University of Minnesota, Chemical Theory Centre, and Minnesota Supercomputing Institute, 207 Pleasant St. SE, Minneapolis, MN 55455-0431, USA.
Phys Chem Chem Phys. 2013 Feb 14;15(6):1837-43. doi: 10.1039/c2cp43305b. Epub 2012 Dec 18.
We predict structures and energies of water clusters containing up to 100 waters with tight-binding density functional theory (DFTB). A per-hydrogen-bond energy correction is found to correct for systematic errors in the DFTB cluster energies. We compare the DFTB structures and energies to density functional theory (DFT) calculations and to the most accurate wave function theoretical (WFT) values available (ranging from coupled-cluster theory to second-order perturbation theory). After including the simple hydrogen bond correction, we achieve a root-mean-square difference of less than one kcal mol(-1) with the best estimates. As DFTB optimizations are orders of magnitude faster than DFT or canonical MP2, it is apparent that DFTB is a very practical method for calculating large water cluster structures and, with the hydrogen bond correction, also energies.
我们使用紧束缚密度泛函理论(DFTB)预测了多达 100 个水分子组成的水团簇的结构和能量。发现每个氢键的能量修正可以修正 DFTB 团簇能量中的系统误差。我们将 DFTB 的结构和能量与密度泛函理论(DFT)计算以及可用的最精确的波函数理论(WFT)值进行比较(范围从耦合簇理论到二级微扰理论)。在包含简单氢键修正后,我们与最佳估计值的均方根差小于 1kcal/mol。由于 DFTB 优化比 DFT 或经典 MP2 快几个数量级,因此 DFTB 显然是计算大水团簇结构和(带有氢键修正时)能量的非常实用的方法。