Collaborative Computational Project on Computational Electronic Structure of Condensed Matter

在20世纪20年代末，有人声称量子力学可以预测任何分子或材料的每一个特性——从本质上讲，量子力学解释了物理学、化学、材料科学和生物学的全部。然而，直到50多年后，计算机才变得强大到足以解决量子力学方程，从而开始证明这种说法是正确的。这些最初的测试只针对原子或只包含2或3个电子的非常小的分子，或只包含1或2个原子的晶体材料。在过去的30年里，更强大的计算机与更好的理论和数值方法的结合使我们能够常规地计算分子和材料的大量特性。例如，我们现在可以确定一个特定的原子组合的最稳定的晶体结构，即使这是未知的实验。此外，我们还可以计算包含数百个原子的系统的振动频率、化学反应的活化能、表面能和许多其他性质。然而，计算机继续变得越来越强大。我们现在将能够在这些日益强大的计算机上使用我们强大的量子力学计算方法来生成关于大量材料的大量信息。这些都是已知的材料，更有趣的是，还没有制造出来的材料。使用这些庞大的数据库，在未来，我们将能够通过识别数据库中的候选材料来为特定的应用选择材料。我们还希望数据库中的数据能够帮助我们设计制造路线，不仅可以制造所需的材料，还可以制造整个设备。因此，在创建基础设施以创建和提供对这些数据库的访问时，我们在计算材料和设备设计的道路上迈出了第一步。我们每天使用的许多产品都是用计算机设计的，人们发现，用计算机上的虚拟设计代替真实设计可以减少产品的开发时间，使其质量更高和/或更安全（在汽车的情况下尤其如此），从而产生显著的社会经济效益。

It was claimed in the late 1920's that quantum mechanics chould predict every property of any molecule or material - essentially that quantum mechanics explained the whole of physics, chemistry, materials science and biology. However, it took over 50 years before computers became powerful enough to solve the equations of quantum mechanics to begin to prove that this claim was correct. These first tests were only for atoms or very small molecules that contained only 2 or 3 electrons or for crystalline materials that contained only 1 or 2 atoms in basic crystal unit cell. Over the last 30 years the combination of more powerful computers combined with better theoretical and numerical methods has brought us to the point where we can routinely compute a vast range of properties of molecules and materials. For instance we can now determine the most stable crystal structures for a particular combination of atoms even when this in not known experimentally. In addition we can calculate vibrational frequencies, activation energies of chemical reactions, surface energies and many more properties for systems containing hundreds of atoms. However, computers continue to get more powerful. We will now be able to use our powerful quantum mechanical computational methods on these increasingluy powerful computers to generate huge amounts of information about a vast number of materials. These will be both known materials and, much more interestingly, materials that have not yet been made. Using these vast databases we will, in the future, be able to choose a material for a particular application by identifying candidates in the databases. We also expect that the data in our databases will allow us to help design the fabrication route that will create not just the required materials but entire devices. Thus in creating the infrastructure to create and provide access to these databases we are taking the first step on the route to computional materials and device design. Many products we use every day are designed using computers and it has been found that replacing real design by virtual design on a computer can reduce the development time for a product, make it higher quality and/or safer (this has been particularly true in the case of cars) and thus produce a significant socio-economic benefit.