Renner-Teller quantum dynamics of the N(2D) + H2-->NH + H reaction

We present the Born-Oppenheimer (BO) and Renner-Teller (RT) quantum dynamics of the reaction (14)N((2)D)+(1)H(2)(X (1)Sigma(g) (+))-->NH(X (3)Sigma(-))+H((2)S), considering the NH(2) electronic states X (2)B(1) and A (2)A(1). These states correlate to the same (2)Pi(u) linear species, are coupled by RT nonadiabatic effects, and give NH(X (3)Sigma(-))+H and NH(a (1)Delta)+H, respectively. We develop the Hamiltonian matrix elements in the R embedding of the Jacobi coordinates and in the adiabatic electronic representation, using the permutation-inversion symmetry, and taking into account the nuclear-spin statistics. Collision observables are calculated via the real wave-packet (WP) and flux methods, using the potential-energy surfaces of Santoro et al. [J. Phys. Chem. A 106, 8276 (2002)]. WP snapshots show that the reaction proceeds via an insertion mechanism, and that the RT-WP avoids the A (2)A(1) potential barrier, jumping from the excited to the ground surface and giving mainly the NH(X (3)Sigma(-)) products. X (2)B(1) BO probabilities and cross sections show large tunnel effects and are approximately four to ten times larger than the A (2)A(1) ones. This implies a BO rate-constant ratio k(X (2)B(1))k(A (2)A(1)) approximately 10(5) at 300 K, i.e., a negligible BO formation of NH(a (1)Delta). When H(2) is rotationally excited, RT couplings reduce slightly the X (2)B(1) reaction observables, but enhance strongly the A (2)A(1) reactivity. These couplings are important at all collision energies, reduce the collision threshold, and increase remarkably reaction probabilities and cross sections. The RT k(A (2)A(1)) is thus approximately 3.3 order of magnitude larger than the BO value, and degeneracy-averaged, initial-state-resolved rate constants increase by approximately 13% and by approximately 47% at 300 and 500 K, respectively. Owing to an overestimation of the X (2)B(1) potential barrier, the calculated thermal rate is too low with respect to that observed, but we obtain a good agreement by shifting down the calculated cross section.