Toward the exact solution of the electronic Schrödinger equation for noncovalent molecular interactions: worldwide distributed quantum monte carlo calculations.

Toward the exact solution of the electronic Schrödinger equation for noncovalent molecular interactions: worldwide distributed quantum monte carlo calculations.
复制标题

非共价分子相互作用电子薛定谔方程的精确解:全球分布式量子蒙特卡罗计算。

DOI:
--
复制
发表时间:
2008
影响因子:
2.9
通讯作者:
S. Grimme
S. Grimme
中科院分区:
化学3区
文献类型:
--
作者:
M. Korth;A. Lüchow;S. Grimme

文献摘要

被引文献

相似文献

量子Monte Carlo(QMC@HOME)是量子化学领域第一个大规模的分布式计算项目,它可以对叠层(st)和沃森/克里克(wc)结合的腺嘌呤/胸腺嘧啶(A/T)和胞嘧啶/鸟嘌呤(C/G)DNA碱基对复合物进行量子Monte Carlo(QMC)计算.相互作用能的结果(wc-A/T = 15.7 kcal/mol,wc-C/G = 30.2 kcal/mol,st-A/T = 13.1 kcal/mol,st-C/G = 19.6 kcal/mol)与最著名的基于耦合簇的估计非常一致。这些值的准确性进一步支持的S22基准集的非共价结合系统,我们得到一个小的平均绝对偏差为0.68千卡/摩尔的计算。我们的研究结果支持以前的说法,堆叠能量是可比的大小,通常讨论的氢键基序的相互作用。此外,我们表明,QMC可以作为一个有利的替代传统的波函数方法为大型非共价键结合的系统。我们还详细研究了QMC模拟的所有技术参数,并建议仔细优化程序的Jastrow相关因子,以获得数值稳定和可靠的结果。
Quantum Monte Carlo (QMC) calculations on the stacked (st) and Watson/Crick (wc) bound adenine/thymine (A/T) and cytosine/guanine (C/G) DNA base pair complexes were made possible with the first large scale distributed computing project in ab initio quantum chemistry, Quantum Monte Carlo at Home (QMC@HOME). The results for the interaction energies (wc-A/T = 15.7 kcal/mol, wc-C/G = 30.2 kcal/mol, st-A/T = 13.1 kcal/mol, st-C/G = 19.6 kcal/mol) are in very good agreement with the best known coupled-cluster based estimates. The accuracy of these values is further supported by calculations on the S22 benchmark set of noncovalently bound systems, for which we obtain a small mean absolute deviation of 0.68 kcal/mol. Our results support previous claims that the stacking energies are of comparable magnitude to the interactions of the commonly discussed hydrogen-bonded motif. Furthermore, we show that QMC can serve as an advantageous alternative to conventional wave function methods for large noncovalently bound systems. We also investigated in detail all technical parameters of the QMC simulations and recommend a careful optimization procedure of the Jastrow correlation factors in order to obtain numerically stable and reliable results.