598
JOHNF. REEDAND 13. S. RADINOVITCH
lower the C.M.C. by a factor of ten and this effect ascribed wholly to the lowering of the electrical free energy of the micelle, the latter term is decreased by only 1.4 Itcal./mole. Therefore it, is bchvcd that in view of the crudeness of the present treatment, we are justified in neglecting 1s
Vol. 61
the electrical terms as a first approximation. Acknowledgments.-The authors wish to thank Mr. P. D'Arcy for invaluable help in the calorimetric determinations. Also they are most indebted to the Lever Brothers Company for providing the sample of sodium octyl sulfate.
"HE SODIUM DIFFUSION FLAME METHOD FOR FAST REACTIONS. 11. REACTIONS OF FLUORINATED METHYL CHLORIDES1sZ BY JOHNF. REEDAND B. S. RABINOVITCH Conlrilrution from the Department of Chemistry of the University of Vashinglon, Seattle, Washington Received November 9, 1868
The reaction of sodium atoms with the series CHaCl, CFH&l, CF2HCl ft"d CFaCl has been studied. After primar abstraction of chlorine, subsequent reactions of sodium with fluoromethyl radicals are of importance. A new upper bouFg ary condition for the limiting visible s o d i u m r has been applied. Various models for the calculation of s p e c ~ u ratc conRtants are diecussed and compared. e activation energies of the rimary step in the series are found to be 9.8 10.1, 10.0 and 9.2 kcal./mole, re8 ectively. These are used to obtain D(Cl!H&l), = 81-82 D CFnH-Cl) = 81-82 and D(CFrC1) = 80-81 based on D(Cft-Cl) = 81.2 kcal./mole. These enabled calcdations of elf,( Fa) = - 119.5, D(C$a-F) = 117.5 and D(CFrCFa) = 64 kcal./mole. Some other relevant thermochemical qunntities for fluorocarbons and limits of rate constants for sodium atom reaction with radicals are calculated and discussed.
d
In a previous paper,a an analysis of the experimental variables of sodium diffusion flames was presented. In the present work, application is made to the reactions of sodium atoms with the series of compounds CHaCl, CFH&l, CF2HCl and CFaCI. The nature and limitations of some of the assumptions made in experimental models used to describe the steady-state phenomena are examined in the light of the experimental findings, in an effort more accurately to define the rate constants. Corrections for radical reactivities are explicitly included. From these reactions, information concerning the effect of fluorine substitution on the magnitude of the carbon-chlorine bond dissociation energy has been obtained, as well as some knowledge of the chemistry of the fluoroalkyl radicals and the magnitude of some bond dissociation energies for fluorocarhons. A hrief report of an early value of the heat of formation of CFI radical obtained in t8hiswork already has appeared.' Experimental Materials.-Methyl chloride was an Eastman Kodak white label product purified by repeated bulb to bulb distillation. Monofluoro-, difluoro- and trifluoromethyl chlorides were obtained from the Kinetic Chemicals Division of the E. I. du Pont de Nemours Co. They were distilled in a low temperature fractionating column and the middle fractions were used. Nitrogen was purified by passage through a metal ketyl solution6 and thon through several cold t r p s , including one containing silica gel. Sodium was cut from a large block and was refluxed under vacuum for about 10 hours a t 500" to drive off volatde irnpurities and was handled undcr dry nitrogen. (1) Presented before the Physical and Inorganis Division of the Amerioan Chemical Hociety at Loa Angelefi, Maroh, 1963. (2) Abstracted in part from B thesis nuhmitted by John F. Reed in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the University of Wsshington, 1953. (3) J. F. Reed and 8. S. Rabinovitah, THIEJOURNAL, 69, 201 (1965).
(4) R. 5. Rabinovitoh and J. F. Reed, J . Chem. Phya., 8 8 , 2092 (1954). (6) L. F.Fieser, "Experirnente in Organic Chemistry," D. C. Heath and Co., New Tork, 2nd Ed., 1941, p. 390.
Apparatus.-A flow system modeled after those of Po1an ie and of Heller' was employed. The reaction vessel, a Jyrex tube 6 cm. in diameter with plane windows, was heated by a furnace provided with plane windows for observation of the flame. Nitrogen carrier gas streamed over sodium a t 529'K. in the carburetor, and through a nozzle of radius 1.00 mm. into the reaction zone. The flow rate of nitrogen, which was maintained by a three stage mercury vapor pump, was determined from the pressure fall across a capil ary, measured by meane of a double McLeod gage. Halide was fed to the reaction zone through a calibrated capillary tlrbe. The reaction zone was illuminated with a sodium lamp monitored and controlled at constant irradiation intensity. Volatile products were collected in cold traps, and salts were collected on a plate under the flame. Reaction system pressure was measured by a McLeod gage. Temperatures wcre measured with calibrated thermocouples. The inside of the reactor was blackened by a soot deposit to providc B uniform dark barkground for flame observations. Procedure.-Temperatirres, nitrogall reRsurc in the systein and carrier gas flow rate were brou&t to constant conditions before the halide was introduced. The fluorescencc zone was reduced to the steady-state aim in a few minutes. The flame radius waR measured by sighting at right angles to the flow axis along a thin steel plate mounted on a lucito t,ravelling arm of a steel rule. The lucite allowed observation of the entire flame or, when covered, only the flame edgc so that the intenso flame center could be obscured. Several readings of flame size together with relevant pressure and temperature readings were made and constituted the measurements of a sin le run. After these measurements, halide flow was closed o f and new conditions were established. There was no chemiluminescence under dark conditions. Product6 were isolated in runs of several hours duration. Sodium Pressure.-The sodium pressure in the carburetor was obtained from the deta of Gordon.8 The limiting visible pressure, PI, was determined by raising the temperature of the carburetor and reaction zone (the latter always 20" highcr than the former), and notin the temperature@when sodium fluorescence was first visibje both within and 'ust outside the nozzle t,ube, while nitrogen was circulated at rates and pressures comparable with a run. The appropriate value to be taken is discussed below. Halide Pressure .-The halide artial pressure was calculated from the nitrogen and halite flow rates and the total reaction pressure. Diffusion Coefficient of Sodium.-The ternary diffusion coefficient of sodium into the mixture of halide and nitrogen (6) H. W. Hartel and M. Polanyi. 2. phyaik. Chem.. B l l , 97 (1030) (7) W. IIrller. Trans. Faraday Soc., SS, 1566 (lQ37). ( 8 ) A . R.Gordon, J . Chem. Phys., 4,100 (1930).
c
L
Tirr., 8 0 1 ) I l I M
M R . ~1957 ,
I)IFFITRTON FLAMF;
WRR cnlculntcd using Wilke's equation,g in which thr hinary diflusion cocficients of Rotlium into nitrogen and into hnlitle are rcqriired d o n g with the mole fractions of the gases. Tho binary codkient in nitrogen is in the litcrature,lOt" but thnt of the halide was cnlculatcd from kinetic theorgla using molecular diclrnetcrs of halides cdculnted from viscosity data.13 Analysis of Products.-Products could \IO grouped as volatile and non-volatile. The former included roducts of reactions of the radicals, the latter included halde salts and carbon. Volatile products wcre analyzed using a Consolidated Model 21-103 mass spectrometer. The mws spectra of the original reactants, of the total products, and, to assist in identification, ?f high and low boiling fractions of the products were obtained. The halide salts were sodium chloride and sodium fluoride. Samples from various portions of the collection plate were annlyacd for the chloride to fluoride ratio. The chloride was determined by a semi-micro Mohr method and the fluoride by a spectrophotometric m ~ t h o d . 1 ~Both methods were calibrated against samples of known composition.
Results Volatile Reaction Products.-Although not all the mass spectral patterns of the products were known, a sufficient number of those of the simpler compounds were available from the API tables to enable deduction of the nature of the products, but not always of their relative amounts. The principal results are listed in Table I. Only trace quantities of higher molecular weight compounds were found. The product of the rcaction of CFH2Cl was C2FH3and no C2F2H4was detected. The latter would arise from radical recombination of CFHz after C1 abstraction by sodium. However, it has been pointcd out that this ethane is very unstable toward HF elimination which occurs even a t O o . I 6 Likewise in the products of the reaction of CF2HC1, both C2FdH2 and C2F3Hwere detected, the latter presumably also arising by HF split from the somewhat more stable tetmfluoroethane. TABLE I
METHODFOR
FAST
REACTIONS
590
in Table 11. The ratio a t any point on the colloction plate fell within 10% of the average. TABLE I1 AVERAQENaF/NaCl DEPOSIT RATIOS,c Reactant
CFaCl
CBEHCI
CFHaCl
C
0.75
0.20
0.05
Calculation of Specific Rate Constants.-Four methods of calculating the specific rate constant were used. The first three models have been discussed previously3 and only their definition and calculation are described here. The fourth model is introduced below. k(").-This is the conventional Polanyi diffusion model rate constant, which employs the boundary condition that the pressure of sodium at the nozzle is the carburetor pressure, p".
where the observables on the right-hand side have their customary meaning. k(').-When the nozzle boundary condition is modified by setting the flow rate of sodium into the system equal t o the total amount of reaction per second, W) is obtained.
where vo is the linear streaming velocity of the gas in the nozzle and Cz = k(')p'/D. Since k(l) occurs in C, it was evaluated by successive approximation. ic(2).-This constant utilizes the same boundary condition as k(1) but with a different model. An average streaming velocity is inserted into the equation of continuity and the solution contains a factor exponential in the coordinate along the streaming axis, with a resultant asymmetric flame shape dependent on the streaming velocity.
VOLATILEREACTION PRODUCTS Products
CF&l CIi'zHCI CFHICl
CaFa and CzF4 (1,OO:O 74); CdFe (trace) C2F4€12,CaFzHzand CzF,H CZFHS, CaH4 (small)
The products of the reaction of sodium with methyl chloride were not analyzed. Non-volatile Reaction Products.-In general the molar ratio of fluoride to chloride salt found decreased slightly with increasing distance from the nozzle. Significant amounts of a dark carbon deposit were evident only with CFaCl. Average values over the entire reaction zone of the molar ratios, for the various reactants studied, are given (9) C. R. Wilke, Chem. E w . Proe., 46, 0 5 (1050). (10) H. V. Hartel, M. Polanyi and N. Me'er, Z.phymk. Chem., 819, 138 (1932).
( 1 1 ) R. J. Cvctanovic and D. J. LeRoy, J . Chsm. Phffa.,20, 343 (1853). (12) E. H. Kennard, "Kinetic Thcory of Gases," McGraw-Hill Book Co.. Inc., New York, N. Y..1938, p, 104. (13) J o h n F. Rced, Thesis, Univcrsity of Wadiington, 1053, available o n microfilm from University Microfilms, Ann Arbor, Michigan. (14) D. Morrier, Vaucher and P. Wagnci, HeEu. Cham. Acta, 81, 828 (1848). (15) A . 1 4 . IIenne and T. Midgley, J . Am. Chem. Soc., 68, 882 (1936).
R.
where Z is the displacement of the flame center along the flow axis, a! = v/2D, R2 = P2 Z2 and r 0 = rovop*/[4D(1 pro a2r02/2)];P i s the flame height; p2 = a2 k ( 2 ) p t / D : u is the average streaming velocity, obtained by averaging an assumed velocity distribution (initially vo and decreasing as an inverse square function of the distance from the flow axis), weighted by the sodium prcssure over the flame. Typical values of v were of the order of O.lvo, while Z was of the order of 0.2 t801.0 cm. Since p contains W ) ,the latter was obtained by successive approximation. I~(~).-Thisconstant was evaluated with the same lower boundary condition as for V I ) , but utilizes a newly applied boundary condition a t the flame edge. Observation of the limiting visible fluorescence pressure of sodium at the flame edge actually represents integration of the intensity over the line of sight by the eye. The edge of the flame is characterized by a value of the integral of the sodium concentration over the line of slght, X, normal to the radius a t the flame edge. The value of this limiting quantity of sodium was determined by observing the first appearance of sodium fluores-
+ + +
+
600
JOHN F. R m n SPECIFIC
Compound No.of runs
CIInCl 14 T,OK. 586 Over-all rate constants k(O) 1.49 (0.08) k(l) 0.55( .05) k(2) 0.50( .04) k(8) 1.26( .05)
AND
l3. S. RARINOVITCH
TARLE TI1 RATEC O N S T A N T 8 FOR RNACTION WIT11 CI~II2CI ci~~rci
SODITIM'
CFaCI
CF,Cll
586
11 586
G 525
1.24 (0.14) 0.52( .06) 0.40( .05) 1.19( .OS)
3.82(0.29) 1.85( .OS) 1.70( .07) 3.46( .15)
1.01(0.17) 0.17( .02) 0.14( .02) 0.07( .OB)
11
11
588 1 13 (0.17) 0.44( .OO) 0.41( .OG) 0.98( .11) I
Primary step rate constants k.(O) 1.49 (0.08) 1.08(0.16) 1.03(0.11) k.(l) 0.55( .05) 0.42( -06) 0 . 4 3 ( .05) ks(*) 0.60( .04) 0.39( .OG) 0.41( .04) 1.26( ,051 0.93( .lo) 0.99( .07) Unite are 10" cm.*/mole sec.; standard deviation of average in parentheses.
cence, under the same irradiation intensity used in rate studies, both within and outside the nozzle tube. Within the tube, the concentration is constant along the line of sight bounded by the tube walls. Outside the tube, the sodium concentration varies in space and this variation was calculated on the basis of a radial diffusion and streaming model, the latter contribution being negligible. The results for a number of measurements within and outeide the tube showed good agreement and the value of 4.7 X mole/cm.* was used in rate calculations. The corresponding values of p1 were 3.2 to 6.5 X mm., depending on reaction vessel age, while the integrated intensit,y showed much less variation. Using this boundary condition, Ida) was determined from the relation
Vol. 61
2.18(0.17) 1.06( -04) 0.97( .04) 1.98( ,091
radical products which also deplete ,sodium in the flame. Since reaction of sodium with these radicals is bimolecular (Discussion), the over-all rate of depletion of sodium at any point in the flame, in the case of CFaC1, is given by
+
~ I I ~ ~ C F &kbppCF, I
+ keppcF, + 6dppOP
If the ratio of formation of sodium chloride to 8pdium fluoride in the flame is c, then the rate of dtsappearance of sodium due to all reactions is k,(l 4C ) ~ ~ C F ~ and C I , the over-all rate constant is equal to (1 c)k,. It will be recalled that the average value of c agreed with the value at any point of the flame within 10%. The value of k, is obtained by dividing the over-all rate constanb in Table 111by (1 c) for the particular chloride. It is to be noted that irrespective of the model employed for calculation, the order of reactivity among the various halides is the same (the minor where C2 = kca)p'/D. A family of plots were discrepancy between CFHzCl and CFsHC1 for the made of the above integral versus C for various Polanyi model actually falls within the limit of exvalues of the parameter, R, the flame radius which, perimental error). reactivity of CF&l ie with the reaction vessel size, determines S. From greater than that of The CHaCl, while CFHlCl and the appropriate value of the integral to fit equation CFdICl react more slowly. 4, C was obtained. Activation Energies.-From the specific rate conFor all halides the correction to the rate constant stants at 586°K. of the primary step, hafa),activafor halide depletion in the flame was found to be tion energies computed assuming a unit steric low, of the order of 2 to 5%) using the diffusion factor and a were value of 0% = 3.5 x 10-16 cmS2,and model'6 treatment of Smith. '7 Instead, the Hellera are given in Table V. The difference between the correction modified for nozzle sizea was employed, activation energies recorded here and those rein as much as this takes into account the effect of ported earlier4 is due to the fact that the latter streaming on the halide impoverishment in the values were based on ka(l). The approximate acflame. tivation energy computed from the temperature The results are summarized in Table 111. The of the rate of CF&1 between 525 and data are too numerous to tabulate." The maxi- dependence 586"K., assuming the same product distribution at mum range of experimental variables employed in any run is as follows: halide pressure in the both temperatures, was 12.7 kcal./mole. flame from 0.08 t o 0.4 mm.; vo/D from 4.9 to 24 Discussion cm.-'; total reaction pressure from 2.5 to 5.5 Evaluation of Experimental Models.-It is seen mm.; flame radius from 1.49 to 1.86 cm. Typical from Table I11 that in all cases W0)> > k(') > data bearing on a discussion of the experimental IC(2), as expected.8 The values of the first two models is given in Table IV for CF3C1. constants are roughly the same as are the last two, Correction of Rate Constants for Radical Re- the first pair being about twice the value of the secactivities.-In the past, corrections have not been ond pair. made for the further reaction of halogen-containing The spcoific ratc constrtnts,-Vo) and W ) ,are re(16) R. J. Cvctanovio and D. J. LeRoy, Can. J . Cham., 2 9 , 597 lated as (1951). (17) F.T. Srnit,h,.I. Chem. P h y s . , Z Z , 1005 (1054); cf. It. J. Cvctano-@!e) ' ' 1.-2 In- 210to/4n . (/c(0))l/z = ( \ c ( l ) ) l / * + 6; (y vio, Can. J . Cham., ac, 54 (IDGO).
+
*
+
(R
- ro)
(5)
a
c
c
60 1
May, 1057
8'.
cm.-l
vo/D.
dynes om.-'
8.00 8.03 13.8 14.5 14.0 10.2 113.3 16.9 17.8 18.4 24.2
158 129 156 273 143 128 145 336 2134 352 245
k (1)
185 178
3.01 3.19 5.31 2.53 4.11 3.23 4.35 2.70 2.82 2.85 3.31
I08 213 138 176 131 199 193 190 1G7
1.51 1.54 1.40 I .49 1.52 1.8F 1 .08 1.70 1.57 1.41 1.48
TABLE V ACTIVATION ENEnQIES OF PRIMARY REACTIONS AND
SOME
PHYSICAL PARAMETERS FOR FLUOROMETHYL CHLORIDES E, kcaI./ Camporand
mole
2;
eQe
(20DK.)~o.o
5.20 5.75 3.12 3.80 4.46 3.62 3.18 3.02 3.14 3.10 3.00
1.66 1.90 1.44 1.88 2.20 1.91 1.09 1.65 I .77 1.81. 2.43
3.49 4.10 3.02 3.33 3.97 3.68 3 . 10 2.90 3.00 2.94 3.91
1.60 1.80 1.31
1.75 2.00 1.71 1.60 1.40 1.64 1.57 2.14
1
3.5
0
sQq (gas)
1
CHIC1 9.8 1.78* 68.4 75.13d CFIIiCl 10.1 1 . 7 6 " 0 ~ 67.6 70.40" CFaHCl 10.0 1.73d 70.5 CFaCl 9.2/ 1.75* 77.6 78.05° W. Gordv. J. W . Simmons and A. G . Smith. PhU8. Rev.. 74, 243 (1948). TA. S. Bartell and L. 0.nrockwav, J : N. Miillor, J . Am. Cham. Chem. Phyls., 23, 1860 (1955). SOC.,75, 860 (1983). I,. 0. Brockway, THISJOURNAL, 41, 747 (1937). e D . IC. Coles and R. €1. Hughes, Phys. Rea., 76, 858 (1949). J. W. Hodgins and R. L. Haincs, Can. J . Chem., 30,473 (1952), report 6.2 kcal.
where 6 arises only from the difference in lower boundary conditions. The experimental range of 6 was greatest (fourfold) for CFaCl at 5Sti°K., and a plot of versus 6 is shown for CF3C1in Fig. 1. If k(O) is a physically correct quantity, then it should show only random variation with 6, hiit if k ( ' )is more precise, tho points should cluster about a line of slope one. It can be seen that the data fit the variation expected on the basis of the model associated with Id1). The slope of the least squares line calculated from the data for each of the halides is .0.9 for CFaCl, 0.8 for CF2HC1, 1.4 for CFHzCl and 0.3 for CH3CI. Although the calculated slopes vary somewhat due t o experimental error, the conclusion may be drawn that there is a positive slope in all cases which is of the order of magnitude predicted from the boundary condition of W . The smaller standard deviation of the average of as opposed t o that of Id0) further emphasizes the better fit of the data t o kc*). The improvement in the standard deviation of Id2)over that of k(') is small enough so that it is felt that the introduction of an average streaming velocity into the equation of continuity has not materially altered the results, even though the model more adequately describcs the flame contour and does provide some improvement. This is expected since the asymmetry of the flame was not pronounced, the ratio of the flame height t o center displacement from the nozzle being of the order of 2 t o 5 in all cases. It may be concluded that within this range of flame asymmetry, there is little need for a velocity model. The effect on halide distribution is, of coursc, an associatcd prollcm.
1 0
I.o
0.5
8 (cm?
dyne-' sec!
.
Fig. 1.-Relation between I c ( O ) l / * and 6; solid line is theoretical fit with slope unity, broken line is least squares fit, for reaction of CFsCI.
Adoption of the new upper boundary condition yields improved precision in the rate constant, W , and also readjusts the magnitude of the rate constant to a greater value. The fortuitous approach toward the value of Ido) stems from the partially compensating errors in both boundary conditions of the model for the latter quantity. The exact exwill depend on tent of the variation of k(O)from experimontal conditions. The values of Ida) are not only the most precise, but are here considered to be conceptually the most nearly correct derived from the models examined. The crtrlier comparison of k(O) and k(') above is still valid, however, since it may be shown that use of the incorrect upper boundary condition in both modcls tends to be mutually compensating. Range of vo/D.-Ttie range of v n / D employed cxtends Ltcyoiid that recomniended by Heller.7
602
JOHNF. R.mn AND B. S. RABINOVITCH
Justification for extending the lower limit has been discussed previously and a correction for the sodium depletion, 8, due to back diffusion of halide, derivcds
Vol. 61
Reactions (b), (c) and (d) are evidenced by the NaF and carbon found in the products. The formation of CF2 by reaction (b) leads by recombination (f) to CZFd, found in the products in significant amount. Any reactions of CF other than (d) were not detected or recognizod. It is now well known that recombination of CF8 radicals (e) occurs at these pressures without any concomitant disproportionation.20 Abstraction of F by CFs is not favored under these conditions.21 Also, pertinent to reactions of the less fluorinated chloride, hydrogen abstraction by fluorinated radicals is not expected, as known specific rates of hydrogen abstraction by both methyl and trifluoromethyl radicals22 are much slower than processes occurring in the flames studied here. This mechanism adequately accounts for all the products of the reaction. Trace amounts of higher compounds, such as C,FS in the reaction of CFaCI, could arise from polymerization of unsaturated products. The extent of each of the reactions in the above scheme, 21, was calculated on the basis of one mole of sodium consumed and is given in Table VI. For the reaction of CFlHCl with sodium, since the extent of fluorine abstraction compared to chlorine abstraction is low (0.20), and since the ratio of reactions (c) to (b) are low for CF8Cl,it was arbitrarily assumed that the extent of reaction (c), stripping of the last fluorine of the radical by sodium, may be taken to be zero; sufficient information was not available to determine the ratio of reaction (c) to (b). Thus the entry in Table VI for this reaction implies an assumed value. A dash in the table indicates that no reaction exists. TABLE VI IZXTBNT OF REACTION, MOLESOF PRODUCT/MOLE OF SODIUMCONSUMED
mhcre PO' is thc pressure of halide at the nozzle and D' is the diffukon coefficient of halide into the gas mixture. Calculations of 8 at the lowest vO/Dvalues employed were made for the various halides and found to be negligible, e.g., for CI13Cl at vo/D = 4.9 crn.-', the lowest value used, the sodium depletion was lees than 1%. From the data for CF3C1 at 525"I, 7 Evans model.2@ First,, the polarizability, a, of the X l O I 3 ~ m . ~ / m o sec. l e Treatment of this quantity various methyl radicals are not the same but inas before yields an upper limit t o the activation crease with increasing fluorine substitution. This effect is greatcst for CF3 relative t o CHa. There is energy of 3.5 kcal./mole. By application of the Schwarz inequality in a no reliable value of a for CFs, but it is calculated similar manner t o the other radical reactions with that, this factor would decrease the activation ensodium, estimates of the lower limits of the rate ergy for CFaCl by an amount of roughly 1.4 ked/ constants per C-F bond may be obtained. The re- mole from that for CHaCl. sults are given in Table VI1 in terms of the radical A second complicating factor is the fact that the recombination rate constant. This latter quantity equilibrium internilclear C-C1 bond distances of is known only for CFa. A good assumption is that these compounds differ. This causes a shift of the all the recombination constants of the fluorine sub- iriit8ialand final state energy curves on a cut along stituted methyl radicals are approximately 2 X the C--Cl bond distance axis through the transition 10lacm.8/mole sec., in line with the values for CFI statc of t,ho potential energy surface representing and CHa.2a,zs It miist be concluded that these the reaction. A decrease in the C-C1 equilibrium radical reactions with sodium occur with small ac- internuclear separation result,s in an increase in activation energy. Furthermore, it appears that tivation energy, for this factor alone. This effect CFa is more reactive than CFZHor CPHZper C-I’ would raise tho act,ivation energy for CJ?,HCl by 2 bond. kcal./mole ovw that of CHaC1, and by 1.2 kca1.l mole for CFaCI. TABLM VI1 A third and smaller effect is the decrease in the LOWER LIMITSO F S P E C I F I C RATE CONSTANTS PER C-F observed activation energy resulting from resonance I h N D I N CM.’/MOLIII 8 E C . FOR R A n r c A L REACTION WITH in tho transition state, which increases with the SODIUM (586°K.) number of fluorine atoms in the molecule. The Rarltmagnitude of this factor was roughly estimated from ral CF: CFIII CF’III CPa the treatment which has been given by Warhurst,19 (7 X kb (2 5 X (1 8 x (5 x 10’akc)’/2 10”ke)’/2 10%)t/s 10’2k,)’/Z in conjunction with some of our preliminary results on the reacttion of sodium with CF3Rr and CH3Br, Activation Energies and D(C- Cl).-The dif- the bond dissociation energies of which are knownz9 rerences in observed activation ciicrgies for the to differ by only 3 kcal. The greatest decrease in primary reactions of sodium atoms can be related activation energy to be expected is for CF3C1 and t o differences in bond dissociation energies of the here this cffect was estiinat,cd a,t 0.5 kcnl./mole. C-CI bond in these molecules. Consideration of (20) (a) R. A. Ong and M. I’olanyi, l’rans. Ipnmday SOC.,31, 1376 the factors affecting an estimate of these differ- (1035): (b) M. G. Evans and E. Warhurst, zbid., 36, 693 (1930); ences is made most easily on the basis of the model (c) howcver, cotnilare also R. P. Smith and 11. EyrinR, J . Am. Chem. stant, ratio. To this approximation, $)CF3 may be taken constant throughout the flame at an average value, POF3, whose magnitude may be obtained from the extlcnt of the recombination reaction (e)
(23) P. €3. Ayscoiigh, J . Chsm. Phys., 14. 944 (1050). (24) €1. Margenau and G . h l . Murphy, “The Mathematics of Physics and Chemistry,” D. Vnn Nofitrnnd Co.,
N. Y., 1943, p. 130.
Inc.,
New York,
(26) A. Slirpp, J . Clem. P ~ T M8.4,, 939 (1050); E. W. R. Steacie, “Atomic end Frcc Rndtnal Reactions,” Vol. 11, 2nd Ed., Reinhold Publ. Corp., New York, N. Y.,1964.
Soc., T4, 229 (1962); R. P. Smith, Am. Chem. 500. Meeting, Buffdo, 1052. (27) E. T. Hutlcr and Polanyi, Trana. Faraday SOC.,39, 19 (1943); A . G. Evans and IT. Walker, %bid.,40, 384 (1944). (28) C. P. Atortimer, H. 0. Pritchard and A. H. Skinner, ibtd., 48, 220 (1082). (29) A. 11. Schon and AT. Szwarc, Proc. R o y . Soe. (London). AaOS 110 (1051).
M.
JOHN F. R,EEl) ANI)
604
n. s. RAnINOVITCH
These various factors operat,c as a whole in the direction of canccllation, and together with the small sprcacl i l l cxperimcntal I!: vnhias Icads to thc conclusion that there is littlo variation of I)(C-Cl) in tho series studied, and that based on D(CIIa-CI) = 81.2 kcal./mole, a reasonable estimate is D(CF~C1) = 80-81 kcal./mole, while D(CFJI--Cl) and D(CFH2-Cl) = 81-82 kcal./mole. Our data are too sparse to justify any det,niled attempt to correlate the bond dissociation energies measured here with the various factors and theories relating to valence binding. It may be remarked that in addition to the qualitative agreement of the observed activation energies, and probable relative order of the dissociation energies, with the variation of the parameter TC-cl in this series, a similar correspondence may be found with the quadrupole coupling constant for C186in these molecules (Table V), of which the relative gas phase values are the more unequivocal. The explanation of this relation may involve the decreasing importance of
€I structures such as F- t C l + with increasing
I
H
fluorine substitution. The variation of eqQ is such, in any case, as to Bhow that inductive factors alone do not determine its magnitude (see reference 2Gc). However, the nature of this binding, particularly in mixed polyhalogenated methanes, is notoriously complex. It is possible to give a qualitatively consistent explanation of the relative magnitudes of carbonhalogen bond dissociation energies of mixed polyhalogenated methanes, based on a competition between two competing factors along the lines first pointed out by Scanlan and Warhurstao: namely, that the greater is n, in a molecule of the type, CH,-.X,, the greater the ionic resonance energy and stabilization, the effect of which is also to reduce the M-X bond length; on the other hand, increasing substitution of H by X or another atom, Y,increases the non-bonded repulsion between the X atom and the other substituents of the central C atom, the more so, the larger the van der Waals radius of Y.sl Such an explanation has been applied by Szwarc to the analogous bromides.2g Some Derived Thermochemical Quantities for Fluorocarbons.-In an earlier note4 we gave the following thermochemical quantities, AHt(CF3) = -120.5, D(CF3-F) = 116.5 and D(CF3-CFa) = 62, all in kcal./molc at 25". From D(CF,-Cl) taken from this study as 80.5 kcal./mole, the above quantities, when recalculated on the same basis rn before, become -119.5, 117.5 and 64 kcal./mole, respectively. Although some question has been raised recentlyR2regarding the possible inexactitude of D(F,) = 38 kcal./mole, this value has been retained in the above calculations. Very recently, Lossing, et U Z . , * ~ from electron impact studies have (30) J. Scnnlan and E. Warhurst, Trana. Faraday Soc., 46, 1000 (1949). (31) D. F. Heath and J. W. Linnett, J . Chem. Phys.. ill, 147 (1950). (32) K.L. Wrag and D. F. Hornig, ;bid., 24, 1271 (1958). (33) J. B. Farmer, I. H. 9. Hcnderaon, F. P. Lossin& and D. 0. H. Marsdcn, ibid., 24, 348 (19513.
Vol. 61
reported AHr(CFS) = - 117 kcal./mole, whil Pritchard, et U Z , , ~ have ~ estimated this quantity a - 119 kcal./mole, both quantities being in satis factory agreement with our values. Correspond' ence for D(CF3-F) and D(CFrCFa) follows also. Some limits may also be set for other related quantities. From the activation energy for the reaction of sodium with CFa, which haa a maximum value of 3.5kcal./mole, a.limit on the heat of formation of CFS may be obtained. Using the value of AHf(CF8) = -119.5 kcal., AHt(Na) = 26.0 kcaLlg6and AHt(NaF) -- -72 k ~ a l values . ~ ~ of AHf(CF2) -18 kcal./mole and D(F&-F) 120.5 kcal./mole are obtained. There is an indication that the second bond in CF4 is weaker than the first, since the data of sodium atom-alkyl halide reactions, in general, show that a reaction characterized by a low activation energy of honphRto will he used. (2) ‘A. Bsont-Qyorpyi, “Cheniintry of Miiscular Contraction,” 1 s t ed.. Acadeniic Press, New York, N. Y.. 1847. (8) A. Szont-Gy6rgyi, “Chemistry of Muscular Contraction,” 2nd ed., Acadornic Press, Now York, N. Y . , 1051. (4) H. H . Wcbor and €1. Portzchl, Pros. niophya. Phy8. Biochem., 4, GO (3054). ( 5 ) Y . Tonomura, J . Reaenrch Inat. /or Cnlalysis (Zfokknido U n i v . ) , 4, 87 (1956). (6) J. J. Rlum. Arch. Rdochem. Rioph,ys., 6S, 480 (1955). (7) W. F. H. M . Mommaerts, J . Den. I’hysiol.. 31, 361 (1948). (8) H.MatRumiya, ysMorita, N. Kitagawn, K. Yagi and Y. Tonomrira, J . Biochem. ( J n p n n l , in press. ( 9 ) Y. Tonorriirrn, 8. Wntnnabe and I
