GENERAL DISCUSSION Dr. M. S. Child (Oxford) said: It is interesting that Prof. Setser should report significant electronic energy transfer R* + X2 + R, + XI for the systems Kr/Cl,, Kr/Br, and Xe/I, because, if this process is a byproduct of the ionic/covalent harpoon mechanism, it implies that a significant fraction of the en- counters on the ionic surface fail to negotiate the corner which leads to the product channel. In the language of the paper of Child and Whaley, this indicates the exist- ence of a second trapped trajectory on the ionic surface even at the thermal energies of Prof. Setser's experiments. Dr. M. A. D. Fluendy and Mr. D. Sutton (Edinburgh) said: The results presented by Setser et al. in which they observed substantial production of electronically excited bromine molecules in collisions with rare gas metastables provide yet another contrast to the behaviour of the analogous alkali metal/halogen molecule systems.Prelimin- ary measurements of the excitation produced in K/Br, collisions using the time of flight crossed beam technique described earlier in this Discussion' are shown in fig. 1. 0 10 A E/eV FIG. 1.-K/Br2 Energy loss profile at a collision energy of 164 eV (centre of mass). The results are the average of 27 separate observations at angles between 0.45 and 1.78". The main peak at AE 21 0 is broadened, presumably by vibrational excitation of the Br,, but possible electronic transitions with AE > 1.4 eV account for (2% of the observed scattering. The second main feature is due to the K41 isotope which occurs with 6% abundance and provides a useful magnitude comparison.In these M. A. D. Fluendy, K. P. Lawley, J. McCall, C. Sholeen and D. Sutton, Faraday Disc. Chem. Soc., 1979, 67, 41.344 GENERAL DISCUSSION alkali metal/halogen systems, interactions between the ion pair potential surface and the surface leading asymptotically to the excited halogen molecule occur at extremely large distances and cannot therefore be an important route to these excited states. In contrast, as suggested, in the rare gas metastable analogue systems the same V(RG+, Br;)/ V(RG, Br;) interaction lies inside the initial electron transfer radius [ V(RG*, Br,)/V(RG+, Bri)]. Dr. W. M. Jackson (Washington, D.C.) said: We have combined the laser induced fluorescence technique with vacuum ultraviolet flash photolysis to study the dynamics of the C + NO reaction.This reaction is summarized in the following equation for ground state atoms : C[~P~(~P,)] + NO -+ CN(X2X, u", N") + O(3P); AH = 1.1 eV. The carbon atoms are produced by using a high pressure argon flashlamp, which It has been shown that the V.U.V. has a CaF, window to photodissociate the C302. photolysis ' of this molecule leads to the following photochemical reactions. -+ C(29,) + 2co --+ C(2'SO) + 2 c o . The relative output of the argon flash lamp is such that it decreases with decreasing wavelength. When the relative output of the flash lamp is folded into the absorption coefficient of C302 one predicts that most of the molecules are dissociated at 165 & 10 nm. Husain and Kirsch2 have shown that most of the carbon atoms produced in this wavelength region are in the (Z3PJ) electronic state.The rate constant for the reaction of carbon atoms in each of these is the same order of magnitude so that any reaction that is observed probably comes from the reaction of 3P atoms with NO. The rate constant2 for the reaction of 3P atoms with NO is 5 x lo-'' cm3 molecule-' s-l. At 0.150 Torr of NO this corresponds to a mean time between collisions of 5 ps. The mean time for non-reactive collisions at this pressure is 1 p s . In 2 ps, few of the CN radicals that are formed will have had time to undergo a relaxing colli- sion. The conclusion can therefore be drawn that any distribution observed under these conditions represents the nascent distribution of the radical.The spectra that are obtained under the above conditions show first that CN radi- cals are produced from the reaction of C atoms with NO, in agreement with the corre- lation arguments of Husain2 and that some of the exotherniicity of the reaction is used in producing vibrationally and rotationally excited radicals. Preliminary analysis of the data indicates that CN radicals are produced with a rotational tem- perature of z 1000 K. No analysis of the vibrational distribution has been made since only two vibrational levels are observed. A surprisal analysis of the observed distribution is calculated from the observed temperature which is then compared with a " prior " distribution. This prior distribution was calculated using the techniques described by Kinsey and LevineO3 The analysis gives a linear surprisal and it shows that the rotational distribution ob- served in these experiments is cooler than the predictions based upon the prior distri- bution.W. Braun, A. M. Bass, D. D. Davis and J. D. Simmons, Proc. Roy. SOC. A, 1969,312,417. D. Husain and L. J. Kirsch, Truns. Faruduy SOC., 1971,67, 3166. R. D. Levine and J. L. Kinsey, Atomic and Molecule Collision Theory: A Guide for the Ex- perimentalist, ed. R. B. Bernstein (Plenum Press, New York, 1978).GENERAL DISCUSSION 345 The prior distributions used in the present surprisal plots were computed using only the conservation of energy as a constraint. The linearity of the surprisal plot suggest that for the present results this constraint is adequate.It is therefore un- likely that the rotational states that are accessed are significantly constrained by con- servation of angular momentum. In conclusion, the present preliminary results show that the dynamics of A + BC reactions may be studied under bulb conditions using short duration flash lamps and LIF. Dr. G. Hancock (Oxford) said: The direct measurement of the quenching of C('S,) with 0, by Husain and Norris results in a rate constant for the process, 9.9 x lo-', cm3 molecule-' s-l, which is considerably larger than the previously reported estimate of 5 x Recently we have used the latter value in an interpretation of the kinetics of vacuum ultraviolet emission observed in the quenching of C,(a311n,) by 02,2 and this needs to be re-examined in the light of the new measurement.C, in the low lying a311, state was produced by the infrared multiple photon dis- sociation of 0.2 mTorr C2H3 CN diluted in 5 Torr Ar, and fig. 2 shows the rate of cm3 molecule-' s-l.l 2 c I * 0 0 50 100 150 200 02 pressure /mTorr FIG. 2.-First order decay rates of C,(a3Xu) in mixtures of 0.2 mTorr CtHJCN with varying amounts of added O2 at a total pressure of 5 Torr with added Ar. 0, laser excited fluorescence signal; 0 , vacuum U.V. emission. decay of the radical, measured by dye laser excited fluorescence, as a function of the pressure of added 0,. The slope of this plot gives a rate constant for the removal of C2(a31-In,) by O2 as 3.4 x lo-', cm3 molecule-' s-l. In view of the substantial exo- thermicity of several reaction schemes which can be written for C2 + 02, a search was made for vacuum U.V.emission accompanying the quenching by 02. Emission was observed in a wavelength region strongly suggesting that it originates from CO(A'I1). Furthermore, the decay of the vacuum U.V. emission followed first order kinetics with the same dependence upon O2 pressure as the removal rate of C,(a3rIu), again illu- strated in fig. 2. Clearly CO(A'll) is either being formed directly from the reaction C2(a3Hn,) + 02(x3Cg) -+ CO(ATI) + CO(x'C+) or by a secondary process involving the products of the C2 + 0, reaction, and taking place at a considerably faster rate. The reaction scheme C2(a311n,) + 02(x3z;) --f Co2(x1cS+) + C('S,) C(lS,) + 0 2 ( X 3 C g ) -+ c o ( ~ ' r I ) + 0(3~), (2) (3) G. M. Meaburn and D.Perner, Nature, 1966, 212, 1042. S. V. Filseth, G. Hancock, J. Fournier and K. Meier, Chern. Phys. Letters, 1979, 61, 288.346 GENERAL DISCUSSION which is exothermic for the production of CO(A'II), was rejected on the grounds that the previously reported value of k, was far too slow to account for the observed vacuum U.V. decay kinetics. Although the new value of k3 is faster than the rate constant for quenching of C,(a31-In,) by O,, reactions (2) and (3) would not imply a single expo- nential decay of the emission from CO(A'II), but would indicate that the signal should initially rise to a maximum value following the infrared laser pulse, and then fall, with a rate at long times equal to that of reaction (2). In all cases only a single ex- ponential decline in signal was experimentally observed ; the time resolution was sufficient to be able to detect any rise in signal following the infrared laser pulse com- patible with this kinetic scheme.On this evidence, the two-step process (2) and (3) can be rejected, and reaction (1) still appears to be the prime candidate for the produc- tion of vacuum U.V. emission in the C2 + 0, system. Dr. R. J. Donovan and Mr. M. C. Addison (Edinburgh) said: We had originally hoped to include work on the reactions of S(31D2) in our paper, but the results were not available in time for the manuscript deadline. Since then we have succeeded in making a number of direct studies of S(3'D2) using time-resolved atomic absorption photometry (1 = 167 nm) in the vacuum ultraviolet and have measured the absolute rate for its reaction with OCS as k = (1.2 & 0.3) x 10-l' cm3 molecule-' s-'.This is a particularly interesting reaction as the product S, is formed in its first electronically excited state, uiz: S(3'D2) + OCS --f S,(a'A,) + CO. We are currently making a detailed study of the state-to-state kinetics in this system and some other systems including reaction of S(3'D2) with O,, which produces SO(b'Z), and with N20, which produces SO(alA). Dr. J. E. Butler (Washington, D.C.) said: We have recently measured the product OH rotational distributions for u = 0, 1 formed in the reactions of electronically excited oxygen atoms, O('D), with H,, HCl, H2016 and H20I8. We have observed that the available energy in these exothermic reactions is not always statistically distributed.In the one case where a statistical distribution agrees with our observa- tions, O('D) + H2 --f OH(u = 0, N I 26; u = 1, N 5 IS), we believe the mechanism to be a distribution of " insertion " and " abstraction " events. Our observations are: (i) For H,, all observable rotational levels of OH(u = 0, N I 26; v = 1, N 15) can be fitted with an infinite temperature Boltzmann like distribution, with relative integrated rotational distribution of N(v = l)/N(v = 0) = 1.1 5 0.40. (ii) For HCl, the OH rotational distributions for u = 0, 1 can be charac- terized as the sum of two Boltzmann like distributions, one with T x 400 K for 0 I N 6, and with T x 4000 K for 7 5 N I 23 or 16 for v = 0 and 1 , respectively. The relative integrated rotational distributions gave N(v = l)/N(u = 0) = 0.3 & 0.15.(iii) For H2016, the OH rotational distributions observed for u = 0, 1 could be approxi- mated by the sum of two Boltzmann like distributions with characteristic tempera- tures of x400 K for 0 < N < 6, x2300 K for 7 I N I 18 or 15, respectively. The relative integrated rotational distributions gave N(v = l)/N(u = 0) = 0.35 & 0.10. (iv) For H2018 + 016('D), the 016H and 018H distributions were measured and could be approximated as above with characteristic temperatures of x 400 and x 3300 K for 016H and x 400 and x 1800 K for 0l8H. The integrated rotational distributions gave N(v = l)/N(u = 0) = 0.44 & 0.15 for 016H and 10.08 for 0"H. Trajectory calculations on the O('D) + H2 system by Sorbie and Murrell' and K.S. Sorbie and J. N. Murrell, Mol. Phys., 1976, 31, 905.GENERAL DISCUSSION 347 Whitlock et aZ.l indicate that one can factorise reactive encounters into two groups which can loosely be described as " insertion " into the H2 bond and " abstraction " of a H-atom by the attacking O(lD) atom. These two types of reactive events were predicted to give noticeably different rotational and vibrational product OH distribu- tions. " Insertion " events gave hot, even inverted rotational distributions and mono- tonically decreasing vibrational distributions (with vibrational energy), while " ab- straction " events gave rotationally cold distributions and vibrationally inverted or hot distributions. These trajectory results also agree with our observed OH distri- butions from O('D) + H2.Since our observed OH(u, J ) distribution for O(ID) + HC1, H20I6, H201' do not agree with the predictions of statistical theories, and since they all gave rotationally cooler distributions than that observed with H2 where " insertion " should be the easiest, we infer that the mechanism in these reactions is probably not the formation of a long lived complex in which energy is scrambled amongst available modes, but rather a distribution of" insertion " and " abstraction " events with the former more important in H2 and the latter in the HCl, H2016*18 systems. Dr. P. A. Gorry, Dr. C. V. Nowikow and Prof. R. Grice (Manchester) said: We wish to comment on the flash photolysis measurements of the O(3P) + CF31 reaction by Addison, Donovan and Garraway and their relation to a recent molecular beam study' of this reaction.The molecular beam data may be explained in terms of hard sphere scattering at small impact parameters combined with significant energy transfer to the CF, radical but do not indicate the existence of a long-lived collision complex. Thus the potential energy surface for the 0 + CF31 reaction may involve a shallow well or even be essentially level. The CF310 molecule3 is stable with respect to dispropor- tionation for temperatures ,<O "C and must correspond to a deep well E, 2 100 kJ mol-' on the potential energy surface. Such a well corresponding to the (presum- ably) singlet CF,IO molecule would give rise to a long-lived collision complex if it participated in the reaction dynamics.Accordingly we feel that the 0 + CF31 reaction dynamics may be explained by motion over a triplet potential energy surface and that it is not necessary to invoke transitions to the singlet surface. This would be in line with the situation which is believed to obtain4 for the reactions of 0 atoms with halogen molecules. The small total reaction cross section Q z 2 A' for the 0 + CF31 reaction indicated by the flash photolysis measurements would be attri- buted to reaction occurring only at small impact parameters together with an orienta- tion requirement for reaction. These conclusions concerning the O(3P) + CF31 reaction are not in conflict with the observation of some O(lD) quenching to O(3P) in the O('D) 4- CF3C1 reaction by Addison, Donovan and Garraway.Reaction of O(lD) with CF31 would yield a long-lived singlet CF310 collision complex persisting for 1 1 50 vibrational periods. Consequently the seam of intersection between the singlet and triplet CF310 surfaces may be traversed 2300 times in the O(lD) + CF31 reaction compared with a single time in an O(3P) + CF31 reaction proceeding via a direct mechanism. Accordingly, P. A. Whitlock, J. T. Muckerman and E. R. Fisher, Research Institute for Engineering Sciences Report (Wayne State University, 1976). P. A. Gorry, C. V. Nowikow and R. Grice, Chem. Phys. Letters, 1978, 55, 24; Mol. Phys., 1979, 38, in press. D. Naumann, L. Deneken and E. Renk, J. Fluorine Chem., 1975,5,509. D. D. Parrish and D. R. Herschbach, J. Amer. Chem. Soc., 1973,95,6133; D. St.A. G. Radlein, J. C. Whitehead and R. Grice, Mol. Phys., 1975, 29, 1813.348 GENERAL DISCUSSION the probability of a singlet ++ triplet transition may be higher in the O('D) + CF31 reaction than in the O(3P) + CF31 reaction by a factor 2300, provided that the transition probability is small for a single traversal of the seam. Dr. J . P . Simons (Birmingham) said: One of the simplest experimental methods of studying molecular photodissociation is to monitor the fluorescence of electronically excited fragments following predissociation from prepared states populated by monochromatic light absorption in the vacuum U.V. The initial wavefunction in the generalised Franck-Condon factors discussed by Freed et. al.' is then that of the photoexcited state rather than the ground state.For example, this technique has been used to determine the distribution over rotational (and vibrational) states in CN(B2C +) following the predissociation of HCN( CIA') from vibronic levels carrying 3 and 6 quanta in the bending mode.2 In this particular example, the distribution followed the simple mapping .Iparent --f jfragment, dictated by the requirement of angular momentum conservation, following the loss of the light H atom. In the case of CS2 however, the rotational energy distribution in CS(A'n) following predissociation from a linear Rydberg state at 130.4 nm,3 qualitatively conforms to the predictions of the full Franck-Condon mode1.l The distribution peaks at j M 16, well below the level j M 30, determined solely by angular momentum conservation (cf.the " bending " term in the generalised Franck-Condon factor)' but close to the level determined solely by the energy conservation requirement (cf. the " stretching " term in the generalised Franck-Condon factor) which, for a Boltzmann distribution in the parent molecule, would produce a maximum at j z 11. In contrast, the rotational distributions in CN(B2C +) produced following excitation of the cyanogen halides in the cc-continuum4 cannot be explained solely in terms of a Franck-Condon model. In these molecules, the topography of the potential energy surface over which the neglected " final state interactions " occur, is thought to be an important factor in determining the final energy d i ~ p o s a l . ~ The importance of adequately characterising the nature of the photoexcited state initially prepared by photon absorption cannot be overemphasised, particular in a Faraday Discussion devoted to the Kinetics of State Selected Species.A useful technique which helps in the spectroscopic assignment of the prepared state has been devised recently5 and termed polarised photofluorescence excitation spectroscopy. Examination of the degree, and particularly the sign, of the polarisation in the fluores- cence of electronically excited fragments, following direct or predissociation, can give valuable information related to the symmetry and the lifetime of the photoexcited parent m ~ l e c u l e . ~ * ~ Fig. 3 shows the excitation and polarisation spectra of the CN- (B --f X ) fluorescence produced through predissociation of BrCN in the B and C band systems.The high positive polarisation in the underlying continuum region indi- cates an upper state of either O+(linear) or A'(bent) symmetry, while the negative polarisation in the banded regions reflects the perpendicular orientation of the transi- tion m ~ r n e n t . ~ ~ ~ The B and C systems are assigned to the 311 and 'll components of the Rydberg transitions 27t -+ 3sa.' It is particularly striking that there is a dramatic K. F. Freed, M. Morse and Y. B. Band, Faraday Disc. Chem. Soc., 1979, 67, 297. M. N. R. Ashfold, A. M. Quinton and J. P. Simons, unpublished work. M. N. R. Ashfold and J. P. Simons, J.C.S. Faraday 11, 1978, 74,280. G. A. Chamberlain and J. P. Simons, J.C.S. Faraday ZZ, 1975,32,355.; M. T. Macpherson and J.P. Simons, Chem. Phys. Letters, 1977, 51, 261. M. T. Macpherson, J. P. Simons and R. N. Zare, Mol. Phys., to be published. ' M. N. R. Ashfold, M. T. Macpherson and J. P. Simons, Chem. Phys. Letters, 1978, 55, 84. ' A. S. Georgiou, M. T. Macpherson and J. P. Simons, unpublished work.GENERAL DISCUSSION 349 I f 0.06 0.04 0.02 0 -0.02 I . 142 144 146 148 150 152 A/nm FIG. 3.-Photofragment fluorescence excitation spectrum of CN(B) from BrCN and detailed polar- ization measurements in the region of the 'TI, 311 (2n +ma) band systems. Dashed curve shows variation in intensity of continuum photolysis source. change in the internal energy disposal in CN(B), on changing the exciting wavelength from 149.4 nm, which lies principally in the continuum region, to 147 nm,' which excites principally the 1JI,3s0 t X band origin.Prof. R. N. Dixon (Bristol) said: I would like to comment on the model for pre- dissociation discussed by Freed, Morse and Band and to suggest a need for a more detailed model. Mr. Noble and I have been studying the predissociation of HNO in its first 'Aff excited state using the technique of laser excited fluorescence at low pressure ( z 1 mTorr). Evidence for predissociation has also been obtained previously through breaking off in the structure of HNO chemiluminescence from the reaction H + NO + M --f HNO* + M,' and through line broadening3 This spectrum is also subject to extensive minor perturbations of the rotational structure. Table 1 shows the highest bound rotational levels of various vibronic states.TABLE HIGHEST BOUND ROTATIONAL LEVELS IN VIBRONIC STATES OF HNO JIA" vibronic highest bound level term vibronic states K' J' value/cm-' origin/cm- 000 13 ? 2 16 440* 13 154 010 10 ? 2 1 6 500* 14 575 020 4 12 16 469 15 956 100 3 14 16 452 16 009 100 4 11 16 485 16 009 101 0 0 16 971 16 971 030 0 0 17 310 17 310 * Ref. (1). M. N. R. Ashfold and J. P. Simons, J.C.S. Faraday ZZ, 1978, 74,280. M. J. Y. Clement and D. A. Ramsay, Canad. J . Phys., 1961,39,205. P. A. Freedman, Chern. Phys. Letters, 1976, 44, 605.350 GENERAL DISCUSSION Theoretical potential energy curves for HNO show a long range attraction in the correlation of the H + NO limit with the ground 2 'A' state of HNO, but a repulsive barrier between this limit and the A ' A " excited state.' The theory of Freed, Morse and Band would be directly applicable to the predissociation of the A state through the crossing of this barrier.However, there are a number of pieces of evidence to suggest that the predissociation of the levels in the table proceeds through a two-step mechanism involving high levels of the ground state: (i) The breaking-off limit is remarkably constant at about 16 488 cm-' even though for the 0,, vibronic level the K, quantum number is 13 and the total rotational energy is ~33300 cm-', whereas for the 100 level and K, = 3 the rotational energy content is only ~ 4 4 0 cm-'. The centrifugal barrier in the effective potential for the exit channels must therefore be small. This would be the case if the predissociation route involved the attractive ground state potential, but not for passage over the excited state barrier, which occurs at a conformation with a substantial a-axis rota- tional constant.(ii) We have found (table 1) that the O,, rotational levels of the 101 and 030 vibronic states are sharp, even though they lie well above the predissociation limit of the lower levels, whereas levels with higher J are predissociated. Furthermore, Freed has found that these levels have a width which is J-dependent.2 Thus the predissociation must be rotationally induced. (iii) By exciting laser induced fluorescence in magnetic fields up to 10 kG we have found that only the perturbed lines are highly sensitive to the field. From the analysis of these effects we conclude that the transition densities between the main and per- turbing levels have magnetic moments of the order of tenths of a Bohr magneton.The excited state alone should be diamagnetic, but the k and A" states together corre- late with a lA state of linear HNO. From intensity anomalies in perturbed levels close to the predissociation limit we conclude that the perturbations and the pre- dissociation are two manifestations of the same " internal conversion " mechanism. During the dissociation of HNO* the atoms will therefore move in the force field of the ground state, and the final distribution over quantum states of the products will not be simply related to the theory of Freed, Morse and Band. Presumably their conclusions concerning the conversion of reactant bending vibrational momentum into product rotational angular momentum should still hold, but the vibrational distribution will be greatly affected by the extra step.We hope to be able to measure this distribution in the NO formed. Internal conversion is known to be an important process in many polyatomic molecules and can provide a general mechanism for circumventing barriers to the dissociation of excited states. We may therefore expect our conclusions concerning the mechanisms of predissociation of HNO to hold for many other molecules. We hope that Prof. Freed will extend his model to include this important process. Prof. K. F. Freed (Chicago) said: We are glad to see the excellent experimental data on photodissociation processes which Dr. Simons and Prof. Dixon have pre- sented. We hope that experiments of this type will help to test and broaden the theories as well as to provide us with detailed information concerning the structure of repulsive potential energy surfaces.The theory described in our paper and previous 0nes~9~ is directly applicable to A. W. Salotto and L. Burnell, Chem. Phys. Letters, 1969, 3, 80. P. A. Freedman, Chem. Phys. Letters, 1976, 44, 605. Y . B. Band and K. F. Freed, J . Chem. Phys., 1975, 63, 3382. K. F. Freed and Y . B. Band, Excited States, 1978, 3, 109.GENERAL DISCUSSION 351 predissociation when the individual rotational states are broadened, but not over- lapping such that states with identical good quantum numbers (like total angular momentum) are nonoverlapping. In this case the initial state, li), in eqn (1) is the predissociating level, and all else remains the same.Dr. Simons' beautiful polariza- tion experiments indicate the more complicated situation in BrCN photodissociation to CN(B2C+) in the B and C band systems where there are contributions from direct and predissociation occurring simultaneously. (Perhaps even the predissociating levels are also overlapping.) This situation corresponds to the general problem discussed by Fanol which can also be applied to photodissociation as mentioned previously by US.^^^ However, this procedure involves the solution of a configura- tion interaction matrix involving mixing of the predissociating levels with the con- tinuum of directly dissociating states. This, of course, represents a very difficult calculation, and it would be preferable to have experiments which involve weak pre- dissociation through readily assignable quantum states, li).The three-dimensional theory in our paper discusses how vibrational and rotational relaxation on the repulsive surface contribute to the state-to-state photodissociation probabilities. Thus, the wavefunction (4) of our paper has a summation over all possible fragment quantum states to describe this half-collision scattering event. The calculations have not yet included this feature. Evidence from collinear calcu- l a t i o n ~ ~ ~ ~ shows how this half-collisional process can yield fragment vibrational excitation, and similar effects are to be anticipated in the full three-dimensional case. In face, ab initio calculations on excited surfaces4-' for HCN, H20, H02 and HCO all display conical intersections between surfaces and this is fairly general when the fragments are radicals with low lying excited states.In the H20 case' the surface leads to large torques on the departing OH fragment, producing high rotational excitation' in the half collision. More effort is required on this important problem. In fact, close coupled half collision scattering calculations would require the use of our Franck-Condon theory if optimal (and different) coordinate systems are utilized for the bound and dissociative surfaces. We should also emphasize that the theory predicts rotational distributions, as in our eqn (9), which can be fit approximately to Boltzmann distributions without any model of rotational energy randomization. These distributions (and ones for higher initial J ) have the exponential dependence on ( j + +)2[and ( J - j)2] due to the gaussian nature of the bending wave-function, having no thermal origins: p in our eqn (8) and (9) is dynamically determined.Dr. W. M. Jackson (Washington, D.C.) said: We have measured the rotational distribution of the CN(X2C) state radical produced by the predissociation of three different vibrational levels of (C'll,) state of C2N2. This information represents the most detailed experimental data currently available on photodissociation. It is therefore a challenge to theory. The data show how a molecule in a given electronic vibrational state predissociates into fragments with a fixed amount of translational energy. A surprisal analysis of the data indicates that excited molecules with 8400 and 10 500 cm-' of excess available U.Fano, Phys. Rev., 1961, 124, 1866. Y. B. Band and K. F. Freed, J. Chem. Phys., 1975,63, 3382. K. F. Freed and Y . B. Band, Excited Stares, 1978, 3, 109. G. J. Vazquez and J. F. Gouyet, Chem. Phys. Letters, 1978,57, 385 and in press. F. Flouquet and J. A. Horsley, J. Chem. Phys., 1974, 60, 3767. S. R. Langhoff and R. L. Jaffe, J . Chetn. Phys., in press. S. Iwata, Chem. Phys. Letters, in press. * T. Carrington, J. Chem. Phys., 1964, 41, 2012.352 GENERAL DISCUSSION energy have the same distribution with respect to the fraction of the available energy that ends up in rotation. At 6300 cm-' there is a difference in the observed distribu- tion. Currently, the experi- mentalists have been unable to measure the ro-vibronic distribution of the CN(A2n) state fragments that are produced and undetected.Theory should be able to do this. I hope that theorists will accept these experimental challenges. Dr. M. S. Child (Oxford University) said: In fig. 4 of his paper Prof. Leone attributes the inaccessibility of states of IBr which correlate with electronically excited Theory should be able to explain these differences. 18 500 18 000 17 500 17 000 c 5 2 16 500 16 000 500 0 FIG. 4.-Potential curves for the X'X+, B3110+ and P O + of IBr. The arrows show how a two photon experiment might break the Franck-Condon selection rule against photodissociation to produce I* atoms from the ground state. iodine atoms to negligibly small Franck-Condon overlap with the zero point wave- function of the IBr ground state.Some years ago I made a detailed analysis of the predissociation from the IBr B'(O+) state,' which suggests a possible two-photon experiment to detect these Franck-Condon forbidden states. Two features of the analysis' are important. The first is that the absorption* and magnetic rotation3 spectra show occasional sharp levels, so sharp in fact that Weinstock and P r e s t ~ n ~ . ~ have been able to observe laser induced fluorescence. The second feature is that the quantitative analysis ' required the adoption of a coupling scheme intermediate between the diabatic and adiabatic limits. This implies substantial wave function amplitude at all three of the turning points shown in fig. 4. M. S.Child, Mol. Phys., 1976, 32, 1495. L. E. Selin, Arkiv. Fys., 1962, 21, 479. W. H. Eberhardt, Wu-Chich Cheng and H. J. Renner, J . Mol. Spectr., 1959,3, 664. E. M. Weinstock, J . Mol. Spectr., 1976, 61, 395. E. M. Weinstock and A. Preston, J . Mol. Spectr., 1978,70, 188.GENERAL DISCUSSION 353 The suggested experiment involves one photon tuned to B' f- X (20- 1) absorption band at 2 17 500 cm-l, used by Weinstock and Preston1i2 to observe the fluores- cence and a second photon to probe the excited state. Dr. C. Fotakis, Dr. M. Martin and Dr. R. J. Donovan (Edinburgh) said: We would like to comment on the work of Baughcum et al., concerning the photolysis of alkyl iodides with a rare gas halide excimer laser. Using the unfocused output of a KrF laser (2 = 248 nm, E = 38 mJ) we have observed chemiluminescence from the reaction, 1(52P+) + I(52P;) + M + 12(B3no+J + M following the photolysis of CF31 and CH31.This luminescence was first observed by Abrahamson et al.3 using a high energy (broad band) flash photolysis system, and more recently by Stephan et al.4 using a quadrupled Nd-YAG laser. In our experiments the 12(B-X) emission, following photolysis of CFJ(0.6 kN mV2) was observed to increase with time over the first few milliseconds after the laser pulse and then decline. This is readily understood in terms of the mechanism: CF31 + hv(A = 248 nm) -+ CF, +- I(5,P,) -+ CF, + I(5,P;) 1(52P+) + Q -+ I(5,P;) + Q I(52P+) + I(52P;) + M -+ 12(B31To+u) + M. The I,@-X) emission is observed immediately after the laser pulse as some ground state atoms are produced in the primary step; however, the maximum in the chemi- luminescence is not observed until 2 2 ms (depending on conditions), when the con- centrations of excited and ground state atoms are equal, as the relaxation of I(52P,) by CF31 is very inefficient.If a sample of CF31 is subjected to a second laser pulse the peak in the chemiluminescence is observed at shorter times, due to the more rapid quenching of 1(5,P,) by I,. For CH31 the 12(B-X) emission is observed to peak shortly after the laser pulse due to the efficient relaxation of I(52P,) by CH31. We would therefore emphasise that the observation of emission from molecular fragments, from molecules such as CH21Z, does not necessarily arise from a primary photochemical step.Indeed, we would expect chemiluminescence from secondary reactions, such as those outlined above, to be a general feature of excimer laser photochemical studies due to the high radical concentrations that can be produced. The excimer laser should therefore be a powerful tool for the study of new chemiluminescent reactions. Dr. A. Ding (Berlin) said: It has been mentioned that certain types of experiments, particularly scattering and spectroscopic polarization experiments, give insight into vector properties of the collision dynamics, which allow predictions about the direction of the angular momenta of the collision encounter. It may be noted that emission spectra of polyatomic, especially triatomic, species contain similar information. The correlation between the two rotational quantum numbers J and K are such a vector property.They describe the mode of rotation of the product and can be used E. M. Weinstock, J . Mot. Spectr., 1976, 61, 395. E. M. Weinstock and A. Preston, J . Mol. Spectr., 1978, 70, 188. E. W. Abrahamson, D. Husain and J. R. Wiesenfeld, Trans. Faraday SOC., 1968, 64, 833. K. H. Stephan and F. J. Comes, in Laser Induced Processes in Molecules, ed. K. L. Kompa and S. D. Smith (Springer-Verlag, Berlin, 1979), p. 301.354 GENERAL DISCUSSION to give information on the collision dynamics. In the case of triatomic molecules, in particular, one is able to distinguish between in-plane and out-of-plane collision geometries. An example for such a system is the ion-molecule reaction Hg + H 2 + H 3 + H for which experiments have been performed by measuring the infrared chemilumines- cence of the H i product.The reaction took place in a large vessel filled with low pressure H2 (w 5 x lo-' Torr), where Hg-reagents were produced by electron impact, and subsequently reacted with the unionized H2-gas. Infrared emission was measured in the 2.5-5.5 pm region with the use of a high throughput double monochromator' specially built for this experiment. So far the spectroscopy of the H i system is not yet completely known, as this is the first time such ion spectra have been recorded. However, with the help of ab initio calculations2 one can reach at least qualitative conclusions, which show a non-statistical behaviour of the appropriate distributions. Particularly one can conclude that the 3rd vibrational level of the asymmetric stretch mode is considerably populated, and the rotational distribution of the H'; shows maxima for levels with J z K , indicating that the rotation is mainly about the axis perpendicular to the molecular plane.This is in agreement with trajectory studies on the same s y ~ t e m , ~ and would lead to the conclusion that the reaction predominantly proceeds via an in-plane collision encounter. Dr. M. Martin, Mr. M. Trainer and Dr. R. J. Donovan (Edinburgh) (communicated) : We would like to mention some recent results which support the findings of Baughcum et al. concerning the low yield of HBr from the reaction of Br(2P+) with H2S. In a wide range of studies involving I(52P+) interacting with hydrides (e.g., CD3CN, C6HsCH3, CH,CHO, C6HSCHO) for which exothermic chemical reaction to produce HI is thermodynamically favoured, we find that the dominant removal process is physical quenching.The fact that the electronic excitation energy in these systems cannot be used efficiently for reaction is not in fact surprising as a non-adiabatic transition from the excited entrance channel hypersurface to the ground state exit channel surface is required before the hydrogen halide product can be formed. Thus adiabatic correla- tion rules provide a good guide to the branching into product channels for these, and many other We would emphasise that where adiabatic reaction on an excited hypersurface is possible, reaction proceeds efficiently. Examples of this behaviour are given by Baughcum et al.for reaction of Br(42P+) and I(52P+) with halogens and interhalogens. Further examples are provided by the reactions of F(22P+) and C1(32P+) with hydrogen halides where the thermal population of the 'P+ state, at 300 K, is sufficient to give rise to substantial yields of the excited halogen atom product via the adiabatic ~ h a n n e l , ~ X("+) + HY -+ HX + Y(2P+). Dr. J. Wanner (Munich) said: In this Discussion Clyne and McDermid reported on an improved method for the determination of bond dissociation energies of the A. Ding and A. Redpath, Proc. I.C.P.E.A.C., 1977,10, 759. G. D. Carney and R. N. Porter, J. Chem. Phys., 1976,65,3547. R. J. Donovan and D. Husain, Chem. Rev., 1970,70,489. C. Fotakis and R. J. Donovan, J.C.S. Faraday 11, 1979, 75, 1553.' J. Muckerman, personal communication.GENERAL DISCUSSION 355 ground state of interhalogen molecules from the observation of the onset of predisso- ciation in the B state under collision free conditions. This method may be compared with independent experiments of laser-induced fluorescence product state analysis. So far we are only able to comment on the bond dissociation energy D: (IF). We have performed a crossed molecular beam study of the reactions F + CH31 --f I F + CH, and F + CF,I -+ I F + CF, using thermal reagent beams at 300 K.' The total available energy for internal I F product excitation E,,, = 45.3 and 54.9 kJ mol- ', respectively, can be calculated using reagent bond dissociation energies Di (CH,-I) = 229.4 kJ mo1-' and Dg (CF,-I) = 219.9 kJ mol-' given by Okafo and Whittle2 and the recently improved value Dg (IF) = 268.0 kJ mol-' by Clyne and M~Dermid.~ Vibrational product excitation in the electronic ground state of I F should thus be possible up to u = 6 and u = 7 for the reactions with CHJ and CF31 as reagents, respectively.This has been found in consistency with our experimental observations. It should be added that the earlier value for 08 (IF) = 272.4 kJ mol-' determined by Clyne and McDermid at a higher pressure4was partially in disagreement with our results since population up to u = 8 should have been observed in the reaction with F + CF31. Dr. M. A. A. Clyne (Queen Mary College, London) said: The elegant crossed- beam experiments of Wanner5 provide confirmation of the value of (22 333 + 2) cm-' for Do0 (IF) reported by Clyne and McDermid., In our work,, the dissociation energy of ground-state I F X'C+ was determined from the energy of onset of predisso- ciation in the rotational structure of I F B3n(O+) - X'C+, observed in laser-induced fluorescence (LIF).Collision-free conditions were used,3 which resulted in observa- tion of a sharp fall in excited-state lifetime between J' = 6 and J' = 7 of the level u' = 9. we observed LIF of I F at higher pressures ( z 1 Torr as compared with \< 1 mTorr in our later, definitive work3). Collisions modified the initial rotation- al ~ t a t e , ~ so that onset of predissociation was not sharp. Rotational (or vibrational) relaxation results in a considerable overestimate of the energy of onset of predissocia- tion, because of the smearing out of the initial energy distributions.This point can be illustrated by the historical decrease in estimated DO, (IF) values, as a function of reducing pressure. Durie's6 value for DO, (IF) was <23 341 cm-', based on data from atmospheric-pressure flames; the first Clyne and McDermid value4 was <(22 700 15) cm-' from LIF data near 1 Torr. The definitive Clyne and McDermid value3 of (22 333 & 2) cm-' was obtained from LIF data below 1 mTorr. Earlier data on other dissociation energies obtained from the onset of predissocia- tion may well prove to give overestimated values. The overestimate may be serious when the excited state is long-lived (7, > 1 p s ) , or when measurements are made at higher pressures. A critical re-examination is desirable in such cases.Thus, very close limits for Do0 (IF) could be obtained. In earlier Prof. P. E. Siska and Prof. M. F. Golde (Pittsburgh) said: The electronic struc- The outer s ture of the metastable rare gases Ne* to Xe* is . . , (n - l)p5 ns. L. Stein, J. Wanner, H. Figger and H. Walther in Laser-induced Processes in Moleciiles, ed. K. L. Kompa and S. D. Smith, (Springer Berlin, 1979), vol. 6, p. 232. E. N. Okafo and E. Whittle, Int. J. Chem. Kinetics, 1978, 7, 273. M. A. A. Clyne and I. S. McDermid, J.C.S. Faraday 11, 1978,74, 1644. M. A. A. Clyne and I. S. McDermid, J.C.S. Faraday II, 1976, 72,2252. J. Wanner, previous comment at this Discussion. R. A. Durie, Canad. J. Phys., 1966, 44, 337.3 56 GENERAL DISCUSSION electron endows these highly energetic species with chemical properties remarkably similar to those of the alkali meta1s.l The imaginative and elegant experiments of Rettner and Simons as well as the copious and informative rate measurements re- ported by Setser thus herald the beginning of a " neo-alkali age ".From such studies we are likely to learn much about the alkali metal reactions that has long lain hidden, as well as a rich new chemistry engendered by the electronic excitation. Though the s electron is the major influence on metastable rare gas properties, the rare gas ion cores are isoelectronic with neutral halogen atoms, and thus are ex- pected to resemble these atoms more than the corresponding alkali metal ions in their chemistry. This halogen-like property may influence van der Waals forces at inter- mediate range involving the metastable atoms, and also may make chemi-ionization channels such as Ar* + Cl, -+ [Ar+Cl-] -+ ArC1+ + C1- energetically possible owing to the likely large bond energy of ArCl+ relative to Ar+ + C1.Evidence for the existence and stability of rare gas halide positive ions comes both from mass spectral plasma studies and ab initio calculation^,^ which suggest they are at least as stable as the isoelectronic halogen or interhalogen molecule. The well-known stability of ArH+ makes the Ar+ + HC1+ ArH+ + C1- reaction exoergic by 1.2 eV. With polyatomic halogen-bearing molecules, more bizarre ion chemistry may occur. We plan to look for production of such ions under single- collision conditions in a crossed-beams apparatus.Prof. D. W. Setser (Kansas) said: Siska and Golde suggest that the isoelectronic nature of the rare-gas ion core and halogen atoms may lead to chemical properties that resemble halogen atoms, such as formation of ArC1+ in some reactions. A re- lated suggestion to this was made by Thrush at the 1972 Faraday Discu~sion.~ We agree that the electron deficient core can be important. For example in some cases incipient ion-molecule chemistry may be important in the quenching mechanisms for the excited states of the rare gases.5 Another example is the results of ab initio calculations6 of excited states of KrF* which predict a high lying bound state that corresponds to the first Rydberg state of the series that terminates in KrF+. We have begun a search for ionic exit channels from quenching of Ar(3P,) using a flowing afterglow monitored by a mass spectrometer (see tables 2 and 3 of the Introductory Lecture to section 3 of this Discussion).The only ion found from the AI-(~P,) + Cl, reaction was trace amounts of C1+. Thus, the chemi- ionization reaction yielding ArCl+ and C1- suggested by Siska and Golde does not occur at thermal energies. The small C l i yield, which arises from Penning ionization is a measure of the colli- sions which stay on the V(Ar*, Clz) potential rather than branch to the V(Ar+, Cl;) potential because Penning ionization occurs primarily at the repulsive wall of V(Ar*, Cl,). The observation of a very low yield of C1+, thus, confirms the adiabatic pathway for the reactive quenching of AI-(~P,) + C12 at thermal energy.By analogy to the For a review, see M. F. Golde in Gas Kinetics and Energy Transfer, (Spec. Period Rep., The Chemical Society, London, 1976), vol. 2, p. 123. I. Kuen and F. Howorka, J . Chem. Phys., 1979,70, 595. For a review, see C. Thomson in Theoretical Chemistry, (Spec. Period. Rep., The Chemical Society, London, 1974), vol. 2, p. 83. B. A. Thrush, Faraday Disc. Chem. Soc., 1972, 53, 121. J. E. Velazco, J. H. Kolts and D. W. Setzer, J . Chem. Phys., 1978, 69, 4367. P. J. Hay and T. H. Dunning, Jr., J . Chem. Phys., 1977, 66, 1306.GENERAL DISCUSSION 357 reactions of alkaline earth metal atoms.' There are likely to be other examples for which chemi-ionization is an allowed exit channel. Prof. M. H. Alexander (Maryland) and Prof.P. J. Dagdigian (Johns Hopkins) said : In connection with the discussion of Simons and coworkers and the subsequent in- formal comment by Herschbach, we wish to point out that the orientation dependence of nonreactive molecular collisions could well provide a sensitive probe of the aniso- tropy in intermolecular potentials, especially if the torque exerted by the collision part- ner is substantially enhanced for particular approach geometries. Several recent theoretical s t ~ d i e s , ~ - ~ of varying degrees of sophistication, indicate that both rota- tionally inelastic and elastic collisions can lead to significant polarization and align- ment of the final molecular rotational angular momentum, even for unpolarized reactants. The nature and magnitude of these effects appear to be quite sensitive to the potential surface.3* The orientation dependence of rotationally inelastic collisions can be studied using for state selection and detection either electric quadrupole fields6 or lasers. As we and others have in the latter case the polarization of the radiation provides a natural means to detect nonuniformities in molecular mj-state distributions. An excellent example of this type of experiment is provided by the beautiful recent work of McCaffery and coworkers9 on collisional depolarization in excited electronic states of diatomic molecules. Prof. J. C. Polanyi (Toronto) said: Rettner and Simons'O note that the " repulsive " component of the energy-release in the exothermic reaction Xe*(3P2) + Br, --f XeBr* + Br could contribute to the XeBr* rotational excitation, J ' .As they surmise, this can be a significant factor under circumstances where the initial orbital angular momentum, L, is small; i.e., when the collision-energy imparted by their paddle- wheel is modest. A model study (3D Monte Carlo trajectories) for an equal-mass exchange reaction A + BC -+ AB + C showed markedly enhanced J' at low L on a more-repulsive surface as compared with a more-attractive p.e. surface." These model surfaces were both of the LEPS variety, and consequently favoured collinear reaction. In the case of Xe* + Br, the intermediate could be bent," thus channelling product repulsion still more efficiently into rotation.'* The same model study was used to examine the factors governing the polarization of J',11913 which is one of Rettner and Simons's variables." Particular attention was paid to the angle x between the product molecule angular-momentum vector J' and G.J. Diebold, F. Engelke, H. U. Lee, J. C. Whitehead and R. N. Zare, Chem. Phys., 1977, 20, 265. M. H. Alexander, J. Chem. Phys., 1977,67,2703; Chem. Phys., 1978,27,229. L. Monchick, J . Chem. Phys., 1977, 67, 4626. S. R. Kinnersley, Mol. Phys., in press. U. Borkenhagen, H. Malthan, and J. P. Toennies, Chem. Phys. Letters, 1976, 41, 222; J. Chem. Phys., in press. ' M. H. Alexander, P. J. Dagdigian, and A. E. DePristo, J . Chem. Phys., 1977,66, 59. D. A. Case, G. M. McClelland, and D. R. Herschbach, Mol. Phys., 1978,35, 541. S. R. Jeyes, A. J. McCaffery, M. D. Rowe, and H. Kato, Chem.Phys. Letters, 1977, 48, 91; M. D. Rowe and A. J. McCaffery, Chem. Phys., 1978, 34, 81 ; Chem. Phys., in press. M. H. Hijazi and J. C . Polanyi, J . Chem. Phys., 1975, 63, 2249. ' M. H. Alexander and P. J. Dagdigian, J . Chem. Phys., 1977,66,4126. l o C. T. Rettner and J. P. Simons, Faraday Disc. Chem. SOC., 1978,67, 329. l2 D. S. Perry and J. C. Polanyi, Chem. Phys., 1976, 12, 37. l3 M. H. Hijazi and J. C. Polanyi, Chem. Phys., 1975, 11, 1.358 GENERAL DISCUSSION the initial relative velocity vector, urel, since the measurement of this angle had been pioneered by Herschbach's The optical method of determining polariza- tion, exemplified in Rettner and Simons' work, is likely to be capable of substantially greater sensitivity and precision than the electric-field deflection method that preceded it.It may, therefore, be timely to point out that, according to the model calculation alluded to above, the best indicator of repulsive energy release as the source of product rotation was a tendency for x = 90" (highly polarised product4) to be observed in the sharply backward-scattered component of the reaction product but not in the less backward-scattered product. What are the prospects for measuring the polarization at more than one scattering angle? It is a pleasure to be in the position of asking for still more-refined data, rather than being asked for it. Dr. J. P. Simons, Dr. Y. Ono and Mr. R. J. Hennessy (Birmingham) said: Since we have extended the work to completing the paper included in this include a study of the reaction Xe(3P2,0) + IC1 --f XeI* + C1 ( 1 4 XeCl* + I.(1b) The chemiluminescence spectrum recorded from the cross-beam interaction at thermal energies shows fluorescence from both XeCl(B, C) and XeI(B, C) but with reaction ( l b ) the dominant branching channel. [In contrast, in the reactions with ClCN and BrCN, no trace of XeC1* or XeBr* has been detected, though there is strong emission from CN(A, B), presumably formed through predissociation of XeCN*, cf. SetserSb]. Preliminary measurements of the polarisation of the XeCl(B+ X ) emission as a function of the collision energy show it to lie close to the curve obtained for the reaction of Xe(3P0,,) with CCl,. In the case of IC1 however, there is no possibility of angular momentum disposal in the atomic product, and the low polarisation can be attributed solely to the " blocking " effect of the bulky I atom.Finally, the close similarity between the behaviour of Xe(3P3,0) and CC1, and Cs and CC1, reported by Riley, Siska and Herschbach,6 is gratifying, particularly on a Jubilee oc~asion.~ C. Maltz, N. D. Weinstein and D. R. Herschbach, Mol. Phys., 1972, 24, 133. D. S. Y . Hsu and D. R. Herschbach, Faruduy Disc. Chem. SOC., 1973, 55, 116. D. S. Y. Hsu, G . M. McClelland and D. R. Herschbach, J. Chem. Phys., 1974, 61,4927. D. S. Y. Hsu, N. D. Weinstein and D. R. Herschbach, Mol. Phys., 1975, 29, 257. (a) C. T. Rettner and J. P. Simons, Furaduy Disc. Chetn. SOC., 1979, 67, 329. (6) D. W. Setser, T. D. Dreiling, H . C. Brashears, Jr, and J. H. Kolts, Foraduy Disc.Chem. SOC., 1979, 67,255. S . J. Riley, P. E. Siska and D. R. Herschbach, Faruduy Disc. Chem. SOC., 1979, 67, 27. ' T. H. Bull and P. B. Moon, Disc. Faraday Soc., 1954, 17, 54.GENERAL DISCUSSION 3 59 ADDITIONAL REMARKS Mr. S. J. Buelow, Mr. D. R. Worsnop and Prof. D. R. Herschbach (Harvard) said: As emphasized by Prof. Siska, metastable argon atoms show a schizophrenic chemical personality: alkali-like at long range, halogen-like at short range. This prompts us to mention recent work which perhaps exemplifies the halogen-like character of argon ions and also may serve as a reminder that physical chemists still occasionally use the word “ state ” to refer to the macroscopic phase of matter. Our study concerns the interaction of argon ions with clusters of perchloroethylene molecules, (C2C14),,.The motivation stems from an exotic solar neutrino experiment.’ This employs a large tank of cleaning fluid (400 000 litres of C,Cl,!) in which a few neutrinos are captured (with cross section M cm’!) by chlorine-37 nuclei to form argon-37 ions via u + 37Cl + 37Ar+ + e-. The expected capture rate is about three neutrinos per week. The experiment requires circulating helium through the tank to flush out the 37Ar atoms (that presumably result from neutralization of the ions), which are detected by radioactive decay. The experi- ment has been pursued and refined for over ten years, with results which perplex physicists and astronomers. The data indicate a neutrino flux only about one third the predicted amount. The possibility that 37Ar+ ions produced by neutrino capture might be trapped in a molecular cage or compound has been suggested’ but discounted in view of a gas phase mass spectrometric experiment3 which found no evidence for argon molecule ions of the form ArC,,Cl,+.In this experiment, which employed a “ high-pressure ” ion source operated at M 1 Torr, the only observed process induced by Arf ions was charge transfer to produce C2Cl,‘ ions followed by drastic fragmentation to form CC1+, C2C1+, CCl,’-, C2Cl,’- and C,Cl;. However, this result may arise merely from the very large difference in ionization potentials of Ar and C2C14, x6.4 eV. The dis- posal of such a large exoergicity is a dominant factor under gas phase conditions, but need not be in the liquid medium which is M 106-times denser.In our experiment, we employed a supersonic nozzle (diameter 0.1 mm, operated at 110 “C with Ar at 20 atm and C2C14 at 10 Torr) in order to generate van der Waals clusters, Arn(C2C14)m. Ionization of such clusters by electron bombardment (at ~ 2 5 V) permits argon-containing molecule ions to be formed without ionizing argon in the clusters and hence avoids entirely the problem of disposing of a large reaction exoergicity. Indeed, we observed large yields of cluster ions of the form Ar,(C,CI,)+, with n = 1-29 and nz = 1-4. Still larger ions were surely present but beyond the range of our mass spectrometer. The role of ions such as Ar(C,Cl,),+ in the solar neutrino detector of course remains an open question. In these ions, the argon atom must be essentially neutral (by virtue of its high ionization potential); the binding nergy results from polarization by the nearby organic cation.On the long time-scale of the solar neutrino experiment, such polarization seems very unlikely to prevent flushing out the 37Ar by the carrier gas. If the 37Ar is nonetheless tied up by an ionic complex, it should be possible to release the argon by electrolytically discharging the ions. For recent reviews, see J. N. Bahcall and R. Davis, Science, 1976, 191, 264; J. N. Bahcall, Rev. Mod. Phys., 1978, 50, 881. K’. C. Jacobs, Nature, 1975, 256, 560. J. J. Leventhal and L. Friedman, Phys. Rev. D., 1972, 6 , 3338.360 GENERAL DISCUSSION Dr. G. M. McClelland (Stanford) and Prof. D. R. Herschbach (Haruard) said: The elegant experiment of Rettner and Simons has provided the first information about the energy dependence of angular momentum alignment in reactive collisions.Such directional or vector properties of reactions offer dynamical information comple- mentary to scalar properties such as rate constants and product energy distributions. Indeed, a potential energy surface in effect acts as a polarizing lens which induces anisotropies and correlations among the directions of relative velocity vectors and angular momentum vectors of the reactant and product molecules. The theory of angular correlations has proved very fruitful in analysis of nuclear reaction processes' and should have a comparable role in treating angular momentum properties of chemical reactiom2 Laser methods now offer the prospect of deter- mining angular correlations among the directions of the initial and final relative veloci- ties k and k' and the rotational angular momenta j and j ' of reagent and product mole- cules.In anticipation, we have calculated angular correlations for an A + BC + AB + C atom exchange reaction for both statistical3 and impulsive models4 Here we give some results calculated for comparison with the data of Rettner and Simons. The rotational alignment coefficient obtained from their data is the second Legendre moment of the angular correlation of j ' and k , defined by a2/ao = 5(P,(j'-k)). This coefficient is large and negative ( j ' tends to be perpendicular to k ) at high collision energies, but the alignment decreases rapidly and becomes very small as the collision energy descends into the thermal regime.Our calculations of the alignment employ a variant 2-4 of the impulsive DIPR model (formulated originally by Kuntz and Polanyi); this amounts to the spectator stripping model with addition of repulsion between the products. The repulsion is released when atom A approaches BC within the covalent-to-ionic crossing distance, R,. The mean repulsive energy released was taken as 35 kJ mol-' for the Br, reaction (by analogy with photodissociation) and 41 kJ mol-' for the CCl, reaction (chosen to simulate product translational energy data reported at this Discussion for the analogous Cs + CCI, reaction). In the latter case, the product radical CCI, was treated as an atom. Except at the limit of complete alignment (a2/a0 -+ -5/2), the DIPR model strongly over- estimates the alignment coefficient.According to this impulsive model, the align- ment persists even when the collision energy becomes negligibly small compared with the product repulsion energy. This result, which we find is essentially unchanged on introducing various forms for the orientation dependence of the reaction, arises from a Jacobian factor. The " dart board " distribution of initial impact parameters must be projected onto a sphere of radius R, centred on the BC molecule. A Jacobian weight- ing proportional to k*R, is thereby introduced and this anisotropic factor produces the residual rotational alignment obtained at low collision energies. We conclude that the experimental observation of near-zero alignment at low energies cannot be accounted for by simply invoking product repulsion.It will be interesting to see if other models or classical trajectory calculations can find a way to obtain nearly isotropic product rotation without going over to a statistical rather than L. C. Biedenharn, in Nuclear Spectroscopy, Part B, ed. F. Azzenberg-Selove (Academic Press, New York, 1960), D. 732. D. A. Case and D. R. Herschbach, J . Chem. Phys., 1978, 69, 150 and references cited therein. G. M. McClelland and D. R. Herschbach, J . Phys. Chem., 1979, 83, 1445 and references cited therein. D. S. Y . Hsu, G. M. McClelland and D. R. Herschbach, J . Chem. Phys., 1974, 61, 4927; G . M. McClelland, Ph.D. Thesis (Harvard University, 1979). Fig. 5 compares our results with the data of Rettner and Simons.GENERAL DISCUSSION + 361 -2.0 - 0 0 - 0 .5 ~ 1 0 20 40 60 80 100 120 collision energy, Et / k J mol-' FIG. 5.-Alignment coefficient for rotational angular momentum of XeBr* and XeC1* from reactions of Xe* with Br2 and CCl,. Points show experimental data of Rettner and Simons. Curves are calculated from DIPR model. direct reaction mechanism. For the pertinent trajectory calculations available at present,' product repulsion has at least qualitatively the same role as in the DIPR model. To the extent that the analogy between reactions of metastable rare gas atoms and alkali atoms is valid, the product angular distributions (correlation of k and k') should be quite anisotropic for the Br, and CCI, systems. Thus it would be parti- cularly interesting to see what kind of trajectories can give nearly isotropic product rotation accompanying strongly anisotropic reactive scattering. Rettner and Simons remark that the mass combination in the M + CH3X reaction forces the product angular momentum to show large polarization and thereby reveals little about the dynamics. Out of loyalty, we are compelled to note that in fact this proves not to be the case; the kinematic constraint resulting from the light mass of the methyl group is not very severe. An electric deflection study' of the Cs + CHJ --f CsI + CH, reaction found that the j ' , k correlation was only modest. However, it proved feasible also to determine a coefficient related to the triple vector correlation of j r , k r , k . This was stronger and corresponds to preferred orientation of the product rotational angular momentum perpendicular to the plane containing the initial and final relative velocity vectors. Model calculations and trajectory results' indicate such alignment is characteristic when strong repulsion occurs between the products. The preliminary results mentioned by Simons for the ICI reaction indicate a particularly striking comparison with the corresponding alkali reaction^.^ The dif- ference in electrophilic character of I and C1 has an interesting consequence. The attacking M atom transfers its valence electron mainly to the I atom, but the charge usually shifts to the C1 atom as the intermediate (IC1)- ion dissociates in the field of the M+ ion. Even when the electron jump occurs, however, the migration of charge to M. H. Hijazi and J. C. Polanyi, Chem. Phys., 1975, 11, 1 ; J. C. Polanyi, remarks at this Discus- sion. D. S . Y . Hsu, G. M. McClelland and D. R. Herschbach, J. Chem. Phys., 1974, 61, 4927; G. M. McClelland, Ph.D. Thesis (Harvard University, 1979). G. H. Kwei and D. R. Herschbach, J . Chem. Phys., 1969, 51, 1742.362 GENERAL DISCUSSION the C1 atom and dissociation of (IC1)- may be inhibited or precluded for a certain class of collisions. This is expected when the closest approach in the reaction tra- jectory occurs for a configuration in which the I (or I - ) atom blocks the excess of M (or M+) to the C1 atom. There is experimental evidence for this “ blocking” mechanism.’ Thus it is pleasing to see that the contrast between the rotational alignment observed for ICl and Br, offers further evidence for such a process. G. H. Kwei and D. R. Herschbach, J . Chem. Phys., 1969,51, 1742.