REACTION BETWEEN CO AND SO. AT ELEVATED TEMPERATURES: A SHOCK-TUBE INVESTIGATION 8. H. BAUER, P. JEFFERS,* ASSA LIFSHITZ,t AND B. P. YADAVAt Department of Chemistry, Cornell University, Tihaca, New York ‘The oxidation of carbon monoxide by sulfur dioxide was studied ina single-pulso shock tube lover the temperature range 1770°-2483°K. The reactant compositions were CO/S0,/Ar = 1/1/98, 4/4/92, 1/4/95, and 4/1/95, with initial reactant driver (He) pressure ranges of 27-170 torr and 50-175 psi, respectively. Reflocted-shock dwell times lasted 376-541 yseo. Gaschromatographic and mass-spectrometric analyses of the shocked samples indiosted that ‘no Os, 80, or 8:0 was produced and only traces of COS and CS; were generated. Close to 2 moles of COs, the principal product, was formed per mole of 803 which disappeared. At the ‘small conversions tested, precise stoichiometry could not be established, and the CO eoncen- trations were not directly measured. Although no quantitative check on the presonce of solid ‘sulfur could be obtained, there were several qualitative indications of sulfur deposition on the sshock-tube walls and on the VPC column packing. ‘An empirical rate expression was found to be (C03) /r = a(r)*(COP™SO)-” ky = 2.7 X 108 exp[—48,300 1,200)/RT} mole em* seo, with error limits of 40.1 on the partial orders, and 2 +t 0.1 for the total order. Three plausible mechanisms were co tionsl excitation process. 1, Introduction Knowledge of the basic chemistry of gaseous sulfur dioxide is meager, particularly where it pertains to reactions with hydrogen, carbon monoxide, and hydrocarbons. Chemists can only vaguely surmise the mechanistic pathways for generation of SO; during the combustion of sulfur containing carbonaceous matter. ‘There fare two recent, but as yet incomplete, studies ‘of the mechanism of its thermal dissocition,! while the most striking discovery regarding the reduction of SO, by saturated hydrocarbons remains the report by Dainton (1950) of the photosctivated production of sulfinie acids? More recently, the attack of SO; by I atoms in a discharge-low system at room temperature * Chemistry Department, SUNY—Cortland, New York 13045. {Department of Inorganie Chemistry, Hebrew University, Jerusalem, Israel t Department of Chemistry, Lucknow University, Lucknow, India The lered; arguments are presented against two of these in favor of a vibra- hhas been studied by Fair and Thrush,’ while Good and Thyme! investigated the reaction between SO: and CHy over the temperature range 25°-164°C. In contrast, the spectroscopy of SOs, its fiuorescence, phosphorescence, quenching re- factions, and chemiluminescent, reactions have been the subject of many publications. Search ‘of the chemical abstracts covering the years 1960 to the present uncovered 2 number of investigations on aerosol formation from gaseous air pollutants due to photosensitization of SO: in the presence of hydrocarbons,* and several ‘papers on the catalytic reduction of SOs. ‘The investigation of SO, reduction by CO is fa direet extension of our current shock-tube program on the kineties of CO-COs conversion by a range of potent oxidizing agents. In five reactions for which the rate constants are known, there is direct transfer of an oxygen atom to the CO in a bimolecular process (Gee Table 1), One may anticipate that, in an en- counter with $0:, carbon monoxide would abstract an oxygen atom, as in the NOs reaction; ar418 AIR POLLUTION TABLET ‘nd kinetic parameters for five CO oxidation reactions Alias? zk log 4 ‘Temp. range (keal/mole) _(cal/mole) (em mole-t see) (°K) (1) N.O + CO = Ns + COs 87.1 28,000 108 1300-1800 @) NOs + CO = NO +00, 54.1 27,600 12,10 1660-1800, @) OH +CO=H+ CO; 24.9 1,000 1.62 300-2000 @ OF $CO-F +00, ~76.2 ° 10.90 300-1400, (@) 0: + CO = 0 + C0, -8.1 50,000 12.55 1700-2600 (2) Lin, M. C. and Bauer, 8 H.: J. Chem. Phys. 60, 3377 (1969). (2) Bureat, A. and Lifshits, A.J. Phys. Chem. 74, 268 (1970); Thomas, J. H. and Woodma ‘Trans. Faraday Soc. 63, 2728 (196 (8) Schofield, K.; Planetary Space Sei. ‘This Symposium. GR: 6, 643 (1967); Brabbs, T. A., Belles, F. E., and Brokaw, RS. (4) Lin, M. C., Henrici H., and Bauer, 8. H.:J. Chem. Phys. 92, 5834 (1970). (8) Dean, A. M. and Kit however, the activation energy would be higher, since the process is endothermic (+83 keal/ mole) and higher shock temperatures would be required. This proved to be the case. Tt is interesting to note that in the combustion of 0 by O;, the initiation step involves the direct transfer of an oxygen atom, although the mechanism is dominated by ‘the subsequent chain reaction. ‘The essential steps appear to be adequately established.” 2, Experimental 21 Shock Tube These studies were conducted in a 1 Pyrex single-pulse shock tube. Its construction and mode of operation have been described previously." The driversection length was. ad- justed for each temperature range to provide 8 flat pressure record over the entire reflected shock dwell period. The test section was evac- uated to less than 5 X 10 torr before each run, and tests showed that the over-all “Teak rate” was less than 0.1 4/min. Diaphragms were of Mylar, 0.5 to 275 mil and mere presure bust (50 t0 175 psig). Initial test. pressures from 27 to 170 or nedentseck speeds were read from oscilloscope pressure traces produced by two Kistler transducers; estimated precision of Ty is -£35°K. Shocked samples were with- drawn from the downsiream end of the tube into 40-ml glass bulbs and were analyzed gas chromatography immodiately after extraction from the shock tube, comsky, G. B.: To appear in J. Chem. Phys. 2.2 Preparation of Samples ‘The materials used ineluded Matheson ultra- high-purity. argon, 99.99%, Matheson C. P. gmde CO, 99.5%, and Matheson anhydrous S0;, 90.98%. The CO was passed through » liquid-nitrogen trap, while the SO, was distilled from a liquid-nitrogen trap into a dry ice acetone trap before use; the argon was not repuified. ‘Aliquots of carbon monoxide and sulfur dioxide were withdrawn to prepare a 30% stock mixture (in argon), which was later used to prepare samples for the shock experiments. Standard manometric procedures were followed ‘n these dilution operations. ‘The glass vacuum system was pumped to 10°F torr, as were the 5:1 glasa bulbs used for storage, Mixtures were allowed to stand at least overnight before they were used, Their compositions were: 1%-1%, Jo-§7o, 8YoV%o, andl 4%-4% of CO and SOs in argon, 2.3 Analytical Techniques ‘The analyses for Ar, COs, COS, SOs, and CS, utilized a Hewlett-Packard Model 700 fas chromatograpb, with a heatod-filament, detector which was operated at 250 mA. Carefully matched dual 0.25 in. X 6 ft columns packed with 80-100 mesh Porapak Q affected separ: of the above species in the order listed. The flow rate of He carrier gas was 60 ml/min with the columns at room temperature. After COs was eluted (4 min at 25°C), the column tem- perature was programmed at 30°/min to 100°CREACTION BETWEEN CO AND $0: and was then held until COS and SO: appeared (@ and 40.5 min, respectively). Further program- ring at 30°/min to 180°C led to elution of CS: after 18 min, Temperature programming yielded ‘a much betier peak shape for SO; than was observed with isothermal operation at 100°C. The chromatograph was fitted with « glass gascous-sumple inlet system, comprising two Sway stopcosks to allow ‘either bypass or sweeping of a 60-ml sample loop, a manometer, 1 joint to accommadate the 40-1] siraple bulbs, and a vacuum pump. ‘The sample loop. was filled to a pressure of 25 torr in the usual analysis. Calibration eurves for CO;, COS, SO, and GS: were linear and passed through the origi. Detection limits for CO, were about 5X 10~ ‘mole, which allowed measurements on shocked samples which had conversions as low as 0.1%. Caleulation of percent conversion was based ‘on ratios of the CO: and $0: peak aress to the area for Ar. Relative sensitivities for CO; and ‘80, were determined from calibration samples. ‘An analytical method for O; was also developed, employing 1 Perkin-Elmer model 154 VPC with a thermistor detector. The column of 0.25 in X 10 ft Porapak Q in series with & 0.25 in X 15 ft column of Porapak R. separated Nz from Ar at room temperature, and separated Nj, Os, and Ar, in that onder, when the column, was immersed in a dry ice-acetone bath. Reten= ‘tion times were 14, 16}, and 18 min with 90 ml/min He carrier flow. ‘The Perkin-Ekner ‘chromotograph also had a glass gaseous-sampling arrangement, but with a 30-ml sample loop, and a vacuum system capable of obtaining 10 torr. A very good vacuum vias required to keep background peaks from masking the samples, Sensitivity for O; was about 100_ppm. This system could be used as a check for air von tamination by measuring the Ne peak at 0°C. The higher temperature shortened the elution time for Nz to about § min with o resultant narrowing of the peak. Detection limits of 10 ppm or less are estimated for this technique. No separation of CO from Ar was observed in either ehzomstographie method. ‘Mass spectrometry was also used to check sample composition, but CO at mass 28 and $0» at, mass 32 made Nz and Os determinations less reliable than was possible by chromatography. 3, Experiments and Observations Preliminary analyses of shocked samples indicated that the major detectable product was COs, The possibility that air contamination might provide O; for oxidation of CO was cheeked by introducing the test gas to the shock 419 tube through # glass spiral trap cooled with liguid nitrogen, No CO: was produced. when the SO, was froven out in this manner, demon- strating that SO; was the sole oxidizer in the system, Carbon disulfide and COS were formed only in trace amounts, less than 1% of the COs produced, Mass spectra disclosed no additional products, especially at mass 80, the parent peak for 80; and 8,0. Some yellowing and clumping of the Porapak Q was observed. The yellow material may have been either solid ‘sulfur filtered out of the gaseous samples or some decomposition products of some of the gascous species. ‘The relation (BO4/Ar)o= (8Oz/Ar), + (1/n)(CO/Ar), (1) was used to help establish the stoichiometry, inserting peak areas for Ar, SOs, and COs, ‘The factor n represents the number of CO; molecules formed for each $0; molecule con sumed. Early experiments suggested (rather roughly) that m~ 1. It was these results which spurred the search for molecular oxygen as a reaction produet, However, gas-chromatographie analysis showed that no O: was produced. ‘The quantitative evaluation of Eq. (1) is difficult, since conversions over 207% were’ not ‘easily realized; for lower conversions one must work with small differences of large numbers. The most reliable values for w in Eq. (1), a8 deduced, wore: 1% SOr-4% CO, n= 1.95, 2.43, 2,08, 2.80, 2.50; 4% S0s-4% CO, n = 1.76, 2.26, 17. An error anslysis indieated that up to 660% error in nis possible for the 4%-4% mixture, with shocks of about 10% conversion even with 2% accuracy in the gas-chromatography analysis, We concluded that the stoichiometry is best represented by 2CO +80; = 200: + 8(9) ‘The next step was to evaluate the exponents in the relation Rate « (Ar)*(COP GO), ‘These were derived by selecting sequences of combinations of exponents, calculating the corresponding rate constants and their least squares deviation from an Arrhenius relation. ‘The combination which gave the best least squares fit was (COs) /r = h(x) (COY™ SO)", @)420 AIR POLLUTION ‘TABLE TL Representative data from 87 experiments. Bimolecular rate constants were calculated with the power-rate expression, using partial orders for Ar, CO, and SQ: 0.57, 0.76, and 0.67, respectively. pro GAR 100 tT % conversion % conversion No. (Gore) (mole/emm!)—— (usec) (PK) tog hs [801),/$0,], at equilibrium 1% CO, 1% 80s, high density 10s 140 0.524 470 2320 7.852 85 20 109155, 0.546 4701925 7.055, 2a m4 wm 160 0.572 4821986 7.238, 36 im 139 0.513 4820 257.419, 46 m4 168 0.578, 494 106 7.022 24 1% CO, 1% 804, low density 2040 0.141 5051925 6.885 0.5 ey 22 30 0.109 505 20757-4837 03 Be B 0.102 BAL 2102 7.495, 12 2640 0.153, 47 458.196 83 20 274s or 482-2316 7.911 5.0 1% €0, 4% 80: m8 0.179 470240 7.761 22 1 9 70 0.251 5171770 8.801 02 19% 40 0.160 4392243 7.794 2.0 20145 0.156 4821673 6.000, o4 2 2945 0.181 47 2205 7.739 2.0 4% CO, 1% 80. 22 Be 0.136 4 2445 8.222 2.8 “ Dot 45, 0.162 S17 2028 7.238, 3.0 48 20740 0.138 5171838 6.497 08 a m4 45 0.167 4942180 7.641 8.0 2845 0.170 ar 57.765, 10.0 6 4% ©, 4% 80: uz 40 0.358 423 2167 —7.680 44 1m 45 0.152 400 2902, 7.787 6.3 29 W245 0.175 423 ost 7.235 7 12 aT 0.108 435 263 7.308 1.0 m2 0.168 423 2ITS 7-608, £0 25 ‘This dwell time was taken as the recorded width of the ps pressure trace, ‘The total order is 2 (to within 0.1), All data 4. Discussion points were reduced on this basis and @ repre- sentative sampling is assembled in Table Il; aan Arrhenius plot is shown in Fig. 1, The least- squares value for the rate constant (mole-tem* sec") is logy = (12.48 + 0.12) = (48 200-4 12901/2.9R7 ‘To obtain some estimates of the driving foree for reaction we calculated the thermodynamics ‘equilibrium compositions anticipated under re- flected-shock conditions for typical choices of initial composition. Mole fractions, as a funetion of reflested-shock temperature are plotted in Fig, 2. In these calculations, the equilibrium-REACTION BETWEEN CO AND 80; ob 045 050 21 055 3 WV, x10 Fic. 1. Temperature dependence of the bimolecular rate constant evaluated according to the empirical ‘expression ‘A(COs)/r = ba(Az}*27(CO}97¥(803)*7; units are mole/em 2 1% 80e-1% CO, high densities; @ 1% SOs-1% CO, low densities; C1 19% SO4% CO; V 4% S0-1%, 60; & 4% 807-4% CO. shock program prepared at CAL? was used. ‘The thermodynamie-fit parameters were taken from JANAF tables, The species inserted in ‘the program are: 80s, CO, COs, 80, 02, 0, 8), 82(9), 8:0, COS, 80;, C8, CS. It is evident from the plots in these figures that complete ‘conversion is not expected for the temperatures ‘covered in these studies except when a significant ‘excess of SO: is present. Neither COS nor CS: should be generated in appreciable amounts (the minute concentrations we detected are probably consequence of a kinetically controlled process), and at clevated temperatures the principal products are COs, $; and SO. In view of the observed less-than-unit-order dependence on $0, and the definite participation ‘of argon in controlling the rate, we must diseard not only the direct bimolecular transfer of an ‘oxygen atom but also that reaction path which appeared as the logical alternative 80:-+$0,+80, +80. Alfie? = 47.83 keal 802+ CO—+ 803+ COs Ang? = ~44.04 Keal ©0480-0048 AHlys? = 2.5 kcal. al Also, because of the observed low activation ‘energy, we did not consider further an initiation involving dissociation of the SO,(De = 1313 eal); E, = 110 keal/mole for S0:(A)— SOCE) + OGP). However, it is possible to devise an energy chain for the net conversiona AIR POLLUTION MOLE FRACTION MOLE FRACTION a 5 Se MOLE FRACTION iy MOLE FRACTION ic) we T T at 2° 38} aa ge es Sa. zal gab fh ah *, ee bet staat Bho th — abo asd \(d) 18 Tk Fro, 2. Equilibrium mole fractions ve 7, calculated with the CAL program. Tntial conditions are ps = 40 torr, and [Link] speeds 08 to 1.3 min/ysee. (a) 1% SO<-4% C0; (b) 2% 801-4% CO; (0) 4% 80s-4% CO; (1) 1% 80s-1% CO; (e) 49% 801-1% CO; (a) 4% 802% CO.REACTION BETWEEN CO AND $0; CO + 80;— CO; + 80: = SO:+M ‘ 801+ M 1 2 80,5—>80 + 0 3 ©CO+ 0+ (M)—+ COs + (M) 4 COs? + M—+ COn+ M 5 (COs! + 80, —+ 802+ COs, [B] By imposing steady-state conditions on S02", O, ‘and CO;#, one readily derives the anticipated rate: d(COn)/dt = Kee (M) 8O1)/Tx 4M) — 1; ts (802) (M)/Lau(M) + ws(SO)e]. (8) Q ‘This relation is over-all first order, and has no ©O dependence. ‘The form of Eq, (2) and the relatively low activation energy suggests that this atom- transfer process is limited by a vibrational excitation proves, Hence, we propoze the follow- ing mechanism: <1 802+ Ar S$ SOs Ar 1 -1 CO+ Ar #5 CO*+4 Ar v 2 SO* + CO—> 00; + 80 2 802+ CO*—+ CO, + 80 3 804+ CO—> CO, + 80 4 $0 +80, 80;-+ 8 803+ CO—+ CO; + 80; SS (lid). [C] 423 Imposing steady-state conditions on $03, $0, $03", and CO*, mechanism [C] leads to 4(C0.)/dt = 2hyke (CO}(S0,) (Ar) X {Uea(Ar) + (CO)T* +a8/Th (Ar) + k'2(60:)]}, 4) where a = Wi/tx and B = Ka/ky. ‘The rate equation is overall second order and permits the allocation of partial orders to exch reactant, provided the rates of deexcitation in step 1 are comparable to the rates of the atom transfer in step 2. Steps 3, 4, and 5 lend to the observed stoichiometry, and we postulate that the solid sulfur generated either deposits on the shock- tube walls or is carried “as finely suspended particles into the VPC, where it is filtered out by the column packing, ‘The production of very small amounts of COS is accounted for on the basis of the infrequent termolecular ste CO+S + Ar COS + Ar, One may argue that the steps which involve CO* play a considerably lesser role than those with 8O;* (Le, « and are both much smaller than unity), The vibrational relaxation times for S02 with argon as a collision partner have not been measured over the temperature range used in these studies. However, at 300°K, the ratio of self-collision relaxation times (atm-see) (pr sos! @riJoo™ 6X 10-7/5 X 10-4, deduced from ultrasonie absorption measure- ‘ments At the higher temperatures, with argon colliders, this ratio is undoubtedly larger. Nevertheless, one may with confidence assume that the useful vibrational excitation is derived from the sulfur dioxide moiety, and itis plausible that such excitations are particularly effective when in the molecule which loses the oxygen atom, This is not inconsistent with the observa- tions by Bureat and Lifshitz," who found no argon depenvdenee for the rate of production of carbon dioxide from the analogous reaction NO, ++ CO—> NO + COs. First, the relative rates of atom transfer from NO,’ vs SO; at 10/T = 0.0[<1606°K] is 104/108, due to the large difference in activa-48 tion energies, Second, although no data are available for’ the vibrational relaxation times of NO», these must be 10 to 50 times faster than for 80s, and hence the nitrogen dioxide reaches its Boltzmann population close to the shock front. While both molecular symmetries are Cx, the ground state for SO: is '4y, while ‘Non isin a 841 state; the stronger intermolecular interaction potential for the latter is demon- strated by its well known dimarisation behavior. The fate of the SO and 8 species may be more complex than is indicated by mechanism [C]. The summary of the chemistry of the oxides of sulfur? directs attention to $0. Although the reaction S0-+S0:~>S:0 ++ 0; appears to be highly improbable, it is exeluded in our case by the demonstrated absence of O2 as a product. ‘The formation of ($0)z and sub- sequent’ attack by @ third $0, could produce 8:0 and S0z; algo the reaction between (80): and 8; gives 2 8:0. However, the equilibrium calculation does indicate that ‘S:(9) is a major product over the entire temperature range of this study, and thus solid sulfur is expected as a final product. To a certain extent the resulta plotted in Fig. 2 ‘were surprising, While SO is a potent oxidising agent such that, at around 1000°K, thermo dynamic eriteria require extensive conversion 2802+ 4CO>4. 002 + &, the rate for homogeneous conversion is low, because of the appreciable activation energy (48.3 keal/mole). At the higher temperatures, where the reaction is signifcantly faster, equic librium eonditions favor the left side of the above equation. This may be noted from the position of the crossovers of the C0./CO curves in the sequence Fig, 2(a,b,c,) and from the last column in Table IZ, which lists the pereent conversion, [80z](final)/(80; J(origi- nal), ealeulated for equilibrium under shock conditions. Clearly, the extensive conversion of these noxious materials to the less-polluting products must be performed at the lowest temperatures possible, with the aid of an ap- propriate catalyst Acknowledgments This investigation was supported by PHS Re- search Grant No. I-ROL-AP00633.01, from Office AIR POLLUTION oof Research Grants, National Air Pollution Controt Administration. REFERENCES 1. (a) Gavnon, A. G., Knenett, G. 1, axp Pauaen, H. B.: Proc. Roy. Soe. (London) A279, 513 (1964); (b) Taox, J. anv Wacnst, 4H, Go.: 28th Propulsion and Energetics Panel Meeting of AGARD, May 1966, Oslo; (c) Ouscnuwext, H.A., Tron, J, axo Wacssn, H. Go.: 2, Phys. Chem. (NF) 44, 178 (1965). 2. Dametox, F. 8. anp Ivin, K. J-: Trans. Fara day Soc. 46, 374, 382 (1950). 3. Farm, R. W, anp Tremuse, B. A: Trans. Para- day Soe. 65, 1550 (1908); see also Katsann, A. 8.:4. Phys. Chem, 70, 2055 (1068). 4. Goon, A. ano Tavwe, C. T.: Trans. Faraday Soe. 64, 2708 (1967). ‘5. A few current references will suffion: McKewzrr, A. anp Tunusu, B. A.: Proc. Roy. Soc. 4808, 139 (1968); Merrie, H. D.: J. Chem. Phys. 49, 1784 (1968); Smarckume, 8. J. and Howsu, B. P: J. Chem. Phys. 49, 1947 (1988); Furtcuen, 8. R, ax Luvrrs, B. P.: Trans Faraday Soc. 65, 1544 (1968); Rao, T. N., Couumn, 8. 8, Axo Cauvenn, J. G.: J. Am. Chem, Soe. 97, 1609, 1616 (1968); Maree, HD. J. Phys. Chem. 73, 1071 (1969); Cons, N. ax Guoss, R. W. F.: J. Chem. Phys. 60, 3119 (1968), 6, For example: Korcerwsxs, 8. L. np ALTSHUL~ us, A. P.r Intern, J. Air Pollution 6, 138 (1962). 7. Drax, A. M. anp Kistiaxowaxy, G. B Re port to ONR, March 1970, Contract Nour 1866(80) ; submited to J. Chem. Phys. 8. Lirsmry, A, Baves, 8. HL, ano Resta, E. J., Jn: J. Chom. Phys. 38, 2058 (1063), 9. Winusaws, M. J. ano Gann, L.: “Equilibrium Normal Shock Computer Program,” Cornell Aeronautical Laboratory, Buffalo, New York, September 1966. 10, Srevex, Brian: Collisional Activation in Gases, Pergamon Press, Oxford, 1967, AL, Buneat, A. ap Lursmmms, A.: J. Phys, Chem. 74, 208 (3970). 12, Seumxk, P. W. ano Srexpet, R.: Ang. Chem. (Dnt. Ba.) 4, 402 (1965). COMMENTS J. Troe, University of Goettingen. In the low= temperature pyrolysis of S02 (7 < 3000°K), there has also been some trouble with respect to reaction order and activation energy (~50 kcal/mole). For these eonditions, participation of electronieally excited SO: has ‘been proposed. ‘What do you think about a similar mechanism in 4 flow system? Authors? Reply. It is our impression that the separation between the ground and first excited lectronic states of SO; is significantly greaterREACTION BETWEEN CO AND $0, than 50 keal. Hence, it is more reasonable to assume that the excited form of $0: involves vibrations. However, as far as the mechanism is concerned, the nature of the excited state need not be specified. H. B. Palmer, Pennsylvania State University. Isit really appropriate to say that a $0: exponent of 0.67 in the rate equation corresponds to a first-order dependence on SO: concentration— ie., that it is compatible with your proposed ‘mechanism? Authors! Reply. See our General Remarks (ollowing Comment). Authors! Comments. We have stated that, by analogy with the NOs + CO reaction, itis prob- able that a-8 <1. Clearly it cannot be 2ez0, since that would foree the predicted partial order Of the rate with respect to [S0,] to be unity, whereas the observed dependence is 0.67. We hresitate, however, to use the observed power dependence as the sole basis for evaluating a8. 125 ‘We have not covered a sufficiently large range of relative concentrations in determining the rate law to give sufficient precision for this type of an evaluation, A. P. Myerson, Reso Rescarch and Engineering Co, Have you considered the alternative mecha- nistie possibility of ehain branching? Although not ‘many oxygen atoms may be formed from $0, not ‘many are noeded when branching is present, for example, via radical and energy ehains which right involve the very energetie step CO + O-> COs + 250,000 eal. The presence of chain branching would also bo consistent with the small overall activation energy. Authors! Reply. We have not considered a branching chain mechanism in any detail. We believe that it does not apply because itis far from. clear as to how any chain mechanism which dominates the reaction would lead to an argon dependence as large as the one which empirically fits our rate data.