Formal and Computational Studies of Molecular Dynamics

理论与计算化学项目支持休斯顿大学的D. Kouri教授。在接下来的拨款期内，我们将继续进行时间无关波包分治（TIWDC）方法的计算实现和测试。直接目标是：a)测试dc程序，其中初始安排也从强相互作用区域解耦。b)测试解耦振动-旋转、纯旋转和弹性最终状态相互作用区域的直流策略，以实现该方法的最大效益。c)将DC程序应用于H+O2 - OH+O，以便尝试首次完全从头计算系统的状态到状态反应截面作为能量的函数。他还将首次对TIWDC方法进行HOD光解的完整3D测试，重点是与HOD振动热时的实验进行比较。其次，他将继续发展分布式近似函数（daf），包括探索与各种插值方案相关的一类新的daf。这些研究还将研究使用有限数量的从头算输入数据的多维势能曲面的拟合，特别强调电子非绝热系统。第三，他将继续研究一种正在开发的新方法，用于对角化具有非常密集特征谱的超大矩阵。最后，他将开始开发和测试新的近似方法，用于解决涉及4个以上原子的反应的量子方程，以模拟更复杂的化学系统。这里支持的研究重点是开发鲁棒、高效和高度稳定的方法来求解用于描述基本化学反应过程的偏微分方程。由于这些方程极其复杂，只有使用功能最强大、速度最快、内存最大的计算机才能解决这些问题。然而，对基本化学反应过程的详细了解的发展可以打开控制化学反应的可能性，以最大限度地提高所需的产物。这些研究也用于模拟地球大气中发生的数值基本过程，以了解和更好地处理污染。这些研究也可以应用于模拟燃烧的详细化学过程，以开发无污染的推进系统。此外，正在开发的一些技术也可以用于解决其他非常重要的偏微分方程。因此，正在开发的技术显示出巨大的希望，可以解决从长期天气预报到用于模拟燃烧的其他方程的各种方程。

The Theoretical and Computational Chemistry Program is supporting Professor D. Kouri of the University of Houston. In this next grant period he we will continue to pursue computational implementation and testing of the Time-Independent Wavepacket Divide-and-Conquer (TIWDC) approach. The immediate objectives are: a) Testing the DC-procedure whereby the initial arrangement is also uncoupled from the strong interaction region. b) Testing the DC strategy for uncoupling the vibrational-rotational, pure rotational, and elastic final state interaction regions so that the maximum benefit of the approach can be realized. c) Applying the DC procedure to H+O2 - OH+O in order to try and carry out the first ever completely ab initio calculation of state-to-state reaction cross sections for the system as a function of energy. He also will carry out the first full 3D tests of the TIWDC method for photodissociation of HOD, with emphasis on a comparison with experiment when the HOD is vibrationally hot. Secondly, he will continue development of the distributed approximating functionals (DAFs), including the exploration of a new class of DAFs that are related to various interpolation schemes. These investigations will also study the fitting of multidimensional potential energy surfaces using finite amounts of ab initio input data, with particular emphasis on electronically non-adiabatic systems. Third, he will continue studies of a new method being developed for diagonalizing extremely large matrices with very dense eigenspectra. Finally, he will begin work on developing and testing new approximation methods for solving the quantum equations for reactions involving more than 4 atoms in order to simulate more complex chemical systems. The research supported here focuses on developing robust, efficient, and highly stable methods for solving the partial differential equations which are used to describe fundamental chemical reaction processes. Because of the enormous complexity of the equations, solutions are feasible only by using the most powerful, high-speed, large-memory computers. However, the development of a detailed understanding of the elementary chemical reaction process can open up the possibilities of controlling chemical reactions in order to maximize desired products. These studies are also used in simulating numerically fundamental processes that occur in the earth's atmosphere, in order to understand and better deal with pollution. Such studies also find application in modeling the detailed chemistry of combustion in order to develop pollution-free propulsion systems. In addition, some of the techniques being developed can also be used to solve other, enormously important partial differential equations. Thus, the techniques being developed show great promise for making it possible to solve equations ranging from long term weather forecasting, to other equations used to model combustion.