Radiation Chemistry-II

perfluorocyclobutane, including the effects of added N 2 0 and 0 2 on the reaction. Mixtures of ... To save time and cost of operating the x-ray machi...
0 downloads 0 Views 1MB Size
6

T h e X - R a d i o l y s i s o f Perfluorocyclobutane and M i x t u r e s o f Perfluorocyclobutane a n d

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

M e t h a n e i n the G a s Phase EDGAR HECKEL1 and ROBERT J. HANRAHAN University of Florida, Gainesville, Fla. 32601 The x-radiolysis of perfluorocyclobutane in the gas phase consists of a combination of fragmentation and telomerization processes. More than 10 products were noted with chain lengths fromC5toC14,as well asC2F4,C3F6, C3F8, and a white polymer. Even-carbon exceeded odd-carbon products by two-fold. The radiolysis was not affected by added N2O, but all heavierfluorocarbonproducts were eliminated by addedO2.With addedCH4no products having more than five carbon atoms were found. It is postu­ red that the radiolysis of perfluorocyclobutane proceeds along three major paths: direct decomposition, giving two molecules of tetrafluoroethylene; a process giving perfluorocyclobutyl radicals plus fluorine atoms; and a process giving C1 andC3species. The radicals then initiate a short chain polymerization of tetrafluoroethylene. perfluorocyclobutane differs from its hydrocarbon analog in many physical and chemical properties. It is more easily formed and more stable thermodynamically than cyclobutane (17). The thermal decom­ position of perfluorocyclobutane proceeds by a direct unimolecular proc­ ess to give two molecules of tetrafluoroethylene, although other processes contribute to some extent (2, 12). Although Doepker and Ausloos have recently described the radiation chemistry of cyclobutane (3), the perfluorocarbon has received only slight attention. Fallgatter and Hanrahan (5) qualitatively studied the γ-radiolysis of liquid F-cyclobutane [F- is A

1

Present address: Chemistry Department, East Carolina University, Greenville, N. C. 27834. 120 Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

6.

H E C K E L AND HANRAHAN

121

X-Radiolysis of Perfluorocyclobutane

used as an abbreviation for perfluoro] and found F-ethylene, F-cyclopropane, and C , C , and C perfluorocarbons, as well as fluoroform, which presumably arose from an impurity. Rajbenbach (19) published a brief account of the effect of fluorocarbons, including F-cyclobutane, on the hydrogen yield from hydrocarbons. The mass spectral cracking pattern of F-cyclobutane has also been published (14). 5

e

7

This paper describes a study of the gas-phase x-radiolysis of pure perfluorocyclobutane, including the effects of added N 0 and 0 on the reaction. Mixtures of methane and F-cyclobutane were also examined over a broad composition range. The latter experiments were intended to shed light on the radiolysis mechanism of pure perfluorocyclobutane as well as to provide information on the radiolytic behavior of gas-phase fluorocarbon-hydrocarbon mixtures. Related studies on liquid phase mixtures of cyclohexane and F-cyclohexane are currently i n progress i n this laboratory.

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

2

2

Experimental Sample Preparation and Irradiation. The F-cyclobutane used i n these studies was obtained from A i r Products and Chemicals, Inc. and contained only small amounts of impurities; it was repurified using preparative gas chromatography. The column was 2.5 meters long, 3/8 inches in diameter, packed with silica gel of 60-200 mesh ( W . H . C u r t i n ) , and was connected between the storage tank of F-cyclobutane and the vacuum line. After the column was evacuated to high vacuum over several hours, the F-cyclobutane was released cautiously from the tank. At the arrival of the first traces of gas, the pressure i n the vacuum line increased. The first portion of the repurified gas was discarded. The main fraction of F-cyclobutane d i d not contain any impurities which could be detected by gas chromatography. The methane used was Phillips Research Grade (99.68%) and contained ethane, ethylene, propane, nitrogen, and carbon dioxide. Repurification was simliar to the method described above. However, molecular sieve 13 X was used as the stationary phase. The purified methane contained only nitrogen and much less of the original ethane. Nickel-plated copper vessels were used in all radiolysis experiments. The vessels were made from copper tubing, 1.5 inches i n diameter. The bottom was a 1-mm. thick copper sheet which was silver soldered to the approximately 10-cm. high vessel. A Hoke brass valve with Teflon seat and phosphorus bronze bellows was attached to the top of the vessel by a 1/4-inch Swagelock fitting. Each vessel and all parts of the valves were nickel plated. The radiation source was a General Electric laboratory model M a x i tron 300 x-ray unit. The machine was operated at 300 kv. and 20 ma. in all experiments. The radiation vessel hung vertically, inverted so that its bottom was as close as possible to the window of the x-ray tube. Thus, the vessel partially shielded its valve from the x-ray beam. The bottom

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

122

RADIATION CHEMISTRY

II

of the vessel served also as a filter for the weak component in the x-radiation. A l l experiments were carried out at 30 ± 3°C. To save time and cost of operating the x-ray machine, we used a special technique to prepare larger amounts of reaction products for microcombustion analysis. A 1-liter round-bottomed flask was provided with two stainless steel electrodes opposite each other and with a FischerPorter Teflon plug needle valve. The vessel was filled with 10 to 50 torr of gas, and an electrical discharge between the electrodes produced good yields of the same kind of products obtained by radiolysis. The discharge was generated by a small Tesla coil of the type used to check the pressure in a vacuum line. To duplicate the chromatograph patterns characteristic of the x-ray work, it was necessary to use the lowest setting of the Tesla coil which would maintain the discharge. Sample Analysis. Qualitative and quantitative analyses of the radi­ olysis products of perfluorocyclobutane and mixtures of perfluorocyclo­ butane and methane were made using a combination of gas-liquid chro­ matography, mass spectrometry, and microcombustion analysis. These techniques, described in detail elsewhere (8), are based on the use of a specially constructed dual column, dual detector gas chromatograph with a vacuum input manifold. A l l products except hydrogen fluoride were analyzed in the "duplex" gas chromatograph. Hydrogen was separated from methane and traces of air on a 4.4-meter column of molecular sieve 5 Α.; detection was by thermal conductivity. A l l other products were analyzed using a temperature-programmed unit (2.1°C. per minute, 15 minutes after injection) which had a flame ionization detector; the column used was silica gel (60-200 mesh). A special approach was necessary to calibrate the flame ionization detector for the various radiolysis products (8) since standard samples of most higher fluorocarbons cannot be obtained. By means of a streamsplitting valve on the duplex gas chromatograph about half of the column effluent was diverted into a modified Miller-Winefordner (15) chromatographic-type microcombustion apparatus. Here all organic compounds are converted to C 0 (and H 0 , if hydrogen is present). The combustion products are analyzed with a thermistor detector. This unit was calibrated on an absolute basis in terms of gram-atoms of carbon per unit chart area (1 sq. cm. of chart = 1.75 Χ 10" mole of carbon). After the molecular formula of each product was found by mass spectrometric analysis, the number of moles of compound in each chromatograph peak could be determined, and the sensitivity of the flame ionization detector could be calculated. A relatively large product yield, produced by several hours of x-irradiation (or by about 20 minutes of Tesla coil discharge) was necessary to obtain sufficient signal on the microcombustion unit. Radi­ olysis times of 15 to 150 minutes were used in other experiments. Hydrogen fluoride was determined by a conventional analytical pro­ cedure. Using basically the qualitative test for fluoride ion as described by Feigel (7), 5 ml. of a solution containing zirconium chloride, alizarin sulfonate, and hydrochloric acid were introduced into the irradiation vessel through its valve. The valve was shut, and after shaking vigor­ ously, an aliquot was placed in a 1-cm. borosilicate glass cuvette to measure the absorbance of the solution at 5200 A . This method is sensi­ tive but not very reliable because the absorbance is time dependent. 2

2

7

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

6.

H E C K E L AND HANRAHAN

X-Radiolysis of Perfluorocyclobutane

123

Successful detection of H F as a radiolysis product apparently depends on using metal radiolysis vessels since no H F could be found i n an experiment using a glass vessel. Dosimetry. The dose rate from the x-ray source was measured by using the hydrogen yield ( G = 1.3) of the ethylene dosimeter ( 9 ) . Using a 104.5-cc. vessel filled with 50 torr ethylene, we found a dose rate of 4.80 Χ 1 0 e.v./gram-hour. H 2

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

20

Energy absorption in F-cyclobutane and in methane-F-cyclobutane mixtures was calculated relative to the ethylene result. Since the vessels were nickel-plated copper, the radiation effects were mainly caused by secondary electrons ejected from the vessel walls. Dose rates i n the gases were calculated using the Bragg-Gray principle and Bethe's formula for the electronic stopping power S . (Since the more complicated equa­ tion for electrons and the much simpler one for heavy charged particles give equivalent results for electrons in the range 50-200 Kev., the latter was actually used in the calculations.) Therefore, the relative stopping power per electron i n two different materials, designated £ , is given by e

P

£e = S / S el

where

S = \η(2Μ οψ/Τ) e

(1)

e2

- l n ( l - 0») - 0*

0

*(2)

In this expression β is the velocity of the electrons divided by the velocity of light, and / is the average excitation potential of the material. Equation 2 differs from the complete Bethe equation in that a combination of con­ stant factors multiplying the logarithmic term, which would cancel i n Equation 1, has been omitted. The Ζ of a compound can be calculated from values of its atomic constituents by the approximate formula: \n(T) =% [N Z \n(T )]/N Z i

i

i

i

i

i

(3)

where Ν is the number of atoms of a given kind per formula unit, and the summation is over the various kinds of elements present ( i ) ( 9 ) . Values of S are given for methane, ethylene, and F-cyclobutane i n Table I. Values of I were found to be 40.8 e.v. for methane, 51.6 e.v. for ethylene, and 102.4 e.v. for F-cyclobutane. e

For two different gases the ratio of the energy absorbed Ε (e.v./ gram) is given b y E

2

NS 2

e2

where N i and N are the number of electrons/gram for each gas, and S i and S are the respective values of stopping power per electron. (To the extent that the Bragg-Gray principle applies, the energy absorp­ tion on a per gram basis is independent of the pressure.) 2

e

e2

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

124

RADIATION CHEMISTRY

II

Table I shows that electron stopping powers vary moderately over the energy range appropriate for secondary electrons from a 300 kv. x-ray set (we use the range 50-200 k v . ) . However, the ratio of stopping powers for any pair of gases varies by only about 2 % for the energy range indicated. Accordingly, calculations were made using an average value of the ratio of S for each gas to S for the ethylene dosimeter. O n this basis the measured absorbed dose rate of 4.80 X 10 e.v./gram-hour in ethylene corresponded to an absorbed dose rate of 3.73 Χ 1 0 e.v./ gram-hour in F-cyclobutane and 5.40 X 10 e.v./gram-hour in methane. For mixtures of methane and F-cyclobutane it was assumed that each gas e

e

20

20

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

20

Table I.

Stopping Power per Electron as a Function of Secondary Electron Energy

Energy, Kev.

β

Methane

Ethylene

F-cyclobutane

50 100 200

0.4128 0.5483 0.6954

8.34 8.96 9.55

8.12 8.70 9.32

7.42 8.05 8.62

absorbed energy independently i n proportion to the mass of that gas present. The total pressure i n these experiments was held constant at 150 torr. Results Product Identification. Several experiments were performed at a relatively high total x-ray dose (ca. 1 Χ 1 0 e.v./gram, or about two hours of irradiation) to identify radiolysis products from pure F-cyclo­ butane. During the chromatographic analysis of each sample, five or six peaks were collected in U-traps and analyzed with a Bendix mass spec­ trometer. Owing to the small amount of material available, the samples were allowed to leak directly into the ion source via a needle valve; a conventional reservoir-gold leak system was not used. The assignments made for the molecular formulas of the various chromatograph peaks are listed in Table I. 21

Consideration of the peaks eluting before the parent F-cyclobutane, identification of Q>F and C F was straightforward since each gave a mass spectrum in good agreement with those tabulated by Majer (14). Perfluoromethane was not detected. Since its response i n the flame detector was 1/1000 that of other fluorocarbons, it may be present despite our failure to detect it. 4

3

8

The molecular formula of C F was definitely established from its mass spectrum, but there is some uncertainty about its structure. The 3

e

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

6.

H E C K E L AND HANRAHAN

125

X-Radiolysis of Perfluorocyclobutane

mass spectrum was not identical to that of either F-propylene (14) or F-cyclopropane ( 5 ) . The difficulty was probably caused by contamination of the sample with the parent F-cyclobutane, which eluted immediately afterwards. The compound is tentatively identified as F-cyclopropane, based on its elution position. The olefin F-cyclobutene, which would be derived from the parent compound by loss of F , definitely is not produced. A standard sample of F-cyclobutene eluted between the peaks identified as C - , F i and C o F i , where no radiolysis product appeared. Since the parent F-cyclobutane tailed badly on the silica gel column used, the presence of other possible C fluorocarbons such as n - C F or n - C F i could have been masked. 2

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

;

4

4

8

4

2

4

0

Identification of the higher products through C i is contingent upon the peculiarities of fluorocarbon mass spectra, which have been reviewed by Majer (14). One problem is the fact that saturated fluorocarbons generally do not give the parent ion. A n ion with one fluorine atom less than the parent usually does occur but is usually only 1 or 2 % of the base peak. As a result, molecular weight verification usually rests on interpreting a few small peaks at considerably higher masses than the main peaks of the spectrum. A doubt often exists as to whether these peaks might arise from impurities i n the sample or from machine background. O n the other hand, perfluoro-olefins usually do show a substantial parent ion peak, as well as other peaks near the parent mass. Since these features do not occur i n any of our higher radiolysis products, we conclude that none of them are olefins. Instead, the several C „ F „ products are believed to contain a C ring, and the compound C i F i must contain two rings. A l l products above C - , F i showed a strong m/e 100 peak in their spectrum, consistent with the presence of a cyclobutane ring (14). 0

2

4

0

8

2

For products listed as C i , C i , and C 2

3

i

4

it appeared that the column

failed to resolve compounds of the same chain length. The mass spectra indicated that each of these peaks contained a mixture of substances. Product Yields. Yields of all products from pure F-cyclobutane were measured as a function of the radiation dose. G values listed i n Table II are taken from the slopes of the resulting graphs. A l l products below C i 2

gave linear plots; the heavier products are probably also linear with dose, but experimental error was greater for these compounds. A t the higher temperatures and longer elution times necssary for the C i , C i , and C products, the peaks became broader, and the chromatograph background increased. In view of the fact that a white, Teflon-like polymer formed during radiolysis, higher products were probably formed but were not detected. The G-values given i n Table II correspond to a total consumption rate of F-cyclobutane of 3.0 molecules per 100 e.v. Owing to the 2

3

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

i

4

126

RADIATION CHEMISTRY

Table II.

Products and Their Yields from X-ray Irradiated F-cyclobutane at 15 0 torr Product

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

II

G Value 0.13 0.008 0.005 0.065 0.091 0.077 0.21 0.13 0.007 0.12 0.18 0.17 0.12

production of polymer, this must be taken as a lower limit of the actual radiation sensitivity of F-cyclobutane. Effect of Sample Pressure. The effect of F-cyclobutane pressure was investigated over 50-300 torr. No change in product G-values was found —i.e., total product yields increased in direct proportion to the number of grams of F-cyclobutane irradiated. This implies both that the reactions involved are not appreciably pressure dependent in the range studied and also that the Bragg-Gray conditions are adequately met over this range. (The occurrence of equal but opposite trends from these two causes, which would cancel, appears unlikely.) Effects of Added Scavengers. Several additives were used to obtain more information about the reaction mechanism. Adding 5 % nitrous oxide had no measurable effect on product yields. However, with 5 % added oxygen or ethylene, all products listed in Table II except C F were eliminated, and new but unidentified products were formed. These results strongly suggest that the final products from pure F-cyclobutane arise from a free radical sequence. Methane-F-cyclobutane Mixtures. Studies were performed on various mixtures of methane and F-cyclobutane with compositions between 3 and 97% methane. Products were also measured for pure methane. Results of these experiments are shown in Figures 1 and 2. Products from pure methane—namely hydrogen, ethane, ethylene, and propane—were found at all intermediate concentrations, although yields were depressed below "ideal mixture" lines. The radiolytic behavior of the F-cyclobutane, however, was drastically modified. A l l perfluorocarbon products beyond the parent F-cyclobutane were completely eliminated, even with 3 % 2

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

4

6.

H E C K E L AND H A N R A H A N

added methane. The fragmentation p r o d u c t s — C F , C were enhanced in yield, maximizing between 30 and 50% H F was found i n substantial amounts. The only other measured was tentatively identified as C H — C F = C F ; mized at about 40% methane. 2

3

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

127

X-Radiolysis of Perfluorocyclobutane

0

4

3

2

0.5

F , and C F — added methane. "cross-product" its yield maxi8

3

0

1

MOLE FRACTION OF METHANE Figure 1. G values of the main products formed in the radiolusis of c-CJîg-CH]^ mixtures. • HF; Ο H ; Ο C H X C H ; Φ C F g

2

R

S

8

2

4

Other products which might be expected were probably masked by the F-cyclobutane peak. In particular, c - C F H may be masked i n this way since it would be anticipated as a product from work on mixtures of cyclohexane and F-cyclohexane (6). One product peak was i n fact ob­ served on the tail of F-cyclobutane, but it was never separated sufficiently to analyze. Since this product appeared to maximize at high methane content, it may be c - C F C H which also would be expected, at least i n small yields (6). 4

4

7

7

3

Discussion Pure F-cyclobutane. The general aspects of the radiolysis results are consistent with other studies of perfluorocarbon systems ( J , 4, 5, 10, 11, 13, 22). Common features include considerable fragmentation of both C — C and C — F bonds, yielding diverse products, low yields for individual products, but a moderate over-all decomposition yield, absence

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

128

RADIATION CHEMISTRY

II

of an olefin product derived from the parent by loss of F , and a mecha­ nism involving free radical precursors for most products. Several observations suggest that some type of chain polymerization is involved in the radiolysis of F-cyclobutane, especially since a white polymeric material was observed on the walls of radiolysis vessels and since measured G values do not decline appreciably with increasing chain length. The experiments with added methane suggest that the primary yield of tetrafluoroethylene may be appreciably greater than the net yield of 0.13 which is observed from pure F-cyclobutane. Furthermore, there appears to be a genetic relationship between several of the radioly­ sis products, based on the addition of successive C F groupings. Accord­ ingly we suggest a branching primary process followed by a sequence of competing free radical steps i n which radicals may add to F-ethylene or react with one another. Because of the results obtained with N 0 , 0 , and C H scavengers, it is assumed that all secondary processes are of the free radical type. In the following reaction sequence, F within the cyclobutane ring designates perfluoro.

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

2

2

4

2

2

2

4

Primary Processes

ΠΗ

|T]

+ F-

(5a)

—WV—•

2CF =CF

—VW—•

CF - + CF2=CF—CF -

2

(5b)

2

3

2

(5c)

Chain Propagation F- + C F — C F 2

4

(6)

2

C F - + C^F,— C F -

(7)

C F - + C F — C„F -

(8)

2

5

4

4

e

2

9

4

na

etc. CF - + CaF -*- C F 3

4

+

3

7

Ç F -^C F 2

4

e

1 1

etc.

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

(9) (10)

6.

HECKEL AND HANRAHAN

Ξ E

+ C F 2

• Γ" ^ — C F JFJ 2



4

X-Radiolysis of Perfluorocyclobutane

C F 2

+ C F 2

(11)

2

r

4

129

r* ^ F s *

(12)

I F I

4

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

etc.

Chain Termination CjF -

+

C F -

+ C3F7 ·

B

2

5

CF --^C F 3

3

(13)

8

C F 5

(14)

1 2

[fJ

+ C F -

+

C F -

+

[ 7 f

C F -

+

Ξ'

C F -

+

1

2

2

3

3

5

5

7

T

+

(Tf c

C s F

'

F

( C e F12)

ι—r < »

''

' _

c

Ι

Ι C F

HJ

e

F

1 3

2

4

-

(15)

(C F 8

i e

(16)

)

(C10F20)

(17)

(C F )

(18)

[7]- ' "

(C F

(19)

ΕΓ '"Ξ

(Ci Fi )

-

τ C F -

ΕΓ

^

T

c

F

1 4

9

Ρ

0

1 8

)

8

(20)

This reaction scheme gives a reasonable account of the formulas of the products and suggests structures for many of them. The radicalradical combination reactions are chosen because they lead to observed products. Other steps could be written, and presumably a l l possible combination steps occur to some extent, depending on the rates of forma­ tion of precursors, on the relative rate constants for the possible reaction channels disposing of each species, and on the concentration of the second species involved in each bimolecular step. Additional products are cer­ tainly formed. A l l major peaks beyond C were double or triple, and 7

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

130

RADIATION CHEMISTRY

II

intervening peaks too small to measure were seen. Furthermore, products beyond C i must be formed since polymer was observed. Radical-radical disproportionation steps are omitted from the mechanism because there is experimental evidence that such processes do not occur i n perfluorocarbon systems (18). 4

h

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

0Ί0

0

05

1

MOLE FRACTION OF METHANE Figure 2. G values of the minor products formed in the radiolysis of c-C F -CH mixtures. • C H ; Ο CH C F ; Ο CF; Φ CF i

e

4

S

2

8

S

i

S

8

3

6

The postulated primary chemical steps can be examined in the light of the mass spectroscopic fragmentation pattern of F-cyclobutane (14), which shows C F as the parent peak (abundance 100 arbitrary units). A 1-3 split is also favorable; C F , has an intensity of 87 units, and C F \ 25 units. These data are consistent with Steps 5b and 5c. Rupture of a C — F bond (Reaction 5a) must be more important in the radiolysis mech­ anism than indicated by the low abundance of 0.1 unit for the C F ion in the mass spectrum. However, this anomaly occurs not only in other fluorocarbon systems ( J , 5, 10, 11, 22) but i n most hydrocarbon systems studied to date. The intensities of C F and C F are also substantial in the mass spectrum, being 13 and 54 units respectively. These results, and the fact that C F has often been found under pyrolytic conditions (2,12), suggest the possibility that difluorocarbene plays a role in the mechanism, perhaps leading to a portion of the odd-carbon products. W e have no evidence on this point, however. Another process which should be considered is ring opening by fluorine atom attack (Reaction 21). 2

4

+

3

r

+

3

4

2

+

7

+

+

2

F- +

|TJ

— -

CF CF CF CF 3

2

2

2

(21)

This reaction would not lead to a unique product since η-butyl radicals

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

6.

131

X-Radiolysis of Perfluorocyclobutane

M E C K E L AND HANRAHAN

can be produced in other ways (cf. Reaction 7). However, Reaction 21 may be an additional source of these radicals. Reaction 21 is favored thermodynamically since the C — F bond strength exceeds the C — C bond strength in fluorocarbons by about 30 kcal./mole (17). There are interesting parallels between the products obtained in this study and in the xenon-sensitized photolysis of F-cyclobutane, reported by Miller and Dacey (16). Products identified included C F , C F , C F , C F , and c - C F , as well as considerable amounts of higher molecular weight fluorocarbons thought to be C compounds, but not positively identified. Although the authors considered a mechanism involving decomposition of excited F-cyclobutane an give F-ethylene, they discarded it i n favor of a scheme involving abstraction of two fluorine atoms to give X e F and excited F-cyclobutene. 4

3

6

3

2

0

3

8

6

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

8

2

The present data are broadly similar to earlier radiolysis results on liquid F-cyclobutane by Fallgatter and Hanrahan (5), who found C , C , C , C , and C compounds. Higher products were not detected. Since the earlier work involved liquid rather than gaseous F-cyclobutane and since the radiolysis samples studied at that time were known to be slightly contaminated with hydrogen-containing materials, a detailed comparison with the present work is not warranted. Methane-F-cyclobutane Mixtures. The present data on the radrblysis of methane-F-cyclobutane mixtures can be compared with several earlier studies. Rajbenbach (19, 20) measured H yields and competition between fluorocarbon and N 0 as electron scavengers in liquid-phase systems containing F-cyclobutane (or other fluorocarbons) in n-hexane and cyclohexane. A marked decrease in the hydrogen yield was demonstrated and attributed to nondissociative electron capture by the fluorocarbon. Fallgatter and Hanrahan (6) studied the full range of liquid mixtures of cyclohexane and F-cyclohexane from 0-100% fluorocarbon and measured several product yields as a function of dose. They found a large yield of c - C F n H ( G = 3.5) and an increase in dicyclohexyl at intermediate concentrations, as well as confirming the drop in the hydrogen yield. It was pointed out that production of c - C F u H appears to rule out nondissociative electron capture as the main explanation of the effect on the hydrogen yield. Sagert (21) made detailed studies of more dilute solutions (0.3M or less) of F-cyclobutane, F-cyclohexane, and F-methylcyclohexane in cyclohexane and also noted extensive conversion of C — F to C — H bonds. H e also interpreted this as evidence that processes other than nondissociative electron capture must be taking place. 2

5

e

3

8

2

2

6

e

Fallgatter and Hanrahan (6) noticed that the radiolysis products from their fluorocarbon-hydrocarbon mixtures resemble the pure hydrocarbon results rather than the products from the pure fluorocarbon. A similar generalization appears to be valid for the present experiments.

Hart; Radiation Chemistry Advances in Chemistry; American Chemical Society: Washington, DC, 1968.

132

RADIATION CHEMISTRY

II

For C H - c - C F mixtures, H , C H , C H , and C H all appear at intermediate concentrations, although their yields are depressed below ideal mixture lines, while all fluorocarbon products except C and C compounds are eliminated. However, a more detailed comparison of this work with the earlier studies shows a major difference in that efficient hydrogen scavenging at low concentrations of fluorocarbon was not observed. Whereas 0.2M added c - C F decreased the hydrogen yield from liquid cyclohexane by 50% according to Rajbenbach (19), it required about 25 mole % F-cyclobutane to decrease the H yield from gaseous methane by the same amount. These contrasting results may be caused by a larger role for ionic processes in the production of H from gaseous methane than from liquid cyclohexane (23). A marked regularity was found between the decrease in the hydrogen yield and the decreases in the yield of hydrocarbon products i n the experiments depicted in Figures 2 and 3. In the absence of complications, stoichiometry indicates that: 4

4

8

2

2

4

2

G

3

8

2

4

3

8

Downloaded by UNIV OF MELBOURNE on July 10, 2016 | http://pubs.acs.org Publication Date: January 1, 1968 | doi: 10.1021/ba-1968-0082.ch006

2

2

A G ( H ) = AG(