Dissociation Dynamics of Perhaloalkoxy Radicals - American

more gratifying than the direct comparison is the agreement in many trends of the geometry when the molecule goes from the gaseous to the solid state...
0 downloads 0 Views 772KB Size
5660

J . Am. Chem. S o t . 1989, 11I , 5660-5661

the available diffraction data alone. Second, the agreement between the theoretical and the favored experimental geometry is excellent. T h e difference is less than one standard deviation for all parameters. Comparison of the calculated and experimental ra0 geometry of the solid state reveals a n rms deviation of 0.015 8, in the bond lengths with a maximum discrepancy of 0.025 A a t C N and a n rms deviation of 1.0' in the valence angles with a maximum discrepancy of 13' at C C N . The agreement is very satisfactory, particularly if one keeps in mind that the uncertainties in the employed K and b values are on the order of 0.01 A. But even more gratifying than the direct comparison is the agreement in many trends of the geometry when the molecule goes from the gaseous to the solid state. First, the experimental increase of the C = O bond length as well as the experimental decrease of the C-N length are matched in direction and magnitude by the calculations. The difference A = gas phase value - solid phase value amounts to A(expt1) = -0.030 8, for the C=O, to be compared to A(calc) = -0.022 A. For the C-N bond the corres onding values are A(expt1) = +0.043 A and A(calc) = +0.021 One may predict that N-H bonds, which take part in hydrogen bonding, will be longer in the solid state than in the gas, whereas C-H bonds, not engaged in hydrogen bonding, are foreseen to be relatively stable. Indeed, these expectations are met in the experimental as well as in the calculated geometry. Second, the valence angles for which the comparison can be made all show the correct trend: for CCO, A(expt1) = +1.9' vs A(calc) = $0.8'; for C C N A(expt1) = -1.4' vs A(calc) = -0.5'; for NCO, A(expt1) = -0.3' vs A(calc) = -0.2O. Third, the experimentally observed rotation of the methyl group is paralleled in our calculations. The difference between the experimental and calculated torsion angle H(3)-C( 1)-C( 2 ) = 0 is only 6.6'. This result strongly indicates that the rotation is indeed a consequence of electrostatic lattice forces, as discussed by Caillet, Claveric, and Pullman.22 Interestingly, short inter-

1.

molecular interactions are absent; all contact distances to the methyl group are larger than or equal to van der Waals distances (e.g., CH,.-CH,, 3.930 A; CH,-O, 3.596 A; CH,.-N, 3.614 A). This is consistent with the large rms of oscillation (15') observed for the methyl group by neutron diffraction. The absence of short intermolecular interactions may have played a role in the successful application of the crystal field approach, which emphasizes electrostatic effects and neglects covalent (short range) interactions with neighboring molecules. The calculations also reproduce some subtle details noted in the experimental geometry: the C-H(3) bond, normal to the molecular plane, is longer than the two other C-H bonds, the valence angle C(2)-C( 1)-H(3) is smaller than the other C C H angles, while the angle H ( 1)-C( 1)-H(2) is larger than the H C H angles involving H ( 3 ) . Although perhaps not statistically significant, but gratifying nevertheless, the experimentally observed pyramidization of the amide group is also reproduced in the calculated model, as follows from the torsion angles H(4)-N-C(2)=0 and H(4)-N-C(2)-C(l).

Acknowledgment. C.V.A. acknowledges support as a Research Fellow by the Belgian National Science Foundation, N F W O . This work was partly supported by N A T O , Research Grant N o . 0409/88. The text presents in part results of the Belgian Program on Interuniversity Attraction Poles initiated by the Belgian State-Prime Minister's Office-Science Policy Programming. T h e scientific responsibility, however, remains with the authors. Registry No. Acetamide, 60-35-5.

Supplementary Material Available: Charges on atoms before and after the optimalization of the solid-state model (1 page). Ordering information is given on any current masthead page. (22) Caillet, J.; Claverie, P.; Pullman, B. Theor. Chim. Acta 1978, 47, 17-26.

Dissociation Dynamics of Perhaloalkoxy Radicals Zhuangjie Li and Joseph S. Francisco* Contribution from the Department of Chemistry, Wayne State University, Detroit, Michigan 48202. Receiued December I , 1988

Abstract: The dissociation dynamics of CCI,-,F,O radicals have been studied by using ab initio molecular orbital theory. The ab initio calculations suggest that chlorine atom elimination reactions from CCl,_,F,O are low activation barrier processes, which dominate over fluorine atom dissociation processes. Calculated dissociation rate constants for both C1 and F elimination reactions predict a lifetime that is less than lO-''s for the CCI,..,F,O radicals. Replacement of two chlorine atoms by two fluorine atoms is found to stabilize the CCI3-,F,O radicals. The atmospheric implication of these calculations is discussed.

The photochemical dissociation of chlorofluoronethane, CX,Y ( X = CI or F, and Y = CI), yields a halogen atom and C X 3 fragment, via CX,Y hu cx, Y (1)

and

Oxidation of the CX, fragment to C X 2 0 with the release of a halogen atom has been suggested to involve the participation of CX,O radicals'-, via

T h e CCI,_,F,O species produced as a result of these oxidation steps are atmospherically important since they are responsible for the release of additional halogens from the initial halomethane and engender the production of C X 2 0 , which have recently been detected in the stratosphere from in situ measurement^.^^^

-

+

CCI,_,F, CC13,F,

+0

+ 0,

-

-

+

CCI,-,F,O

CCI,-,F,O

(2)

+ O2

+0 2

CCI,_,F, CC13-,FX02

+ NO

-

+

CC13-,FXO2

CCI,_,F,O

+ NO2

(4) (5)

(3)

( I ) Simonaitis, R . In Proceedings of the NATO Advanced Study Institute on Atmospheric Ozone; Report FAA-EE-80-20; Aikin, A, C., Ed.; Washington, D.C., 1980; pp 501-515.

0002-7863/89/1511-5660$01.50/0

~~

~

(2) Francisco, .I.S.; Williams, I. H. Inr. J . Chem. Kiner. 1988, 20, 455. (3) Francisco, J. S.; Williams, 1. H. In Proceedings of the 16th Annual Meeting of the National Organization of Black Chemists and Chemical Engineers, Chicago, IL, 1989.

0 1989 American Chemical Society

J . Am. Chem. SOC.,Vol. 11 I , No. 15, 1989 5661

Dissociation Dynamics of Perhaloalkoxy Radicals The kinetics of reactions 4 and 5 have been experimentally studied in detail:,' but evidence of the CCI,,F,O was not reported. Indeed, little is known about these radicals. The simplest perhaloalkoxy radical, C F 3 0 , has only recently been isolated and identified by infrared matrix studies,8 partly on the basis of our theoretical prediction^.^ The role CF,O radicals play in possible stratospheric reactions of catalytic importance has been examined previously'0 but not those for other CCl,-,F,O radicals. Furthermore, the role CCI,-,F,O radicals play is determined by the competition between the unimolecular dissociations given by CCI,-,F,O

-

+

CC12-,F,O

CCl,-,F,

+0

+ C1 (8 1

and bimolecular reactions with other atmospherically important species. There have been theoretical decomposition studies of halogenated alkoxy However, those studies employed the semiempirical M N D O method, which did not adequately describe the activation energy barrier for fluorine atom extrusion.]' Here we report a n a b initio calculation of the dissociation of CCI,-,F,O radicals and supplement our study with R R K M calculations to determine rates of dissociation. W e comment on atmospheric implications of the R R K M rates.

Computational Methods A b initio molecular orbital calculations were carried out by using the GAUSSIAN86 program.I3 All equilibrium geometries and transition structures were fully optimized by using a n analytical gradient method a t both the 3-21GI4 and 6-3 lG*I5 level. Using Schle el's method,I6 the geometries were optimized to less than 0.001 for distances and 0 . 1 O for angles; both maximum force and rms force were less than 5 X 1O4 au for the optimized structures. T o find the transition structure for these reactions, several additional structures along the reaction path were first optimized a t the UHF/3-21G level by fixing the C-CI' or C-F' bond and minimizing the energy with respect to all other coordinates. The maximum in the energy profile yielded a suitable initial guess for the Schlegel algorithm. Electron correlations were performed to fourth order17*'*along with spin p r o j e c t i ~ n 'using ~ Merller-Plesset perturbation theory, including single, double, and quadruple excitation (PMP4SDQ, frozen core). Cartesian force constants were calculated analytically for the reactants and for the transition structures a t the UHF/3-21G level. Vibrational frequencies were computed by determining second derivatives of

i

(4) Rinsland, C. P.; Zonder, R.; Brown, L. R.; Farmer, C. B.; Park, J. H.; Norton, R. H.; Russel 111, J. M.; Ruper, 0. F. Geophys. Res. Lett. 1986, 11, 327. (5) Wilson, S. R.; Crutzen, P. J.; Schuster, G.; Griffith, D. W. T.; Helas, G.Nature 1988, 334, 689. (6) Caralp, F.; Lesclaux, R. Chem. Phys. Lett. 1983, 102, 56. (7) Ryan, K. R.; Plumb, 1. C. J . Phys. Chem. 1982, 86, 4678. (8) Andrews, L.; Hawkins, M.; Withnall, R. Inorg. Chem. 1985, 24, 4234. (9) Francisco, J. S.; Williams, I . H. Chem. Phys. Lett. 1984, 110, 240. (IO) Francisco, J. S.; Li, 2.; Williams, I. H. Chem. Phys. Lett. 1987, 140, 531. ( 1 1 ) Rayez, J. C.; Rayez, M. T.; Halvick, P.; Duguay, B.; Lesclaux, R.; Dannenberg, J. J. Chem. Phys. 1987, 116, 203. ( 1 2) Rayez, J. C.; Rayez, M. T.; Halvick, P.; Duguay, B.; Dannenberg, J. J. Chem. Phys. 1987, 118, 256. (13) Frisch, M. J.; Binkley, J. S.; DeFrees, D. J.; Raghavachari, K.; Schlegel, H. B.; Whiteside, R. A.; Fox, D. J.; Martin, R. L.; Fleuder, E. M.; Melius, C . F.; Kahn, L. R.; Stewart, J. J. P.; Bobrowicz, F. W.; Pople, J. A. GAUSSIAN 86.

(14) Binkley, J. S.; Pople, J. A.; Hehre, W. J. J . Am. Chem. SOC.1980, 102, 939. (IS) Gordon, M. S . ; Binkley, J. S.; Pietro, W. J.; Hehre, W. J. J . Am. Chem. SOC.1982, 104, 2797. (16) Schlegel, H. B. J. Comput. Chem. 1982, 3, 214. ( 1 7 ) Krishnan, R.; Pople, J. A . Int. J . Quantum Chem. 1980, 14, 91. (18) Krishnan, R.; Frish, M . J.; Pople, J. A. J . Chem. Phys. 1980, 72, 4244. (19) Schlegel, H . B. J . Chem. Phys. 1986, 84, 4530.

the energy with respect to the Cartesian nuclear coordinates and then transforming them to mass-weighted coordinates.20

Results and Discussion Geometries. Optimized geometries for all reactants and products are listed in Table I. Although the geometries for some of the reactants and products have been published previously, they are reproduced here to facilitate comparison with other structures. Comparisons with the available experimental structures indicate that the overall agreement is very good: h0.02 8, for bond lengths and *1.8' for angles a t the UHF/6-31G* level. Geometry optimizations of CF,O, CF2C10, C F C I 2 0 , and CC1,O radicals suggest the ground state possesses C, symmetry with the unpaired electron localized mainly on the oxygen atom in the symmetry plane p orbital. Preference for the C, symmetry may be due to the Jahn-Teller effect, which is well-known to occur in CH3021-23and C F , 0 . 9 The equilibrium geometries of the 'A' and 2A/' states on the C, potential energy surface have been examined. W e find that generally the lowest energy equilibrium geometry corresponds to the 'A' state with the exception of CCI'FO, where the 2A" state is lower than the 2A' state by ca. 5 kcal mol-' at the UMP4/6-31G* level. We note that dissociation of CC1,FO on the 2A'' surface result in electronically excited products that are of higher energy than the ground state. Coupling of the 'A" to the 'A' surfaces through a radiationless path would allow the 2A" state to dissociate CC1,FO on the 'A' surface. Consequently, all dissociation pathways considered in this study have been examined on the 'A' potential surface. The minimum energy pathway for chlorine atom elimination in CC130 mimics fluorine atom dissociation in CF,O" in that the extrusion process occurs in the symmetrical plane of the molecule containing the oxygen, carbon, and the leaving chlorine atom. The dissociating CCY bond in the transition state is predicted to be 2.121 8, (6-31G*) and makes a n angle of 93.4' (6-31G*) with the oxygen. For C C I F 2 0 and CCI2FO, there are two different channels through which the halogen atoms may be eliminated; either through fluorine or through chlorine atom elimination. Chlorine atom elimination from C C l F 2 0 and fluorine atom elimination from CC1,FO occur, as with CCI,O, in the symmetry plane of the molecule, thereby retaining the C, symmetry. However, fluorine atom elimination from C F 2 C I 0 and chlorine atom elimination from CC12F0 possess no plane of symmetry in the extrusion process, thereby resulting in a reduction of symmetry in the transition state. W e note that the loss of the symmetry plane in the dissociation process does not significantly alter geometrical parameters. For example, the predicted dissociating C-CI' bond length in CC1F20 (symmetrical transition state) is 2.114 8, (6-31G*) as compared with 2.1218, (6-31G*) for CC12F0 (unsymmetrical transition state). Similarly, the predicted C-F' bond length for the extruding fluorine in CC12F0 is 1.778 8, (6-31G*) as compared with 1.775 8, (6-31G*) for C C I F 2 0 . However, the angle made by the oxygen, carbon, and the dissociating halogen atom reveals an interesting trend. For fluorine atom dissociation from C C I F 2 0 and CC12F0, no significant change occurs. However, for chlorine atom elimination, the angle widens by 0.7" (6-31G*) in going from C C l F 2 0 to CCI2FO. This may result as a consequence of the repulsion between two chlorine atoms, which is larger than that between a fluorine atom and a leaving chlorine atom. The undissociating CCI and C F bonds in the transition states of CC1F20 and CC1,FO are quite similar to the product CCIFO. Furthermore, the substitution of one chlorine atom in CCI2FO by a fluorine atom does not result in significant changes in the geometries, but it does affect the activation energy barrier and enthalpy of dissociation for the extrusion of either chlorine or fluorine. (20) Schlegel, H. B.; Binkley, J. S.; Pople, J. A . J . Chem. Phys. 1984, 80, 1976. (21) Cowell, S. M.; Amos, R. D.; Handy, N. C. Chem. Phys. Lett. 1984, 109. 525. ~~

(22)~Yarkony, D. R.; Schaefer 111, H. F.; Rothenburg, S. J. Am. Chem. SOC.1974, 96, 656. (23) Saebo, S.; Radom, L.; Schaefer 111, H. F. J. Chem. Phys. 1983, 78, 845.

5662 J . Am. Chem. Sot., Vol. 11 I , No. IS, 1989

Li and Francisco

Table I. Optimized Geometries for Reactants, Products, and Transition States for CCI,_,F,O Radial Dissociation“

basis set basis set ~~~. 3-21G 6-31G* expt species 3-21G 6-31G* expt CO 1.169 1.157 1.170’ LFCF 110.0 109.7 CF 1.322 1.290 1.316 [F’***CClFO]* co 1.248 1.227 LOCF 125.8 125.9 126.2 cc1 1.810 1.720 CCIFO co 1.166 1.158 1.173c C F’ 1.789 1.775 CCI 1.814 1.720 1.725 CF 1.330 1.299 CF 1.327 1.300 1.334 LOCCl 121.1 122.2 LOCCl 125.9 125.7 127.5 LOCF’ 94.1 90.4 LOCF 125.5 124.1 123.7 LOCF 121.3 120.4 CCI20 co 1.164 1.159 1.179d LClCF’ 101.7 104.1 CCI 1.830 1.735 1.742 LClCF 110.6 111.4 LOCCl 124.9 123.4 124.1 LF‘CF 101.2 100.5 CF 1.301 1.329 1.318‘ CC12F0 (X2A’) co 1.387 1.359 c F3 LFCF 111.6 111.3 Ill.lf CF 1.344 1.320 CCIF2 CCI 1.726 1.832 cc1 1.844 1.760 CF 1.328 1.303 LOCF 108.0 105.0 LClCF 112.7 113.6 LOCCl 110.4 111.0 LFCF 112.0 110.6 LFCC! 109.4 109.4 CC1,F CF 1.309 1.333 LClCCl 110.8 109.2 CCI 1.719 1.812 [F’***CCI20]* co 1.264 1.241 LFCCl 112.9 113.2 C F’ 1.772 1.778 LClCCl 117.7 115.2 CCI 1.820 1.732 CCI, CL 1.714 1.793 LOCF‘ 96.1 90.8 LClCCl 117.1 116.3 116.0s LOCCl 119.9 119.6 CF,O (X2A’) co 1.353 1.382 LF‘CCI 101.2 102.6 C F’ 1.307 1.332 LClCCl 114.1 112.3 CF 1.307 1.332 [CI’.. .CCIFO]* co 1.291 1.250 LOCF’ 106.8 107.0 CF 1.334 1.304 LOCF 111.3 111.3 CCI‘ 2.121 2.128 LFCF 107.9 108.2 CCI 1.727 1.820 LFCF’ 109.5 109.7 LOCF 118.8 118.5 [ F’. *CF,O]* co 1.237 1.218 LOCCI’ 98.1 93.3 C F’ 1.762 1.791 117.2 LOCCl 119.8 CF 1.289 1.322 LFCCI‘ 102.8 104.2 LOCF‘ 89.6 91.7 LFCCl 110.1 110.9 LOCF 122.6 122.3 LCI‘CCI 105.9 107.6 LFCF 109.8 109.8 CCI,O (X2A’) co 1.377 1.350 LFCF’ 101.9 101.4 CC1‘ 1.851 1.768 CCIF20 (X2A’) co 1.344 1.369 cc1 1.839 1.768 CCI 1.755 1.856 LOCCI‘ 104.5 103.8 CF 1.314 1.336 LOCCl 111.2 110.6 LOCCl LClCCl 119.7 106.1 104.4 109.5 LOCF LCICCI’ i 10.2 111.0 111.2 112.6 LClCF [CY. * .CCI2O]* co 1.296 1.263 110.5 109.2 CC1‘ 2.132 LFCF 2.121 107.3 108.6 [CI’.. *CF,O]* 1.241 1.290 CCI co 1.827 1.740 CCI‘ 93.4 2.1 14 2.136 LOCCI’ 98.1 CF 117.5 1.292 LOCCl 117.1 1.326 113.2 LOCCI‘ LClCCl 111.6 92.6 96.6 LOCF LCICCI‘ 105.1 105.8 120.5 118.7 LCI‘CF 104.5 104.9 ”Bond lengths in angstroms, angles in degrees; refer to Figure 1 for atom labeling. *Carpenter, J. H. J . Mol. Spectrosc. 1974, 50, 182. COberhammer, H. J . Chem. Phys. 1980, 73, 4310. dMirri, A. M.; Guarneri, A,; Favero, P.f Zulian, G. Nuouo Cimento 1962, 25, 265. ‘Yamada, C.; Hirota, E. J . Chem. Phys. 1983, 78, 1703. ffessenden, R. W.; Schuler, R. H. J . Chem. Phys. 1965, 43, 2704. gHesse, C.; Leray, N.; Roncin, J. Mol. Phys. 1971, 22, 137. ~

species CF2O

~

Normal Modes. Tables 11, 111, and IV contain unscaled frequencies and intensities a t the UHF/3-21G level for all the reactants, products, and transition states for CC13-,F,0 dissociation. All the CCl,-,F,O radicals, in their equilibrium ground state, posses C,symmetry. Their vibrations span the representation 6a’ 3a”, and all are infrared and Raman active. Halogen atom dissociation occurring in the symmetry plane also spans the same vibrational representation. Two transition states for CC12F0 and CCIF,O halogen atom dissociation that take place with C1symmetry span the vibrational representation 9a. All vibrations in this representation are also infrared and Raman active. Calculated frequencies of the products from CCI,,F,O dissociation compared to experimental ones are overestimated by -10-15% owing to the use of the harmonic approximation, truncation of the basis set, and neglect of electron correlation in the UHF/3-21G frequency calculation. Nevertheless, calculated frequencies a t this level provide correct predictions of the mode ordering. Complemented with estimates of relative intensities, these data provide

+

useful spectroscopic guides to the spectroscopic observations of these species. Vibrational frequencies for the transition structures are all characterized by one imaginary frequency. O n e common characteristic of the vibrational frequencies of the transition states is the complex coupling between the modes; this causes some difficulties in defining mode descriptions. In such cases, only the major motions are described. The imaginary frequency for C-CI’ bond dissociation in CCI3O, CCl2FO, and C C I F 2 0 appears to range from 502i to 5761‘cm-I. The transition vectors consist mainly of the CCI’ stretching mode. T h e remaining frequencies of the transition structure, for the most part, lie between the frequencies of the reactants and products. For C-F’ bond extrusion processes, the imaginary frequency for C F 3 0 , C C I F 2 0 , and CClzFO transition states are 1059i,1° 877i, and 7781’cm-I, respectively. In this case, the transition vectors for CC1F20 and CCI2FO show considerable mixing of the C F mode. Interestingly, the substitution of CI with F, in going from CC130 to C F 3 0 , increases the C O

J . A m . Chem. SOC.,Vol. 111, No. 15, 1989 5663

Dissociation Dynamics of Perhaloalkoxy Radicals

Table 11. Calculated UHF/3-21G Vibrational Frequencies (cm-I) and Intensities (km mol-,) for Reactants, Products, and Transition State for CCI,O Dissociation freq int mode abs re1 sym descr calcd exptl molecule no. 364.5 0.64 2023 1827“ a1 CO str CClz0 1 2 b2 CCI str, asym 789 849 574.0 1 .oo 586 585 20.14 0.04 3 bl CCl, wag 505 570 18.66 0.03 4 a, CC1 str, sym 0.004 0.00 43 1 440 5 b2 CC1, rock 285 285 0.990 0.00 6 a, CCl, scissor 858 898’~‘ 154.4 1 .oo CCI, I e CC1 str, asym 154.4 1 .oo 858 2 e CC1 str, asym 1.513 0.01 a, CC1 str, sym 456 460d 3 0.982 0.01 314 4 a, CC1, umbr 1.454 0.01 256 5 e CC1, twist 1.454 0.01 256 240d 6 e CC1, wag 101.3 0.58 1075 CCI,O (X2A’) 1 a‘ CO str 176.1 1 .oo 808 2 a” CCI, rock 169.1 0.96 747 3 a‘ CCI str 7.730 0.04 489 4 a’ CC1 str, sym 4.937 0.03 371 5 a’ CC1, wag CC1, umbr 331 1.182 0.01 6 a’ CC1, rock 318 3.302 0.02 7 a” 8 a’ OCCI’ def/CCI, scissor 243 0.051 0.00 9 a’‘ CC1, twist 196 7.222 0.04 [CY. * .CCI,O] * 1 a’ CC1‘ str 502i 2.982 0.01 CO str 1167 216.9 0.82 2 a’ CC1 str 830 265.6 1.oo 3 a’ CC1 str, sym 515 24.70 0.09 4 a’ CC1, rock 388 0.507 0.00 5 a” OCCl rock 343 7.565 0.03 6 a’ 7 a” CCI3 umbr 298 1.579 0.0 1 CCI2 twist 213 0.144 0.00 8 a’ 9 a“ OCC1’ def/CClz scissr 183 0.113 0.00 JANAF Themrochemical Tables, 3rd ed.; Chase, Jr., M. W., Davies, C. A., Downey, Jr., J. R., Frurip, D. J., McDonald, R. A,, Syverud, A. N. J . Phys. Chem. ReJ Data, Suppl. 1985, 14. bHesse, C.; Leray, N.; Roncin, J. Mol. Phys. 1971, 22, 137. ‘Andrews L. J . Chem. Phys. 1968, 48, 972. dBenson, S. W. J . Chem. Phys. 1965, 43, 2044.

stretching mode for each chlorine replaced.

Energetics Heats of Reaction. The total energetics for the various equilibrum structures and transition states are listed in Table V. Calculated heats of reaction are provided in Table VI and compared with experimental results; and individual barrier heights are collected in Table VII. The heats of reaction (Table VI) have been calculated with extended (3-21G) and polarization (6-31G*) basis sets a t the Hartree-Fock level with electron correlation. The effect of d orbitals on the heats of reaction is small; but for oxygen atom extrusion, a 6-1 1 kcal mo1-I increase in the energy is predicted. However, electron correlation has a significant effect on the energetics of oxygen atom extrusion; a n increase of 41 kcal mol-’ is predicted in going from UHF/6-31G* to UMP2/6-31G* level of theory for CCI,O dissociation to CCI, 0. No significant differences are produced in the energetics for the transition states for halogen atom elimination as a consequence of electron correlation and spin projection. For fluorine atom elimination processes, electron correlation and spin projection increases the heat of reaction by 6 kcal mol-’, while heat of reaction for chlorine atom elimination processes decreases, except for CC1,O + CI. Experimental values for the heat of formation for CF302and CCI3O3are -1 56.7 f 1.5 and -2.8 f 5 kcal mol-I, respectively. The A H r ( C F 2 0 ) is -1 52.7 f 0.4 kcal mol-’, and that for CCI2O is -52.6 f 0.8 kcal mol-’. Experimental heats of reactions for CF,O CF20 F (9)

+

CCI,O

-

+ C C l 2 0 + CI

(10)

are 23.0 f 1.6 and -20.8 i 5.1 kcal mol-, a t 298 K. Assuming a 0.5 kcal mol-‘ change of heat of formation for both C F , O and CC130 going from 298 to 0 K, the heats of reaction AHo (0 K) become 23.1 f 2.0 kcal mol-’ for (9) and -20.8 i 5.6 kcal mol-’ for ( I O ) . Theoretical heats of reactions are in reasonable agreement with these experimental estimates.

On purely thermodynamic grounds, the energetically most favorable reaction for C C l F 2 0 and CC12F0 radicals is chlorine extrusion. Fluorine atom extrusion is the next thermodynamically favored reaction; while oxygen atom elimination is the most thermodynamically unfavorable channel of those considered in this study. There are interesting trends that emerge from these data. The heat of reaction for fluorine atom elimination shows a decreasing trend in going from C F 3 0 , CClF,O, and CC12F0 corresponding to 25.2, 21.8, and 13.2 kcal mol-’, respectively. These data suggest that increasing chlorination of C F 3 0 thermodynamically destabilizes the fluorine atom extrusion process. Activation Energies. Calculated activation energies for various transition state structures for CCl,-,F,O dissociation are given in Table VI1 and are calculated from the total energies given in Table V. Effects of basis sets, higher order electron correlation, and spin projection on the barrier show similar trends to those calculated for the heat of reactions, e g , changes of f 1 0 kcal mol-’. The only experimentally determined activation energy for the CCl,-,F,O system is trifluoromethoxy radical, CF,O. Kennedy and Levy24measured a value of 3 1.Of 0.5 kcal mol-’. Descamps and measured a value of 26 kcal mol-I, but this value is believed to be too low since a Lindemann extrapolation was used to obtain the rate constant; it is known that this procedure tends to underestimate activation energy. Nevertheless, the present theoretical estimate for the activation energy of 29.1 kcal mol-’ is in excellent agreement with that observed by Kennedy and Levy; this also suggests that calculations of the activation energies for CC13-,F,0 halogen atom extrusion processes a r e probably underestimated by ca. 2 kcal mo1-I at the PMP4/6-31G* level of theory. With chlorine substitution on C F 3 0 , there is a large systematic decrease in the activation barrier for fluorine elimination a t the (24) Kennedy, R. C.; Levy, J. B. J . Phys. Chem. 1972, 76, 3480 (25) Descamps, B.; Forst, W. Can. J . Chem. 1975, 53, 1442.

5664

J . Am. Chem. Soc., Vol. 1 1 1, No. 15. 1989

Li and Francisco

Table 111. Calculated UHF/3-21G Vibrational Frequencies (cm-I) and Intensities (km mol-’) for Reactants, Products, and Transition State for

CCI,FO Dissociation freq int sym descr calcd exptl abs re1 1 a’ CO str 2086 1868“ 406.9 0.97 C F str, asym a’ 1238 1095 421.6 1 .oo 3 CFO scissor a‘ 776 150.0 0.36 760 4 CFCl wag a“ 0.11 690 667 46.32 5 CC1 str a’ 501 2.746 0.01 48 1 6 CFCl scissor a’ 393 415 5.052 0.01 CC1,O 1 CO str 2023 1827“ 364.5 0.64 a, 2 CC1 str, asym 789 849 574.0 1 .oo b2 CC1, wag 3 586 585 20.14 0.04 bi 4 CC1 str, sym 570 18.66 0.03 505 a1 5 CC1, rock 440 0.004 0.00 43 1 bz CC1, scissor 6 285 285 0.990 0.00 a, CC12F 1 CF str a‘ 1303 1 143b,C 217.1 0.71 2 CC1 str, asym a’’ 1 .oo 850 919 304.5 3 CC1 str, sym a’ 0.1 1 574 747 33.78 4 a‘ 42 1 CC1,F umbr 5.957 0.02 5 CFCl rock a” 355 1 .os 1 0.00 6 CC1, scissor a’ 0.00 1.312 266 CCI,FO (X2A’) 1 a‘ 1289 CF str 145.6 0.26 2 1113 CO str a’ 229.4 0.41 3 CC1 str a” 1 .oo 553.7 864 4 OCF scissor a‘ 0.08 43.70 599 5 a” 437 CC1, rock 0.00 2.052 6 0.00 41 1 OCF def/CC12 scissor a’ 0.353 7 a‘ 400 0.02 9.733 CC1,O umbr 8 a” 342 0.00 1.016 OCF/CCl, twist 9 a’ CCI, wag 26 1 0.00 0.463 a‘ [F’. *CCI20]* 1 CF str 7781’ 0.28 94.83 2 a‘ CO str 1267 0.53 180.9 3 a” CC1 str 1 .oo 841 342.6 4 a‘ CC1 str, sym 0.06 519 22.06 5 a” CC1, rock 0.00 400 1.147 6 0.02 a’ 369 5.448 OCF rock 7 0.01 a’ 302 1.948 CC1,F umbr 8 0.00 a’ 225 0.183 CC1, wag 9 0.00 a’’ 209 0.040 CC1, twist [Cl’. * CCIFO] * 1 0.03 a 5351 10.41 CC1 str 2 CF str 1 .oo a 1396 302.2 3 CO str 0.96 a 1113 289.7 4 FCO rock 0.23 a 697 70.89 0.01 5 CCIO/FCO twist a 453 3.357 0.03 6 CClF twist a 385 10.38 0.02 4.678 7 CFO wag a 371 0.00 0.873 8 FCO twist a 257 0.00 0.228 a 9 FCCl rock 195 JANAF Thermochemical Tables, 3rd ed.; Chase, Jr., M. W., Davies, C. A,, Downey, Jr., J. R., Frurip, D. J., McDonald, R. A., Syverud, A. N. J . Phys. Chem. Refer. Data. Suppl. 1985, 14. *Milligan, D. E.; Jacox, M. E.; McAuley, J. H.; Smith, C. E. J . Mol. Spectrosc. 1973, 45, 377. ‘Prochaska, F. T.; Andrews, L. J . Chem. Phys. 1978, 68, 5568. molecule CCIFO

mode no.

;

-

+

PMP4/6-31G* AZPE level of theory. Substitution of one fluorine in C F 3 0 by one chlorine decreases the barrier by ca. 1 kcal mol-’; the second reduces the barrier by an additional 6.5 kcal mol-’. Neither basis set effects nor correlation energy corrections are large enough to alter the trend of decreasing activation energy with greater chlorine substitution. Figure 2 (top) emphasizes the trends in the energy profile along the reaction path for fluorine atom dissociation in CF,O, C C I F 2 0 , and CC12FO. An increase in fluorine substitution on CC1,O appears to increase the activation energy for chlorine atom elimination, except for CC12F0, which shows a negative barrier due to a n overcorrection by the spin projection method for a small barrier. Nevertheless, the calculation suggests that the addition of chlorine to CCIFO (i.e., the reverse of the chlorine atom extrusion process from CC1,FO) may be barrierless. However, with two fluorine substitutions, the activation energy for chlorine atom extrusion is generally larger than in both CC1,O and CC1,FO a t all levels of calculations. Figure 2 (bottom) emphasizes the trends in the energy profile along the reaction pathway for chlorine atom extrusion in CCI3O, CCI,FO, and CC1F20. Comparison of activation energies for chlorine and fluorine atom extrusion processes indicate that the former is essentially more

(a) Figure 1. Geometry model of CCl,-,F,O:

(b) (a) ground state; (b) transition

state. favorable than the latter, because of low activation barriers. Thus for reactions that generate CCI,O, CCI2FO, and CC1F20 exothermically, the excess energy would be sufficient to drive these radicals to dissociate via reaction 6. RRKM Rate Constants T o assess the relative importance of chlorine and fluorine atom substitution on the decomposition kinetics of CC13-,F,0 radicals, energy-dependent unimolecular dissociation rates are needed. Such rates are obtained from R R K M theory, in which the microca-

J . A m . Chem. Soc., Vol, 11 I , No. 15, 1989 5665

Dissociation Dynamics of Perhaloalkoxy Radicals

Table IV. Calculated U H F / 3 - 2 1 G Vibrational Frequencies (cm-I) and Intensities (km mol-') for Reactants, Products, and Transition State for CCIF20 Dissociation freq int mode molecule no. SYm descr calcd exptl abs re1 0.97 1868" 406.9 2086 CO str a' 1 .oo 1095 421.6 1238 C F str, asym a' 0.36 776 150.0 760 CFO scissor a' 0.1 1 667 46.32 690 CFCl wag a" 0.01 501 2.746 48 1 CCI str a' 0.01 415 5.052 393 CFCl scissor a' 1 .oo 1928" 450.2 2132 CO str CF2O a, 0.90 1294 407.4 1455 C F str, asym b2 0.08 965 35.17 1053 C F str, sym a1 0.18 774 83.17 818 CF, wag bl 0.06 626 24.92 674 CF, rock b2 0.02 584 9.646 598 CF2 scissor a1 0.69 1208bsc 235.6 1421 C F str, asym a" CFZCI 1 .oo 1148 343.3 1244 C F str, sym a' 0.43 761 148.6 719 CCI str a' 0.06 599 19.82 590 CF, rock a" 0.00 1.098 384 CF, wag a' 0.01 4.342 341 CF, twist a'' 276.3 0.90 1412 C F str, asym a" CF2CI0 (X2A') 1 .oo 1360 305.4 CO str a' 0.64 1008 195.0 C F str, sym a' 0.45 138.1 735 C F 2 0 umbr a' 16.6 1 0.05 574 CF, def a' 0.04 12.09 541 CF, rock af' 0.00 0.092 426 CCI str/CF, wag a' 0.01 2.832 329 CF, wag a' 0.00 1.193 183 CF, twist a'' 0.06 5761' 32.97 1 CCI' str a' [CI'. .CF,O]' 1 .oo 1459 548.5 2 CO str a' 0.54 1457 298.9 CCI str a" 3 0.03 916 16.23 4 CCI str, sym a' 0.03 583 15.03 5 CCI, rock a" 0.03 572 14.83 6 OCCl rock a' 419 6.857 0.01 7 CF2 wag a' 27 1 2.122 0.00 8 OCCl def/CF, scissor a' 0.640 0.00 9 234 CF2 twist a' 197.2 0.56 877i [F'**CCIFO]* 1 C F str a 0.72 254.5 1493 2 CO str a 351.7 1 I77 1 .oo CCI str 3 a 90.13 722 0.26 4 CFO scissor a 2.261 46 1 0.01 5 CClO rock a 10.25 6 410 0.03 CFO rock a 4.476 7 388 0.01 CF,CI umbr a 1.009 268 0.00 CFO twist 8 a 212 0.1 11 0.00 9 CF, twist a " JANAF Thermochemical Tables, 3rd ed.; Chase, Jr., M. W., Davies, C. A,, Downey, Jr., I . R., Frurip, D. J.; McDonald, R. A,; Syverud, A. N. J . Phys. Chem. Refer Data, Suppl. 1985, 14, in box. bMilligan, D. E.; Jacox, M. E.; McAuley, J. H.; Smith, C. E. J . Mol. Spertrosc. 1973, 45, 377. 'Prochaska, F. T.; Andrews, L. J . Chem. Phys. 1978, 68, 5577.

CCIFO

1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 7 8 9

Table V. Total Energies (Hartrees) for CC13_,F,0 Dissociation Reactions 6-31G*

3-21G

system reactions CF,O CF,O [F'. * *CF,O]' CF20 + F CF, + 0 CCIF2O CCIF2O [CI'. * *CF,O]' [F'.**CCIFO]' CF2O + C1 CCIFO + F CCIF, + 0 CCI2FO CCI2FO [ F'. * *CCI20]' [CI'. * *CCIFO]' CCl20 + F CClFO + CI CCI,F + 0 CCI,O CCl,O [CI'*.CCI20]' CCI20 + CI CCll+ 0

UHF

UHF

-408.78660 -408.72303 -408.74929 -408.70702 -767.14618 -767.1 3938 -767.08687 -767.18083 -767.11638 -767.07410 -1 125.5075 -1 125.4638 -1 125,5081 -1125.4930 -1125.5479 -1125.4517 -1483.8955 -1483.8891 -1483.9245 -1483.8426

-411.01325 -410.94246 -410.98027 -410.91512 -771.03745 -771.01 525 -770.97127 -771.06327 -771.01093 -770.94721 -1 131.0583 -1 131.0037 -1 131.0454 -1131.0445 -1131.0939 -1130.9854 -149 1.0944 -1491.0789 -1491.1275 -1491.0321

UMP2 -411.77437 -41 1.71312 -411.73685 -411.62167 -771.76679 -771.75344 -771.70291 -771.80024 -771.73305 -771.61757 -1 131.7532 -1 131.6956 -1 131.7454 -1131.7330 -1131.7982 -1131.6195 -1 491.7569 -1491.7407 -1491.7981 -1491,6292

UMP3 -411.77922 -41 1.71593 -41 1.73764 -411.63208 -771.78928 -771.76644 -771,71925 -771.80746 -771.74551 -771.64112 -1 131,7813 -1 131.7392 -1 131.7713 -1131.7569 -1131.8169 -1131.6558 -149 1,7952 -1491.7787 -1491.8283 -1491.6776

UMP4 -411.79395 -41 1.73467 -411.75418 -411.64765 -771.79606 -771.78273 -771.73623 -771.82260 -771.75917 -771.65427 -1 131.7918 -1 131.7397 -1 131.7857 -1131.7684 -1131.8304 -1131.6663 -1 491.8042 -1491.7909 -1491.8396 -1491.6856

PHF -411.01599 -410.96099 -410.98234 -410.91904 -771.04060 -771.03084 -770.99261 -771.06567 -771.01302 -770.95155 -1 13.0612 -1 131.0275 -1 131.0626 -1131.0466 -1131.0963 -1130.9904 -1 491,07979 -1491.0970 -1491.1300 -1491.0380

PMP2 -411.77589 -41 1.72820 -411.73603 -411.62378 -771.76871 -771.76616 -771.72087 -771.80134 -771.73401 -771.63256 -1 131.7549 -1 131.7 159 -1 131.7597 -1131.7339 -1131.7993 -1131.6224 -1 491.759 1 -1491.7559 -1491.7992 -1491.6328

PMP3 -411.78007 -41 1.72615 -411.73699 -411.63318 -771.78448 -771,77537 -771.731 86 -771.80789 -771.74647 -771.64244 -1 131.7823 -1 131.7398 -1 131.7816 -1131.7579 -1131.8174 -1131.6575 -1491.7967 -1491,7900 -1491,8287 -1491,6797

PMP4 -411.79480 -41 1.74489 -41 1.75232 -411.64875 -771.79725 -771.79166 -771.74884 -771.82304 -771.76013 -771.65561 -1 131.7929 -1 131.7546 -1 131.7960 -1131.7963 -1131.8309 -1131.6679 - 1 491.8056 -1491.8021 -1491.8322 -1491.6877

5666 J . Am. Chem. SOC.,Vol. 111, No. 15, 1989

Li and Francisco

$4-

CCIzO

F

j P. B

d

Io-

%& W:J%5 !PB&:lP

1

A

-I,-

LLz

-20-

ULU.:,~

"JB -25

,

,

I

1

,

>

'

'

'

R n a l h COOrdhah

Figure 2. Energy profile for chlorine and fluorine atom elimination processes from CCI,-,F,O

nonical unimolecular rate constant, K ( E ) ,of an isolated molecule possessing the total energy E is given by

K ( E ) = G(E - Eo)/hN(E)

radicals.

N ( E ) is the density of states for the reactant molecule, and h is Plank's constant. Harmonic-state counting by the Hase and Bunker RRKM program26was used to compute these quantities.

(1 1)

where G(E-E,) is the sum of states for the transition structure,

(26) Hase, W. L.; Bunker, D. L. Program 234, Quantum Chemistry, Program Exchange, Indiana University, Bloomington, 1973.

J . Am. Chem. SOC.,Vol. 111, No. 15, 1989 5661

Dissociation Dynamics of Perhaloalkoxy Radicals 14 7

0

0

0

A

A X

0 A

X

X

A X

0

Legend A cf20 + f

30

50

40

80

70

60

00

x

CClFO

+F

0

CCI2O

+F

100

E (kcal moi.')

Figure 3. Energy-dependent RRKM rate constants for fluorine atom elimination from CF30, CCIF20, and CCI2FO radicals.

X

A O X

11-IV) were employed. Plots of the resulting R R K M rate constants against total energy E are shown in Figures 3 and 4. With the total energy below 100 kcal mol-', the dissociation rate constants are larger for the CI atom extrusion processes than for the F atom extrusion processes. A CCI,,F,O radical with about 1.O kcal mol-' of energy in excess of Eo has a chemical lifetime of less than lo-" s. As shown in Figure 3, greater chlorine substitution on the radical can increase the rate for fluorine elimination, as a consequence of lowering the barrier for the extrusion. Furthermore, the substitution of one chlorine atom by a fluorine atom in CCI,O, as shown in Figure 4, increases the rate constants of the chlorine atom elimination reaction, but the substitution of two chlorine atoms by two fluorine atoms has the opposite effect, slowing the rate of release of chlorine atom. In this case the reduced density of available states for dissociation which results from higher frequencies modes in the transition state for CC12F0 contributes to the slower dissociation rate in addition to the increased barrier height. The implications a r e that on the carbon fragment there should be a t least two fluorine atoms to complement the chlorine atom; this would act to retard the release of the chlorine atom. Consequently, C C l F 2 0 may be sufficiently long lived enough to undergo bimolecular reaction. However, we do note that if the CCI,-,F,O radicals are produced by highly exothermic processes, unimolecular decomposition would dominate. Additionally, comparison of calculated unimolecular rates for C F and CCI bond fission processes (Figures 3 and 4) clearly indicates that the relative rates of chlorine are greater than fluorine eliminations, CI > F. Concluding Remarks

Legend A 0

A

CCl,O

x

cciro

+ CI + CI

0 CF.0 + CI 0

10

10

30

40

50

60

70

80

E (kcal m0l.l)

Figure 4. Energy-dependent RRKM rate constants for chlorine atom

elimination from CC130, CCI2FO, and CCIF20 radicals. The critical energy E, for halogen atom eliminations from CCI,,F,O was taken to be the best estimate of the dissociation AZPE/3-21G). T h e moments barrier (PMP4SDQ/6-31G* of inertia for both reactants and transition states were obtained from the UHF/6-31G* optimized geometries (see Table I) and the unscaled U H F / 3 - 2 1 G vibrational frequencies (see Tables

+

The dissociation dynamics of the CCI,-,F,O ( x = 0, 1, 2) radicals have been studied by using a b initio method with spin annihilation. F atom extrusion from CI-containing species of CCl,-,F,O radicals is unlikely. The chlorine atom elimination processes are predicted to be low activation energy processes, and a t low pressure, these processes will dominate over bimolecular reactions of CCI,-,F,O with other molecules. The replacement of CI by F affects the stebility of CCI,-,F,O species. These calculations suggest that the release of chlorine atom from the CCI,-,F,O could be retarded by the addition of two fluorines to the carbon fragment.

Acknowledgment. W e are grateful to Wayne State Universtiy Computer Center and Chemistry Department for provision of computing resources and to the National Science Foundation for a Presidential Young Investigator Award.