J. CHEM. SOC. FARADAY TRANS., 1994, 90(13), 1811-1817 181 1 Ab lnitio Quantum Chemistry Study of the Gas-phase Reaction of CIO with HO, David Buttar and David M. Hirst* Department of Chemistry, University of Warwick, Coventry, UK CV4 7AL The reaction of CIO with HO, has been studied using ab initio molecular orbital theory with electron correlation being taken into account by Msller-Plesset (MP) perturbation theory. Saddle-point geometries, energies and harmonic vibrationai wavenumbers have been calculated at the MP2/6-31 G** level and barrier heights at the MP4/6-311G** level. The study shows that the products and mechanism depend on which potential-energy surface the reaction occurs. If the reaction proceeds along the triplet reaction surface, ground-state HOCl ('A') and ground-state 0, (3Xg-)are formed.If the reaction takes place on the singlet reaction surface the same products are formed, but in this case the molecular oxygen is formed in a singlet state. The CIO + HO, reaction proceeds via a multi-step reaction mechanism on the singlet surface and via a direct hydrogen-abstraction mechanism on the triplet surface. The reaction of chlorine monoxide (C10) and the hydro- peroxy radical (HO,) has been studied '-*extensively because of its importance in atmospheric chemistry and its unusual rate behaviour. The reaction can proceed through a number of possible reaction channels C10 + HO, +ClOO + OH OClO + OH HOCl + 0, HClO + 0, HCl + 0, ArH (298 K)9-11 channel /kJ mol-' 1 ca.12 2 ca. 25 3 CU. -191 4 ca. 71 5 CU. -62 leading to a wide range of products. However, the reaction enthalpies indicate that channels (3) and (5) should be the principal reaction channels. The formation of ozone through channel (5) has not been detected in experimental studies of this system and this pathway is therefore thought to be of minor importance. The principal reaction channel is that leading to the formation of HOCl and 0,.This is supported by experimental observation of HOCl as a product of the C10 + HO, The formation of ClOO + OH through channel (1) is only slightly endothermic and may be a minor reaction pathway under suitable conditions. The remaining channels are too endothermic to be of importance to the C10 + HO, reaction. The C10 + HO, reaction is believed to be involved in the destruction of stratospheric ozone.The destruction cycle is a consequence of the photolysis of HOCl and is outlined below, C10 + HO, +HOCl + 0, HOCl + hv +Cl + OH Cl + 0,+ClO + 0, OH + 0, +HO, + 0, net 2 0, +3 0, This reaction scheme is thought to be particularly important in the middle and lower stratosphere. Experimental studies'-6 of the C10 + 0,H reaction have also found that the reaction displays a complicated tem-perature dependence, which is inconsistent with a simple H-atom metathesis reaction. The rate of the reaction is very fast and the experimentally measured rate constant at 298 K is about (4.5-6.5) x cm3 molecule-' s-l, and has a negative temperature dependence.The experimental results suggest that the reaction proceeds by a multi-step mech- anism, possibly via a reaction intermediate. This postulate is supported by the Arrhenius A factor determined from the experimental studies. This is found to be 4.6 x 10-l2 cm3 molecule-' s-'. The range in magnitudes of the rate constant arises from the different experimental techniques that have been employed to study this system. The concept of a multi- step reaction mechanism has been postulated by a number of author^,^^^.^-* but to date there is no experimental evidence for the existence of reaction intermediates. The study of short-lived reaction intermediates using experimental techniques is extremely difficult. However, theo- retical techniques can be used to study the possibility of the existence of such intermediates.There have been two theo- retical studies of the ClO + HO, system. Mozurkevich' has studied the reaction using RRKM theory and achieved good agreement with experimental rate parameters by assuming that the reaction proceeds by a multi-step mechanism involv- ing a weakly bound intermediate. The structures and proper- ties of the intermediates were determined from established thermochemical methods. The only ab initio theoretical study, to date, that has performed a full characterisation of reaction intermediates is the work of Toohey and Anderson.8 In this study various ab initio techniques were used to study the triplet reaction surface, under the constraint that any reaction intermediates would be formed with C, symmetry.The authors report the existence of a single saddle point that appears to connect the reactants to the products HOCl and 0,. The Arrhenius A factor derived from this work is found to be in the range 8.7 x 10-l3-5.4 x lo-', cm3 molecule-' s-', in reasonable agreement with experiment. Here we report a complete characterisation of the C10 + HO, reaction surface using second-order Mlaller-Plesset perturbation theory. In this work we have studied the singlet and triplet potential surfaces and report the structures, ener- gies and harmonic vibrational wavenumbers of various inter- mediates and transition states for the C10 + HO, reaction on both surfaces. In the calculations reported here no sym- metry constraint has been imposed.The stationary points located on each surface have been further studied using higher-order perturbation theory. ComputationalDetails Searches for minima and saddle points at the self-consistent field (SCF) level employed the restricted and unrestricted Hartree-Fock methods (RHF, UHF)l2,l3with the 6-31G** basis of Pople and co-worker~.~~*~~ To study the effect of electron correlation on the singlet and triplet potential- energy surfaces of the C10 + HO, reaction, minima and saddle-point calculations were also made using second-order Msller-Plesset perturbation theory (MP2).16 All stationary points located on the singlet and triplet surfaces were charac- terized by harmonic vibrational frequency calculations.The highest level of calculation reported is fourth-order Msller-Plesset (MP4)’ 7*1 calculations using the larger 6- 311G** basis set.” The MP4 calculations include all single, double, triple and quadruple excitations. Barrier heights for the C10 + HO, reaction are reported at all levels of theory. At the highest level, MP4/6-3 1lG**, basis-set superposition error calculations have been performed using the Boys- Bernardi Although the UHF wavefunctions are not true eigen-functions of the (S2> operator through contamination by higher spin states, the largest values of (S2) for the doublet and triplet states considered here were 0.77 and 2.07, respec- tively. There is only a small deviation from the expected values of 0.75 and 2.0, indicating only minor spin con-tamination.The study of the singlet reaction surface required the struc- ture and energy of the oxygen molecule in the ‘8, state to be considered. This calculation was performed using complex orbitals at the RHF and MP2 levels with the 6-31G** basis set. The optimum bond length for this molecule was deter- mined using a non-gradient optimization routine. The results are in good agreement with previously reported results.22 Intrinsic reaction coordinate (IRC)23724 calculations were performed at the MP2/6-31G** level to confirm that the transition states on each reaction surface connect with stable intermediates, reactants or products. The IRC calculations follow the minimum-energy pathway from the transition states.The initial search direction is determined from the imaginary frequency computed at the transition-state geometry. Searches were performed on both the forward and reverse sides of the potential surface with a stepsize of 0.1-0.3 a, u1I2. All calculations were performed using the GAUSSIAN 9225 and GAMESS-UK26,27 software packages on the Intel iPSC/860 and the Convex 220 at the SERC, Daresbury Laboratory, and a HP9000/735 workstation. Results The ClO + HO, reaction can proceed through two com-peting reaction pathways. The first involves the direct attack on the ooHbond of HO, by the singly occupied n* (ClO) orbital. The ground-state configuration for the HO, molecule 5a’),( la“)2(6a’)2(7a’)2(2a’’)1.is (1 a’)2(2a’)2(3a’)2(4a’)2( The oOH orbital of HO, corresponds to the doubly occupied, in plane, 7a’ orbital.The interaction of these orbitals results in the for- mation of a planar, slightly bent, triplet intermediate. The transition state formed from this interaction decomposes to form the ground-state products HOCl + O,, as studied pre- viously by Toohey and Anderson.8 The second reaction pathway arises from the interaction of the singly occupied 2a“ orbital of HO,, which is predominantly a n*(02) orbital, with the singly occupied n*(ClO) orbital. This pathway results in the formation of a singlet intermediate, which is believed to be entropically unfavourable.8 Decomposition of the singlet intermediate leads to the formation of HOCl + 0,.However, due to spin conservation rules, one of the products is formed in an excited state. Fig. 1 shows the orbital interactions for both reactions considered here. Here we present a comprehensive ab initio study of the singlet and triplet potential surfaces for the reaction of C10 with HO,. The results discussed are concerned with the for- J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 CIO + H02 -HOCl + 02 Fig. 1 Orbital diagram for the C10 + HO, reaction mation of hydrochlorous acid and molecular oxygen through reaction channel (3). Note that all possible reaction channels discussed previously were considered, but at the level of theory employed here only reaction channel (3) was found to be significant. There was no evidence, in this computational study, of the formation of any other products through the other possible reaction channels.Triplet Surface The calculated geometries, energies and harmonic vibrational wavenumbers of the reactants and products considered in this work are presented in Table 1. This table lists the results from calculations performed at the UHF/6-31G** and MP2/ 6-31G** levels. Also reported are results of single-point MPqSDTQ)/6-311G** calculations at the MP2 optimized geometries. The highest level of calculation reported for the singlet oxygen species is a MP2/6-3 1G** optimization. MP4 calculations on this species were not possible due to the SCF calculation requiring the use of complex orbitals. The MP4 results will be used later to compute ab initio barrier heights for the C10 + HO, reaction.The enthalpy of reaction for the ClO + HO, -,HOCl + 0, reaction under consideration is calculated, at the MP4/6-311G** level, to be ArH (0 K) = -241 kJ mol-I. This value is 50 kJ mol-’ larger than the experimental value of A,H (298 K) = -191 kJ mol-’. The computed heat of reaction could be improved by using higher levels of theory. Comparison of the experimental and calculated heat of reaction indicates that the C10 + HO, reaction surface is reasonably described at the MP4/6-311G** level. Table 2 summarises the results of the study of the triplet potential surface at the UHF/6-31G** and MP2/6- 3 1G** levels. The harmonic vibrational wavenumbers used to characterize the stationary points at the MP2/6-31G** level are tabulated in Table 3.This table also reports the har- monic vibrational wavenumbers for stationary points located on the singlet reaction surface. These results will be discussed in the next section. As discussed by Toohey and Anderson,* for the C10 + HO, reaction the UHF/6-31G** results provide a poor representation of the true MP2 saddle points. The results in Table 2 show that at the MP2 level the OH bond of the HO, species in transition state 1 is shorter than the bond length reported at the UHF level. Similarly the HO(3) bond length in the transition state is longer at the MP2 level than at the UHF level, indicating the transition state, at the MP2 level, occurs earlier in the reaction channel. The UHF transition state is found to occur later in the reaction channel, as a con- sequence of the UHF calculation underestimating the reac- tion exothermicity. These results follow the Hammond postulate2’ that relates the position of the transition state to the enthalpic change of the reaction.These results show the importance of using a correlated method to determine the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1813 Table 1 Reactants and products of the C10 + HO, reaction UH F/6- 3 1 G **O RClO RHO Roo LHOO LHOCl E/Eh 1 c10 1.620 -534.232 27 2 HO, 0.949 1.308 105.9 -150.17663 3 HOCl 1.669 0.947 105.2 -534.848 30 4 02(3%-) 1.168 -149.61791 5 O,(W 1.166 -149.557 49 MP2/6-31G** species &lo RHO Roo LHOO LHOCl E/Eh 6 CIO 1.607 (1.546)d -534.518 65 -534.668 49 7 HO, 0.975 (0.977) 1.325 (1.334) 104.4 (104.1) -150.513 23 -150.650 14 8 HOCl 1.713 (1.689) 0.970 (0.960) 102.4 (100.8) -535.18078 -535.331 43 9 0,(3q) 1.246 (1.207) -149.954 32 -150.079 13 10 O2(lA) 1.264 -149.917 50 harmonic vibrational wavenumbers'/cm -~~ 1 v1 832 6 v1 854 (vl 866)d 2 v1 1251, v, 1601, v3 4074 7 v1 1235, v2 1458, v3 3720 (vl 1098, v, 1392, v3 3436)3 v1 853, v, 1396, v3 4149 8 v1 756, V, 1298, v3 3823 (vl 739, v2 1242, v3 3626)4 v1 1996 9 v1 1411 (vl 1580) " All bond lengths in 8, and angles in degrees.MP4/6-311G** at MP2/6-3 1G**-optimized geometries. Harmonic vibrational wavenumbers calculated at the UHF/6-3 1G** and MP2/6-31G** levels. Experimental values in par en these^.^,^',^**'^*^'-^^. position of the saddle points accurately on the C10 + HO, ducts HOCl + 0,.The HO and C10 bond lengths of HOCl reaction surface. are computed to be 0.966 and 1.711 A, respectively, at the The triplet surface is found to be relatively simple with final point of the IRC calculation. The molecular oxygen only two stationary points being located at the MP2 level. bond length computed from the IRC calculation is 1.220 A. The first corresponds to the transition state discussed above, These results are in good agreement with the ground-state and was reported previously by Toohey and Anderson.* The parameters given in Table 1. In the reverse direction of the second stationary point, a loosely bound local minimum, was IRC calculation, stationary point 2 was located.The internal found as a result of performing an intrinsic reaction coordi- coordinates for this intermediate species are given in Table 2 nate calculation, using point 1 as a starting point. In the and the harmonic vibrational wavenumbers given in Table 3 forward direction the saddle point breaks down into the pro- indicate that this point corresponds to a local minimum. The Table 2 Triplet reaction surface" UHF/6-3 1G ** 1 0(1)0(2)H0(3)C1 1.245 1.110 1.265 1.627 109.7 172.6 107.6 0.0 0.0 -684.367 79 MP2/6-31G** 1 0(1)0(2)H0(3)CI 1.287 1.054 1.414 1.568 105.5 163.1 104.7 0.0 0.0 -685.029 59 -685.316 19 2 0(1)0(2)H0(3)C1 1.323 0.976 2.073 1.591 104.0 144.7 132.4 0.0 0.0 -685.041 24 -685.32435 ~~ ~~ a All bond lengths in A and angles in degrees. MP4/6-311G** results at MP2/6-31G**-optimized geometries.Table 3 Harmonic vibrational wavenumbers computed at the MP2/6-31G** level triplet surface/cm -stationary point V1 V2 v3 v4 v5 '6 v7 '8 v9 1 -1487 154 160 350 814 972 1496 1765 5853 2 46 63 88 150 343 858 1251 1479 3695 singlet surface/cm-' -65 262 395 493 714 797 962 1439 3679 -2205 104 120 28 1 515 782 1424 1531 2075 -152 106 194 367 459 84 1 1374 1759 3696 -291 155 368 517 640 833 854 1366 3791 -467 153 344 512 615 815 836 1401 3787 -281 138 274 361 459 860 1219 1334 3813 -1588 182 247 422 709 876 977 1207 2633 198 239 426 532 685 835 1300 1449 3626 171 182 274 392 683 88 1 1151 1332 3627 -148 120 257 329 520 836 1227 1323 3648 -1670 240 274 473 706 846 895 1467 1913 1814 mechanism for the C10 + HO, reaction on the triplet surface is shown in Scheme 1.$3) O(3j”-CI CI’ 2 1 Scheme 1 A gradient search was performed in an attempt to locate a saddle point that connects minimum 2 to the reactants. At the MP2 level of theory employed here no stationary point could be located on this region of the triplet surface. This could be a consequence of the intermediate species being formed with a negligible activation barrier. This abinitio study of the triplet reaction surface, supports and reproduces the results of Toohey and Anderson.* These results are reported here in order to facilitate comparison with the singlet reaction surface discussed in the next section.In both studies the C10 + HO, reaction on the triplet poten- tial surface is found to proceed through direct attack of the C10 radical on the OH bond of the HO, radical. This reac- tion leads to the formation of ground-state HOCl and 0, through a direct hydrogen-abstraction mechanism. Singlet Surface Here we report the first abinitio study of the singlet potential surface for the C10 + HO, reaction. The calculated geom- etries and energies of the stationary points located on the singlet surface are tabulated in Table 4. The harmonic vibra- tional wavenumbers computed at the MP2/6-3 1G** level to characterize the stationary points on the singlet surface are listed in Table 3. The MP2 singlet surface is more complex than the RHF/6- 31G** surface and the corresponding MP2 triplet surface.Unlike the triplet surface, where the MP2 stationary points were found to be formed earlier in the reaction channel than the corresponding UHF stationary points, the MP2 station-ary points on the singlet potential surface appear to occur later in the reaction channel than the corresponding RHF stationary points. This result and the large number of station- ary points reported in Table 4 indicate that the C10 + HO, reaction on the singlet surface occurs via a different mech- anism from that described for the triplet reaction surface. Eleven stationary points have been located and characterized on the singlet surface. However, IRC calculations indicate that stationary points 1, 4, 5 and 6 are not involved in the C10 + HO, reaction.These stationary points correspond to transition states arising from internal rotation in the reaction intermediates. Stationary point 1 corresponds to a saddle point arising from rotation about the central ClO-00H bond. The barrier to internal rotation about this bond is computed, at the MP2 level, to be ca. 20 kJ mol-l. The remaining rotational isomers, points 4, 5 and 6, correspond to saddle points that arise from rotation about the C100-OH bond. The barrier to internal rotation about this bond is similar in magnitude to the barrier about the C10-00H bond. Fig. 2 plots the relative energies of the stationary points, with respect to the reactants, against the reaction coordinate. The stationary points corresponding to rotational saddle points have been omitted for reasons of clarity.The figure also indicates the stationary points that are found to be con- 50/ J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 products ---, I I I reaction coordinate Fig. 2 Singlet potential surface (MP2/6-31G**). 0,0,see text. nected from IRC calculations. The figure shows that the ClO + HO, reaction on the singlet surface involves the formation of stable reaction intermediates. The reaction path shown in Fig. 3 also indicates that the reaction occurs via a multi-step reaction mechanism. The principal reaction channel on the singlet surface con- nects the local minima, points 8 and 9, to the saddle points 10,7 and 11.In Fig. 2, stationary point 10 does not appear to be a saddle point as it is lower in energy than the reactants to which it is connected. However, the harmonic vibrational wavenumbers given in Table 3 characterize stationary point 10 as a saddle point and the IRC calculation that connects point 10 to point 9 on the singlet reaction surface also shows that point 10 breaks down, in the reverse direction, to form the reactants C10 + HO,. This discrepancy in the energies probably arises from the potential surface not being corrected for the effects of zero-point vibrational energy and basis-set superposition error. The reaction path shown in Fig. 2 is shown schematically in Fig. 3. This reaction scheme shows that the reaction initially involves the formation of a loosely bound species, with the reactants separated by 2.7 A.The remaining stages involve a shortening of the 0(2)0(3) bond length and a migration of the H atom from the HO, species to the C10 radical. The Cl-0(1) distance in stationary point 11 is 2.4 A, whereas the corresponding distance in the triplet &l0(3) 3.3A hO(3)3.0 A 640(3) 2.5 A 10 9 7 H hO(3) .2 A RH0(3) 2.1 A 11 8 Fig. 3 Singlet surface reaction mechanism UHF/6-3 1G* * 1 C10( 1)0(2)0(3)H 1.512 2.615 1.404 0.949 112.1 107.7 100.2 105.2 159.6 -684.30604 2 ClO(1)0(2)0(3)H 1.683 1.343 1.396 0.949 110.0 105.4 100.6 86.0 172.9 -684.363 73 3 ClO( 1)0(2)0(3)H 1.659 1.409 1.356 0.949 105.7 103.2 103.8 184.8 274.2 -684.363 13 4 C10( 1)0(2)0(3)H 1.68 1 1.360 1.368 0.949 109.8 107.8 103.7 -86.1 96.4 -684.369 93 MP2/6-31G** 1 C10(1)0(2)0(3)H 1.702 1.528 1.402 0.972 103.0 101.9 101.4 159.8 86.9 -685.073 51 -685.13704 2 C10(1)0(2)0(3)H 1.671 2.856 1.378 1.087 131.2 43.6 107.2 -9.8 -30.8 -685.039 82 -685.313 64 3 C10(1)0(2)0(3)H 1.646 3.317 1.428 0.980 70.6 40.1 100.9 172.6 75.0 -685.072 96 -685.346 48 4 C10(1)0(2)0(3)H 1.757 1.382 1.487 0.973 109.2 103.7 96.9 -83.3 -169.0 -685.073 57 -685.35030 5 C10(1)0(2)0(3)H 1.768 1.393 1.478 0.974 109.2 107.6 100.3 -88.6 12.6 -685.071 63 -685.34808 6 C10( 1)0(2)0(3)H 1.508 2.807 1.489 0.971 116.8 98.7 96.6 102.4 160.8 -685.035 65 -685.339 03 7 0(1)0(2)H0(3)Cl 1.476 1.221 2.584 1.659 49.3 62.3 73.4 265.2 122.7 -685.02771 -685.291 93 8 0(1)0(2)HO(3)Cl 1.390 0.983 2.157 1.636 102.9 75.0 99.1 259.7 75.7 -685.081 48 -685.340 19 9 0(1)0(2)H0(3)Cl 1.373 1.857 3.054 1.613 31.1 48.8 61.2 268.9 139.9 -685.06624 -685.33692 10 0(1)0(2)HO(3)Cl 1.408 1.864 3.281 1.628 31.1 56.1 60.1 267.9 138.6 -685.064 71 -685.342 33 11 0(1)0(2)HO(3)Cl 1.395 1.164 1.175 1.639 104.5 124.6 105.4 -83.1 75.1 -685.059 65 -685.324 43 a All bond lengths in 8, and angles in degrees.MP4/6-311G** energies at MP2/6-31G**-optimized geometries. stationary point is 3.1 A. This shortening of the C1-0(1) dis-tance indicates that the Cl atom of C10 is involved in this reaction mechanism. The reaction scheme shown in Fig. 3 is a four-centre addition of HO, to C10, followed by hydrogen- atom migration leading to the formation of HOCl + 0,.Two additional stationary points have been located on the singlet surface. Point 3 corresponds to a loosely bound saddle point and point 2 corresponds to a tightly bound transition state. Scheme 2 shows that these stationary points arise from a two-centre addition of C10 to HO,. The Cl-0(2) distance in structures 2 and 3 is 4.1 and 3.2 A, respectively, indicating that in this case the C1 atom of C10 is not involved in the formation of a reaction transition state. From IRC calcu- lations saddle point 2 was found to be connected to the pro- ducts HOCl + 0, in the forward direction and to intermediate 8 in the reverse direction. At the level of theory employed here no stationary point could be located in the reverse direction from saddle point 3.However, point 3 was found to be connected to intermediate 8 in the forward direc- tion. It is believed that saddle points 2 and 3 are involved in an H-atom migration mechanism as discussed above. 3 2 Scheme 2 The singlet surface presented here shows that there are two possible reaction pathways leading to the formation of HOCl + 0,. JRC calculations were performed using saddle points 2 and 11 as starting points. These saddle points were found to decompose to form HOCl + 0,. At the final point of the IRC calculation the products are found to be separated by ca. 3 A and the OH and C10 bond lengths of HOCl were found to be about 0.97 and 1.71 A, respectively. The molecular oxygen species was found to have an internuclear distance of about 1.27 A.Comparison of these values with those given in Table 1 indicate that HOCl is formed in the ground state and that molecular oxygen is formed as a singlet. This conclusion is supported by the results of the IRC calculation on the triplet reaction surface, where the internuclear distance of the 0, product was found to be considerably shorter indicating the formation of the more stable, ground-state, triplet 0,. Discussion The results reported in the previous sections show that the C10 + HO, reaction can take place on the triplet and singlet reaction surfaces. The reaction energies, at the MP4/6-311G** level, for the triplet and singlet surfaces are -241 and -174 kJ mol-', respectively. The reaction energy for the singlet surface is an estimate obtained by assuming that the electron correlation recovered for singlet 0, on going from the MP2 to the MP4 levels is similar to that recovered for triplet 0,.The C10 + HO, reaction on both the triplet and singlet reaction surfaces is computed to be exothermic and energetically favourable. The energies reported in Tables 2 and 4 show that the energies of the two surfaces are very similar. The lowest energy, intermediate species are formed on the singlet surface. This indicates that on the singlet surface there is the possi- bility that the C10 + HO, reaction involves the formation of J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 relatively long-lived intermediates. The barrier heights, with respect to the reactants, for the stationary points located on each surface are given in Table 5.The barrier heights are reported at the MP2/6-31G** and MP4/6-311G** levels. The MP4 results have been corrected for zero-point vibrational and basis-set superposition error effects. The barrier heights at this level of theory tend to overestimate the true barrier height^.^'.^' Brown and Truhlar have shown that even large- scale configuration interaction calculations can overestimate barrier heights by several kcal mol-'.3' For the singlet reac- tion surface the corrected barrier heights reported in Table 5 are slightly higher, less negative, than those shown in Fig. 2. However, the structure of the reaction surface is not dramati- cally altered from that shown in Fig.2 and these results support the reaction mechanism discussed in the previous section. The barrier height between stationary points 9 and 7 on the singlet surface is 112 kJ mol-', indicating that inter- mediate species 9 may be relatively long-lived. The corrected barrier heights for the triplet surface are slightly higher than those computed at the MP2 level. The MP4 results indicate that on the triplet reaction surface the CIO + HO, reaction proceeds with a negligible or small activation energy. These results indicate that the observed negative tem-perature dependence of the ClO + HO, rate constant cannot be explained by considering only the direct reaction mech- anism that takes place on the triplet reaction surface. Stimp- fle et al.' found that the experimental rate constant could be represented by the rate expression, k, = 3.3 x exp(-850/T) + 4.5 x lo-" (T/300)-3.7 cm3 molecule-' s-'.This expression shows that the Arrhenius plot of the C10 + HO, reaction is strongly curved and that at high tem- peratures the activation energy tends to zero. This result sug- gests that at high temperatures the triplet reaction surface is dominant, leading to the formation of ground-state HOCl + 0,. At low temperatures the Arrhenius plot displays a strong negative temperature dependence, indicating the influ- ence of a multi-step reaction mechanism. Therefore at low temperatures the singlet reaction surface must play an impor- tant role in the C10 + HO, reaction.Dynamics calculations will determine whether these initial conclusions are valid for the C10 + HO, reaction. However, to perform such calcu- lations accurately the energies of the stationary points pre- sented here will have to be recalculated at a higher order of theory. The singlet reaction mechanism results in a decrease in the formation of triplet molecular oxygen and a corresponding increase in the formation of singlet molecular oxygen. The role of singlet molecular oxygen, a relatively long-lived excited state of O,, in the atmosphere has been explored by Table 5 Barrier heights (kJ mol-') corrected stationary barrier point AEa AEb BSSE ZPVE' /kJ mol-' singlet surface 2 -20.3 13.1 25.3 -9.0 29.4 3 -107.3 -73.2 14.6 2.8 -55.8 7 11.5 70.1 19.8 -6.4 83.5 8 -129.6 -56.6 20.3 5.7 -30.6 9 -89.6 -48.0 17.0 2.2 -28.8 10 -85.6 -62.3 12.5 -0.4 -50.2 11 -72.3 -15.2 37.9 -9.1 13.6 triplet surface 1 -6.8 6.3 27.3 19.3 52.9 2 -23.9 -15.0 7.9 -2.1 -9.2 a MP2/6-31G** barrier heights.MP4/6-311G** barrier heights.'Determined at the MP2/6-31G** level. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 1817 Handwerk and Zellner.,, Singlet oxygen was believed to be 10 JANAF Thermochemical Tables, ed. D. R. Stull, H. Prophet, involved in the production of ‘odd’ oxygen through the for- mation of ClO, but the recent work by Rauk et aL3, has shown that this is unlikely. The HOCl produced from the C10 + HO, reaction has an important role in stratospheric chemistry. If the hydrochlorous acid formed has a long life- 11 12 13 14 National Bureau of Standards, New York, 1979. G.Hirsch, P. J. Bruna, S. D. Peyerimhoff and R. J. Buenker, Chem. Phys. Lett., 1977, 52, 442. C. C. J. Roothaan, Rev. Mod. Phys., 1951,23,69. J. A. Pople and R. K. Nesbet, J. Chem. Phys., 1954,22,571. W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys., 1972, time it can be considered to act as a sink for ‘active’ chlorine, 56,2257. therefore decreasing the effects of stratospheric ozone destruction. However Guo~~has shown that HOCl can absorb a UV photon and dissociate to form radical pro- ducts., 15 16 17 P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213. C. Msller and M. S. Plesset, Phys. Rev., 1934, 46, 618. R. Krishnan and J.A. Pople, Int. J. Quantum Chem., 1978, 14, 91. HOCl + hv -+ OH + C1 18 R. Krishnan, M. J. Frisch and J. A. Pople, J. Chem. Phys., 1980, 72, 4244. OH and C1 are the principal products of photodissociation of HOCl whereas the quantum yield of alternative products such as 0 + HC1 has been shown to be negligible.36 The pro- ducts formed from the photodissociation of HOCl would 19 20 21 R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem. Phys., 1980, 72,650. J. H. van Lenthe, J. G. C. M. van Duijneveldt-van de Rijdt and F. B. van Duijneveldt, Adv. Chem. Phys., 1987,69, 567. S. F. Boys and F. Bernardi, Mol. Phys., 1970,19, 553. result in accelerated ozone destruction as discussed in the 22 K. Nahm, Y. Li, J. D. Evanseck, K. N. Houk and C.S. Foote, J. introduction. The importance of the ClO + HO, reaction to stratospheric chemistry is therefore dependent on the rate of photolysis of HOCl under stratospheric conditions. In this work we have reported the results of a com-prehensive ab initio study of the C10 + HO, reaction. It is 23 24 25 Am. Chem. Soc., 1993,115,4879. C. Gonzalez and H. B. Schlegel, J. Phys. Chem., 1989,90,2154. C. Gonzalez and H. B. Schlegel, J. Phys. Chem., 1990,94,5523. GAUSSIAN 92, Revision A, M. J. Frisch, G. W. Trucks, M. Head-Gordon, P. M. W. Gill, M. W. Wong, J. B. Foresman, B. G. Johnson, H. B. Schlegel, M. A. Robb, E. S. Replogle, R. Gom- found that the reaction can proceed on both the triplet and singlet reaction surfaces and that the unusual temperature dependence of the rate constant can be explained only through the consideration of the multi-step singlet reaction mechanism.It has also been shown that the products of the HO, + C10 reaction could have a detrimental effect on 26 27 perts, J. L. Andres, K. Raghavachari, J. S. Binkley, C. Gonzalez, R. L. Martin, D. J. Fox, D. J. DeFrees, J. Barker, J. J. P. Stewart and J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1992. M. F. Guest, P. Sherwood, GAMESS-UK and User’s Guide and Reference Manual, SERC Daresbury Laboratory, 1992. M. F. Guest, R. J. Harrison, J. H. van Lenthe and L. C. H. van Corler, Theor. Chim. Acta, 1987,71, 117. stratospheric ozone concentrations. 28 29 G. S. Hammond, J. Am. Chem. Soc., 1955,77, 334. W. J. Hehre, L. Radom, P.vR. Schleyer and J. A. Pople, Ab The authors gratefully acknowledge the SERC for generous grants of computer time at the Rutherford Appleton and Daresbury laboratories, for a grant to purchase a HP735 30 31 Initio Molecular Orbital Theory, Wiley, New York, 1986.A. A. Nanayakkara, G. G. Balint-Kurti and I. H. Williams, J. Phys. Chem., 1992,%, 3662. F. B. Brown and D. G. Truhlar, Chem. Phys. Lett., 1985, 117, workstation and funding for DB. 32 307. V. Handwerk and R. Zellner, Ber. Bunsenges. Phys. Chem., 1986, References 33 10,92. A. Rauk, E. T. Tschuikow-Roux, Y. Chen, M. P. McGrath and B. Reimann and F. Kaufman, J. Chem. Phys., 1978,69,2925. R. M. Stimpfle, R. A. Perry and C. J. Howard, J. Chem. Phys., 1979,71,5183. M-T. Leu, Geophys. Res. Lett., 1980, 7, 173. T. J. Leck, J-E. L. Cook and J. W. Birks, J. Chem. Phys., 1980, 72,2364. J. P. Burrows and R. A. Cox, J. Chem. Soc., Faraday Trans. 1, 1981,77,2465. F. C. Cattell and R. A. Cox, J. Chem. Soc., Faraday Trans. 2, 1986,82,1413. M. Mozurkewich, J. Phys. Chem., 1986,90,2216. D. W. Toohey and J. G. Anderson, J. Phys. Chem., 1989, 93, 1049. J. D. Cox, D. D. Wagman and V. A. Medvedev, CODATA. Key 34 35 36 37 38 39 40 41 L. Radom, J. Phys. Chem., 1993,97,7947. H. Guo, J. Phys. Chem., 1993,97,2602. M. J. Molina, T. Ishiwata and L. T. Molina, J. Phys. Chem., 1980,84,821. P. J. D. Butler and L. F. Phillips, J. Phys. Chem., 1983,87, 183. K. Hedberg and R. M. Badger, J. Chem. Phys., 1951,19,508. A. M. Mirri, F. Scappini and G. Gazzoli, J. Mol. Spectrosc., 1971, 38, 218. J. W. C. Johns, A. R. W. McKeller and M. Riggin, J. Chem. Phys., 1978,68,3957. K. Nagai, Y.Endo and E. Hirota, J. Mol. Spectrosc., 1981, 89, 520. C. Yamada, Y. Endo and E. Hirota, J. Chem. Phys., 1983, 78, 4379. Values for Thermodynamics, Hemisphere Publishing, New York, 1989. Paper 3/07015H; Received 25th November, 1993