Biomaterials and membranes designed at the computer

On Professor Dieter Hofmann’s screen we see a confused mass of balls and lines. “When a hydrogen atom makes its way through a layer of synthetic material, it is as if a tiny marble were moving aimlessly through a bundle of pearl necklaces,” Hofmann explains. He can predict approximately which path the marble will take. For many practical applications a transport process of this kind is of great significance. For example, when water molecules are supposed to dissolve the thread with which surgeons have sewn up a wound. Or furthermore for membranes, which in a fuel cell permit hydrogen atoms to pass through to the oxygen in carefully measured doses, which separate alcohol from beer, or release medicines from capsules in very small amounts. Designing biomaterials and membranes for special purposes at the computer is a speciality of Dieter Hofmann at the GKSS Research Centre in Teltow near Berlin. The physicist coordinates the European research project Multimatdesign. With reference to concrete examples such as refining natural gas or biologically degradable suture threads, eleven European partners are working on new strategies for computer-aided material development.

The demands made on these materials are enormous. Their task is to separate mingled substances precisely, or they must be easily degradable in wet surroundings. To achieve this they must block certain molecules, and let others through. And depending on the application, they are supposed to be mechanically stable, compatible with other substances and if possible simple to produce. Hitherto such highly specialised synthetic materials were developed with the help of systematic experimentation. Starting out with an idea of what might be suitable chemical components, researchers synthesised polymers in the laboratory – long chains of complex molecules – and then used them to make a membrane, for example, whose properties could then be tested. “This is a lengthy and expensive process,” says Hofmann. So why not let a computer do all the hard work? “In the last twenty years there has been tremendous progress in the programming of molecular processes,” Hofmann explains. And the Multimatdesign project is building on these developments.

For the virtual test tube, too, the computer chemists select molecules with favourable properties. They add information about their type and size, and about the strength and geometry of their bonding. Then the computer sets to work, and calculates how the chains of molecules would finally end up arranging themselves. “This modelling procedure takes a lot of know-how,” Hofmann emphasises. The mutual interaction of the many atoms in such a process is very complex. At various size levels different mechanisms come into play. These range from quantum chemistry, which influences the atoms and their bonding, to thermodynamics, which describes larger collections of particles in statistical terms. “In the Multimatdesign project we have brought together leading international experts from all these fields,” Hofmann is pleased to tell. “It enables us to work at a problem in its entirety.”

For example, the physicist and his team at the GKSS Zentrum für Biomaterialentwicklungdesigned various models for a material that can be used to separate gases. In refining natural gas it could filter the unwanted components, butane and propane, out of the methane.The properties of the virtual membrane are also tested on the computer. Above all, the researchers calculate the path and the speed with which a propane molecule passes through the polymer. This depends, for example, on the size and distribution of the available volume between the swirling chains of pearls, and on how propane interacts with the membrane polymer – whether it is bound, permitted to pass, or is actively transported elsewhere. An Italian team from the research institute ITM-CNR participated in the project segment on gas refining.

The French company Air Liquide wants to use the results in the very near future for quite specific tasks in the field of gas separation, for example, refining natural gas. For Multimatdesign is certainly not restricted to virtual chemistry. At each stage of the development chain there are experiments designed to test whether the models and simulations actually coincide with reality. However, the use of computers does not only make it possible to develop new materials more efficiently. “The simulations also make it possible for us to improve our understanding of fundamental processes,” emphasises Hofmann. “And that is the precondition for the ongoing development of wholly new ideas and rules pertaining to effective material screening.”