A theoretical study of rate coefficients for the O + NO vibrational relaxation

Quasi-classical trajectories have been integrated to study the vibrational relaxation of the O + NO(v) process as a function of the initial vibrational quantum number for T = 298 K, 1500 K, and 3000 K. Two reliable potential energy surfaces have been employed for the A' and A' doublet states of NO2. The calculated vibrational relaxation rate constants show a nearly v-independent behavior at room temperature and a moderate increase with v for higher temperatures. Although deviating significantly from the recommended values, good agreement with recent experimental results has been obtained. The importance of multi-quantum transitions is also analyzed.     

Vibrationally resolved rate coefficients and branching fractions in the dissociative recombination of O2+

We have studied the dissociative recombination of the first three vibrational levels of O(2) (+) in its electronic ground X (2)Pi(g) state. Absolute rate coefficients, cross sections, quantum yields and branching fractions have been determined in a merged-beam experiment in the heavy-ion storage ring, CRYRING, employing fragment imaging for the reaction dynamics. We present the absolute total rate coefficients as function of collision energies up to 0.4 eV for five different vibrational populations of the ion beam, as well as the partial (vibrationally resolved) rate coefficients and the branching fractions near 0 eV collision energy for the vibrational levels v=0, 1, and 2. The vibrational populations used were produced in a modified electron impact ion source, which has been calibrated using Cs-O(2)(+) dissociative charge transfer reactions. The measurements indicate that at low collision energies, the total rate coefficient is weakly dependent on the vibrational excitation. The calculated thermal rate coefficient at 300 K decreases upon vibrational excitation. The partial rate coefficients as well as the partial branching fractions are found to be strongly dependent on the vibrational level. The partial rate coefficient is the fastest for v=0 and goes down by a factor of two or more for v=1 and 2. The O((1)S) quantum yield, linked to the green airglow, increases strongly upon increasing vibrational level. The effects of the dissociative recombination reactions and super elastic collisions on the vibrational populations are discussed.     

On the role of dynamical barriers in barrierless reactions at low energies: S(1D) + H2

Reaction probabilities as a function of total angular momentum (opacity functions) and the resulting reaction cross sections for the collision of open shell S((1)D) atoms with para-hydrogen have been calculated in the kinetic energy range 0.09-10 meV (1-120 K). The quantum mechanical hyperspherical reactive scattering method and quasi-classical trajectory and statistical quasi-classical trajectory approaches were used. Two different ab initio potential energy surfaces (PESs) have been considered. The widely used reproducing kernel Hilbert space (RKHS) PES by Ho et al. [T.-S. Ho, T. Hollebeek, H. Rabitz, S. D. Chao, R. T. Skodje, A. S. Zyubin, and A. M. Mebel, J. Chem. Phys 116, 4124 (2002)] and the recently published accurate double many-body expansion (DMBE)/complete basis set (CBS) PES by Song and Varandas [Y. Z. Song and A. J. C. Varandas, J. Chem. Phys. 130, 134317 (2009)]. The calculations at low collision energies reveal very different dynamical behaviors on the two PESs. The reactivity on the RKHS PES is found to be considerably larger than that on the DMBE/CBS PES as a result of larger reaction probabilities at low total (here also orbital) angular momentum values and to opacity functions which extend to significantly larger total angular momentum values. The observed differences have their origin in two major distinct topographic features. Although both PESs are essentially barrierless for equilibrium H-H distances, when the H-H bond is compressed the DMBE/CBS PES gives rise to a dynamical barrier which limits the reactivity of the system. This barrier is completely absent in the RHKS PES. In addition, the latter PES exhibits a van der Walls well in the entrance channel which reduces the height of the centrifugal barrier and is able to support resonances. As a result, a significant larger cross section is found on this PES, with marked oscillations attributable to shape resonances and/or to the opening of partial wave contributions. The comparison of the results on both PESs is illustrative of the wealth of the dynamics at low collision energy. It is also illuminating about the difficulties encountered in modeling an all-purpose global potential energy surface.     

Fully Coriolis-coupled quantum studies of the H + O2 (upsilon i = 0-2, j i = 0,1) --> OH + O reaction on an accurate potential energy surface: integral cross sections and rate constants

We present accurate quantum calculations of the integral cross section and rate constant for the H + O2 --> OH + O combustion reaction on a recently developed ab initio potential energy surface using parallelized time-dependent and Chebyshev wavepacket methods. Partial wave contributions up to J = 70 were computed with full Coriolis coupling, which enabled us to obtain the initial state-specified integral cross sections up to 2.0 eV of the collision energy and thermal rate constants up to 3000 K. The integral cross sections show a large reaction threshold due to the quantum endothermicity of the reaction, and they monotonically increase with the collision energy. As a result, the temperature dependence of the rate constant is of the Arrhenius type. In addition, it was found that reactivity is enhanced by reactant vibrational excitation. The calculated thermal rate constant shows a significant improvement over that obtained on the DMBE IV potential, but it still underestimates the experimental consensus.     

Experimental and theoretical study of the reaction of the ethynyl radical with nitrous oxide, C2H + N2O

We investigated the rate constants and reaction mechanism of the gas phase reaction between the ethynyl radical and nitrous oxide (C(2)H + N(2)O) using both experimental methods and electronic structure calculations. A pulsed-laser photolysis/chemiluminescence technique was used to determine the absolute rate coefficient over the temperature range 570 K to 836 K. In this experimental temperature range, the measured temperature dependence of the overall rate constants can be expressed as: k(T) (C(2)H + N(2)O) = 2.93 × 10(-11) exp((-4000 ± 1100) K/T) cm(3) s(-1) (95% statistical confidence). Portions of the C(2)H + N(2)O potential energy surface (PES), containing low-energy pathways, were constructed using the composite G3B3 method. A multi-step reaction route leading to the products HCCO + N(2) is clearly preferred. The high selectivity between product channels favouring N(2) formation occurs very early. The pathway corresponds to the addition of the terminal C atom of C(2)H to the terminal N atom of N(2)O. Refined calculations using the coupled-cluster theory whose electronic energies were extrapolated to the complete basis set limit CCSD(T)/CBS led to an energy barrier of 6.0 kcal mol(-1) for the entrance channel. The overall rate constant was also determined by application of transition-state theory and Rice-Ramsperger-Kassel-Marcus (RRKM) statistical analyses to the PES. The computed rate constants have similar temperature dependence to the experimental values, though were somewhat lower.     

Influence of the potential energy surface on the reaction cross section: the K + HF → KF + H system

The effect of the potential energy surface on the K + HF → KF + H cross section has been studied using reasonable Sorbie-Murrell (bent saddle point) and LEPS (collinear saddle point) potential energy surfaces (PESs). Trajectory calculations for selected initial conditions (translational energies, rovibrational levels (v, J) of HF, as well as initial parallel or perpendicular alignments between the HF rotational angular momentum and the reactants relative velocity vectors) have been performed on these PESs to compare them with experiments. The Sorbie-Murrell and LEPS-4 PESs lead to steric effect ratio results quite close to the experimental ones, once the error margins are included. The resllts point towards a bent K-F-H saddle point although the PES is very isotropic. This could explain why experimental determinations lead to suggest a collinear saddle point. The K + HF → KF + H reaction exhibits an enormous vibrational enhancement of reactivity with one quantum HF vibrational excitation, even at translational energies well above the HF(v=0) threshold, where tunnelling effect contribution to reactivity can be neglected. This behaviour has not been reproduced in the trajectory calculations and no satisfactory explanation has been obtained for this fact. Nevertheless, the HF(v=1)/HF(v=0) cross section ratio at translational energies not far from the HF(v=0) threshold and the relative cross section for HF(v=0) have been satisfactorily descibed. In what regards rotation, the best theoretical results are those corresponding to the Sorbie-Murrell PES (the cross section increases with J), although important differences with experiment appear for the J = 0–3 interval at the lower translational energy values considered (0.54 and 0.77 eV).     

Coupled-states statistical investigation of vibrational and rotational relaxation of OH(2pi) by collisions with atomic hydrogen

We report state-to-state cross sections and thermal rate constants for vibrational and rotational relaxation of OH(2pi) by collision with H atoms. The cross sections are calculated by the coupled-states (CS) statistical method including the full open-shell character of the OH + H system. Four potential energy surfaces (PESs) ((1,3)A' and (1,3)A') describe the interaction of OH(X2pi) with H atoms. Of these, three are repulsive, and one (1A') correlates with the deep H2O well. Consequently, rotationally and ro-vibrationally inelastic scattering of OH in collisions with H can occur by scattering on the repulsive PESs, in a manner similar to the inelastic scattering of OH by noble gas atoms, or by collisions which enter the H2O well and then reemerge. At 300 K, we predict large (approximately 1 x 10(-10) cm3 molecule(-1) s(-1)) vibrational relaxation rates out of both v = 2 and v = 1, comparable to earlier experimental observations. This anomalously fast relaxation results from capture into the H2O complex. There exists a significant propensity toward formation of OH in the pi(A') lambda-doublet level. We also report state-resolved cross sections and rate constants for rotational excitation within the OH v = 0 manifold. Collisional excitation from the F1 to the F2 spin-orbit manifold leads to an inverted lambda-doublet population.     

Quantifying the non-RRKM effect in the H + O2 ⇄ OH + O reaction

In the present investigation the non-RRKM behavior in the title reaction is quantified in two different ways: (1) Quasiclassical trajectory calculations of the thermal rate coefficient are compared with results from a microcanonical variational transition-state theory/RRKM model. Results on both the Varandas DMBE IV and Melius-Blint potentials indicate that the non-RRKM behavior acts to reduce the thermal rate coefficient by about a factor of two, independent of temperature from 250 K to 5500 K. The QCT thermal rate coefficients on the two potentials are in remarkably good agreement with experiment and with each other over the entire temperature range. (2) The non-RRKM behavior as a classical phenomenon is demonstrated and quantified on both potentials by a direct test of the fundamental assumption. Complex-forming classical trajectories, started as either O + OH or H + O₂, are shown preferentially to return to the region of configuration space from which they were started. This test is discussed in detail in the text. The transition of the non-RRKM behavior from classical to quantum mechanics is also discussed. © 1997 John Wiley & Sons, Inc. Int J Chem Kinet 29: 275–287, 1997.     

Study of the H+O2 reaction by means of quantum mechanical and statistical approaches: the dynamics on two different potential energy surfaces

The possible existence of a complex-forming pathway for the H+O(2) reaction has been investigated by means of both quantum mechanical and statistical techniques. Reaction probabilities, integral cross sections, and differential cross sections have been obtained with a statistical quantum method and the mean potential phase space theory. The statistical predictions are compared to exact results calculated by means of time dependent wave packet methods and a previously reported time independent exact quantum mechanical approach using the double many-body expansion (DMBE IV) potential energy surface (PES) [Pastrana et al., J. Phys. Chem. 94, 8073 (1990)] and the recently developed surface (denoted XXZLG) by Xu et al. [J. Chem. Phys. 122, 244305 (2005)]. The statistical approaches are found to reproduce only some of the exact total reaction probabilities for low total angular momenta obtained with the DMBE IV PES and some of the cross sections calculated at energy values close to the reaction threshold for the XXZLG surface. Serious discrepancies with the exact integral cross sections at higher energy put into question the possible statistical nature of the title reaction. However, at a collision energy of 1.6 eV, statistical rotationally resolved cross sections managed to reproduce the experimental cross sections for the H+O(2)(v=0,j=1)-->OH(v(')=1,j('))+O process reasonably well.     

Kinetic study of the reaction Ca+ + N2O from 188 to 1207 K

Ion-molecule reactions involving metallic species play a central role in the chemistry of planetary ionospheres and in many combustion processes. The kinetics of the Ca(+) + N(2)O --> CaO(+) + N(2) reaction was studied by the pulsed multiphoton dissociation at 193 nm of organo-calcium vapor in the presence of N(2)O, followed by time-resolved laser-induced fluorescence spectroscopy of Ca(+) at 393.37 nm (4(2)P(3/2) <-- 4(2)S(1/2)). This yielded k(188-1207 K) = 5.45 x 10(-11) (T/300 K)(0.53) exp(282 K/T) cm(3) molecule(-1) s(-1), with an estimated accuracy of +/-13% (188-600 K) and +/-27% (600-1207 K). The temperature dependence of this barrierless reaction, with a minimum in the rate coefficient between 400 and 600 K, appears to be explained by the role of N(2)O vibrational excitation. This is examined using a classical trajectory treatment on a potential energy surface calculated at the B3LYP/6-311+g(2d,p) level of theory.     

The relaxation of OH (v=1) and OD (v=1) by H2O and D2O at temperatures from 251 to 390 K

We report rate coefficients for the relaxation of OH(v=1) and OD(v=1) by H2O and D2O as a function of temperature between 251 and 390 K. All four rate coefficients exhibit a negative dependence on temperature. In Arrhenius form, the rate coefficients for relaxation (in units of 10(-12) cm3 molecule-1 s-1) can be expressed as: for OH(v=1)+H2O between 263 and 390 K: k=(2.4+/-0.9) exp((460+/-115)/T); for OH(v=1)+D2O between 256 and 371 K: k=(0.49+/-0.16) exp((610+/-90)/T); for OD(v=1)+H2O between 251 and 371 K: k=(0.92+/-0.16) exp((485+/-48)/T); for OD(v=1)+D2O between 253 and 366 K: k=(2.57+/-0.09) exp((342+/-10)/T). Rate coefficients at (297+/-1 K) are also reported for the relaxation of OH(v=2) by D2O and the relaxation of OD(v=2) by H2O and D2O. The results are discussed in terms of a mechanism involving the formation of hydrogen-bonded complexes in which intramolecular vibrational energy redistribution can occur at rates competitive with re-dissociation to the initial collision partners in their original vibrational states. New ab initio calculations on the H2O-HO system have been performed which, inter alia, yield vibrational frequencies for all four complexes: H2O-HO, D2O-HO, H2O-DO and D2O-DO. These data are then employed, adapting a formalism due to Troe (J. Troe, J. Chem. Phys., 1977, 66, 4758), in order to estimate the rates of intramolecular energy transfer from the OH (OD) vibration to other modes in the complexes in order to explain the measured relaxation rates-assuming that relaxation proceeds via the hydrogen-bonded complexes.     

A quasiclassical trajectory study for the N(4S)+O2(X 3Sigmag-)-->NO(X 2Pi)+O(3P) reaction on the new 2A' and 4A' potential-energy surfaces

A quasiclassical trajectory study with the sixth-order explicit symplectic algorithm of the N(4S)+O2(X 3Sigmag-)-->NO(X 2Pi)+O(3P) atmospheric reaction has been performed by employing the new 2A' and 4A' potential-energy surfaces reported by Sayos et al. [J. Chem. Phys. 117, 670 (2002)]. For the translational temperature considered up to 10,000 K, the larger relative translational energy and the higher rovibrational levels of O2 molecule with respect to the previous works have been taken into account, and a clearer database about the character of the total reaction cross section has been presented in this work. The dependence of microscopic rate constants on the vibrational level of O2 molecule at T=3000, 5000, and 10,000 K has been exhibited, and we can see that the values of log10 k(T,v,J) vary almost linearly with the vibrational level of O2 molecule. The thermal rate constants at the translational temperature between 300 and 10,000 K have been evaluated and compared with the experimental and previous theoretical results. It is found that the thermal rate constants determined in this work have a better agreement with the experimental data and can provide a more valid theoretical reference at the translational temperature considered for the title reaction.     

Quasiclassical trajectories on a finite element density functional potential energy surface: The C++H2O reaction revisited

A new method for the representation of potential energy surfaces (PESs) based on the p version of the finite element method is presented and applied to the PES of the [COH2]+ system in order to study the C++H2O-->[COH]++H reaction through the quasiclassical trajectory method. Benchmark ab initio computations have been performed on the most relevant stationary points of the PES through a procedure that incorporates basis set extrapolations, the contribution of the core correlation energy, and scalar relativistic corrections. The electronic structure method employed to compute the many points needed to construct the PES is a hybrid density functional approach of the B3LYP type with geometry-dependent parameters, which improves dramatically the performance with respect of the B3LYP method. The trajectory computations shed light on the behavior of the COH2+ complex formed in the collision. At a fixed relative translational energy of 0.62 eV, which corresponds to the crossed beam experiments [D. M. Sonnenfroh et al., J. Chem. Phys. 83, 3985 (1985)], the complex dissociates significantly into the reactants (37%). However, the behavior for a thermal sampling at T=300 K is significantly different because only 9% of the trajectories where capture occurs lead to dissociation into the reactants. The latter kind of behavior is coherent with the view that simple ion-molecule reactions proceed quite often at the capture rate provided it is corrected by the fraction of the electronic states which, being nearly degenerate for the reactants, become attractive at short distances. For both T=300 K and crossed beam conditions, the trajectory computations indicate that COH2+ is the critical intermediate, in agreement with a recent work [Y. Ishikawa et al., Chem. Phys. Lett. 370, 490 (2003)] and in contrast with the interpretation of the crossed beam experiments. Besides, virtually all trajectories generate COH++H (>99%), but a significant proportion of the isoformyl cation is formed with enough vibrational energy as to surmount the COH+-HCO+ isomerization barrier, about 37% at T=300 K.     

State-to-state reaction probabilities for the H+O2(v,j)-->O+OH(v',j') reaction on three potential energy surfaces

We report state-to-state and total reaction probabilities for J=0 and total reaction probabilities for J=2 and 4 for the title reaction, both for ground-state and initially rovibrationally excited reactants. The results for three different potential energy surfaces are compared and contrasted. The potential energy surfaces employed are the DMBE IV surface by Pastrana et al. [J. Phys. Chem. 94, 8073 (1990)], the surface by Troe and Ushakov (TU) [J. Chem. Phys. 115, 3621 (2001)], and the new XXZLG ab initio surface by Xu et al. [J. Chem. Phys. 122, 244305 (2005)]. Our results show that the total reaction probabilities from both the TU and XXZLG surfaces are much smaller in magnitude for collision energies above 1.2 eV compared to the DMBE IV surface. The three surfaces also show different behavior with regards to the effect of initial state excitation. The reactivity is increased on the XXZLG and the TU surfaces and decreased on the DMBE IV surface. Vibrational and rotational product state distributions for the XXZLG and the DMBE IV surface show different behaviors for both types of distributions. Our results show that for energies above 1.25 eV the dynamics on the DMBE IV surface are not statistical. However, there is also evidence that the dynamics on the XXZLG surface are not purely statistical for energies above the onset of the first excited product vibrational state v'=1. The magnitude of the total reaction probability is decreased for J>0 for the DMBE IV and the XXZLG surfaces for ground-state reactants. However, for initially rovibrationally excited reactants, the total reaction probability does not decrease as expected for both surfaces. As a result the total cross section averaged over all Boltzmann accessible rotational states may well be larger than the cross section reported in the literature for j=1.     

A kinetic study of the reactions of Fe+ with N2O, N2, O2, CO2 and H2O, and the ligand-switching reactions Fe+.X + Y --> Fe+.Y + X (X = N2, O2, CO2; Y = O2, H2O)

A series of reactions involving Fe(+) ions were studied by the pulsed laser ablation of an iron target, with detection of ions by quadrupole mass spectrometry at the downstream end of a fast flow tube. The reactions of Fe(+) with N(2)O, N(2) and O(2) were studied in order to benchmark this new technique. Extending measurements of the rate coefficient for Fe(+) + N(2)O from 773 K to 185 K shows that the reaction exhibits marked non-Arrhenius behaviour, which appears to be explained by excitation of the N(2)O bending vibrational modes. The recombination of Fe(+) with CO(2) and H(2)O in He was then studied over a range of pressure and temperature. The data were fitted by RRKM theory combined with ab initio quantum calculations on Fe(+).CO(2) and Fe(+).H(2)O, yielding the following results (120-400 K and 0-10(3) Torr). For Fe(+) + CO(2): k(rec,0) = 1.0 x 10(-29) (T/300 K)(-2.31) cm(6) molecule(-2) s(-1); k(rec,infinity) = 8.1 x 10(-10) cm(3) molecule(-1) s(-1). For Fe(+) + H(2)O: k(rec,0) = 5.3 x 10(-29) (T/300 K)(-2.02) cm(6) molecule(-2) s(-1); k(rec,infinity) = 2.1 x 10(-9) (T/300 K)(-0.41) cm(3) molecule(-1) s(-1). The uncertainty in these rate coefficients is determined using a Monte Carlo procedure. A series of exothermic ligand-switching reactions were also studied at 294 K: k(Fe(+).N(2) + O(2)) = (3.17 +/- 0.41) x 10(-10), k(Fe(+).CO(2) + O(2)) = (2.16 +/- 0.35) x 10(-10), k(Fe(+).N(2) + H(2)O) = (1.25 +/- 0.14) x 10(-9) and k(Fe(+).O(2) + H(2)O) = (8.79 +/- 1.30) x 10(-10) cm(3) molecule(-1) s(-1), which are all between 36 and 52% of their theoretical upper limits calculated from long-range capture theory. Finally, the role of these reactions in the chemistry of meteor-ablated iron in the upper atmosphere is discussed. The removal rates of Fe(+) by N(2), O(2), CO(2) and H(2)O at 90 km altitude are approximately 0.1, 0.07, 3 x 10(-4) and 1 x 10(-6) s(-1), respectively. The initially formed Fe(+).N(2) and Fe(+).O(2) are converted into the H(2)O complex at approximately 0.05 s(-1). Fe(+).H(2)O should therefore be the most abundant single-ligand Fe(+) complex in the mesosphere below 90 km.     

Quantum mechanical rate constants for H + O2 <--> O + OH and H + O2 --> HO2 reactions

Canonical rate constants for both the forward and reverse H + O(2) <--> O + OH reactions were calculated using a quantum wave packet-based statistical model on the DMBE IV potential energy surface of Varandas and co-workers. For these bimolecular reactions, the results show reasonably good agreement with available experimental and theoretical data up to 1500 K. In addition, the capture rate for the H + O(2) --> HO(2) addition reaction at the high-pressure limit was obtained on the same potential using a time-independent quantum capture method. Excellent agreement with experimental and quasi-classical trajectory results was obtained except for at very low temperatures, where a reaction threshold was found and attributed to the centrifugal barrier of the orbital motion.     

Non-RRKM dynamics in the CH3O2 + NO reaction system

Quasi-classical trajectory (QCT) calculations on a model potential energy surface (PES) show strong deviations from statistical Rice-Ramsperger-Kassel-Marcus (RRKM) rate theory for the decomposition reaction (1) CH3OONO* --> CH3O + NO2, where the highly excited CH3OONO* was formed by (2) CH3O2 + NO --> CH3OONO*. The model PES accurately describes the vibrational frequencies, structures, and thermochemistry of the cis- and trans-CH3OONO isomers; it includes cis-trans isomerization in addition to reactions 1 and 2 but does not include nitrate formation, which is too slow to affect the decay rate of CH3OONO*. The QCT results give a strongly time-dependent rate constant for decomposition and damped oscillations in the decomposition rate, not predicted by statistical rate theory. Anharmonicity is shown to play an important role in reducing the rate constant by a factor of 10 smaller than predicted using classical harmonic RRKM theory (microcanonical variational transition state theory). Master equation simulations of organic nitrate yields published previously by two groups assumed that RRKM theory is accurate for reactions 1 and 2 but required surprising parametrizations to fit experimental nitrate yield data. In the present work, it is hypothesized that the non-RRKM rate of reaction (1) and vibrational anharmonicity are at least partly responsible for the surprising parameters.     

Quasiclassical trajectory study of the vibrational quenching of hydroxyl radicals through collision with O atoms

The collisional removal of vibrationally excited OH radicals by O atoms is studied by the quasiclassical trajectory method. To evaluate the effect of different topological features on the scattering processes two different global potential energy surfaces, DMBE IV and TU, are used. Results for reactive, exchange, and inelastic scattering probabilities are reported for central collisions (with zero total angular momentum) with a fixed relative translational energy for vibrational levels of OH ranging from nu=1 to v=8. Vibrational state distributions of product molecules are also compared on the two potential energy surfaces. Both surfaces predict higher probabilities for reaction than for exchange or inelastic scattering. The vibrational state distributions of the product diatomic molecules are different on the two surfaces. In particular, the two surfaces give substantially different probabilities for multiquantum OH vibrational relaxation transitions OH(v)+O-->OH(v')+O.     

Quasi-classical trajectory study of Si+O2-->SiO+O reaction

Quasi-classical trajectory calculations for the Si(3P)+O2(X 3Sigmag-)-->SiO(X 1Sigma+)+O(1D) reaction have been carried out using the analytical ground 1A' potential energy surface (PES) recently reported by Dayou and Spielfiedel [J. Chem. Phys. 119, 4237 (2003)]. The reaction has been studied for a wide range of collision energies (0.005-0.6 eV) with O2 in its ground rovibrational state. The barrierless PES leads to a decrease of the total reaction cross section with increasing collision energy. It has been brought to evidence that the reaction proceeds through different reaction mechanisms whose contributions to reactivity are highly dependent on the collision energy range. At low collision energy an abstraction mechanism occurs involving the collinear SiOO potential well. The associated short-lived intermediate complex leads to an inverted vibrational distribution peaked at v'=3 and low rotational excitation of SiO(v',j') with a preferentially backward scattering. At higher energies the reaction proceeds mainly through an insertion mechanism involving the bent and linear OSiO deep potential wells and associated long-lived intermediate complexes, giving rise to nearly statistical energy disposals into the product modes and a forward-backward symmetry of the differential cross section.