Quantum and classical studies of the O(3P)+H2(v=0–3,j=0)→OH+H reaction using benchmark potential surfaces

Rate coefficients for the O + H2 and O + D2 reactions: how well ring polymer molecular dynamics accounts for tunelling

We present here extensive calculations of the O(3P) + H2 and O(3P) + D2 reaction dynamics spanning the temperature range from 200 K to 2500 K. The calculations have been carried out using fully converged time-independent quantum mechanics (TI QM), quasiclassical trajectories (QCT) and ring polymer molecular dynamics (RPMD) on the two lowest lying adiabatic potential energy surfaces (PESs), 13A' and 13A'', calculated by Zanchet et al. [J. Chem. Phys., 2019, 151, 094307]. TI QM rate coefficients were determined using the cumulative reaction probability formalism on each PES including all of the total angular momenta and the Coriolis coupling and can be considered to be essentially exact within the Born-Oppenheimer approximation. The agreement between the rate coefficients calculated by using QM and RPMD is excellent for the reaction with D2 in almost the whole temperature range. For the reaction with H2, although the agreement is very good above 500 K, the deviations are significant at lower temperatures. In contrast, the QCT calculations largely underestimate the rate coefficients for the two isotopic variants due to their inability to account for tunelling. The differences found in the disagreements between RPMD and QM rate coefficients for the reactions for both the isotopologues are indicative of the ability of the RPMD method to accurately describe systems where tunelling plays a relevant role. Considering that both reactions are dominated by tunelling below 500 K, the present results show that RPMD is a very powerful tool for determining rate coefficients. The present QM rate coefficients calculated on adiabatic PESs slightly underestimate the best global fits of the experimental measurements, which we attribute to the intersystem crossing with the singlet 11A' PES.

Unexpected dynamical effects change the lambda-doublet propensity in the tunneling region for the O(3P) + H2 reaction

One of the most relevant features of the O(³P) + H₂ reaction is that it occurs on two different potential energy surfaces (PESs) of symmetries A' and A'' that correlate reactants and products. The respective saddle points, which correspond to a collinear arrangement, are the same for both PESs, whilst the barrier height rises more abruptly on the ³A' PES than on the ³A'' PES. Accordingly, the reactivity on the ³A'' PES should be always higher than on the ³A' PES. In this work, we present accurate quantum-scattering calculations showing that this is not always the case for rotationless reactants, where dynamical factors near the reaction threshold cause the ³A' PES to dominate at energies around the barrier. Further calculation of cross sections and Λ-doublet populations has allowed us to establish how the reaction mechanism changes from the deep tunneling regime to hyperthermal energies.

Quantum reactive scattering of O(3P)+H2 at collision energies up to 4.4 eV

We report the results of quantum scattering calculations for the O((3)P)+H2 reaction for a range of collision energies from 0.4 to 4.4 eV, important for astrophysical and atmospheric processes. The total and state-to-state reactive cross sections are calculated using a fully quantum time-independent coupled-channel approach on recent potential energy surfaces of (3)A' and (3)A″ symmetry. A larger basis set than in the previous studies was used to ensure single-surface convergence at higher energies. Our results agree well with the published data at lower energies and indicate the breakdown of reduced dimensionality approach at collision energies higher than 1.5 eV. Differential cross sections and momentum transfer cross sections are also reported.

Complete state-resolved non-adiabatic dynamics of the O((3)P) + D2 → OD(X(2)Π) + D reaction

The first quantum-state-resolved distributions over the full range of available product levels are reported for any isotopic variant of the elementary reaction of O((3)P) with molecular hydrogen. A laser-detonation source was used to produce a hyperthermal oxygen-atom beam, which allowed for sufficient collision energy to surmount the reaction barrier. This beam was crossed by a supersonic beam of D2. The nascent OD products were detected by laser-induced fluorescence. OD rotational distributions in vibrational states v' = 0, 1, and 2 at a collision energy of 25 kcal mol(-1) are reported, together with distributions for the dominant product vibrational level, v'= 0, at lower collision energies of 20 and 15 kcal mol(-1). The OD product is highly rotationally excited, to a degree that declines as expected for the higher vibrational levels or for reductions in the collision energy. The measured rovibrational distributions at the highest collision energy are in excellent agreement with previous theoretical predictions based on quantum scattering calculations on the triplet potential energy surfaces developed by Rogers et al. (J. Phys. Chem. A 2000, 104, 2308-2325). However, no significant OD spin-orbit preference was observed, in contrast to the predictions of most existing theoretical models of the non-adiabatic dynamics based on the widely used reduced-dimensional four-state model of Hoffmann and Schatz (J. Chem. Phys. 2000, 113, 9456-9465). Furthermore, a clear observed preference for OD Π(A') Λ-doublet levels is not consistent with a simple extrapolation of the calculated relative reaction cross sections on intermediate surfaces of (3)A' and (3)A″ symmetry.

Product-state-resolved dynamics of the elementary reaction of atomic oxygen with molecular hydrogen, O(³P) + D₂ → OD(X²Π) + D

Elementary three-atom systems provide stringent tests of the accuracy of ab initio theory. One such important reaction, O((3)P) + H2 → OH(X(2)Π) + H, has eluded detailed experimental study because of its high activation barrier. In this reaction, both the ground-state reactant atom and product diatomic molecule have open-shell character, which introduces the intriguing complication of non-Born-Oppenheimer effects in both the entrance and the exit channels. These effects may be probed experimentally in both the fine-structure and the Λ-doublet splittings of the OH product. We have used laser-induced fluorescence to measure OD internal product-state distributions from the analogous reaction of O((3)P) with D2, enabled by a unique high-energy O((3)P) source. We find that the OD (ν' = 0) product is rotationally highly excited, in excellent agreement with earlier theoretical predictions. However, the distributions over the OD(X(2)Π) fine-structure and Λ-doublet states, diagnostic of electronic non-adiabaticity in the reaction, challenge the prevailing theoretical understanding.

Nonadiabatic state-to-state reactive collisions among open shell reactants with conical intersections: the OH((2)Pi) + F((2)P) example

Accurate wave packet calculations on the OH((2)Pi) + F((2)P) → O((3)P) + HF((1)Sigma(+)) reactive collisions are performed using a recently proposed coupled diabatic states. Adiabatic and nonadiabatic dynamics are compared in detail, analyzing the final state distribution of products. It is found that with the new surfaces a significant increase of the rate constant is obtained, with noticeable nonadiabatic effects. The inclusion of the spin-orbit splittings for the calculation of the electronic partition function produces an important increase of the reaction rate constants, yielding a rather good agreement with the experimental results. It is also concluded that spin-orbit couplings are also necessary in the entrance channel to describe this reaction.

Significant Nonadiabatic Effects in the S(1D) + HD Reaction

A nonadiabatic quantum dynamics calculation involving four coupled potential energy surfaces (two degenerate 3A' ', one 3A', and one 1A') and the spin-orbit coupling matrix for these states is reported for the title reaction. The results show that the important discrepancy between theoretically calculated and experimentally measured intramolecular isotope effects can at least in part be attributed to significant nonadiabatic effects.

Dynamics of the O(3P) + H2 reaction at low temperatures: comparison of quasiclassical trajectory with quantum scattering calculations

Quasiclassical trajectory and quantum-mechanical scattering calculations are reported for the O((3)P) + H(2)(X (1)Sigma(g) (+);upsilon = 1-3,j = 0)-->OH(X (2)Pi) + H((2)S) reaction at energies close to the reaction threshold. The dynamics of the reaction have been investigated for zero total angular momentum using the lowest (3)A" potential-energy surface developed by Rogers et al. [J. Phys. Chem. A 104, 2308 (2000)] and its recent extensions by Brandao et al. [J. Chem. Phys. 121, 8861 (2004)] which provide an improved description of the van der Waals interaction. Good agreement is observed for this system between quasiclassical and quantal results for incident kinetic energies above the tunneling regime. Quantum-mechanical calculations also confirm recent theoretical predictions of a strong collision-energy dependence of the OH(v(') = 0)OH(v' = 1) product branching ratio in the O((3)P) + H(2)(v = 1) reaction, which explains the differences observed in OH vibrational populations between experiments using different O((3)P) sources.

An ab initio investigation of the O(3P)–H2(1Σ+g) van der Waals well

We report an ab initio study of the van der Waals region of the O(3P)-H2 potential energy surface based on RCCSD(T) calculations with an aug-cc-pVQZ basis supplemented by bond functions. In addition, an open-shell implementation of symmetry-adapted perturbation theory (SAPT) is used to corroborate the RCCSD(T) calculations and to investigate the relative magnitudes of the various contributions to the van der Waals interaction. We also investigate the effect of the spin-orbit coupling on the position and depth of the van der Waals well. We predict the van der Waals minimum to occur in perpendicular geometry, and located at a closer distance than a secondary well in colinear geometry. The potentials obtained in the present study confirm the previous calculations of Alexander [M. H. Alexander, J. Chem. Phys., 1998, 108, 4467], but disagree with the earlier work of Harding and co-workers [Z. Li, V. A. Apkarian and L. B. Harding, J. Chem. Phys., 1997, 106, 942] as well as with recently refitted surfaces of Brandão and coworkers [J. Brandão, C. Mogo and B. C. Silva, J. Chem. Phys., 2004, 121, 8861]. Inclusion of spin-orbit coupling reduces the depth of the van der Waals minimum without causing a change in its position.

Experimental and Theoretical Investigations of the Inelastic and Reactive Scattering Dynamics of O(3P) + D2

This paper presents a combined experimental and theoretical study of the dynamics of O((3)P) + D(2) collisions, with emphasis on a center-of-mass (c.m.) collision energy of 25 kcal mol(-1). The experiments were conducted with a crossed-molecular-beams apparatus, employing a laser detonation source to produce hyperthermal atomic oxygen and mass spectrometric detection to measure the product angular and time-of-flight distributions. The novel beam source, which enabled these experiments to be conducted, contributed unique challenges to the experiments and to the analysis, so the experimental methods and approach to the analysis are discussed in detail. Three different levels of theory were used: (1) quasiclassical trajectories (QCT), (2) time-independent quantum scattering calculations based on high-quality potential surfaces for the two lower-energy triplet states, and (3) trajectory-surface-hopping (TSH) studies that couple the triplet surfaces with the lowest singlet surface using a spin-orbit Hamiltonian derived from ab-initio calculations. The latter calculations explore the importance of intersystem crossing in the dynamics. Both experiment and theory show that inelastically scattered O atoms scatter almost exclusively in the forward direction, with little or no loss of translational energy. For the reaction, O((3)P) + D(2) --> OD + D, the experiment shows that, on average, approximately 50% of the available energy goes into product translation and that the OD product angular distributions are largely backward-peaked. These results may be interpreted in light of the QCT and TSH calculations, leading to the conclusion that the reaction occurs mainly on triplet potential energy surfaces with, at most, minor intersystem crossing to a singlet surface. Reaction on either of the two low-lying reactive triplet surfaces proceeds through a rebound mechanism in which the angular distributions are backward-peaked and the product OD is both vibrationally and rotationally excited. The quantum scattering results are in good agreement with QCT calculations, indicating that quantum effects are relatively small for this reaction at a collision energy of 25 kcal mol(-1).

A quantum wave-packet study of intersystem crossing effects in the O(P2,1,03,D21)+H2 reaction

We present for the first time an exact quantum study of spin-orbit-induced intersystem crossing effects in the title reaction. The time-dependent wave-packet method, combined with an extended split operator scheme, is used to calculate the fine-structure resolved cross section. The calculation involves four electronic potential-energy surfaces of the 1A' state [J. Dobbyn and P. J. Knowles, Faraday Discuss. 110, 247 (1998)], the 3A' and the two degenerate 3A" states [S. Rogers, D. Wang, A. Kuppermann, and S. Wald, J. Phys. Chem. A 104, 2308 (2000)], and the spin-orbit couplings between them [B. Maiti, and G. C. Schatz, J. Chem. Phys. 119, 12360 (2003)]. Our quantum dynamics calculations clearly demonstrate that the spin-orbit coupling between the triplet states of different symmetries has the greatest contribution to the intersystem crossing, whereas the singlet-triplet coupling is not an important effect. A branch ratio of the spin state Pi32 to Pi12 of the product OH was calculated to be approximately 2.75, with collision energy higher than 0.6 eV, when the wave packet was initially on the triplet surfaces. The quantum calculation agrees quantitatively with the previous quasiclassical trajectory surface hopping study.

Potential energy surface for H2O(3A″) from accurateab initiodata with inclusion of long-range interactions

A new potential energy surface for the O((3)P) + H(2) system in the lowest (3)A(") state is built using ab initio data calculated by Rogers et al. [J. Phys. Chem. A 104, 2308 (2000)] and the double many-body expansion formalism. It incorporates a semiempirical model of long-range interactions, which should play an important role at low collision energies. Preliminary quasiclassical trajectory results at 12.6 kcal/mol collision energy, show that the deeper van der Waals region described in this new surface translates into a four times higher cross section than that of Rogers' (3)A(") surface. To confirm this hypothesis, a second surface was calibrated. The two surfaces are fitted with rmsd<0.5 kcal/mol and differ mainly on the depth of the van der Waals region. That difference in the van der Waals region corresponds to a 22% lower cross section of the less deep surface, which is still three times higher than the equivalent results from Rogers' (3)A(") surface. This study reflects the importance of a correct description of van der Waals forces on potential energy surfaces.

Quantum calculations of the O(3P)+H2→OH+H reaction

Quantum scattering calculations are reported for the O(3P)+H2(v=0,1) reaction using chemically accurate potential energy surfaces of 3A' and 3A" symmetry. We present state-to-state reaction cross sections and rate coefficients as well as thermal rate coefficients for the title reaction using accurate quantum calculations. Our calculations yield reaction cross sections that are in quantitative accord with results of recent crossed molecular beam experiments. Comparisons with results obtained using the J-shifting calculations show that the J-shifting approximation is quite reliable for this system. Thermal rate coefficients from the exact calculations and the J-shifting approximation agree remarkably well with experimental results. Our calculations also reproduce the markedly different OH(v'=0)/OH(v'=1) branching in O(3P)+H2(v=1) reaction, observed in experiments that use different O(3P) atom sources. In particular, we show that the branching ratio is a strong function of the kinetic energy of the O(3P) atom.