Citation
The photolysis, radiolysis, and mass spectrometry of 1,1,2,2-tetrafluorocyclobutane

Material Information

Title:
The photolysis, radiolysis, and mass spectrometry of 1,1,2,2-tetrafluorocyclobutane
Creator:
Ravishankara, Akkihebbal Ramaiah, 1949-
Publication Date:
Language:
English
Physical Description:
xiii, 130 leaves : ill. ; 28cm.

Subjects

Subjects / Keywords:
Atoms ( jstor )
Dosage ( jstor )
Fluorides ( jstor )
Hydrogen ( jstor )
Ions ( jstor )
Mass spectroscopy ( jstor )
Molecules ( jstor )
Photolysis ( jstor )
Quenching ( jstor )
Radiolysis ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Tetrafluorocyclobutane ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 123-129.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Akkihebbal Ramaiah Ravishankara.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
025216976 ( ALEPH )
02672732 ( OCLC )

Downloads

This item has the following downloads:


Full Text











THE PHOTOLYSIS, RADIOLYSIS, AND MASS SPECTROMETRY
OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE










By

AKKIHEBBAL RAMAIAH RAVISHANKARA


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY








UNIVERSITY OF FLORIDA


1975



















mg ad
C


d)-Zedc?
















ACKNOWLEDGEMENTS


The author expresses his deepest gratitude to Dr. R.J. Hanrahan for his encouragement, advice, help and patience during the course of this work. He also thanks Dr. J.R. Eyler for providing access to the ICR instrument, Dr. W.S. Brey, Jr., for obtaining and interpreting the N.M.R. Spectra of some of the compounds, and Mr. R.J. Dugan for designing the circuit shown in Chapter II.


iii
















PREFACE


The decomposition dynamics of cyclobutane and substituted cyclobutanes have been investigated by numerous workers under conditions of direct photolysis, I'2 mercury sensitized photolysis,'4 thermolysis, -7 radiolysis, 8-11 and electron bombardment in the mass spectrometer.12 A point of special interest is the considerable ring strain in these compounds, causing a propensity to undergo ring opening and fragmentation into two ethylenic products. Even though the radiolysis of both cyclobutane and perfluorocyclobutane has been studied, no partially fluorinated cyclobutane has been investigated. In fact, even the mass spectral cracking patterns of partially fluorinated cyclobutanes have not been reported. However, a study of the decomposition of such a compound would be very interesting since it is expected to "bridge" the behavior of cyclobutane and perfluorocyclobutane. Hence, we undertook the study of the decomposition of 1,1,2,2-tetrafluorocyclobutane.

This work describes a study of the gas-phase gammaradiolysis of 1,1,2,2-tetrafluorocyclobutane (Chapter IV). To help understand the energetics and mechanisms involved in the radiolysis of this compound three additional investigations were carried out. They are:









(1) An electron impact investigation of 1,1,2,2tetrafluorocyclobutane,* described in Chapter I.

(2) Ion-molecule reactions in 1,1,2,2-tetrafluorocyclobutane, described in Chapter II.

(3) Mercury 6(3P1) sensitized decomposition of 1,1,2,2tetrafluorocyclobutane, described in Chapter III.





*Chapter I is reproduced in part with permission from the Journal of Physical Chemistry, 79, 876 (1975). Copyright 1975 by the American Chemical Society.
















TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS ...................................... iii

PREFACE ............................................... iv

LIST OF TABLES ........................................ viii

LIST OF FIGURES ....................................... ix

ABSTRACT .............................................. xi

CHAPTER I - AN ELECTRON IMPACT INVESTIGATION OF
1,1,2,2-TETRAFLUOROCYCLOBUTANE ............... 1

Introduction ........................... 1

Experimental ........................... 2

Results and Calculations .................. 5

Discussion ............................. 17

CHAPTER II - ION-MOLECULE REACTIONS IN 1,1,2,2TETRAFLUOROCYCLOBUTANE ...................... 23

Introduction ........................... 23

Experimental ........................... 23

Results ................................ 26

Discussion ............................. 35

CHAPTER III - THE Hg6(3 P) PHOTOSENSITIZED DECOMPOSITION OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE.. 44

Introduction ........................... 44

Experimental Section ...................... 45

Sample Preparation ..................... 45

Sample Irradiation ..................... 47



vi










CHAPTER III - Continued


Page

Product Identification and
Analysis ............................ 48

Results ................................ 51

Quenching Cross Section Measurements and Actinometry ................. 51

Photolysis Products .................... 53

Discussion ............................. 64

CHAPTER IV - GAM!A-RADIOLYSIS OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE IN THE GAS PHASE ............... 75

Introduction ........................... 75

Experimental ........................... 76

Reagents ............................ 76

Sample Irradiation and Dosimetry .... 76 Sample Preparation for Hydrogen and Organic Yield Measurements ............ 77
Sample Preparation for HF Yield Measurements ........................ 78

Product Identification ................ 79

Product Analyses .................... 80

Results ................................ 81

Material Balance ....................... 92

Discussion ............................. 96

APPENDIX I - IDENTIFICATION OF PRODUCTS ................ 109

APPENDIX II - UNIMOLECULAR FRAGMENTATION OF CYCLOBUTANE COMPOUNDS ........................ 118

REFERENCES AND NOTES .................................. 123

BIOGRAPHICAL SKETCH ................................... 130


vii
















LIST OF TABLES


Table Page

I Mass Spectral Cracking Pattern of c-C4H4F4
at 70 V ....................................... 8

II Appearance Potential Data ..................... 9

III Thermochemical Data ........................... 10

IV Rate Constants for Ion-Molecule Reactions in
1, 1,2,2-Tetrafluorocyclobutane ................... 31

V Mass Spectral Fragmentation Patterns of
1,1,2,2-Tetrafluorocyclobutane Under Analytical and Ion-Molecule Reaction Conditions ...... 36

VI Yields of Products Formed in a 120 Sec Photolysis of 50 Torr of c-C H F ...................... 61
4 44
VII Yields of Products Formed in a 300 Sec Photolysis of a Mixture of 400 Torr of H2 and 20
Torr of c-C H F ........................ ...... 63
4 4 4
VIII G Values in the Radiolysis of 1,1,2-2-Tetrafluorocyclobutane: Products Eluting on the
Silica Gel Column ............................. 89

IX G Values in the Radiolysis of 1,1,2,2-Tetrafluorocyclobutane: Products Eluting on the
SE-30 Column .................................. 94
X G Values of Products in 12 Scavenged Radiolysis of 1,1,2,2-Tetrafluorocyclobutane ........ 95

XIa Mass Spectral Cracking Patterns of Products... 115

XIb Mass Spectral Cracking Patterns of Products... 117


viii















LIST OF FIGURES


Figure Page

1 R.P.D. curves for CH4 from CH4, N2 from
N2, and C3H4F+ from C-C4H4F4, measured on
a Bendix T.O.F. mass spectrometer ...................6

2 R.P.D. curves for C2H2F2, C2F4, and C2H
from c-C4H4F4, measured on a Bendix T.O.F.
mass spectrometer ............................... 7

3 Circuit diagram of the pulse generator for
ion grid Number 1 ............................... 25

4 Normalized ion intensities as a function of
delay time in the T.O.F. spectrum of 1,1,2,2tetrafluorocyclobutane at 26.8 microns and
500 C ........................................... 28

5 Normalized ion intensities as a function of
delay time in the T.O.F. spectrum of 1,1,2,2tetrafluorocyclobutane at 53.0 microns and
500 C ........................................... 29

6 Normalized ion intensities as a function of
pressure in the T.O.F. spectrum of 1,1,2,2tetrafluorocyclobutane at 250 C .................... 32

7 The photolysis set up ........................... 46

8 Titration curves for quantitative analysis of F ion .......................................... 50

9 Modified Stern-Volmer plot; a plot of the reciprocal of the number of moles of N2 against
the ratio of the pressure of c-C4H4F4 to that
of N20 .......................................... 52

10 Yield of HF and H2 as a function of time in the
photolysis of c-C4H4F4 .......................... 55

11 Yield of C3H2F4 and C2HF3 as a function of time
in the photolysis of c-CtihFr ...................... 56









LIST OF FIGURES - Continued

Figure Page

12 Yield of C6H4F6(I) and C6H4F6(II) as a function
of time in the photolysis of C-C4H4F4 .............. 57

13 Yield of C2H6 and CH4 as a function of time in
the photolysis of c-C4H4F4 ....................... 58

14 Yield of C8H6F8 and the C8 compound as a
function of photolysis time ...................... 59

15 Plots of yields of C8HeF8 and the C8 compound
as a function of pressure in the photolysis
vessel ........................................... 62

16 Production of hydrogen fluoride as a function
of dose in the radiolysis of c-C4H4F4 .............. 82

17 Production of hydrogen as a function of dose
in the radiolysis of c-C4H4F4........................ 83

18 Gas chromatogram of irradiated 1,1,2,2-tetrafluorocyclobutane: low molecular weight products eluting on a Silica Gel column ................ 84

19 Production of I,I-C2H2F2 and C2H4 as a function
of dose in the radiolysis of c-C4H4F4 .............. 85

20 Production of C2F4 as a function of dose in
the radiolysis of c-C4H4F4 ....................... 86

21 Production of C2H2 as a function of dose in
the radiolysis of c-C4H4F4 ....................... 88

22 Gas chromatogram of irradiated 1,1,2,2-tetrafluorocyclobutane: high molecular weight products eluting on a SE-30 column ...................... 90

23 19F NMR spectrum of product Number 21, C6H4F6 .... 91

24 Production of C4H3F3 and C6H4F6 as a function
of dose in the radiolysis of c-C4H4F4 ............ 93
















Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy

THE PHOTOLYSIS, RADIOLYSIS, AND MASS SPECTROMETRY

OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE

By

Akkihebbal Ramaiah Ravishankara

December, 1.975


Chairman: Dr. R.J. Hanrahan Major Department: Chemistry


The fragmentation of 1,1,2,2-tetrafluorocyclobutane

under electron bombardment was investigated using a Bendix Time-of-Flight mass spectrometer. Using the Fox Retarding Potential Difference technique, measurements were made of the appearance potentials of major ionic fragments including C H F+ (12.15 V), C F+ (12.60 V), C H+ (13.15 V), and C H F+
2 2 2 2 4 2 4 3 4 (12.85 V). From these results it is found that AH0 for the C HfF+ ion is < +207 kcal/mole and AHf for the parent c-C H F
3 4 f 4 4 4 is < -202 kcal/mole. The symmetrical rupture of the molecule into C H F+ plus C H F is a sequential process with the sum
2 2 2 2 2 2
of the energy of breaking two nonequivalent C-C bonds equal to 126 kcal/mole, while the nonsymmetrical rupture into C H+
2 4
and C F is concerted, and the energy of breaking each of
2 4
the two equivalent C-C bonds between CF and CH groups is
2 2
61 kcal/mole.








To obtain kinetic data on the ion-molecule reactions in l,l,2,2-tetrafluorocyclobutane, measurements of the variation of the ion intensities with the changes in delay time were carried out using a Bendix Time-of-Flight mass
+
spectrometer. The parent ion C H F had very low abundance
4 4 4
and was found to be nearly non-reactive. However, ionmolecule reactions were observed for the olefinic ions C H F+
2 2 2
and C H+. It was found that C H F becomes the most abundant
2 4 2 2 2
ion in this system when the pressure is increased. Using the ICDR technique, it was established that reaction of C H F+ with parent c-C H F produces the species C H F+ and
2 2 2 4 4 4 6 6 6 C H F+.
5 6 4
The Ig6(3 P1) photosensitized decomposition of 1,1,2,2tetrafluorocyclobutane (c-C H F ) was carried out by irradiating a mixture (50 or 150 Torr) of c-C H F and mercury vapor with 253.7 nm radiation from a low pressure mercury lamp. The photolysis of a mixture of hydrogen and a small amount of c-C H F was also investigated to establish the reactions of hydrogen atoms with c-C H F . Using the Chemical method 44
of Cvetanovic, the quenching cross section of c-C H F
4 [ 4
0
was determined to be 0.3 A2 relative to that of N 0.
2
Major products in the photolysis of the pure system

include HF, two isomers of C H F , a C compound, and C H F
6 4 6 8 8 6 8
the yields of these species and six others, formed in moderate quantities, were followed as a function of irradiation time.

It is suggested that the rupture of the C-H bond is the major primary process in the chemical quenching of IIg6(3P1)


xii









by l,l,2,2-tetrafluorocyclobutane. A reaction scheme involving C H F 4 and H" (formed in the primary quenching process), is proposed to explain the formation of the observed products.

The gamma-radiolysis of gaseous 1,1,2,2-tetrafluorocyclobutane was studied at 50 Torr pressure and 240C, both pure and with added oxygen. The major products were HF, H ,
2
1,1-C H F , C F , C H , and C H F ; C H was seen only in
2 2 2 2 4 2 2 4 3 3 2 4
the presence of oxygen. The addition of approximately 4% oxygen dramatically increased the G values of two olefins, 1,1-C H F (from 0.046 to 0.47) and C H (from -0 to 0.125),
2 2 2 2 4
decreased the yield of H , C F and C H. Under the same
2 2 4 2 2
conditions most heavier products were either eliminated or considerably reduced. The yield of C H F , however, was not changed substantially. Addition of both C H and I as
2 4 2
scavengers decreased the yield of HF by 25%. The observations are interpreted in terms of reactions of both ions (especially C H F+ and F-) and neutral species.
2 2 2


xiii















CHAPTER I


AN ELECTRON IMPACT INVESTIGATION OF
1,1,2,2-TETRAFLUOROCYCLOBUTANE


Introduction


Information on the mass spectral cracking pattern of a compound is very helpful to an understanding of its radiolytic behavior, since slow electrons are responsible for the decomposition in both the cases. Additionally, bond energy data are very useful in interpreting the decomposition dynamics of any compound. One reliable and accurate method for obtaining bond energy information is the measurement of appearance potentials of fragment ions of a compound in a suitably equipped mass spectrometer using Retarding Potential Difference (R.P.D.) techniques. The first goal of the present investigation was to obtain the mass spectral cracking pattern and bond dissociation energies in 1,1,2,2-tetrafluorocyclobutane (subsequently referred to as c-C H F
4 4 4

F F
F-i-

I I
II H


since these data were not available.

Secondly, it has been proposed that the c-C H F ion
444





2



constitutes the transition complex in ion-molecule reactions of C H F+ with C H F , C F with C H , and C H+ with C F
2 2 2 2 2 2 2 t 2 4 2 4 2 4 It was postulated that the complex is very short-lived and during its existance maintains its cyclic structure. Accordingly, a study of the fragmentation of this compound under electron impact is an obvious experiment to shed light on whether the predictions of the model are right and assumptions plausible.

In the present work, the appearance potentials of C F
2 4
C H , C H F +, and C H F+ from c-C H F were measured. From
2 4 2 2 2 3 4 4 4 4 these data we can calculate some of the bond dissociation energies in this compound and estimate some others. The experiments were done on a Time-of-Flight mass spectrometer using standard R.P.D. techniques.15'16


Experimental


The mass spectral cracking pattern of c-C H F was
4 4 4
obtained on the Bendix Time-of-Flight mass spectrometer at 70 eV. The R.P.D. experiments were also done on the Bendix, as described by Melton and Hamill16 with some minor changes. Potentials on the five grid electron gun were as follows: No. 1 grid (electron control grid, nearest to the filament),

-6 V with a +12 V pulse of 3 psec duration repeated at a frequency of 10 kHz; No. 2 grid, +0.01 V; No. 3 grid (Fox retarding potential difference grid), -1.000 V with an intermittent AV of -0.200 V; No. 4 grid, +0.01 V; and No. 5 grid was grounded to the ion source structure (and therefore









to the frame of the instrument). Potentials on grids 1-4 are referenced to the electron filament, and hence their absolute value relative to ground varies along with the filament. The choice of -1.000 V for the Fox arid and -0.200 as AV was made after preliminary optimization experiments to obtain the best compromise of ion current I and its variation AI in response to AV applied to the R.P.D. grid. The purpose of the R.P.D. procedure is to measure an amount of ion current AI attributable to electrons in a narrow voltage range AV taken as a "slice" out of the much broader thermal electron energy distribution from the filament.

The ion focus pulse was triggered 0.7 psec after the control grid pulse; we found this measure essential to obtaining reliable results. (It is possible to use the "timelag-focusing" control for this purpose, on Bendix instruments so equipped.) The potential on the electron collecting trap was reduced to +8 V to minimize extraneous additional ionization of the experimental compound in the target region. Several minor modifications to the control electronics of the Bendix were made as recommended by Melton and Hamill.16

The ion current signal from the electrometer circuit

in the "scanner" unit was fed to a potentiometric chart recorder through a potential divider. The compound under study was contained in a 5 Z. vessel and was leaked into the ionization chamber through a gold leak. The pressures in the ionization chamber could be changed by changing the pressure in the 5 k. vessel. The usual operating pressure









in the ionization chamber was 6-8 x 10-6 Torr. The normal trap current was 6 pA and the change in trap current was less than 5%. (The trap current was not regulated.)

The electron beam accelerating potential, which was

applied to the filament, was generated by standard circuitry in the Bendix and monitored on the 40-V scale of a Digitec Model Z-204-B digital volt meter (resolution 0.005 V, accuracy 0.01 V). However, due to contact potentials, surface charges, etc., it can not be assumed that the effective accelerating potential of the electron gun is identical with that measured externally. Accordingly, an R.P.D. curve was measured for Ar+ ion from argon gas and the apparent appearance potential of Ar + (corresponding to the first break) was determined from a graph made against the experimental voltage scale. The difference between the measured value and the accepted literature value gave the necessary correction for the displacement of the voltage scale. Using the scale calibrated as described, cross checks were done by measuring the appearance potentials of C F+ (from C F ) and N+ (from N ). Our measured values
2 4 2 4 2 2
agreed with the literature to within 0.1 V, indicating that our voltage scale is linear at least within the approximately 6-V range tested. In later experiments C F+ and N+ were used
2 4 2
as secondary calibration standards.

We could reproduce exactly the shapes of ionization efficiency curves for CH from CH and C H , C H , and C H+ from
4 4 2 6 2 4 2 5 C H , which were reported by Melton and Hamill;16 all re2 6
ported "break points" agreed within a 0.1-V range. (The




5




CH+ curve is shown in Figure 1.) Although we always had
4
a moderate air background, N+ (28) does not interfere with
2
C H+ (28) as the appearance potential of N+ is 15.6 V and
2 4 2 N+ has zero intensity below about 15.1 V (well above the
2
region of interest for C H +). In the spectrum of c-C H F 2 4 4 4 4 the appearance potential of CH+ could not be carried out
2
because of the interference by the CH+ in the background and
2
that of CF+ and C H F+ could not be done because of very low
2 3 3 2
intensities.

The 1,1,2,2-tetrafluorocyclobutane was obtained from

Columbia Organic Chemicals Co. It was dried and deaerated on a vacuum line through several freeze-pump-thaw cycles at liquid nitrogen and Dry Ice temperatures, and vacuum distilled into a glass bulb for transfer to the mass spectrometer.


Results and Calculations


The Retarding Potential Difference curves which we obtained for C H F+, C H F+, C F+, and C H+ from c-C H F are
3 4 2 2 2 2 4 2 4 4 4 4
shown in Figures (1) and (2) along with N+ as a standard and
2
CH for comparison with the work of Hamill and Melton.16 The
4
mass spectral fragmentation pattern which we obtained for

c-C H F is shown in Table I. In Table II we present the
4 4 4
appearance potential values which were found for the several major product ions from c-C H F , along with values for
4 4 4
certain other pertinent species from the literature.

From the appearance potential values given in Table II, along with the standard enthalpy of formation for several species from the literature 17-1 (given in Table III), it is






























ZD
-4

14J,80


(0

1 2 3







15,55









12 13 14 15 16 17 eV


Figure 1. R.P.D. curves for CH+ from CH4, N2 from
N2, and C3H4F+ from c-C4H4F4, measured on
a Bendix T.O.F. mass spectrometer.





























4-)




-4
-P1 2


H




14,76









11 12 13 14 15 16 eV

Figure 2. R.P.D. curves for C2H2F+, C2F4, and C2H4
from c-C4H4F4, measured on a Bendix T.O.F.
mass spectrometer.










TABLE I


Mass Spectral Cracking Pattern of
c-C H F at 70 V.
4 44


Mass Intensity Assignment Mass Intensity Assignment


14 18 Cif 57 7 C H F
2 3 2 16 7 CH 59 16 C H F
4 3 4 26 9 C H 64 100 C H F
2 2 2 2 2 27 11 C H 69 5 CF or C 11 F
2 3 3 4 2 28 23 C H 75 4 C HF
2 4 3 2 31 13 CF 77 9 C H F
3 3 2
32 37 CHF 89 16 C H F 4 3 2
39 9 C H 100 23 C F
3 3 2 4 40 30 C H 109 9 C H F
3 4 4 4 3 44 12 C HF 113 3 C HF
2 3 4
45 18 C H F 127 2 C 1I F
2 2 4 3 4 50 6 CF 128 3 C H F
2 4 4 4 51 16 CHF
2










TABLE II


Appearance Potential Data


Ion Reactant Product AP C F+ c-C H F C F+ + C H 12.60 Va
2 4 4 4 4 2 4 2 4


C H c-C H F C 1+ + C F 13.15 Va
2 4 4 4 4 2 4 2 4


C H F+ c-C H F C H F + C H F 12.15 Va
2 2 2 4 4 4 2 22 2 2 2


C 11 F c-C H F C H F + CF 12.85 Va
3 4 4 4 4 3 4 3


C H+ C H C i+ + H + H 16.3 Vb
2 4 2 6 2 4 C H+ C H 10.48 Vc
2 4 2 4


C H F+ H C-CF C H F + HF 10.30 Vc
2 2 2 3 3 2 2 2


aThis work

bMelton and Hamill, Reference I. (16) cNSRDS complilation, Reference I. (18)










TABLE III


Thermochemical Data


Ion or Radical


C F+
2 4

CH
2 4

C H+
2 4

CF
2 4

C H F+
2 2 2

C H F
2 2 2

CF
3


AH 0 in Kcal/mole
f 298

78


12.49


253


-155.5


159


-78.6


-112.6


Reference


18 18 18 18 18 18 19









possible to calculate heats of formation of the ion C H F+
3 4
and the neutral parent molecule c-C H F , as well as values
4 44

for the energy of dissociation of C-C bonds in the C ring.
4
The heat of formation of the parent molecule can be formulated in three different ways. First, using our measurements of the appearance potential of C F+ we obtain:
2 4
+
c-C H F - C F + C H + e I.(la)
4 44 2 4 2 4


A(C F+) > AH (C F+) + AH (C H) -AH (c-C H F
2 4 - 2 4 f 2 4 f 4 4 4
I. (lb)

Therefore

AH0 (c-C H F ) > AH0 (C F+ ) + AHO (C H ) - A(C+
f 4 4 4 - f 2 4 ~f 2 4 2 4
I. (1c)

AHf (c-C H F ) > 78 kcal/mole + 12.49 kcal/mole - 12.60 eV
4 4
I. (ld)

AH0 (c-C H F ) > -200 kkal/mole I.(le)
f 4 4 4


Next, based on the appearance potential of C H we can
2 4
calculate:

c-C H F C H++ C F + e- I.(2a)
4 4 4 2 4 2 4

A(C H+) > AH 0(C H1 ) + A1(CF AH 0 (c-C H F
2 4 - f 24 f 2 4 f 4 4 4
I. (2b)

Therefore

AH 0 (c-C H F ) > AH (C H+) + AHO (CF) - A(C H+)
f 4 4 4 f 2. 4f 2 4 2 4
I. (2c)
AHf (c-C H F ) > 253 kcal/mole + (-155.5 kcal/mole
f 444

13.1 eV I. (2d)

AH0 (c-C H F ) > -206 kcal/mole I.(2e)
4 4 4 -









Finally, the measured appearance potential of C H F+ gives
2 2 2

c-C H F C 11 F+ + C H F + e- I.(3a)
4 4 4 2 22 2 2 2

A(C11F+) > A (C H F+), AH0 ( H F)- H (c-C 11 F)
2 2 2f 2 22 f 2 22 4 4 4 I. (3b)

Therefore

AH (c-C H F )> All (C H F+) + AHf (CHF) -A(C H F+)
f 4 4 4-f 2 2 2f 2 2 2 2 2 2 I. (3c)

AHf (c-C H F ) > 159 kcal/mole + (-79 kcal/mole) -12.15 eV
4 4 4
I. (3d)

AH 0 (c-C H F ) > -200 kcal/mole I. (3e)
f 4 4 4

Accordingly, the average value for AHf (c-C H F ) is >
4 4 4
-202 kcal/mole. The satisfactory agreement between the three calculations is gratifying because the calculations constitute a rather severe test of the consistency of the published thermochemical data as well as our appearance potential measurements.

Using our value of -202 kcal/mole for the heat of formation of c-C H F and the appearance potential of C H F +, we
4 44 3 4
can calculate a value for the heat of formation of this ion as follows:


c-C H F C H F+ + CF + e I. (4a)
4 44 3 4 3

A(C H F+) > AHf (C H F+) + AHf (CF -AHf (c-C H F)
3 4 - 3 4 f 3 f4 4 4
I. (4b)


Therefore










AHf0 (C H F+) < A(C i F+) - AH� (CF-) + AH� (c-C HF )
f 3 4 3 4f 3 f 4 4 4
I. (4c)

AH0 (C H F+) < 12.85 eV-(-112.6 kcal/mole) + (-202
f 3 4
kcal/mole) I. (4d)

AH0 (C H F+) < +207 kcal/mole I. (4e)
f 3 4

In order to consider the energy of breaking C-C bonds

in the c-C H F ring, it is necessary to define our notation.
4 44
First, we let D(F/F),D(J/Hand D (H/F)bethe bond dissociation


energy of the bonds shown below. (Note that DA/B represents a C-C dissociation energy of a bond where "A" and "B" are substituents on the carbon atoms.)


I I I I
F- C-C-F F- C-C-F


H H H TH


D (F/F)


D (11/H)


F F F F I I I I
F-CC-F F-C-i -F
H- - -H H-C-C-H
I I I I
H 1t H 11


D (F/H)


We also let AH(=H/H) and All(=II/F) represent the energies for the reverse of the following processes:


H H
C-C
7 N\

H H


H F

C-C
7 F H F


H C=C
H


-AH (=H/F)


H F
C=C
H F


In order to determine the enthalpy change in these reactions,


1. (5)


I. (6)


-AH









useful energy cycles can be written as follows:


H H 2DcH H H
- K
> C-C< + + Hi.7

H H H H


HHI +.(8 KH y (=H/H)


H H I (C H +/C H) C=C 24 24 H + eH H

I. (9)

and

H F DC-H + DC-F -F /- F

_ _-__-F __ N-C. + HF I. (10)
H F H F



H F (=F/H) J (= 1 I. (11) H F (C H F+/C 2 H22


H F

C H1 F + e
2 2 2
I. (12)


where DCH is the C-H bond dissociation energy in C H , or C 2H 3F 3, DC-F is the C-F bond dissociation in CH 3CF 3, and DHF is the H-F bond dissociation energy (equal to the bond energy in this case). We will use I (CH+/C2H ) to indicate the ionization potential (or appearance potential) of C H+ from
2 4









C H and similarly for other daughter ion/parent molecule
2 4
pairs which are mentioned subsequently.

Then we have


A(C H/C H )2D - Al I(C H+/ H1 1(3
2 4 2 6 C-H (=H/H) + 2 4 4 4

A(C H F+/C H F +D -D - All +
2 2 2 2 3 3 C-H C-F H-F (=F/H)

I (C H F+/C H F ) I. (14)
2 2 2 2 2 2

Therefore, application of the energy cycle (I.(7)-I.(9)) to Reaction I. (5) gives

H+C +
AH H= 2D + I(C H /C H ) - A(C 1 +/C H ) I.(15)
(=H/H) C-Hi 2 4 2 42 4 2 6

Al (=H/H)= 2(4.16) + 10.48 - 16.3 = 2.5 eV I.(16)


A similar calculation for Reaction I. (6) using the energy cycle (I.(10)-I.(12)) gives:


Ali = D + D - D + I(C H F+/C H2F
(=11/F) C-H C-F 1H-F 2 2 2 2 2 2

-A(C H F +/C H F ) I. (17)
2 2 2 2 3 3

= (4.16 + 4.6 - 5.82 + 10.12 - 11.2) = 1.86 eV I.(18)


It would be of interest to calculate the quantity AH(F/F) for formation of tetrafluoroethylene from the corresponding diradical, but the necessary data are not accessable - C F+
24
has zero abundance from C F and C F H.20
2 6 2 5
We next consider the energetics of the unsymmetrical splitting of c-C H F into C H+ and C F
4 4 4 2 4 2 4

c-C H F - C H+ + C F + e- I. (19)
4 4 4 2 4 2 4









It is possible to make two reasonable, limiting assumptions about the energetics of this process (and the similar symmetrical split into C II F+ and C H F ): either the re2 2 2 2 2 2
organization of the neutral fragment into a double-bonded species occurs before the split from the ion, or else it occurs afterwards. If the latter is the case, then the corresponding reorganization energy is not available to decrease the energy of formation of the observed ionic fragment:


A(C H+/c-C H F ) = 2D - AH(H + I(C H+/C H
2 4 4 4 4 (F/H) (=H/H) +12 4 2 4
I. (20)

This assumption leads to the following evaluation of D (F/H)' the energy per bond to break the CF ... CH carbon-carbon
2 2
bonds in the c-C H F ring:
4 4 4

2D(F/H) A(=H/H) I(C H+/C H ) + A(C H+/c-C H F
(FH = )2 4 2 4 2 4 44 4
I. (21)

2D (F/H) 2.5 eV- 10.48 eV+ 13.25 eV I.(22) and

D(F/H) = 2. 64 eV= 61 kcal/mole I. (23) The magnitude of the resulting bond dissociation energy appears reasonable; the significance of this result will be considered further in the Discussion section.

In the case of the symmetrical dissociation process, the alternative limiting assumption appears to give reasonable results:









c-C H F 4C H F+ + C H F 1.(24)
4 44 2 22 22 2


Assuming that all of the reorganization energy is available to drive the process, we obtain:


A(C H F/c-C HF ) D +D -2AHi
2 2 2 4 4 4 (F/F) (H/H) (=F/H)

+ I(C H F+/C H F ) I. (25)
2 2 2 2 2 2

which rearranges to give


(F/F) (H/H) = A(CH /c-C H F ) + 2AH(=F/H)


- I(C H F +/C H F ) 1. (26)
2 2 2 2 2 2

D(F/F) + D(H/H) = 12.05 eV+ 2(1.86 eV) - 10.30eV I.(27) D (F/F) + D(H/H) = 5.47 eV= 126 kcal/mole 1.(28) Accordingly, the average of these two bond dissociation energies is the reasonable value of 63 kcal/mole. This result will also be considered in the Discussion.

In order to carry out a calculation for the third

dissociation process, into C F+ and C H , we would need the
2 4 2 4
energy of a tetrafluoroethylene diradical rearranging to tetrafluoroethylene. As noted above, we were not able to calculate this quantity.


Discussion


The overall mass spectral fragmentation pattern of c-C H F is consistent with what might be expected for a
44
saturated, cyclic C compound.21 Most compounds of this type









give a small parent peak along with considerable fragmentation 13
into lower molecular weight ions. Furthermore, Jennings has postulated that the cyclic C H F+ ion, formed as an intermediate complex in several ion-molecule reactions, has a very short lifetime. It appears reasonable to suggest that the structure of the c-C H F + system reached upon electron
4 4 4
bombardment of the stable molecule is the same one formed in the ion-molecule reactions, although there could be differences in internal energies (especially vibrational energy). We note that c-C H F shows a small abundance of the parent ion,
4 4 4
c-C H F+ ; this is consistent with the fact that Jennings
4 4 4
also saw a small amount of c-C H F + at high pressures (pre4 4 4
sumably, involving some collisional stabilization in the ionmolecule experiments).

As can be seen from the cracking pattern of c-C H F
4 4 4
there is no peak corresponding to C H F+ or C HF+. This fact
2 3 2 3
lends some credibility to the assumption made by Jennings that the transition complex formed in the systems [1,1-CF CH
2 2
on 1,1-C H F , C H+ on C F , and C F+ on C 1H maintains a
2 2 2 2 4 2 4 2 4 2 4
stable structure and does not scramble F and H atoms during its brief existence. We note a number of species which must involve some rearrangement, however, such as CHF+. It is our suggestion that most (and perhaps all) such rearrangement occurs in connection with ring opening, which produces unsaturated carbon centers.

We calculate the heat of formation of c-C H F + to be
4 4 3
about 190 kcal/mole by Franklin's group contribution method.2









This value, when combined with our value for All of c-C H F f4 4 4
gives 14.4 V for the appearance potential of C H F+ if we
4 4 3
assume concurrent elimination of F-, and a value of 18.0 V if we assume F atom elimination. We find an approximate appearance potential of c-C H F+ of 13.5 + 1 V. Although
4 4 3
our data were not of high quality, due to the low abundance of the ion, its appearance potential is certainly not higher than 15.0 V. Accordingly, we conclude that c-C H F + is formed
4 4 3
by F elimination.

The two modes of splitting the cyclobutane ring into an ethylenic ion and an ethylenic molecule (Reactions I. (19) and I.(24)) clearly must proceed by two different mechanisms. For the unsymmetrical break (Reaction I. (19)) we utilized in the calculation only the reorganization energy of -CH CH into
2 2
C H , and obtained 61 kcal/mole each for breaking two CF ...
2 4 2
CH -bonds. If we had also utilized the reorganization energy
2
of -CF CF - into CF =CF the resulting bond dissociation
2 2 2 2
energies would increase by at least 15 to 20 kcal each, which would surely be too high for the badly strained c-C H F molecule.

On the other hand, we utilized the reorganization energy for two -CF CH - fragments into CF =CH , in order to inter2 2 2 2
pret the data for the symmetric split. We obtained 126 kcal for D(F/F) plus D(H/H), or an average of 63 kcal/mole for dissociation of CF ...CF and CH ... CH , which is of a reasonable
2 2 2 2
magnitude. (If we apportion the two 'energies similar to the known dissociation energies of CF ... CF and CH ... CH , the
3 3 3 3









resulting values are 57 kcal/mole and 69 kcal/mole respectively.) Use of only one reorganization energy term would give an impossibly low value of 46.5 kcal/mole for the average of D (F/F) and D (H/H)* In fact, the postulate that anything much less than 100% of the two reorganization energy terms is available to drive Reaction I. (24) would lead to an unrealistically low result.

Our specific suggestion for the mechanistic difference

in Reactions I. (19) and I. (24) (which forms the foundation of the different energetics) is that the former is a one-step dissociation giving C H+ and an energy-rich -CH CF � species
2 4 2 2 (which later forms CF =CF without aiding in the progress of
2 2
the original dissociation) while Reaction I. (24) is sequential, proceeding through an open-chain intermediate.

+
F F F F+
F-- -- F F-

H-T--H H - - + eI. (24a)
H H

+
F F F F+
/o


+ (-e-) I. (24b)
H H \H H


Consistent with the conclusions reached by Jennings et al., as well as our own observations, the open-chain species formed in Reaction I. (24a) would have to be short-lived (106 sec or less). Atom-migration in the open-chain intermediate could explain occurrence of several additional species in the









cracking pattern of c-C H F , in particular CF and C H F A similar ring-opening of C H F followed by a split into 443
1-carbon and 3-carbon fragments would explain the occurrence of CFH+ (37%) and C H F+ (9%). Several other similar
3 3 2
decomposition sequences can be suggested.

The more favorable reaction path and energetics leading to C H F+ compared to C H+ (and probably also C F +) are very
2 2 2 2 4 2 4
likely related to the relative abundances of 100%, 23%, and 23% for these three species. Although the intensity of C H F+ should equal the sum of the intensities of C F+ and
2 2 2 2 4 C H+ due to statistical considerations, it is in fact even
2 4
higher by a factor of two.
+
Our appearance potential value (12.85 V) for the C H F
314
ion gives a value of 207 kcal/mole for AH0 of this species,
f
which is closer to a previously reported experimental value of 210 kcal/mole than to a theoretically estimated value of 175 kcal/mole.23 The calculated AHf is apparently for a linear species; we suggest that the C H F+ species formed in
314
this work may have a cyclopropane structure. (Other C
3
species formed might also be cyclic.)

The utility of the electron impact method for determining thermodynamic quantities rests on the implicit assumption that the relevant but unmeasured corrections for excess energy in ionic and molecular fragments are small.16-18'24 This situation is recognized by inequality signs in Equations I. (1)-I. (4). In several instances (see Equations I. (15), I. (17), I. (21), and I. (26)) two ionization or appearance potential









terms enter with opposite signs, so that partial cancellation of excess energy effects can be expected. (In these cases no inequality is indicated since its sign is not unambigous.) An assumption of little or no excess energy is supported by Stevenson's Law25 in the case of C F+ and C 11+, since
2 4 2 4
the appearance potential of C F+ is 0.55 V less than that of
2 4
C H+. This argument is not applicable in the case of C H F+,
2 4 2 2 2 but the good agreement of AHf (c-C 11 F ) from Equations I. (1),
4 4 4
I. (2), and I.(3) suggests that any excess energy correction needed for A(C 11 F +) must be small in the latter case as
222
well; our results suggest that AH0 (c-C H F ) = -202 kcal/ f4 4 4
mole with little or no correction required for excess energy.

Although the R.P.D. curves for C H F+ and C 11+ both show
3 4 2 4
well-defined breaks at higher energies (respectively 2.70 and 1.61 V above the appearance potentials) we have not attempted to interpret the electronic states or processes involved. However, Hamill and coworkers have had considerable success in understanding data of this type.26















CHAPTER II


ION-MOLECULE REACTIONS IN
1,1,2,2-TETRAFLUOROCYCLOBUTANE


Introduction


This work was undertaken to obtain supplementary

information on ionic reactions in the gamma-radiolysis of gaseous 1,1,2,2-tetrafluorocyclobutane. Most of the reactions leading to stable products in the radiolysis of this compound appear to be radical reactions or molecular elimination processes. These radicals may have ionic precursors, however, since one of the major primary events in the gamma-radiolysis is the ionization of the compound. Accordingly, examination of ion-molecule reaction pathways in this system, using the technique of high pressure single source mass spectrometry and ion cyclotron resonance mass spectrometry, can shed light directly on the mechanism of the radiolytic decomposition.


Experimental


In this investigation the following two kinds of experiments were carried out using two different instruments:

(1) measurement of the variations of the ion intensities as a function of delay time (i.e., delay between ion generation









and ion extraction for detection) using a Bendix Time-ofFlight mass spectrometer and (2) investigation of the reaction pathways using the ion cyclotron double resonance (ICDR) technique in a Varian "Synotron" ion cyclotron resonance instrument.

The Bendix Time-of-Flight mass spectrometer (model 14107) was modified as described byFutrell et al.27 The electron energy was 100 eV. Ion grid No. 1 was at +18 V and the repeller grid at +6 V with respect to ground. Distance from the repeller grid to exit orifice was 0.533 cm, and the drift distance from the electron beam position to the exit orifice was 0.267 cm. Under these conditions the field strength in the reaction chamber is 11.25 V/cm. A new circuit shown in Figure (3) was used to provide a positive pulse to ion grid No. 1. This circuit provided for the variation of the pulse height from 0 to +25 V and the width from 2 to 6 isec while maintaining a very fast rise time (ca. 2 nsec) and very short delay (<1 nsec) relative to the ion focus pulse. In the experiments reported here the pulse height was at +23 V and the width was 4 Hsec. All other ion source potentials were identical to those reported by Futrell.27 Further details about setting up and adjusting the instrument were described by Heckel and Hanrahan ;28 similar procedures were employed in the present work. All ion intensity measurements taken under the pulsed conditions were corrected for diffusional discrimination using as a correction term the inverse square root of the ion mass (i/Vm).27'29 Mass














out put


0

0


-250VI
2p1s
ION FOCUS PULSE


o. 1 PF


2-6 iAs


Figure 3. Circuit diagram of the pulse generator for ion grid Number 1.









spectrometric data acquisition and data reduction, including the square root correction, were carried out using a General Automation SPC-12 mini-computer as described in a previous report.3�

The ion source was operated at 500 C. The reactant

c-C H F gas was leaked into the ion source through a gold
44
leak from a 5k.reservoir. Variation of the reservoir pressure allowed adjustment of the pressure in the ion source, which was monitored using an MKS Baratron capacitive micromanometer (model 77) equipped with a 3 mm head. The tetrafluorocyclobutane was obtained from Columbia Organic Chemicals, Columbia, South Carolina, and was purified by trap-to-trap distillation.

Ion cyclotron resonance experiments were carried out

on a Varian "Synotron" (model 5900) equipped with a flat cell. In the ion cyclotron double resonance (ICDR) experiments the second radio frequency field was applied to the source region. The procedure described by Jennings31 was followed to obtain the double resonance signals.


Results


It is possible to obtain rate data in a suitably equipped single-source mass spectrometer using a time delay mode, as originally described by Tal'roze.32 In this procedure ions are generated by a brief electron pulse (ca. 1 psec wide), allowed to react with the substrate gas for an appropriate time (typically 1 to 10 psec), and then extracted for analysis









by a pulse delayed with respect to the electron pulse.33

The procedure in our instrument is substantially the same, except that a small portion of the reactant gas product ion mixture from the source region continually effuses from the exit orifice, and these effusing ions are sampled by the ion focus pulse at a specified, variable delay time after the electron pulse.29'34 In either event, the ions undergo a pseudo-first order reaction with the substrate gas. A plot of the log of the normalized, corrected ion intensity versus delay time gives a straight line with slope equal to the pseudo-first order rate constant; multiplying by the ambient gas number density (molecules/cc) gives the second order rate constant directly.

Figure (4) shows the variation of the normalized, square root corrected intensities of five ions, C H+, C H F+ CF
24 222
C F+, and C H F+, as a function of delay time at a pressure of
2 4 4 4 3
27 microns. (Smooth curves through the experimental data points in Figures (4), (5), and (6) are visual best fits.) Intensities of C H+ and CF+ decay to zero within 3 jisec and
2 4
C H F+ intensity grows to a near constant value at approximately 222
3 psec. After 3 psec C F+ and C H F+ ion intensities continue
2 4 2 2 2
to grow, while the abundance of C H F+ levels off and declines
2 2 2
somewhat. A similar plot is obtained at a pressure of 53 microns as shown in Figure (5). Semilogaritlic plots of normalized, corrected intensities for loss of C H+ and CF
2 4
+ + + 0
and formations of C H F , C F , and C H F , measured at 50 C
2 2 2 2 4 4 4 3
against delay time gave good straight lines; resulting rate














0.60w





0.50.
0 .5 0 . 0C 11 F + ( 6 4 )
2 2 2




0.40'





N 0.30

+
C H F (109)
H 4 4 3


0.20.
C H + (28)
2 4 + (io


0.10 (00





0.00.
I I I I

1.0 2.0 3.0 4.0 5.0 6.0 Delay time in psecs Figure 4. Normalized ion intensities as a function of delay
time in the T.O.F. spectrum of 1,1,2,2-tetrafluorocyclobutane at 26.8 microns and 500 C.






















0.50
C2H F+ (64)

0

0.40





H 0.30



C 11 F+ (109)
S4 4 3
0.20





0.10 C F + (100)
2 4


CF+ (31)C (28)
2 4
0.00
0.0 1.0 2.0 3.0 4.0 5.0 Delay in psecs


Figure 5. Normalized ion intensities as a function of
delay time in the T.O.F. spectrum of 1,1,2,2tetrafluorocyclobutane at 53.0 microns and
500 C.









constants (based on a linear least squaresfit of the data points) are given in Table 1V.

It is also possible to obtain ion-molecule reaction

rate data in a single source instrument using the classical pressure variation method. This technique consists of measuring the ion intensities as a function of pressure in the presence of a known field strength, produced by the repeller grid. Although ion-molecule rate data were first taken using the pressure variation method, there are difficulties with interpreting the results of this procedure.28'35 When continuous ionization is used, the essence of the problem is that the effective reaction time for light ions is shorter than for heavy ions, since the former are lost sooner in collisions with the wall. By operating our instrument at a fixed, intermediate value of the ion-extraction delay parameter, it is possible to mitigate this difficulty to some extent; most intermediate and heavy mass ions then have the same effective reaction time, while some but not all of the light ions are lost. By operating in this manner with a fixed delay of 1.6 psec we were able to take pressure dependence data which gave a reasonable comparison to the timedelay data, without recourse to extensive diffusion calculations.28 Results thus obtained are shown in Figure (6); the rate constants are shown in Table IV.

Several ionic species observed in the Bendix instrument had intensities too low for quantitative work. Many of these (CIIF+, C 11+,etc.) are fragment ions which apparently disappear
3 3









TABLE IV


Rate Constants for Ion-Molecule Reactions
in 1,1,2,2-Tetrafluorocyclobutane


Pressure in Rate Constant in Microns cc molecule-' Reaction sec-1 x loll Methoda


Disappearance of 26.8 11.0 TD C H+ 53.0 7.76 TD
2 4
16.2 PD Disappearance of 26.8 17.9 TD CF+ 53.0 11.4 TD 17.7 PD Formation of 26.8 4.3 TD C H F+ 53.0 0.89 TD
2 2 2
1.74 PD Formation of 26.8 5.1 TD C F+ 53.0 2.9 TD
24
3.4 PD Formation of

C H F+ 26.8 7.5 TD
4 4 3
53.0 4.0 TD 5.2 PD


aTD = Rate constants obtained from the time delay experiments. PD = Rate constants obtained from the pressure dependence experiments.










0.60 0.50




0.40 0.30




0.20 0.10




0.00


0 10.0


20.0


30 40 50 60 70


Pressure


in microns


Figure 6. Normalized ion intensities as a function of pressure in
the T.O.F. spectrum of 1,1,2,2-tetrafluorocyclobutane at
250 C.









by reaction with the substrate. Of more interest, however, are two ions with molecular weights greater than the parent molecule; the species appear at m/e 164 (C H F+) and 192
42 2 6
(C H F+). The former ion is formed with moderate intensity
6 6 6
under conditions of very low repeller grid voltage; it is formed in an ion-molecule reaction in which an ion of m/e 64 (C H F+) is consumed by reaction with the substrate.
2 2 2
Although no further information on the ion of mass 192 was obtained using the Bendix instrument, its mode of formation was characterized by the ICDR technique described below.

In the original ion cyclotron method for the observation of ion-molecule reaction as developed by Baldeschweiler, an investigation of the variation of ion intensities as a function of pressure also allows calculation of rate constants.36 Although we investigated the ICR spectrum of 1,1,2,2-tetrafluorocyclobutane over the entire pressure range available on the Varian instrument, no variation in ion intensities was observed. (The instrument operates well at pressuresfrom 1 x 10-4 Torr to 1 x 10-6 Torr; although the
-7
region 1 x 10 Torr is accessible, the poor signal-to-noise ratio prevents quantitative measurements in this region.) Throughout the readily accessible pressure region, the spectrum observed on the ICR instrument resembles highpressure or long delay-time spectra seen using the Bendix machine. Apparently due to the very long residence times in the ICR instrument (several milliseconds), the more reactive ions, such as ethylene in the present case, have already reacted.










A more recent technique using the ICR instrument is the double resonance method, in which kinetic energy of a chosen reactant ion is enhanced by irradiation with a second radio frequency field. The product ion intensity will increase provided the formation reaction of the product ion is endoergic. However, a decrease in the product ion intensity with an increase in the kinetic energy of the reactant ion indicates either an exoergic or thermoneutral reactions. Such a decrease in intensity is most commonly seen in exoergic reactions.

In the ICR instrument two ions heavier than the parent ion, one at m/e of 192 and the other at m/e of 142, were
-6
formed above 6 x 10 Torr. Although the ion with m/e of 192 (C H F ) was seen in experiments carried out using the Bendix machine, the ion of m/e 142 (C H F ) was not seen
5 6 [4
probably due to a very low formation rate constant. Using the ICDR technique, the reactant ion in the formation reactions of these two heavier ions was established to be C H F+. The ICDR signal for the reactions leading to the formation of C H F+ and C H F+ at different power-levels of
6 6 6 5 6 4
the second radio frequency field, applied to the source region, were of the shapes shown below (vertical axis, intensity; horizontal axis, double resonance frequency):


(For ion with m/e of 192)


(For ion of m/e of 142)









The power absorption by the product ions decreases when the intensity of the second radio frequency field of frequency corresponding to that of C H F+ is increased, indicating
2 2 2
that both the reactions are exoergic.


Discussion


An examination of the low pressure mass spectrum of c-C H F (Table V) shows that the compound fragments extensively under electron impact. The six most intense ions are all C and C fragments (abundancies from 33% down to
1 2
+
ca. 6%). The parent C H F ion (not necessarily cyclic)
+
has an abundance of only 0.1%; the other C species, C H F
4 4 4 3
and C H F +, occur to the extent of 2.9 and 5.3%, respectively.
1+32
In all, five C fragments are seen, but the abundance of the
3
most important is only 5.3%.

In previous researches by Jennings et al.'3'114 and

Anicich and Bowers37 the ion C H F+ was formed by the reaction of an ethylenic ion (C H F+, C H , or C F+) on an ethylenic
2 2 2 2 4 2 4
target molecule (C H F , C F , or C H , respectively). It
2 2 2 2 4 2 4
was found that the intermediate complex ion C H F is very
4 4 4
unstable (life time < 2 psec) and decomposes unless collisionally stabilized. The major products are an ethylenic ion/ethylenic molecule pair:


C H F + C H F + C H F II.(la)
4 4 4 2 2 2 2 2 2

C H F+ .... > C 11+ + C F II. (lb)
4 4 4 214 214

C H F+ C F+ + C H+ II. (lc)
414+ 4 214 214






TABLE V

Mass Spectral Fragmentation Patterns of 1,1,2,2-Tetrafluorocyclobutane
Under Analytical and Ion-Molecule Reaction Conditions


m/e Assignment


Intensitya


b
Low Pressure Spectrum


c
High Pressure Spectrum


ICR Spectrum atd
6.5 x 10-5 Torr


27 Microns 53 Microns
Short Time Long Time Short Time Long Time


7.0

20.5 14.7 7.3 6.0





30.0





5.1


6.7





6.1 3.6 5.9 48.3 7.8 7.6 7.4 6.6


5.1 7.9 6.6


5.4 3.3 4.3 44.7

5.0 5.4 5.8 5.1


6.0



6.0 49.7 7.8 8.3 9.9 12.2


12 27 28 31 32 39 51 59 64 77 89 100

109


CH
2
C H
2 3 C H
2 4 CF

CHF

C H
3 3
CHF
2
C H F
3 4 C H F 2 2 2 C H F 3 3 2

C H F
43 2 C F
2 4 C H F 4 4 3


5.8 4.1 7.9 4.2 12.0

2.9 6.6 5.3 32.7 2.6

5.3 7.5

2.9


84.2

2.1e

1.5 7.9 1.2e







113 C HF 0.1 0.9e
34
128 C H F 0.1 1.0 1.0 1.5 192 C H F 0.1 0.7
6 b 6



a All intensities are mass corrected and normalized to the total intensity of 100. Some minor ions not shown in the table. bFrom Chapter I.

CTaken from delay time experiments. Mass correction is proportional to the inverse square root of the mass.
dMass correction is proportional to the reciprocal mass. eSeen by Jennings et al., Reference 13.









From the fragmentation pattern shown in Table V it is clear that the behavior of the stable molecule c-C H F under
4 44
electron impact is very similar to that of the ion-molecule collision complex (as shown in Chapter I).

Under conditions of elevated pressure, long residence time, or both, the primary fragmentsfrom tetrafluorocyclobutane are available to enter into subsequent ion-molecule reactions with substrate. The abundance of the parent ion is so low, however, that it cannot make a significant contribution to the observed ion-molecule reactions. Several of the C species are of moderate abundance (ca. 6 to 12%);
1
they react so rapidly with substrate that is difficult to ascertain the details of the processes involved. It appears likely (Figures (4) and (5)) that they form one or more of the ethylenic ions. It can be seen in Figures (4), (5), and

(6) that the predominant reactions at intermediate pressures and/or residence times involve formation and decay of the ethylenic ions C +H C H F , and C F
2 4 2 22 2 4
The changes in the abundance of the major ions at low

pressures due to increases in pressure or residence time are shown in Table V . At lower pressures (of the order of 27
+
microns) and short delay times, the ions heavier than C F
24
are absent. When the delay time is increased the heavier ions increase in abundance. At higher pressures (of the order of 50 microns), all the ions seen at lower pressures are present throughout the range of delay times studied, including the heavy ions which at 27 microns were seen only at long









+
delay time. As seen in the table, the intensities of C H F
222
+ F+
C F , and C H F increase with increases in both the pressure
2 4 4 43
and delay time, while C H+ and CF+ abundancies decrease.
2 4
The same trend continues at higher pressures, as shown by

the ICR spectrum. (Even though the pressure in the ICR mass spectrometer was lower than in the Bendix mass spectrometer,

the very long residence times in the former instrument allow for a large number of collisions before the ions are lost to the walls or removed from the cell.) This result clearly shows that C H F+ is the major ion in this system at higher
2 2 2
pressures (for example, pressures of 10 to 50 Torr in radiolysis experiments).
+
Although C H F was found to react with the substrate,
2 2 2
the rate constant for the process must be quite small. In

experiments using the Bendix T-O-F, the C H F ion was the 222
most abundant even at high pressures and long delay times. The low reactivity of C H F + is substantiated by the experi2 2 2
ments using ICR, where mass corrected ion intensity fraction

of this ion was measured as a function of pressure (from 1 x 10-6 to 1 x 10-4 Torr). There was no measurable change in intensity of the ion with the change in pressure.

In contrast, C I1+ is a very reactive ion. The intensity
2 4
of C H+ decays to zero within a few microseconds with the
2 4
concomitant increase in the intensity of C H F+ (Figures '4)
2 2 2
and (5)). The same is true in the pressure dependence experiments where the growth of C H F + accompanied the decay
222
+
of C H with increasing pressure (Figure (6)).
2 4









The reaction leading to the decay of C 1+ could be a
2 4
charge transfer from C H1+ to c-C H F , followed by the de2 4 4 4 4
composition of the C H F+ intermediate to give C 1H F+
4 4 4 2 2 2

C H + c-C H F C 11F + C H II.(2a)
2 4 4 4 4 4 4 4 2 4

C H F+ C H F+ + C H F II.(2b)
4 4 4 2 2 2 2 2 2

However, this reaction would be thermodynamically unfavorable by 42 kcal, unless C H1+ in Equation II. (2) is in
2 4
an excited state. Hence, it is to be concluded that C H+
2 4
reacts with C I1 F+ to give C H F+ and C H F , the product
4 44 2 2 2 4 6 2
C H F having a heat of formation of at least -108 kcal/ 462
mole. A heat of formation of -108 kcal/mole is a reasonable

value to be expected for a C compound.
4

CH+ + c-C H F >pp (C H C H F+
2 4 4 4 4 6 8 4 2 2 2

+ C H F II. (3)
4 6 2

Reaction II. (3) could proceed through either a six center

intermediate or attack of C H+ on a C-C a bond in c-C H F
2 4 4 4 4 There is considerable evidence for the reaction of carbocation on a C-C u bond, as presented by Olah, whereas a

six center complex has not been established. Accordingly,

we expect Reaction II. (3) to be a simple radical ion (C H+
24
attack on the C-C a bond connecting two methylene groups in c-C H F , followed by the rupture of the C ion to give
4 44 6
C H F+. Since the reaction occurs in the gas phase, the
2 2 2
energy-rich C intermediate would tend to decompose, unlike
6









the situation in Olah's "magic acid" solvents, where the

intermediate undergoes collision deactivation followed by

atom migration to give a stable product. Several intermediates and ultimate products are probable, depending on which of the different C-C bond types is attacked, and how

the transient C ion ruptures. However, attack on the most
6
electron-rich of the three bond types (-CH --CtH -) would lead
2 2
to formation of the most abundant product C H F+ while attack
2 2 2
at the next richest site (-CF --CH -) would explain a small
2 2
production of C F
24"

Although the agreement is only moderate, the rate

constant for the disappearance of C H in Table IV can be
2 4
+
equated to the sum of rate constants for formation of C H F
2 2 2
and C F+. It will be seen that the sum of the two formation
24
rate constants is somewhat smaller than the C H+ decay rate
2 4
constant at 27 microns, and considerably smaller at 53 microns. We suggest that this discrepancy is due to a reaction

in which C H F + is lost by reaction with substrate; several
2 2 2
products are possible including C H F+, which would be formed
4 4 3
by a fluoride ion transfer process.


C H F+ + c-C H F C H F + C H F+ 11.(4)
2 2 2 4 4 4 2 2 3 4 4 3

As predicted by this model, the discrepancy between kloss

(C H+) and k (C H F + C F) increases with an increase
2 4 formation 2 2 2 2 4
in pressure. The reactions of C 11+ mentioned above could not
2 4
be established using ICDR since the intensity of C F+ was
2 4
very small.









The small abundance of species heavier than 128 (the parent) shows the relative unimportance of ion-molecule condensation reactions in this system. Of the three heavy ions (m/e 142, 164, and 192), only the one at 192 was seen both in the Bendix and ICR. The relatively small intensity of this ion forbade any kind of time or pressure dependence measurements. However, using the ICDR technique we could show that at least one of the routes of formation is by the reaction of C H F+ with c-C H F As mentioned in the Results
2 2 2 4 4 4
section, this ion is formed by an exoergic reaction.


C H F + c-C H F > C HF II. (5)
222 444 666

m/e = 64 m/e = 128 m/e = 192


The ion of mass 142 was seen only in the ICR and the one at 164 only in the Bendix. Again, the intensities of these ions were very small indicating small rate constants for the formation of these products. The ion at mass 142 (C H F+ 564
was also formed by the reaction of C 11 F+ with c-C H F as
2 2 2 4 4 4
shown by ICDR.


C H F + c-C H F C H F + CF II.(6)
2 2 2 4 44 > 5 6 4 2

m/e = 64 m/e = 128 n/e = 142 m/e = 50


Again, this ion (C H F +) is formed through an exoergic 564
reaction. This observation is compatible with the absence of this ion in the Bendix experiments where the electron energy was 100 eV, which would tend to increase the energy




43




of the reactant ion, C H F+.
2 2 2
The route for the formation of the ion at mass 164 (C H F+) was not established because this ion was absent
4 2 6
in the ICR. However, it is very likely that C H F+ is
2 2 2
responsible for this ion also. The occurrence of C H F+
426
in the Bendix and not in the ICR suggests that the lifetime of the ion is very small, perhaps of the order of 2 jisec or less.
















CHAPTER III


THE Hg6(3pI) PHOTOSENSITIZED DECOMPOSITION
OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE Introduction

3
Mercury 6( P1) photosensitized decomposition studies on a large number of alkanes and cycloalkanes have established that C-H bond rupture is the primary quenching process in these systems. ' Work on a few partially fluorinated hydrocarbons has indicated that in these systems as well, C-H bond scission is the primary mode of quenching the excited mercury.41 Only a limited amount of work has appeared dealing with mercury sensitized decomposition of perfluorocarbons (whether open chain or cyclic); details of the decomposition mechanism have not been established.4 There7i have been no previously reported investiqations of the mercury sensitized decomposition of partially fluorinated cyclobutanes or other cycloalkanes.

The present investigation of the mercury sensitized

photolysis of 1,1,2,2-tetrafluorocyclobutane was undertaken in parallel with studies of the gamma radiolysis, mass spectrometric fragmentation, and ion-molecule reactions of this compound. It was expected that mechanistic information derived from the photolytic (and mass spectrometric) studies would aid









in interpretation of the more complex radiolytic system.


Experimental Section


Sample Preparation

The vessel used in photolysis experiments had a volume of 91 cc; it was constructed of 18 mm ID GE type 204 clear fused quartz and was equipped with a glass break seal and side arm (Figure (7)). To clean the vessel between experiments, it was soaked in nitric acid overnight to remove inorganic material (especially mercury), rinsed with distilled water, and annealed at 5650 C to pyrolyze any organic materials on the walls. Before introducing the sample a drop of triple-distilled mercury was placed in the vessel, and the vessel pumped on for at least six hours. In the case of photolysis of the pure system, samples of c-C H F
4 4 4
were measured in a standard volume using PVT technique, transferred to the photolysis vessel, and sealed under vacuum. For photolysis of a mixture of hydrogen and c-C 11 F , the
4 44
tetrafluorocyclobutane was introduced as above, the photolysis vessel immersed in liquid nitrogen, and hydrogen metered directly into the photolysis vessel. After the introduction of hydrogen, the vessel was very cautiously sealed. The pressure of hydrogen in the vessel at room temperature (2970 K) would be 3.87 (i.e., 297 K/770 K) times the measured pressure. For quenching cross section and actinometry experiments, c-C H F was introduced as in the case of the pure system; a
4 4 4
premeasured amount of N 0 from a standard vessel was vacuum
2









A



QZZIG


C


JILl"


Figure 7. The photolysis set up: A, mercury drop; B, break seal; C, reaction cell;
D, glass tube to support the photolysis vessel; E, desk lamp fixture;
F, germicidal lamp; G, graded seal.









transferred into the photolysis vessel and sealed under vacuum. The purification of c-C H F has been described
4 4 4
in Chapter IV. N 0 obtained from Matheson Gas Company
2
was subjected to at least ten freeze-pump-thaw cycles to remove air before use; C.P. grade Matheson H was used as
2
received.

Sample Irradiation

To maintain a constant mercury vapor pressure in

mercury sensitization experiments, it is common to provide a small drop of liquid mercury in direct contact with the vapor phase. To obtain reproducible results, however, we found it necessary to make sure that the drop was in a small side arm, and not in the main reaction cell exposed to the full lamp intensity.

The source of 253.7 nm light was a General Electric 15 W Germicidal lamp (type GI5T8) which is essentially a low pressure mercury lamp in a high silica envelope, which transmits 253.7 nm and not 184.9 nm (Figure (7)). This lamp was mounted inside a standard desk lamp fixture equipped to hold two fluorescent tubes. The second lamp was replaced by a stack of Pyrex tubes which provided reproducible positioning of the sample vessel. To avoid intensity variations due to AC line voltage fluctuations, a Sola saturable core transformer was used to power the lamp. The photolysis arrangement was covered with a wooden eye shield and air was blown through this set-up to remove any ozone formed and to keep the temperature constant. The temperature was recorded using a









thermometer with its bulb resting inside the photolysis cavity. This experimental arrangement is essentially the same as the one described in detail by Dr. Arthur J. Frank of this laboratory.42 After photolysis, the sample was cooled down to liquid nitroqen temperature to prevent any dark reactions.

Product Identification and Analysis

The organic product identification was done using an on-line gas chromatograph-mass spectrometer combination. Hydrogen was measured by the conventional Toepler pump MeLeod gauge arrangement and the organic products by flame ionization gas chromatography. The details of identification and analyses are given in Chapter IV and Appendix I.

Preparatory to a wet-chemical analysis of HF, an extractant solution of 2 ml of 66% ethanol in water was introduced into the tubulation extending beyond the break-seal on the photolysis vessel (Figure (7)). With the cell contents frozen, the break seal was opened with a rod introduced through a piece of rubber tubing used as a gland. This procedure was necessary to avoid condensation of oxygen inside the vessel. The vessel was warmed to room temperature and the ethanolic solution was left in contact with the walls of the vessel for at least 30 minutes to extract all the fluoride ion into the solvent. This solution was transferred into a 5 ml polyethylene beaker, and the vessel was rinsed out into the beaker with another 2 ml of 66% ethanol.

The determination of HF was based on a previously described









electrochemical procedure; ,3 however, it was necessary to make a minor modification to permit analysis in the presence of mercury or mercuric ions. A drop of dilute aqueous NaCl was added to the fluoride ion solution prepared as described above.

To make sure that the added chloride ion was not interferring with the fluoride ion determination, a series of experiments using known amounts of NaF was carried out. (The Orion fluoride ion selective electrode, model 94-09, used in these experiments is claimed to have a selectivity of 1000:1 over chloride ions.44) In these verification experiments 40 pk of 5.00 x 10- 3M NaF was titrated against 5.00 x 10- 3M La(NO ) solution, in the presence of chloride ion and/or
33
mercuric ion and mercury. Figure (8) shows the titration curves for these analyses. In case of the titration in the presence of mercuric ions, the EMF changed from -0.052 V to

-0.003 V soon after the addition of mercuric ion indicating the complexation of fluoride ions with mercuric ions; the subsequent addition of chloride ion changed the EMF back to

-0.051 V, indicating the release of fluoride ion into the solution. These experiments show that the chloride ions (which can be used to release fluoride ions from mercuric ions and/or mercury) do not interfere with the determination of fluoride ions.

In all these experiments, the procedure prescribed by Hecke144 was followed. It was essential to condition the fluoride ion selective electrode by performing at least two




















0
0 0 +0.060 o 0 013
0
+0.040- 0 0


0

+0.0200

+0.000
> 0



-0. 02Q




-0.040




-0.060

0.0 5.0 10.0 15.0 20.0 25.0 30.0
-3
Volume in microliters of 5.0 x 10 M La(NO )
3 3

Figure 8. Titration curves for quantitative analysis of F- ion: 40 Ol of NaF + HG2+ ion + CI- ion, 40 pl of NaF + CI- ion, ; 40 p1 of NaF, solid line; photolysis, 0.










titrations using this electrode before it could be used for

a quantitative determination. During all the titrations,

the analyte was continuously stirred using a magnetic stirrer,

and the EMF was monitored using a high input impedance (10

Mega Ohm) Hickock digital Volt-Ohm meter.

Nitrogen formed in the photolysis of a mixture of N 0
2
and c-C 41 F was transferred from the photolysis vessel immersed in liquid nitrogen via a Toepler pump into a McLeod

gauge where it was measured.


Results


Quenching Cross Section Measurements and Actinometry

For quenching cross section measurements and actinometry

four mixtures with different pressure ratios of c-C H F to
4 4 4
N 20 (P c-C H F /PN O) were photolysed for 30 sec. In each
4 4 4 2
sample, the total pressure was 50 Torr. A plot of the reciprocal of the number of moles of nitrogen formed in each

mixture against the quantity (Pc-C H F /PN 0) gave a straight
L L 2
line as shown in Figure (9). The intercept of this line gives

the reciprocal of the number of moles of nitrogen formed in

pure N 0. The amount of nitrogen formed in the absence of
2
c-C Il F was found to be 9.597 pmoles. Assuming a quantum
4 4 4
yield of 0.8 for N production in mercury sensitized photol2
ysis of N 0,40 the absorbed light intensity was calculated
2
to be 2.4 x 1017 quanta/sec in the 91 cc vessel (i.e., 24

oEinsteins/min). The quenching cross section of c-C H F
4 4 4
was determined relative to that of nitrous oxide.4 0 The
























0.112 0.110


N 0.108


0.106


0.104





0.102

2.0 4.0 6.0 8.0 P c-C H F /P N 0
4 4 4 2

Figure 9. Modified Stern-Volmer plot; a plot of the reciprocal
of the number of moles of N2 against the ratio of the
pressure of c-C4114F4 to that of N20.










ratio, 1, of the relative quenching rates for c-C H F and
4 4 4
nitrous oxide was calculated from the slope of the line in Figure (9) to be


S= (slope) x (number of photons absorbed in 30 sec)/

Avagadro's number

S= 0.0102


The quenching cross section of c-C H F was calculated using
4 4 4
the equation40


o 2 1 + (Mg
c-C HF HgP4N0


c-a 2I 1 K (MH/MC I~
N0 Hg/c-c H F
2 L4 4 4

where G2 is the quenching cross section and M is the molecular

weight of the species shown by the subscripts. Assuming the

quenching cross section of N 0 determined by the chemical
2
method to be 18 A (Reference (40)), a 2 is calculated 'c-C H F
to be 0.27 A2. Recently Gleditsch and Michael45 have given

022
a value of 22.1 A as the preferred value for N 0* Using
2
this value of 1N O, the corresponding value of GcC H F is
2 4 4 14
0
0.31 A2.

Photolysis Products

The major products formed in the photolysis of c-C H F
4 4 4
at 50 Torr are HF, C H , C F H , two isomers of C H F (here2 6 3 4 2 6 4 6 after referred to as C 1 F (I) and C H F (II)), an unidentified
6 4 6 6 4 6
C compound, and C H F (a bicyclic compound); the minor prod8 8 6 8 ucts were H , CH , C F , 1,1- difluoroethylene (1,1-C H F
2 4 2 4 2 2 2









and C F 31. There were at least two more products whose
2 3
yields were not recorded since they were formed in very small quantities and their compositions were unknown. Based
19
upon the F NMR and mass spectral cracking pattern the probable structure of C H F (I) is as follows:
6 4 6

H 1- _ F



FF


The other isomer, C H F (II), is probably non-cyclic as indicated by its proton NMR. We could determine neither the structure nor the exact composition of the C compound.
8
Further discussions of the composition and structure of some of the photolysis products are given in Chapter IV and Appendix II.

A series of experiments was performed to investigate

the dependence of product yields on photolysis time in pure 1,1,2,2-tetrafluorocyclobutane. These experiments were carried out at both 50 and 150 Torr of the tetrafluorocyclobutane to reveal a possible pressure effect. The yields of C H and two isomers of C H{ F increased linearly with
2 6 6 4 6 irradiation time. The rest of the products were formed at a decreasing rate at longer irradiation times and many of the yields approached limiting values after two minutes of photolysis. Figures (10) - (14) show the yield-Dhotolysis time plots for all products at both pressures. Quantum yields for all products at 120 sec photolysis time for a 50




















0.20


1.5 0
oH
4D
0
4-4 0


4-4



1.0





0.)0. 0 1.I.0304. .
wx










Fiue1.Yedo H H t5 or �a 5 ot n 0.5







0.0

0.0 1.0 2.0 3.0 4.0 5.0

Photolysis time in minutes


Figure 10. Yield of HF ( 0 at 50 Torr; * at 150 Torr) and
H2 (0 at 50 Torr; 0 at 150 Torr) as a function
of time in the photolysis of c-C41tIF4.














6.0


0D
0
4.0



to
4i

U
04






2.





0.0 1







0.0 1.0 2.0 3.0 4.0 5.0 Photolysis time in minutes


Figure 11. Yield of C3112F4 ( 0 at 50 Torr; 9 at 150 Torr) and C2HF3 ( 0 at 50 Torr; n at 150 Torr) as a function of time in the photolysis of c-C4H4F6.





















0.10

C H F (I)
6 4 6



4J
o 0.08 ro
0

2c H F (I
0 6 4 6 E 0.06

w
0
0


0.04





0.02








0.0 1.0 2.0 3.0 4.0 5.0

Photolysis time in minutes


Figure 12. Yield of C6H4F (I) (0 at 50 Torr;� at 150 Torr)
and C6H4F6(II) (C at 50 Torr; 0 at 150 Torr) as a function of time in the photolysis of c-C41H4F4

























































1.0 2.0 3.0


4.0


12.0











8.0 0 (D 0



4:
o


4.0 0.0


5.0


Photolysis time in minutes


Figure 13. Yield of C2 11 ( [] at 50 Torr; X at 150 Torr) and
CH4 (0 at 50 Torr;@ at 150 Torr) as a function
of time in the photolysis of c-C 4H F4


2.0


1.0


0.0


0.0




























CO 6.0
1-iC 11 F
x8 6 8/150
~Torr
-.4

0
0 4
4-J 4.0
U




0
U24

CHF
0 C Compound
D8
8 5 Torr
-0 2.00
00





0 . 0 / . ,
0.0 1.0 2.0 3.0 4.0 5.0

Photolysis time in minutes


Figure 14. Yield of C8H6F8 ( 0 at 50 Torr; 0 at 150 Torr)
and the C. compound (0 at 50 Torr;1 at 150
Torr) as a function of photolysis time.









Torr sample are shown in Table VI. Since the absolute yields are quite small, a set of relative yield values normalized to HF yield = 1.00 are also given in the Table, in order to give a clearer picture of the behaviors of the system.

Since the yields of CH , the C compound and C H F at
4 8 8 6 8
150 Torr were different from those at 50 Torr, a series of 120 sec irradiations were carried out using pure c-C H F at
4 4 4
three more pressures ranging from 10 to 150 Torr. Yields of all the products except CH , the C compound and C H F were
4 8 8 6 8
independent of pressure above 30 Torr, within experimental error. The yields of methane decreased with increase in pressure while those of the C compound and C H F increased
B 8 6 8
linearly as shown in Figure (15).

To establish whether any of the products were due to

the reactions of hydrogen atoms, a mixture of 400 Torr of H
2
and 20 Torr of c-C H F was photolysed for 300 sec. All
4 4 4
the products formed in the pure system were formed except C F , C 11 F (II), the C compound and C If F The yields
2 4 6 4 6 8 8 6 8 of products formed in this experiment are given in Table VII.

The material balance in the photolysis is reasonably

good. The ratio of C/H/F is 4.00/3.65/4.48 when the photolysis has progressed for 120 sec. This shows a shortage of carbon and hydrogen. However, at shorter photolysis times the material balance is better, and at longer photolysis times worse. This suggests the formation of polymers or products which were not analyzable by the methods used in these experiments.










TABLE VI


Yields of Products Formed in a 120 Sec
Photolysis of 50 Torr of c-C H F
4 4 4


Product Yield at 120 sec Quantum Yield Relative
in p moles Yield


HF 0.091 0.0019 1.0

H 0.006 0.00012 0.063
2
CH 0.0016 0.00003 0.016
4
C H 0.0095 0.000197 0.10
2 6
C F 0.0015 0.00003 0.016
2 4
C H F 0.00092 0.00002 0.010
2 2 .2
C HF 0.00161 0.000033 0.017
2 3
c-C H F 0.0310 0.0006 0.34
3 2 4
C H F (I) 0.046 0.000954 0.50
6 4 6
C H F (II) 0.031 0.000643 0.34
6 4 6
C ? 0.009 0.000186 0.10
8
C H F 0.020 0.00041 0.22
8 6 8

















5.0 0






4.0


C H F
CO 8 6 8
C


4J 3.0

roO
0
tl-4

0
02 2.0 C compound a) 8
-.
0




1.0






0.0

0.0 50.0 100.0 150.0 Pressure in Torr


Figure 15. Plots of yields of C8H6F8 ( 0 ) and the C
compound ( 0 ) as a function of pressure in
the photolysis vessel.










TABLE VII


Yields of Products Formed in a 300 Sec Photolysis
of a Mixture of 400 Torr of H and 20 Torr of c-C H F
2 4 4 4



Product Amount in pmoles Relative Yield


HF 0.2348 1.0000

CH 0.0019 0.0082
4
C H 0.0125 0.0530
2 6
C H F 0.0021 0.0090
2 2 2
C HF 0.0308 0.1310
2 3
c-C H F 0.0025 0.0106
3 2 4
C H F (I) 0.0019 0.0082
6 4 6










Discussion


The possibility of direct photolysis of c-C H F by
4 4 4
253.7 nm light can be ignored since c-C H F has no absorp4 4 4
tion above 180.0 nm.46 Hence, all of the observed products

in these experiments must be due to the quenching of Hg6(3 P1)

to the 6( S0) ground state by c-C H F . (Although it is

possible to quench Hg6(3 P1) to the 6(3 P0) state, this process

has been shown to be unimportant in many systems47 and we

ignore it for lack of any direct evidence.) That c-C H F is
4 4 4
not very efficient in quenching the Hg6(3 P) is shown by its
02
relatively small value of 0.3 A2 for the quenching cross
0 2 Thi inficec
section (based on a value of 18 A2 for oN This inef iciency
2
in quenching is in accordance with the generalization, discussed by Cvetanovic, 40 that cycloalkanes are analogous to
0
alkanes in their low quenching efficiency. A value of 0.3 A2

for G2 is similar to the reported quenching cross
c-C H F
4 4 4
sections of three partially fluorinated alkanes - CH CH F
3 2
0 20 20
(0.46 A2), CH CHF (0.45 A2), and CH CF (0.45 A2).[4 As
3 2 3 3
expected this value for the quenching cross section of c-C H F

is considerably smaller than those of comparable alkanes or

cycloalkanes but larger than those of perfluoroalkanes (e.g.,

2 = 6.7 2 and a2 << 0.005 A c-C H A CF4
18
The primary photophysical event is shown in Equation III. (1)

and the subsequent quenching in Equation III. (2).


Hg6( S0) + hv (253.7 nm) >- Hg6(3PI) III.(i)









Hg6(3PI) + c-C H F 4 C H F +H- + Hg6( 1S0)
P1 4 4 4 3 4

III. (2)
It is suggested that the quenching of Hg6(3 P) by c-C H F
444
leads to the scission of a C-H bond (Reaction III.(2)). Rupture of the C-C bond is not suggested since we saw no evidence for it. Reaction III. (2) is suggested to be the major quenching process because of the following three reasons: (a) C-H bond rupture has been established to be a major quenching process in cycloalkanes ;3,40 (b) the products formed in photolysis of pure c-C H F and those in a mixture
4 4 4
of a large amount of hydrogen and a small amount of c-C H F
4 4 4
are similar, suggesting that hydrogen atoms are formed in the quenching process in this system; and (c) when a mixture of c-C H F and I was photolysed in the presence of mercury
4 4 4 2
vapor, all the products in the photolysis were eliminated, and a new product identified as C H F I (on the basis of
4 3 4
GC retention time and the mass spectral fragmentation pattern) was formed. These three observations argue very heavily in favor of Reaction III. (2) as the primary quenching process.

The two radicals C H F - and H- (formed in the primary
4 34
process) can react with each other or with the substrate gas leading to the formation of the observed photolysis products. We were able to study the simple reactions resulting from reactions of H- and fluorobutyl radicals in the absence of the more complex reactions resulting from mutual reactions of fluorobutyl radicals by photolyzing a mixture of H and
2
c-C H F in the presence of mercury vapor. These experiments
4 4 4









are discussed first, followed by consideration of the more complex pure system.

In the photolysis of a mixture of 400 Torr of hydrogen and 20 Torr of c-C H F , essentially all the excited mercury
4 4 4
atoms will be quenched by hydrogen since its quenching cross section (8.6 A2)47 and relative amount are each twenty times larger than those of c-C H F ; thus only about one quenching
4 4 4
event in 400 will involve the fluorocarbon directly.


Hg6(3 P) + H 2H- + Hg6(IS ) III.(3)


It is possible that this reaction proceeds via formation of HgH, although leading to the same final products. Since Hwould be the only reactive species formed in the quenching process, all the products seen in the mixture of H and
2
c-C H F must be due to the reactions of H- with c-C I F
4 44 4 4 4 Expected reaction of hydrogen atoms would be either an abstraction of H- from the substrate or an addition to substrate with simultaneous ring opening.


H" + cHC4H4F4 H 2 + c-C II F III.(4)


H" + c-C H F (CH CH CF CF )* III.(5)
4 4 4 3 2 2 2

Production of H- in Reaction III. (4) can not be observed directly due to the presence of excess hydrogen. However, results from the pure system discussed below suggest that Reaction III. (4) is considerably less efficient than Reaction III.(5).









We have postulated the addition of hydrogen atom to a carbon bonded to hydrogen, since in 1,1-difluoroethylene hydrogen atoms add almost exclusively at the nonfluorinated carbon end of the molecule,48 indicating that the addition to the methylene group is more favorable than to the perfluoromethylene group. Furthermore, the radical site in the product of Reaction III. (6) is expected to be the terminal CF
2
group, since fluorine atoms are known to stabilize radicals.9

Extended considerations of the observed product distribution requires that we postulate a substantial contribution from a third hydrogen atom reaction, namely abstraction of a fluorine atom from substrate.


H" + c-C H F > IF + c-C H F III. (6)
4 4 4 4 4 3

Although such a reaction is thermodynamically favorable in most cases, it is generally not observed, presumably due to unfavorable kinetic factors. The reaction should be especially favorable in a cyclobutane; however, the near - 900 ring angles prevent good sp3 hybridization, decreasing sigma bond energies, and enhance accessibility of both the carbon atom and substituents to attack.

The fates of the three radicals, c-C I F -, (CH CH CF CF )*
4 3 4 3 2 2 2
and c-C H F * (formed in Reactions III.(4) through III.(6))
4 4 3
determine the products formed and their relative distribution observed in the mixture. Since Reaction III. (5) involves formation of a primary C-H bond (ca. 98 kcal/mole17) and rupture of a weak C-C bond (61 kcal/mole according to our










measurements ) , the product radical will be excited by approximately 37 kcal, probably mostly in the form of vibrational energy. This amount of energy is easily sufficient to cause HF elimination, which is not more than 15 kcal endoergic:


(H CCH CF CF)* HF + CH CH = CFCF
3 2 2 2 3 2

III. (7)
The reaction probability is further enhanced by the fact that the product is a relatively stable vinyl radical. Reaction III. (7) would occur in competition with thermalization of the precursor species and we expect that it is a moderately important process.

The final fate of the radicals in Reaction III. (7) should be reaction with hydrogen atoms, which are present in much larger concentration than any other radicals in the system. The product (CH CH = CFCF H) would initially contain excess
3 2
energy due to C-H bond formation which could either lead to further fragmentation or be lost in thermalizing collisions. Although no product having the formula CH CHCFCF H was observed,
3 2
it might easily be masked by elution with the parent c-C H F 4 Since the measured product yields indicate a surplus of fluorine relative to carbon and hydrogen, formation of a moderate amount of this compound would improve the material balance.

The decomposition of (CH 3CH 2CF 2CF 2-)* to qive either a methyl or ethyl group is thermodynamically unfavorable









(endoergic by at least 30 kcal/mole for methyl radical formation and 21 kcal/mole for ethyl radical elimination). However, the product of H- and CH3CH2CF2CF2- (either excited or collisionally stabilized) certainly contains sufficient energy to undergo decomposition.


CH CH CF CF + H- > (CH CH CF CF H)* III.(8)
3 2 2 2 3 2 2 2

(H CCH CF CF H)* > CII + c-C H F III.(9)
3 2 2 2 4 32 4


(CH CH CF CF H)* C H - + C F H- III. (10)
3 2 2 2 2 5 2 4

Reactions III. (9) and III. (10) are very similar to those reported for the pyrolytic decomposition of n-butane.5s


n-C H > CH + C H III. (11)
41I0 4 3 6

and

n-C H 10 2C H 5III.(12)
4 10 / 2 5

Although the present work was done at room temperature, the energy made available by Reaction III. (8) should be sufficient to drive either Reaction III.(9) or Reaction III.(i0). Since Reaction III.(9) involves rupture of a C-C bond (33 kcal) and formation of a C-H bond (102 kcal) and a new C-C bond, the overall process is energetically very favorable. (If the heavier product in Reaction III. (9) is an open chain olefin, the second link in a C=C bond will be formed with an energy release of 63 kcal.)17 The amount of CH formed is almost
4
equal to that of C H F as expected according to Reaction
3 24
III.(9). The sequence of Reactions III.(8), III.(10) is










exothermic only to the extent that the C-H bond energy (ca. 98 kcal/mole) exceeds the energy of the C-C bond (83 kcal/mole). Nevertheless, both types of reactions are well established under high temperature conditions50 and appear plausible here as well.

The fate of C F H" (from Reaction III.(10)) should be
2 4
reaction with H', the most abundant radical in the mixture.


C 2F 4H- + H" > (C2 F H )* III. (13)
24 4 2

This energy rich ethane molecule could readily undergo HF elimination leading to the formation of C F H and HF:
2 3

(C F H )* C F H + HF 111.(14)
2 4 2 2 3


Having considered in some detail the fate of the open chain butyl radical formed in Reaction III.(5), it is now necessary to consider further reactions of cyclic butyl radicals formed by hydrogen and fluorine abstractions (Reactions III.(4) and III.(6), respectively),. In the presence of H. as the major steady-state species, the primary fate of c-C H F and c-C H F should be combination, fre434 443
quently followed by decomposition of the energy rich cyclobutanes.
: C H F + HF
4 3 3
III.(15a)


c-C H F + H'--- (c-C 11 F )*> 2C H F III.(15b)
43 4 4H4 4 2 2 2
M c-C H F III. (15c)
- 4 4 4










C H + C HF
2 4 2 3
H 11 Ill. (16a) HC-CF
c-C H F + H- (HC-C - C H F + C H F
F F/ III. (16b)

M
--L c-C H F III. (16c) 453

The relative probabilities of the above six reactions are hard to assess since two of the products (C H F and C H F
4 33 45 3
were not observed (since these would probably elute with the parent c-C4 H 4F in the gas chromatograph) and the olefins
44
would be labile under photolytic conditions. However, the relatively large yield of C HF indicates that Reaction III.
2 3
(16a) is efficient. One factor which determines the abundance of products formed through Reactions III. (15) and III. (16) is the amount of precursor radical formed in the mixture. Since the amount of C 2HF3 formed is relatively large in these experiments the concentration of c-C H F must be substantial.
4 43
Hence, we infer that a substantial fraction of the hydrogen atoms in the system undergo Reaction III. (6). Like all the other olefins in the system, C HF can undergo H. atom attack.
othr oefis n te sste, 2 3

Although C 2HF3 may be less reactive than some of the other olefins in the system, at least a portion of it should undergo H' addition followed by HF elimination to give C2HF2-.

F F F F
I I I I
C HF + H------ "C = CH--- "C = C + HF 111. (17)
2 3 1 1
F H{

The product C HF - formed in the above reaction could combine
2 2









with any of the radicals in the system. However, we postulate only Reaction III.(18) since C H F (I) is the
6 4 6
only such product of combination observed in the mixture.


F F
I I
c-C H F * + "C = C(-. C6H F6(I) III.(18)

H

With the above reaction scheme in mind, we proceed with the interpretations of results in the photolysis of pure c-C H F . The major difference between the pure system and
444
the mixture is the relative amounts of the radicals formed. As discussed earlier, H, is the predominant active species in the mixture. However, in the pure system the amount of H" formed initially is equal to that of C4 H 3F 4. Accordingly, most radicals combine with H- in the mixture, whereas in the pure system there is also a substantial probability of radical combination with c-C4 H 3F , leading to considerably greater production of high molecular weight species.

Since C H F 4 would be the predominant organic radical
434
in the pure system, a combination of two such radicals should be expected.


2C H F C H F III.(19)
434 868

This type of addition product of two rings has been observed in the mercury sensitized photolysis of both cyclobutane3 and perfluorocyclobutane, as well as in the radiolyses of both compounds.9''0 The C 11 F in Reaction III. (19) would
8 6 8










certainly be vibrationally excited and might need collisional stabilization.

In addition to C H F 8, another C8 compound is also
6 88

formed. Since its composition and structure are not known, it is inappropriate to propose a mechanism for its formation. However, we expect it to be a condensation product of C H F 4 with one of the many C4 radicals which are formed.

We propose that C6H4 F6(I) is formed by the same mechanism in the pure system as in the mixture. The large yield of C H F (I) in the pure system is not surprising since it is formed by reaction of C H F - which is present in larger
434
quantities in this case. A second compound with the same empirical formula but (probably) an open chain structure is C H F (II). One possible route for its formation would be
6 4 6
combination of the CH CH = CFCF � radical from Reaction III.
3 2
(7) with C F ", formed by HF elimination from C HF subse2 3 2 4
quent to Reaction III. (10).

It can be noted that the postulated mechanism calls for formation of a variety of olefinic products which were not observed. The olefins would be labile under photolytic conditions, and can undergopolymerization. In fact we did see some white solid on the walls of the vessel when a large amount of c-C H F was photolyzed for a long time. Also, we observed the formation of three products heavier than C compounds at longer irradiation times and higher pressures.

A final point of difference between the pure system and the mixture is the relatively large yield of C H F in the
3 2 4










former case; this is especially noteworthy since no other odd-carbon species is found in nearly as large a yield (34% relative to HF). It appears that a contribution by a route other than Reaction III. (9) is important in the pure system. One such route could be the decomposition of a C compound; two C products are formed with much higher
6 6
relative yields in the pure system than in the mixtures. A detailed explanation will not be proposed, however, due to lack of further evidence on this point.

In summary, we conclude that Hg6(3P1) reacts with

c-C 4H F to cause hydrogen atom production. The resulting free H- atoms react with tetrafluorocyclobutane by hydrogen atom abstraction, fluorine atom abstraction, and addition to the ring (causing ring opening). The radicals formed in these processes combine with each other or with hydrogen atoms; the resulting energy release frequently causes decomposition of the product molecule, such as elimination of HF or rupture of the C ring into two ethylenic compounds.
4
(Radical disproportionation reactions probably also occur but were not specifically identified.) As expected from the postulated mechanism, addition of I scavenger eliminates
2
all products observed in the pure system, and only c-C H F I
4 3 4
is detected.
















CHAPTER IV


GAMMA-RADIOLYSIS OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE IN THE GAS PHASE


Introduction


1,1,2,2-tetrafluorocyclobutane differs from both cyclobutane and perfluorocyclobutane in many physical and chemical properties. The thermodynamic stability of tetrafluorocyclobutane is between those of c-C H and c-C F .52 Although 48 48
there has been a considerable amount of work on the radiolysis of both c-C H 8,9 and c-C F , as well as a study of the
4 8 4 8
radiolysis of mixtures of the two compounds,1" there has been no work reported on a partially fluorinated cyclobutane. Tetrafluorocyclobutane was chosen for this study for several reasons, including the fact that it is empirically equivalent to an equimolar mixture of c-C H and c-C F
4 8 4 8
This chapter describes a study of the gas-phase gammaradiolysis of pure 1,1,2,2-tetrafluorocyclobutane, including the effect of added oxygen on the reactions of the system. To help understand the mechanisms involved in the radiolysis of this compound, parallel investigations of mercury sensitized photolysis and ion-molecule reactions in tetrafluorocyclobutane were also carried out.










Experimental


Reagents

The 1,1,2,2-tetrafluorocyclobutane used in these studies was obtained from Columbia Organic Chemicals Co., Columbia, South Carolina; it contained small amounts of impurities of which a C H compound was the most abundant. This
6 6
material was purified by a gas chromatographic technique. A glass column 0.5 m long, 3/8 inch in diameter, and packed with (Analabs) 60-200 mesh silica gel was connected between the storage vessel of c-C H F and a chemical high vacuum
4 4 4
line. After the column was evacuated overnight, the vessel was cooled in a dry ice-isopropanol bath and pumped on for an hour. The sample was then allowed to warm up to room temperature. With the arrival of the first traces of gas the pressure in the vacuum line increased. The first portion of the gas entering the vacuum line was discarded. The middle fraction of the purified c-C H F , which did not contain any
444
impurities that could be detected by gas chromatography, was collected for subsequent use.

Matheson Company C.P. grade ethylene was passed through a drying tube of barium oxide into a storage vessel on the vacuum line. Air was removed by repeated freeze-pump-thaw cycles. Matheson Company research grade oxygen was bled into the vacuum system through a 1/4 inch drying column of 60-200 mesh reagent grade silica gel. Sample Irradiation and Dosimetry

The irradiations of the samples were carried out using










a Cobalt-60 gamma irradiator which has been described previously.53 The dose rate in 400 Torr ethylene was determined by measuring hydrogen production, assuming G(I ) =1.2.4
2
Gases non-condensible at -196 0C were collected via a Toepler pump, measured on a McLeod gauge, and analyzed for methane and ethylene content on a flame ionization gas chromatograph using a silica gel column at room temperature. Hydrogen analysis was by difference. Energy absorbed in c-C H F was
4 4 4
assumed proportional to S, the electron stopping power per mole, 55r56 relative to ethylene, and S(c-C H F )/S(c-C H ) was
4 4 4 2 4
calculated to be 3.81. The dose rates were 2.14 x 1019 eV mole-I hr- in the metal vessel using a 230 Curie source and

5.59 x 1019 eV mole-1 hr- using a 600 Curie source. The dose rate in the glass vessel was 6.68 x 1021 eV mole-I hr-; only the 600 Curie source was utilized. Sample Preparation for Hydrogen and Organic Yield Measurements

The radiolysis vessels were made of pure nickel cylinders which were helium arc welded to Monel valves.29 For unscavenged radiolyses, c-C4H4F4 was directly metered into the nickel vessel which had been pumped down and flamed under vacuum to decompose any solid products formed in the previous radiolysis. Then the vessel was immersed in a Dewar containing liquid nitrogen and pumped on for five minutes to remove any air which might have leaked into the vessel during sample preparation. This procedure was essential to get reproducible yields for the olefins, since their yields are very sensitive to the presence of any scavenger. In the case of oxygen









scavenged radiolyses the sample was introduced as above, and 0 was transferred into the liquid nitrogen cooled
2
radiolysis vessel from a standard volume container filled with 0 at known pressure. By knowing the volume of the
2
vacuum line, the standard container, and the radiolysis vessel, the pressure of 0 in the radiolysis vessel at room
2
temperature was calculated. In several blank experiments, the amount of 0 metered out as described was checked using
2
a Toepler pump - McLeod gauge apparatus; the error was found to be less than 5%. In all the irradiations the pressure of c-C H F was 50 Torr, and the pressure of the added 0 was 4% of the total pressure in the vessel. Also,
2
a few semi-quantitative experiments were carried out with I 2
2
C H , or 1,1-C H F added to the system as scavengers.
2 4 22 2
Sample Preparation for HF Yield Measurements

For HF analyses a glass vessel equipped with a break

seal was cleaned and annealed at 565 C to pyrolyze any remaining organic residues. In the cases of unscavenged radiolyses required amounts of c-C H F were metered by pressure
4 4 4
measurements into a standard vessel of known volume and later transferred into the radiolysis vessel cooled in liquid N
2
The radiolysis vessel was sealed under vacuum.

In cases of 0 scavenged radiolyses, the c-C H F was
2 4 4 introduced as above, a premeasured amount of 0 from another
2
standard vessel was introduced into the vacuum line, and the vessel was sealed off. The amount of 0 introduced into the
2
vessel was measured in one blank run utilizing the procedure










described earlier in connection with the metal vessel. Product Identification

The formation of hydrogen fluoride in the radiolysis of c-C H F was established by potentiometric titration of
44
the irradiated sample using a fluoride ion selective electrode. Hydrogen was identified by its retention time on a 13-X molecular sieve column in a gas chromatograph equipped with a thermal conductivity detector. The organic products were identified by their mass spectral cracking pattern

4 2
using a gas chromatograph - mass spectrometer combination. When possible the identifications were confirmed by comparison of retention times with known standards and by matching the 5 7 19
mass spectra with spectra given in the API Tables.57 The F NMR spectrum of peak 21 recorded using a Varian XL-100 spectrometer equipped with NTC Fourier transform accessories. The spectrum was scanned 2000 times. In most of these identification procedures, products formed by a spark-discharge in c-C H F were used, since the yields of the products are larger than in the radiolysis and the tailing by the parent peak could be minimized. (In several previous researches in this laboratory it has been found that products formed in a very weak Tesla Coil discharge of organic gases are qualitatively and even semi-quantitatively similar to those formed during gamma radiolysis.)10'5' Standards were available for most of the lighter products analyzed on the silica gel column, but not for C H F or C H F ; there were no available stan342 324
dards for several of the products measured on the SE-30 column.









In such cases, product identification is somewhat tentative. Product Analyses

1) Hydrogen fluoride. Hydrogen fluoride was determined by potentiometric titration using a fluoride ion selective electrode. The details of the experiment are given in connection with the photolysis of tetrafluorocyclobutane (Chapter III). Addition of chloride ions to the analyte, necessary for the photolysis work due to the presence of mercury, was not required for the radiolysis samples.

2) Hydrogen. In the unscavenged system, hydrogen was measured as in the case of the dosimetry experiments. However, in the oxygen scavenged experiments, all gases noncondensible at -196� C were Toeplered into a McLeod gauge, and this mixture was analyzed for both the methane and hydrogen content using a gas chromatograph equipped with a flame ionization detector (FID) in tandem with a thermal conductivity detector (TCD). The FID was used for measuring the amount of methane and the TCD for that of hydrogen. The column used was a 13-X molecular sieve column mentioned above, and the carrier gas was nitrogen.

3) Low Molecular Weight Products. The low molecular

weight organic products (all compounds with one or two carbon atoms and one compound with three carbon atoms) were analyzed on a 4.6 m stainless steel column packed with 60-80 mesh (Analab) silica gel using a Tracor 550 research gas chromatograph equipped with a flame ionization detector.

4) High Molecular Weight Products. The high molecular









weight organic products (molecules with three or more carbon atoms) were analyzed on a 12 m glass column packed with Chromasorb-W coated with 30% SE-30. The same Tracor gas chromatograph was used. In the chromatographic analysis of both low and high molecular weight organic products, helium was used as the carrier gas. The molar response of all C and C compounds in a flame ionization detector were
1 2
obtained by using standards which were available. However, the molar response factors of the compounds were estimated from the response of similar compounds.


Results


In all, 24 products were identified and their G values

measured in the radiolysis of c-C H F . The inorganic species,
4 4 4
hydrogen fluoride and hydrogen,were the most abundant products. The yields of both hydrogen fluoride and hydrogen were linear with dose, within experimental error (Figures (16) and

(17), respectively). A typical gas chromatogram of the products eluting on the silica gel column is shown in Figure

(18). The three major olefins, 1,1-difluoroethylene (1,1C H F ), tetrafluoroethylene and ethylene are shown in Figures
2 2 2
(19) and (20). C H was produced in negligible quantities.
2 4
The yield of C F showed a small induction period and later
2 4
increased linearly with dose. The yield of 1,1-C II F was
2 2 2
small and the dose-yield plot did not extrapolate to zero yield at zero dose. However, an unirradiated sample did not show even a trace of 1,1-C H F . The yield-dose plot of
2 2 2






























1.5 *



0
,
o /






0
o4 1.0/ ro




H
0
0

0 0.5











0.0

0.0 1.0 2.0 3.0 Dose in eV x 1019



Figure 16. Production of hydrogen fluoride (pure, 0 ; 4% 02, 0 ) as a fundtion of dose in the radiolysis of c-C4 H4F 4.












0.40


0
ro 0.30


0
q l


o 0.20

0


0.10





0.00

0.0 1.0 2.0 3.0 19
Dose in eV x 10


Figure 17. Production of hydrogen (pure,] ;4% 0 ) as a function of dose in the radiolysis of c-C H F 2
4 4 4






































ELUTION TIME IN MINUTES


Figure 18.


Gas chromatogram of irradiated 1,1,2,2-tetrafluorocyclobutane: low molecular weight products eluting on a Silica Gel column.


1. CH
4
2. C l
2 6 3. C F
2 4 4. C H
2 4


5. 1,1-C H F
2 2 2
6. C H
2 2
7. C H F
2 3
8. 1,2-C H F (trans)
2 2 2


9. c-C H F
3 4 2
10. 1,2-C 11 F (cis)
2 2 2















0.15





-- 2 220


1, 1-C H F /0




-4
0
~In
0
0




0.05







1,1-C Il F 2 2 2



0.00

0.0 1.0 2.0 Dose in eV x 1019


Figure 19. Production of 1,1-C2H2F2 (pure Q ; 4% 02,
0 ) and C2H4 (4% O2,&) as a function of dose in the radiolysis of c-C H F
4 4
























0.05 .
U
44c 2F


0

0
f)
O4 2 42

'H










0.00
0.0 1.0 2.0 3.0

Dose in eV x 1019
0O
Figure 20. Production of C2F4 (pure U; 4% 02, ) as a function of
dose in the radiolysis of c-C4H4F4.










acetylene (Figure (21)) showed a distinctive break in slope, the yield increasing at a higher rate at longer doses. The G values of all the products eluting on the silica gel column are shown in Table VIII.

A chromatogram of the products eluting on the SE-30

column is shown in Figure (22). The first three peaks were made up of all the products analyzed on the silica gel column except C H F . Products 9 and 11 through 14 were identified to be C H F , C H F , C H F , C H F and C H F , respectively,
342 324 424 422 433
based on their mass spectral cracking patterns, obtained using the GC/MS combination as mentioned earlier. The details of the identification are given in Appendix I. Peak 15 was small and neither the composition nor structure of this product could be determined. Peak 16 is probably c-C H F . Peaks 424
18 and 19 were very small and in most of the runs smeared out by the tail of the parent peak. The 19F NMR of Peak 21 is shown in Figure (23). The proton NMR showed the presence of the olefinic hydrogen. Based on the NMR and the mass spectral data Peak 21 was assigned to be: H H F F
I I I I
H-Y - C = C
F-C- C-F H
I I
F F


Peak 22 is non-cyclic as shown by its proton NMR. A discussion of the structural determination of many of the other compounds is given in Appendix I.

The G values of all the products measured on the SE-30




Full Text

PAGE 1

THE PHOTOLYSIS, RADIOLYSIS, AND MASS SPECTROMETRY OF 1,1,2, 2-TETRAFLUOROCYCLOBUTANE By AKKIHEBBAL RAMAIAH RAVISHANKARA A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA

PAGE 2

g,S3Cil) di3'8®e/n‘

PAGE 3

ACKNOWLEDGEMENTS The author expresses his deepest gratitude to Dr. R.J. Hanrahan for his encouragement, advice, help and patience during the course of this work. He also thanks Dr. J.R. Eyler for providing access to the ICR instrument. Dr. W.S. Brey, Jr., for obtaining and interpreting the N.M.R. Spectra of some of the compounds, and Mr. R.J. Dugan for designing the circuit shown in Chapter II. Ill

PAGE 4

PREFACE The decomposition dynamics of cyclobutane and substituted cyclobutanes have been investigated by numerous workers under conditions of direct photolysis , ^ ^ mercury sensitized photolysis,^'** thermolysis,^"^ radiolysis , ^ ^ and electron bombardment in the mass spectrometer.^^ A point of special interest is the considerable ring strain in these compounds, causing a propensity to undergo ring opening and fragmentation into two ethylenic products. Even though the radiolysis of both cyclobutane and perf luorocyclobutane has been studied, no partially fluorinated cyclobutane has been investigated. In fact, even the mass spectral cracking patterns of partially fluorinated cyclobutanes have not been reported. However, a study of the decomposition of such a compound would be very interesting since it is expected to "bridge" the behavior of cyclobutane and perf luorocyclobutane . Hence, we undertook the study of the decomposition of 1 , 1 , 2 , 2— tetraf luorocyclobutane . This work describes a study of the gas-phase gammaradiolysis of 1, 1, 2 , 2-tetraf luorocyclobutane (Chapter IV). To help understand the energetics and mechanisms involved in the radiolysis of this compound three additional investigations were carried out. They are: IV

PAGE 5

(1) An electron impact investigation of 1, 1,2,2tetraf luorocyclobutane, * described in Chapter I. (2) Ion-molecule reactions in 1 , 1 , 2 , 2-tetraf luorocyclobutane, described in Chapter II. 3 (3) Mercury 6( P^) sensitized decomposition of 1, 1,2,2tetraf luorocyclobutane , described in Chapter III. *Chapter I is reproduced in part with permission from the Journal of Physical Chemistry , 79 , 876 (1975) . Copyright 1975 by the American Chemical Society. V

PAGE 6

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS iii PREFACE iv LIST OF TABLES viii LIST OF FIGURES ix ABSTRACT xi CHAPTER I AN ELECTRON IMPACT INVESTIGATION OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE 1 Introduction 1 Experimental 2 Results and Calculations 5 Discussion 17 CHAPTER II ION-MOLECULE REACTIONS IN 1,1,2,2TETRAFLUOROCYCLOBUTANE 2 3 Introduction 23 Experimental 2 3 Results 26 Discussion 35 CHAPTER III THE Hg6(^Pj_) PHOTOSENSITIZED DECOMPOSITION OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE.. 44 Introduction 44 Experimental Section 45 Sample Preparation 45 Sample Irradiation 47 VI

PAGE 7

CHAPTER III Continued Page Product Identification and Analysis 48 Results 51 Quenching Cross Section Measurements and Actinoraetry 51 Photolysis Products 53 Discussion 64 CHAPTER IV GAMHA-RADIOLYSIS OF 1 , 1 , 2 , 2-TETRAFLUOROCYCLOBUTANE IN THE GAS PHASE 7 5 Introduction 75 Experimental 76 Reagents 7 6 Sample Irradiation and Dosimetry.... 76 Sample Preparation for Hydrogen and Organic Yield Measurements 77 Sample Preparation for HF Yield Measurements 78 Product Identification 79 Product Analyses 80 Results 81 Material Balance 92 Discussion 96 APPENDIX I IDENTIFICATION OF PRODUCTS 109 APPENDIX II UNIMOLECULAR FRAGMENTATION OF CYCLOBUTANE COMPOUNDS 118 REFERENCES AND NOTES 12 3 BIOGRAPHICAL SKETCH 130 Vll

PAGE 8

LIST OF TABLES Table Page I Mass Spectral Cracking Pattern of C-C 4 H 4 F 4 at 70 V 8 II Appearance Potential Data 9 III Therraochemical Data 10 IV Rate Constants for Ion-Molecule Reactions in 1 . 1 . 2 . 2Tetraf luorocyclobutane 31 V Mass Spectral Fragmentation Patterns of 1 . 1 . 2 . 2Tetraf luorocyclobutane Under Analytical and Ion-Molecule Reaction Conditions 36 VI Yields of Products Formed in a 120 Sec Photolysis of 50 Torr of c-C H F 61 4 4 4 VII Yields of Products Formed in a 300 Sec Photolysis of a Mixture of 400 Torr of H 2 and 20 Torr of c-C H F 63 444 VIII G Values in the Radiolysis of 1 , 1 , 2-2-Tetraf luorocyclobutane : Products Eluting on the Silica Gel Column 89 IX G Values in the Radiolysis of 1 , 1 , 2 , 2-Tetraf luorocyclobutane : Products Eluting on the SE-30 Column 94 X G Values of Products in I 2 Scavenged Radiolysis of 1 , 1 , 2 , 2-Tetraf luorocyclobutane 95 XIa Mass Spectral Cracking Patterns of Products... 115 Xlb Mass Spectral Cracking Patterns of Products... 117 viii

PAGE 9

LIST OF FIGURES Figure Page 1 R.P.D. curves for Cut from CH4, ut from N2, and CsHitF"^ from C-C4H4F4, measured on a Bendix T.O.F. mass spectrometer 6 2 R.P.D. curves for C 2 H 2 Ft, C 2 Pt, and C 2 H 4 from C-C 4 H 4 F 4 , measured on a Bendix T.O.F. mass spectrometer 7 3 Circuit diagram of the pulse generator for ion grid Number 1 25 4 Normalized ion intensities as a function of delay time in the T.O.F. spectrum of 1, 1,2,2tetraf luorocyclobutane at 26.8 microns and 50° C 28 5 Normalized ion intensities as a function of delay time in the T.O.F. spectrum of 1, 1,2,2tetraf luorocyclobutane at 53.0 microns and 50° C 29 6 Normalized ion intensities as a function of pressure in the T.O.F. spectrum of 1, 1,2,2tetraf luorocyclobutane at 25° C 32 7 The photolysis set up 46 8 Titration curves for quantitative analysis of F ion 50 9 Modified Stern-Volmer plot; a plot of the reciprocal of the number of moles of N 2 against the ratio of the pressure of C-C 4 H 4 F 4 to that of N 2 O 52 10 Yield of HF and H 2 as a function of time in the photolysis of C-C 4 H 4 F 4 55 11 Yield of C 3 H 2 F 4 and C 2 HF 3 as a function of time in the photolysis of C-C 4 H 4 F 6 56 IX

PAGE 10

LIST OF FIGURES Continued Figure Page 12 Yield of C6Hi+F6(I) and C6H4F6(II) as a function of time in the photolysis of C-C4H4F4 57 13 Yield of C2H6 and CH4 as a function of time in the photolysis of C-C4H4F4 58 14 Yield of CsHsFs and the Cs compound as a function of photolysis time 59 15 Plots of yields of CsHsFe and the Ca compound as a function of pressure in the photolysis vessel 52 16 Production of hydrogen fluoride as a function of dose in the radiolysis of C-C4H4F4 82 17 Production of hydrogen as a function of dose in the radiolysis of C-C4H4F4 83 18 Gas chromatogram of irradiated 1 , 1 , 2 , 2 -tetraf luorocyclobutane : low molecular weight products eluting on a Silica Gel column 84 19 Production of 1,1-C2H2F2 and C2H4 as a function of dose in the radiolysis of C-C4H4F4 85 20 Production of C2F4 as a function of dose in the radiolysis of C-C4H4F4 86 21 Production of C2H2 as a function of dose in the radiolysis of C-C4H4F4 88 22 Gas chromatogram of irradiated 1 , 1 , 2 , 2 -tetraf luorocyclobutane : high molecular weight products eluting on a SE -30 column 90 19 23 F NMR spectrum of product Number 21 , C6H4F6.... 91 24 Production of C4H3F3 and C6H4F6 as a function of dose in the radiolysis of C-C4H4F4 93 X

PAGE 11

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE PHOTOLYSIS, RADIOLYSIS, AND MASS SPECTROMETRY OF 1,1,2,2-TETRAFLUOROCYCLOBUTANE By Akkihebbal Ramaiah Ravishankara December, 1975 Chairman: Dr. R.J. Hanrahan Major Department: Chemistry The fragmentation of 1 , 1 , 2 , 2-tetraf luorocyclobutane under electron bombardment was investigated using a Bendix Time-of-Flight mass spectrometer. Using the Fox Retarding Potential Difference technique, measurements were made of the appearance potentials of major ionic fragments including C H f'^ (12.15 V), C f"^ (12.60 V), C H^ (13.15 V), and C H f"^ 222 24 24 34 (12.85 V). From these results it is found that AH^ for the C H F ion is < +207 kcal/mole and AH^ for the parent c-C H F 3 4 — f 4 4 is ^ -202 kcal/mole. The symmetrical rupture of the molecule into C H F plus C H F is a sequential process with the sum 222 222 of the energy of breaking two nonequivalent C-C bonds equal to 126 kcal/mole, while the nonsymmetrical rupture into C h"^ 2 4 and C F is concerted, 2 4 the two equivalent C-C 61 kcal/mole. and the energy of breaking each of bonds between CF and CH groups is 2 2 4 XI

PAGE 12

To obtain kinotic data on the ion— molecule reactions in 1 , 1 , 2 , 2-tetraf luorocyclobutane , measurements of the variation of the ion intensities with the changes in delay time were carried out using a Bendix Time-of-Flight mass spectrometer. The parent ion C H had very low abundance 4 ^ 1 * and was found to be nearly non-reactive. However, ionmolecule reactions were observed for the olefinic ions C H F^ 2 2 2 and C H • It was found that C H F becomes the most abundant 2 4 2 2 2 ion in this system when the pressure is increased. Using the ICDR technique, it was established that reaction of C H F with parent c-C H F produces the species C H f"*” and 2 2 2 4 4 4 6 6 6 C H F"^. 5 6 4 3 The Hg6 ( photosensitized decomposition of 1, 1,2,2tetraf luorocyclobutane (c-C H F ) was carried out by irradiat4 4 4 ing a mixture (50 or 150 Torr) of c-C H F and mercury vapor 4 4 4 with 253.7 nm radiation from a low pressure mercury lamp. The photolysis of a mixture of hydrogen and a small amount of c-C H F was also investigated to establish the reactions of 4 4 4 hydrogen atoms with c-C H F . Using the Chemical method 4 4 4 of Cvetanovic, the quenching cross section of c-C H F 4 4 4 o was determined to be 0 . 3 A relative to that of N 0. 2 Major products in the photolysis of the pure system include HF, two isomers of C H F , a C compound, and C H F ; 6 4 6 8 8 6 8 the yields of these species and six others, formed in moderate quantities, were followed as a function of irradiation time. It is suggested that the rupture of the C-H bond is the major primary process in the chemical quenching of Hg6(^P^) Xll

PAGE 13

1 / 1 f 2 , 2-tetraf luorocyclobutane . A reaction scheme involving and H* (formed in the primary quenching process), is proposed to explain the formation of the observed products. The gamma-radiolysis of gaseous 1 , 1 , 2 , 2-tetraf luorocyclobutane was studied at 50 Torr pressure and 24°C, both pure and with added oxygen. The major products were HF, H , 2 1.1C H F , C F , C H , and C H F ; C H was seen only in 2222422 43324 the presence of oxygen. The addition of approximately 4% oxygen dramatically increased the G values of two olefins, 1.1C^H^F^ (from 0.046 to 0.47) and (from ~0 to 0.125), decreased the yield of H , C F and C H . Under the same 2 2 4 2 2 conditions most heavier products were either eliminated or considerably reduced. The yield of C H F , however, was 4 3 3 not changed substantially. Addition of both C H and I as 2 4 2 scavengers decreased the yield of HF by 25%. The observations are interpreted in terms of reactions of both ions (especially C H f"*^ and F ) and neutral species. 2 2 2 Xlll

PAGE 14

CHAPTER I AN ELECTRON IMPACT INVESTIGATION OF 1 , 1 , 2 , 2-TETRAFLUOROCYCLOBUTANE Introduction Information on the mass spectral cracking pattern of a compound is very helpful to an understanding of its radiolytic behavior, since slow electrons are responsible for the decomposition in both the cases. Additionally, bond energy data are very useful in interpreting the decomposition dynamics of any compound. One reliable and accurate method for obtaining bond energy information is the measurement of appearance potentials of fragment ions of a compound in a suitably equipped mass spectrometer using Retarding Potential (R.P.D.) techniques. The first goal of the present investigation was to obtain the mass spectral cracking pattern and bond dissociation energies in 1 , 1 , 2 , 2— tetraf luorocyclo— butane (subsequently referred to as c-C H F ) II H since these data were not available. Secondly, it has been proposed that the c-C H F ion F F 4 4 4 1

PAGE 15

2 constitutes the transition complex in ion-molecule reactions of C H with C H F , C with C H , and C h'*' with C F . 222 2222M 24 24 24 It was postulated that the complex is very short-lived and during its existance maintains its cyclic structure. Accordingly, a study of the fragmentation of this compound under electron impact is an obvious experiment to shed light on whether the predictions of the model are right and assumptions plausible . In the present work, the appearance potentials of C f’*’, 2 4 “f" "f* C H , C II F , and C H F from c-C H F were measured. From 24222 34 444 these data we can calculate some of the bond dissociation energies in this compound and estimate some others. The experiments were done on a Time-of-Flight mass spectrometer using standard R.P.D. techniques . ^ ^ ^ ^ Experimental The mass spectral cracking pattern of c-C H F was 444 obtained on the Bendix Time-of-Flight mass spectrometer at 70 eV. The R.P.D. experiments were also done on the Bendix, as described by Melton and Hamill*® with some minor changes. Potentials on the five grid electron gun were as follows: No. 1 grid (electron control grid, nearest to the filament) , -6 V with a +12 V pulse of 3 y sec duration repeated at a frequency of 10 kHz; No. 2 grid, +0.01 V; No. 3 grid (Fox retarding potential difference grid), -1.000 V with an intermittent AV of -0.200 V; No. 4 grid, +0.01 V; and No. 5 grid was grounded to the ion source structure (and therefore

PAGE 16

3 to the frame of the instrument) . Potentials on grids 1-4 are referenced to the electron filament, and hence their absolute value relative to ground varies along with the filament. The choice of -1.000 V for the Fox grid and -0.200 as AV was made after preliminary optimization experiments to obtain the best compromise of ion current I and its variation Al in response to AV applied to the R.P.D. grid. The purpose of the R.P.D. procedure is to measure an amount of ion current Ai attributable to electrons in a narrow voltage range AV taken as a "slice" out of the much broader thermal electron energy distribution from the filament. The ion focus pulse was triggered 0.7 ysec after the control grid pulse; we found this measure essential to obtaining reliable results. (It is possible to use the "timelag-focusing" control for this purpose, on Bendix instruments so equipped.) The potential on the electron collecting trap was reduced to +8 V to minimize extraneous additional ionization of the experimental compound in the target region. Several minor modifications to the control electronics of the Bendix were made as recommended by Melton and Hamill.^® The ion current signal from the electrometer circuit in the "scanner" unit was fed to a potentiometric chart recorder through a potential divider. The compound under study was contained in a 5 S. vessel and was leaked into the ionization chamber through a gold leak. The pressures in the ionization chamber could be changed by changing the pressure in the 5 Jl. vessel. The usual operating pressure

PAGE 17

in the ionization chamber was 6-8 x 10 ^ Torr . The normal trap current was 6 pA and the change in trap current was less than 5%. (The trap current was not regulated.) The electron beam accelerating potential, which was 3-Pplisd to the filament, was generated by standard circuitry in the Bendix and monitored on the 40-V scale of a Digitec Model Z-204-B digital volt meter (resolution 0.005 V, accuracy 0.01 V). However, due to contact potentials, surface charges, etc., it can not be assumed that the effective accelerating potential of the electron gun is identical with that measured externally. Accordingly, an R.P.D. curve was measured for Ar ion from argon gas and the apparent appearance potential of Ar (corresponding to the first break.) was determined from a graph made against the experimental voltage scale. The difference between the measured value and the accepted literature value gave the necessary correction for the displacement of the voltage scale. Using the scale calibrated as described, cross checks were done by measuring the appearance potentials of (from C F ) and N (from N ) . Our measured values z 4 2 4 2 2 agreed with the literature to within 0.1 V, indicating that our voltage scale is linear at least within the approximately 6-V range tested. In later experiments C f'*’ and n'*’ were used 2 4 2 as secondary calibration standards. We could reproduce exactly the shapes of ionization efficiency curves for Ch'^ from CH and C h'^ , C H^, and C from 4 2624 25 *“ 2 ^ 6 ' were reported by Melton and Hamill;^^ all reported "break points" agreed within a 0 . 1-V range. (The

PAGE 18

5 CH^ curve is shown in Figure 1.) Although we always had a inoderate air background, N (28) does not interfere with 2 C H (28) as the appearance potential of n'*’ is 15.6 V and + ^ has zero intensity below about 15.1 V (v/ell above the region of interest for C h"*") . in the spectrum of c-C H F , 2 4 4 4 4 the appearance potential of CH could not be carried out 2 because of the interference by the CH^ in the background and that of CF^ and C^H^F^ could not be done because of very low intensities. The 1, 1, 2 ,2-tetrafluorocyclobutane was obtained from Columbia Organic Chemicals Co. It was dried and deaerated on a vacuum line through several freeze-pump-thaw cycles at liquid nitrogen and Dry Ice temperatures, and vacuum distilled into a glass bulb for transfer to the mass spectrometer. Results and Calculations The Retarding Potential Difference curves which we obtained for C H F^, C H F^, C F+ , and C from c-C H F are 34 222 24 24 444 shown in Figures (1) and (2) along with as a standard and 2 CH^ for comparison with the work of Hamill and Melton.^® The mass spectral fragmentation pattern which we obtained for c-C H t is shown in Table I. In Table II we present the 4 4 4 appearance potential values which were found for the several major product ions from c-C H F , along with values for 4 4 4 certain other pertinent species from the literature. From the appearance potential values given in Table II, along with the standard enthalpy of formation for several species from the literature (given in Table III), it is

PAGE 19

AI (Arbitrary Units) 6 12 13 14 15 16 17 eV Figure 1. R.P.D. curves for CHÂ’Jj' from CH^^, from N 2 , and C 3 H 4 F+ from c-C 4 H^F^, measured on a Bendix T.O.F. mass spectrometer.

PAGE 20

AI (Arbitrary Units) 7 eV Figure 2. R.P.D. curves for CzHzPt, CzFt, and Cjut from C-C4H4F4, measured on a Bendix T.O.F. mass spectrometer.

PAGE 21

Mass 14 16 26 27 28 31 32 39 40 44 45 50 51 8 TABLE I Mass Spectral Cracking Pattern of c-C H F at 70 V. 4 4 4 Intensity Assignment Mass Intensity Assignment 18 CH 57 7 CHF 2 3 2 7 CH 59 16 CHF 4 3 4 9 C H 64 100 CHF 2 2 2 2 2 11 C H 69 5 CF or C H : 2 3 3 4 2 23 C H 75 4 C HF 2 4 3 2 13 CF 77 9 CHF 3 3 2 37 CHF 89 16 CHF 4 3 2 9 C H 100 23 C F 3 3 2 4 30 C H 109 9 CHF 3 4 4 4 3 12 C HF 113 3 C HF 2 3 4 18 CHF 127 2 CHF 2 2 4 3 4 6 CF 128 3 CHF 2 4 4 4 16 CHF 2

PAGE 22

9 TABLE II Appearance Potential Data Ion Reactant Product AP c f'*Â’ 2 4 c-C H F 4 4 4 c f"^ + 2 4 C H 2 4 12.60 c 2 4 C-C H F 4 4 4 c + 2 4 C F 2 4 13.15 C H f'*' 2 2 2 c-C H F 4 4 4 C H f"^ + 2 2 2 C H F 2 2 2 12.15 C H f'*' 3 4 c-C H F 4 4 4 C H f"^ 3 4 + CF 3 12.85 c h'*' 2 4 C H 2 6 c + 2 4 H + H 16.3 c h'*' 2 4 C H 2 4 10.48 C H f'*' 2 2 2 H C-CF 3 3 C H f"^ 2 2 2 + HF 10.30 ^This work Melton and Hamill, Reference I. (16) Q NSRDS complilation. Reference I. (18)

PAGE 23

10 TABLE III Thermochemical Data Ion or Radical in Kcal/mole Reference C F 2 4 78 18 C H 2 4 12.49 18 C H 2 4 253 18 C F 2 4 -155.5 18 C H F 2 2 2 159 18 C H F 2 2 2 -78.6 18 CF 3 112.6 19

PAGE 24

ii possible to calculate heats of formation of the ion C H F"*” 3 4 the neutral parent molecule c~C H F , as well as values 4 4 4 for the energy of dissociation of C-C bonds in the C ring, 4 The heat of formation of the parent molecule can be formulated in three different ways. First, using our measurements of the appearance potential of C F'*' we obtain: 2 4 c-C H F 4 4 4 -V CF + CH +e' ^ 2 4 2 4 I. (la) A(C F ) > AH° (C f"^) + AH° (C H ) AH° (c-C H F ) 24 r 24 r 24 f 4 4 4 I. (lb) Therefore tO AH^ (c-C H F ) > AH? (C f'^) + Ah? (C H ) A(C f'^) r 444 — t 24 f 24 24 I. (Ic) AH° (c-C H F ) > 78 kcal/mole + 12.49 kcal/mole 12.60 eV I. (Id) AH_ (c-C H F ) > -200 kcal/mole r 4 4 4 — I. (le) Next, based on the appearance potential of C h"^ we can 2 4 calculate : c-C H F 4 4 4 C H"^ + C F + e' 2 4 2 4 I. (2a) 1 (c: h"^) + AH? (C F ) AH? (c-C H F ) 2 4 2 4 2 4 4 4 4 I. (2b) Therefore tO ah (c-C H F ) > ah? (C ll'^) + AH? (C F ) A(C H+) 4 444 — r 24 t 24 2 4 , O I • (2c) AH^ (c-C H F ) > 253 kcal/mole + (-155.5 kcal/mole 4 U U — ' 4 4 4 13.1 eV AH° (c-C H F ) > -206 kcal/mole L 4 4 4 — 4 4 4 I. (2d) I. (2e)

PAGE 25

12 Finally, the measured appearance potential of C H f'*' gives 2 2 2 c-C H F 4 4 4 >CHF^+CHF +e ^ 2 2 2 2 2 2 I. (3a) A(C II F^) > AH° (C H f"^) + AH° (C H F ) 222 f 222 f 2 2 2 AH° (c-C H F ) r 4 4 4 I. (3b) Therefore AH° (c-C II F ) r 4 4 4 AH° (c-C H F ) r 4 4 4 ^ AII° (c H f"^) + AH? (C H F ) A(C H f'*') ^ 222 f 222 222 I. (3c) ^ 159 kcal/raole + (-79 kcal/mole) -12.15 eV AH^ (c-C^H^F^) ^ -200 kcal/mole I. (3d) I. (3e) Accordingly, the average value for AH° (c-C H F ) is > f 4 4 4 — -202 kcal/mole. The satisfactory agreement between the three calculations is gratifying because the calculations constitute a rather severe test of the consistency of the published thermochemical data as well as our appearance potential measurements . Using our value of -202 kcal/mole for the heat of formation of c-C H F and the appearance potential of C H f"^, we can calculate a value for the heat of formation of this ion as follows: k + CF^+ eI. (4a) A(C H F"^) > 4H° (C H F'*') + AhJ (CF ) AH? (c-C H F ) 3“* t34 f 3 f 44 4^ I. (4b) Therefore

PAGE 26

IJ AH AH AH o f o f o f (C H f"^) < A(C H F^) AH° (CF-) + AH° 3 4 — 3 4 f 3 ' f (C^H^F ) 12 . 85 eV (-112 . 6 kcal/raole) + kcal/raole) (C H F ) < +207 kcal/mole 3 4 — (c-C H F ) 4 4 4 I. (4c) (-202 I. (4d) I. (4e) In order to consider the energy of breaking C-C bonds in the c-C^H_^F^ ring, it is necessary to define our notation. First, we let D^,^^p^bethe bond dissociation energy of the bonds shown below. (Note that represents a C-C dissociation energy of a bond where "A" and "B" are substituents on the carbon atoms.) F I C — F I ^ C — H I H °(F/F) °(H/H) °(F/H) We also let and represent the energies for the reverse of the following processes: H H \ C-C \ H H -AH (=H/H) H H \ C=C \ H H I. (5) H F / \ H F -AH (-H/F) > H F I. (6) In order to determine the enthalpy change in these reactions.

PAGE 27

14 useful energy cycles can be written as follows: ? T ^°C-H H — C — C — H ^ -C-C H rf H H + H + H H H H -AH (=H/H) I. (7) I. (8) ^ H I + \ '(C hVc H ) C=C h'" + e' 2 4 and H F H— C — C— F T I H F °C-H °C-F I. (9) '^H-F ^ ^C-C'^ + HF I. (10) H Y -Ah (=f/h) I. (11) ,+ \ ^(C H F"/C H F ) V=P 2 2 2 2 2 2 V / \ ^ > H F C H f’*’ + e' 2 2 2 I. (12) where D is the C-H bond dissociation energy in C H , or 2 6 °r-F bond dissociation in CH CF , and D is the II— F bond dissociation energy (equal to the bond energy in this case) . We will use ^ (q \ to indicate the ionization potential (or appearance potential) of C h’*’ from 2 4

PAGE 28

15 and similarly for other daughter ion/parent molecule pairs which are mentioned subseguently . Then we have ,+ A(C H /C H ) = 2D C-H AH (=H/H) + I (C H VC H ) I. (13) A(C H F^/C H F 2 2 2 2 3 3 °C-H + °C-F °H-F " ^^^(=F/n) + I (C H F Vc H F ) 2 2 2 2 2 2 I. (14) Therefore, application of the energy cycle (I . ( 7 ) -I . ( 9 ) ) to Reaction I. (5) gives ^^(=H/H) = + I(C^hVc^H^) A(C^nyc^Hj 1.(15) ^^^(=H/H) 2(4.16) + 10.48 16-3 = 2.5 eV 1.(16) A similar calculation for Reaction 1.(6) using the energy cycle (I. (lO)-I. (12)) gives : ^^^=H/F) °C-H ^ °C-F °H-F ^ ^^2^2^2^^2^2^2 ^ -A(C H fVc H F ) 2 2 2 2 3 3 = (4.16 + 4.6 5.82 + 10.12 11.2) 1.86 eV I. (17) I. (18) It would be of interest to calculate the guantitv AH (=F/F) for formation of tetraf luoroethylene from the corresponding diradical, but the necessary data are not accessable C f"*" 2 4 has zero abundance from C F and C F H.^° 2 6 2 5 * We next consider the energetics of the unsymmetrical splitting of c-C H F into C h"^ and C F : 4 4 4 2 4 2 4 c-C H F ^ C h"^ + C F + e“ 4 4 4 2 4 2 4 I. (19)

PAGE 29

16 It xs possible to make two reasonable, limiting assumptions about the energetics of this process (and the similar symmetrical split into C II F+ and C H F ) : either the re2 2 2 2 2 2 organization of the neutral fragment into a double-bonded species occurs before the split from the ion, or else it occurs afterwards. If the latter is the case, then the corresponding reorganization energy is not available to decrease the energy of formation of the observed ionic fragment : 2 4 4 4 4 A(C_H /c-C H F ) 2 d Ah, „ + I (C H^/C H ) 4 I. (20) 2 4 2 4 (F/H) ““(=H/H) This assumption leads to the following evaluation of D (F/H) ' the energy per bond to break the CF . . . CH carbon-carbon 2 2 bonds in the c-C H F ring: 4 4 4 2D 2D (F/H) “ ^^(=H/H) " A(C hVc-C H F ) <^*+^4 2 4 444 (F/H) 2.5 eV10.48 eV+ 13.25 eV I. (21) I. (22) and ^(F /H ) 2.64 eV — 61 kcal/mole j _ (23) The magnitude of the resulting bond dissociation energy appears reasonable; the significance of this result will be considered further in the Discussion section. In the case of the symmetrical dissociation process, the alternative limiting assumption appears to give reasonable results :

PAGE 30

17 c-C H F '4 4 4 C H f'*’ + c H F 2 2 2 2 2 2 I. (24) Assuming that all of the reorganization energy is available to drive the process, we obtain: A(C H F /c-C H F 2 2 2 4 4 4 + °(F/F)^ °(H/H) + I (C H F /C H F ) 2 2 2 2 2 2 I. (25) which rearranges to give “(F/F) ’(H/H) = I (C H F'^/C H F ) 2 2 2 2 2 2 ^(F/F) ^ ^(H/H) ~ 12.05 eV+ 2(1.86 eV) 10.30 eV ^(F/F) ^ ^(H/H) ~ 5.47 eV = 126 kcal/mole I. (26) I. (27) I. (28) Accordingly, the average of these two bond dissociation energies is the reasonable value of 63 kcal/mole. This result will also be considered in the Discussion. In order to carry out a calculation for the third dissociation process, into C F and C H , we would need the 2 4 2 4 energy of a tetraf luoroethylene diradical rearranging to tetraf luoroethylene . As noted above, we were not able to calculate this quantity. Discussion The overall mass spectral fragmentation pattern of is consistent with what might be expected for a saturated, cyclic compound. Most compounds of this type

PAGE 31

18 give a small parent peak along with considerable fragmentation into lower molecular weight ions. Furthermore, Jennings has postulated that the cyclic C H f'^ ion, formed as an 4 4 4 intermediate complex in several ion— molecule reactions , has a very short lifetime. It appears reasonable to suggest that the structure of the c-C H f"*" system reached upon electron 4 4 4 bombardment of the stable molecule is the same one formed in the ion-molecule reactions, although there could be differences in internal energies (especially vibrational energy) . VJe note that c— C H F shows a small abundance of the parent ion, / this is consistent with the fact that Jennings also saw a small amount of c-C H F^ at high pressures (presumably, involving some collisional stabilization in the ionmolecule experiments) . As can be seen from the cracking pattern of c-C H F , 4 4 4 there is no peak corresponding to C H f’*’ or C HF^. This fact 2 3 2 3 lends some credibility to the assumption made by Jennings that the transition complex formed in the systems [1,1-CF CH j"*” 2 2 ori 1,1-C H F , C H on C F , and C F on C H maintains a 222 24 24 24 24 stable structure and does not scramble F and H atoms during its brief existence. We note a number of species which must involve some rearrangement, however, such as CHF*^. It is our suggestion that most (and perhaps all) such rearrangement occurs in connection with ring opening, which produces unsaturated carbon centers. We calculate the heat of formation of c-C H f'*’ to be 4 4 3 about 190 kcal/mole by Franklin's group contribution method.

PAGE 32

i9 This value, when combined with our value for AH° of c-C H F f 444 ' gives 14.4 V for the appearance potential of C H f’*’ if we 443 assume concurrent elimination of F , and a value of 18.0 V if we assume F atom elimination. We find an approximate appearance potential of c-C_^H^F^ of 13.5 + 1 V. Although our data were not of high quality, due to the low abundance of the ion, its appearance potential is certainly not higher than 15.0 V. Accordingly, we conclude that c-C H F'*' is formed 443 by F elimination. The two modes of splitting the cyclobutane ring into an ethylenic ion and an ethylenic molecule (Reactions I. (19) and 1.(24)) clearly must proceed by two different mechanisms. For the unsymmetrical break (Reaction I. (19) ) we utilized in the calculation only the reorganization energy of • CH CH • into 2 2 C H , and obtained 61 kcal/mole each for breaking two CF ... 2 CH^‘ bonds. If we had also utilized the reorganization energy of -CF^CF^into CF^=CF^ the resulting bond dissociation energies would increase by at least 15 to 20 kcal each, which would surely be too high for the badly strained c-C H F 444 molecule . On the other hand, we utilized the reorganization energy for two -CF CH • fragments into CF =CH , in order to inter22 22 pret the data for the symmetric split. We obtained 126 kcal ^(F/F) ^(H/H) ' average of 63 kcal/mole for dissociation of CF . . . CF and CH . . .CH , which is of a reasonable 22 22 magnitude. (If we apportion the two energies similar to the known dissociation energies of CF . . . CF and CH ...CH , the 3 3 3 3

PAGE 33

resulting values are 57 kcal/mole and 69 kcal/inole respec— Use of only one reorganization energy term would *9^^® sn impossibly low value of 46.5 kcal/mole for the average of and In fact, the postulate that anything much less than 100% of the two reorganization energy terms is available to drive Reaction I. (24) would lead to an unrealistically low result. Our specific suggestion for the mechanistic difference in Reactions 1.(19) and 1.(24) (which forms the foundation of the different energetics) is that the former is a one-step dissociation giving C and an energy-rich • CH CF • species 2 4 2 2 (which later forms CF^=CF^ without aiding in the progress of the original dissociation) while Reaction I. (24) is seguential, proceeding through an open-chain intermediate. + + e~ ^ I. (24a) I. (24b) Consistent with the conclusions reached by Jennings et al., as well as our own observations, the open-chain species formed in Reaction I. (24a) would have to be short-lived (lO"^ sec or less). Atom-migration in the open-chain intermediate could explain occurrence of several additional species in the

PAGE 34

cracking pattern of c-C H F , in particular CF^ and C H F^. 3 3 4 * A similar ring-opening of C H f'*' followed by a split into 1-carbon and 3-carbon fragments would explain the occurrence of CFH (37%) and C^II^F^ (9%). Several other similar decomposition sequences can be suggested. The more favorable reaction path and energetics leading compared to C h'*’ (and probably also C f'*’) are very ^ ^ ^ 24 24 ^ likely related to the relative abundances of 100 %, 23%, and 23-6 for these three species. Although the intensity of equal the sum of the intensities of C f''" and *“ 2^4 statistical considerations, it is in fact even higher by a factor of two. Our appearance potential value (12.85 V) for the C H f"^ 3 4 ion gives a value of 207 kcal/mole for AH^ of this species, which is closer to a previously reported experimental value of 210 kcal/mole than to a theoretically estimated value of 175 kcal/mole. The calculated AH° is apparently for a linear species; we suggest that the C H F species foimed in 34 this work may have a cyclopropane structure. (Other C 3 species formed might also be cyclic.) The utility of the electron impact method for determining thermodynamic quantities rests on the implicit assumption that the relevant but unmeasured corrections for excess energy in ionic and molecular fragments are small . ^ ^ ® ^ This situation is recognized by inequality signs in Equations I.(l)-I.(4). In several instances (see Equations 1.(15), 1.(17), I. (21) , and I. (26) ) two ionizati on or appearance potential

PAGE 35

22 terms enter with opposite signs, so that partial cancellation of excess energy effects can be expected. (In these cases no inequality is indicated since its sign is not unambigous.) An assumption of little or no excess energy is supported by Stevenson's Law^^ in the case of C f’*’ and C since 2 4 2 4 the appearance potential of is 0.55 V less than that of C 2 H 4 . This argument is not applicable in the case of C H f"^, but the good agreement of AH° (c-C H F ) from Equations 1.(1), I. (2) , and I. (3) suggests that any excess energy correction needed for A(C 2 H 2 F 2 ) must be small in the latter case as well; our results suggest that Ah° (c-C H F ) = -202 Real/ mole with little or no correction required for excess energy. Although the R.P .D. curves for C H F and C both show 34 2 4 well-defined breaks at higher energies (respectively 2.70 and 1.61 V above the appearance potentials) we have not attempted to interpret the electronic states or processes involved. However, Hamill and coworkers have had considerable in understanding data of this type.^® success

PAGE 36

CHAPTER II ION-MOLECULE REACTIONS IN 1 , 1 , 2 , 2-TETRAFLUOROCYCLOBUTANE Introduction This work was undertaken to obtain supplementary xnformation on ionic reactions in the gamma-radiolysis of gaseous 1 , 1 , 2 , 2-tetraf luorocyclobutane . Most of the reactions leading to stable products in the radiolysis of this compound appear to be radical reactions or molecular elimination processes. These radicals may have ionic precursors, however, since one of the major primary events in the gamma— radiolysis is the ionization of the compound. Accordingly , examination of ion— molecule reaction pathways in this system, using the technique of high pressure single source mass spectrometry and ion cyclotron resonance mass spectrometry, can shed light directly on the mechanism of the radiolytic decomposition. Experimental In this investigation the follov/ing two kinds of expsriments were carried out using two different instruments: (1) measurement of the variations of the ion intensities as a function of delay time (i.e., delay between ion generation 23

PAGE 37

24 and ion extraction for detection) using a Bendix Time-ofFlight mass spectrometer and (2) investigation of the reaction pathways using the ion cyclotron double resonance (ICDR) technique in a Varian "Synotron" ion cyclotron resonance instrument. The Bendix Time-of-Flight mass spectrometer (model 14107) was modified as described by Futrell et al.^^ The electron energy was 100 eV. Ion grid No. 1 was at +18 V and the repeller grid at +6 V with respect to ground. Distance from the repeller grid to exit orifice was 0.533 cm, and the drift distance from the electron beam position to the exit orifice was 0.267 cm. Under these conditions the field strength in the reaction chamber is 11.25 V/cm. A new circuit shown in Figure (3) v/as used to provide a positive pulse to ion grid No. 1. This circuit provided for the variation of the pulse height from 0 to +25 V and the width from 2 to 6 nsec while maintaining a very fast rise time (ca. 2 nsec) and very short delay (<1 nsec) relative to the ion focus pulse. In the experiments reported here the pulse height was at +23 V and the width was 4 ysec. All other ion source potentials were identical to those reported by Futrell. Further details about setting up and adjusting the instrument were described by Heckel and Hanrahan;^® similar procedures were employed in the present work. All ion intensity measurements taken under the pulsed conditions were corrected for diffusional discrimination using as a correction term the inverse square root of the ion mass (l//m).^^'29 Mass

PAGE 38

25 L-jci o Figure 3. Circuit diagram of the pulse generator for ion grid Number

PAGE 39

spectrometric data acquisition and data reduction, including the square root correction, were carried out using a General Automation SPC-12 mini-computer as described in a previous report. ^ ° The ion source was operated at 50*^ C. The reactant leaked into the ion source through a gold leak from a 5 £, reservoir . Variation of the reservoir pressure allowed adjustment of the pressure in the ion source, which was monitored using an MKS Baratron capacitive micromanometer (model 77) equipped with a 3 mm head. The tetraf luorocyclobutane was obtained from Columbia Organic Chemicals, Columbia, South Carolina, and was purified by trap-to-trap distillation. Ion cyclotron resonance experiments were carried out on a Varian "Synotron" (model 5900) equipped with a flat cell. In the ion cyclotron double resonance (ICDR) experiments the second radio frequency field was applied to the source region. The procedure described by Jennings^ ^ was followed to obtain the double resonance signals. Results It is possible to obtain rate data in a suitably equipped single-source mass spectrometer using a time delay mode, as originally described by Tal'roze.^^ In this procedure ions are generated by a brief electron pulse (ca. 1 ysec wide) , allowed to react with the substrate gas for an appropriate time (typically 1 to 10 psec) , and then extracted for analysis

PAGE 40

Zj by a pulse delayed with respect to the electron pulse. The procedure in our instrument is substantially the same, except that a small portion of the reactant gas product ion mixture from the source region continually effuses from the exit orifice, and these effusing ions are sampled by the ion focus pulse at a specified, variable delay txme after the electron pulse. in either event, the ions undergo a pseudo-first order reaction with the substrate gas. A plot of the log of the normalized, corrected ion intensity versus delay time gives a straight line with slope equal to the pseudo-first order rate constant; multiplying by the ambient gas number density (molecules/cc) gives the second order rate constant directly. Figure (4) shows the variation of the normalized, square root corrected intensities of five ions, C H^, C H F^, CF"*” ^ 2 4 2 2 2 ' ' C^F^,and as a function of delay time at a pressure of 27 microns. (Smooth curves through the experimental data points in Figures (4), (5), and (6) are visual best fits.) Intensities of and CF decay to zero within 3 ysec and ^ 2 ^ 2^2 qrows to a near constant value at approximately 3 ysec. After 3 ysec C F and C H F ion intensities continue 2 4 2 2 2 to grow, while the abundance of C H f'*' levels off and declines 2 2 2 somewhat. A similar plot is obtained at a pressure of 53 microns as shown in Figure (5). Semilogaritlimic plots of normalized, corrected intensities for loss of C and Cf"*^ 2 4 ' and formations of C H f"^, C f"^, and C H f"^, measured at 50° C 22224 44a' against delay time gave good straight lines; resulting rate

PAGE 41

(I//ra) /Z (I//m) 2.0 3.0 4.0 5.0 Delay time in ysecs Normalized ion intensities as a function of delay time in the T.O.F. spectrum of 1 , 1 , 2 , 2-tetraf luorocyclobutane at 26.8 microns and 50° C. Figure 4. < ©

PAGE 42

(I//m) /Z (I//m) 29 Figure 5. Normalized ion intensities as a function of delay time in the T.O.F. spectrum of 1, 1,2,2tetraf luorocyclobutane at 53.0 microns and 50° C.

PAGE 43

30 constants (based on a linear least squares fit of the data points) are given in Table IV. It is also possible to obtain ion-molecule reaction rate data in a single source instrument using the classical pressure variation method. This technique consists of measuring the ion intensities as a function of pressure in the presence of a known field strength, produced by the repeller grid. Although ion-molecule rate data were first taken using the pressure variation method, there are difficulties with interpreting the results of this procedure. ^ ® ^ ® When continuous ionization is used, the essence of the problem is that the effective reaction time for light ions is shorter than for heavy ions, since the former are lost sooner in collisions with the wall. By operating our instrument at a fixed, intermediate value of the ion-extraction delay parameter, it is possible to mitigate this difficulty to some extent; most intermediate and heavy mass ions then have the same effective reaction time, while some but not all of the light ions are lost. By operating in this manner with a fixed delay of 1.6 ysec we were able to take pressure dependence data which gave a reasonable comparison to the timedelay data, without recourse to extensive diffusion calculations.^® Results thus obtained are shown in Figure (6); the rate constants are shown in Table IV. Several ionic specxes observed in the Bendix instrument had intensities too low for quantitative work. Many of these (CUF , C^H^,etc.) are fragment ions which apparently disappear

PAGE 44

TABLE IV RstG Constants foir Ion~Molscul 0 Roactions in 1,1,2, 2-Tetraf luorocyclobutane Reaction Pressure in Microns Rate Constant in cc molecule"^ sec~^ X 10 * ° Method^ Disappearance of 26.8 11.0 TD c h ''' 2 4 53.0 7.76 TD 16.2 PD Disappearance Of 26. 8 17.9 TD cf'*’ 53.0 11.4 TD 17.7 PD Formation of 26.8 4.3 TD C H f"^ 2 2 2 53.0 0 . 89 TD 1.74 PD Formation of 26. 8 5.1 TD c f "^ 2 4 53.0 2.9 TD 3.4 PD Formation of C H f"^ 4 4 3 26.8 7.5 TD 53.0 4.0 TD 5.2 PD TD = Rate constants obtained PD = Rate constants obtained experiments . from the time delay experiments, from the pressure dependence

PAGE 45

32 (E//I) 2/(^/l) Figure 6. Normalized ion intensities as a function of pressure in the T.O.F. spectrum of 1 , 1 , 2 , 2-tetraf luorocyclobutane at 250 C.

PAGE 46

JJ by reaction with the substrate. Of more interest, however, are two ions with molecular weights greater than the parent molecule; the species appear at m/e 164 (C H f'*') and 192 ^ 4 2 6 former ion is formed with moderate intensity under conditions of very low repeller grid voltage; it is formed in an ion— molecule reaction in which an ion of m/e 64 (C^H^F^) is consumed by reaction with the substrate. Although no further information on the ion of mass 192 was obtained using the Bendix instrument, its mode of formation was characterized by the ICDR technique described below. In the original ion cyclotron method for the observation of ion— molecule reaction as developed by Baldeschweiler , an investigation of the variation of ion intensities as a function of pressure also allows calculation of rate constants.^® Although we investigated the ICR spectrum of 1 / 1 f 2 , 2— tetraf luorocyclobutane over the entire pressure range available on the Varian instrument, no variation in ion intensities was observed. (The instrument operates well at pressuresfrom 1 x 10“^ Torr to 1 x 10“^ Torr; although the -7 region 1 x 10 Torr is accessible, the poor signal-to-noise ratio prevents quantitative measurements in this region.) Throughout the readily accessible pressure region, the spectrum observed on the ICR instrument resembles highpressure or long delay-time spectra seen using the Bendix machine. Apparently due to the very long residence times in the ICR instrument (several milliseconds) , the more reactive ions, such as ethylene in the present case, have already reacted.

PAGE 47

34 A more recent technique using the ICR instrument is the double resonance method, in which kinetic energy of a chosen reactant ion is enhanced by irradiation with a second radio frequency field. The product ion intensity will increase provided the formation reaction of the product ion is endoergic. However, a decrease in the product ion intensity with an increase in the kinetic energy of the reactant ion indicates either an exoergic or thermoneutral reactions. Such a decrease in intensity is most commonly seen in exoergic reactions. In the ICR instrument two ions heavier than the parent ion, one at m/e of 192 and the other at m/e of 142, were formed above 6 x 10 ^ Torr. Although the ion with m/e of 192 (C^H^F^) was seen in experiments carried out using the Bendix machine, the ion of m/e 142 (C H F ) was not seen 5 6 4 probably due to a very low formation rate constant. Using the ICDR technique, the reactant ion in the formation reactions of these two heavier ions was established to be ^ 2 ^ 2 ^ 2 * ICDR signal for the reactions leading to the formation of C H F and C H f'*Â’ at different power-levels of the second radio frequency field, applied to the source region, were of the shapes shown below (vertical axis, intensity; horizontal axis, double resonance frequency) : (For ion with m/e of 192) (For ion of m/e of 142)

PAGE 48

35 The power absorption by the product ions decreases when the intensity of the second radio frequency field of frequency corresponding to that of is increased, indicating that both the reactions are exoergic. Discussion An examination of the low pressure mass spectrum of (Table V) shows that the compound fragments extensively under electron impact. The six most intense ions are all and fragments (abundancies from 33% down to ca. 6%). The parent ion (not necessarily cyclic) has an abundance of only 0.1%; the other C species, C H f"^ + ‘*443 and C^H^F^, occur to the extent of 2.9 and 5.3%, respectively. In all, five fragments are seen, but the abundance of the most important is only 5.3%. In previous researches by Jennings et al.^^'^** and Anicich and Bowers'^ the ion C^H^F+ was formed by the reaction of an ethylenic ion (C H F^, C h''' , or C f"^) on an ethylenic target molecule (C^H^F^, C^F^, or C^H^, respectively). It was found that the intermediate complex ion C H F^ is very unstable (life time < 2 ysec) and decomposes unless collisionally stabilized. The major products are an ethylenic ion/ethy lenic molecule oair: C H f’*’ 'V. CUP 4 4 4 c H f"*" * / 2 2 c 4 4 4 C H F^ ^ 2 4 4 4 4 ^ 2 4 + C H F 2 2 2 + C F 2 4 + C 2 4 2 II. (la) II. (lb) II. (Ic)

PAGE 49

Mass Spectral Fragmentation Patterns of 1 , 1 , 2 , 2-Tetraf luorocyclobutane Under Analytical and Ion-Molecule Reaction Conditions 36 4J M (T3 M O B P 4-1 I U O CU iH di cn x (U o t— I LD CTi (N • • • . . ^ (N rH rH 00 « in U . H VT> to >1 4-1 •H U) C Q) 4-1 C H u dJ p m U3 0) g u p 4-1 x; o Cn O •H d K tn m m e o p d 4-1 0 S o O d 1-^ 00 4-1 p (U e p gt •H O U3 o o m CO rH • • un K U m X C4 u (N O' a 04 CN O a 04 1 1 JK Cm CM Cm E u U U 00 1 — I CN CN 00 00 CJ^ 'X) oo 04 x> E fO CM Ph jCO •-rt
PAGE 50

113 37 QJ C3^ in • • • OHO (U e o o o o o H H O 4J *H W c Q) 4J C •H 0) cn (D > c: •H (U s: -p o 4-) I — I c o o I — I 1 — I • • o o JPh IX u S) IJO K E PO OJto u u u 00 (N CN CTl (C 4-1 O 4-> 0) 44 O 44 'd CJ N to e p o c c to 'O (D . 44 (U U H QJ XI M to >4 44 O O QJ 40 m 44 cn to c S -H QJ C to o 40 to to o •H 44 44 O •H fO to C to QJ c 44 o C -H H P QJ 44 O. to 40 U S4 rH O -H to . QJ • a • to 0 1 — 1 QJ to U to g to Cu •H g 44 44 to QJ QJ •H >i40 to to 44 to tjt 1 — 1 0 to QJ 44 •H *H 13 0 44 lO O c e 44 QJ QJ O 0 U E M 0 U 44 P 0 Oh O 40 QJ t4 to to 4<^ to to tu to 3 to QJ Eh tr' S cn U to 13 QJ

PAGE 51

38 Firorti th 0 f iT3.gin0nt3tion p3tt03rn shown in T 3 bl 0 V it is cl 03 .]r that th 0 b 0 havior of tho stablo moloculo c-C H F undor 4 4 4 ^l^^tiTon impact is V 0 iry similar to that of tho ion— moloculo collision complex (as shown in Chapter I) . Under conditions of elevated pressure, long residence time, or both, the primary fragments from tetraf luorocyclobutane are available to enter into subsequent ion-molecule reactions with substrate. The abundance of the parent ion is so low, however, that it cannot make a significant contribution to the observed ion-molecule reactions. Several of the species are of moderate abundance (ca. 6 to 12%); they react so rapidly with substrate that is difficult to ascertain the details of the processes involved. It appears likely (Figures (4) and (5)) that they form one or more of the ethylenic ions. It can be seen in Figures (4), (5), and (6) that the predominant reactions at intermediate pressures and/or residence times involve formation and decay of the ethylenic ions C h'*' , C H f'*’ , and C f'*'. 2 4 2 2 2 2 4 The changes in the abundance of the major ions at low pressures due to increases in pressure or residence time are shown in Table v . At lower pressures (of the order of 27 microns) and short delay times, the ions heavier than C F^ 2 4 are absent. When the delay time is increased the heavier ions increase in abundance. At higher pressures (of the order of 50 microns) , all the ions seen at lower pressures are present throughout the range of delay times studied, including the heavy ions which at 27 microns were seen only at long

PAGE 52

39 delay time. As seen in the table, the intensities of C H f"*", + + , _ 2 2 2 ^ 2 ^ 4 ' with increases in both the pressure and delay time, while and Cf"^ abundancies decrease. The same trend continues at higher pressures, as shown by the ICR spectrum. (Even though the pressure in the ICR mass spectrometer was lower than in the Bendix mass spectrometer, the very long residence times in the former instrument allow for a large number of collisions before the ions are lost to the walls or removed from the cell.) This result clearly shows that C^H^F^ is the major ion in this system at higher pressures (for example, pressures of 10 to 50 Torr in radiolysis experiments) . Although C^H^F^ was found to react with the substrate, the rate constant for the process must be quite small. In experiments using the Bendix T-O-F, the C H f"*" ion was the 222 most abundant even at high pressures and long delay times. The low reactivity of C^H^F^ is substantiated by the experiments using ICR, where mass corrected ion intensity fraction of this ion was measured as a function of pressure (from 1 x -6 -4 10 to 1 X 10 Torr) . There was no measurable change in intensity of the ion with the change in pressure. In contrast, is a very reactive ion. The intensity of C H decays to zero within a few microseconds with the 2 4 concomitant increase in the intensity of C H f'*' (Figures M) 222 and (5) ) . The same is true in the pressure dependence experiments where the growth of C H f"*" accompanied the decay of v/ith increasing pressure (Figure ( 6 )).

PAGE 53

40 The reaction leading to the decay of C n’’’ could be a 2 4 Charge transfer from C to c-C H F , followed by the decomposition of the C H F intermediate to give C H f"*” 4 4 4 ^ + c-C H F 4 4 4 ^ C H f"^ + C H ' 4 4 4 2 4 II. (2a) C H F"^ 4 4 4 ^ C H f”^ + C H F ^ 2 2 2 2 2 2 II. (2b) However, this reaction would be thermodynamically unfavorable by 42 heal, unless in Equation 11.(2) is in an excited state. Hence, it is to be concluded that C H^ 2 4 reacts with C H f"^ to give C H f"^ and C H F , the product ^ 4 ^^ 6^2 a heat of formation of at least -108 kcal/ mole. A heat of formation of -108 kcal/mole is a reasonable value to be expected for a C compound. C h'*' + c-C H F 2 4 4 4 4 / + * (C H F ) 6 8 4 -^C H F 2 2 + 2 Reaction 11.(3) could proceed through either a six center iritermediate or attack of C H on a C-C o bond in c-C H F . 2 4 4 4 4 * There is considerable evidence for the reaction of carbocation on a C-C a bond, as presented by Olah,^® whereas a six center complex has not been established. Accordingly, we expect Reaction II. (3) to be a simple radical ion (C h"*”) 2 4 attack on the C-C a bond connecting two methylene groups in ^~^ 4 ^\^ 4 ' followed by the rupture of the C ion to give ^ 6 *^ 2 ^ 2 ^ 2 ‘ Since the reaction occurs in the gas phase, the energy rich C^ intermediate would tend to decompose, unlike

PAGE 54

41 the situation in Olah's "magic acid" solvents, where the intermediate undergoes collision deactivation followed by atom migration to give a stable product. Several intermediates and ultimate products are probable, depending on which of the different C-C bond types is attacked, and how the transient ion ruptures. However, attack on the most electron-rich of the three bond types (-CH — CH -) would lead 2 2 to formation of the most abundant product C H while attack 2 2 2 at the next richest site (-CF^ — would explain a small production of C F . 2 4 Although the agreement is only moderate, the rate constant for the disappearance of C in Table IV can be 2 4 equated to the sum of rate constants for formation of C H f"*” ^ 2 2 2 and C^F^. It will be seen that the sum of the two formation rate constants is somewhat smaller than the C decay rate 2 4 constant at 27 microns, and considerably smaller at 53 microns. We suggest that this discrepancy is due to a reaction in which C^H^F^ is lost by reaction with substrate; several products are possible including C H f"^, which V'/ould be formed 4 4 3 by a fluoride ion transfer process. C H F"^ + c-C H F 2 2 2 4 4 4 ^CnF.+CHF"*" 2 2 3 4 4 3 II(4) As predicted by this model, the discrepancy between k loss (C^HJ and ^^ 2 ^ 2^2 *^ 2 ^ 4 ^ increases with an increase in pressure. The reactions of mentioned above could not be established using ICDR since the intensity of C F^ was 2 4 very small.

PAGE 55

42 The small abundance of species heavier than 128 (the parent) shows the relative unimportance of ion-molecule condensation reactions in this system. Of the three heavy ions (m/e 142, 164, and 192), only the one at 192 was seen both in the Bendix and ICR. The relatively small intensity of this ion forbade any kind of time or pressure dependence measurements. However, using the ICDR technique we could show that at least one of the routes of formation is by the reaction of with c-C^H^F^. As mentioned in the Results section, this ion is formed by an exoergic reaction. C H F + c-C H F ^ c H f"*" tt 2 2 2 4 4 4 ^ 6 6 6 H.(5) rn/e = 6 4 m/e = 12 8 m/e 192 The ion of mass 142 was seen only in the ICR and the one at 164 only in the Bendix. Again, the intensities of these ions were very small indicating small rate constants for the formation of these products. The ion at mass 142 (C H F^) 5 6 4 was also formed by the reaction of C H f'*Â’ with c-C H F as 2 2 2 4 4 4 shown by ICDR. C H f"^ + c-C H F 2 2 2 4 4 4 C H f'*' + CF ^ 5 6 4 2 II. (6) m/e 64 m/e 128 m/e = 142 m/e = 50 Again, this ion (C^H^F^) is formed through an exoergic reaction. This observation is compatible with the absence of this ion in the Bendix experiments where the electron energy was 100 eV, which would tend to increase the energy

PAGE 56

43 of the reactant ion, C H F*". 2 2 2 The route for the formation of the ion at mass 164 established because this ion was absent in the ICR. However, it is very likely that C H is 2 2 2 responsible for this ion also. The occurrence of C H f"^ 4 2 6 in the Bendix and not in the ICR suggests that the lifetime of the ion is very small, perhaps of the order of 2 ysec or less.

PAGE 57

CHAPTER III 3 THE Hg6( P^) PHOTOSENSITIZED DECOMPOSITION OF 1,1,2, 2 -TETRAFLUOROCYCLOBUTANE Introduction 3 Mercury 6( P^) photosensitized decomposition studies on a large number of alkanes and cycloalkanes have established that C-H bond rupture is the primary quenching process in these systems. ^ 5 ° Work on a few partially fluorinated hydrocarbons has indicated that in these systems as well, C-H bond scission is the primary mode of quenching the excited mercury.'*^ Only a limited amount of work has appeared dealing with mercury sensitized decomposition of perfluorocarbons (whether open chain or cyclic) ; details of the decomposition mechanism have not been established.** There have been no previously reported investigations of the mercury sensitized decomposition of partially fluorinated cyclobutanes or other cycloalkanes. The present investigation of the mercury sensitized photolysis of 1,1,2,2-tetrafluorocyclobutane was undertaken in parallel v/ith studies of the gamma radiolysis, mass spectrometric fragmentation, and ion-molecule reactions of this compound. It was expected that mechanistic information derived from the photolytic (and mass spectrometric ) studies v/ould aid 44

PAGE 58

45 in interpretation of the more complex radiolytic system. Experimental Section Sample Preparation The vessel used in photolysis experiments had a volume of 91 cc; it was constructed of 18 mm ID GE type 204 clear fused quartz and was equipped with a glass break seal and side arm (Figure (7)) . To clean the vessel between experiments, it was soaked in nitric acid overnight to remove inorganic material (especially mercury) , rinsed with distilled water, and annealed at 565° C to pyrolyze any organic materials on the walls. Before introducing the sample a drop of triple-distilled mercury was placed in the vessel, and the vessel pumped on for at least six hours. In the case of photolysis of the pure system, sam.ples of c-C H F 4 4 4 were measured in a standard volume using PVT technique, transferred to the photolysis vessel, and sealed under vacuum^. For photolysis of a mixture of hydrogen and c-C II F , the 4 4 4 tetraf luorocyclobutane was introduced as above, the photolysis vessel immersed in liquid nitrogen, and hydrogen metered directly into the photolysis vessel. After the introduction of hydrogen, the vessel was very cautiously sealed. The pressure of hydrogen in the vessel at room temperature (297° K) would be 3.87 (i.e., 297° K/77° K) times the measured pressure. For quenching cross section and actinometry experiments, c-C H F was introduced as in the case of the pure system; a 4 4 4 premeasured amount of N 0 from a standard vessel was vacuum 2

PAGE 59

46 Figure 7. The photolysis set up: A, mercury drop; B, break seal; C, reaction cell; D, glass tube to support the photolysis vessel; E, desk lamp fixture; F, germicidal lamp; G, graded seal.

PAGE 60

47 transferred into the photolysis vessel and sealed under vacuum. The purification of c-C H F has been described in Chapter IV. n^O obtained from Matheson Gas Company was subjected to at least ten freeze-pump-thaw cycles to remove air before use; C.P. grade Matheson H was used as 2 received. Sample Irradiation To maintain a constant mercury vapor pressure in mercury sensitization experiments, it is common to provide a small drop of liquid mercury in direct contact with the vapor phase. To obtain reproducible results, hov/ever, we found it necessary to make sure that the drop was in a small side arm, and not in the main reaction cell exposed to the full lamp intensity. The source of 253.7 nm light was a General Electric 15 W Germicidal lamp (type G15T8) which is essentially a low pressure mercury lamp in a high silica envelope, which transmits 253.7 nm and not 184.9 nm (Figure (7)). This, lamp was mounted inside a standard desk lamp fixture equipped to hold two fluorescent tubes. The second lamp was replaced by a stack of Pyrex tubes which provided reproducible positioning of the sample vessel. To avoid intensity variations due to AC line voltage fluctuations, a Sola saturable core transformer was used to power the lamp. The photolysis arrangement was covered with a wooden eye shield and air was blown through this set-up to remove any ozone formed and to keep the temperature constant. The temperature was recorded using a

PAGE 61

48 thermometer with its bulb resting inside the photolysis cavity. This experimental arrangement is essentially the same as the one described in detail by Dr. Arthur J. Frank of this laboratory. ** 2 After photolysis, the sample was cooled down to liquid nitroqen temperature to prevent any dark reactions. Product Identification and Analysis The organic product identification was done using an on-line gas chromatograph-mass spectrometer combination. Hydrogen was measured by the conventional Toepler pump MeLeod gauge arrangement and the organic products by flame ionization gas chromatography. The details of identification and analyses are given in Chapter IV and Appendix I. Preparatory to a wet-chemical analysis of HF , an extractant solution of 2 ml of 66% ethanol in water was introduced into the tubulation extending beyond the break-seal on the photolysis vessel (Figure (7)). With the cell contents frozen, the break seal was opened with a rod introduced through a piece of rubber tubing used as a gland. This procedure was necessary to avoid condensation of oxygen inside the vessel. The vessel was warmed to room temperature and the ethanolic solution was left in contact with the walls of the vessel for at least 30 minutes to extract all the fluoride ion into the solvent. This solution was transferred into a 5 ml polyethylene beaker, and the vessel was rinsed out into the beaker with another 2 ml of 66% ethanol. The determination of HF was based on a previously described

PAGE 62

49 electrochemical procedure;''^ however, it was necessary to make a minor modification to permit analysis in the presence of mercury or mercuric ions. A drop of dilute aqueous NaCl was added to the fluoride ion solution prepared as described above . To make sure that the added chloride ion was not interf erring with the fluoride ion determination, a series of experiments using known amounts of NaF was carried out. (The Orion fluoride ion selective electrode, model 94-09, used in these experiments is claimed to have a selectivity of 1000:1 over chloride ions.****) in these verification experiments 40 of 5.00 x 10 NaF was titrated against 5.00 x 10~^M La(N 0^)3 in the presence of chloride ion and/or mercuric ion and mercury. Figure (8) shows the titration curves for these analyses. In case of the titration in the presence of mercuric ions, the EMF changed from -0.052 V to -0.003 V soon after the addition of mercuric ion indicating the complexation of fluoride ions with mercuric ions; the subsequent addition of chloride ion changed the EMF back to -0.051 V, indicating the release of fluoride ion into the solution. These experiments show that the chloride ions (which can be used to release fluoride ions from mercuric ions and/or mercury) do not interfere with the determination of fluoride ions. In all these experiments, the procedure prescribed by Heckel'*'" was followed. It was essential to condition the fluoride ion selective electrode by performing at least two

PAGE 63

EMF 50 3 3 Figure 8. Titration curves for quantitative analysis of F~ ion: 40 yl of NaF + HG^+ ion + Cl“ ion, 40 yl of NaF + Cl“ ion, ; 40 ylof NaF, 'solid line; photolysis, O.

PAGE 64

51 titrations using this electrode before it could be used for a quantitative determination. During all the titrations, the analyte was continuously stirred using a magnetic stirrer, and the EMF was monitored using a high input impedance (10 Mega Ohm) Hickock digital Volt-Ohm meter. Nitrogen formed in the photolysis of a mixture of N 0 2 and c-C H F was transferred from the photolysis vessel 4 4 4 rr j immersed in liquid nitrogen via a Toepler pump into a McLeod gauge where it was measured. Results Quenching Cross Section Measurements and Actinometry For quenching cross section measurements and actinometry four mixtures with different pressure ratios of c-C H F to 4 4 4 N^O (P ^_0 p /Ppi q) were photolysed for 30 sec. In each 4 4 4 2 sample, the total pressure was 50 Torr. A plot of the reciprocal of the number of moles of nitrogen formed in each mixture against the quantity (P(^_q pj p /Ppj q) gave a straight 4 4 4 2 line as shown in Figure (9) . The intercept of this line gives the reciprocal of the number of moles of nitrogen formed in pure N^O. The amount of nitrogen formed in the absence of c-C^H^F^ was found to be 9.597 ymoles. Assuming a quantum yield of 0.8 for production in mercury sensitized photolysis of the absorbed light intensity was calculated 17 to be 2.4 X 10 quanta/sec in the 91 cc vessel (i.e., 24 UEinsteins/min) . The quenching cross section of c-C H F 4 4 4 was determined relative to that of nitrous oxide. The

PAGE 65

moles 52 P c-C H F /P N 0 4 4 4 2 9* Modifisd StG]rn"*VoliTiG 2 r plot; a. plot of' thG iTGcipirocal of the number of moles of against the ratio of the pressure of c-C^H^F^ to that of N 2 O.

PAGE 66

53 ratio, 3, of the relative quenching rates for c-C H F and 4 4 4 nitrous oxide was calculated from the slope of the line in Figure (9) to be 3 (slope) X (number of photons absorbed in 30 sec)/ Avagadro's number 6 = 0.0102 The quenching cross section of c-C H F was calculated using 4 4 4 ^ the equation** ° a c-C H F 4 4 4 a 2 N 0 2 " ^ o' 1 + (M /M Hg/c-C H F ) 4 4 4 '2 Where is the quenching cross section and M is the molecular weight of the species shown by the subscripts. Assuming the quenching cross section of N^O determined by the chemical method to be 18 A^ (Reference (^O)) , Q c-C K F to be 0.27 A^ . Recently Gleditsch and Michael 0 a value of 22.1 A as the preferred value for this value of q, the corresponding value of 0 ^ 0.31 A^ . IS calculated ** ^ have given a N,0c-C H F is 4 4 4 Photolysis Products The major products formed in the photolysis of at 50 Torr are HF, C H , C F H , two isomers of C H 2 6 3 4 2 c c-C H F 4 4 4 F (here6 after referred to as C H F (I) and C H F (II)), an unidentified O 4 fa 6 4 6 C^ compound, and C^H^F^(a bicyclic compound); the minor products were H^, CH^, C^F^, 1,1dif luoroethy lene (1,1-C H F )

PAGE 67

54 and C^F^H. There v/ere at least two more products whose yields were not recorded since they were formed in very small quantities and their compositions were unknown. Based 19 upon the F NMR and mass spectral cracking pattern the probable structure of C H F (I) is as follows: 6 4 6 The other isomer, C H F (II) , is probably non-cyclic as 6 4 6 indicated by its proton NMR. We could determine neither the structure nor the exact composition of the C compound. 8 Further discussions of the composition and structure of some of the photolysis products are given in Chapter IV and Appendix II. A series of experiments was performed to investigate the dependence of product yields on photolysis time in pure 1 , 1 , 2 , 2-tetraf luorocyclobutane . These experiments were carried out at both 50 and 150 Torr of the tetraf luorocyclobutane to reveal a possible pressure effect. The yields of C H and two isomers of C H F increased linearly with irradiation time. The rest of the products were formed at a decreasing rate at longer irradiation times and many of the yields approached limiting values after tv;o minutes of photolysis. Figures (10) (14) show the yield-photolysis time plots for all products at both pressures. Quantum yields for all products at 120 sec photolysis time for a 50

PAGE 68

Micromoles of HF formed 0-0 1.0 2.0 3.0 4.0 5.0 Photolysis time in minutes Figure 10. Yield of HF ( O at 50 Torr; # at 150 Torr) and H 2 ( at 50 Torr; | at 150 Torr) as a function of time in the photolysis of c-C^H^F^. Moles of H formed

PAGE 69

Moles of products b fa Photolysis time in minutes Figure 11. Yield of C 3 H 2 F;, ( O at 50 Torr; # at 150 Torr) and C 2 HF 3 ( Q at 50 Torr; at 150 Torr) as a function of time in the photolysis of c-C^H^Fg,

PAGE 70

Micromoles of Products 57 0-0 1-0 2.0 3.0 4.0 Photolysis time in minutes Figure 12. Yield of CgH.Fgd) ( Q at 50 Torr; # at 150 Torr) and CgH^Fg dl) ( at 50 Torr; at 150 Torr) as a function of time in the photolysis of c—

PAGE 71

Moles of 58 Figure 13. Yield of C 2 Hg ( at 50 Torr; at 150 Torr) and CH 4 ( O at 50 Torr;# at 150 Torr) as a function of time in the photolysis of c-C H F Moles of CH

PAGE 72

Moles of product formed 59 Figure 14. Yield of CgH^Fg ( O at 50 Torr; • at 150 Torr) and the Cg compound ( at 5 0 Torr;B at 15 0 Torr) as a function of photolysis time.

PAGE 73

60 Torr sample are shown in Table VI. Since the absolute yields are quite small, a set of relative yield values normalized to HF yield = 1.00 are also given in the Table, in order to give a clearer picture of the behaviors of the system. Since the yields of CH , the C compound and C H F at 8 8 6 8 150 Torr were different from those at 50 Torr, a series of 120 sec irradiations were carried out using pure c-C H F at 4 4 4 three more pressures ranging from 10 to 150 Torr. Yields of all the products except CH , the C compound and C H F were ** 8 8 6 8 independent of pressure above 30 Torr, within experimental error. The yields of methane decreased with increase in pressure while those of the C compound and C H F increased 8 8 6 8 linearly as shown in Figure (15) . To establish whether any of the products were due to the reactions of hydrogen atoms, a mixture of 400 Torr of H 2 and 20 Torr of c-C H F was photolysed for 300 sec. All 4 4 4 the products formed in the pure system were formed except C F , C n F (II) , the C compound and C H F . The yields 24646 8 868 of products formed in this experiment are given in Table VII. The material balance in the photolysis is reasonably good. The ratio of C/H/F is 4.00/3.65/4.48 when the photolysis has progressed for 120 sec. This shows a shortage of carbon and hydrogen. However, at shorter photolysis times the material balance is better, and at longer photolysis times worse. This suggests the formation of polymers or products which were not analyzable by the methods used in these experiments .

PAGE 74

61 TABLE VI Yields of Products Formed in a 120 Sec Photolysis of 50 Torr of c-C H F 4 4 4 Product Yield at 120 sec in n moles Quantum Yield Relative Yield HF 0.091 0.0019 1.0 H 2 0.006 0.00012 0.063 CH 4 0.0016 0. 00003 0.016 C H 2 6 0.0095 0.000197 0.10 C F 2 4 0.0015 0.00003 0.016 C H F 2 2 .2 0.00092 0.00002 0.010 C HF 2 3 0.00161 0.000033 0.017 c-C H F 3 2 4 0. 0310 0.0006 0.34 C H F (I) 6 4 6 0.046 0.000954 0.50 C H F (II) 6 4 6 0.031 0.000643 0.34 c ? 8 0.009 0.000186 0.10 C H F 8 6 8 0.020 0.00041 0.22

PAGE 75

Moles of product 62 Figure 15. Plots of yields of ( O ) and the C compound ( ) as a function of pressure in the photolysis vessel.

PAGE 76

63 TABLE VII Yields of Products Formed in a 300 Sec Photolysis of a Mixture of 400 Torr of H and 20 Torr of c-C H F 2 1+4 4 Product Amount in ymoles Relative Yield HF CH I* C H 2 6 C H F 2 2 2 C HF 2 3 c-C H F 3 2 4 C H F (I) 6 4 6 0.2348 0.0019 0.0125 0.0021 0.0308 0.0025 1.0000 0.0082 0.0530 0.0090 0.1310 0.0106 0. 0019 0.0082

PAGE 77

64 Discussion The possibility of direct photolysis of c-C H F by 4 4 4 253.7 nm light can be ignored since c-C H F has no absorp4 4 4 tion above 180.0 nm. ® Hence, all of the observed products in these experiments must be due to the quenching of Hg6(^P^) to the 6 (^Sq) ground state by c-C^H_^F^. (Although it is possible to quench Hg6(^P^) to the 6 (^Pq) state, this process has been shown to be unimportant in many systems'*^ and we ignore it for lack of any direct evidence.) That c-C H F is 4 4 4 very efficient in quenching the Hg6(^P^) is shown by its relatively small value of 0.3 for the quenching cross section (based on a value of 13 A^ for . This inefficiency in quenching is in accordance with the generalization, discussed by Cvetanovic , ° that cycloalkanes are analogous to alkanes in their low quenching efficiency. A value of 0.3 A^ for pj p is similar to the reported quenching cross 4 4 4 sections of three partially fluorinated alkanes CH CII F 0 0 ^ ^ (0.46 A^), CH CHF (0.45 A^), and CH CF (0.45 As 3 2 3 3 expected this value for the quenching cross section of c-C H F 4 4 4 is considerably smaller than those of comparable alkanes or cycloalkanes but larger than those of perf luoroalkanes (e.g., ‘^c-C H ^ 0.005 A^).‘^0''*2 4 8 I, , The primary photophysical event is shown in Equation III.(l) and the subsequent quenching in Equation 111.(2). Hg6 (^Sq) + hv (253.7 nm) ^ Hg6(^P^) III(1)

PAGE 78

65 Hg6 ( P ) + c-C H F ^ C H F • + H+ Hg6 (^S^) III. (2) It is suggested that the quenching of Hg6 (^P, ) by c-C H F 1 i, 4 4 leads to the scission of a C-H bond (Reaction III. (2)). Rupture of the C-C bond is not suggested since v/e saw no evidence for it. Reaction III. (2) is suggested to be the major quenching process because of the following three reasons: (a) C-H bond rupture has been established to be a major quenching process in cycloalkanes ; ^ ° (b) the products formed in photolysis of pure c-C H F and those in a mixture 4 4 4 of a large amount of hydrogen and a small amount of c-C H F 4 4 4 are similar, suggesting that hydrogen atoms are formed in the quenching process in this system; and (c) when a mixture of and I^ was photolysed in the presence of mercury vapor, all the products in the photolysis were eliminated, and a new product identified as C H F I (on the basis of 4 3 4 GC retention time and the mass spectral fragmentation pattern) was formed. These three observations argue very heavily in favor of Reaction III. (2) as the primary quenching process. The two radicals C^H^F • and H* (formed in the primary process) can react with each other or v/ith the substrate gas leading to the formation of the observed photolysis products. We were able to study the simple reactions resulting from reactions of H* and fluorobutyl radicals in the absence of the more complex reactions resulting from mutual reactions of fluorobutyl radicals by photolyzing a mixture of H and 2 in the presence of mercury vapor. These experiments

PAGE 79

66 are discussed first, followed by consideration of the more complex pure system. In the photolysis of a mixture of 400 Torr of hydrogen and 20 Torr of c— C H F , essentially all the excited mercury atoms will be quenched by hydrogen since its quenching cross 0 section (8.6 ) “* ^ and relative amount are each twenty times larger than those of c-C^H^F^; thus only about one quenching event in 400 will involve the fluorocarbon directly. Hg6 ( P^) + ^ 2H+ Hg6(^SQ) 111.(3) It is possible that this reaction proceeds via formation of HgH, although leading to the same final products. Since H* would be the only reactive species formed in the quenching process, all the products seen in the mixture of H and 2 c-C H F must be due to the reactions of Hwith c-C H F . 4 4 4 Expected reaction of hydrogen atoms would be either an abstraction of H* from the substrate or an addition to substrate with simultaneous ring opening. H* + c-C, H F 4 4 4 > H + c-C II F • 4 3 4 III. (4) H* + c-C H F 4 4 4 ^ (CH CII CF CF •)* 3 2 2 2 III . (5) Production of H* in Reaction 111.(4) can not be observed directly due to the presence of excess hydrogen. However , results from the pure system discussed below suggest that Reaction 111.(4) is considerably less efficient than Reaction III. (5) .

PAGE 80

67 We have postulated the addition of hydrogen atom to a carbon bonded to hydrogen, since in 1 , 1-dif luoroethylene hydrogen atoms add almost exclusively at the nonf luorinated carbon end of the molecule,**® indicating that the addition to the methylene group is more favorable than to the perf luoromethylene group. Furthermore, the radical site in the product of Reaction III. (6) is expected to be the terminal CF 2 group, since fluorine atoms are known to stabilize radicals.**® Extended considerations of the observed product distribution requires that we postulate a substantial contribution from a third hydrogen atom reaction, namely abstraction of a fluorine atom from substrate. H* + c-C H F ^ HF + c-C H F * III. (6) Although such a reaction is thermodynamically favorable in most cases, it is generally not observed, presumably due to unfavorable kinetic factors. The reaction should be especially favorable in a cyclobutane; however, the near 90° ring 3 angles prevent good sp hybridization, decreasing sigma bond energies, and enhance accessibility of both the carbon atom and substituents to attack. The fates of the three radicals, c-C H F *, (CH CH CF CF • ) * 4 3 4 3 2 2 2 and c-C_^H^F^* (formed in Reactions 111.(4) through III. (6)) determine the products formed and their relative distribution observed in the mixture. Since Reaction III. (5) involves formation of a primary C-H bond (ca. 98 kcal/mole^’) and rupture of a weak C-C bond (61 kcal/mole according to our

PAGE 81

68 measurements ) , the product radical will be excited by approximately 37 kcal, probably mostly in the form of vibrational energy. This amount of energy is easily sufficient to cause HF elimination, which is not more than 15 kcal endoergic: (H CCH CF CF • ) * ^ HF + CH CH = CFCF • 3 2 2 2 3 2 III. (7) The reaction probability is further enhanced by the fact that the product is a relatively stable vinyl radical. Reaction 111.(7) would occur in competition with thermalization of the precursor species and we expect that it is a moderately important process. The final fate of the radicals in Reaction III. (7) should be reaction with hydrogen atoms, which are present in much larger concentration than any other radicals in the system. The product (CH^CH = CFCF^H) would initially contain excess energy due to C-H bond formation which could either lead to further fragmentation or be lost in thermalizing collisions. Although no product having the formula CH CHCFCF H was observed, it might easily be masked by elution with the parent c-C H F . 4 4 4 since the measured product yields indicate a surplus of fluorine relative to carbon and hydrogen, formation of a moderate amount of this compound v/ould improve the material balance . The decomposition of (CH^CH^CF^CF^ ) * to give either a methyl or ethyl group is thermodynamically unfavorable

PAGE 82

69 (endoergic by at least 30 kcal/mole for methyl radical formation and 21 kcal/mole for ethyl radical elimination) . However, the product of H* and CH ^CH^CF^CF^ • (either excited or collisionally stabilized) certainly contains sufficient energy to undergo decomposition. CH CH CF CF • + H^ (CH CH CF CF H) * 3 2 2 2 / 3 2 2 2 (H CCH CF CF H) * CH + c-C H F 3 2 2 2 ^ 4 3 2 4 (CH CH CF CF H)* ^ C H • + C F H3 2 2 2 2 5 2 4 III. (8) III. (9) III. (10) Reactions 111.(9) and 111.(10) are very similar to those reported for the pyrolytic decomposition of n-butane.^° +CH 111.(11) 4 10 / 4 3 6 and ^ ^ ‘ 111 .( 12 ) 4 10 2 5 Although the present work was done at room temperature, the energy made available by Reaction 111.(8) should be sufficient to drive either Reaction 111.(9) or Reaction 111.(10). Since Reaction III. (9) involves rupture of a C-C bond (83 kcal) and formation of a C-H bond (102 kcal) and a new C-C bond, the overall process is energetically very favorable. (If the heavier product in Reaction III. (9) is an open chain olefin, the second link in a C=C bond will be formed with an energy release of 63 kcal.)^ ’ The amount of CH formed is almost 4 equal to that of C H F as expected according to Reaction 111.(9). The sequence of Reactions 111.(8), 111.(10) is

PAGE 83

70 exothermic only to the extent that the C-H bond energy (ca. 98 kcal/mole) exceeds the energy of the C-C bond (83 kcal/raole) . Nevertheless, both types of reactions are well established under high temperature conditions^ ° and appear plausible here as well. The fate of (from Reaction III. (10)) should be reaction with H* , the most abundant radical in the mixture. C F H* + H* 2 4 -V ( C F H ) * X 2 4 2 III. (13) This energy rich ethane molecule could readily undergo HF elimination leading to the formation of C F H and HF : 2 3 ( C F H ) * 2 4 2 F H + HF 3 III. (14) Having considered in some detail the fate of the open chain butyl radical formed in Reaction III. (5), it is now necessary to consider further reactions of cyclic butyl radicals formed by hydrogen and fluorine abstractions (Reactions 111.(4) and III. (6), respectively). In the presence of H* as the major steady— state species, the primary fate of c-C H F • and c-C H F * should be combination, fre4 3 4 4 1* 3 quently followed by decomposition of the energy rich cyclobutanes . c-C H F • + H* 4 3 4 + HF III. (15a) III. (15b) III. (15c) 4 ^ (c-C H F ) * ' 4 4 4

PAGE 84

71 c-C H F • + H‘ 4 ^ 3 f I" H H HC — CF HC — CF F F -> C H + C HF 2 4 2 3 III. (16a) -> CHF + CHF 2 2 2 2 3 M > c-C H F 4 5 3 III. (16b) III. (16c) The relative probabilities of the above six reactions are hard to assess since two of the products (C H F and CHF) 4 3 3 4 5 3 were not observed (since these would probably elute with the parent c-C^H^F^ in the gas chromatograph) and the olefins would be labile under photolytic conditions. However, the relatively large yield of C HF indicates that Reaction III. 2 3 (16a) is efficient. One factor which determines the abundance of products formed through Reactions III. (15) and III. (16) is the amount of precursor radical formed in the mixture. Since the amount of C^HF^ formed is relatively large in these experiments the concentration of c-C^H^F^* must be substantial. Hence, we infer that a substantial fraction of the hydrogen atoms in the system undergo Reaction 111.(6). Like all the other olefins in the system, C^HF^ can undergo H* atom attack. Although C^HF^ may be less reactive than some of the other olefins in the system, at least a portion of it should undergo H* addition followed by HF elimination to give C^HF^*. F F F F C HF + H^ -C = CH ^ -C = C + HF 2 3 ^ I I ^ 1 H III. (17) The product C HF • formed in the above reaction could combine 2 2

PAGE 85

72 with any of the radicals in the system. However, we postulate only Reaction 111.(18) since C H F (I) is the 6 4 6 only such product of combination observed in the mixture. F F ''"^4«3^4’ + 'C = C ^ C^H^F^(I) 111.(18) H With the above reaction scheme in mind, we proceed with the interpretations of results in the photolysis of pure . The major difference between the pure system and the mixture is the relative amounts of the radicals formed. As discussed earlier, H* is the predominant active species in the mixture. However, in the pure system the amount of H* formed initially is equal to that of C^H^F^*. Accordingly, most radicals combine with H* in the mixture, whereas in the pure system there is also a substantial probability of radical combination with c-C^H^F^*, leading to considerably greater production of high molecular weight species. Since C^H^F^would be the predominant organic radical in the pure system, a combination of two such radicals should be expected. 2C H F • ^ C H F 111.(19) 4 3 4 ^868 This type of addition product of two rings has been observed in the mercury sensitized photolysis of both cyclobutane ^ and per f luorocyclobutane , as well as in the radiolyses of both compounds.®''® The C H F in Reaction 111.(19) would , 8 6 8

PAGE 86

73 certainly be vibrationally excited and might need collisional stabilization . In addition to C H F , another C compound is also 8 6 8 8 formed. Since its composition and structure are not known, it is inappropriate to propose a mechanism for its formation. However, we expect it to be a condensation product of C H F * 4 3 4 with one of the many radicals which are formed. V7e propose that CgH^Fg(I) is formed by the same mechanism in the pure system as in the mixture. The large yield of C H F (I) in the pure system is not surprising since it is formed by reaction of C H F • which is present in larger quantities in this case. A second compound with the same empirical formula but (probably) an open chain structure is C H F (II) . One possible route for its formation would be 6 4 6 combination of the CH CH = CFCF • radical from Reaction III. 3 2 (7) with C F •, formed by HF elimination from C HF • subse2 3 2 4 quent to Reaction 111.(10). It can be noted that the postulated mechanism calls for formation of a variety of olefinic products which were not observed. The olefins would be labile under photolytic conditions, and can undergo polymerization . In fact we did see some white solid on the walls of the vessel when a large amount of c-C^H^F^ v/as photolyzed for a long time. Also, we observed the formation of three products heavier than C 8 compounds at longer irradiation times and higher pressures. A final point of difference between the pure system and the mixture is the relatively large yield of C H F in the 3 2 4

PAGE 87

74 former case; this is especially noteworthy since no other odd-carbon species is found in nearly as large a yield (34% relative to HF) . It appears that a contribution by a route other than Reaction 111.(9) is important in the pure system. One such route could be the decomposition of a C compound; two C products are formed with much higher relative yields in the pure system than in the mixtures. A detailed explanation will not be proposed, however, due to lack of further evidence on this point. In summary, we conclude that Hg6(^P^) reacts with c-C H F to cause hydrogen atom production. The resulting free H* atoms react with tetraf luorocyclobutane by hydrogen atom abstraction, fluorine atom abstraction, and addition to the ring (causing ring opening) . The radicals formed in these processes combine with each other or with hydrogen atoms; the resulting energy release frequently causes decomposition of the product molecule, such as elimination of HF or rupture of the ring into two ethylenic compounds. (Radical disproportionation reactions probably also occur but were not specifically identified.) As expected from the postulated mechanism, addition of I scavenger eliminates 2 all products observed in the pure system, and only c-C H F i 4 3 4 is detected.

PAGE 88

CHAPTER IV GAMMARADIOLYSIS OF 1 , 1 , 2 , 2-TETRAFLUOROCYCLOBUTANE IN THE GAS PHASE Introduction 1 , 1 , 2 , 2-tetraf luorocyclobutane differs from both cyclobutane and perf luorocyclobutane in many physical and chemical properties. The thermodynamic stability of tetraf luorocyclobutane is between those of c-C H and c-C F Although "<8 1 * 8 there has been a considerable amount of work on the radiolysis of both c-C H and c-C F as well as a study of the 4 8 4 8 radiolysis of mixtures of the two compounds, there has been no work reported on a partially fluorinated cyclobutane. 1st raf luorocy c lobutane was chosen for this study for several reasons, including the fact that it is empirically equivalent to an equimolar mixture of c-C H and c-C F . 4 8 4 8 This chapter describes a study of the gas-phase garamaradiolysis of pure 1 , 1 , 2 , 2-tetraf luorocyclobutane , including the effect of added oxygen on the reactions of the system. To help understand the mechanisms involved in the radiolysis of this compound, parallel investigations of mercury sensitized photolysis and ion-molecule reactions in tetraf luorocyclobutane were also carried out. 75

PAGE 89

76 Experimental Reagents The 1,1,2,2-tetrafluorocyclobutane used in these studies was obtained from Columbia Organic Chemicals Co. , Columbia, South Carolina; it contained small amounts of impurities of which a C H compound was the most abundant. This 6 6 material was purified by a gas chromatographic technique. A glass column 0.5 m long, 3/8 inch in diameter, and packed with (Analabs) 60—200 mesh silica gel was connected between the storage vessel of c-C H F and a chemical high vacuum 4 4 4 iin®» After the column was evacuated overnight, the vessel was cooled in a dry ice-isopropanol bath and pumped on for an hour . The sample was then allowed to warm up to room temperature. With the arrival of the first traces of gas the pressure in the vacuum line increased. The first portion of the gas entering the vacuum line was discarded. The middle fraction of the purified c-C H F , which did not contain any 4 4 4 impurities that could be detected by gas chromatography, was collected for subsequent use. Matheson Company C.P. grade ethylene was passed through a drying tube of barium oxide into a storage vessel on the vacuum line. Air was removed by repeated f reeze-pump-thaw cycles. Matheson Company research grade oxygen was bled into the vacuum system through a 1/4 inch drying column of 60-200 mesh reagent grade silica gel. Sample Irradiation and Dosimetry The irradiations of the samples were carried out using

PAGE 90

77 a Cobalt 60 gamma irradiator which has been described previously.^^ The dose rate in 400 Torr ethylene was determined by measuring hydrogen production, assuming G(H ) = 1.2.®** 2 Gases non-condensible at -196°C were collected via a Toepler pump, measured on a McLeod gauge, and analyzed for methane and ethylene content on a flame ionization gas chromatograph using a silica gel column at room temperature. Hydrogen analysis was by difference. Energy absorbed in c-C H F was 4 4 4 assumed proportional to S, the electron stopping power per mole, relative to ethylene, and S (c-C H F )/S(c-C H ) was 4 4 4 2 4 calculated to be 3.81. The dose rates were 2.14 x 10^^ eV mole ^ hr ^ in the metal vessel using a 230 Curie source and 19 -1 -1 5.59 X 10 eV mole hr using a 600 Curie source. The dose rate in the glass vessel was 6.68 x 10^^ eV mole"^ hr“^; only the 600 Curie source was utilized. S ample Preparation for Hydrogen and Organic Yield Measurements The radiolysis vessels were made of pure nickel cylinders which were helium arc welded to Monel valves. For unscavenged radiolyses, c-C^H^F^ was directly metered into the nickel vessel which had been pumped down and flamed under vacuum to decompose any solid products formed in the previous radiolysis. Then the vessel was immersed in a Dewar containing liquid nitrogen and pumped on for five minutes to remove any air which might have leaked into the vessel during sample preparation. This procedure was essential to get reproducible yields for the olefins, since their yields are very sensitive to the presence of any scavenger. In the case of oxygen

PAGE 91

78 scavenged radiolyses the sample was introduced as above, and 0 ^ was transferred into the liquid nitrogen cooled radiolysis vessel from a standard volume container filled with 0^ at known pressure. By knowing the volume of the vacuum line, the standard container, and the radiolysis vessel, the pressure of 0 ^ in the radiolysis vessel at room temperature was calculated. In several blank experiments, the amount of 0 ^ metered out as described was checked using a Toepler pump McLeod gauge apparatus; the error was found to be less than 5%. In all the irradiations the pressure of c-C^H^F^ was 50 Torr, and the pressure of the added 0^ was 4% of the total pressure in the vessel. Also, a few semi-quantitative experiments were carried out with I , ^ 2 ^ 4 ' system as scavengers. Sample Preparation for HF Yield Measurements For HF analyses a glass vessel equipped with a break seal was cleaned and annealed at 565*^ C to pyrolyze any remaining organic residues. In the cases of unscavenged radiolyses required amounts of c-C H F were metered by pressure measurements into a standard vessel of known volume and later transferred into the radiolysis vessel cooled in liquid N 2 * The radiolysis vessel was sealed under vacuum. In cases of 0 scavenged radiolyses, the c-C H F was ^ 444 introduced as above, a premeasured amount of O from another 2 standard vessel was introduced into the vacuum line, and the vessel was sealed off. The amount of 0 introduced into the 2 vessel was measured in one blank run utilizing the procedure

PAGE 92

79 described earlier in connection with the metal vessel. Product Identification The formation of hydrogen fluoride in the radiolysis of c-C^H_^F^ was established by potentiometric titration of the irradiated sample using a fluoride ion selective electrode. Hydrogen was identified by its retention time on a 13-X molecular sieve column in a gas chromatograph equipped with a thermal conductivity detector. The organic products were identified by their mass spectral cracking pattern using a gas chromatograph mass spectrometer combination.**^ When possible the identifications were confirmed by comparison of ^stention times with known standards and by matching the mass spectra with spectra given in the API Tables.®’ The NMR spectrum of peak 21 recorded using a Varian XL-100 spectrometer equipped with NTC Fourier transform accessories. The spectrum was scanned 2000 times. In most of these identification procedures, products formed by a spark-discharge in c-C^H^F^ were used, since the yields of the products are larger than in the radiolysis and the tailing by the parent peak could be minimized. (In several previous researches in this laboratory it has been found that products formed in a very weak Tesla Coil discharge of organic gases are qualitatively and even semi-quantitatively similar to those formed during gamma radiolysis ^ ° ^ ^ Standards were available for most of the lighter products analyzed on the silica gel column but not for C^H^F^ or C^H^F^; there were no available standards for several of the products measured on the SE-30 column

PAGE 93

80 In such cases, product identification is somewhat tentative. Product Analyses 1) Hydrogen fluoride . Hydrogen fluoride was determined by poten tiome tr ic titration using a fluoride ion selective electrode. The details of the experiment are given in connection with the photolysis of tetraf luorocyclobutane (Chapter III). Addition of chloride ions to the analyte, necessary for the photolysis work due to the presence of mercury, was not required for the radiolysis samples. 2) Hydrogen . In the unscavenged system, hydrogen was measured as in the case of the dosimetry experiments. However, in the oxygen scavenged experiments, all gases noncondensible at —196 C were Toeplered into a McLeod gauge, snd this mixture was analyzed for both the methane and hydrogen content using a gas chromatograph equipped with a flame ionization detector (FID) in tandem with a thermal conductivity detector (TCD) . The FID was used for measuring the amount of methane and the TCD for that of hydrogen. The column used was a 13-X molecular sieve column mentioned above, and the carrier gas was nitrogen. 3) Low Molecular Weight Products . The low molecular weight organic products (all compounds with one or two carbon atoms and one compound with three carbon atoms) were analyzed on a 4.6 m stainless steel column packed with 60-80 mesh (Analab) silica gel using a Tracer 550 research gas chromatograph equipped with a flame ionization detector. 4) High Molecular Weight Products . The high molecular

PAGE 94

81 weight organic products (molecules with three or more carbon atoms) were analyzed on a 12 m glass column packed with Chromasorb-W coated with 30% SE-30. The same Tracer gas chromatograph was used. In the chromatographic analysis of both low and high molecular weight organic products, helium was used as the carrier gas. The molar response of sll and compounds in a flame ionization detector were obtained by using standards which were available. However, the molar response factors of the compounds were estimated from the response of similar compounds. Results In all, 24 products were identified and their G values measured in the radiolysis of c-C H F . The inorganic species, 4 4 4 hydrogen fluoride and hydrogen, were the most abundant products. The yields of both hydrogen fluoride and hydrogen were linear with dose, within experimental error (Figures (16) and (17) , respectively) . A typical gas chromatogram of the products eluting on the silica gel column is shown in Figure (18) . The three major olefins, 1 , 1-dif luoroethylene (1,1^ 2 ^ 2 ^ 2 ^ ' and ethylene are shown in Figures (19) and (20). ^^^4 produced in negligible quantities. The yield of C F showed a small induction period and later 2 4 increased linearly with dose. The yield of 1,1-C H F was 222 small and the dose-yield plot did not extrapolate to zero yield at zero dose. However, an unirradiated sample did not show even a trace of 1,1-C H F . The yield-dose plot of 222

PAGE 95

Micromoles of hydrogen fluoride 82 Figure 16. Production of hydrogen fluoride (pure, Q ; function of dose in the radiolysis of c-C^H^F^.

PAGE 96

83 c H 0 ) U) o 4-1 o c o •H -p u c 3 4 -( (d U) n 3 o oV QJ P • P JCifa •>—4 C J(1) u Cn I O U P T) >4-1 >1 O tn IW -H O u) >1 C -H o o •H *H P T 3 O fO ^3 P T! O 0) P 4:: CU -P r~ 0 ) p 3 Cn *H [p uaboapAq jo sepouiojOTH

PAGE 97

84 LU 12 o ltI OJ . c c to g -P 3 3 rH Xi o O O •H O H >1 (U u o o P 3 0 CJ 3 -H r— ) iH 4-1 -H 3 U3 U -P to (U •P 1 (N ^ O' (N C c o I — I -p 3 1 — I I — I 0) 03 OJ W P +J 3 U •H 3 03 03 3 O U P u & •H •p 4-) 43 O tji •H e oj 3 s p Cri P O 3 -P H 3 3 e u O 0) p I — I to •H o IP CM IP 53 j53 U I U CTi U I CM IP c K c U IP cn c 3 P p) 53 < U 43 0 rH t— H hi-l 53 1 CN U s CM CM iH u u iH 3 3 0 O rH • • • • in VD r-00 • 00 i-H
PAGE 98

Micromoles of products 85 Figure 19. Production of 1,1-C2H2F2 (pure Q ; 4% O 2 , # ) and (4% 02 /.^^) as a function of dose in the radiolysis of c-C H F . 4 4 4

PAGE 99

86 o c o •H 4-) o o • c m 3 4-1 (0 Ui (0 O OP O U • 1 04 — w OJ >i rH c Cm 0 •H CM-H u ^ 0) to CO 4-1 !m 0 0 Q 0) G x: O 0 4-) • •H rH 4-1 C u T3 o o o o CN (U u 0 tj> •H Cm o o i z d D JO saxouiojOTw

PAGE 100

87 acetylene (Figure (21) ) showed a distinctive break in slope, the yield increasing at a higher rate at longer doses. The G values of all the products eluting on the silica gel column are shown in Table VIII. A chromatogram of the products eluting on the SE-30 column is shown in Figure (22). The first three peaks were made up of all the products analyzed on the silica gel column except C^H^F^. Products 9 and 11 through 14 were identified to be C^H^F^, C^H^F^, C^H^F^ and C^H^F^, respectively, based on their mass spectral cracking patterns, obtained using the GC/MS combination as mentioned earlier. The details of the identification are given in Appendix I. Peak 15 was small and neither the composition nor structure of this product could be determined. Peak 16 is probably c-C H F . Peaks 4 2 4 * 18 and 19 were very small and in most of the runs smeared out by the tail of the parent peak. The ^^F NMR of Peak 21 is shown in Figure (23) . The proton NMR showed the presence of the olefinic hydrogen. Based on the NMR and the mass spectral data Peak 21 was assigned to be: Peak 22 is non-cyclic as shown by its proton NMR. A discussion of the structural determination of many of the other compounds is given in Appendix I. The G values of all the products measured on the SE-30

PAGE 101

88 Figure 21. Production of C 2 H 2 (pure Q ; 4% 02,H ) as a function of dose in the radiolysis of c-Ci,H^F^.

PAGE 102

89 TABLE VIII G Values in the Radiolysis of 1 , 1 , 2 , 2-Tetraf luorocyclobutane : Products Eluting on the Silica Gel Column Product Pure System O 2 Scavenged System Low Dose^ High Dose Low Dose High Dose H 2 0.52 0.52 0.205 0.682 HF 2.3 2.3 8.6 8.6 CH 4 0.0082 0.0082 0.008 0.0 C H 2 6 0.0057 0.0129 0.0057 0.0057 C F 2 4 0.093 0.171 0.109 0.109 C H 2 4 0.0 0.0 0.053 0.125 1,1-C H F 222 ('-.47) 0.0464 0.471 0.471 C H 2 2 0.0853 0.196 0.0186 0.114 C H F 2 3 0.0125 0. 0 0.024 0.0397 1,2-C H F 222 0.0071 0.0071 0.0 0.0 C H F 3 4 2 0.023 0.0487 0.0 0.0 1,2-C H F 222 0.0546 0.0546 £0.001 £0.001 C HF 2 3 0.01^ 0.01^ b — C H F 3 2 4 0. 0035 0.0165 £0.001 £0.001 CHF 3 0.01 0.01^ — — ^Yields measured at doses below 1 x 10^^ eV (2 x 10^^ eV in the case of C H ) . 2 6 Dash indicates product not measured. Q C 2 HF 3 eluted with C 2 H 2 /' yield estimated from mass spectrometric analysis . CHF3 eluted with 1 , 1 -C 2 H 2 F 2 ; yield estimated from mass spectrometric analysis.

PAGE 103

90 o 0) C fO -P 3 • XI a o E H 13 U rH >1 O U O O P O O rn 0 I H w M-i cn (0 +j 04 O e 43 g •H MH tP 0 tn 0 -H u TO CU c G e s CO 3 O (d u 0 •H p p 1 a o 4J ct> (d JCO lO 6 ro :3 0 rH u Pm Pm O rH 4-) 3 jJ0 w 3 U K X g (U • J rj CM cn X 33 Ch M-4 3 cn jJP u U u u 3 Cn *H • • « • O Pm 1— ^ CN ro rH rH rH iH esuodsen xo:;oeq.aa

PAGE 104

104.0 115.6 110.8 114.2 117.7 121.4 128 91

PAGE 105

92 column are shown in Table IX. The yield— dose curves for C H F (Peak 14) and C H F (Peak 21) are shown in Fiqure (24). The yields of both C II F and C H F showed an 4 3 3 6 4 6 induction period and at higher doses increased linearly with dose. Addition of oxygen to the system substantially altered the G values of all but two of the major products, C H F 4 3 3 and C F . The G value of C H increased considerably (from zero to 0.125). The yield showed an induction period and increased linearly with dose above 0.5 x 10^^ eV. The most dramatic change was in the yield of 1,1-C H F , which increased 2 2 2 ten-fold from 0.046 to 0.471. The yield of this olefin increased linearly with dose. The acetylene yield was decreased upon addition of oxygen. The yield of methane, a minor product, showed an interesting plateau above 1 x 10^^ eV in the presence of oxygen. The G values of products eluting on the SE-30 column in the oxygen scavenged system were small; moreover, they elute between unidentified oxygenated compounds. Hence, we could only estimate an upper limit for these yields. The G values of HF in C^H and I scavenged systems were 1.6 and 1.8, respectively. The G values of all measured products in the scavenged system are given in Table X. Material Balance Material balance is not very satisfactory in this system. The ratio of C/H/F is 4. 0/6. 2/6. 3. It is evident that there is a shortage of carbon. We attribute this shortage mostly

PAGE 106

93 fO CTi > QJ C •H OJ U 5 O Q 1 I — ! o ' — ' -H Og (D M (U P E a-p E ro E j•H 0 ) m o U T) 4 ^ 4 -^ o o c o •H p p u o p 'd o P u E d CN (1) P P tJi *H E peuiao5 sd-onpoad jo sepouioaoxH

PAGE 107

TABLE IX G Values in the Radiolysis of 1 , 1 , 2 , 2-Tetraf luorocyclobutane Products Eluting on the SE-30 Column Product Pure System 0^ Scavenged System Low Dose High Dose C H F^' 3 4 2 0.032 0.047 0.005 C H F 3 2 4 0.013 0.021 0.005 C H F 4 2 4 0.0062 0.017 <<0.001 C H F 4 2 2 0.013 0.013 <<0.001 C H F 4 3 3 0.063 0.197 0.13^ C H F 5 6 4 0.007 0.05 c 7 0.01 0.01 — ? — 0.05 — C H F 6 4 6 0.05 0.09 -0.05^ C H F 6 4 6 0.02 0.06 ~0.05^ ? 0.0 0.015 <<0.001 C H F 8 6 8 0.01 0.0 0.0 Measured on silica gel column also. Based on the first three points. (The yield at longer doses becomes zero and then becomes negative.) c Dash indicates product not measured. ^Irreproducible .

PAGE 108

95 TABLE X G Values of Products in 1 2 Scavenged Radiolysis of 1,1,2, 2-Tetraf luorocyclobutane Product G Value HF 1.8 H 2 0.28 C H <0.005 2 6 C F 2 4 0.10 C H 2 4 0.01 1 , 1 -C H F 222 0.60 C H 2 2 0.14 C H F 2 3 0.09 C H F 433 0.12 The G value of HF in C H scavenged system is 1.6.

PAGE 109

96 to the formation of polymers which were not analyzable by the methods employed in this research. The formation of polymers was indicated by the evolution of a large amount of fluorinecontaining organic fragments when a vessel which had been irradiated with the compound inside it was attached directly to the mass spectrometer and flamed strongly with a hand torch. We expect some contribution to this carbon deficit from un^^^.lyzed products which may have eluted with the parent compound. Discussion From the results of our studies of ion— molecule reactions in c-C^H^F^ it is evident that the most abundant positive ion at higher pressures is C^H^f"^. If we assume a typical W value of 33 eV/ion pair for this system, the value of 2 ^ t>e approximately 2.5 because the abundance *^ 2 ^ 2^2 higher pressures is of the order of 85% of the total ion intensity. A W value of 33 would imply a G value of 3 for the decomposition of c-C H F through ionic primary processes. The value G(-M) = 3 is in accordance with that expected for a partially fluorinated hydro-carbon where C-C and C— H bond ruptures predominate and the net decomposition is moderate. Stoichiometric considerations lead to a G value of 2.5 for C^H^F^ in the primary step. c-C H F ^VA/V^ C H f"^ + C H F + e IV (1) 444 222 222 xv.vx; Whereas C^H^F^ is observed to be the predominate positive

PAGE 110

97 ion in the system, theoretical arguments indicate that F~ must be the most abundant negative entity. The electron affinity of fluorine atom is so large (3.6 eV) that dissociative electron attachment leading to the formation of F would be a very low energy process. > C H F • + F" IV. (2) 4 4 4 4 4 3 A reasonable value for the C-F bond strength in a cyclobutane would be about 110 kcal/mole. Using this value of bond strength and the electron affinity of Fwe calculate the threshold of Reaction IV. (2) to be not greater than 2 eV. Hence, the neutralization of these two ions, C H f”^ and f", 2 2 2 is expected to play a major role in the radiolysis. C H F+ + f" ^ (C H F )* ^ C HF • + HF IV. (3) This reaction could also proceed directly via h"*" transfer to F (i.e., a stripping process): C H F^ + F ^ C HF • + HF IV. (4) 222 ^ 22 The HF formed in Reactions IV. (3) and IV. (4) would be unscavengeable by radical scavengers since it is formed in an ionic acid-base reaction. If we attribute most of the unscavengeable HF to this reaction, the G value of F~ would be approximately 1.5, which appears reasonable. The C H F not neutralized by F~ should be neutralized 2 2 2 by the only other plausible negative species in the system. the electron.

PAGE 111

98 ^ > ^^2^^2^* IV. (5) The magnitude of G(C^H^F^)* would be about 1.0 since the rest of the (G=1.5) would have been neutralized by F . The excited C^H^F^ formed in Reaction IV. (5) should have an excess energy of 10.12 eV (the ionization potential ^ 2 ^ 2 ^ 2 ^’ amount of energy is sufficient to break any of the bonds in C^H^F^. However, the C=C double bond is considerably stronger than either C-F or C-H (145 kcal versus 108 and 98 kcal, respectively) and its rupture appears unlikely in spite of the considerable energy excess. (C H F ) * ^ (C H F ) * ^ In dif luoroethylene the C-F bond strength should be about 110 kcal/mole and the C-H bond strength about 95 kcal/mole. The internal energy of (C^H^F^)* would be so large compared to the difference in the bond strength of C-F and C-H bonds that there might be very little specificity in breaking of one bond over the other. However, we expect Reaction IV. (6) to be somewhat more probable than Reaction IV. (7) since C HF 2 2 would be more stable than C^H^F. Another possible mode of decomposition of (C H F )* would be HF elimination 2 2 2 C HF • + HIV. (6) 2 2 "^2^^^* ^ IV. (7) (C H F ) * ^ H-CeC-F + HF 2 2 2 IV. (8) It is interesting to note that we did see a product whose mass spectrum was similar to that of C HF , ® ^ but the yields

PAGE 112

99 of this compound were very small and irreproducible , probably due to its lability. Since the postulated G value of (C H F )* is 1.0, the 2 2 2 G values for production of F* and C HF • in Reaction IV. (7) 2 2 would be less than 0.5 and those for H* and C HF • in Reaction 2 2 IV. (6) would be greater than 0.5. Hence, the total G value for C^HF^in Reactions IV. (3) and IV. (4) should be greater than 2.0. We will come back to the discussion of the fate of these radicals later in this section. Equation IV. (1) shows the "symmetrical" rupture of parent c-C^H^F^ into two dif luoroethylenic species. The other two olefins, C F and C H are also certainly formed by the direct decomposition of c-C H F (in either an ionic or ex4 4 4 cited state). The increase in the G value of 1,1-C H F 2 2 2 (from 0.046 to 0.47) and C H (from 0.0 to 0.125) upon 2 4 addition of oxygen to the systems indicates that these two olefins are labile under radiolytic conditions. However, even the measured yields in the 0 scavenged system do not 2 reflect the primary yields of these species, especially that of C^H^F^. The relatively small yields of ethylenic products indicate that there may be polymerization of these compounds under free radical attack leading to products not detected in our experiments. As discussed in the section on material balance, we did see a large amount of fluorinated compounds coming off the vessel walls when heated, indicating the formation of polymers. Oxygen is known to react with fluorinated radicals to form peroxy radicals which usually decompose

PAGE 113

100 to give other radicals.®^ These other radicals can initiate the polymerization even in the presence of oxygen. Hence, the small yield of olefins in the presence of 0 is not 2 very revealing. In the pure system, however, we expect C HF • (the most 2 2 abundant radical in the system) to initiate polymerization of the olefins. C H F + C HF 2 2 2 2 2 C H F • 4 3 4 C H F + C H F • 2 2 2 4 3 4 -> C H F • 6 5 6 Chain termination reactions would include 2 C HF ' 2 2 -> C H F 4 2 4 C HF • + C H F ' 2 2 4 3 4 C H F ^ 6 4 6 2 C H F 4 3 4 C H F 8 6 8 IV. (9) IV. (10) IV. (11) IV. (12) IV. (13) — etc. — The yield-dose plots of C H F and product 23 level off at 8 6 8 higher doses indicative of further reactions. This type of polymerization has been seen in many olefinic systems, such as C F ^ and C H A similar polymerization scheme 2 4 2 4 has been postulated in the radiolysis of c-C F . ^ “ in 4 8 addition, it is very likely that there could be ionic polymerization of C^H^F^ similar to that seen in the ethylene system. Although the limited data available do not justify further speculation concerning polymerization processes, the

PAGE 114

101 proposGd reaction scheine does provide a reasonable explanation of the formation of several products including C H F , C H F , C H F , etc. 8 6 8 In the presence of oxygen, the yields of 1,1-C H F and 2 2 2 are markedly increased relative to the pure system. In the unscavenged system the yield of C F shows an initial 2 4 linear region with G = 0.093, shifting to a second essentially linear region with the considerably larger yield of G = 0.171 at high dose. With added oxygen, the tetraf luoroethylene yi^ld is linear throughout, with G = 0.109. Since a direct route to production would be independent of dose range, it is apparent that there must be a second route to tetra— f luoroethylene production which is blocked by the presence of scavengers. It should be noted that the yield of C H F 2 2 2 is much greater than those of either C F or C H • the 2 4 2 4 difference is considerably beyond the statistical expectation of 2:1:1. An interpretation of the relatively large yield of *2 ^ 2 ^ 2 ' I’usod upon unimolecular fragmentation theory, is given in Appendix II. The decrease in the yield of in the I scavenged system (G(H^)=0.28) and at lower doses in the oxygen scavenged system shows that a part of the hydrogen yield is scavengeable . The remainder must come from molecular elimination reactions, ionic processes, or hot hydrogen atom abstraction from the substrate. C H F * formed either by direct excitation or 4 4 4 usutralization of C H F could undergo a C— H bond scission to 4 4 4 give either hot or thermal atoms.

PAGE 115

102 c-C H F 'VW— ^ (C H F )* 4 4 4 441, C n f'*' + e“ ^ (C II F )* 4 4 4 ^ 4 4 4 (c H F )* ^ C H F • + H4 4 4 ^ 434 IV. (14) IV. (15) IV. (16) The sequence (Reactions IV. (14) and IV. (16) ) may be moderately important but Reactions IV. (15) and IV. (16) probably contribute very little to the hydrogen yield, in comparison to Reaction IV. (6) , since C H F ion has a low abundance in the 444 mass spectrum of c-C H F . 444 The molecular processes leading to the formation of H 2 would include (C H F )* C H F + H 444 4 2 4 2 and (C H ) * >c H + H 2 4 2 2 2 IV. (17) IV. (18) Molecular hydrogen elimination by excited ethylene is well known. In Chapter I, we have postulated that c-C H f"*" under444 goes fission through a concerted reaction to give C f"^ and + ^ C 2 H 4 or C^H^ and C^F^. The excess energy content of the reaction intermediate would be probably greater than the endoergicity of the reaction and would leave behind an excited ^ 2^4 ^ 2 ^ 4 * , Reaction IV. (18) (or its ionic analog) could certainly follow. Since we contend that most of the acetylene yield is formed through hydrogen elimination by ethylene, the sum of the yields of C H and C H F should 22 2 2 4 approximately equal the yield of H in the scavenged system;

PAGE 116

103 in fact, this is found to be true. We do not have a definite explanation for the increase in hydrogen yield at larger doses in the presence of oxygen; this may be due to the reactions of oxygenated radicals. Hydrogen fluoride is the most abundant product in the radiolysis of c-C H F with G(HF) = 2.3. About 25% of this 4 4 4 yield is scavengeable by either C H or I ; this HF must come 2 4 2 from the reaction of H* atoms formed in reactions such as in Equation IV. (6) . H+ c-C H F ^ hF + C H F • IV. (19) 4 4 4 4 4 3 The increase in HF yield in the presence of oxygen is most likely due to oxygenated radicals. It is known that 0 reacts 2 with F, H* , etc. forming very reactive radicals,"*^ which can attack the substrate leading to the formation of CO, FCOF , etc., many of which will end up on the wall as fluorides. The unscavengeable HF is probably formed mainly through molecular HF elimination, ionic reactions, and abstraction by F atoms. Fluorine atoms are certainly formed in this system and could react with the substrate through either Hatom abstraction or ring opening. The HF formed by H* atom abstraction would be unscavengeable by any additive at moderate concentrations, since the reaction is exoergic by approximately 25 kcal; there is no reason to expect an unfavorable steric factor. The four— centered hydrogen halide elimination is well known.®® This HF elimination, presumably by C H F * , would 4 4 4 give either a cyclic C H F or a straight chain isomer of 4 3 3 C H F . 4 3 3

PAGE 117

104 (C H F )* ^ C H F + HF 4 4 4 4 3 3 IV. (20) The product identified to be C H F (product 14) is formed 4 3 3 mostly by HF molecular elimination as suggested by the fact that its G value is not substantially altered in the presence 1,1-C H F additives. However, we contend that most of the unscavengeable HF comes not from molecular eliminations but from ion neutralization (Reactions IV. (3) and IV. (4) ) . This ionic yield would also be unscavengable by C H and I . 2 4 2 There are two compounds and the sum of their yields [G (sum)=0. 068] is very much greater than the sum of the yields of compounds. However, if product 18 is formed through Reaction IV. (25) only, then the sum of the yields of C H F 5 6 4 ' CH^, and CHF^ would be 0.068 which shows mass balance for these products. In this argument it is assumed that C H F 4 4 4 undergoes a scission to give a C and a C species 1 3 (C H F ) * 4 4 4 (C H F ) * 4 4 4 C HF • + CH 3 4 3 C H F+ CF 3 4 3 IV. (21) IV. (22) CH • + c-C H F 3 4 4 4 CF • + c-C H F 3 4 4 4 -> CH + C H F • 4 4 3 4 -> CHF + C H F ' 3 4 3 4 IV. (23) IV. (24) CH • + C H F • 3 4 3 4 C HF • + c-C,H o 4 4 C II F+ c-C H 3 4 4 > CHF 5 6 4 — > C,H,F, CHF 3 4 2 + C^H3F + C H F 4 3 IV. (25) IV. (26) IV. (27)

PAGE 118

105 It is very likely that other abstraction, combination and disproportionation reactions occur. Formation of quite likely proceeds through hydrogen atom reaction with C H . 2 4 C H + H2 4 IV. (28) C H • + c-C H F 2 5 4 4 1, C H + C H F 2 6 4 3 4 IV. (29) The source of these hydrogen atoms would be Reactions IV. (6) and IV. (16). We cannot rule out a two step ring opening/ fragmentation under hydrogen atom attack to form C H • 2 5 in the case of mercury sensitized photolysis. The amount of C^H formed is too large compared to CH to consider two 4 methyl radicals combining to give C H as a significant 2 S process. The formation of other minor products, C H F, C F H, 2 3 2 3 ^'^~^ 2 ^ 2 ^ 2 ' ^ 4 ^ 2 ^ 2 ' explained either by the reaction of C HF • or C H Fabstracting either a For H2 2 2 2 from the substrate or by radical combinations. C HF • + 2 2 c-C H F 4 4 4 > C HF 2 3 + C H F • 4 4 3 IV. (30) C IIF • + 2 2 > 1,2-C H F^ + C H F • 2 2 2 4 3 4 IV. (31) C H F* + 2 2 c-C H F 4 4 4 > C H F 2 3 + C II F • 4 3 4 IV. (32) C II F* + 2 2 C HF • 2 2 — > (c H 4 F ) * 3 3 V C H F + HF IV. (33) The radical C^H^Fcould also abstract a hydrogen from the

PAGE 119

106 substrate . C H F • + c-C H F ^ CHF + CHFjy ( 34 ) However, we have not seen this CHF compound. It probably elutes with the parent as discussed in the section on mass balance . A comparison of the radiation chemistry of cyclobutane, 1 , 1 , 2 , 2-tetraf luorocyclobutane , and perf luorocyclobutane is of interest in understanding the effect of fluorine atoms in a cyclobutane ring. (It is to be noted that, unlike the cases of other halogens, the C-F bond is stronger than the C-H bond.) The stability of the cyclobutanes increases with the increase in the number of fluorine atoms in the compound (i.e., -AH. (c-C H ) < -AH. (c-C H F ) < -AH^ (c-C F )),^^ “ which is reflected by the appropriate trend in the net radiolytic decomposition of these compounds, since G(c-C H ) 4 8 4, G(c-C H F ) 3, and G (c-C F ) < 3. One of the 4 4 4 4 8 primary modes of decomposition in all three cyclobutanes is a scission of the ring to give two olefins. This scission is most prominent, in c-C H and least in c-C F . (In both 4 8 4 8 the c-C^H^ and c-C^F^ systems only one olefin can be formed by the concerted rupture of the ring, unlike c-C H F where 4 4 4 three different olefins, C H , 1 , 1-c H F , and C F , are 2 4 2 2 2 2 4 produced. ) Another similarity is the ring opening under radicalatom attack which takes place in all three above mentioned cyclobutanes. Both Hand Fatoms can be formed in

PAGE 120

107 tetraf luorocyclobutane , and both the atoms are expected to open up the ring. There is one very marked difference between c-C H F 4 4 4 on the one hand and c-C H and c-C F on the other, in that 4 8 4 8 HF is the most abundant product in the c-C H F system, 4 4 4 while it can not be formed in the other two cases. The system resembles other partially fluorinated alkanes in the formation of HF as a major product of radiolysis . ^ °

PAGE 121

APPENDICES

PAGE 122

APPENDIX I IDENTIFICATION OF PRODUCTS In this sBction the deduction of the composition and/ where possible, the structure of the products formed in the radiolytic and/or photolytic studies is discussed. The mass spectral cracking patterns of the products formed in a sparkdischarge of c-C^H^F^ are the main source of data. A second general approach is based upon the injection of a series of known compounds on a SE-30 GLC column and determination of the resulting elution times. This procedure was helpful in determining the number of carbon atoms in each product, since elution times on a SE— 30 column are dependent mainly on the points of the compounds. Thus, it was possible to correlate the elution times with the carbon numbers of the compounds. In several instances, comparison of the elution times on two or more gas chromatographic columns was helpful in product identification. This procedure was particularly ^nluable in establishing the olefinic character of several low molecular weight products, which are selectively retarded on a column made of propylene glycol saturated with silver nitrate. Comparison of tentative product identifications with predictions of a plausible reaction mechanism is often 109

PAGE 123

110 informative, but this approach must be used with caution since it involves cyclical arguments. Ultimately, however, most of the identifications depend on mass spectral cracking patterns. Unfortunately, only a few partially fluorinated compounds are listed in the API mass spectral tables” or in Majer's article on the mass spectroscopy of fluorine compounds. Accordingly, several of the product assignments listed below are tentative. In all, 24 organic products were detected in the gammaradiolysis of 1 , 1, 2 , 2-tetraf luorocyclobutane. Since this group includes all of the products formed in the mercury sensitized photolysis, it is possible to discuss product identification comprehensively with reference to the radiolysis experiments alone. In the case of the majority of the products which eluted on the silica gel column, standards were available and product identifications are definite. Included in this group are the following: Peak 1, CH ; Peak 2, C H ; Peak 3, 4 2 6 C F ; Peak 4, C H ; Peak 5, 1,1-C H F ; Peak 6, C H ; Peak 7, 24 24 222 22 Peaks 8 and 10, 1,2-C H F (trans and cis isomers, 2 3 2 2 2 respectively) . The section below describes our conclusions concerning the identify of the remaining radiolysis products, all of which can be separated on the SE-30 column. As mentioned above, the identifications are largely based on mass spectral cracking patterns. Good spectra were obtained for Peaks 9-14 and 21-24; these are listed in Table XI. The spectrum listed for Peak 10 is actually that of a known sample of cis 1,2-

PAGE 124

Ill difluoroethylene. It was not possible to obtain useful spectra of Peaks 15-20 due to interference by the parent tetraf luorocyclobutane . Peak No. 9, C ^H^F^. Peak 78 clearly corresponds to the molecular ion. The loss of a hydrogen atom or a fluorine atom gives Peak 77 or 59. The loss of both fluorine atoms gives Peak 40 and a loss of CF^ (50) gives the base peak. This compound was retained by silver nitrate column indicating that it is an olefin. Peak No. 10, cis 1,2-C H F . A standard was available for — ^ 2 — 2—2 this compound. The mass spectrum and the elution time of this product corresponds to those of the standard. Peak No. 11, c-C ^H^F^. The base peak is the parent peak. Elimination of a CF^ group (50) gives Peak 64 which is quite intense. This compound was not retained by the silver nitrate column and hence is not an olefin; the only plausible identification is c-C^H^F^. Furthermore, the moderate abundance of the parent ion indicates the cyclic nature of this compound. Peak No. 12, C H F . The base peak is C H f"^, formed by 4 2 4 3 2 the elimination of CF^. The parent ion is at 126 and the loss of 19 (fluorine) gives 107. This type of fluorine loss is very common in fluorocarbon mass spectral cracking patterns. The retention time of this compound on the SE-30 column shows that it is species. The compound is not 1 , 1 , 2 , 2-tetrafluorocyclobutene since the retention times of the two do not match.

PAGE 125

112 Peak No. 13, We are not sure of the composition of this product. Its yield was too small to obtain a good spectrum. It eluted between two C compounds on the SE-30 4 column and hence we assume it to be a C species. The 4 heaviest peak is at mass 88 and therefore we suggest that it has the empirical formula C H F . 4 2 2 Pea k No. 14, . The heaviest peak is at mass 108, the parent peak. The base peak is at mass 50, and is probably CF or more likely a combination of CF"^ and C h’*’. The peak ^ ^ 2 4 2 at 57 (C^H^F ) is formed by the elimination of CHF , probably indicating the presence of such a group in the molecule. The absence of a peak at 64 (C H f"^) indicates that this is 2 2 2 not 2,3,3 -trif luorocyclobutene . Peak s 15-2 0 . These products eluted either just before, or on the tail of the parent compound. Hence, we could not obtain a useful mass spectrum of any of them. Since Peaks 15 and 16 eluted just before the parent, a C compound, we 4 assume that these two are C species. Peak 20 could be a 4 species with the others being either or C compounds. Peak No. 21, C HF. The heaviest ion is the parent ion, *3 H F . The base peak is at mass 89 (C H F^) formed by the 4 3 2 elimination of F during formation of the radical ion C H f"*” 4 3 3 (elimination of fluorine atom is also possible). The second largest peak is at 127 (CHF) which is formed by the elimination of C HF from CHF. The similarity between the cracking pattern of this compound and that of c-C H F indi4 4 4 cates the presence of a cyclic ring similar to that in c-C H F

PAGE 126

113 The absence of a peak at 28 (C H ) and the presence of peaks at 100 (C F ) and 64 (C H F ) indicates that there 2 4 2 2 2 is no C H group in the molecule (like that in c-C H F ) 2 4 1 . 1 . ' and that c-C^H^F^ is a part of this compound. This information leads us to write the structure of Peak No. 21 as firins this structure and indicates that the two fluorine atoms in the side chain are on two different carbon atoms in a position cis to each other. Hence, the structure of this product is: Peak No. 22, C ^H^F^ . The cracking pattern of this compound and the one in Peak No. 21 are very similar except for the base peak which is one mass heavier. This compound was shown to be linear by its proton NMR spectrum. Peak No. 24, a compound . The mass spectral cracking pattern of this compound is not very helpful in assigning a structure for it. The abundance of this product was not sufficient to obtain an NMR spectrum for it. The elution time on the SE-30 column indicated that it is a C compound. 8 Peak No. 24, C H F . The base peak is at mass 89 and the heaviest peak is at 127. This spectrum is very similar to that of c-C^H^F^, Peak No. 21 and 22. This indicates the presence of c-C^H^F^ group in the structure of this compound. 19 ^~*' 4 ^ 3 ^ 4 ~^ 2 ^^^ 2 ' ^ spectrum of this compound conF H F F F-C C C=C H

PAGE 127

114 Its elution time shows that it is a C compound. 8 we conclude that it is C H F , a dimer of C H F 8 6 8 4 3 4 Hence ,

PAGE 128

Mass Spectral Cracking Patterns of Products 115 X csj (tl I (N QJ CU 00 CN (N CO I iH I — I I I to ( 1 ) a, H to 0 ) a, CN I I I VO I I I I I ~4 CM Cm Cm K H F 2 CO u U U u u U CO nj ro CM ro rn ro ro ro CO JCm CM K S pL4 E K CO fO CM CM CM u C c u u
PAGE 129

116 00 o ^ O I H I I I I O rH I I I CN I (M IT) I 1 O CN 1 1 o 1 1 1 1 fH \D cn 1 1 1 1 CO 1 1 1 1 1 1 1 rH 1 1 1 1 CO (N LTi in in in ro HI I H I I O I O I I I I r I — I I I H I I
PAGE 130

117 TABLE Xlb Mass Spectral Cracking Patterns of Products m/e Assignment Intensities^ Peak 21 Peak 22 Peak 23 Peak 24 14 CH 2 3 27 c n 2 3 7 7 17 1 28 C H 2 1* 13 39 C H 3 3 12 13 • 19 19 51 CHF 2 18 18 58 25 54 c n 4 6 9 57 CUE 3 2 14 19 59 C II F 3 4 39 64 C II F 2 2 2 , 10 21 88 36 69 CF , C H F 16 14 3 4 2 75 C HF 3 2 14 15 77 CHF 3 3 2 4 15 89 CHF 4 3 2 100 100 90 CHF 4 4 2 100 95 CHF 3 2 3 20 23 7 24 100 C F 2 4 9 8 29 _ 113 C HF 3 4 12 100 121 C HF 5 4 10 127 CHF 4 3 4 53 53 43 141 CHF 5 5 4 21 177 CHF 5 '3 6 21 190 CHF 6 4 6 10 12 Intensities normalized to the largest peak=100.

PAGE 131

APPENDIX II UNIMOLECULAR FRAGMENTATION OF CYCLOBUTANE COMPOUNDS The thermal decomposition of cyclobutanes (cyclobutane itself and substituted cyclobutanes) into two olefinic products has been extensively studied. Benson has shown that these results are in agreement with the formation of a tetramethylene biradical intermediate followed by a C-C bond rupture to give two olefins. This sequential two step scission is aided by the electronic rearrangements in the tetramethylene biradical. Also, this postulate is compatible with Woodward-Hof fman rules as the C symmetry 2 of the cyclobutane ring can be conserved throughout the reaction . We expect excited c-C H F to follow a similar decom4 4 4 position scheme even under radiolytic conditions. In radiolysis the secondary processes follow thermal equilibrium kinetics, unlike the primary processes. Assuming that one of the species formed in the primary processes is an excited c-C H F molecule (c-C H F ) * the above mentioned scheme of 4 4 4 4 4 4 reaction, and symmetry considerations, can be applied in explaining the amounts of olefinic end products formed in the radio lysis . We can apply the Woodward-Hof fman rules to the 118

PAGE 132

ii9 decomposition of (c-C^H^F^)* assuming it to have a C symmetry like the ground state c-C H F . As the C axis is the only 4 4 4 2 element of symmetry, this symmetry has to be conserved throughout the reaction. Therefore, the cleavage to give 2 CH^CF^ molecules would be symmetry allowed. Even though the symmetry rules would not favor the formation of C^F^ and through a sequential mechanism, it cannot be ruled out. Accordingly, the biradicals which can be formed are: •F C-CH -CH -CF • (I) 2 2 2 2 •H C-CF -CF -CH • 2 2 2 2 (ID •H C-CH -CF -CF • (III) 2 2 2 2 depending on which C-C bond in (c-C H F ) * breaks first 4 4 4 abundance of these three biradicals would depend on the energetics and the kinetics of the reactions leading to these biradicals and the availability of energy in (c— C H F ) * 4 4 4 to drive these reactions. The availability of energy for these reactions would depend on how much energy is channelled the particular side of the molecule containing the C— C bond to be broken in (c-C H F )*. The kinetics of the 4 4 4 reactions accordingly depend on the energetics of the system and in particular the strengths of the bonds to be broken. A qualitative discussion of these two factors, energy availability and kinetics, is given below. Even if energy is put into the molecule initially as

PAGE 133

120 excitation energy , or via recombination of positive ions with negative ions or electrons, most of it would end up as vibrational energy via the usual route of internal conversion to a high vibrational level of the ground electronic state. There are nine active vibrations in Four of these vibrations, which are the most active, are C—F stretches. Thus, the fluorine containing side of the molecule will get more energy than the hydrogen containing side. A simplified Kassel type calculation using nine active vibrations, an average excitation energy of 5 eV and bond energies given in Chapter I shows that the fission of the F C . . .CF bond is three times more probable than the H C... ^ 2 CH^ bond. The combination of these two factors, the tunneling of the energy to the fluorine containing side and relatively larger rate constant for the breakage of the F C 2 CF^ bond, would help cleave this preferentially. Hence, the biradical (I) would be more abundant than either (II) or (III). These biradicals could undergo electronic rearrangement, with concurrent atom migration, leading to a linear butene (or butene precursor) which could either be stabilized or eliminate a CH F (n = 0—3) . Alternatively, 3-n the biradical could fragment to give two ethylenic species. The amount of ethylenic products formed depends on which of the three biradicals would undergo the scission of the middle C-C bond to a greater extent. The four factors which influence the scission of the

PAGE 134

121 middle C-C bond in the biradical are the centrifugal forces, the bond energy of the middle C-C bond, the number of energy states available in that bond vibration and the energy content of the biradicals. The centrifugal forces pulling the biradical apart would be maximum in biradical (I) leading to the cleavage of the middle -H C-CH • bond and minimum in 2 2 biradical (II) leading to the cleavage of the middle *F C-CF • 2 2 bond. Because the bond strength of the middle bond in (I) is larger than in either (II) or (III) the centrifugal force enhances and bond strength inhibits the cleavage of the middle C-C bond. Furthermore, the middle bond in (I) will have a larger number of available states because the CH -CH bond 2 2 vibration frequency is higher than that of CF -CF , although 2 2 it is quite possible that these C-C bond vibrations are not very active. Lastly, since the bond strength of F C CF 2 2 bond is smaller than others, the biradical (I) would have more available (surplus) energy, on the average, than either (II) or (III). We could expect the inhibition due to bond energy would be at least partially cancelled by the excess energy of that biradical. Hence, the centrifugal effect would enhance the rupture of the middle bond in (I) more than in (II) or (III), and the formation of CH = CF would 2 2 be larger than C H + C F . 2 4 2 4 We have previously presented arguments that the reactions C H F ^ C h'*' + C F + e“ 444 ^^24 24 S C f'^ + C H + e" / 2 4 2 4

PAGE 135

follow a concerted mechanism, while the reaction C H F ^ V CHF'^ + CHF + e 444 ^ 222 222 goes through a sequential route as indicated.

PAGE 136

REFERENCES AND NOTES 1. R.D. Doepker and P. Ausloos, "Photolysis of Cyclobutane at Photon Energies below and above Ionization Energies," J. Chem . Phys . , £3, 3814-19 (1965). 2. A.S. Gordon, S.R. Smith, and C.M. Drew, "Photolysis of a Mixture of Cyclobutane and Acetone-de. Reactions of Cyclobutyl Radical," J. Chem . Phys . , 36 , 824-29 (1962). 3. D.L. Kantro and H.E. Gunning, "The Reaction of Cyclobutane with Hg6(3pj Atoms," J. Chem. Phys., 21, 179799 (1953) . ^ “ — ^ o 4. R.L. Cate and T.C. Hinkson, "2537 A Mercury-Sensitized Photochemical Decomposition of Perf luorocyclobutane , " J. Phys . Chem . , 7^, 2071-72 (1974). 5. C.T. Genaux and W.D. Walters, "Thermal Decomposition of Cyclobutane," J. Chem . Soc . , 73 , 4497-98 (1951). 6. J.N. Butler, "The Thermal Decomposition of Octaf luorocyclobutane, " J. Chem . Soc . , 84, 1393-98 (1962). 7. A. Lifshitz, H. Carroll, and S. Bauer, "Studies with a Single-Pulse Shock Tube II. The Thermal Decomposition of Perfluorocyclobutane, " J. Chem. Phys., 39, 1661-65 (1963). ~ — 8 . R.D. Doepker and P. Ausloos, "Gas-Phase Radiolysis of Cyclobutane," J. Chem . Phys . , 44, 1641-47 (1966). 9. E. Heckel and R.J. Hanrahan, "The y-Radiolysis of Cyclobutane in the Gas-Phase," Int. J. Radiat. Phys. Chem . , 5, 271-79 (1973). 10. E. Heckel and R.J. Hanrahan, "The X-Radiolysis of Perfluorocyclobutane and Mixtures of Perfluorocyclobutane and Methane," Adv. Chem. Series, #82, "Radiation Chemistry-II, " 120-33 (1968). 11. E. Heckel and R.J. Hanrahan, "Radiolytic Processes in Mixtures of Cyclobutane and Perfluorocyclobutane," Int. J. Radiat . Phys . Chem . , 5, 281-91 (1973). 123

PAGE 137

124 12. B.M. Hughes and T.O. Tiernan, "Ionic Reactions in Gaseous Cyclobutane," J. Chem. Phys., 51, 4373-84 (1969) . — 13. R.M. O'Malley, K.R. Jennings, M.T. Bowers, and V.G. Anicich, "ICR Mass Spectra of Fluoroalkenes II. IonMolecule Reactions in the 1 , 1-Dif luoroethylene System," Int. J. Mass Spectrom . Ion Phys . , 11 , 89-98 (1973). 14. A.J. Ferrer-Correia and K.R. Jennings, "The ICR Mass Spectra of Fluoroalkenes IV. Ion— Molecule Reactions in Mixtures of Ethylene and Fluoroethylene , " Int. J. Mass Spectrom . Ion Phys . , 11 , 111-26 (1973). ~ 15. R.E. Fox, W.M. Hickam, D.J. Grove, and T. Kjeldaas, Jr., "Ionization in a Mass Spectrometer by Monoenergetic Electrons," R^. Sci. Instrum . , 1101-07 (1955). 16. (a) C.E. Melton, "Ionization Processes by Monoenergetic Electrons," Ph.D. Dissertation, University of Notre Dame (1964). (b) C.E. Melton and W.H. Hamill, "Appearance Potentials of Positive and Negative Ions by Mass Spectrometry," J. Chem . Phys . , 41 , 546-53 (1964). 17. T.L. Cottrell, "Strength of Chemical Bonds," 2nd Ed., Academic Press, New York, N.Y. (1958) . 18. J.L. Franklin, J.G. Dillard, H.M. Rosenstock, J.T. Herron, K. Draxl, and F.H. Field, Nat. Std. Ref. Data Ser., Nat. Bur. Std., No. 26 (1969). 19. J. Heicklen, "Gas Phase Oxidation of Perhalocarbons , " Adv . Photochem . , 1 _, 57-148 (1969). 20. The fragmentation pattern of C 2 HF 5 is not listed in the API tables. We have examined the mass spectrum of this compound and found that there is no fragment at m/e 100 . 21. J.R. Majer, "Mass Spectrometry of Fluorine Compounds," Adv . Fluorine Chem . , 55-103 (1961). 22. J.L. Franklin, "Prediction of Heat and Free Energies of Organic Ccpmpounds," Ind . Eng . Chem . , 1070-76 (1949); "Calculations of the Heats of Formation of Gaseous Free Radicals and Ions," J. Chem . Phys . , 21 , 2029-33 (1953). Both the experimental value of 210 kcal/mole and the theoretical value of 175 kcal/mole are given by Jennings and coworkers, ref. 13. V.G. Anicich and M.T. Bowers have transmitted to us a preprint of an article currently in press, in which they argue that AH° (C 3H4F+) < 185 23 .

PAGE 138

125 kcal/mole based upon the assumption that the observed ion/molecule reaction C 2 Ht + C 2 H 3 F C 3 H 4 F''' + CH 3 * must be exothermic. Using Ah 5 (C 2 H 3 F) = -32.5 kcal/mole and AHf (CH 3 *) = +33.0 kcal/mole as given by Jennings (ref. 13) , and AH£(C 2 Hi*) = 253 kcal/mole as given in the NSRDS tables (ref. 18), we obtain 187.5 kcal/mole in essential agreement with Bowers. This result clearly depends on the accuracy of the thermodynamic input data, as well as the validity of Bower's assumption. 24. F.W. McLafferty, "Mass Spectrometry of Organic Ions," 145-56, Academic Press, New York, N.Y. (1963) . 25. R.W. Kiser, "Introduction to Mass Spectromy and Its Applications," Prentice-Hall, New York, N.Y. (1965). 26. • S. Tsuda and V7.H. Hamill, "Structure in IonizationEfficiency Curves Near Threshold from Alkanes and Alkyl Halides," J. Chem . Phys . , 41 , 2713-21 (1964). 27. J.H. Futrell, T.O. Tiernan, F.P. Abramson, and C.D. Miller, "Modification of Time-of-Flight Mass Spectrometer for Investigation of Ion-Molecule Reactions at Elevated Pressures," Rev. Sci. Instrum., 39, 340-45 (1968). 28. E. Heckel and R.J. Hanrahan, "Ion-Molecule Reactions in the Systems CF^-CH^ and CF 4 -C 2 H 6 ," J. Chem. Phys., 62, 1027-39 (1975) . ~ 29. C. Chang, "Radiation Chemistry of Hexaf luoroacetone in the Gas Phase," Ph.D. Dissertation, University of Florida (1970) . 30. M.B. Fallgatter and R.J. Hanrahan, "SPC-12 Spectrometer Routines, Real Time Data Acquisition and Data Reduction for Bendix T.O.F. Instruments," University of Florida, 1969, Available from COSMIC Program Library, University of Georgia, as number COS 02260. 31. G.C. Goode, A.J. Ferrer-Correia, and K.R. Jennings, "The Interpretation of Double Resonance Signals in Ion Cyclotron Resonance Mass Spectrometry," Int. J. Mass Spectrom. Ion Phys. , 5, 229-40 (1970). 32. V.L. Tal'roze and E.L. Frankevich, "Measurements of Rate Constants of Ionic Molecular Reactions by the Impulse Method," Izv. Akad. Nauk SSSR, Otd. Khim. Nauk No. 7, 1351 (1959) ; "Pulse Method of Determining the Rate Constants of Ion-Molecule Reactions," Zhur. Fiz. Khim., 34 , 2709 (1960); M. I . Markin, G.N. Karachevitzev, and V.L. Tal'roze, "Investigation of Charge Exchange of

PAGE 139

126 Thermal Ions A , Kr , and Xe"*” by the Impulse Method of Molecules CH 4 , CaHe, and C 2 H 4 ," Izv. Akad. Nauk SSR, Otd. Khim. Nauk No. 8, 1528 (1961) . (The above three papers are from English translations of Russian journals . ) 33. J.L. Franklin, "Positive-Ion-Molecule Reaction Studies in a Single ElectronImpact Source," "Ion Molecule Reactions," Ed. J.L. Franklin, Plenum Press, New York, N.Y. , 9-32 (1972) . 34. M.B. Fallgatter and R.J. Hanrahan, "High Pressure Mass Spectrometry of Simple Hydrocarbon and Alkyl Halide Systems Using Pulsed and Continuous Ionization," Paper No. 145, Am. Chem. Soc. Division of Physician Chemistry, 158th National Meeting, New York, N.Y., September 8-12, 1969. 35. G. Sroka, C. Chang, and G.G. Meisels, "Arrival Time Distribution in High Pressure Mass Spectrometry. I. Residence Time of Ions in Chemical Ionization and the Measurement of Reaction Rate Constants," J. Am. Chem . Soc . , 95 , 1052-56 (1972); "Arrival Time Distribution in High Pressure Mass Spectrometry III. Effect of Ion Source Geometry on Arrival Times and Their Distribution," Int. j. Mass Spectrom. Ion Phys., 11, 367-82 (1973). — 36. J.D. Baldeschweiler , "Ion Cyclotron Resonance Spectroscopy," Science , 159 , 263-73 (1968). 37. V.C. Anicich and M.T. Bowers, "Energy Transfer in Excited Ionic Species, Collisional Stabilization of the Dimer Ions (C4H4FX-)* and (CiaHia-)* in 1 , 1-Dif luoroethylene and Benzene," J. Am. Chem. Soc., 96, 1279-84 (1974). ~ — — 38. George A. Olah, "Carbocations and Electrophilic Reactions," Verlag Chemie, O'ohn Wiley and Sons, 1974. 39. H.E. Gunning and O.P. Strausz, "Isotopic Effects and the Mechanism of Energy Transfer in Mercury Photosensitization," Adv . Photochem . , 1, 209-74 (1963). 40. R.J. Cvetanovic, "Mercury Photosensitized Reactions," Progr. Reaction Kinetics , 39-130 (1964). 41. P.M. Scott and K.R. Jennings, "The Mercury-Photosensitized Decomposition of Ethyl Fluoride, 1 , 1-Dif luoroethane, and 1 , 1 , 1-Trif luoroethane , " J. Phys. Chem., 73, 1513-21 (1969). “ — A.J. Frank, "The Radiation Chemistry and Photochemistry of Ethyl Bromide in the Gas Phase," Ph.D. Dissertation, University of Florida (1975). 42 .

PAGE 140

127 43. E. Heckel and P.F. Marsh, "Titration of Subnanomole Quantities of Fluoride Ions in Polar Nonaqueous Solvents," Analyt . Chem . , 44 , 2347-51 (1972). 44. Orion lonalyzer. Instruction Manual, Fluoride Electrode Model 94-09, Orion Research, Inc . (1967). 45. S.D. Gleditsch and J.V. Michael, "Further Mercury (^P-, ) Quenching Cross Sections," J. Phys. Chem., 79, 409^ 13 (1975). 46. We recorded the ultraviolet absorption spectrum of c-Ci,H 4 F 4 using a McPherson Model 218 Vacuum Ultraviolet monochromator with double-beam attachment and ratiorecording electronics and found that this compound does not absorb above 180.0 nm. 47. J.G. Calvert and J.N. Pitts, Jr., "Photochemistry," John Wiley and Sons, Inc. (1966). 48. P.M. Scott and K.R. Jennings, "The Addition of Hydrogen to Vinyl Fluoride, 1 , 1-Dif luoroethylene , and Trifluoroethylene," J. Phys . Chem . , 75 , 1521-25 (1969). 49. Due to resonance effects fluorine atoms are good electron donors. Since an electron donor stabilizes the radical (See, G. March, "Advanced Organic Chemistry," McGraw-Hill, 1968), fluorine atoms are expected to stabilize radicals. 50. E.W.R. Steacie, "Atomic and Free Radical Reactions," 2nd Ed., ]^, Reinhold Publishing Co., New York, N.Y. (1954) . 51. S.W. Benson and H.E. O'Neal, "Kinetic Data on Gas Phase Unimolecular Reactions," NSRDS-NBS 21, National Bureau of Standards (1970) . 52. C.R. Patrick, "The Thermochemistry of Organic Fluorine Compounds," Adv . Fluorine Chem. , 2, 1-34 (1961). 53. R.J. Hanrahan, "A Co Gamma Irradiator for Chemical Research," Inter. J. Appl. Radiation Isotopes, 13, 254-55 (1962). — 54. G.G. Meisels, "Gas-Phase Dosimetry by Use of Ionization Measurements," J. Chem . Phys . , 41 , 51-56 (1964). 55. R.E. Marcotte and R.J. Hanrahan, "Radiation Chemistry of CF4-CCI4 Mixtures in the Gas Phase," J. Phys. Chem., 76 , 3734-41 (1972). “ 56. G.N. Whyte, "Principles of Radiation Dosimetry," John Wiley and Sons, Inc., New York, N.Y. (1959).

PAGE 141

128 57. "Mass Spectral Data," American Petroleum Institute Research Project 44. 58. The Spark-Discharge technique has been extensively used in our laboratory. Mr. T. Hsieh, Mr. M.D. Scanlon, Dr. A.J. Frank, and Dr. D. Johnson have carried out such discharges in CF3I and C2F5I, C2F6-C2H6, C2HsBr, and C2H4 + O2 systems, respectively; they found that the products of a low-voltage spark-discharge are qualitatively and even semiquantitatively the same as in the radiolysis of the compound. 59. C.E. Kolts, "Energy Deposition Mechanisms," "Fundamental Processes in Radiation Chemistry," Ed. P. Ausloos, Interscience Publishers (1968) . 60. K.L. Hall, R.D. Bolt, and J.G. Carroll, "Radiation Chemistry of Pure Compounds," "Radiation Effects on Organic Materials," Ed. R.O. Bolt and J.G. Carroll, Academic Press, 63-125 (1963) . 61. L.G. Christaphorou, "Atomic and Molecular Radiation Physics," Wiley-Interscience , 1971. 62. A.W. Kirk and E. Tschuikow-Roux , "Vacuum Ultraviolet Photolysis of Fluoroethylenes . I. Vinyl Fluorine at 1470 A," J. Chem. Phys . , 1924-30 (1970). 63. H. Sutcliffe and I. McAlpine, "The Radiation Chemistry of Polyfluorinated Organic Compounds," Fluorine Chem. Rev. , 6, 1-42 (1972). 64. B. Atkinson, "The Mercury Sensitized Reactions of Tetrafluoroethylene, " Nature, 163, 291 (1949). 65. Y.L. Khmelnitski, V.V. Nesterovski, I. I. Melekhonova, E.M. Kononova, and V.M. Nikitina, "Radiolytic Oxidation of Paraffin Wax and other Hydrocarbons. Radiation-Induced Plymerization and Co-Polymerization of Low Molecular Weight Olefins," Proceedings of the 1962 Tihany Symposium in Radiation Chemistry, Akademiai KiaDo, Budapest, 51-65 (1964). 66. G.G. Meisels, "Formation and Reactions of Ions in Ethylene Radiolysis," "Ion-Molecule Reactions in the Gas Phase," Adv . Chem . Series , 58 , 243-63 (1966). 67. In our experiments HF was actually measured as fluoride ions. These fluoride ions are attached to the walls of the vessel as silicon fluorides. 68. S.W. Benson, "Thermochemical Kinetics," John Wiley and Sons (1968) .

PAGE 142

129 69. The AHj (c-C 4 Hi,F 4 ) -202 kcal/mole as shown in Chapter I. 70. H. Carmichael and Y.K. Lau^ "Radiolysis of 1 , 1— Dif luoro— ethane," J. Phys . Chem . , 78 , 2183-86 (1974). 71. R.J. Hanrahan, "Mechanism of the Radiolysis of Alkyl Iodides," Ph.D.Dissertation, University of Wisconsin (1957) . 72. J.R. Durig and W.C. Harris, "Vibrational Spectra and Structure of Small Ring Compounds — XXI. 1, 1,2,2Tetrafluorocyclobutane, " Spectrochim. Acta, 27, 649-62 (1971). —

PAGE 143

BIOGRAPHICAL SKETCH Akkihebbal Ramaiah Ravishankara was born in Shiinoga, India, on November 16, 1949. He graduated from the University of Mysore, India with a Bachelor of Science degree with a composite major in Physics and Chemistry in 1968, and with a Master of Science degree in Chemistry in 1970. After a year of research at the Indian Institute of Science, Bangalore, he entered the graduate program in Chemistry at the University of Florida in September, 1971. During his work towards the degree of Doctor of Philosophy, he has held research and teaching assistantships . 130

PAGE 144

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deg ~ . I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1975 Professor of Chemistry, Chairman Associat^ Profe^or, Nuclear Engineering Sciences Dean, Graduate School