodised coatings and are widely used as a primer for paint to form a protective surface. Such coatings can be applied in thin layers and adhere well to aluminium alloys. They provide a high degree of protection against marine and humid environments and have the added advantage that, if the surface is scratched, the surrounding coating "leaks" to cover the exposed metal.

However, there are increasing concerns about the environmental and health-and-safety implications of chromates. Their toxic effects have been known for some time, but it is only recently that they have been confirmed as carcinogens. They are dangerous when they come into contact with skin, and when chromate dust is inhaled. The risk to workers who apply chromate primers is greatest when sanding the surface of an aircraft to prepare it for priming, when removing the dust generated by sanding, and when applying the coating using high-volume low-pressure (HVLP) manual spray guns. Further to the workplace risks, chromate-based conversion coatings also result in the discharge of a toxic heavy metal waste. The continued use of such coatings is thus not viable in the long term. BAe Systems predicts that environmental legislation will soon restrict their use.

A strong candidate for the next generation of anti-corrosive coatings comes from a group of materials known as rare earths. Rare earths are the series of elements of atomic number 57 to 71. They are used in applications such as permanent magnets, automotive catalytic converters, nickel-hydride batteries, polishing compounds, structural ceramics, and as additives in iron and steel[1]. Cerium is the most abundant of the rare earth metals - there is more cerium in the earth's crust than there is aluminium.

The Australian Aeronautical Research Laboratory has published results of work with cerium oxide-based coatings that have shown interesting properties with three potential uses. The first is as a decorative coating on aluminium, zinc, and steel sheet. The second is as an anti-corrosion coating on metal construction operating in moist situations such as air-conditioning cooling towers. The third application is as a replacement for chromium oxide anti-corrosion coatings on aircraft parts. After evaluation by salt-spray and electro-chemical tests used in the aircraft industry, paint adhesion to the new conversion coating is found to be extremely good. The material has very low toxicity[2]. However, the properties of cerium are still not fully understood, and there is still a great deal of research to be done before cerium can be used as a commercial anti-corrosive aircraft coating.

Using computation at BAe Systems

BAe Systems is using molecular modelling to assess whether rare earths really can provide an environmentally friendly alternative to toxic chromate anti-corrosive aircraft coatings. The problem in gaining a full understanding of their properties and likely performance is with the complex structure that results when rare earths are deposited in thin films. This has proved difficult to decipher. The layers of cerium required for anti-corrosive coatings are about one micron thick. The films are complex both structurally and chemically, containing two different oxidation states of cerium - Ce(III) and Ce(IV)[3]. Steve Harris and Chris Somerton of BAe Systems are working to understand the nature of these depositions. They are using X-ray photoelectron spectroscopy (XPS) to discern the oxidation states in a non-destructive manner. The results of XPS can reveal the electronic properties of the solid, the composition of the solid, and the chemical nature of its surfaces.

However, the spectra of Ce(III) and Ce(IV) are complex and hard to interpret. Somerton and Harris are using computational chemistry software from Molecular Simulations Inc. (MSI) in order to predict and understand spectra. DMol³ is a density functional theory (DEr) program. This quantum mechanical method is able to predict the structure and properties of either molecular clusters or periodic systems[4]. An XPS spectrum involves the removal of a core electron from an atom. Using DMol³, computational chemists can compute the total energies of both the neutral and the ionised states of the atoms. The energy difference between these states corresponds to a peak in the XPS spectrum. Thus DMol³ can be used to predict XPS. Predicted spectra are compared with the experimental spectra produced from the cerium layers, enabling Somerton and Harris to establish the composition of these layers. Comparison with known systems has validated this process, which is now being used to elucidate the properties of alternative materials.

Better materials by design

BAe Systems provides one example of a company effectively combining computation with experimental techniques in the design of new materials. This example shows how far computational materials science has come from its origins in the pharmaceutical industry, dealing as it does with metallic systems in complex thin film structures. Many other examples are available to reinforce this point. At Schlumberger Research, for example, researchers used molecular modelling in the design of speciality cements for oilfield applications[5]. At Zeneca, chemists designed additives to prevent corrosion and scale formation[6]. At ALCOA, researchers have also used quantum mechanical techniques to study coatings on aluminium[7]. Such work demonstrates the emergence of computation as an important research tool across the diverse problems posed to researchers in the materials-based industries. As computer power increases, making such technology accessible from standard desktop PCs, its importance can only increase.

Notes and references

1. 1.

   For more information on rare earths, including market information, see the Rare Earths Profile from The World Metals Information Network at http://www.amm.com/ref/rare.htm

2. 2.

   Hughes, T., "Cerium based anti-corrosion coatings", Australian Rare Earth Newsletter, CSIRO.

3. 3.

   Kotani, J.P. (1988), Advances in Physics, Vol. 37, pp. 37-85.

4. 4.

   DMol³, Molecular Simulations Inc., 230-250 The Quorum, Barnwell Drive, Cambridge CB5 8RE, UK. http://www.msi.com/

5. 5.

   Coveney, P.V. and Humphries, W. (1996), J. Chem. Soc., Faraday Trans., Vol. 92 No. 5, p. 831.

6. 6.

   Bromley, L.A., Buckley A.M., Chlad, M., Davey, R.J., Drewe, S. and Finlan, G.T. (1994), J. Colloid lnt. Sci., Vol. 164, pp. 498-502.

7. 7.

   For this and many other case studies, see http://www.msi.com/materials/cases/