New Stress Test for Ring Polymer Molecular Dynamics: Rate Coefficients of the O(3P) + HCl Reaction and Comparison with Quantum Mechanical and Quasiclassical Trajectory Results.

In the past decade, ring polymer molecular dynamics (RPMD) has emerged as a very efficient method to determine thermal rate coefficients for a great variety of chemical reactions. This work presents the application of this methodology to study the O(3P) + HCl reaction, which constitutes a stringent test for any dynamical calculation due to rich resonant structure and other dynamical features. The rate coefficients, calculated on the 3A' and 3A″ potential energy surfaces (PESs) by Ramachandran and Peterson [ J. Chem. Phys. 2003 , 119 , 9590 ], using RPMD and quasiclassical trajectories (QCT) are compared with the existing experimental and the quantum mechanical (QM) results by Xie et al. [ J. Chem. Phys. 2005 122 , 014301 ]. The agreement is very good at T > 600 K, although RPMD underestimates rate coefficients by a factor between 4 and 2 in the 200-500 K interval. The origin of these discrepancies lies in the large contribution from tunneling on the 3A″ PES, which is enhanced by resonances due to quasibound states in the van der Waals wells. Although tunneling is fairly well accounted for by RPMD even below the crossover temperature, the effect of resonances, a long-time effect, is not included in the methodology. At the highest temperatures studied in this work, 2000-3300 K, the RPMD rate coefficients are somewhat larger than the QM ones, but this is shown to be due to limitations in the QM calculations and the RPMD are believed to be more reliable.

Automated calculation of thermal rate coefficients using ring polymer molecular dynamics and machine-learning interatomic potentials with active learning.

We propose a methodology for the fully automated calculation of thermal rate coefficients of gas phase chemical reactions, which is based on combining ring polymer molecular dynamics (RPMD) and machine-learning interatomic potentials actively learning on-the-fly. Based on the original computational procedure implemented in the RPMDrate code, our methodology gradually and automatically constructs the potential energy surfaces (PESs) from scratch with the data set points being selected and accumulated during the RPMDrate simulation. Such an approach ensures that our final machine-learning model provides a reliable description of the PES that avoids artifacts during exploration of the phase space by RPMD trajectories. We tested our methodology on two representative thermally activated chemical reactions studied recently by RPMDrate at temperatures within the interval of 300-1000 K. The corresponding PESs were generated by fitting to only a few thousand automatically generated structures (less than 5000) while the RPMD rate coefficients showed deviation from the reference values within the typical convergence error of RPMDrate. In future, we plan to apply our methodology to chemical reactions that proceed via complex-formation thus providing a completely general tool for calculating RPMD thermal rate coefficients for any polyatomic gas phase chemical reaction.

A ring polymer molecular dynamics study of the OH + H2(D2) reaction.

In this work we have performed a ring polymer molecular dynamics (RPMD) study of the OH + H2 and OH + D2 reactions at temperatures ranging from 150 K to 2000 K using two available ab initio potential energy surfaces (PESs) that have been termed as the YZCL2 and NN1 PES, respectively. The YZCL2 PES was developed by Yang et al. [J. Chem. Phys., 2001, 115(1), 174] which is based on points fitted by a modified Shephard interpolation method and calculated with unrestricted coupled-cluster theory with all single and double excitations and a perturbative account of the triple excitations (UCCSD(T)) method with an aug-cc-pVQZ basis. The NN1 PES was constructed by Chen et al. [J. Chem. Phys., 2013, 138(15), 154301] using a neural networks method to fit ab initio energies calculated at the UCCSD(T)-F12a/AVTZ level of theory. We show that both techniques provide reliable PESs. The RPMD thermal rate coefficients and the kinetic isotope effects (KIEs) calculated using these two PESs are in very good agreement with each other as well as with previous experimental values available to date. Besides, we have shown that these two procedures for fitting PESs can yield even more similar RPMD rate coefficients when the same level of ab initio theory is employed, at least for the present OH + H2 reaction. Comparison of the previous theoretical calculations on the NN1 PES, namely, instanton theory and canonical variational theory with microcanonical optimized multidimensional tunneling, shows that the present RPMD results are more consistent and accurate. Future experimental measurements of the KIEs and accurate quantum mechanical calculations on these PESs are highly desirable, especially at low temperatures.

A ring polymer molecular dynamics study of the Cl + O3 reaction.

We have performed ring polymer molecular dynamics (RPMD) calculations on the Cl + O3 → ClO + O2 reaction at temperatures ranging from 200 K to 400 K, and compared the results with previous theoretical studies and also with the available experimental data. This reaction presents a couple of features which makes it a particularly interesting and challenging case to be studied using RPMD. First, classically, this is essentially a barrierless reaction, with a saddle point located below the reactants. However, the free energy profiles along the reaction coordinate display small barriers due to the fact that the decrease in enthalpy from reactants to the TS is somewhat compensated by a decrease in entropy. To our knowledge this is the first time such a process is studied using this technique. Second, the transition state is located early in the reactant valley, therefore the inclusion of the recrossing correction in the RPMD calculations is crucial to determine rate coefficients. Regarding quantum effects, our calculations show that RPMD results are within the error bars of the purely classical ones. This implies that tunnelling is negligible in this reaction at the temperatures studied, not surprisingly for a system including oxygen and chlorine atoms, and that the zero point energies of reactants, transition state and products are practically the same. Finally, the rate coefficients presented in this work are in a fairly good agreement with the recommended experimental values, somewhat better than those obtained using other approaches.

Theoretical kinetics study of the F((2)P) + NH3 hydrogen abstraction reaction.

The hydrogen abstraction reaction of fluorine with ammonia represents a true chemical challenge because it is very fast, is followed by secondary abstraction reactions, which are also extremely fast, and presents an experimental/theoretical controversy about rate coefficients. Using a previously developed full-dimensional analytical potential energy surface, we found that the F + NH3 → HF + NH2 system is a barrierless reaction with intermediate complexes in the entry and exit channels. In order to understand the reactivity of the title reaction, thermal rate coefficidents were calculated using two approaches: ring polymer molecular dynamics and quasi-classical trajectory calculations, and these were compared with available experimental data for the common temperature range 276-327 K. The theoretical results obtained show behavior practically independent of temperature, reproducing Walther-Wagner's experiment, but in contrast with Persky's more recent experiment. However, quantitatively, our results are 1 order of magnitude larger than those of Walther-Wagner and reasonably agree with the Persky at the lowest temperature, questioning so Walther-Wagner's older data. At present, the reason for this discrepancy is not clear, although we point out some possible reasons in the light of current theoretical calculations.