Development And Application Of Transition State Search Methods

With the improvement of computer performance and the development of quantum chemical methods,quantum chemical calculations have received more and more attention in the study of chemical reaction mechanisms.The search for equilibrium geometries and transition states on potential energy surface is the basis and focus of theoretical study of chemical reaction mechanisms.Compared with equilibrium geometries,the search of transition states is much more difficult.In the past two decades,a variety of transition state search methods have been proposed.Transition state search methods can be roughly divided into single-ended and double-ended methods.The single-ended methods require an approximate transition state as initial guess structure.The quality of the initial guess determines the efficiency and even success of the transition state search.The initial guess of the transition state can be artificially given,or obtained through constrained optimization or relaxed scan.These methods are very dependent on the researcher’s experience and intuition.The doubleended transition state search methods use the reactant and product structures as the initial structures and does not need initial guess of the transition state.The double-ended methods can automatically search for the transition state between the reactant and product.Compared with the single-ended methods,the double-ended methods are more convenient to use,but require more computational resources.In order to overcome the shortcomings of the large amount of calculation required by the double-ended methods and retain the advantages of ease to use,we developed a variety of efficient transition state search methods that do not rely on the experience or intuition of researchers,and these methods have been successfully applied in the study of transition metal catalytic reaction mechanisms.The key research subjects discussed in this study are the following:1.A double-ended transition state search method is proposed: the elastic image pair method(EIP).In this method,two images start from the reactant and product side respectively,and approach the transition state during the optimization.The spring forces are added between the two images to prevent the images from falling into the area of local minimum on the potential energy surface.When the optimization is completed,the midpoint of the two images is the structure of transition state.In order to test the efficiency and accuracy of the proposed method,we constructed a test set containing 45 different reactions,and searched transition states using the EIP method and nudged elastic band(NEB)method.The results show that the EIP method we developed is better than the NEB method in terms of efficiency and accuracy.And also,the EIP method uses reactant and product as input,it is very easy to use.In summary,the EIP method is efficient,accurate and simple to use.This method is expected to become a convenient and efficient tool in searching transition state.2.The EIP method is modified and the improved elastic image pair method(iEIP)is proposed.The original EIP method can locate transition states in small-and medium-size systems,but it does not perform well on very flat potential energy surfaces or in large systems.This is mainly because the spring forces added between the image pair does not adapt well to different potential energy surfaces,which makes it difficult to control the distance and relative energy between the two images during the optimization.In the i-EIP method,the displacement of the image pair is decomposed into three independent displacements,which are used to minimize the energy of the image pair,control the distance and relative energy between the two images respectively.The total displacement of the image pair in each optimization step is calculated using micro iteration.And this method is combined with the rational function optimization to give accurate transition states.The test results show that the i-EIP method if more efficient than the original EIP method.We also tested the stability of the i-EIP method in complex systems,five reactions contain different transition metal atoms were selected for use as a test set.All transition states were successfully located with the iEIP pair method.The above results show that the i-EIP pair method is more efficient and stable.This method could become a powerful transition stat search tool.3.To search a large number of transition states of the same type reactions,we proposed a transition state search method named constrained minimization rational function optimization method(CM-RFO).This method uses known transition state as templet,and the transition vector of the template is used to establish constrains.The constrained optimization method is used to refined the structures to give initial guesses of transition states.Combined with rational function optimization method to obtain the accurate transition state.In order to verify the effectiveness of the CM-RFO method,the Diels-Alder reaction was selected as the test example,and 703 different substituted butadiene and ethylene were used as test set.The corresponding transition states were searched with the CM-RFO method.The results proved that all transition states were successfully located and are consistent with expectations.We also calculated and studied the energy barrier and concerted asynchronous mode of the reactions.Through the analysis of bond length differences,relative vibration components and intrinsic reaction coordinates of the transition states,concerted asynchronous mode of DielsAlder reactions with different substrates were verified.4.The mechanism of asymmetric amination of a racemic alcohol catalyzed by a Ru-pincer complex was studied using density functional theory(DFT),and the developed transition state search method,the i-EIP method,was used to search transition states in the reaction path.This reaction uses racemic phenylethyl alcohol and tert-butanesulfinamide as substrates.The chiral amine product is generated through three key steps: alcohol dehydrogenation,imine formation and imine hydrogenation.In the first step,racemic alcohol undergoes dehydrogenation step to generate acetophenone.Then dehydration coupling between acetophenone and tertbutanesulfinamide forming tert-butanesulfinimide in imine formation step.Finally,imine is reduced via imine hydrogenation step to give the chiral amine product.The calculated results show that the hydride transfer step during the hydrogenation of imine is the chirality-determining step in the whole catalytic cycle.In the hydride transfer step,the energy of the transition state in favorable pathway is 7.6 kcal/mol lower than the energy of the corresponding transition state in unfavorable pathway,the energetic span of favorable pathway is 18.9 kcal/mol,which is consistent with the experimental observations.The origin of diastereoselectivity of this reaction is also studied.Through the electrostatic potential surface analysis,cis-and trans-imine hydrogenation calculation,and sulfinyl group modification calculation,it was found that the diastereoselectivity mainly stems from the hydrogen bonding interactions between the subtract and the ligand of the catalyst.The effects of different substituents on the ligands on the imine hydrogenation step were explored,and the CM-RFO method is used to search for the corresponding transition states.The calculated results show that when-Cl group is the substituent at the para position of the benzene in the ligand,it is most favorable in the imine hydrogenation step.We also found that the relative energy of the intermediated formed in the hydride transfer step have good correlation with the electronic effect of the substituent.And based on this feature,we proposed a fast method to estimate the relative energies with semi-empirical calculated results.