809
PHOTOLYSIS OF PERFILJOROCYCLOBUTANE
(or x = -3 amu). The molecular orbital in eq 1 should then be !l!322-T2 = No{ds.2-T2 l / N u (0.18 9 4 8 - 0.032 *as) (-21 2,) - aI(s1 s2)}. The Nu given in Table I1 are not corrected for this small effect. The participation of the metal s orbitals may be considered a general phenomenon in the complexes with (das2-_r2)1configuration, where the fractional s character should in large part be determined by the relative
+
+
+
heights of the s and d orbitals and, consequently, may depend on the nature of the axial ligand. The explanation may be sought along this line for the observed large variation in 6 Q Chf~ parameters in Co-phthalocyanines.'S2 Acknowledgment. The authors are grateful to Mr. R. Shrager and Mrs, 13. McNeel for the least-square curve fitting.
Photolysis of Perfluorocyclobutane by John E. Davenport1 and Glenn H. Miller Department of Chemistry, University of California, Santa Barbara, California $8106
(Received June $4, 1968)
Perfluorocyclobutanemixed with small amounts of xenon was photolyzed using 1470-A radiation froin a xenon resonance lamp. No evidence was found for XeFz formation. Reaction products included CF4, CzF4, CtFe, C3Fs,C3Fs,1-C4Fs,2-C4Fs,C4Fl0,and higher fluorocarbons to polytetrafluoroethylene. trans-Z-CaFg was the main product. A reaction mechanism is proposed to account for the experimental observations.
The 1470-A xenon-sensitized reaction of CFd was studied by Dacey and Hodgins2in 1950. Fluorine mas shown to be present in the reaction products and a quantum yield of one eliminated chain processes as possible steps leading to the assumed product polytetrafluoroethylene. The only other reported photolysis of a perfluoroalkane was the preliminary note by Miller and DaceyS on the photolysis of perfluorocyclobutane. This work indicated the presence of XeF2 and a dimer as important reaction products. The mercury-sensitized photolyses of tetrafluoroe t h ~ l e n e , perfluoropropene,B ~,~ and cis- and trans-2perfluorobutene' have been reported as well as the thermal decomposition of perfluorocyclobutane8 and tetrafluor~ethylene.~~'~ The purpose of the present investigation was to confirm the presence or absence of XeF2 as a reaction product or intermediate in the xenon-photosensitized photolysis of perfluorocyclobutane and to elucidate the reaction mechanism,
In the first arrangement, the reaction system consisted of a closed loop containing the reaction cell, lamp, an induction-operated circulating pump (Teflon bearings), a Wallace-Tiernan pressure gauge (Model FA-141) , U-trap, and a KBr-windowed 10-cm infrared cell. The latter was positioned in the beam of a Beckman IR-10 spectrometer. Product samples were condensed and transferred for mass spectrographic analysis. I n the second arrangement, the infrared cell was replaced with a vpc sampling valve (Teflon stopcocks) by means of which a sample could be isolated and passed directly from the reaction loop into the column of the chromatography apparatus. The 90-P3 Aerograph gas chromatograph was equipped with a Sargent SR
Experimental Section
(1) Taken in part from a dissertation presented to the Graduate Division, UCSB, in partial fulfillment of the requirement for the M.A. degree. (2) J. R. Dacey and J. W. Hodgins, Can. J . Res., 28B, 173 (1950). (3) G. H. Miller and J. R. Dacey, J . Phys. Chem., 69, 1434 (1965). (4) B. Atkinson, J . Chem. SOC.,2684 (1952). (5) J. Heicklen, V. Knight, and S. A. Greene, J . Chem. Phys., 42,
Equipment. The vacuum system was of conventional design and utilized an oil diffusion pump, ionization gauge, and Teflon stopcocks with %ton 0 rings. Two arrangements were used for the reaction system; both, however, utilized the xenon resonance lamp and reaction chamber previously described.li
221 (1965). (6) J. Heicklen and V. Knight, J . Phys. Chem., 69, 3600 (1965). (7) D. Saunders and J. Heicklen, {bid., 69, 3205 (1965). (8) B. F. Gray and H. 0. Pritchard, J . Chem. SOC.,1002 (1956). (9) B. Atkinson and V. A. Atkinson, ibid., 2086 (1957). (10) B. Atkinon and A. B. Trenwith, ibid., 2082 (1953). (11) G. H. Miller and J. R. Dacey, Rev. Sci. Instrum., 36, 1041 (1965). Volume Y3,Number 4
April 1060
JOHN E. DAVENPORT AND GLENNH. MILLER
810 recorder, disk integrator, and fraction collector to remove samples for mass spectrographic analysis. The total volume of the reaction loop was 285 ml. The ambient temperature was 23", the temperature of the reaction chamber about 150", and the surface of the lamp was estimated to be approximately 240" ; thus during the photolyses the cell had a temperature gradient from 23" on the outer mall to 240" on the lamp window. Several chromatography columns were used; all were constructed from 1/4-in. copper tubing. Column A, 20 ft, consisted of CHz=CHC02CH2(CFzCFz),"2 on Chromosorb W. It was prepared as previously described13and operated at 0". Column B was a 1-m, 40-60 mesh silica gel column operated at 200". Column C was a tandem arrangement consisting of column A plus a 1-m squalane on silica gel section. Helium at 60 cc/min was used as the carrier gas.
Perfluorocyclobutane from Columbia Organic Chemicals Inc. was purified by bulb-to-bulb distillation. The ms and vpc analyses showed no detectable impurities. Airco ultrapure xenon was used without further purification.
Experimental Results and Techniques Table I contains a summary of the reported experiments. Experiments 2, 3, and 4 were made with the ir cell as part of the reaction system loop. The system Torr, the gases were adwas evacuated to less than mitted from their storage bulbs, the lamp was turned on, and after thermal equilibrium was established the total pressure was followed as a function of time. Typical total pressure-time curves are shown in Figure 1. The
30
' L:
Initial pressures.a Expt -Torr--Xe c-CaFs no.
2 3 4
5 6 7 8 11 12 13 14 15 16 17 18 19
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 a
0.32 trace 12.42 6.90 15.10 3.56 0.18 0.22 0.13 1.2 3.01 3.03 3.01 3.01 3.01 3.29 3.29 3.34 3.29 3.29 3.31 3.31 3,34 3.34 3.35 3.34 3.34 3.34 3.01 0.78 0.80 0.51 0.51 0.52 0.51 0.53
31.28 10.51 18.78 28.61 11.73 31.32 35.07 17.62 25.57 28.6 22.43 22.61 22.7 22.60 22.66 26.72 30.01 22.72 26.75 26.70 26.93 26.99 27,21 27.21 27.24 27.24 27.23 27.23 27.00 29.26 29.26 18.78 18.79 18.92 18.79 19.39
Re-
action time, min 5600 1040 1501 76 20 90 97 50 1987 246 5 10 20 33 35 20 60 32 50 62 40 585
10 20 95 125 20 303 7194 1000 2840 15 31 45 60 100
h 20 -
Final pressures: --Torr-Xe
...
0.51 0.51 9 52 0.51 0.53
.f:
I-
Analysis
c-CbFs
Total 10.09 Total 9.35 Total 24.73 Total 33.98 Total 26.80 Total 34.80 Total 35.20 Total 18.20 Total 17.25 Total 27.30 3.0 21 3.0 20.8 3.0 18.2 3 . 1 18.9 3.2 22.2 3 . 1 25.6 3.3 24.1 3 . 3 22.6 3.29 26.1 3,34 26.0 . . . 25.9 3.3 17.4 3.32 26.9 3.34 26.0 3.36 25,58 3.35 25.0 3.33 26.8 3.35 23.2 Total 16.20 Total 21.4 18.70 18.74 18.37 18.51 18.37
Ir Ir Ir Vpc column A Vpc column A Vpc column A Vpc column A Vpc column A Vpc column A Vpc column A Vpc column B Vpc column B VpccolumnB VpccolumnB Vpc column B Vpc column B Vpc column B Vpc column C Vpc column C Vpc column C Vpc column C Vpc column C Vpc column c Vpc column C Vpo column C Vpc column C Vpc column C Vpc column C Vpc column C Vpc column C
Vpc column C Vpc column C Vpc column C Vpc column C Vpc column C
Partial pressures were determined from vpc data.
The Journal of Physical Chemistry
25-
6 :
Table I : Summary of Experiments
15
-
t t,
i I
,
*
"
'
I
I'
1
"
'
" "
'1, 2
'
I
' '
" 2 ' '
'
1
' ' I
3 Time, min. x 10'
'
"
1,
4
"
I
'
" "
l
~
,
'
j I
~
5
Figure 1. Total pressure of reaction system as a function of time for experiments 2, 33, and 34.
lamp had to be turned off and the reaction interrupted to obtain ir spectra; all attempts to shield the spectrograph from microwave interference were unsuccessful. No ir absorption peaks due to XeFz were observed. The peaks previously believed to be due to XeFz were shown to originate from various fluorocarbons in the reaction product mixture. For all the remaining experiments the ir cell was replaced by the vpc sampling valve arrangement. Experiments 5-13 were made to determine the composition of the reaction product mixture. Following a run the total product was collected in the liquid nitrogen cooled trap of the valve system and after warming to room temperature it was passed directly into the vpc column. The pressure due to noncondensables never exceeded 0.5 Torr. A total of 47 vpc peaks were observed for experiment 12 over a total elution time of 920 min. Xenon, C ~ F O , CzF4, CsF8, C3Fs, c-C4F8(8.7 min), C4F10,t~uns-2-CeF8 (14.8 min), and CsFlz's were identified. Additional major product peaks (relative areas greater than 6000, compared to t ~ u n s - 2 - C ~at F ~30,800) were observed at elution times of 34.0, 35.1, 37.8, 338, and 685 min. (12) Courtesy of the E. I. Dupont de Nemour and Co. (13) S. A. Greene and F. M. Wachi, Anal. Chem., 35, 928 (1963).
PHOTOLYSIS OF PERFLUOROCYCLOBUTANE Other product peaks with relative areas greater than 2000 were eluted after 43, 104, 115, 137, 380, 470, and ,500 min. For experiments 13-40 the Xe and c-C4F8 were premixed and stored in a glass storage bulb. The vpc column was calibrated with known amounts of Xe and +CAFEand the partial pressures of these gases were determined quantitatively before and after the reaction. Within experimental error no loss of xenon was observed. Photolyses were made for various lengths of time in order to obtain plots of c-CeF8disappearance as a function of time; several of these are shown in Figure 2. Attempts were made to obtain the quantum yield for c-CIF8 disappearance by using COz actinometry. The method, however, is subject to many restrictions and the results can only be considered approximate. The values obtained varied from 0.75 to 1.5. Exposure of c-C4Fsto a helium discharge in the same apparatus produced no pressure change. Thus, thermal decomposition of c-C4F8 did not occur on the lamp window. Calibration of the vpc column with available fluorocarbons gave the following relative partial pressures for the listed reaction products: truns-z-C4F8 > C2Fe > C3F8 > C4Flo > Cd'4 N l-C4F8 N C3F6 > individual higher molecular mass products. Some free carbon was found and polytetrafluoroethylene was identified by ir analysis. CF4 production mas followed by observing an ir absorption peak. The concentration of CFd increased rapidly during a photolysis to a relatlively high partial
811
2500
2000
Wave number, Cm-'.
Figure 3. Infrared spectra of CF4. A, 10.4-Torr standard CFI sample; B, Xe and c-CdFs a t start of reactmion;C, reaction mixture after 189 min; D, after 1202 min; E, after 1925 min.
pressure and then decreased slowly. These data are shown in Figure 3. Separate photolysis experiments showed that truns2-CdF8 is photolyzed directly by 1470-A radiation. These results will be presented in a separate communication.
Discussion On the basis of the above experimental results it is possible to propose a tentative mechanism for the cC4F8 photolysis. The initial step is undoubtedly Xe" c-C4F8 Xe c-C4Fs" (1) followed by C - C ~ F+ ~ " (--CF&F&F&F2-) *
L
2 25
-
+
f I
t?
n
20
E --+
0
100
50 Time, min.
Figure 2. Total and c-C4Fs pressures for experiments 27-31 (top) and experiments 36-40 (bottom).
+
CFsCF=CFCF3 CF2
+ CFsCF===CF2 (2)
2C2F4
4CF2 Step a is favored. truns-2-CrF8is the major product. Also, there is no overall pressure decrease during the initial stages of the reaction, Figure 2, but actually a slight pressure increase due to fragmentation reactions which precede polymerization. During this time, Volume 79, Number 4
April 1968
JOHNE. DAVENPORT AND GLENNH. MILLER
812 c - C ~ is F ~being converted to 2-CdFs and the overall pressure decrease occurs only when the latter is photolyzed and polymerization reactions become important. Photolysis of the 2-C4Fsproduces CF3CFspecies
-
2-C4F8 -+ (2-CdFe) * These react as 2CF;ICF
2CFaCF
(4)
C4Fs
-
CFaCF *C2F.j wall
CFaCF
CF4
(3)
(5)
+ carbon
(6)
Dalby14 in his study of CF2 reactions assumed that no CF&F radicals rearranged to C2F4; however, Heicklen and Knighte have proposed this reaction for some energetic state of CF3CF. Step 6 accounts for the formation of CF4 and free carbon. The decrease in CF4 partial pressure during photolysis can be explained by assuming this reaction to be dependent on available wall space; large areas are available a t the beginning of the reaction, but as other products are formed and adsorbed the available space decreases. CF4 also disappears by photolysis2 CFd
+ Xe*
- + CF2
Fz
+ Xe
(7)
Polymer is formed by polymerization of CzF4*species nC2F4*
(C2F4) f~
C'he?nialr$~
CFaC (CFa)=CFz might occur CFaCF=CFCFa
CFs
+ CF&F=CF
-+ CFa
CFaCFECF
(9)
This would be followed by 2CFa
CzFa
It is necessary to assume the presence of either free fluorine, CFS radicals, or both to account for the products. Further speculation concerning the reaction mechanism must await additional experimental evidence. Acknowledgment. We wish to thank the National Science Foundation (Grant GP-4346) for support of this work.
(8)
and not by combination of CF2 radicals since the pre-
The Journal of Ph&aZ
ferred reaction is the recombination of CF2 radicals to form C2F4. The formation of the major products C2F6, CaFs, and other saturated fluorocarbons is more difficult to explain. It is doubtful if sufficient free fluorine from reaction 7 is present. A reaction analogous to the formation of CFa radicals from perfluoroisobutene proposed by Atkinsons
(14) F.W. Dalby, J. Chern. Phus., 41, 2297 (1964).
