یک روش عمومی حمل و نقل سنکرون جهت انتقال حالت گذار

This paper describes a generalized synchronous transit method for locating transition state structures or first-order saddle points. The algorithm is based on the established scheme of combining the linear or quadratic synchronous transit method with conjugate gradient refinements but generalized to deal with molecular and periodic systems in a seamless manner. We apply the method to a study of the early stage atomic layer deposition (ALD) growth of ZrO2.

Transition state structures play a crucial role in the understanding and design of chemical reactions. Despite their fundamental significance transition state structures are very challenging to locate accurately and conventional algorithms are typically computationally demanding and are not universally applicable. Conceptually the easiest approach to locate a transition state structure or first-order saddle point is to sample the potential energy surface (PES) and map out the minimum energy path (MEP) [1] connecting reactant and product structure. This, however, requires a reasonably accurate PES sampling. This method, though simple-minded, becomes impractical for systems with a large number of degrees of freedom. Over the last several years a number of methods have been proposed [2] to locate transition state structures. One good example is the first derivative based nudged elastic band (NEB) method [3]. The method uses a guess for the reaction pathway and optimizes the geometries of several images along this path simultaneously. The resulting path is the MEP. Although the NEB approach greatly reduces the computational load compared with a direct evaluation of the PES, substantial resources are still directed towards optimizing geometries at several points along the adiabatic path, even though the point of greatest interest is the transition state geometry itself. This raises the question whether a method can be formulated where the search is still performed along a guessed path, but where the main effort is directed towards accurately locating the transition state. A very promising attempt to solve this problem was proposed by Halgren and Lipscomb [4] in 1977 via the linear (LST) and quadratic synchronous transit methods (QST), respectively. In the LST approach a series of single point energy calculations are performed on a set of linearly interpolated structures between a given reactant and product. The maximum energy structure along this path provides a first estimate of the transition state structure. A single refinement in a direction orthogonal to the LST path is then performed, which is then used as an intermediate to define a QST pathway. This procedure yields a further refined estimate for the transition state geometry. In this paper we propose a generalized transition state location scheme [9] where both molecular and periodic models can be treated within the same framework. Our first derivative based approach closely follows and builds upon the ideas of the traditional LST/QST method in conjunction with conjugate minimization ideas of Bell and Crighton [5] and those of Fischer and Karplus [6]. We first present the details of the method followed by an application where we investigate the reaction of ZrCl4 with partially hydroxylated and hydrated Si(1 0 0) surfaces using both periodic slab and cluster models.

Transition state structures play a crucial role in the understanding and design of chemical reactions. Despite their fundamental significance transition state structures are very challenging to locate accurately and conventional algorithms are typically computationally demanding and are not universally applicable. Conceptually the easiest approach to locate a transition state structure or first-order saddle point is to sample the potential energy surface (PES) and map out the minimum energy path (MEP) [1] connecting reactant and product structure. This, however, requires a reasonably accurate PES sampling. This method, though simple-minded, becomes impractical for systems with a large number of degrees of freedom. Over the last several years a number of methods have been proposed [2] to locate transition state structures. One good example is the first derivative based nudged elastic band (NEB) method [3]. The method uses a guess for the reaction pathway and optimizes the geometries of several images along this path simultaneously. The resulting path is the MEP. Although the NEB approach greatly reduces the computational load compared with a direct evaluation of the PES, substantial resources are still directed towards optimizing geometries at several points along the adiabatic path, even though the point of greatest interest is the transition state geometry itself. This raises the question whether a method can be formulated where the search is still performed along a guessed path, but where the main effort is directed towards accurately locating the transition state. A very promising attempt to solve this problem was proposed by Halgren and Lipscomb [4] in 1977 via the linear (LST) and quadratic synchronous transit methods (QST), respectively. In the LST approach a series of single point energy calculations are performed on a set of linearly interpolated structures between a given reactant and product. The maximum energy structure along this path provides a first estimate of the transition state structure. A single refinement in a direction orthogonal to the LST path is then performed, which is then used as an intermediate to define a QST pathway. This procedure yields a further refined estimate for the transition state geometry. In this paper we propose a generalized transition state location scheme [9] where both molecular and periodic models can be treated within the same framework. Our first derivative based approach closely follows and builds upon the ideas of the traditional LST/QST method in conjunction with conjugate minimization ideas of Bell and Crighton [5] and those of Fischer and Karplus [6]. We first present the details of the method followed by an application where we investigate the reaction of ZrCl4 with partially hydroxylated and hydrated Si(1 0 0) surfaces using both periodic slab and cluster models.