Reactions of methylene with dichloromethane in the presence of

Unimolecular HCl and HF Elimination Reactions of 1,2-Dichloroethane, 1,2-Difluoroethane, and 1,2-Chlorofluoroethane: Assignment of Threshold Energies...
1 downloads 0 Views 1MB Size
Download PDF
1670

W. G. CLARK,D. W. SETSER,AND E. E, SIEFERT

Reactions of Methylene with Dichloromethane in the Presence of Carbon Monoxide and the Collisional Deactivation of Vibrationally Excited 1,2=Dichloroethaneby Carbon Monoxide and Perfluorocyclobutanel

by W. G. Clark,2 D. W. Setser,a and E. E. Siefert Chemistry Department, Kansas State University, Manhattan, Kansas 66602

(Received November 21, 1969)

The reactions of CH2, produced by ketene photolysis, with CHzClzin the presence of variable quantities of CO was investigated. The addition of small quantities of CO to mixtures of ketene and dichloromethane reduced yields of products associated with triplet CH2reactions; however, large amounts of CO (GO: CHnGlz = 30: 1) were required to decrease the triplet level to below 5%. The results indicate that singlet CH1 abstracts only chlorine and triplet CH, abstracts only hydrogen from CHzClz; the rate constants are of comparable magnitude. Singlet methylene reacts with the double bond of butene-2 and with CHnCll at the same rates. The ratio of reaction rate constants for CH2(?Zg)with CO and CHzClz was estimated at ~ 3 The . collisional deactivation of chemically activated 1,2-CzH4C12by carbon monoxide and perfluorocyolobutane was studied, and the average amount of vibrational energy lost per collision was measured as 6 -f 2 and -12 kcal mol-l for CO and G F g , respectively.

Introduction Methylene, when produced by photolysis of ketene or diazomethane, reacts with chloro- and bromomethanes by H and C1 or Br abstraction but not to an appreciable extent by insertion,*s6 Since both singlet ('A1) and triplet (3&-) methylene are produced from photolysis of ketene6 or diazomethane,' it has not been possible to unambiguously assign elementary reactions to the singlet or triplet methylene with the halomethanes, Carbon monoxide preferentially removes triplet methylene,* and the addition of CO to a ketene photolysis system enhances the singlet CH2 reactions. Bamford and coworkers9 used the CO technique to study the reactions of singlet methylene with chloroethane; they concluded that CH2('A1) reacted a t least 16 times faster by C1 abstraction than by H abstraction while the CH2(3Z:,-) mainly reacted by H abstraction. We wish to report photolysis experiments of ketene with CHzClz using CO to remove selectively CH2(3Z2,-). The ratio of singlet to triplet methylene was directly monitored by adding cis-2-butene t o the reaction mixture and measuring the ratio of cis- to trans-1,2dimethylcyclopropane.'' Results will be presented which generally agree with the conclusions of previous workers*>Swho used CO to remove CH2(3Zg-). However, even for large ratios of CO: CHLYz, up to 30: l, a small amount of CH2(3Z,-) was still present, and some care must be used in deducing singlet vs. triplet methylene reaction mechanisms and rate constant ratios by the CO addition technique. For example, it seems likely that the C-H abstraction measured by Bamfordg was due to residual triplet methylene, and that the rate The Journal of Physical Chemistry

constant ratio for H us. C1 abstraction by CHt('A1) is less than 1: 50. I n spite of the slight complication mentioned in the above paragraph, the addition of convenient amounts of CO can greatly simplify CH2 plus chloroalkane reaction systems because secondary radical-radical processes associated with the H abstraction reaction are reduced. Thus, ketene plus chloro- or bromoalkane photolysis systems with added CO provide a way for studying nonequilibrium unimolecular reactions of chemically activated chloro- and bromoalkanes. For suchs tudies with haloalkanes or other CHAnitiated (1) Abstracted from the M.S. thesis of W.G. Clark, submitted in partial fulfillment of requirement of the Master of Science Degree, Kansas State University, Manhattan, Kansas, 1969. (2) NASA Predoctoral Fellow. (3) Alfred P. Sloan Foundation Fellow. (4) (a) J. C. Hassler and D. W.Setser, J . Chem. Phys., 45, 3237 (1966); (h) J. C. Hassler, D. W.Setser, and R. L. Johnson, ibid., 45, 3231 (1966); (c) C. H. Bamford, I. E. Casson, and R. P. Wayne, Proc. Roy. Soc., A289, 287 (1965); (d) K. Dees and D. W. Setser, J . Chem. Phys., 49, 1193 (1968). (5) R . L. Johnson and D. PV. Setser, J . Phys. Chem., 71,4366 (1967). (6) (a) R. W.Carr, Jr., and G. B. Kistiakowsky, ibid., 70, 118, 1970 (1966); (b) S. Y. Ho and W. A. Noyes, Jr., J . Amer. Chem. Soc., 89, 5091 (1967) : ( c ) A . N. Strachan and D. E. Thornton, Can. J . Chem., 46, 2353 (1968): (d) P. G. Bowers, J . Chem. Soc., A, 466 (1967). (7) (a) B. M.Herzog and R. W.Carr, Jr., J . Phys. Chem., 71, 2688 (1967); (b) F. H. Dorer and B. S. Rabinovitoh, ibid., 69, 1952 (1965). (8) (a) B. A. DeGraff and G. B. Kistiakowsky, ibid., 71, 3984 (1967): (b) . . R. A . Cox and R . J. CvetanoviE, ibid., 72, 2236 (1968). (9) C. H. Bamford, J. E. Casson, and A . N. Hughes, Proc. Roy. Soc., A306,135 (1968). (10) T . W.Eder and R. 15'. Carr, Jr., J . Phus. Chem., 73, 2074 (1969). This reference summarizes the problems and controversies of the CHz * cis- or trans-butene technique of measuring singlet and triplet CHz ratios.

REACTIONS O F METHYLENE WITH DICHLOROMETHANE I N THE chemical activation systems, the deactivating efficiency of CO for the vibrationally excited molecules is required so that the rate constants observed with CO as the deactivating gas can be related to the intrinsic unimolecular reaction rate constants, which are the rate constants obtained with gases having near unit deactivation efficiencyall The nonequilibrium HC1 elimination rate constant for 1,2-C2H4Cl2*formed by combination of CHzCl radicals served as a prototype, and the collisional deactivation efficiency of CO was measured and compared to an efficient gas, perfluorocyclobutane.

Experimental Section Photolysis samples used to study the effects of variable CO : CH2Cl2 ratios were prepared by transferring the CHzCO and CHtCl2 into vessels of desired sizes using standard vacuum techniques. For some experiments cis-2-butene was added to the CH2C12-CH2CO mixtures; in other cases cis-2-butene replaced CH2C12. The noncondensable CO was then metered into the vessel, which was immersed in liquid nitrogen, until the desired pressure of CO was obtained. The experiments were done at constant pressure, usually 350 Torr, with a nearly constant amount of CH2Clz (2.1 cc) and constant CHzCO:CHzClzratio of 1:7. Thus the vessel size increased as CO : CHzClzwas increased. Dichloromethane was Fisher Certified Reagent Grade; the cis-2-butene was Phillips Research Grade and contained no detectable trans-2-butene; carbon monoxide was Matheson CP grade. Mass spectral analysis shows the CO to contain less than 1 part per thousand 0 2 . Ketene was prepared and purified as described p r e v i ~ u s l y . ~Photolyses were done in Pyrex vessels at room temperature with the unfiltered light of a General Electric A-H6 high-pressure mercury lamp. A rough calculation, using the lamp output, molar absorbancy of ketene, quantum yields of ketene,12 and transmittance curves of Pyrex, showed that -75% of the ?ethylene originated from photolysis at 30003200 A and -25% at 3200-3400 A. The condensable products were recovered by pumping the photolyzed sample through a glass wool packed trap at 77°K. These products were then transferred to a gas chromatograph inlet and injected onto the columns described below, which were followed by a thermal conductivity detector. For the products from the CHz CHzClzreaction, the column consisted of 3 f t of Porapak S and 6 ft of Porapak T. For the products from the reaction of CH2 with CH2Clz and cis-C4H8 mixtures a double-pass analysis scheme was necessary. The entire sample was first passed through the Porapak columns for analysis of chloroethanes. The CsHlo products along with CHZCl2were trapped from the glpc effluent and cis and trans-dimethylcyclopropane were analyzed on a column consisting of 15 f t of 12% diisodecyl phthalate on Chromosorb P in series

+

PRESENCE O F

CARBON MONOXIDE

1671

with 15 ft of 40% of AgNOa-saturated ethylene glycol on Chromosorb P. Gas chromatographic peaks were identified from retention times of pure samples and from mass spectra cracking patterns of samples trapped from the glpc effluent. Empirical calibrations were made with prepared standard samples which were near replicas of the photolyzed samples. Also the prepared calibration samples were subjected to the same recovery procedure that was used for the photolyzed samples as a check for possible handling losses. Series of experiments were done with constant CO: CHzC12:CHzC0 ratios (10: 1:0.2), but a t variable total pressure, in order to measure the rate constant for HC1 elimination from vibrationally excited CzH&12*. CH&l Some experiments were also done using CH2 rather than CH2 CH.#212 as the source of CH2Cl radicals. The samples were photolyzed as described above. Since low-pressure experiments are needed to obtain collisional deactivation efficiencies, vessels ranging from 10 to 10,000 cm3 in volume were used. Constant quantities of CHzClz (2.0 cm3 of gas) were used to improve reliability of analyses, which consisted of Porapak columns in a gas chromatograph unit equipped with hydrogen flame detectors and a temperature programmer. Careful empirical calibrations of the sensitivity of the gas chromatograph unit for 1,2-C2H4Cl2 and CzH3Clwere made. A double-pass analysis was necessary for the c-C4F8experiments. The C4Fs was removed by putting the samples through a Porapak column (thermal conductivity detector) ; the other components were trapped from the He effluent and were subsequently run back through the glpc with the H2flame detector.

+

+

Results Methylene Reactions with Dichloromethane in the Presence of Carbon &''onoxide. The photochemical processes (eq 1 and 2) are the formation of singlet and triplet methylene, which may react by either H or C1 abstraction (eq 3-6) .* The photodissociation of CH2CO is undoubtedly not as simple as implied by (1) and (2). Triplet ketene is involved in the mechanism and 41 and 43 are wavelength dependenta6 On the basis of previous data, it was not possible to eliminate any of the four following reactions although reactions 4 and 5 were thought to be the most important. The methyl and chloromethyl radicals subsequently combine in all possible binary combinations, but the relevant reactions for our present purposes are CHzCO

+ hv +aCHz + CO

(1)

+ CO

(2)

$8

$1

+ 'CHz

(11) (a) D. W. Setser and J. C. Hassler. J. Phus. Chem.. 71. 1364 i1967);'(b) H. W.Chang and D. W. Setser, JI Amer. Chem: SOC., 91, 7648 (1969). (12) G.A. Taylor and G. B. Porter, J . Chem. Phgs., 36, 1353 (1962). Volume 74, Number 8

April 16, 1970

W. G. CLARK,D. W. SETSER,AND E. E. SIEFERT

1672

+ CH2C12--%! 2CH2C1

corrected (see a later section for details) to give the infinite high-pressure yield. The results from the experiments in which the ratio -% CH2 CHClz (4) CO :CH2C12was varied and the C2HsC1and 1,2-C2H4C12 lCHz CHzC12-% 2CH2C1 (5) product yields were measured are shown in Figure 1; the C2H5C1yield has been corrected for the unimolecular 3 CHI CHC12 reaction of C2H5Cl*, The C2H&1/1,2-C2H&12 ratio with no added CO was 0.9 and agrees with previously 2CHzC1+ CH2CICH&1* (7) measured values.4* The addition of two parts of CO CHa CHzClCHaCHzCI* (8) relative to CH2Clz reduced the yield of C2H6C1 from Further addition of CO reduced -50% t o -15%. 2CH3 ---f C2He* (9) the C2H6C1but not to the same degree. The detailed interpretation of this curve is delayed until the DisAt high pressure the C2H6to 1,2-CzH4Cl2ratio gives k s [CH3I2/k7[CH2C1J2and the C2H5C1t o 1,2-CzH4Cl2 cussion. However, it is clear from Figure 1 that CO drastically reduced the yield of C2H&1, which is ratio gives ICs [CH3]/k7 [CHZC~].From previous work equivalent to reducing the concentration of CHg it is known that kT = ks = 1/24, as would be expected radicals. Since small amounts of CO apparently reduce for simple radical combination reactionsOi3 Thus these [CH3] and since CO is known to remove CH2(3Zg-),it is two product ratios can be used to obtain relative steadyobvious, on a qualitative basis, that reaction 4 is favored state radical concentrations; this technique only gives an over reaction 3. The experiments with large amounts approximation to the true instantaneous relative radical of CO, ;.e., small [CH2(3Zg-)],show that CH, radicals concentration, since integration of the ratios of reaction are not present; hence, reaction 5 is greatly favored rates to give product ratios assumes that the relative over reaction 6. radical concentrations do not vary with time. Some experiments were done to measure the ratio of The molecules formed in steps 7-9 are highly vibraC2He t o C2H4C12 yields as a function of [CO]/ [CH2CI]. tionally excited, -90 kcal mol-', but a t the pressure The results15are very similar to Figure 1 and show that used in these experiments, 350 f 50 Torr, the vibrathe concentration of CH3 radicals is drastically tionally excited 1,Bdichloroethane formed in step 7 is reduced by small quantities of CO. With no added completely s t a b i l i ~ e d 350 ; ~ ~ Torr also is sufficient to CO, the CHzCl and CHClz radicals combine t o give quench the ethane dissociation reaction.14 Since the C Z H ~ C ~this ~ ; ~product " was absent from experiments half-quenching pressure of the C2HsCl*formed by reacwith ten parts of added CO. tion 8 is 350 TorrSdin an efficient quenching gas, about Reactions of Methylene with Mixtuyes of CH2C12and half of these molecules will decompose to form HC1 and cis-d-C4H8and CO. To confirm that singlet CH2 only ethene at the pressure used in these experiments. If abstracted chlorine from CH2C12, it was desirable to the yields of chloroethane were used to estimate methyl measure [3CH2]/[1CH2] a t high concentrations of CO radical yields, the CzHsCl measured a t 350 Torr was and see whether the residual [CH,] was due to small concentrations of CH2(Q,-) or to a slow C-H abstraction reaction by CH2(lAAI).The stereospecific reactions with singlet CH2with cis- and trans-2-butene have been I used by many investigators1°to measure [%H2]/ ['CH,]. 0 We used cis-butene because it provides the most sensiT 06t O't tive measurement of the triplet component. Most investigators agree that the singlet CH2 addition is stereospecific providing the pressure is sufficiently high to quench the unimolecular cis-trans isomerization of O!Ol OiO2 d.03 o d the chemically activated 1,2-dimethylcyclopropanes u 0 . (DMC).lo I n contrast there is little agreement for the 3CHz

(3)

+

+

+

+

i-l

1

$1

1

0

2I

4I

6I

B I

1I0

p,C~2J/[COl Figure 1. T h e ratio of the CZHECl yield (corrected t o high pressure) t o the sum of the CtHjCl plus 1,2-CzH&lz yields from photolysis of ketene with CO and CHzClzmixtures. The insert shows t h e d a t a for high values of [CO]/[CH&lz] : 0 , P = 350 Torr; A, P = 1600 Torr: The Journal of Physical Chemistry

(13) J. M . Tedder and J. C. Walton, Proor. React. Kinet., 4 , 37 (1967). (14) M . L. Halberstadt and J. R. McNesby, J . Amer. Chem. Soc., 89, 3417 (1967). (15) W. G. Clark, Master's Thesis, Kansas State University, 1969. Some of the conclusions reached in this reference differ from the ones in the text of the paper; the latter takes precedence. (16) (a) J. W. Simons and G. W. Taylor, J . P h y s , Chem., 73, 1274 (1969); (b) D. W. Setser and B. S. Rabinovitch, Can. J. Chem., 40, 1426 (1962): ( c ) J. W. Simons and B. 8. Rabinovitch, J . P h y s . Chem., 78, 1322 (1964).

' 4-J

12

1673

REACTIONS OF METHYLENE WITH DICHLOROMETHANE IN THE PRESENCE OF CARBON MONOXIDE ratio of cis- and trans-DMC resulting from addition of Although the CH2(38,-) to cis- or trans-2-b~tene,~~ suggestion that CH2(3Z,-) plus either cis- or trans-2butene gives the same ratio (2.9) of trans- to cis-DMC is attractive,17most investigations have found different ratios. We will use the recent valuelo of 1.8 for the ratio of trans- to cis-DMC from reaction of CH2(3Zg-) with cis-butene; this ratio is 2.7 for CH2(YZCg-)plus trans-2-but ene. Experiments were done at 350 and 800 Torr total pressure with various combinations of CO, CH2C12, cisand cis-C& The trans-DMC/(Irans-DMC DMC) was measured as 0.18 in the absence of CO or CH2C12. This ratio indicated a 28% triplet methylene reaction system. Such a calculation assumes 'ICC,HJ 3 k ~ a=~ 1.0, 8 a view which, although widely held, does not seem to have been proven. Our 28% value agrees with conclusions from other photolysis studies of CH2Cis-DMC) C0.6-8 The trans-DnlIC/(trans-DMC ratio was unchanged by the addition of CH2C12, and it follows that [CH2('A1)]/[CH2(38;g-) ] also was unchanged. This has several ramifications. (1) The chloromethanes apparently do not induce fast collisional crossing of CH2(lA1)--t CHz(3Xg-). (2) Since reaction of CH2(lA1) and CH2(%Zg-) with CHzClz did not affect [CH2(1A1)]/[CH2(3Zg-)1, it seems that 3 k ~ / 1 k ~=1 3kCaHs/1kCaHs.' However, it is always possible that this overall result may be the fortuitous combination of several effects, Adding CO to mixtures of CHzC12 and cis-CeH8 affects the yield of trans-DMC as shown in Figure 2. Ten parts CO reduced trans-DMC to a value which corresponds to 15% triplet methylene; further addition of CO, up to 50 parts, reduced the triplet level to 5 7%. This residual triplet CH2 presumably also will be present in the CO-CH2C12 system. Our results differ from DeGraff and Kistiakowsky in that 10 parts CO was not entirely sufficient to remove CHz(32,-).

+

+

Figure 3. Ratio of CzHaClzlcis-dimethylcyclopropane yields us. the composition of the CH2Clz and C4H8 mixture, the [CO] was 10-20 times the sum of [CHzClzl and [CaH8] and the total pressure was 350 Torr.

This difference may arise from slightly different experimental conditions (different light intensity or spread in wavelength) or because DeGraff and IGstiakowsky used trans-CeHg to measure CHz(32g-). Ten per cent CH2(3L1g-)would give [cis-DMC]/ [cis- trans-DMC] = 0.02 (using [trans-DMC]/[cis-DRIIC] = 3.6 from the CH2(3Z,-) trans-C4H8reaction), and this small quantity could have been overlooked. DeGraff and Kistiakowsky also found a more rapid removal of the nonstereospecific component of the reaction at intermediate [CO] than shown in Figure 2. The ratio of reaction rates of CHZ('A1) with cisC4Hs (double-bond addition only) and CHzClz is proportional to the ratio of cis-DMC and 1,2-CzH4CI2products. Photolyses were done for various [CH2C12]/ [cis-C4H8]with 10-20 parts of added CO. The results, corrected for a small contribution to cis-DMC from CH2(3Zg-), are shown in Figure 3. The slope of the line gives % C I / ~ ~ = C ~1.0 H ~f 0.1. Loss of CHzCl radicals by H abstraction from CHzCl2was monitored by observing the formation of CH3C1. Under most conditions the CH3Cl yield was negligible relative to the C2H,Cl2 yield. The two points denoted by the squares of Figure 2 have quite different ratios of CH2C12 to cZ'S-C~H~, but the [trans-DMC I/ [transcis-DMC] ratio was unchanged. The same observation was made when no CO was present, Le., the addition of CHzClt does not alter the steady-state ratio, [CHZ('Al) I/ [CH2(32g-)], in a ketene-butene-2 photolysis mixture. The information from Figures 1-3 will be integrated with published information in the Discussion to obtain rate constant ratios for CH2(%Zg-) with CHzC12relative to

+

+

+

0.12

0 0

A 0.06

A

c

0.0

I

0.0

Figure 2 .

I

I

I

0.04 0.08 ( CHzCIz t C,H,)/ CO

Ratio of [trans-DMC]/[trans-

+ cis-DMC] yields

us. amount of CO added to mixtures of C H & and cis-butene: 0, [CH2C12]/[cis-C4H8]= l, P = 350 Torr; [CH~Clzl/[cis-C4Hs]was either 3 or 0.3, P = 350 Torr; A, [ C H Z C ~= ~ ]0, P = 800 Torr.

,.

cos

Collisional Deactivation Eflciency of CO and C4F8 with l,W-CzH4C&*. Since CO was not expected to be a highly efficient deactivating gas, the competition be(17) (a) C . MoKnight, P. 8.T. Lee, and F. S. Rowland, J. Amer. Chem. Soc., 89,6802 (1967); (b) D. C. Montague and F. S. Rowland, J.Phys. Chem., 7 2 , 3705 (1968).

Volume 74, Number 8 April 16, 1970

W. G. CLARK,D. W. SETSER, AND E. E. SIEFERT

1674

t

/ O

CH2 and chloroalkanes; our earlier s t u d i e ~ ~ ~of, l l ~ CH2 CHzC12 always showed excess vinyl chloride yields due to complicating secondary radical reactions. These complications were removed by addition of CO because the product yields associated with the hydrogen abstraction reaction of CH@Zg-) were diminished. Assuming that CO has the same deactivating efficiency toward C2H6C1*as it does for l,2-C2H4C12*,the measured yield of CzH&l can be corrected for decomposition in mixtures of CH2C12and CO. This was done by using D / S = ( C U X C O X C H W ~ W with P [C2H~C11m given by ( D / S ) S S. The deactivation efficiency on a pressure for pressure basis of CO relative to CH2C12 is a: = 1.55; x denotes the respective mole fraction of bath gas, S is the measured yield of C2H5C1,IC, is the specific rate constant for chloroethane decomposition (35 cm in pressure units),4dand [CzH5C1], is the yield of chloroethane a t infinitely high pressure.

+

+

I /Pressure,

Torr-'

Figure 4. Ratio of the product yields of vinyl chloride-l,2dichloroethane us. 1/P for photolysis of (a) 0, CHZCl or 0, CHzC1z,_CH&O, and CO mixtures in the ratio of 10: 1.5:105; (b) A, CHSCl, CHzCO, and c-CeHs in the ratio of 10: l.5:82. The curve through the CO bath gas points was calculated from a stepladder deactivation model with AE = 5.7 kcal and adjusted to the experimental k," of 27 Torr. The curve through the c-CtFs points i s just a straight line through the d a t a points and is not related to a model.

tween collisional deactivation and unimolecular reacwas studtion of the chemically activated 1,2-CzH4C12* ied and compared t o a more efficient gas, perfluorocyclobutane. Figure 4 is a plot of the ratio of the CzH3C1 decomposition product, D , to the C2H4CI2stabilized product, 8,us. 1/P. The analysis of these plots has been previously discussed;ll the specific rate constant for the unimolecular reaction may be expressed as IC, = COD/#,where w is the collisional frequency of CzHdClz* with bath gas molecules. The two deactivating gases illustrate the difference between an effi, a less efficient, CO, deactivator. cient, c - C ~ F ~and The cascade deactivation by CO is evident from the curvature of the plot. I n the region for D I E 5 0.5, both sets of data are essentially linear and the slopes of the lines in this pressure region are a measure of the overall relative deactivating efficiencies. These slopes CHzCl2 are 27 Torr and 13.5 Torr for CO and c-C~FB; had a value of 17.2 T ~ r r . ~ " ,Converting ~* these values to units of sec-' by using Lennard-Jones collisio? dia m e t e r ~(u(CzH4C12) ~~ = 5.5 A, u(C0) = 3.69 A, u(CHtC12) = 4.90 A, ~ ( C ~ F = B )7.03 8) gives ka = 3.09 X lo8, 1.63 x lo8, and 1.76 X lo8 sec-I for CO, GF8, and CH2C12. A more detailed analysis based upon the curvature of the CO plot, which gives the average energy transferred per collision, is presented in the Discussion. It should be noted that experiments using both CH3C1-C0 and CH2Cl2-COmixtures were done and the results, Figure 4, were identical, which increases our confidence in the data. Also it shows the usefulness of CO in simplifying the chemistry of systems involving The Journal of Physical Chemistry

+

Discussion General Reactions of Methylene with Dichloromethane. The curve in Figure 1 has an intercept, [CO] 03, that is very small. Since CO removes CH2(32Cg-), -+

it is apparent that CH2('A1) reacts with dichloromethane by abstracting a chlorine atom. We attribute the small amount of CzHsCl remaining at high concentrations of CO to residual CH2(s2,-) and not t o a small amount of H abstraction by CH2('A1). Insertion reactions, either C-H or C-C1, were not observed in this study or in previous work4,6which sets a limit to these reactions of less than 3% of the main chlorine abstraction reaction. It is more difficult t o show directly that CH2(SZg-) only abstracts hydrogen. However, mass balance arguments require all of the CH2(3&-) for generation of CHS in order to properly account for the methyl products. Triplet CH2 has some properties that are similar to conventional radicalsjZ0and abstraction of H rather than C1 is similar to the behavior of CF3 with halomethanes. I n fact the abstraction of C1 by CFs has a 5 kcal mol-' higher activation energy than abstraction of Hn2' Triplet CH2 also abstracts secondary H more readily than primary H which is consistent with normal radical behavior.20 There are two studies which must be considered before our interpretation that CHZ('A1) abstracts only C1 and CH2(3&-) abstracts only H can be taken as entirely correct. DeGraff and Kistiakowskysa found that only 10 parts CO was needed to reduce the triplet (18) D. W. Setser, J . Amer. Chem. SOC., 90, 582 (1968). (19) J. 0. Hirschfelder, C. F. Curtis, and R. B. Bird, "Molecular Theory of Gases and Liquids," John Wiley and Sons, New York, N. Y., 1964, pp 1210-1214. (20) S. Krzyzanowski and R. J. CvetanoviE, Can. J . Chem., 45, 665 (1967). (21) (a) TV. G. Alcock and E. Whittle, Trans. Faraday Soc., 62, 664 (1966); (h) J. I. G. Cadogan, D. H. Hey, and E'. G. Hihhert, J. Chem. SOC., 3939 (1965).

REACTIONS OF METHYLENE WITH DICHLOROMETHANE IN THE PRESENCE OF CARBON MONOXIDE methylene to zero in their reaction system of methylene with trans-2-butene. We have already pointed out that their analytical technique was less sensitive than the one used here; they could easily have overlooked the small yield of cis-DMC associated with the residual triplet CHz. Bamford, et al. J 9 have reported that singlet methylene abstracted chlorine a t least 16 times faster than hydrogen, and triplet methylene abstracted hydrogen seven times faster than chlorine. However, their conclusions were based on data obtained with only six parts of added CO, instead of using very large amounts of CO as was done in this study. It seems likely that Bamford, et al., had insufficient CO to remove all of the triplet CHz; hence, the results were obtained from a mixture of singlet and a small amount of triplet methylene. We feel that the room temperature reactions 4 and 5 are faster than 3 and 6 by factors of nearly 50. Explaining the fast C1 abstraction reaction of CHZ(lAl), even on a qualitative basis, presents a challenge. The discrimination for C1 abstraction us. H abstraction cannot be explained on the basis of the heats of reactions, since both reactions would have about the same values. We suggest that CHz('A1) initiates the C1 abstraction reaction by accepting an electron pair from chlorine into its vacant orbital. This is followed by bond cleavage of Cl-CH&l and internal electronic rearrangement of the newly formed CHt-CI. I n this sense the CHZ(1A1) is behaving as a Lewis acid and should show electrophilic character for this type of reaction. This is consistent with the observed5 greater reaction rate with CH3Br6 relative to CHSCI. Singlet methylene does not exhibit electrophilic character in the reaction with olefins.20 However, the double bond addition reaction involves attack by the pair of nonbonded electrons, In contrast, we propose, as did Bamford and coworkers, that the vacant orbital is important for reaction with the C1 or Br in chloro- and bromomethanes. Reactions of Methylene in Mixtures of CO, CHzC12, and CHzCO. The effect of small amounts of CO upon the product yields is quite dramatic and is explained by the more rapid reaction of CO with CH2(3Zg-)than with CHZ(lA1). We shall try to formulate a mechanism which matches the observed variation of the CZH&I/ CZH~CIZ product yield ratio with [CO]/ [CHZCL]. Ignoring possible collisional effects of the CHzCO* formed by reaction of CHZand CO, the following equations may be written

Collisional induced intersystem crossing is a way of preparing CH2(3Zg-),and quantitative dat'a have been

1675

published for rate constants with various third bodies.22 cH2(1A1)

+ nf -% C H ~ ( ~ Z ~+- nil )

(12)

We include reaction 12 with carbon monoxide, although it is not demanded by our data. The main reactionsz0 of CHQAI) and CH&2,-) with CHzCO also are needed

+ CHzCO +CzH4 + CO CHS(~Z,-)+ CHzCO +CzH4 + CO kCH200

CHz('A1)

(13)

%onzoo

(14)

(eq 13 and 14). Removal of CHzCl and CH3 by abstraction reactions is slow under our photolysis conditions and such reactions will be neglected. Steadystate expressions for ['CH2] and [%HZ] may be obtained from reactions 1, 2, 4, 5 , and 10-14. Substitution for ['CH2] in the [aCH2] expression and taking the ratio indicated below leads to eq I. The second ' ~ H [ ~ C H-Z ] lkci ['CHzI

[CO1 'kci [CHZC~ZI (1) 'kco [co] 3 k ~[CHzCO] ~ z ~ ~ 3kH[CHZCl2] [CHZCIZI 'k3

+

term arises from reaction 12 as a source of CHz(3Zg-). Only rate constant ratios are involved and it is possible to assign values to most of them; Table I gives a summary. Setting [CO]/[CHzC12] = 0 in eq I gives eq 11.

[CHzCIt] [CH&O]]/[

' ~ 'kci CH~CO

' k [CHzCL] ~ [CHzCOI]

+ '~CH,CO

(11)

The right-hand side of (11) is the ratio of triplet to singlet quantum yields modified by various rate constant ratios, and if 3kH = llccl, the right-hand side of (11) gives the steady-state ratio of singlet-triplet methylene. If the rate constants for CHzClzare changed to those for C4Hs, eq I1 applies to the butene-ketene experiments. The rate constant ratios are available from Table I, the experimental value, Figure 2, of %C4H8[3CH2]/1kc,~,[1CH2] is 0.39, and &/+I = 0.37 or 0.38 depending upon which 3 k ~ ~ / 3 value k ~ 4 ~is sused to determine 3kCH2C0/3kCIH8. It is clear that for small ratios of (22) (a) T. W. Eder, R. W. Carr, Jr., and J. W. Loenst, Chem. Phys. Lett., 3 , 520 (1969); (b) R. A. Cox and K. F. Preston, Can. J . Chem., 47, 3345 (1969); (c) R. D. Koob, J . Phys. Chem., 7 3 , 3169 (1969).

Volume Y 4 , Number 8

April 16, 1970

W. G. CLARK,D. W. SETBER, AND E, E. SIEFERT

1676

yield ratio and 3kH [3CH2]/1k~l[1CH2]. The factor was 0.89. This factor was assumed to be independent of the amount of CO added to the reaction mixture, Rate constant and the data from Figure 1 were used to construct the ratios Value Referenoe experimental points of Figure 5. It would have been 3 k ~ ~ / 3 k ~ 3~ . 6$ ~ ~ See ref 8a ~~ desirable to obtain values of 3 k H [ 3 C H s ] / k[lCHz] 0.14 4~0.02 'kco/ 'kcRzco See ref 8a more directly from experimental measurements of See ref 22b 0.121 ir 0 . 5 total product yields from the CH3, CH2Cl, and CHClz See ref 88 1 . 3 =t0 . 3 3kCO/akCaHs radicals. Without careful and simultaneous control 2.08 See ref 8b of light intensity, pressure, radical concentrations, etc., 0.10 =t0.02 See ref 8a 1.0 rt 0.1 This work, Figure 3 which are very difficult t o do while adding large quan0.10 'kC1 'kC4He tities of inert gas, such experiments would not be worth'kco ' ~ C H ~ C O -.while. These problems emphasize the need for develop%Haco/'kci 0.71 'koi -. 'krn __ ment of methods for direct monitoring of both [3CHt] 'ka/'kcHzco 0.01 See ref 8a and ['CH2]as a function of all experimental variables. 3kE/8kCaHs 1 .o See a Equation I was used to assign values of 3kco/3kH by '-.-.-. k ~ r ~ as k ' ~~ C H~~ C O 'kCHzCo/3kH 0.36 or 0.58 comparing the calculated and experimental values of %H a k ~ ,'kco ~ ~' also see b 3 k [3CHz]/1k~~ ~ ['CHZ]at various [CO]/ [CHzC12]. The a This ratio was derived from 1koi/'ko4as = 1.0 and 3 k ~ / 1 k =~ ~ comparisons are summarized in Figure 5. The ratios of rate constants for 'CH2 reactions are listed in Table * k ~ ~ ~ ~ / the ' k error c ~ ~ limits ~ ; are probably no better than 307,. (See text for further discussion.) ' The two values correspond I. The value of 3kCH&3/3kH can be estimated as to t h e two values for ~kco/*kcrac This ratio also can be obindicated in Table I or it can be expressed as ( 3 k ~ o / o ) . text tained from the relationship ( a k c o / a k ~ ) ( 3 k ~ ~ z c o / a k g(See 3kH)(3kcH,co/3kco) The only unknown remaining in for further discussion.) eq I is 3kCo/3kH,and a series of curves were calculated with this ratio varying from 1.3-9.0; the best fit obtained for a k ~ ~ / = 3 k3-5. ~ The effect of changing [CH2CO]/ [C&Hs]the quantum yields are hardly al'k3/'kCl from 0.01 to 0.1 is indicated by comparison to tered by the rate constant ratio expression of eq 11. the dotted line of Figure 5. Our data do not set a The value of 43/41 determined above also was used clearly defined limit to this rate constant ratio; howfor the CHZC12 experiments. Since the yield of [transever, best agreement seems to be for 3kc0/3kHof -5 and DMC]/ [cis-DMC] was independent of the ratio of 'kg/'kCI of -0.1. This value is a little higher than [CH2C12]/ [ c ~ s - C ~ Hwe ~ ] ,can use the data points with the 'k3/'kCHpCO value of 0.01 suggested by Kistiakowsky no added CO with the relationship that 3 k ~ [ 3 C H 2 ] / but is not inconsistent with the rate constants for to evaluate 'kcl ['CHZ] = 3k~4~s[3CH~]/11c~4~s[*CHZ] similar third bodies.22 The initial rapid drop in the the proportionality factor between the CzH6C1/CzH4C12 curves of Figure 5 is due to the importance of the term, 3~co/31cH, in the denominator of eq I. For very large [CO]/ [CHZCl2], eq I reduces t o Table I: Rate Constant Ratios a t 300°K for Yetliylene Radicals

I

1

0.44

ak~[3CHz] - (ba ' h o -!'lcc1[lCHz] - $1 %GI

[

- - -; -* 3s 0o I3

-*-a0

1'0

do

-40

5'0

6'0

Figure 5. Comparison of calculated, eq 11, and experimental values of 3 k ~ [ 3 C H ~ ] / 1 k ~ ~ [us. 1 C[CO] H ~ ] /[CHzClz]. The points represent the experimental data; the lines are calculated for values of rate constant ratios appearing on the right side of the graph. The dashed curve shows the effect of increasing 'ka/'kcl from 0.01 to 0.1 with 'kco/'k~ = 5.0.

Xhs Journal o j Physical Chemistry

'"I $

3lcn

-b lkcl

akco

('I1)

For 3 k ~ / 3 k=~ 5.0 ~ and 'k3/lkcl = 0.1, the limiting ratio is 0.04; this is reduced to 0.01 for 'kS/'le~l = 0.01. It is interesting to note that even if %3/'kC1 = 0, eq I11 has a finite intercept of -0.01. Therefore, the CO addition technique could never give more than a 99% singlet methylene reaction system. We concluded that, within our experimental error, CH2(lA1) added t o the double bond of cis-CeHs at the same rate as C1 was abstracted from CHzClz. Qualitative arguments also indicated that ' k c ~ / ~ = k~ 'kCaHs/%C4Hn, and since 'kcl = ' ~ C , H ~ ,we should find %C4Hs = %H. A more direct comparison can be obtained from the rate constants for C4Hs and CHzClz relative to CO. We obtained 3 k ~ ~ / values 3 k ~ of 3-5 which are similar to the higher value for 3kC0/3C& of 2.08 measured by Cox and Cvetanovi6. It is usually assumed that kc,^^ = 1 k ~ 4and ~ s that the quantum

REACTIONS OF METHYLENE WITH DICHLOROMETHANE IN THE PRESENCE OF CARBON MONOXIDE yield ratios of eq I1 are equivalent to [aCH2]/[1CHz], Since the same value for a h ~ [3CHz]/ik~1~s[1CH2] ,~a is obtained under a wide variety of reaction conditions in several different laboratories, we agree that 'kC4H8 N 'kCdH8,which in turn forces the conclusion that = 'kcl. Cox and Prestonzzbfound that CHZ(aZg-) reacts much faster with O2 than with CO. Thus these H and C1 abstraction reactions may be relatively fast but reaction certainly does not occur on every collision and may be as slow as 1/102-10a collisions.z3 Other Reaction Systems. We have made some preliminary investigations of the reactions of CH2 with 1,2-C2H4C12 in the presence of ten parts carbon monoxide. At pressure above 5 Torr the system behaves as expected and the only products found were fromradicals formed by chlorine abstraction by singlet CH2. However, a t lower pressures complications were encountered which can be attributed to inefficient removal of CHZ (3Eg-) by reaction 11. This suggests that collisional stabilization and reverse dissociation of triplet CH&O* formed in reaction 11 may be competitive in this pressure range. Another problem which can arise in the CHzCO-CO-chloroalkane systems is secondary abstraction processes. These also tend to be more important at low pressure because the radical concentrations decline since vessels become very large in order to hold reactant proportions constant. We also encountered problems with secondary radical reactions in the CHzCO-CD3C1-C0 reaction system a t high pressures for unknown reasons. Collisional Deactivation of CzH4C12*by CO, CHzC1z, and c4F8. The exothermic reactions of singlet CH2 provide numerous chemical activation systems. The addition of CO as a bath gas can serve t o simplify the chemistry of such studies, and we have measured the relative efficiencies of CO, CH2C12,and c-C4F8which are 0.50:0.77 and 1.0 on a pressure for pressure basis. The use of conventional collision diameters converts this to a collision for collision basis of 0.53 :0.93 : 1.0. These efficiencies were measured for D / S I 0.5 and are sufficient t o give an estimate of the unit deactivation nonequilibrium rate constant for chemically activated molecules in a CO bath gas. The methods used t o analyze the D / S plots in order to obtain detailed vibrational energy transfer information from the high- and low-pressure regions have been presented e 1 ~ e w h e r e . l ~ ~ experimental If ~ ~the rate constant for some gas, A, is defined as k,(A) = W A D / Sthe , high-pressure rate constant ratio for gases A and B according t o simple cascade deactivation model with initially formed monoenergetic CzH4C12* has the form

The p term allows for uncertainties in the values chosen for the collision diameters needed to calculate W A and

1677

u B ;T a and T g are the number of steps, of size AE, in the cascade, If the gas used for comparison has nearly a unit deactivation efficiency (a Ah' of -14 kea1 mol-' is equivalent t o unit deactivation) and if the cross sections are known, the high-pressure ratio can be used t o assign values of AE. These AE values correspond to the average energy, (AE), removed from the active degrees of freedom of C2H4C12* per collision (according to a symmetric model of collision transition probabilities). A calculated plot of eq IV os. (AE) for simple stepladder cascade of C2H4C12* with the p ratio equal to unity is shown in Figure 6. The full 300°K distribution function of the activated CzH&l2* was used in the calculations. This curve may be used to obtain an estimate for (AB); for example k,"(CO)/k,"(C4H8) gives (AB) = -5 kcal mol-'. The analysis just described for finding (AE) is an approximate procedure, a t best, because of the uncertainties associated with the collision diameters. Analysis of the curvature of the low pressure data, Le., variation of k, with pressure, is a more reliable way of finding (AE),11,24 and it is independent of the values chosen for the collision diameter or of the specific rate constants, k,. This was done for the CO data, and the calculated results from a simple cascade model with (AE) = 5.7 kcal mol-' are shown as the solid line in Figure 4. Since high- and low-pressure data gave the same answer, the average energy removed from C2H4C12*per collision with CO can be placed with confidencezsaas 6 f 2 kcal mol-*. Problems of analysis made the low-pressure data for C4F8 difficult t o obtain. However, no evidence for an increase in k , with pressure (or any c o r r e ~ p o n d i n g ~ ~ ~ ~ ~ curvature in the D / S plot) was found for D / S 5 2.0. This can be interpreted as ( A E ) 12-14 kcal mol-'. We previously found that CHzClz and CHsC1 removed 11 + 2 lical mol-' from CzH4C12*. Furthermore, the rather uncertain data previously obtained

-

(23) W. Braun, A. M. Bass, and hf. Pilling, Fifth International Conference on Photochemistry, 1969. This flash photolysis study suggests that C-H abstraction by CHZ(~Z:,-)is a relatively slow reaction. (24) (a) G. H. Kohlmaier and B. S. Rabinovitch, J . Chem. Phus., 38, 1709, 1692 (1963); (b) D. W.Setser, B. S. Rabinovitch, and J. W. Simons, ibid.,40, 1751 (1964); (c) D. C . Tardy and B. S. Rabinovitch, ibid., 48, 5194 (1968). (25) (a) This AE is from a simple cascade model, which is equivalent t o the average energy, ( AE), transferred per collision for a symmetric distribution of energy transfer probabilities. For values of ( A E ) near 6 kea1 mol-', the curvature of the D / S ~ 8 1/P . plot is very similar for even the two extreme types of energy transfer probabilities, L e . , the exponential or the symmetric distribution models Our data, a t the present time, are not capable of distinguishing between these two types of models; (b) N. L. Craig and D. W. Setser, unpublished data. Preliminary work with chemically activated CH3CF3 in a CaFs bath shows ( A E ) = 14 kcal. This conclusion is based upon the actual curvature of lowpressure data, such as shown in Figure 4 for CO. Reinvestigation of the deactivation of chemically activated dimethylcyclopropane by cis-butene-2 (J. D. Rynbrandt and B. S. Rabinovitch, J . Phys. Chem., 74,1679 (1970) has confirmed24b the previously measured ( A E ) of 11-12 kcal mol-'. These data suggest that the large deactivating molecules, such as CaFs and C4H8, may reach an upper limit of perhaps 12-14 kcal mol-' for the amount of energy removed per collision: (0) large values of ( A E ) , such as 12 kcal, are only compatible with a symmetric distribution of energy transfer pr0babilities.24~2' Volume 74, Number 8 April 16, 1970

W. G.CLARK,D,W. SETSER,AND E, E.SIEFERT

1678

.I

\,,

20

,

hcal mole''

Figure 6. T h e calculated values of ksm(AE)/kam (unit), vs. (AB) for chemically activated CnHdCl%*according to a stepladder deactivation model.

for K2,(AB) % 6 kcal mol-', are in agreement with the present result with CO. We conclude that the change from an optically inactive vibration in NZ to an optically active vibration in CO had no effect upon the amount of energy removed from 1,2-CzH&12* by collisions at 300°K:. This must mean that the longrange dipole-dipole interactionsz6 are not very important for this type of energy transfer. This was anticipated since Ar was only slightly less effective than N2 in removing energy from C2H4C12*or C3HB*,24b which implies that vibrational-vibrational transfer is not a significant mechanism for Nz or CO colliding with molecules such as C2H4CI2* or C3HBn27 The conclusions regarding efficient collisional vibrational energy transfer with polyatomic molecules recently have been challenged by Thrush and Atkinson.28 They interpret their data from the photoexcitation of cycloheptatriene in terms of removal of small quantities of energy rather than the large values derived

The Journal of Physical Chemistry

from studies of chemically activated chloroethanes, hydrocarbon radicals, or cyclopropanes. The work reported here and elsewhere for 1,2-CzH.&lZ* 2 9 and lib new work30 with 3,3-dimethylhexyl-2* 1,2-C2H4F2*, and 3-methylhexyl-2* radicals, and studies 31 with NOz* all favor efficient collisional energy transfer with collision diameters near gas kinetic values and large values of (AE)per collision. At present the evidence is overwhelmingly in favor of the conclusions drawn from chemical activation studies. It should be emphasized that the curvature of the D / S us. (pressure)-' plots provides information regarding (AE)which is independentZ4 of the values for k , and of collision diameters. The same type of information from the cycloheptatriene study should be i n t e r e ~ t i n g . ~ ~ Acknowledgment. Acknowledgment is made t o the donors of the Petroleum Research Fund, administered by the American Chemical Society, for support of this research. (26) (a) B. H. Mahan, J . Chem. Phys., 46, 98 (1967); (b) J. T. Yardley, ibid., 50, 2464 (1969). (27) D. C. Tardy and B. S. Rabinovitch, ibid., 48, 1282 (1968). These authors analyzed the collisional efficiency of several gases from thermal activation experiments with CHaNC; Nz transferred somewhat more energy than Ar in this case which involves excitation of CHsNC to about 40 kcal mol-'. (28) R . Atkinson and B. A. Thrush, Chem. Phys. Lett., 3, 684 (1969). (29) E. E. Siefert and D. W. Setser, unpubIished data for Ne with CzH4Clz* presented at the Gordon Conference on Molecular Energy Transfer, 1969. (30) C. W. Larson and B. Rabinovitch, J. Chem. Phys., 61, 2293 (1969). (31) S. E. Schwartz and H. S. Johnston, ibid.,51, 1286 (1969). (32) NOTEADDEDI N PROOF.Recent studies of the reactions of CH2 with chloropropanes [C. H. Bamford and J. E. Casson, Proc. Roy. SOC.,A312, 141, 163 (1969)] are in accord with the main conclusions of this paper. Additional evidenceZSfor slow abstraction reactions (>lo9 collisions with CD4) by CH2(?Z,-) has been reported [P. S. T. Lee, R. L. Russel, and F. s. Rowland, Chem. Comm., 18 (1970)l.