Citation
The Syntheses of derivatives of iminosulfur difluorides

Material Information

Title:
The Syntheses of derivatives of iminosulfur difluorides
Creator:
Johar, Jogindar Singh, 1935-
Publication Date:
Copyright Date:
1966
Language:
English
Physical Description:
vi, 94 l. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Absorption spectra ( jstor )
ATMs ( jstor )
Atoms ( jstor )
Boiling ( jstor )
Fluorides ( jstor )
Fluorine ( jstor )
Infrared absorption ( jstor )
Infrared spectrum ( jstor )
Liquids ( jstor )
Sulfur ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Fluorides ( lcsh )
Fluorocarbons ( lcsh )
Organic compounds -- Synthesis ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis - University of Florida.
Bibliography:
Bibliography: l. 91-93.
Additional Physical Form:
Also available on World Wide Web
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
030085549 ( AlephBibNum )
11021045 ( OCLC )
ACG9881 ( NOTIS )

Downloads

This item has the following downloads:


Full Text















THE SYNTHESES OF DERIVATIVES

OF IMINOSULFUR DIFLUORIDES












By

JOGINDAR SINGH JOHAR


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


August, 1966












ACKNOWLEDGMENTS


The author takes this opportunity to express his

sincere appreciation and deep gratitude to his research

director, Dr. Richard D. Dresdner, for his inspiration and

assistance in carrying out this work and in the writing of

this dissertation. His friendship along with his sincere

desire to help have been a constant source of encouragement

throughout the course of this work.

Appreciation is extended to the other members of the

author's Graduate Supervisory Committee for making sugges-

tions in the writing of this dissertation.

The author is very grateful to Dr. W. S. Brey, Jr.

and coworkers for obtaining and interpreting the nuclear

magnetic resonance spectra for this study. The author also

expresses his gratitude to Dr. R. J. Hanrahan under whose

direction the mass spectra were determined and interpreted.

Acknowledgment is made to the Army Research Office-

Durham for support of a part of this work.

The author wishes to thank the faculty, fellow gradu-

ate students and staff of the Chemistry Department who have

been most helpful during the course of this study. He also

wishes to thank Mrs. Shirley Ingram for typing the final copy

of this dissertation.













TABLE OF CONTENTS

Page
ACKNOWLEDGMENTS . . . ... . . . . ii

LIST OF TABLES . . . . . .. . .iv

LIST OF FIGURES . . . . .. . . ... V

INTRODUCTION . . . . . . . . . . 1

Statement of the Problem . . . . ... 7

EXPERIMENTAL . . . . . . . . . . 10

Apparatus . . . . . . . . . 10

Materials . . . . . . . . . 12

General Procedure . . . . . . .. 16

DISCUSSION . . . . . . . . .59

"Thermal" Stability of Perfluoroalkyl
Iminosulfur Difluoride . * . 59

Fluorination of RfN = SF with High
Valency Metal Fluories . . ... . . 61

The Preparation of RfN = S---CF-C . .. 63
I I
F CF3
The Syntheses of Several Perfluoromines . 65

The Reactions of SF4 and SOF2 with Olefins .71

The Reaction of Rf N = SF2 with CF3CEC CF3 81

SUMMARY . . . . . . . . . 87

BIBLIOGRAPHY . . . . . . . . .91

BIOGRAPHICAL SKETCH . . . . . . . . 94


iii












LIST OF TABLES


Table Page

1. Reaction of CF3N=SF2 with Metal Fluorides . 19

2. Nuclear Magnetic Resonance (N.M.R.) Spectra . 56

3. The Reactants and Products of CsF
Catalyzed Reactions . . . . . ... 89












LIST OF FIGURES


Figure

1.


Page


Infrared absorption spectrum of
CF NS-CF-F . . . .
F CF3


2. Infrared absorption spectrum of
CF CF2N=S-CF-CF ...
F CF

3. Infrared absorption spectrum of
CF -CF2-CFR-N=S-CF-CF .
F CF5

4. Infrared absorption spectrum of
CF -CF-C-S-CF-CF .
S1i 11 / I 1 3
CF3 N F CF3

5. Infrared absorption spectrum of
CF -CF-C -S-CF-CF,. .
51 2 ii 3'
N F CF5

6. Infrared absorption spectrum of
CF N=0-CF2-CF. . . . .
CFN=-CF2-CF.
CF5


7.

8.

9.

10.

11.


. . . .21



. . . 22



. . . .. 23




* . . . 24




. . . o o 31


. . . .


Infrared absorption spectrum of C8F14S2 .

Infrared absorption spectrum of C8F18S202 .

Infrared absorption spectrum of mixture of
C8F14S2+C8F18S202 * * * * * .

Infrared absorption spectrum of C10F22N2S2.

Infrared absorption spectrum of
CF-N=S C-CF-CF . . . . .
F CF5


. 35



S. 37

. 40


. 44

S. 48


S. 49








Figure


Page


12. Infrared absorption spectrum of
CF3-CF2-N=S--CCF-CF .......... 51

F CF3

13. Infrared absorption spectrum of
3CCCFCF2-N=S-C -CF . . . . 54
F CF3












INTRODUCTION


As a result of numerous investigations begun over a

hundred years ago, many (1) sulfur-nitrogen halides are

known. Michaelis (2) described the preparation of the first

iminosulfur dihalide, phenyl iminosulfur dichloride,

C6H5N = S C12 by the reaction of thionylanilide with phos-

phorous pentachloride. This substance is stable only in

solution and cannot be isolated.

The synthesis of a sulfur-nitrogen-fluorine compound

having the empirical formula F NS was reported by Glemser

(3). The compound was obtained as a main product during

the reaction of gaseous mixture of SNF and SN2F2 with A F2

at room temperature. This compound, F NS, was assigned the

structure F-N = SF2 on the basis of chemical evidence and

was the first example of an iminosulfur difluoride reported

in the literature. But later (4), on the basis of infra-

red and nuclear magnetic resonance data the assignment of

the structure of F3NS was changed to NSSF .

A new class of compounds, the organo iminosulfur

difluorides was discovered in the investigations (5) of the

chemistry of sulfur tetrafluoride, SF4. These sulfur

fluoride derivatives have the general formula R-N=SF2 and







2

are prepared by the reaction of SF4 with compounds that
have carbon nitrogen triple bonds. The organo iminosulfur

difluorides have a high thermal stability as indicated by

the temperature required for their syntheses. A feature of

the infrared spectrum of the iminosulfur difluoride struc-

ture -N=SF2 is an absorption band in the 7.15-7.35/1 (1400-

1350 cm-1) region which is associated with the N=S grouping.
Reactions of the S-F bonds (5) of phenyl iminosulfur di-

fluoride with sodium methoxide and with phenyl lithium

using absolute methanol as a solvent resulted in two new

derivatives, C6H5N=S(OCH3)2 and C6H5N=S(C6H5)2, having N=S

linkages.

The present work was started with the intention of

studying the chemistry of perfluoroalkyl iminosulfur di-

fluorides.

Miller (6) found that F- ion shows a strong nucleo-

philic reactivity toward the carbon-carbon double bond of

fluoroolefins. The initial attack is on the carbon of

double bond, which then results in either addition of F-

to form a carbanion,
CF CF=CF2+F--- CF -CF-CF3

or fluoride ion catalyzed rearrangement,









F F F
S'I (
F-C=:C-C- C-- F-C-C= C -C- + F-
T I I I I I I I
F F F F F F

Andreades (7) conclusively established the formation
of perfluoro carbanions by trapping the hepta fluoro-n-propyl
and hepta fluoro-iso-propyl carbanions, generated from the
corresponding monohydro fluorocarbons in ether solution as
lithium salts.

0
II
CH.LiCH CH2-CH
CF CFH C2- CF -CF-CH--CH2CH
3 Et20-700 I
CF3 CF3 OH

Stepwise:
CF -CFH CF -CF + H
3 1 3-1
CF3 OF3

CH Li CH + Li+
3 5


CH CH2CH + CF -CF- -- CH CH2CH-- F-CF3 (Li salt)
I I
Cacid hydrolysis
OH
CH CH2-CH-CF-CF
CF3









The fluoro carbanions produced by the addition of

fluoride ion to fluoroolefins may undergo secondary reactions

depending upon the experimental conditions and the reactants

present. The choice of reaction conditions suitable for

studying the reactions of fluoride ion with fluoroolefins

presents some difficulty. Stable inorganic fluorides tend

to be insoluble in inert non-polar aprotic organic solvents,

while in protogenic polar solvents, hydrogen ion transfer

to fluoro carbanions limits the usefulness of such systems.

Christie (8) observed that certain alkali metal

fluorides act as catalyst in the oxidation of unsaturated

fluorocarbons with oxygen. Tetrafluoro-2-O-fluorosulfate

propionyl fluoride was synthesized (45) by F- ion catalyzed

defluorosulfurylation of hexafluoro 1,2-bis(o-fluorosulfate)-

propane:

OCFCF(OSO2F)CF2(OSO2F) 6K h SO2F2 + CF CF(OSO2F)COF

Although the authors did not report any mechanism, it is

believed that F- ion from KF initiates the reaction:

CF-- CF --- CF --F
3 1
O 0 CF- CF- CF + SO2F2
O=S=0 3 1 11
0=5=0 0 0 + F-
F /I F I
FF O=S=0
F F

Dresdner (9) investigated the relative catalytic activity

of cesium and sodium fluorides and found that whereas the







5

reaction between C3F6 and NF3 proceeded at 3200 when CsF was

catalyst, 5200 was required when NaF was the catalyst.

Van Artsdalen (10) confirmed that CsF is more active than

all other alkali metal fluorides. In a survey of relative

rates of exchange of fluorine atoms between alkali fluorides

and fluorocarbons, conducted at moderate temperatures, the

order of reactivity per square meter of surface was observed

to be

Cs> Rb, K = Na = Li in the exchange process

C3F6 + F18 CF F18 + MF

Miller (11) and Dresdner (12) found that reaction

between cesium fluoride and fluorocarbon dienes yield

fluorocarbon acetylenes at moderate temperature in the ab-

sence of a solvent. Dresdner (12) found that cesium

fluoride in the solid state is a good fluoride ion catalyst

for the reactions requiring the formation of a carbanion

intermediate. The reaction of CF N = CF2 in presence of F-

is a good example of.the difference between the behavior of

solid catalyst and that of F- ion in a solvent. When

CF3N = CF2 vapors were passed over solid CsF (12) in a flow

reaction, (CF3)2N-CF = NCF3 was obtained and its formation

was explained via the ionic mechanism:







4M'
CF 3-N-=CF2 CF3-N~-CF3

F

F
CF--N=C--F CF3-N=CF-N(CF3)2 + F-

CF -N--CF3
When CF N = CF2 was treated (13) with COF2 in a polar sol-

vent containing F ion, there was no report of the dimer
CF3N = CF-N(CF3)2, having been formed, although the reaction
between COF2 and CF3N=CF2 was very slow. Consequently there

must be an important solvent effect.

The degree of catalytic activity of solid CsF is
related to (a) its surface area, (b) defects in the solids,

(c) temperature, and (d) pressure. Perfluoropropene can be

dimerized (12) to the cis and trans isomers of

(CF3)2CF CF = CF CF3 (I) and (CF3)2C = CF CF2 CF3 (II), i.e.,

CF3CF F2 ) CF CF CF3

F- (from CsF)
F

CF3-CF = C F CF3CF=CF-CF(CF3)2 + F
CF -CF--CF3

(I)

Compound (I) can add another F- to form a fluorocarbanion
which after rearrangement loses F- to give (II),viz,









CF-CF=CF-C(CF3)2.--- F CF CF2CF=C(CF )2 + F

F
In a flow reaction involving only FCF=CFF2 and CsF at

3500C, with a contact time of 123 seconds, there was only
4 to 5 per cent conversion to products (I) and (II). But

in a typical static reaction, in a 500 ml vessel containing

75 grams of dry CsF, 74 per cent of the olefin was converted
to the products (I) and (II) in a 40-hour period. This

clearly demonstrated that static reactions give better

conversions.

Statement of the Problem

The main purpose of this research was to prepare new

derivatives of perfluoroalkyl iminosulfur difluorides by

their reaction with perfluoro carbanions generated from

perfluoroolefins and dry powdered CsF. All reactions were
to be carried out in closed vessels. No solvents were to

be used.

It was found that at high temperatures (2500-3000)
and relatively low pressures (20 to 30 atm.) the perfluoro
alkyl iminosulfur difluorides decompose to SF and the

corresponding nitrile. As most of the reactions required
heating to various temperatures (500-3000), there was al-

ways a possibility that some Rf N = SF2 would decompose and

the resulting SF4 might react with the carbanion at these








moderate temperatures. Sulfur tetrafluoride is known (14)

not to attack a carbon-carbon double bond, for example,

H2C=CHCOOH + SF4 ) H2C=CHCF3

HC = CCOOH + SF4 HC = CCF
0
If
F2C C-C-OCH S F2C --C-CF20OCH,
I II + SF -- I I
F2C CH F2C CH

CC -- C0O C1C -- CF
II O + SF --- II 2
CIC--- 0 C CC
CIC CF2

Thus one of the secondary objectives of this work was to

determine if SF4 would react with olefins other than per-

fluoropropene (43), under the same conditions used in re-

actions with perfluoro alkyl iminosulfur difluorides.

The sulfur atom in sulfur tetrafluoride and thionyl
fluoride is formally isoelectronic with the sulfur atom in

the iminosulfur difluorides,

R-N = SF2 F2 SF2 0 = SF2

It was expected that SF4 and SOF2 might behave in an

analogous manner to -N = SF2. Loth and Michaelis (15)

treated C6H5OCH3 with SOC12 in presence of ZnC12 and obtain-

ed CH 0*C6H-S---C6H OCH When Smith et al. (5) used

phenyl iminosulfur difluorides in place of thionyl fluoride,

they obtained the same product. Similarly, reactions of

SOC12 (16) and C6H5N = SF2 (5) with o-C6H4(NH2)2 produced






9
C6H N S. Thus, this behavior of SOC12 and C6H5N=SF2

suggested that reactions of SOF2 and carbanions be studied.
In part, the studies with SOFZ were a matter of practical
expediency since the grade of SF4 purchased for this work
invariably contained from 10 to 40 mole per cent SOF2 which
could not always be completely separated from the SF4.
Thus it became necessary to evaluate the product from
reactions involving SOF2 so that the products involving

SF4 could be uniquely identified. This work is composed
of the experiments which involve the following reagents:
1. Effect of heat and CsF on RfNaSF2
2. CF3N=SF2 + metal fluorides

3. CF3N=SF2 + CF CF=CF2
4. CF3N=SF2 + CF3CF=CFCF3

5. CF3N=SF2 + CF3CCCCF3
6. C2F5N=SF2 + CF3CF=CF2

7. C2F5N=SF2 + CF3CF=CFCF3
8. C2F5N=SF2 + CF3C5CCF3
9. C3F7N=SF2 + CF CF=CF2
10. C3F N=SF2 + CF3CCCF3
11. SF4 + CF3CF=CFCF3
12. SF4 + CF3CECCF3
13. SOF2 + CF3CF=CFCF3
14. SOF2 + CF3CECCF3











EXPERIMENTAL


All temperatures in this dissertation are in degrees

centigrade. The chemical analyses were performed by the

Schwarzkopf Microanalytical Laboratory, Woodside, New York,

unless otherwise stated.

Apparatus


Distillation of substances with boiling points below

room temperature was carried on in a 55 cm long column with

an evacuated silvered jacket. This column was packed with

0.16 cm single turn, nickel helices and had about 30

theoretical plates and would have somewhat lower number (17)

for fluorocarbons. The head was cooled with a dry-ice-

acetone mixture and was connected to vacuum system for

gaseous take off. The temperatures were recorded using

an iron-constantan thermocouple.

The vacuum line was made up of Pyrex glass and was

designed for ball point connections. Kel-F No. 90 grease

(Minnesota Mining and Manufacturing Co.) was used for all

joints and connections. The pump could attain a vacuum of

0.1 mm of mercury.

A Perkin-Elmer Model 137B double beam infracord

spectrophotometer was employed for the determination of









infrared spectra. A 5 cm long gas cell with sodium chloride

windows was used for recording the infrared spectra of the

gaseous samples.

The F19 Nuclear Magnetic Resonance (n.m.r.) spectra

of the compounds were obtained using a varian D.P. 60

spectrometer operating at 56.4 Mc. Trifluoroacetic acid

was used as the external reference. All chemical shifts

are reported in ppm.units.

The mass spectra of the new compounds were obtained

with a Bendix Model 14-107 Time-Of-Flight spectrometer.

The mass spectra were run for the sole purpose of finding

the molecular weights.

Molecular weights of the compounds, having a vapor

pressure more than 10 mm at room temperature, were determined

by the vapor density method. A weighed bulb (volume 209.4-34

ml) was filled with the vapors at known pressure and

temperature and then weighed.

Vapor Phase Chromatography (V.P.C.) of gases was

accomplished with a 9 foot column packed with a mixture of

silicone oil on chromosorb P. operating at room tempera-

ture using helium at 60 ml/min as a carrier gas. The

apparatus was fitted with an adapter for gaseous in-take.

The preparatory scale separation of gases was accomplished

with an 8 foot glass column, 1/2 inch diameter and packed

with a mixture of silicone oil on chromosorb P. The









preparative scale separation of liquids was accomplished in

a vapor phase chromatograph apparatus designed by Professor

Paul Tarrant and coworkers. A two meter glass column with

an inside diameter of one inch, packed with silicone gum

W-95 (Union Carbide) on chromosorb P kept at a constant
temperature 200 below the boiling point of the fraction to

be collected, was used. Nitrogen was the carrier gas.

The refractive indices of liquids were measured with

an Abbe's refractometer at the given temperatures.

Materials

Sulfur tetrafluoride (SF4) was purchased from E. I.

Dupont De Nemours and Co. It was 85-90 per cent SF4, the

impurities being SOF2 and S2C12. It was difficult to ob-

tain pure SF4 from this mixture. In most cases, SF4 was

used as purchased unless otherwise stated.

Thionyl fluoride (SOF2) was obtained by distillation

of this technical sulfur tetrafluoride.

Cesium fluoride (CsF) was purchased from the American

Potash and Chemical Corporation, Los Angeles. It is a very

hygroscopic material which melts at 7030 (18) when pure.

Perfluoropropene (CF CF=CF2) was purchased from

Peninsular ChemResearch, Inc., Gainesville. It has a

boiling point of -290 (19) and its infrared spectrum re-

sembled that reported (20).






13
Perfluorobutene-2 (CF3CF=CFCF3) was purchased from

the Matheson Company, Murrow, Georgia. It was a mixture of

cis and trans isomers in the ratio 1:3 and had a boiling

point of +10. Its infrared spectrum resembled that re-

ported (21).

Preparation of hexafluorobutyne-2.--Hexafluorobutyne-2

(CF3CCCF3) was prepared (22, 26) by the dehalogenation

of 2, 3 dichlorohexafluorobutene-2. Powdered zinc (130 g.,

2.00 mole) treated with dilute (0.1N) hydrochloric acid

was washed with water and dried with 2-propanol. The ac-

tivated zinc was placed in a 2-liter, 3-necked flask and

was covered with 500 ml of dry isopropanol. The olefin

CF3CC1=CC1CF3 (233 g., 1.00 mole) was added dropwise and

refluxed with continuous stirring. The resulting product

(105 g.) was collected in a trap cooled in a dry ice-acetone
mixture. The crude product on distillation gave (88 g.,
0.54 mole) of CF3C=CCF3; b. p. -240 (19, 26). Sixty-two

grams (0.25 mole) of the starting olefin was recovered.
The yield of the hexafluorobutyne-2 was 72 per cent based

on unrecovered dichloro hexafluorobutene-2. The infrared

spectrum of C4F6 resembled .that described by Haszeldine (23).

Preparation of perfluoroalkyl iminosulfur difluo-
rides, RfN=SF2.--The perfluoroalkyl iminosulfur difluorides

were prepared by the method described by Smith et al. (5).

A nitrile or thiocyanate and excess SF4 were









loaded in a 300 ml stainless steel autoclave. The autoclave

was heated to the recommended temperature. After completion

of the reaction, the volatile contents of each reaction were

distilled to recover the products and unused SF4. The imino-

sulfur difluorides were dried over P4010, degassed and stored.

Trifluoromethyl iminosulfur difluoride, CF N=SF2.--

Forty-two grams (0.51 mole) of NaSCN was loaded into the

autoclave, dried and degassed at 1100 for six hours. Then

SF4 (188 g., 1.75 mole) was condensed into the vessel.

The vessel and contents were heated at 2000 for two hours,

at 2500 for four hours, at 3000 for four hours and at 3500

for two hours, a total of 12 hours. A maximum pressure of
200 atm. was noted. When the vessel was slowly cooled to

room temperature no large drop in pressure was observed.

Some elemental sulfur had condensed in the pressure gauge

inlet line.
The volatile contents of the vessel were transferred

to a vacuum system. They amounted to 169 grams and were

fractionated. The unused SF4 was stripped off, a major
fraction was recovered between -90 and -60. It amounted to

44 grams and based on the starting amount of NaSCN, this

represented a yield of 57 per cent. The infrared spectrum

shows these absorptions in cm-1, 1500 (w), 1385 (v.s.),

doublet at 1205 and 1175 (v.v.s.), 817 (m), 760 (s).









F19 N. M. R. data:

Peak

Shift

Area

Group


a

-126.5

2

N=SF2


b

-28.46

3
CF -N


Preparation of pentafluoroethyl iminosulfur di-

fluoride, C2F5N=SF2.--A 300 ml steel autoclave was loaded

at liquid nitrogen temperature with (76 g.,.0.80 mole) of

CF3CN (b. p. -600) and (108 g., 1.00 mole) of SF4. The

autoclave was heated for 24 hours at 3500. The volatile

products were distilled and (106 g., 0.52 mole) of C2F5N=SF2

was obtained at 230. This represented a yield of 65 per

cent based on the CF3CN.

The following infrared absorptions (in cm-1) were

observed for C2F5N=SF2.

1410 (v. s.), 1350 (v. v. s.), 1240 (v. v. s.), 1165 (v. v. s.),

1050 (v. v. s.), 785 (s.), 758 (v. s.), 715 (v. s.).

Preparation of heptafluoropropyl iminosulfur di-

fluoride, C3F N=SF2.--This material was prepared by reacting

C2F5CN (116 g., 0.80 mole) and SF4 (108 g., 1.00 mole) in an

autoclave. The vessel was heated at 2000 for two hours and

at 2500 for four hours. The cooled vessel was attached to

the vacuum system and the more volatile components, mainly

unreacted starting materials, were stripped off. The de-

sired product boiled at 510. Its infrared spectrum showed









the following absorptions:

1390 (v.s.), 1346 (v.s.), 1287 (s), 1250 (v.v.s.), 1202 (s),
1144 (s), 1094 (s), 979 (s), 940 (s), 775 (s), 750 (s) and

712 (s).

General Procedure

All the catalytic reactions were carried out in a

reactor which was a round-bottomed stainless steel (Hoke)

cylinder of 500 ml capacity, 13 inches tall, and 2.5 inches

external diameter. Powdered cesium fluoride (100 g.) was

placed in the cylinder which was then fitted with a 15 inch

long portion of a 1/4 inch pipe, a pressure gauge and an

outlet valve. The reactor was heated to 3500-4000 for two

days under vacuum to remove any residual moisture from the

CsF. It was absolutely essential to remove the moisture

completely otherwise traces of blue products (9) are formed

in the reactions.

The upright furnace for heating the reactor was an

18 inch tall, 3-inch diameter iron pipe which accommodated

the reactor snugly. The iron pipe was heated and the

temperature was controlled by varying the voltage of the

heater. The temperature inside the furnace was measured

using an iron-constantan thermocouple.

All volatile materials were stored in weighed glass

ampoules and kept cold in a dry ice-acetone mixture. These









ampoules could be connected to a vacuum system and an

appropriate amount could be removed by condensation into a

weighed graduated ampoule. The gases were then expanded

into a large bulb and an infrared spectrum and a vapor phase

chromatogram of a representative sample recorded. The

gases were then transferred into the reactor containing CsF

at -196.

The reactor was heated by placing it in the furnace.

The temperature inside the furnace and reactor pressure were

recorded periodically and the results plotted. Normally if

there is no condensation or decomposition, as the temperature

is raised, the pressure-temperature relation is linear. Any

sharp break in the curve is indicative of a reaction. A rise

in pressure indicates decomposition whereas the fall in

pressure indicates the formation of less volatile products.

A reaction was presumed to be complete when the rate of

change of pressure became negligible.

After the reaction, the reactor was connected to the

vacuum system through a weighed trap which was cooled in

liquid nitrogen. The products were vented and any gas which

did not condense was separated for identification. The

products were allowed to warm. The volatile compounds were

expanded into a large flask and the liquids were left in the

trap. The liquid and gaseous fractions were analyzed

separately.









Attempted reaction of CF3N=SF2 with CsF.--Trifluoro-

methyl iminosulfur difluoride (40 g., 0.26 mole) was loaded

in the 500 ml reactor containing dry cesium fluoride. The

temperature was raised to 3000 and maintained at 3000 for

another three days. The recovered material (34.2 g.) was

expanded into a 22 1. bulb. An infrared spectrum and V.P.C.

analysis of a representative sample did not show any new

product. There was an apparent loss of 5.8 grams of

starting material. When the reactor was heated above 2000

under vacuum, 2 to 3 grams of very reactive volatile material

was collected which was not identified.

Reaction of CF N=SF2 with metal fluorides.--The

fluorination of CF3N=SF2 was attempted using HgF2, Ag.F2,

and CoF3 at temperatures between 1000 to 2000 in an auto-

clave. In general, the metal fluoride was loaded into a

300 ml autoclave (except in a final reaction) and after
evacuation, the CF3N=SF2 was added. Table 1 summarizes

the conditions and results of several reactions.

Reaction of CF N=SF2 and CF CF=CF2 over CsF.--Tri-

fluoromethyl iminosulfur difluoride (41 g, 0.27 mole) and

perfluoropropene (96.0 g., 0.64 mole) were loaded into the

reactor. At room temperature, the pressure in the vessel

was 6.2 atm. The system was heated to a reaction tempera-

ture of 830 over a five hour period and maintained at this









Table 1

Reaction of CF3N=SF2 with Metal Fluorides


Metal Maximum
Fluo- CFN=SF2 Pressure Time
ride Moles Ioles atm. Temp. Hours Products

HF2 0.35 0.11 14 100 22 No reaction.
0
HgF2 0.35 0.11 a6 170 48 No reaction.
0 2
AgF2 0.75 0..1 13 100 24 Slight reaction.
Q 2
AgF2 0.75 0.11 16 150 48 SF6, CF4, NF3,

CF3CN.

CoF3 1.09 0.14 15 100 16 SF6, CF4.

Co3 1.09 0.37 40 200 10 SF6, CF4,

(CF3)NN(CF3)

traces.

Cc3 1.32 0.26 131 200 24 N2 (undetermined),

CF4 (0.06 mole),

SF6 (0.1 mole),

SOF2 (0.02 mole),

CF3SF3 (0.03

mole),

(CF3)2 NF (0.01 mole), (CF3)2 N-N (CF3)2

(0.02 mole).






20
temperature for 115 hours. During this interval, the pres-

sure dropped from 17 atm. to 13 atm. and heating was dis-

continued. The final pressure at room temperature was

4.1 atm.

The volatile contents of the reactor were trans-

ferred to a vacuum system. This material amounted to 136

grams, of which 53 grams (0.35 mole) was CF3CF=CF2. From

the fraction boiling above 250, which amounted to 70 grams,

only two products were isolated.

The major product amounted to 43.5 grams and had

the following physical properties.

Boiling point 65.50.

The infrared absorption spectrum (Fig. 1) showed

a strong peak at 1298 cm-1 which is associated with -N=S

stretching.

F19 N. M. R. data:

Peak a b c d

Shift -66.6 -25.1 -6.3 +92.0

Area 1 3 6 1

Group SF CF3 2xCF3 CF


Analysis: Calculated for: C4F11NS, F, 69.2; C,

15.7; N, 4.72; S, 10.1. Found: F, 69.0; C, 15.8; N, 4.63;

S, 99.








0
0 0 0 0 0
- o 0 CN O


Or-



O--tO

o


O4
co .

0





O -I
0 0
CO k
00
<-



0
o 4

0
0
P4


0





H


r-1


O
0-





















0 o





0--




o ) 0
o -0 O

I Co
l -"--- +o





_rd







ou
1 - c-t

-o -H
ZC)





0







O-
O-
o 0
r- ^r:..::^-^- o o










.-_ -_ __ .^ :=::=r~:= : >
-^ --- <





_ _ _ _ _ ^
0 :
o r ho
i 0 L_ _ _ _
C:

















-0
I




Cu
I
-\


Px4
n




.ri

"-




co
O
F-r


a4
0



*r4


o
02
,0
CO

d




H




i-


O
0
o0




0
0
0.



0

0















0
0-









CN
O

O















O
-







0O
O
0-
0U


oz

O0








z
Z





CO 0
-j



OLI
I



--r










0
0 0 0 0 0


r4-











0


0
Fr




















.4-F
0
12
















rd
0












4
,0












4--4
I
^i

a-








Molecular weight: Calculated for C4F11NS, 303.

Found (vapor density method), 306, 305.

This product was assigned the structure

CF N = S-CF-CF
3 3
F CF3

The yield of C4F11NS was 52 per cent, based on starting

difluoride.

It was observed that the new material became slightly

discolored on standing in glass over a period of two weeks.

Upon gas transfer, it became water white and had the same

infrared spectrum as in Figure 1.

The minor product (3.0 g.) had the following phy-

sical properties:

Boiling point 75.50.

The infrared absorption spectrum (Fig. 4) showed

a strong peak at 1780 cm-1 which was associated with ring

C = N stretching mode.

F19 N. M. R. data:

Peak a b c d e

Shift -60.87 -0.46 +3.85 72.98 102.46

Area 1 6 6 1 1

Group SF 2xCF3 2xCF3 CF CF




26
Analysis: Calculated for C F15NS, C, 20.24; N,

3.37. Found: C, 20.39; N, 3.34.
Molecular weight: Calculated for CF 15S 415.

Found (Mass spectrum), 415.

This product was assigned the structure

(CF3)2CF-C SF-CF (CF3)2.

N

The reaction of C2F5N=SF2 and CF3CF=CF2 over CsF.--

Thirty-six grams (0.18 mole) of C2F5N=SF2 and 51 grams

(0.34 mole) of CF CF=CF2 were condensed in the reactor.
The vessel was heated in an upright furnace to a maximum

temperature of 890. The heating was discontinued when pres-

sure remained constant at 890. The volatile contents were

transferred to a vacuum system and the distillation gave

the following mass distribution:

CF CF=CF2 30.3 g.

A pure product 52.5 g.

Residue 2.2 g.

Apparent loss 2.0 g.
The product (52.5 g.) has the following physical

properties:

Boiling point 82.30.

Analysis: Calculated for C15F13NS, F, 70.0; C, 17.1;

N, 4.0; S, 9.1. Found: F, 69.6; C, 17.0; N, 3.8; S, 9.2.

Molecular weight: Calculated for C5F13NS, 353.

Found (vapor density method), 352, 354.









Infrared absorption spectrum (Fig. 2) showed a

strong peak at 1279 cm-1 which was associated with N = S

stretching mode.

F19 N. M. R. data:

Peak a b c d e

Shift -52.3 -8.4 +14.5 17.4 90

Area 1 6 3 2 1

Group SF 2xCF3 CF CF2 CF


The product was assigned the structure

CF CF2N=S -CF --CF3
3 3
F CF3


The yield of the pure product was about 83 per

cent based on the fact that no C2F5N=SF2 was recovered.

The infrared spectrum of the residue (2.2 g.) showed

an abosrption band at 1715 cm-1, suggesting a compound

with C = N or C = C bond. The yield of this product was

very small (approx. 0.5 g.) so further attempts at an iden-

tification were abandoned.

The reaction of C3F N=SF2 with CsF.--Twenty-eight

grams (0.08 mole) of C3F7N=SF2 was condensed into a 500 ml

stainless steel vessel containing 100 g. of dried CsF. The

vessel and its contents were heated for a 2-day period up

to a temperature of 2000 with a linear increase in pressure.









At 2500 a sharp increase in the pressure was observed and

heating was discontinued. The volatile contents of the

vessel were removed and were found to consist mainly of

C2F5CN, SF4 and C3F7N=SF2. The amounts of nitrile and sul-

fur tetrafluoride recovered were equivalent.

The reaction of C F N=SF2 with CF CF=CF2 over CsF.--

The reactor, containing 100 g.of dry CsF was loaded with 27

grams (0.11 mole) of C3FSN=SF2 and 59 grams (0.40 mole) of

CCF-CF=C2. The pressure in the vessel at 270 was 6.1 atm.

The temperature was raised to 820 and a maximum pressure of

13.9 atm. was observed. The temperature was maintained be-
tween 750 to 850 for another 88 hours, when the observed

pressure was 10.7 atm. at 77.50. On cooling the reaction

vessel, the pressure was 5.5 atm. at 270. The products were

vented and distilled. The following fractions were obtained:

CF2CF2CF3 + CFCF=CF2 45 g.

new products 38.7 g.

No C F7N=SF2 was recovered. The products were separated by

V.P.C. and the two new compounds were isolated and identified.

The major product amounted to 27 grams and had the following

physical properties:.

Boiling point 103.50.

Analysis: Calculated for C6F15NS, F, 70.81; C, 17.86;

N, 3.47; S, 7.94. Found: F, 70.40; C, 17.90; N, 3.40;

S, 8.31.








Molecular weight: Calculated for C6F15NS, 403.

Found (by vapor density method) 400, 401.

Infrared absorption spectrum (Fig. 3) showed a

strong peak at 1272 cm-1 which was associated with the

N=S stretching mode.

F19 N. M. R. data:

Peak a b c d e f

Shift -67.6 -6.6 +7.7 8.7 52.7 91.2

Area 1 6 3 2 2 1

Group SF (CF3)2 CF3 CF2-N CF2-C CF


Based on the above data, the major product of the

reaction between C F N=SF2 and CF3CF=CF2 over CsF was as-

signed the structure

CF3CF2CF2N=S-- CF- CF3
i 1
F CF3


The yield of the product was 65.1 per cent based on

CF 7N=SF2 used. The minor product isolated amounted to

6.5 grams and had the following properties:

Boiling point 1000

Analysis: Calculated for C6F13NS: F, 67.7; C,

19.8; N, 3.83; S, 8.78. Found: F, 67.1; C, 19.9; N, 3.61;

S, 9.11.






30
Molecular weight: Calculated for C6F13NS, 365.
Found (by vapor density method) 368, 370.

Infrared absorption spectrum (Fig. 5) showed a
strong absorption at 1707 cm-1 which was associated with

ring C = N stretching.

F19 N. M. R. data:

Peak a b c d e

Shift -45.1 -4.4 +7.7 42.7 87

Area 1 6 3 2 1

Group SF (CF3)2 CF3 CF2 CF

Based on the above data, the minor product of the
reaction between C3F7N=SF2 and CF3CF=CF2 over CsF was as-

signed the structure


CF -CF2-C -S- CF CF
11/1 I 3
N F CF3

The yield of C6F13NS was 16 per cent, based on C3F7N=SF2.

The attempted reaction of CF3CF=CFCF3 with CsF.--
When Young (24) passed a mixture of cis and trans perfluoro-

butene-2 over CsF heated to 3500 in a flow system he did

not report any dimerization or isomerization.

A mixture of about 30 per cent cis and 70 per cent
trans perfluorobutene-2 (80.0 g., 0.40 mole) was loaded into

the reaction vessel containing dry CsF. The temperature was




















0
u
NM
0-0
I
r4









1\
I
Fr

0











,o

C-4






H
I
4
rx







H

0

.0
r
Fr









raised to 3350 in a period of two days and no irregular pres-

sure changes were observed. The temperature was maintained

at 3350 for another two days. After expanding the cooled

products into a 22 liter bulb, the infrared spectrum and

V.P.C. analyses of representative samples indicated that no

chemical reaction had occurred. But V.P.C. analysis did

indicate a slight increase in the trans isomer.

Attempted reaction of CF3N=SF2 and CF3CF=CFCF3 (with-

out CsF).--Perfluorobutene-2 (11.0 g, 0.52 mole) and tri-

fluoromethyl iminosulfur difluoride (4.0 g., 0.026 mole) were

condensed in a 100 ml stainless steel autoclave and heated

to 3500 for 24 hours. Almost all, about 98 per cent by
weight, of the starting material was recovered unchanged,

with a small amount of black crystalline material. Thus,

there was no appreciable reaction.

The reaction of CF3N=SF2 with CF CF=CFCF over CsF.--

Trifluoromethyl iminosulfur difluoride (26.0 g., 0.17 mole)

and perfluorobutene-2 (34.0 g., 0.17 mole) were condensed in

the reaction vessel containing dry CsF. The temperature was

raised slowly in intervals to allow for thermal equilibrium.

No pressure drop was noticed at 2250 for 12 hours and the

temperature was raised. A small pressure drop was observed

at 2360. The temperature was raised to between 2800 to 2900

to speed up the reaction. After heating for 96 hours, the










pressure remained almost constant and heating was discon-

tinued. As only 50 grams of volatile products were collected,

the reactor was heated to 4000 under vacuum. A mixture con-

taining 5.0 g. of a liquid and solid sulfur was collected in

a liquid nitrogen cooled trap. The liquid was very volatile

and very reactive toward glass. No attempt was made to

identify it.

The volatile products (50 g.) were expanded in 22 liter

bulb and V.P.C. analysis and infrared spectra indicated the

following fractions:

CF N = SF2 6.9 g.

CF CF = CFCF3 17.0 g.

new product 25.0 g.

two unidentified pro-
ducts 3.0 g,
The new product had the following physical properties:

Boiling point the vapor pressure at different tempera-

tures was determined as

240 459 mm -110 87 mm

110 267 mm -780 0 mm
00 164 mm

1
From the plot of log P vs 1 the boiling point of the
product was found to be 36.60 at 760 mm pressure.

Analysis: Calculated for CsF11N: F, 73.85; C,
21.20, N, 4.95. Found: F, 74.00; C, 20.84; N, 5.00.







Molecular weight: Calculated for C5F11N 283.
Found (vapor density method), 277.5, 278.8. (Mass spec-
trum), 283.

Infrared spectrum (Fig. 6) showed a strong peak
at 1725 cm-1 which was associated with C = N stretching.
F19 N. M. R. data:
Peak a b c d
Shift -16.88 -10.78 +5.85 40.19

Area 3 3 3 2

Group CF3-N CF3-C CF3-C C-CF2-C

The above data was consistent with the following

proposed structure:

CF3-N=C CF2-CF3

CF3

The yield of C5F11N was 69 per cent with a conver-

sion of 73.6 per cent of CF3N=SF2 to products.

The reaction of C2F5N=SF2 and CF3CF=CFCF3 over CsF.--
Pentafluoroethyl iminosulfur difluoride (35.0 g., 0.17 mole)
and perfluorobutene-2 (61.0 g., 0.30 mole) were condensed in
a 500 ml reaction vessel containing dry CsF. The reaction
vessel was heated to 2820 over a 25 hour period and a maxi-
mum pressure of 24.2 atm. was observed. On raising the









0
0 0 0 0 0
- co '0 CN 0


r4
0
I




0



o



-*H






.H
0










-p


0



*d




H


I4





36
temperature further to nearly 3000, a rapid rise in pres-
sure was noticed. The temperature was lowered to 2930
'and a maximum pressure of 29.1 atm. was observed. The

pressure started dropping and within the next 72 hours

it dropped to 24.5 atm. at 2900. The products (85 g.) were

removed. The gases and the liquids were analyzed and the

following fractions identified:

CF3CF = CF CF3 28 g.

CF3CN 13 g.

CF3CF2 CF2 CF3 10 g.

Liquid product 32.2 g.

Unidentified 2.8 g.

Apparent loss 10.0 g.
No C2F5N = SF2 was recovered. The following physical data
on the product was obtained.

Boiling point.- 135.

Refractive index at 200 1.3554.

Analysis: Calculated for C4F7S1: F, 62.44; C,
22.54; S, 15.02. Found: F, 62.14; C, 22.91; S, 14.97.

Molecular weight: Calculated for (C4F7S1)2 426.
Found (by mass spectrum), 426. Molecular formula C8F14S2.

Infrared absorption spectrum (Fig. 7) did not show
any peak in the C = C or C = N region. F19, N. M. R. data:

Peak a b c d

Shift -19.9 -4.7 +2.0 32.0

Area 6 3 3 2

Group (CF3)2 CF3 CF CF2








0
0


-: ~::.' :;: :, I:-- -, ilr ~l : .' t ;! Citi
*: I -. . . ,

... - i, .'- j t .j _, ,, l ' t I
*t rt ,i-i [ ,- --- 1 t '" I I-I-
... ---- --------'r-; -: ---, ~7, { -. '

i i.'-- .. ' '- .- . .' -, - -'!"i ii. .

. I . ... 2. -kL, .;4 1 ,ti



-i ., fT ki 2 I "-I j '1''' 'i I :j 4 ii
r. ", I ',-i .! i r



t ] -. i T 1. I ,I:. 1 ..

. ,I. I L-


hlis

.1
* r I
1,1'-11


"-:' '7 L -+ [ I 1 . 1!1
I JL:4



: ,

"i' r--. '- i ~:i,, ;' :


~ I Ili ll


tI

t


I CI

'I


-II ------I





.I 1 .. I , .; . I


S,

NT


oz
CO



0-


z
-j
LU


C





38
The product of the reaction between C2 F5N = SF2

and CF CF = CF CF3 over CsF at 2930 was assigned the struc-

ture

CF C = C- CF

S' S

CF3 -CF2-C -CF3

which is consistent with the physical data.

The reaction of a mixture of SF, and SOF2 with

CF CF = CF CF3 (without CsF).--A mixture (25 g.) consist-

ing of sulfur tetrafluoride and thionyl fluoride whose

composition was indefinite and perfluorobutene-2 (30 g.,

0.15 mole) were condensed in a 500 ml. steel vessel. The

pressure at room temperature was 7 atm. The reaction ves-

sel was heated to a temperature of 2590 and no abnormal rise

or fall in pressure was observed. The P versus T relation

was a straight line. After venting the gases, the infrared

spectrum and V. P. C. analyses of representative samples

indicated that no reaction had occurred.

The reaction of SOF2 with CF3 CF = CF CF3 over

CsF.--Thionyl fluoride (41.0 g., 0.535 mole) and perfluoro-

butene -2 (47.0 g., 0.23 mole) was condensed in the re-

actor containing dry CsF. The temperature was raised to

2020 over a period of 16 hours when a maximum pressure of

28.0 atm. was observed and then started dropping. The

temperature was held between 200 and 2050. The pressure

dropped to 13.2 atm. in a period of 120 hours. The fol-

lowing fractions were recovered:







Volatile products 47 g.

New product (liquid) 25.2 g.

Unidentified liquid 2.8 g.

Apparent loss 18 g.

The volatile gases contained perfluorobutene-2,

thionyl fluoride and one unidentified substance which was

very reactive to glass, Kel-F grease, the NaC1 windows of

infrared cell, mercury and the packing in the V. P. C.

column. On standing in a 22 1tr bulb, the reactive gas

was converted to SOF2 and SiF4. Although the material

was never isolated its chemical behavior suggested that

it was probably SF3OF. The exact composition of the vola-

tile gases could not be determined. No perfluorobutane

was obtained.

The liquid product was 95 per cent pure and was

further purified by V. P. C. The purified product had

the following physical properties:

Boiling point 1490.

Refractive index at 200 1.3318.

Analysis: Calculated for C4 F9 S 0: F, 64.05;

C, 18.00; S, 42.00. Found: F, 64.47; C, 18.29; S, 12.23.

Molecular weight: Calculated for (C4F9 SO)2 534.

Found (by mass spectrum), 534. Molecular formula C8 F18 S2 02.

Infrared absorption spectrum (Fig. 8) showed a strong

doublet at 1351 Cm-1 and 1333 Cm-1 which is associated with

one oxygen bonded to sulfur as in SOF2.

























0
0
O


0


0
0,
O



0
9-

0.
0
















0
0-



O
r-
0
0
0
CN


O
0L7T,


z
0
-O





U
Qi:
0u






LU


o
0

co
1-I



-r
C)

0




C)
0t





0
02
,0



.rj
C-'






F19, N. M. R. data:

Peak a b c d e

Shift -2.52 -1.64 +3.00 40.8 85.8

Area 3 3 6 4 2

Group CF3 CF3 2 CF3 2 CF2 2 CF

The liquid product of the reaction between thionyl fluoride

and perfluorobutene-2 over CsF at 2020 was assigned the

structure

CF 0 CF
I 3 If I 3
CF-S-0- S -- CF
I I
CF2 CF2
I I
CF3 CF3

and was consistent with the above data. The yield of the

product Cg F18 S202 was 40 per cent based on perfluorobutene-2.

The material balance indicated a loss of 18 grams.

To recover it, the reactor was heated to 4000 under vacuum,

when 12 to 13 grams of very reactive and volatile material

condensed in the liquid nitrogen cooled trap. It was dis-

carded.
The reaction of SF4 with CF3CF = CF CF3 over CsF.--

Sulfur tetrafluoride (40D g.) containing some thionyl fluo-

ride and perfluorobutene-2 (57 g., 0.285 mole) was con-

densed in a 500 ml. steel reaction vessel containing dry

CsF. The temperature of the system was raised to 2520

and a pressure of 22.6 atm. was observed. A further rise

in temperature was accompanied by a drop in pressure. The

temperature was held between 2600 to 2700 and the pressure





42
fell to 15.3 atm. in a period of 12 hours. The products

were removed from the reactor and the following fractions

were collected.

Volatile gases 64 g.

Liquid product 29 g.

Apparent loss 4 g.

The volatile fraction was expanded into a 221. bulb.

There was evidence that the glass walls and Kel -F grease

were being attacked. When it was passed through the V. P. C.

column only perfluorobutane, perfluorobutene-2 and some

SOF2 appeared. The failure of the reactive components)

to exit from any V. P. C. column (Silicone oil, Kel -F

ester, D. N. P.) made an analysis difficult. To find the

amounts of each component, comparison infrared spectra of

known samples were used and compared to the corresponding

spectra of the mixture of volatile materials. The quantity

of each in the volatile fraction were crudely estimated to

be

Reactive gas + SOF2 18 g.

CF3 CF = CF CF3 28 g.

CF3 CF2 CF2 CF3 18 g.

The V. P. C. chromatogram of the crude liquid product

had a major peak with a shoulder which did not resolve on

any of V. P. C. columns used; silicone oil, silicone gum,

Kel -F ester and D. N. P., even at temperature lower than

200 below the boiling point and at lower helium flow rates.

When the material represented by this peak in the chromato-





43

gram was separated from the mixture by preparative scale

V. P. C., its boiling point was determined to be 1420.

From its infrared spectrum (Fig. 9) it was presumed to be

a mixture of

CF CF CF 0 CF
1 3 3 3 I I 3
C----S I CF-S-O-S-CF
II C, I I
C S I + CF2 CF2

CF2 CF3 CF3

CF3 CF3

(Infrared spectrum Fig. 7.) (Infrared spectrum (Fig. 8.)

To be sure that this product was a mixture of above

two components, the two compounds were mixed in the ratio

of 1:3 and an exactly superimposable infrared spectrum

was obtained. The boiling point of this synthesized mix-

ture was 1430. The retention time and position of the shoul-

der of its V. P. C. chromatogram were exactly the same as

that for the separated product.

Attempted reaction of CF3N = SF2 with CF3C = C CF3

(without CsF).--Trifluoromethyl iminosulfur difluoride (25.0

g., 0.16 mole) and hexafluorobutyne-2 (15.0 g., 0.09 mole)

were condensed in a 500 ml. Hoke cylinder fitted with a

stem, pressure gauge and an outlet. The reaction vessel

was heated to 1000.for 6 hours.- The infrared spectrum

and V. P. C. chromatogram of the reactor contents indicated

that no reaction had occurred.

The reaction of CF3N = SF2 with CF C = C CF3 over

CsF. Trial A.--Trifluoromethyl iminosulfur difluoride





44



0
0 0 0 0 0


o



cO
-:jt Ilit HER
.iir --+H di
-L- I-I I '

14,


ol I - t : I' 1;d,
-ir '' 00

r- r L -* tt, j








-,i,' ... -
v












*1 *tI". 1 i ti I 1 -
0 II:; -Z 77
;Ill







-- -, -4 H
,- -- i: ^ ^ i: Tlid 'X^ r4 -- o
-, I- TI I'hl uj
E Tt- It '4r'

+ 4 14V
0, 1: 1 1 11 4 *





45

(20.0 g., 0.13 mole) and hexafluorobutyne-2 (15.6 g., 0.10

mole) were condensed in a graduated glass ampoule and al-

lowed to melt. A slightly milky mixture but no separate

layers were formed. This mixture of CF N = SF2 and

CF -CEC-CF3 was loaded into the reaction vessel contain-

ing CsF. The pressure in the reactor at 230 was 2.6 atm.

A prior calibration of the furnace had shown that a 10

volt potential on the heater coil resulted in an equilibrium

temperature of 50 50 In this aase the equilibrium tempera-

ture was 720 indicating an exothermic reaction, so heating

was discontinued. Even then the temperature continued in-

creasing and the pressure was observed to fall. After 3

hours at 800 the observed pressure was 3.33 atm. and then

the temperature started to drop. Electrical heating was

resumed and the temperature held at 800. After another

3 hours the pressure was 1 atm. The reactor was bled and

products condensed but the mass of recovered material was

only a fraction of the mass of reactants. The reactor was
heated to 1500 with pumping and a less volatile liquid col-

lected. To remove the last traces of material, the reactor

was heated to 3500 with pumping when the decomposition

products of absorbed species came out. Following fractions

were recovered:

CF3 N = SF2 9 g.

CF3 C=C CF3 3 g.

Liquid product 20.3 g.

Residue 3-4 g.





46
The V. P. C. chromatogram of the liquid product

indicated it to be 90 per cent for one component and two
more components each approximately 5 per cent. The major
component was separated by preparative V. P. C. and had the
following properties:
Boiling Point 1670.
Refractive index at 200 1.3315.
Analysis: Calculated for C5Fll NS: F, 66.35; C,

19.5; N, 4.44; S, 10.15. Found: F, 67.27; C, 19.45; N,
4.58; S, 10.10.
Molecular weight: Calculated for C5 F11 NS 315.
Found (by mass spectrum), 630. Molecular formula C10 F22 N2S2.
Infrared absorption spectrum (Fig. 10) did not show

any peak in the C = C or C = N region. F19, N. M. R. data:
Peak a b c d e f g
Shift -16.6 -32 -2.4 +3.4 39.5 40.6 84.6
Area 3 6 7 4 2
Group CF -N 2xCF3 2xCF3+CF 2xCF2 2xCF
The above data was consistent with the proposed
structure
CF CF CF
I 3 I 3 I 3
CF- S- N S-- -- CF
I II II I
CF2 CF -N CF2
I I
CF CF3
The yield of the product C10 F22 N2S2 from the reaction
of CFPN = SF2 with CF3C a C CF3 over CsF at 800 was 74 per
cent based on hexafluorobutyne-2.







Trial B.--In order to prepare the monomer
CF N = S C = CF CF3, CF3N = SF2 (25 g., 0.16 mole) and

F CF3

hexafluorobutyne-2 (15.0 g., 0.09 mole) were condensed in

the reactor vessel containing CsF and maintained the tempera-

ture at 560 for 4 hours. The products indicated 85 per

cent of C10 F22 N2 S2 and 8 per cent of another compound,

which after separation by V. P. C., showed a strong infra-

red absorption band at 1748 cm-1 (Fig. 11). The total quantity

of a second product that was separated amounted to only 2-3

drops and thus further identification was abandoned. A

third produce (7%) showed a weak infrared absorption, at

1718 cm-1.

Trial C.--In another attempt to prepare the expected

monomer, CF3N = S C = CF CF3, from CF3N = SF2 (16.0 g.,

F CF3

0.11 mole) and CF 3CC CF3 (7.5 g., 0.047 mole), the sub-

stances were condensed into the reactor and allowed to stand

at room temperature for 38 hours. In the interval the pres-

sure in the vessel dropped from 3.27 atm. to 0.8 atm. Nine

grams of a crude liquid product which consisted of 81 mole
percent C10 F22 N2 S2 and 18 mole percent of the third product,

in trial B, were prepared. Apparently no monomer was formed.

The reaction of C2 F5 N = SF2 and CF3 CECCF3 over

CsF.--Perfluorobutyne-2 (17.2 g., 0.11 mole) and pentafluoro-

ethyl iminosulfur difluoride (28.0 g., 0.14 mole) were con-

densed in the reactor containing CsF. The temperature was

raised slowly and at 1240 a pressure of 8.67 atm. was ob-






























Z
0





0
z
LIl
uJ
LUi





49

0
0 0 0 0 C
c o -- 0 0 LO M- !
t --..-, -- I . ^ ,-'

- r f' r2 -_ -' I 1-, '


'ir C,
0 --\---I-- -- .4_ *-_-l" '-
... .-- .. : -t -' : ,.~ I
0r-i,4^^^^ C ._,
o -' -^ : "~- 7 1 i: : "
4i 4
-1 4


"- ; t -T .. . I 1 .) I C 'I


-II 'f L j l -I
3 ",; Tf" '-
4' I T!






0 Li I-, 4i_' I--t- t.._ IC)


t ,, I .d
3- ,- I ,1,41r 14 r

,^ T1 f, >,
- < I . -. ... ..
--tI E - 4

..-..--..--+_ __ .._.-_

- 'O- -, 1 I ,



Sf T 1 "-LO'+ i,,- '
-4, ,
C' 4 Ur






50
served before the pressure started dropping. The tempera-

ture was maintained between 1240 and 1270 for a total of

52 hours and the pressure became constant at 3.2 atm. The
products were removed and the following fractions were re-

covered:

CF3C C CF3 2 g.

C2F5 N = SF2 7 g.

New liquid product 33.7 g.

Unidentified 0.5 g.
Thr product was almost pure and needed no further purifi-

cation. It had the following physical properties:

Boiling point 980.

Refractive index at 200 1.3158.

Analysis: Calculated for C6F13 NS: F, 67.67;

C, 19.73; N, 3.84; S, 8.77. Found (I): F, 63.9; C, 20.59;

N, 5.07; S, 10.48. Found (II): F, 66.9; C, 19.9; N, -;

S, -.

Molecular weight: Calculated for C6 F13 NS 365.

Found (by vapor density method), 367.4, 368.

Infrared absorptions spectrum (Fig. 12) had a strong

peak at 1724 cm1 which was associated with C = C stretch-

ing mode.
19
F9 N. M. R. spectral data:

Peak a b c d e f

Shift -40.06 -3.92 -3.15 +3.59 41.79 87.5

Area 1 3 3 3 2 1

Group SF CF3 CF3 CF3 CF2 CF
3 CF CF2










0 0 0 0 0
- g000N
,--v c a


I --


S't

- I


SI ;J' .lit
b:t7 '~
+ ___


0
0
OC
N.







0
0
OD





0
0
0

0
C-














0
0
o


C 4
'. "IT -.,











',4y.


Vrr


tPr


1:111,


--4 -fi'---'---


S,:
II,'
I


I-


': '' i I 'i j'--

7' Hj-E
L,
L, I ,.
L--!- 4- ,1 4
. . .r I I .. f.- i - -- I -




ZZZZZ t .4 I
I [
tL
I I ,[





... : i_, -;

1i 1



-4j
_ :+ T/ : -, ,-,, 1 .1.' \4



,^ --i, l I I-rl-,-
I 'n ,, I ,
tl, '
ti


I rl!1 1 11 4 = L


-1--I ,
I -. :


L ,-;


I' I

I I $


O
O





I

Z
u-u

-.


-


r


o N'
0-0


I



r\J
(N

0




0




-H
<4-
0







0
-1)
44
O







H
I
bD


r;tt-


I~~C~t~'
1
"; i


.II "1 - ; :, I


I* o i' ,11

-U ,: ..- Lid1,i ... ... --





52
The above data were consistent with the proposed structure

CF3CF2 N = S-C = CF CF3

F CF3
The yield of the product C16 F13 NS from the reaction of

C2 F5 N = SF2 with CF C C CF over CsF at 1240 was 85.5
per cent based on the perfluorobutyne-2.

The reaction of C3 F7 N = SF2 with CF3 C C CF3
over CsF.--Heptafluoropropyl iminosulfur difluorlde (32.0 g.,

0.125 mole) and perfluorobutyne-2 (32.8 g., 0.2 mole) were

condensed in a 500 ml. reaction vessel containing dry CsF.

It was slowly heated and a maximum pressure of 7.8 atm.

was recorded at 1420. The temperature was held between

1500 to 1550 and the pressure continued dropping. After
a total of 72 hours, the pressure dropped to 6.2 atm. at

1490, when the products were taken out. Analysis indicated

the following fractions:

CF3 C = C CF3 15.5 g.

New liquid product 46.0 g.

Unidentified 3.5 g.
No C3 F7 N = SF2 was recovered. The V. P. C. separated

product had the following properties:
Boiling point 1110.
Refractive index at 200 1.3142.

Analysis: Calculated for C7 F15 NS: C, 20.24;

S, 7.71. Found: C, 20.68; S, 8.00.
Molecular weight: Calculated for C7 F15 NS 415.

Found (by vapor density method), 420, 423.





53

Infrared absorption spectrum (Fig. 13) showed a

strong peak at 1715 cm-1 which was associated with C = C

stretching mode.

F19, N. M. R. data:

Peak a b c d e f

Shift -45.16 -3.14 +3.81 7.12 41.58 42.1

Area 1 3 3 3 2 2

Group SF CF N CF3 CF CF2-N CF2

On the basis of the above data, the product of the react

between C3 F7 N = SF2 and CF C C CF3 over CsF at 150

was assigned the structure


6




ion


CF3 CF2 CF2 N = S-C = CF CF .

F CF3

The yield of the product C F15 NS was 88 per cent based

on C3 F N = SF2.

The reaction of SOF2 and SF4 with CF3C=C CF3 over

CsF.--A mixture (14 g.) of thionyl fluoride and sulfur

tetrafluoride and perfluorobutyne-2 (10.8 g., 0.067 mole)

was condensed in the reaction vessel containing CsF and

allowed to stand at room temperature. A very small amount

of reaction was observed to occur at room temperature, as

indicated by drop of about 0.01 atm. over a 4 hour period.

When the reaction vessel was heated a maximum pressure of

8.67 atm. was observed at 900. The temperature was held

at 800 for next 24 hours and the pressure dropped to 2.99

atm. The following fractions were recovered:


Cr

87.27

1

CF
1









O
o 0 0 0 0
LO-CO CN
^ 1-. T.T,*,ii i I i r i 1 1 11 1 1 i iT ,1, i i i-


0

00





0
OC








0
C-

0








0



co
CN
r-*


N
0--0
In



II


r4

0
U)
Ife













.4-




4r,
Foe
1




O-




a







FI
O











ri






a







r-

60
c







Colorless gas 10 g.

Colorless liquid 9.5 g.

Apparent loss 6 g.

In order to recover the material (6 g.) which re-

mained in the vessel, it was heated to 2000 under vacuum

and an intensely blue-colored material (4.7 g.) condensed

in the liquid nitrogen cooled trap.

The colorless liquid product had a boiling range

of 1470-1480 and the V. P. C. analysis indicated that 95

per cent of the material was a mixture of two components

which were later shown to be C8F18 S2 02 and C8 F14 S2.

The blue-colored volatile liquid which was stored

at -780 was blood red in color after 24 hours. When the

red-colored product was warmed, the red color started dis-

appearing, and as a colorless gas, it reacted with the

Kel -F grease and glass. It also attacked the NaC1 windows

of the infrared cell and the V. P. C. column packing. When

this expanded gas was recondensed at -1950 the solid was

colorless. The infrared absorption spectrum of a repre-

sentative sample indicated a strong doublet at 1260 cm-1

and 1241 cm-1 along with the characteristic frequencies of

SOF2, SF4.

The infrared and V. P. C. analyses of the colorless

volatile product indicated that it contained 0.025 moles

of perfluorobutyne-2 and smaller amount of perfluorobutane.












i- 00i-
Drd OOc






0 -
d 00O-4


crc


0-1
C *-


*0\
-3-

i \r>


OO 4-O

O0

00
0







r-


c .d


co vI%
Cnr-, C J-

\0 fHl I +

I I

\D ro C



I % I



(0 0' t
LdPo


00 r-U
r-1













V 4
r^> 0


rd

0)
>
1--
0 rd

0- 0 0



00 3

-4 0
0 C\
N O
DNU a


4->
0+
aI

0 -



0 o
O .
C X)


r-1\ CN H i I r \O Cc'N (NJr"


Nco
CO O\
i,103



'-t I
!~o


I I

C\l C\




Q) r H
0+


..0 OQ rd
















C C 0r.
II








rd
0















0 ^


56

4.)



co



> -
4I o




C\N 4


4->
0)
a
bO
-


rd
0
0


0
4-)


H
c1


a 4)





o a

a do


C\ Cm i-4 C\\O Hr


-4P




cO
0

( 1

























o
rd










-I
E




co








0




0

0
c0
c,
3t

P<
0d


r-l
et A





-*


cO c-



I +



C,-


cO o\
*I






cOC\O

+


n

~cr



I)


U I












I <


rr 4
cc'



N -V


W od 0 ord a (












rl
P.
W-I
04

,r-lr f--1 4-I 3 )
4 0 0
O ki p r


-4 0



4)-
4 O -


o.


N C.--\O
\n D-O




Nl)~Vr


{>-
OT 0 C-W\ON

.-r C- C- I N
H\ I +
I


00 C 0v
NO I + --0
I r-i



s. Od C)


1-i
CO bo
cdb


0 W-I

rd0 HH

0
rd
H0





1-1
*tQ
0\ir


4-p



O

4-Dv

0 C)
0 0
rO


4-p 4-) 4-3
Q)4- (1) 0
4-3 -I -3

0202 c4
k c0h~


N\o


NHi


rd


Q)
r- H -i 4--) 4-)
hO P4 3 4
r-li 3
M 4 0


cnr-4 C- 0\
o CO C'-O\
'-i


rl \0 \0 rl i


d0 0 0 C r 4)


03c0




I I


ONCO c'-C^\
I t -t
ONC cO A
C\ 00 A




S+ -a




(4z. 0 d


*I *


00

+ cm


0 o o0


IN









Hrd
cH\
U0.















u~
L -



MU


f 0





S vU












0-
L


I









z



0 a
(r\
L.,
U /-~


0 0
c\ v-'



0 t/
/\ a





u



U3(d-'
0 CL0 CO


cf)
H
C-

0


c1v




a o
0 .-0


I









c o
0


NO
0












r-
cl





43
P,
V2<


0




wo
a-
4-)


X 4Z 4-3 4-)
a) a) )
r-l r-i a5 r-i

E 3 cxl 3
00 0
o d rd


(N o\ 0 CN 0 0
r-1 i-i
NW CHN C;C


\0 \0 C \O -t CHO






NI CI 0I C -- G\ r- 0 i I 0

'-Ct- nC + G\ c c- n cn +




* 4.) 4 4- *
%JHOnzCOCr, z \0 fN C-lO A a
1 1 + W-r auo ) r-i i i + on:r 00
0 001


H-i cn clcn C\ r-i rN t- c- C\) MN r-


\0 C'-O
-O O\N C-

cl (


r-i )C C- C

-. I I I
I +


C r





o0o


co o
...00


C- -0
Q\N
CO CM

I I





I +


COM

\0 C C\)
III


H-j\. CN
* *
C~r-l M
~-d-.


r r--- r--,
ac () () rO d a) W


cn
0



u -^

Ct, a) C
I C-



O c
I C02


I U


0 -O
I
c
I



0O

C


co5.0 d ( <)-i hbD


C.)
I

0M ---60




-, 0 I I


Ct


N

0 a)
0 0)


0-0 0^



0:4 Cr)
[I


.0 0



II U
Z








C a
Cv-'
Cr rd
0 *


ca .0o 'd d a) ho


bhO


II c

0-0
I
0-0
II


04



3
s:


Eo

x04
() 0 cl


Ei 0

0 E-4 0


a)
. X 4-> 4.

rri r-1 WCO
S ,



3 E 3'cr
000
CO0 (rt


Id

4-l 4-3
4. 4.) $

:3 4-> :3

tO


N C- N r-
emee~n


C00 o C-0

r-i


C

C


C

C












DISCUSSION


The most fundamental basis for mechanistic specu-

lation is the identification of the reaction products.

Vapor phase chromatography is very valuable for separation

of materials. Infrared and nuclear magnetic resonance

spectroscopy are among the more important techniques that

the chemist uses to gain information about a particular sub-

stance. But, then, there are instances when a highly re-

active and corrosive mixture defies all the "routine" ana-

lytical tools. During the progress of this work, two such

mixtures (A) and (B) were encountered which could not be

separated and identified.


"Thermal" Stability of Perfluoroalkyl
Iminosulfur Difluoride

The perfluoroalkyl iminosulfur difluorides were

synthesized (5) from nitriles and sulfur tetrafluoride

or sodium thiocynate and sulfur tetrafluoride at tempera-

tures ranging from 2000 to 3500 and pressures ranging from

120 to 200 atm. Obviously these sulfur tetrafluoride de-

rivatives are stable up to these temperatures under these

pressures. However, most of the reactions in this work were

expected to involve a maximum pressure of only 35 to 40

atm. Thus it was important to determine the stability

of Rf N=SF2 under the proposed experimental conditions.

59




60
Trifluoromethyl iminosulfur difluoride, CF3 N=SF2,

and perfluorobutene-2, CF3CF=CF CF3, were heated in a steel

vessel at 3500 for 24 hours without any decomposition.

When CF3N=SF2 was heated over CsF at 3000 (maximum pres-

sure observed was 15 atm.) for a period of three days,

no reaction was observed. This clearly demonstrated that

CF3N=SF2 is stable over CsF up to 3000 in a closed steel

vessel. Nearly 15 per cent of the starting material could

not be pumped out of the reaction vessel at room tempera-

ture. It remained in the vessel presumably as a complex

with CsF, i.e.,

CsF + CF3N = SF2---Cs+ (CF N = SF3-).

This ionic solid complex decomposes on heating above 2000

under vacuum.

When pentafluoroethyl iminosulfur difluoride, C2F5N = SF2,

was heated with CF3CF = CF CF3 over CsF, no reaction was

observed up to 2820. The maximum pressure at this tempera-

ture was 24.2 atm. Thus it is safe to conclude that C2 F5 N = SF2

is stable over CsF up to 2820. But when the temperature

was raised to 3000, a rapid rise in pressure was observed.

This abnormal rise in pressure is indicative of some de-

composition. When the products were analyzed, a quantita-

tive yield of CF3C=N was obtained,whereas no C2F5N = SF2

or SF4 was recovered. The liquid recovered from this re-

action was later found to be a product of the reaction in-

volving SF4 and CF3CF = CF CF3. Thus C2F5N = SF2 apparently

decomposes to SF4 and CF3CN, viz,






C2F5N = SF2 above 2820 CF CN + SF4.
Heptafluoropropyl iminosulfur difluoride is even less stable

than the ethyl derivative. When C3F7N = SF2 was heated

over CsF, no decomposition was observed at 2000. At 2500

a sudden rise in pressure occurred and hence decomposition.

The resulting products were an equimolar mixture of C2F5CN

and SF4, viz,

C3F7N = SF2 2500 > C2F5CN + SF4

The decomposition data indicates that the pyrolytic

stability of the fluorocarbon iminosulfur difluorides de-

creases with increasing chain length. For example,

CF N = SF2 is stable up to 3000 over CsF;

C2F5N = SF2 decomposes above 2820 over CsF; and

C3F N = SF2 decomposes at 2500 over CsF.

Fluorination of RfN = SF2 with High Valency
Metal Fluorides

Attempts to form an N-F bond in the RfN = SF2 com-

pounds with high valency metal fluorides tended to be frus-

trated by (a) rupture of the N = S bond and (b) reversible

decomposition of difluorides to nitrile and SF4,which in

turn frequently underwent fluorination. The metal fluorides

used were H F2, AgF2 and CoF3. The iminosulfur difluoride

mainly used in these studies was CF3N = SF2 because (a)

it is thermally most stable, (b) most readily prepared and

(c) easiest to handle in a vacuum system. In a typical re-

action 0.26 mole of CF N = SF2 and 1.32 mole of cobalt (III)

fluoride, CoF3, were loaded in a 100 ml. autoclave and heated





62

to 2000 for 24 hours. The maximum pressure recorded was

130 atm. Seven compounds were identified CF4, SF6, SOF2,
CF3SF3, (CF3)2NF, (CF3)2N -N(CF3)2 and N2.

A reasonable mechanism to account for these products

assumes decomposition of the starting material in presence

of CoF3.


CF N = SF2 CoF3

CF3N + F (from CoF3)


CF3 + SFF
3 3
CF3-N + CF'


2CF3 -N'

CF3

N' + NF

CF3- N' + (F')

CF3

SF3 + F'

SF4 + 2F (from CoF )

SF4 + (H20)


CF N + SF3
CF4 + N'
CF' + (NF)

CF3 SF3

CF3 -N'
3 I
CF3

(CF3)2 N-N(CF3)2


N2 + F'
CF3 -N -F

CF3


- SF4

) SF6

SSOF2 + HF


When SF4 is passed through the V. P. C. column, it is con-

verted to SOF2. Thus the appearance of SOF2 in the products

may result from reaction of SF4 with V. P. C. column pack-

ing.

There appeared to be zero retention of C N = S

backbone in the reaction of CF3N = SF2 and CoF3. It is con-

cluded that the N = S bond is probably the most reactive

bond in the system and is easily polarized and ruptured


--~-)


---------~~





63
in the presence of CoF3. Slight decomposition of CF3N = SF2

was observed as low as 1000 when AgF2 and CoF3 were used

as reagent.

The compounds RfN = SF2 have been fluorinated (44)

by a direct reaction with elemental fluorine to give CF3NF SF5,

CF3NF2,(CF3N = )2SF2 and SF6.

The Preparation of RfN = S--CF-- C3

F CF3

The N- perfluoroalkyl S- perfluoroisopropyl imino-

sulfur monofluorides, RfN = S(F) CF (CF3)2, were prepared

by reaction of the corresponding iminosulfur difluorides

with perfluoropropene over dry powdered cesium fluoride in

a stainless steel vessel at about 850. The reaction time

was several days. The autogenous pressure varied with time

but was in the order of 10 to 20 atm. The course of reaction
CsF
R N = SF2 + CF CF = CF2 -85- > RfN = S(F) CF (CF3)2

can be represented by the formation of the perfluoroiso-

propyl anion (CF3)2 CF which then attacks, as a Lewis base,
an empty d orbital of the sulfur atom in the iminosulfur

difluoride with subsequent elimination of a fluoride ion,

viz,

CF CF = CF2 + F- CF3-CF-CF3

Rf N = S-F -- RfN = S--CF-- CF3 + F-
I I
\ F CF3
CF CF CF
3 3

One problem associated with the identification of

these new compounds was that the characteristic infrared





64
spectral absorption associated with the -N = SF2 system

generally found around 1390 cm-1 (7.2Af) seemed to have

disappeared for the isopropyl derivative. The shift ap-

pears to have occurred to longer wave length and was prob-

ably obscured by the strong C-F stretching bands (27) which

lie in the range of 1400 to 1000 cm1. This hypothesis
was checked using the compound C6H5N = S (OCH3)2,which had

no interfering CF absorptions and in which the N = S ab-

sorption (28) was observed at 1283 cm-1. Absorption bands

were observed for CF3N = SF CF (CF3)2 at 1298 cm-1, for

C2F5N = SF CF (CF3)2 -at 1279 cm-1 and for C3F7N = SF CF (CF3)2
at 1272 cm-1 which are assigned to the N = S stretching

mode.

A second problem arose as a result of the n. m. r.

spectral splitting attributed to the fluorine atoms on

the carbon attached to nitrogen. In both the perfluorethyl

group in C2F5N = SF-CF (CF3)2 and perfluoropropyl group in

C3F N = SF-CF (CF3)2, the spectral splitting for -CF2-N-

is a non-equivalent quartet. Although CF3 on N in per-

fluoromethyl CF3N = SF-CF (CF3)2 appears to be a quartet,

it is actually two doublets so placed as to appear to be

an equivalent quartet presumably as a result of spin-spin

coupling with both the SF and CF fluorine atoms. It is

believed that quartets in the ethyl and propyl derivatives

are actually two doublets. The alkyl groups next to the

-CF2 group have a tendency to bend so as to be cis to one
and trans to the other fluorine atom of the -CF2-N group.





65
This causes both the fluorines to be non-equivalent which

results in two doublets which appear as a non-equivalent

quartet.


The Syntheses of Several
Perfluoromines

A second product which was isolated in the reaction

between trifluoromethyl iminosulfur difluoride and perfluoro-

propene over CsF had a molecular formula C7F15NS based on

its molecular weight and elemental analysis. The infrared

absorption spectrum showed a strong peak at 1780 cm-1 which

can be associated only with C = C or C = N stretching for

the available elements. Its n. m. r. spectrum indicated

five kinds of fluorine, one of them being attached to

sulfur and none to nitrogen. Normal valency considerations

ruled out the possibility of a -C = C- bond and indicated

the possibility of a C = N bond in a ring. The absorption

frequencies of the cyclic systems, nonafluoropyridine at

1754 cm-1 (29) and heptafluoropyroline at 1721 cm-1 (30)

have been associated with C = N stretching mode. On the

basis of elemental analysis, molecular weight, infrared and

n. m. r. spectra the compound C7F15NS has been assigned the

structure

CF--CF--C--S--CF-CF3

CF N F CF3

A probable mechanism of the reaction is through the

formation of CF3N = SF-CF (CF3)2. A fluoride ion attacks

an empty d orbital of the sulfur atom with subsequent re-






arrangement followed by elimination of fluoride ion,
CF -N CsF
CF N = SF2 + CFCF = C2 F CF N = S F CF
3 =S2 CF CF 3 1 1 3
F- CF2 '= SF CF (CF3)2 F CF
F




F

The intermediate has a'terminal difluoromethylene group

(CF2 = N) which is very susceptible to attack (31) by an-

other Carbonion,
F- CF = N SF CF (CF)2

CF CF CF3

--> (CF3)2 CF CF = N S (F) CF (CF )2 + F-
I
F

Abstraction of 2 fluorine atoms (CF CF-C=N-SF CF (CF)
and ring closure 3 2 3 2

The two fluorine atoms were taken up by CF3 CF = CF2
to form C3 F .

When C3F7N = SF2 and CF3 CF = CF2 were heated over
CsF, the main product was C3F7N = S(F) CF (CF3)2 along with

a minor product (16 mole %) which had a molecular formula

C6F13NS based on elemental analysis and molecular weight.
It has two fluorine atoms less than the major product C6F15NS.
The infrared absorption spectrum showed a bond at 1707 cm-1
which can be associated with C = C or C = N stretching






mode, but the normal valency considerations ruled out a

C = C bond and indicated the possibility of a C = N bond

in a ring. Its n. m. r. spectrum indicated five kinds of

fluorine, one of them being attached to sulfur. On the

basis of these observations it is possible to propose two

structures, viz,

CF 3- CF2 C -S- CF --CF (b)
3 11 1 (e)
N F CF3
(c) (d) (a)
(b)

(III)

and

CF-- C -CF2 (d)
3 2 (d)
(c) N--S -F (a)
CF3 CF CF

(b) (e) (b)

(IV)

The fine structure of the F19 n. m. r. spectrum shows that

peak (d) is an equivalent doublet. In structure (IV) the

S F fluorine atom can be oriented cis and trans to the

CF2 fluorines which would split the CF2 doublet in a non-

equivalent manner. Peak (c) in (IV) would be singlet and

not doublet. Moreover, the reaction temperature was only

820 and as C3F7N=SF2 did not isomerize over CsF even up

to 2000. The structure (IV) requires the rearrangement of

CF3CF2CF2N=SF2 to CF3-C-CF2 which might not be expected. There
El I
N-SF4

is one serious flaw in structure (III), that is, there is




68
no coupling between CF3 CF2- in peaks c and d. Both are
doublets. On the basis of evidence against structure (IV),

coupled with the fact that the minor product of reaction

between CF N = SF2 and CF CF = CF2 has a three membered

ring, the product C6F13NS is assigned the structure (III).
A probable mechanism to (III) is via the formation

of CF 7N = SF CF (CF3)2. A fluoride ion attacks on,empty

d orbital of the sulfur atom with subsequent rearrangement

followed by elimination of fluoride ion, vig,

C3FN = SF2 + CF CF = CF2 CsF C3F7N = SF CF (CF)2


> CF3 CF2--CF-N = SP CF (CF3)2



CF3- CF2- C- N = SF CF -(CF,)2 r F-
F F
F F

Abstraction of two fluorine atoms CSF CF
by CF3 CF = CF2 CF3-CF2-C -SF CF (3)2


The two fluorine atoms were taken up by perfluoropropene

which was converted to perfluoropropane.
Another fluoro imine was prepared when CF N = SF2

was heated with CF CF = CF CF3 over CsF at 2800 to 2900.

It was expected that the reaction would follow the usual

course; formation of carbanion which would attack the sul-

fur of iminosulfur difluoride with subsequent elimination

of F- ion, i.e.,

CF3 CF = CF CF3 + F- CF3 CF2 CF CF3

CF3 S-F
CF3 N S-F CFN = SF CF CF2CF + F
CF CF2 CF CF CF





69

but instead the imine, CF3N = C CF2 CF3 was obtained.

CF3

Perfluorobutere-2 is already in its most stable state

with respect to CsF isomerisation and the absence of a

terminal CF2- group makes acceptance of a F- more diffi-

cult. In that case formation of CF3-CF2- CF-CF is ex-

pected to require a high activation energy requiring tempera-

tures in the order of 2500 to 3000. It is possible that

the expected compound, CF3N = SF CF CF2 CF3, even if

CF3

formed, might be susceptible to decomposition at 3000

It has been shown that -N = S bond is susceptible

to rupture. It is postulated that the carbanion first at-

tacks the nitrogen atom of -N = S bond and SF2 is eliminated

which reacts with CsF to form the complex Cs SF3, viz,

CF N SF2 CF N SF
3 32 2
CF3 CF CF CF3 CF CF -CF CF3

C N-F FN N = C CF2CF3

CF CF C --- CF

CF
CF3 + SF* + F-


C + F- + SF' Cs SF

It would be expected that CsSF3 is a solid ionic species.

It appears to decompose on heating at 4000 under vacuum.

The decomposition products were sulfur and other unidenti-

fied volatile species some of which attacked glass, the sodium

chloride windows of the infrared cell, Kel-F grease and

mercury.





70
The perfluoroimine was identified on the basis of
elemental analysis, molecular weight, n. m. r. spectrum
and by comparison with the infrared spectra of the several
known imines (29, 32). The new imine shows infrared absorp-
tion at 1725 cm-1 and has boiling point 36.60. Haszeldine

(29) reported the synthesis of the isomer CF N = CF-CF2CF2CF3
(b. p. 390) and compared the infrared absorption fre-
quencies of the double bond in the following compounds:


Absorption
(i) CF3 N = CF2 C = N 1808 cm-1 (29)
(ii) CF3-N=CF CF3 C = N 1786 cm-1 (29)
(i11) CF3-N=CF CF2CF2CF3 C = N 1698 cm-1 (29)
(iv) CF3 CF2 CF = CF2 C = C 1789 cm-1 (21)
(v) CF3--C = CF2 C = C 1751 cm-1 (21)
CF3
Comparison of (i), (ii) and (iii) indicates that as the
chain length increases, the C = N absorption frequency de-
creases. Similarly comparison of (iv) and (v) indicates
that as branching increases, absorption frequency decreases.
A comparison of the properties of two isomers of C5 F1 N
further confirms that the isomer prepared in this work is
CF3 -N = C--CF2-CF .
CF3
In order to determine whether the syntheses like
that of CF3 N = C (CF3) C2 F5 prepared from CF N = SF2
and CF CF = CF CF over CsF is a general preparation of imines,

C2F5N = SF2 and CF3CF = CF CF3 were heated over CsF. No




71

C2F5N = C (CF3) C2F5 was obtained. The reason is that al-
though C2F5N = SF2 and CF3CF = CF CF3 were expected to re-

act at temperature around 3000, the decomposition of C2F5N=SF2
to SF4 and CF3CN took priority at temperature 2800 to 2900.

The Reactions of SF4 and SOF2
with Olefins

The reaction between C2F5N = SF2 and CF CF = CF CF3
over CsF was expected by analogy to result in either

C2F5N = S(F) CF(CF3)C2F5 or C2F5N = C(CF3)C2F5 but a product
with a molecular weight 426 and molecular formula C8gFS2
resulted instead.
If we consider that the iminosulfur difluorides

basically have the shape of the SOF2 molecule, the added

steric hindrance of the C2F5 group over that of CF3 group

might require a higher activation energy for a reaction
between C2F5N = SF2 and CF3CF = CF CF3. A sudden rise

in pressure at a temperature between 2900 to 3000 suggested
that a decomposition was occurring. The analysis of the
gaseous product indicated that CF3CN in quantitative yield,
with respect to C2F5N = SF2, had been formed.

C2F5N = SF2-->CF3 C =N + SF4
No SF4 was found. It had reacted with perfluorobutene-2.
When sulfur tetrafluoride and perfluorobutene-2 were heated

up to 3000 without CsF, no reaction was observed. The ex-
periment was repeated with SF4 and CF3CF = CF CF3 over CsF

and a reaction occurred at 2500. The product (C8F14S2)




72
obtained was the same as that formed when C2F5N = SF2 and

CF CF = CF CF were reacted.

The F19 n. m. r. spectrum of this new compound had

a single peak located at -19.9 ppm. with respect to trifluoro-

acetic acid and is in the range of CF3 groups attached to

a carbon atom which in turn is doubly bonded to another

carbon atom. The presence of C = C bond is confirmed by

a band in the infrared spectrum at 1626 cm-1. Krespan (33)

prepared CF3 C = C CF3 by heating perfluorobutyne-2
S-S
with sulfur in presence of iodine at 2000. The infrared

spectrum showed a band at 1629 cm-1. The F19 n. m. r. spec-

trum exhibited a single peak in the trifluoromethyl region

at (-811 cps.) -14.4 ppm.

On the basis of elemental analysis, molecular weight,

infrared and n. m. r. spectra the new compound was assigned

the structure

CF3 C = C CF3
I 1
S S

CF3 CF2 C -CF3

The reaction can be represented by these equations:

2 CF3CF = CF CF3 + 2 SF CsF C8F14S2 + 5F2

CF3 CF = CF CF3 + F2 CF> CF2CF2CF3C

The compound C8F14S2 was also obtained when CF C = C CF3

and SF4 were heated over CsF. The reaction occurred at

1500. Perfluorobutane was also recovered.






CF3 C =C CF3 + SF4--- C8F14S2 + F2
CF3 C C CF3 + F2--- CF3 CF2 CF2 CF3

The lower temperature required for perfluorobutyne-2 reac-

tion (1500) than for the perfluorobutene-2 reaction (2500)

is a result of the smaller steric effect about the triple
bond. Moreover perfluorobutyne-2 might be a stronger re-

ducing agent than perfluorobutene-2 and so less activation
energy is required for the reaction with perfluorobutyne-2.

In the syntheses of C8g14S2, the sulfur of SF4

has been reduced from +4 to +2 oxidation state. Loth and

Michaelis (15) treated thionyl chloride with anisole in
presence of zinc chloride and obtained 4,4 dimethyl diphenyl

sulfide in which sulfur has been reduced from the +4 to the
+2 oxidation state, viz,

C6 H5.OCH3 + SOC12 ZnC12 CH30. C6H4- S--C6H4. OCH
Recently Smith et,al. (5) treated phenyl iminosulfur di-

fluoride with anisole in presence of zinc chloride and
obtained 4,4 dimethyl diphenyl sulfide. They did not re-

port a mechanism. But when perfluoropropene was heated
with SF4 over CsF (43) at 1500, the products obtained were

(CF3)2 CF SF3 and (CF3)2 CF--SF2--CF(CF3)2. Though the
authors did not report any mechanism, it is apparent, the
following course of reaction takes place:

CF3CF = CF2 + F- CF3CF CF3

(CF3)2CF + SF3 (CF3)2CF SF + F-

(CF )2jF + (CF3)2CF SF2 F-- (CF)2CF-S2-CF(C)2+ F
S(VI)





74
The sulfur of (VI) which is in +4 oxidation state can be

reduced to +2 state by treatment with TiC14 (43).

Perfluoropropene and perfluorobutene-2 react with

SF4 at 1500 and 2500 and produce different products.
Probably the propene has a terminal =CF2 group which is

very susceptible to fluoride ion attack and so lower reac-

tion temperature is needed. But the lower temperature

may not be high enough for the propene to act as a reduc-

ing agent. It is believed that in this work perfluorobutene-2

and perfluorobutyne-2 are acting as reducing agents and

accept fluorine.

The material balance of the reactants and products

showed a loss of 10 grams which was SF4. The SF4 reacted

with CsF (34) to form CsSF5 which is an ionic solid and

.thus remained in the reactor when the volatile products

were removed. The solid complex decomposes reversibly on

strong beating under vacuum to give SF4 and CsF.

The reaction between C2F5N = SF2 and CF CF = CF CF3
over CsF at 2900 was in fact a catalysed reaction between

SF4 and CF CF = CF CF3. In order to justify this observa-

tion SF4 and CF3CF = CF CF3 were heated over CsF and a re-

action occurred at 2500. The V. P. C. analysis of the liquid

product showed a shoulder that could not be resolved on

any V. P. C. column available, silicone oil, silicone gum,

silicone elastomer, Kel-F ester or D. N. P., at tempera-

tures below 700 and at very low helium flow rate. This

mixture had a boiling point 1420 whereas C8F14S2 boils at





75

1350. Comparison of infrared spectra of pure C8F14S2 (Fig.
7) and the mixture (Fig. 9) indicates that the mixture
did contain C8F14S2. It was assumed that one of the other

products in the mixture was the product of a reaction be-

tween SOF2 and CF CF = CF CF3, since commercial sulfur

tetrafluoride (b. p. -380) contains some thionyl fluoride

(b. p. -450).

In order to test this hypothesis, SOF2 was reacted

with CF CF = CF CF3 over CsF. At the reaction temperature

of 2000, the main product obtained boiled at 1490. The

molecular formula calculated from the elemental analysis

and molecular weight was C8F18 S202. The infrared absorp-

tion spectrum (Fig. 8) did not show any bond in C = C ab-

sorption region. The compounds C8F14 S2 and C8F18S202

were mixed in the ratio 1:3 and the mixture boiled at 1430

The infrared spectrum of the synthetic mixture was similar

to that in Figure 9. This proved that the mixture, obtained

as a result of reacting SF4 with CF CF = CF CF3 over CsF

contained C8F14 2 and C8F18S202 in the ratio of approximately

1:3. The boiling point of the mixture varied as a function
of its composition.

The n. m. r. spectrum of C8F18S202 consisted of

4 pairs of peaks which were all multiplets. One pair of

peaks (CF3 groups) was resolved more than another. A pair

of peaks (CF2 groups) could not be resolved completely

but the fourth pair of peaks (CF groups) were resolved.

On the basis of above information it was estimated that there





76
are two isobutyl groups, CF3-CF2 -CF-CF3 in slightly dif-
ferent magnetic environments in the compound. The two
possible structures are
CF 0 CF CF 0 CF
I II I 11
CF-- S--O-S-CF CF- S-S-CF

CF2 CF2 and CF2 0 CF2
CF3 CF3 CF3 CF3
(VII) (VIII)
Structure (VIII) was ruled out on the basis of infrared
spectral data. The infrared spectrum of sulfuryl fluoride

SO2F2 (20) shows a strong 0 = S = 0 stretching frequency
o o
at 1500 cm-1. Similarly 0 = S = 0 in II II
0 = S 0-- CF -CF-0-S-0
I I I I
F CF2 CF2 F

CF2
and CF3CF2 CF CF3 has stretching frequencies at 1496
I
-- S C0
I
F
cm-1 (35) and 1500 cm-1 (36), respectively. The infrared
spectrum of C8F18S2 02 does not show any absorption near
1500 cm-1 which is characteristic of two oxygens doubly
bonded to one sulfur atom (37). Therefore structure (VIII)
is ruled out. Thionyl fluoride shows a S = 0 stretching
frequency at 1308 cm-1 (38) and SOF4 shows a S = 0 stretch-
ing at 1383 cm-1 (39) whereas oxygen singly bonded to sulfur,
S 0 X, as in fluoroformyl sulfuryl fluoride, shows a

medium intensity band at 757 cm-1 (37). The compound C8F18S2 02

shows a strong doublet at 1351 cm-1 and 1333 cm-1 which is





77
characteristic of an infrared absorption of a grouping con-

sisting of one oxygen doubly bonded to sulfur.

The mass spectral data for C8F18S202 shows a very

strong peak at mass number 64 which indicates either 0-S=0

or -S-S. As the -S-S- bond is much weaker than the -S=0
+
bond, statistically an 0-S-O ion has better chance of sur-

vival under mass spectral conditions. M. Goehring prepared

cyclic S3N202 (40) (by heating SOC12 with S4N4) having a

-0-S=0 linkage which on reaction with moist nitrogen de-

composed (41) to give S Nq and SO2. Based on theseargu-

ments, structure (VII)

CF 0 CF3
3 II
CF- S-0-S-CF
I I
CF2 CF2

CF3 CF3

seems to be more plausible.

A major by-product of the reaction between SOF2

and CF3CF = CF CF3 over CsF was extremely reactive. This

gaseous component could not be separated by V. P. C. as it

reacted with the column packing. The infrared absorption

spectrum was not reliable since it reacted with the NaC1

windows and glass walls of the infrared cell. It could not

be distilled because it reacted with glass and Kel-F grease.

Thus, this reactive species was not identified and hence

it is not possible to be certain about the mechanism of the

reaction.

2 SOF2 + 2 CFCF = CF CF3 CsF- C S 0 (A)
200o C8 F18 S2 02 + (A)





78
No perfluorobutane was detected among the products which

indicates that CF3CF = CF CF3 did not act as a reducing

agent as it did in case of the reaction between SF4 and

CF CF = CF CF3 over CsF at 2500. The only other possible

reducing species was SOF2 although no SOF4 or F5S- OF was

detected. Thus the reactive product (A) must be an oxidized

species.

The first step in the reaction between SOF2 and

CF CF = CF CF3 involves the nucleophilic attack of F~ ion

on a carbon atom in the C = C bond of perfluorobutene-2

to form the carbanion:

CF3CF = CF -CF3 + F- )CF3 CF2 CF -CF3

The carbanion attacks sulfur with subsequent elimination

of F- viz,

0 0
11 2000 II
F S -F -- ) F S
t- I +F~
CF CF CF2 CF3 CF3 CF CF2 CF3

Rf = CF CF2 CF CF3

F S = 0 + F--->F3S--0"

F

0 0

Rf--S-F --- f S 0 SF + F

F3S o0
0 0
Rf- I1 II
Rf 0 SF F-- Rf- S- 0- S -Rf
T- F2
Rf






0 0
II II
Rf S 0 S- Rf + SOF2---Rf S 0 S Rf + (A)
F2
It is just possible that (A) is F3S 0 F.

Another possible mechanism involves free radicals:

CF CF = CF CF3 + F -- CF3CF2 CF CF3
0 0
SOF2 R S F 2000 1
SOF2 F 2000 ) R S. + F*

:0:

Rf -- rearrangement > Rf- -- '
0 0
II It
Rf---S. + 'O--S-Rf--- S O S Rf

F' + F -- F2

Elementary fluorine is known to react with SOF2 in presence

of CsF (42) to produce SF5OF in quantitative yield. The

infrared absorption spectrum of SF5OF shows a band at 935

cm-1 (39),but no such band was detected in the spectra of

the products. Fluorine also reacts with SOF2 without any

catalyst (42) and the product is F4S = 0 which was not de-

tected either. Though it is possible that SOF4 might have

reacted with CsF (46) and the solid ionic product Cs-o-SF5

remained in the reactor when the volatile products were re-

moved. Thus the free radical mechanism does not seem to be

more plausible than the ionic mechanism.

A material balance on the above reaction indicated

a loss of 18 grams. This, probably, is a result of the re-

action of CsF and SOF2 to give Cs- 0 -SF3 (42). This sup-

posedly solid ionic compound remained in the reactor after





80

removal of volatile components. The compound Cs-OSF3 should

decompose on heating to 3000 to 4000 to yield volatile

species.

Perfluorobutyne-2 and thionyl fluoride were heated

over CsF and the liquid product was C8 F18 S2 02. The vola-

tile products did not contain the extremely reactive by-

product (A) which was obtained from the reaction of SOF2

and CF CF = CF CF over CsF. The reason is that CF C C CF3
3 3 3
requires three atoms of fluorine for conversion to the

perfluorobutyl radical and hence no excess of fluorine

is involved.

2CF3C =CCF3 + 3 SOF2 -- C8F18 S202 + (B)

The by-product (3) remained in the reaction vessel.

It may have been absorbed on CsF or reacted to form a com-

plex similar to that formed by SOF4 (46), SOF2 (42) and

S- (34) with CsF. When the reaction vessel was heated to

3000 under vacuum, the effluent obtained had an intense

blue color which changed to red on standing at -780 for 24

hours. On warming, the red-colored species changed to a

colorless gas. Apparently the color transformation was

associated with a free radical. But it is not clear whe-

ther the free radical obtained is the by-product (B) or

a decomposition product of a complex (CsF + (B).). As the

ide tity of (B) is not known, it is not possible to specu-

late intelligently about the mechanism of the reaction be-

tween SOF2 and CF 3C CCF3 over CsF.

Since C3F7 N = SF2 was known to be less "thermally"




81

stable than C2F5N = SF2 and since the extra steric effect

of the C3F7 group over that of C2F5 group would call

for a higher activation energy, it seemed apparent that

any reaction that occurred would be between SF4 and CF3CF=CFCF3,

the results of which have been established above. Thus,

no attempt was made to react C3F N = SF2 and CF3CF = CF CF3

over CsF.

The Reaction of Rf N = SF2 with

CF C= C CF3

Miller (25) heated CF2 CF-CF = CF2 over CsF at

1500 and obtained CF3 C CCF The reaction clearly demon-

strated that perfluorobutyne-2 is stable over CsF at least

up to 150.

The reaction between RfN = SF2 and CF3CsCCF3 over

CsF was expected to yield a product having conjugated double

bond, RfN = SF -C = CF- CF3. It was true in case of ethyl

and propyl iminosulfur difluorides and the derivatives

formed were CF3N S C = CF -CF3 and CF3CF2N = S -C = CF CF3.

F CF3 F CF3

The course of the reaction appears to follow the

same route as that postulated for the preparation of

RfN = SF CF (CF3)2. The fluoride ion attacks the olefin

to produce a carbanion,

CF C=C CF3 + F-- CF3CF = C- CF3

The carbanion adds to the sulfur atom of the iminosulfur

difluoride with the elimination of fluoride ion.






CF3 CF2 N = SF F
T_
CF CF = C CF --- CF3CF2 N = S C = CF CF

F CF3
The boiling point of ethyl and propyl derivatives,980 and

1110 respectively, were extrapolated to determine the ex-

pected boiling point for the methyl derivative

CF N = S C = CF CF3 (approximate b. p. 850).

F CF3

However, the major product of the reaction between CF3N = SF2

and CF C3 C CF3 over CsF boiled at 1670 and its infrared

absorption spectrum (Fig. 10) did not show any absorption

characteristic of the C = C or C = N bond. Obviously it

was not CF3N = SF-C (CF3) = CF CF3. The molecular weight

and the elemental analysis of the pure product indicated

that it had a molecular formula C10F22 N2 S2 which is a

dimer of C5Fll NS.
When the reaction vessel containing CF3N = SF2,

CF C=C CF3 and CsF was heated to 800, the heat evolved

was enough to maintain the system at that temperature un-

til the reaction was over. The reaction to produce

C2F5N = SF-C (CF3) = CF CF3 and C3F7N = SF C (CF3) = CF CF3
were not observed to be exothermic and the process to produce

CF N = SF C (CF3) = CF CF3 was not expected to be exothermic.

Presumably the secondary reaction which resulted in the

dimerisation accounts for the observed exothermicity. The

infrared absorption spectrum of C10 F22 N2 S2 (Fig. 10)

bears a strong resemblance to infrared spectrum of CgF18 S202




83

(Fig. 8) because of the secondary butyl groups. The n. m. r.

spectrum showed the compound C10F22N2S2 was a mixture of two

isomers in the ratio 2:3. The spectrum was so complex and

unresolved enough so that attempts to differentiate between

the two isomers cis and trans, had to be abandoned. How-

ever the n. m. r. spectrum did show that both isomers had

two isobutyl groups in slightly different magnetic environ-

ments as in case of C8F18S202.

The peak (a) in the n. m. r. spectrum (Table 2)

appeared at -16.6 ppm. which was assigned to fluorines in

CF3-N- group. This assignment is confirmed by the fact

that fluorine resonances in the CF3N group of the compound

CF N = C (CF ) CF2 CF3 appear at -16.88 ppm.

A minor problem arose as a result of the low field

value of +3.4 ppm. for peak (d) in the new compound C10 F22 N2 S2

which from its relative area was attributed to a single fluorine

atom on a carbon atom. The reason that this CF resonance

appears at the normal region for CF resonance, far down

field with respect to a normal CF resonance (approximately

80 to 100 ppm.) (Table 2) is attributed to its position

in the molecule. Based on elemental analysis, molecular

weight, boiling point, infrared and n. m. r. spectrum the

compound C10 F22 N2 S2 is assigned the structure

CF CF CF
I 3 1 I 3
CF-- S- N-- S-CF
I II II I
CF2 CF N CF2

CF3 CF3





84
The anamolous resonance of CF in the ring is due to the

decrease of the electron density around the fluorine atom;

the lower the electron density around the fluorine, the

lower the shielding and the lower the field at which the

fluorine resonance is observed. The fluorine in the ring

is deshielded as a result of the following factors: (1)

Presence of the electronegative nitrogen atom next to carbon

atom drains off some electron density. (2) The cyclic part

contains a large loop of 7T electrons in which strong dia-

magnetic currents are induced by the magnetic field. This

effect deshields the fluoride. (3) The relatively electro-

negative isobutyl groups could possibly further drain off

the electron density from the ring resulting in further

deshielding of the fluorine nucleus.

The reaction of perfluorobutyne-2 and CF N = SF2

over CsF at room temperature yielded only two products,

81 mole per cent C10 F22 N2 S2 and 19 mole per cent of an-

other product whose n. m. r. spectrum did not indicate an

SF bond.

Howeve; in another reaction in which the tempera-

ture was maintained at 560 for 4 hours, one of the products

which was in 8 mole per cent showed a strong infrared spectral

absorption at 1748 cm-1 (Fig. 11). Extrapolation of C = C

bond absorption frequencies of the propyl and ethyl deriva-

tives, Rf -N = SF C (CF3) = CF CF3, indicates that the

methyl derivative should have a C = C bond spectral absorp-

tion between 1735 to 1750 cm1.






CF N = SF C = CF CF 1715 cm-1

CF3
C2F5N = SF C = CF CF3 1724 cm-1

CF3

CF N = SF C = CF CF3 (1735-1750 cm-1)

CF3

The spectrum (Fig. 11) of the product present in 8 mole

per cent resembles that in Figures 12 and 13 in all essen-

tial features. The quantity obtained was so meager that

further structural confirmation of the material was not

possible. All subsequent attempts to increase its yield
were unsuccessful. It appears that although C5F11NS is

formed, it largely dimerises in presence of CsF. Probably
the presence of conjugated double bond system facilitates

the dimerisation process. The ethyl and propyl derivatives
show little tendency to dimerise under similar conditions

as a result of the greater steric effect of bulkier groups.
The dimerisation of C5F11NS to C10F22N2S2 in pres-
ence of CsF can be regarded to occur as follows:

CF N = S C = CF CF
S I --->CF3N = S = C-CF2CF3 + F-
CF CF T
F-

CF3 -N = S = C (CF3) CF2 CF3
CF N = S- CF2CF3
F CF3

Rf = CF3 CF2 CF CF

Two moles of anion condense with elimination of F-.





CF3
S-S =N
f a


CF

S=N
I -s
I II
CF2 N


CF
13
B S = N--- CF
I- S- Rf I3
II R S = N S
CF -CN N
I C F--N
0F


- Rf + F


Rf
Rf


- Rf











SUMMARY


Some reactions of perfluoroalkyl iminosulfur difluo-

rides, sulfur tetrafluoride and thionyl fluoride with per-

fluoroolefins in presence of dry powdered cesium fluoride

have been studied.

The thermal stability of perfluoroalkyl iminosulfur

difluories over CsF was determined. The perfluoromethyl

iminosulfur difluoride, CF3N = SF2, was most stable and did

not decompose even at 3000 for 72 hours. The perfluoro-

ethyl homolog, C2F5 N = SF2, did not decompose below 2820

whereas the perfluoropropyl homolog, C3F7N = SF2, decom-

posed to the corresponding nitrile and SF4 at as low as

2500. The thermal stability decreases with the increasing

chain length.

A new class of fluorocarbon compounds, N perfluoro-

alkyl S perfluoroisopropyl iminosulfur monofluoride, in-

volving a -C-N=S-C bonding arrangement has been prepared

by the reaction of the analogous fluorocarbon iminosulfur

difluorides with perfluoropropene over solid powdered CsF

at 80-900 in a closed vessel. A minor product of these re-

actions was isolated and identified as a unique cyclic imine

involving a three membered rings,-C-S-.
H/
N

The reaction between CF3N = SF2 and perfluorobutene-2,

87




88
CF3 CF = CF CF3, at 3000 did not follow the trend observed
with perfluoropropene; instead an imine CF N = C(CF3)C2FS

'is produced. The reaction between C2FS N = SF2 and
CF3 CF = CF CF3 failed to occur up to 2820 as the prior

decomposition of iminosulfur compound to CF CN and SF4

prevailed. The product formed was exactly the same as that

obtained from the reaction of SF4 and CF3 CF = CF CF3 over

CsF and was identified as the five membered ring compound

CF3 C S The reaction of perfluorobutyne-2
II C(CF3) C2F5.
CF3 -C -S

and SF4 over CsF at 1500 also yielded the same substance.

Thionyl fluoride reacted readily with CF3 CF = CF CF3
at 2000 and with CF3 C=C CF3 at 1500. In both reactions

the major liquid product was identified as
0
C2F5 (CF3) CF S 0 S CF (CF3) C2F5. The gaseous

product was extremely reactive and was believed to be
SF OF.

Another new class of compounds bearing a conjugated
double bond system, -N = S C = C was prepared. The
reaction of C2 F5 N = SF2 and C3 F7 N = SF2 with CF3 C=CCF3
over CsF at 100 to 1500 produced sulfur monofluorides of the
structure Rf N = S C = CF CF3, in almost quantitative

F CF3
yield. But the reaction of CF3N = SF2 and CF3 C =C CF3
over CsF produced only a trace of sulfur monofluoride,

CF3N = S C = CF CF3, the main product being its dimer

F CF3






C10 F22 N2 S2 identified as
/N(CF )

C2 F5 (CF3) CF S S CF (CF3) C2 F5.
%CF N
The following table summarizes the pairs of reagents
and products respectively involved in each of the CsF cata-
lyzed reactions described in this dissertations.

Table 3
The Reactants and Products of CsF Catalyzed Reactions

Reactants Reaction Isolated Products boiling
Temp. point
CO


CF3N=SF2 and
CF CF=CF2




C2F5N=SF2 and
CF CF=CF2

C3F7N=SF2 and
CF3CF-CF2


830






890



820


1. CF3-N=S-CF-CF3
F CF3

2. CF3-CF -C-S-CF-CF3
I 11/1 I
CF3 N F CF3

C2F5N=S-CF-CF3
F CF3

1. C3F N=S-CF-CF3

F CF3


2. CF3-CF2-C-S-CF-CF
SF/ CF3
N F CF3


65.50


75.50


82.30


103.50


1000






Table 3 (continued)


CF3N=SF2 and
CF CF=CF CF3

C2F5N=SF2 and
CF3CF=CF CF3


SF4 and


280


2900


2500


CF CF=CF CF3


SF4 and


1500


CF3CiC CF3


CF N=C-CF2-CF3
3 F
CF3
CF


CF3-C-S,
11 C-CF2CF



CF -C-S
SII C-CF2 CF3
CF3-C-S' I
CF3


CF3-C-S
It C-CF2CF3
CF3-C-S I
CF3


SOF2 and


2000


CF3CF=CF CF3


SOF2 and


1500


CF C-C CF3


CF N=SF2 and
CF3 CC CF3


C2F5N=SF2 and
CF 3CC CF3

C3F7N=SF2 and
CF3CEC CF3


560


C2F5-CF-S-0-S-CF-C2F5
S III I
CF3 0 CF3

C2F5-CF-S-0-S-CF-C2F5
5 I 11 1
CF3 0 CF3

1. IF3
N
C2F5-CF S "S-CF C2F5
I II II
CF3 CF-N CF3
2. CF3-N=S-C = CF CF3
F CF3
F CF3


1240 C2F5-N=S C=CF CF3
F C
F CF3


1500


C3F7N=s C=CF CF3
F CF3


36.60


135



1350


1350


1490



1490


1670


980


1110











BIBLIOGRAPHY


1. 0. Glemser, Angew. Chem. Internat. Edit., ., 530 (1960).

2. A. Michaelis, Liebig Ann. Chem., 274, 201 (1893).

3. 0. Glemser and H. Schroder, Z. anorg. allgem. Chem.,
284, 97 (1956).
4. 0. Glemser and H. Richert, Z. anorg. allgem. Chem.,
307, 313 (1961).
5. W. C. Smith, C. W. Tullock, R. D. Smith and V. A. Engelhardt,
J. Am. Chem. Soc., 82, 551 (1960).

6. W. T. Miller, J. H. Fried and H. Goldwhite, J. Am. Chem.
Soc., 82, 3091 (1960).

7. S. Andreades, J. Am. Chem. Soc., 86, 2003 (1964).
8. W. H. Christie, Doctoral Dissertation, University of
Florida (1958).
9. R. D. Dresdner, F. N. Tlumac and J. A. Young, J. Am.
Chem. Soc., 82, 5831 (1960).
10. E. R. Van Artsdalen, A. R. Brosi, T. A. Gens and J. A.
Wethington, J. Am. Chem. Soc., 2, 1001 (1957).
11. W. T. Miller, W. Frass and P. R. Resnick, J. Am. Chem.
Soc., 83, 1767 (1961).
12. R. D. Dresdner, F. N. Tlumac and J. A. Young, J. Org.
Chem., 30, 3524 (1965).

13. F. S. Fawcett, C. W. Tullock and D. D. Coffman, J. Am.
Chem. Soc., 84, 4275 (1962).
14. W. R. Hasek, W. C. Smith and V. A. Engelhardt, J. Am.
Chem. Soc., 82, 543 (1960).

15. F. Loth and A. Michaelis, Ber., 27, 2540 (1894).
16. A. Michaelis, Ann., 274, 262.(1893).







17. L. C. Yen, Master of Science Thesis, University of Florida
(1957).
18. P. A. Bond and D. A. Williams, J. Am. Chem. Soc., 53,
34 (1932).
19. A. M. Lovelace, D. A. Rausch and W. Pastelnek, Editors,
"Aliphatic Fluorine Compounds." Reinhold Publishing
Corporation, New York, 1958.

20. J. H. Simons, Editor, "Fluorine Chemistry," Vol. II.
Academic Press, Inc., New York, 1950.

21. T. J. Brice, J. D. LaZerte, L. J. Hals and W. H. Pearlson,
J. Am. Chem. Soc., 25, 2698 (1953).

22. R. D. Dresdner, T. J. Mao and J. A. Young, J. Am. Chem.
Soc., 80, 3007 (1958).

23. R. N. Haszeldine, J. Chem. Soc., 2504 (1953).

24. J. A. Young, Private Communication, Dunn Research Inst.

25. W. T. Miller, W. Frass and P. R. Resnick, J. Am. Chem.
Soc., 83, 1767 (1961).
26. A. L. Henne and W. G. Finnegan, J. Am. Chem. Soc., 71,
298 (1949).

27. M. Stacey, J. C. Tatlow and A. G. Sharpe, Editors,
"Advances in Fluorine Chemistry." Butterworths, Washing-
ton, 1965, Vol. 4, pp. 253-313.

28. V. A. Engelhardt, Private Communication.

29. R. E. Banks, W. M. Cheng and R. N. Haszeldine, J. Chem.
Soc., 3407 (1962).
30. B. C. Bishop, J. B. Hynes and L. A. Bigelow, J. Am. Chem.
Soc., 84, 3409 (1962).

31. J. H. Fried and W. T. Miller, J. Am. Chem. Soc., 81
2078 (1959).

32. J. B. Hynes, B. C. Bishop, P. Bandophay- and L. A.
Bigelow, J. Am. Chem. Soc., 85, 83 (1963).

33. C. G. Krespan, J. Am. Chem. Soc., 83., 3434 (1961).

34. C. W. Tullock, D. D. Coffman and E. L. Mutterties, J.
Am. Chem. Soc., 86, 359 (1964).







35. J. M. Shreeve and G. H. Cady, J. Am. Chem. Soc., 83,
4521 (1961).
36. J. D. Delfino and J. M. Shreeve, Inorg. Chem., 5, 308
(1966).
37. W. B. Fox and G. Franz, Inorg. Chem., 3., 946 (1966).
38. R. J. Gillespie and E. A. Robinson, Canad. J. Chem.,
2., 2171 (1961).
39. F. B. Dudley, G. H. Cady and D. F. Eggers, J. Am. Chem.
Soc., Z8, 1553 (1956).
40. M. Goehring and U. J. Heinke, Z. anorg. u. allgem.
Chem., a2. 297 (1953).
41. M. Goehring and U. J. Heinke, Z. anorg. u. allgem. Chem.,
278, 53 (1955).
42. J. K. Ruff and Max Lustig, Inorg. Chem., 3, 1422 (1964).

43. R. M. Rosenberg and E. L. Muetterties, Inorg. Chem.,
1, 756 (1962).
44. M. Lustig and G. K. Ruff, Inorg. Chem., 4-, 1444 (1965).
45. M. Lustig and J. K. Ruff, Inorg. Chem., 3, 287 (1964).
46. W. C. Smith and V. A. Engelhardt, J. Am. Chem. Soc.,
82, 3838 (1960).












BIOGRAPHICAL SKETCH


Jogindar Singh Johar was born in Rawalpindi, India

(now in West Pakistan) on January 1, 1935, and attended

G. N. Khalsa A. V. Middle School, Dhamyal (near Rawalpindi)

up to 1947. Because of political disturbances in the Indian

Subcontinent at the time of Independence, he was forced to

leave his home town in September, 1947, and spent one year

in Refugee Camps. He joined school again in 1948 and passed

the Matriculation Examination held by Panjab University in

1951. He obtained Bachelor of Science degree in 1955 and

received B. Sc. (Hons. School) degree in Chemistry in 1947.

After completing all the requirements towards a Master's

degree in Chemistry (by Thesis) in 1958 he joined Arya

College Ludhiana as a lecturer in Chemistry, though the

M. Sc. (Hons.) degree was officially awarded to him in

August, 1959.

He was selected to Panjab Education Service in

November, 1959, and was appointed as a Lecturer in Chemistry

at the Government College Gurdaspur. He resigned that po-

sition when he came to the United States to join the University

of Florida in the fall of 1962.

He is married to the former Manjit Kaur Sodhia and

is the father of one son.




Full Text

PAGE 1

THE SYNTHESES OF DERIVATIVES OF IMINOSULFUR DIFLUORIDES By JOGINDAR SINGH JOHAR A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTL^L FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA August, 1966

PAGE 2

ACKNOWLEDGMENTS The author takes this opportunity to express his sincere appreciation and deep gratitude to his research director, Dr. Richard D. Dresdner, for his inspiration and assistance in carrying out this work and in the writing of this dissertation. His friendship along with his sincere desire to help have been a constant source of encouragement throughout the course of this work. Appreciation is extended to the other members of the author's Graduate Supervisory Committee for making suggestions in the writing of this dissertation. The author is very grateful to Dr. V. S. Brey, Jr. and coworkers for obtaining and interpreting the nuclear magnetic resonance spectra for this study. The author also expresses his gratitude to Dr. R. J. Hanrahan under v;hose direction the mass spectra were determined and interpreted. Acknowledgment is made to the Army Research OfficeDurham for support of a part of this work. The author wishes to thank the faculty, fellov/ graduate students and staff of the Chemistry Department who have been most helpful during the course of this study. He also wishes to thank Mrs. Shirley Ingram for typing the final copy of this dissertation. 11

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGMENTS H LIST OP TABLES iv LIST OP PIGURES v INTRODUCTION 1 Statement of the Problem 7 EXPERIMENTAL 10 Apparatus 10 Materials 12 General Procedure 15 DISCUSSION 59 "Thermal" Stability of Perf luoroalkyl Iminosulfur Difluoride 59 Pluorination of R^n = SPp with High Valency Metal Pluorides 6I The Preparation of RfN = S — CP — CP^ 63 I , , 3 F CP^ The Syntheses of Several Perf luoromines .... 65 The Reactions of SP^^. and SOPg with Olefins • • .7I The Reaction of R^ N = SP2 with CF3CSC CP3 . . 81 SUMMARY S7 BIBLIOGRAPHY ^1 BIOGRAPHICAL SKETCH 94 iil

PAGE 4

LIST OF TABLES Table Page 1. Reaction of CF,N=SF2 with Metal Fluorides ... 19 2. Nuclear Magnetic Resonance (N.M.R.) Spectra . . 56 3. The Reactants and Products of CsF Catalyzed Reactions 89 iv

PAGE 5

LIST OF FIGURES Figure Pag© 1. Infrared absorption spectrum of CF,N=S-CF-CF, 21 F CF, 2. Infrared absorption spectriim of CF,CF2N=S-CF-CF, 22 F OF, 3. Infrared absorption spectrum of CFj-CF2-CF2-N=S-CF-CF5 23 F OF;. ^. Infrared absorption spectrum of CF^-CF — C — S-CF-CF2 24 OF, N*^ F OF, ^ 3 5. Infrared absorption spectrum of CFj-CF^-C — S-CF-CF, 31 N^ F OF, 3 6. Infrared absorption spectrum of CF,N=C-CF2-CF5 '. . . 35 CF3 7. Infrared absorption spectrum of CoF^^Sp .... 37 8. Infrared absorption spectrum of CqF, gSoOo ... 40 9. Infrared absorption spectrum of mixture of ^8^14^2"^°8^18^2^2 ^^ 10. Infrared absorption spectrum of C,qF22N2S2. . . 48 11. Infrared absorption spectrum of CF^-N=S — C — CF-GF2 49 F OF, 3 V

PAGE 6

Figure Page 12. Infrared absorption, spectrum of CF-,-CFo-N=S C=CF-CF, 51 5 2 I , 5 F CF^ 15. Infrared absorption spectrum of CF,CF2CF2-N=S C=CF-CF, 54 F CF, vl

PAGE 7

INTRODUCTION As a result of numerous investigations begun over a hundred years ago, many (1) sulfur-nitrogen halides are known. Michaelis (2) described the preparation of the first iminosulf\ir dihalide, phenyl iminosulfur dichloride, C^Hr-N = S Cl^ by the reaction of thionylanilide with phos6 5 <=phorous pentachloride. This substance is stable only in solution and cannot be isolated. The synthesis of a sulfur-nitrogen-fluorine compound having the empirical formula F;,NS was reported by Glemser (5). The compound was obtained as a main product during the reaction of gaseous mixture of SNF and SN2F2 with AgF2 at room temperature. This compound, F^NS, was assigned the structxire F-N = SFp on the basis of chemical evidence and was the first example of an iminosulfur difluoride reported in the literature. But later (4), on the basis of infrared and nuclear magnetic resonance data the assignment of the structure of F^^NS was changed to N=SF^. A new class of compounds, the organo iminosulfur difluorides was discovered in the investigations (5) of the chemistry of sulfur tetrafluoride, SF^. These sulfur fluoride derivatives have the general formula R-N=SF2 and

PAGE 8

are prepared "by the reaction of SF^ with, compoimds that have carbon nitrogen triple bonds. The organo iminosiilfur difluorides have a high thermal stability as indicated by the temperature required for their syntheses. A feature of the infrared spectrum of the iminosulfur difluoride structure -N=SF2 is an absorption band in the 7.15-7.55/^ (1^001350 cm" ) region which is associated with the N=S grouping. Reactions of the S-F bonds (5) of phenyl iminosulfur difluoride with sodium methoxide and with phenyl lithium using absolute methanol as a solvent resulted in two new derivatives, C^H^N=S(0CH,)2 and CgHcN=S(CgHc)2» having N=S linkages. The present work was started with the intention of studying the chemistry of perfluoroalkyl iminosulfur difluorides. Miller (6) found that F~ ion shows a strong nucleophilic reactivity toward the carbon-carbon double bond of fluoroolefins. The initial attack is on the carbon of double bond, which then results in either addition of F" to form a carbanion, CF,CF=CF2+F" — > CF^-CF~ CF^ or fluoride ion catalyzed rearrangement,

PAGE 9

F (f • F « ,^ N f f I . F-C=^C— C — G F— C — C=C — C— + F till i I I I F* F F F F F Andreades (7) conclusively established the formation of perfluoro carbanions by trapping the hepta fluoro-n-propyl and hepta fluoro-iso-propyl carbanions, generated from the corresponding monohydro fluorocarbons in ether solution as lithium salts, II CH,Li,CH^CH5-CH ^^3-^^^ ' \t^6j%^ CF^-CF-CH-CH^CH^ OF, OH CF,-CF + H"^ I CF3

PAGE 10

Tlie fluoro carbanions produced by the addition of fluoride ion to fluoroolefins may undergo secondary reactions depending upon tbe experimental conditions and the reactants present. The choice of reaction conditions suitable for studying the reactions of fluoride ion with fluoroolefins presents some difficulty. Stable inorganic fluorides tend to be insoluble in inert non-polar aprotic organic solvents, while in protogenic polar solvents, hydrogen ion transfer to fluoro carbanions limits the usefulness of such systems. Christie (8) observed that certain alkali metal fluorides act as catalyst in the oxidation of unsaturated fluorocarbons with oxygen. Tetrafluoro-2-O-fluorosulfate propionyl fluoride was synthesized (^5) by F~ ion catalyzed defluorosulfurylation of hexafluoro l,2-bis(o-fluorosulfate)propane: CF^CF(0S02F)CFg(0S02F) ^(^a^g ^ SO2F2 + CF,CF(0S02F)C0F Although the authors did not report any mechanism, it is believed that F~ ion from KF initiates the reaction: CF5— OF CF-^F 5 I ^0 CF^— CF CF + SOoF. o=s=o ^^1 ^ ^ ^ » I o=s=o y + F ^ / F 0=S=0 Dresdner (9) investigated the relative catalytic activity of cesium and sodium fluorides and found that whereas the

PAGE 11

reaction between C^F^ and NF^, proceeded at 520" when CsF was catalyst, 520** was required v;hen NaF was the catalyst. Van Artsdalen (10) confirmed that CsF is more active than all other alkali metal fluorides. In a survey of relative rates of exchange of fluorine atoms between alkali fluorides and fluorocarbons, conducted at moderate temperatures, the order of reactivity per square meter of surface was observed to be Cs^ Rb^ K = Na = Li in the exchange process o^Fg + m^^ > c^F^F-"-^ + m Miller (11) and Dresdner (12) found that reaction between cesium fluoride and fluorocarbon dienes yield fluorocarbon acetylenes at moderate temperature in the absence of a solvent. Dresdner (12) found that cesium fluoride in the solid state is a good fluoride ion catalyst for the reactions requiring the formation of a carbanion intermediate. The reaction of CF^N = CFp in presence of F" is a good example of , the difference between the behavior of solid catalyst and that of F" ion in a solvent. When CFjN = CF2 vapors were passed over solid CsF (12) in a flow reaction, (CF^)2N-CF = NCF^, was obtained and its formation was explained via the ionic mechanism :

PAGE 12

F CF3-N— CF2 > CF3-N~— CF^ CF3 — N=C-i-F > CF3-N=CF-N(CF3)2 + F~ CFo-N'-CF^ When CF3N = CFg was treated (I3) with COF2 in a polar solvent containing F Ion, there was no report of the dlmer CF3N = CF-N(CF3)2, having been formed, although the reaction between COF2 and CF3N=CF2 was very slow. Consequently there must be an Important solvent effect. The degree of catalytic activity of solid CsF is related to (a) its surface area, (b) defects in the solids, (c) temperature, and (d) pressure. Perf luoropropene can be dlmerlzed (12) to the els and trans Isomers of (CP3)2CF CP = CF CF3 (I) and (CF3)2C = CF CF2 CF3 (II), i.e., CF3CP = CF2 > CF3CF CF3 F" (from CsF) F CF3-CF = C -I F >CF3CF=CF-CF(CF3)2 + F" CF -CF--CF3 (I) Compound (I) can add another F~ to form a f luorocarbanion which after rearrangement loses F~ to give (II), viz.

PAGE 13

CF^— CF=CF— C(CPx)o > CF^CF^CF=C(CF,)o + F" F" In a flow reaction involving only CF^CF=CF2 and CsF at 330°Gj with, a contact time of 123 seconds, there was only 4 to 5 per cent conversion to products (I) and (II). But in a typical static reaction, in a 500 ml vessel containing 75 grams of dry CsF, 7^ per cent of the olefin was converted to the products (I) and (II) in a ^0-hour period. This clearly demonstrated that static reactions give better conversions. Statement of the Problem The main purpose of this research was to prepare new derivatives of perfluoroalkyl iminosulfur difluorides by their reaction with perfluoro carbanions generated from perfluoroolefins and dry powdered CsF. All reactions were to be carried out in closed vessels. No solvents were to be used. It was found that at high temperatures (250'*-500°) and relatively low pressures (20 to 50 atm. ) the perfluoro alkyl iminosulfur difluorides decompose to SF^ and the corresponding nitrile. As most of the reactions required heating to various temperatures (50''-300°), there was always a possibility that some R^ N = SF^ would decompose and the resulting SF^ might react with the carbanion at these

PAGE 14

8 moderate temperatures. Sulfur tetrafluoride is known (1^) not to attack a carbon-carbon double bond, for example, H2C=CHC00H + SF^ > H2C=CHCF, HC ^ CCOOH + SF^ > HC = CCF^ It F^C C-C-OCH, F^C C-CF^OCH, 2| n 5 ^ SF. > 2, n 2 3 F2C CH ^ F^C CH CIC C^^° CIC CF^ II > -S?^— * 11 Xq2 cxc— c =0 Q,!J _ ^ Thus one of the secondary objectives of tbis work was to determine if SF^ would react with olefins other than perfluoropropene (4-5), under the same conditions used in reactions with perfluoro alkyl iminosulfur difluorides. The sulfur atom in sulfur tetrafluoride and thionyl fluoride is formally isoelectronic with the sulfur atom in the iminosulfur difluorides, R-N = SF2 , F2 SF2 , = SF2 . It was expected that SF^ and SOF2 might behave in an analogous manner to -N = SF2# Loth and Michaelis (15) treated Cg^H^OCH, with SOCI2 in presence of ZnCl2 and obtained CH,0'CgH^S-— CgH^OCH,. When Smith et al . (5) used phenyl iminosulfur difluorides in place of thionyl fluoride, they obtained the same product. Similarly, reactions of SOCI2 (16) and C^HcN = SF2 (5) with o-CgH^(NH2)2 produced

PAGE 15

C^Hl^"^ ^ S. Thus, this behavior of SOCI2 and C^H5N=SP2 suggested that reactions of SOF2 and carbanions be studied. In part, the studies with SOP2 were a matter of practical expediency since the grade of SF/^ purchased for this work invariably contained from 10 to ^0 mole per cent SOF2 which could not always be completely separated from the SFi^. Thus it became necessary to evaluate the product from reactions involving SOF2 so that the products involving SFi|, could be uniquely identified. This work is composed of the experiments which involve the following reagents: 1. Effect of heat and CsF on RfN-SFg 2. CF3N=SF2 + metal fluorides 3. CF3N=SF2 + CF3CF=CI^ k, CF^N=SF2 + CFoCF=CFCF3 5. CP-jN=SF2 + CF3CSCCP3 6. C2F^N=SF2 + CF3CF=CP2 7. C2F^N=SF2 + CF3CF=CFCP3 8. C2F^N=SF2 + CF3C5CCF0 9. C3Fr;N=SF2 + CFoCF=CF2 10. C3F N=SF2 + CF3C5CCF3 11. SFj^ + CF3CF=CFCF3 12. SF4 + CP3CHCCF3 13. SOF2 + CF3CF-CFCF3 14. SOF2 + CF3C5CCF3

PAGE 16

EXPERIMENTAL • All temperatures in this dissertation are in degrees centigrade. Tlie chemical analyses were performed by the Schwarzkopf Microanalytical Laboratory, Woods ide, New York, unless otherwise stated. Apparatus Distillation of substances with boiling points below room temperature was carried on in a 55 cm long column with an evacuated silvered jacket. This coliunn was packed with 0.16 cm single turn, nickel helices and had about 50 theoretical plates and would have somewhat lower number (17) for fluorocarbons. The head was cooled with a dry-iceacetone mixture and was connected to vacuum system for gaseous take off. The temperatures were recorded using an iron-const ant an thermocouple. The vacuum line was made up of Pyrex glass and was designed for ball point connections. Kel-F No. 90 grease (Minnesota Mining and Manufacturing Co.) was used for all joints and connections. The pump could attain a vacuum of 0.1 mm of mercury, A Perkin-Elmer Model 157B double beam infracord spectrophotometer was employed for the determination of 10

PAGE 17

11 infrared spectra. A 5 cm long gas cell with sodium chloride windows was used for recording the infrared spectra of the gaseous samples, 19 The F ^ Nuclear Magnetic Resonance (n.m,r.) spectra of the compounds were obtained using a varian D.P. 60 spectrometer operating at 56.^ Mc. Trifluoroacetic acid was used as the external reference. All chemical shifts are reported in ppm. units. The mass spectra of the new compounds were obtained with a Bendix Model 1^-10? Time-Of-Flight spectrometer. The mass spectra were run for the sole purpose of finding the molecular weights. Molecular weights of the compounds, having a vapor pressure more than 10 mm at room temperature, were determined by the vapor density method, A weighed bulb (volume 209,^-34 ml) was filled with the vapors at known pressure and temperature and then weighed. Vapor Phase Chromatography (V.P,C,) of gases was accomplished with a 9 foot column packed with a mixture of silicone oil on chromosorb P. operating at room temperature using helium at 60 ml/min as a carrier gas. The apparatus was fitted with an adapter for gaseous in-take. The preparatory scale separation of gases v/as accomplished with an 8 foot glass column, 1/2 inch diameter and packed with a mixture of silicone oil on chromosorb P. The

PAGE 18

12 preparative scale separation of liquids was accomplislied in a vapor phase chromatograpli apparatus designed by Professor Paul Tarrant and coworkers. A two meter glass column with, an inside diameter of one inch, packed with silicone gum V-95 (Union Carbide) on chromosorb P kept at a constant temperature 20° below the boiling point of the fraction to be collected, was used. Nitrogen was the carrier gas. The refractive indices of liquids were measured with an Abbe's refractometer at the given temperatures. Materials Sulfur tetrafluoride (SF^) was purchased from E. I, Dupont De Nemours and Co, It was 85-90 per cent SF^, the impurities being SOF2 and S2CI2. It was difficult to obtain pure SF^ from this mixture. In most cases, SF^ was used as purchased unless otherwise stated. Thionyl fluoride (SOFp) was obtained by distillation of this technical sulfur tetrafluoride. Cesium fluoride (CsF) was purchased from the American Potash and Chemical Corporation, Los Angeles. It is a very hygroscopic material which melts at 703° (18) when pure. Perfluoropropene (CF,CF=CF2) was purchased from Peninsular Chem Re search. Inc., Gainesville. It has a boiling point of -29° (19) and its infrared spectriim resembled that reported (20).

PAGE 19

13 Perf luorobutene-2 (CFoCF=CFCFo) was purchased from the Matheson Company, Murrow, Georgia. It was a mixture of cis and trans isomers in the ratio 1:3 and had a boiling point of +1 . Its infrared spectrum resembled that reported (21). Preparation of hexaf luorobut.Yne-2 . — Hexaf luorobutyne-2 (CFoC=CCFo) was prepared (22, 26) by the dehalogenation of 2 , 3 dichlorohexaf luorobutene-2. Powdered zinc (I30 g., 2.00 mole) treated with dilute (O.IN) hydrochloric acid was washed with water and dried with 2-propanol. The activated zinc was placed in a 2-liter, 3-necked flask and was covered with 50O ml of dry Isopropanol. The olefin CF3CC1=CC1CF3 (233 g., 1.00 mole) was added dropwise and refluxed with continuous stirring. The resulting product (105 g.) was collected in a trap cooled in a dry ice-acetone mixture. The crude product on distillation gave (88 g., 0.54 mole) of CF3CSCCF-J; b. p. -24° (19, 2.6). Sixty-two grams (0.25 mole) of the starting olefin was recovered. The yield of the hexaf luorobutyne-2 was 72 per cent based on unrecovered dichlo'ro hexaf luorobutene-2 . The Infrared spectrum of Ci^F^ resembled that described by Haszeldine (23). Preparation of perf luoroalkyl iminosulfur difluorldes, Rf.N=S?2'--The perf luoroalkyl iminosulfur difluorides were prepared by the method described by Smith et al . (5). A nitrile or thiocyanate and excess SF^,, were

PAGE 20

1^ loaded in a 300 ml stainless steel autoclave. The autoclave was heated to the recommended temperature. After completion of the reaction, the volatile contents of each reaction were distilled to recover the products and unused SF^, The iminosulfur difluorides were dried over P/iO-iq, degassed and stored, Trifluoromethyl iminosulfur difluoride, CF^N=SF2. — Forty-two grams (0,51 mole) of NaSCN was loaded into the autoclave, dried and degassed at 110** for six hours. Then SF^ (188 g. , 1,75 mole) was condensed into the vessel. The vessel and contents were heated at 200° for two hours, at 250** for four hours, at 500° for four hours and at 350° for two hours, a total of 12 hours, A maximum pressure of 200 atm, was noted. When the vessel was slowly cooled to room temperature no large drop in pressure was observed. Some elemental sulfur had condensed in the pressure gauge inlet line. The volatile contents of the vessel were transferred to a vacuum system. They amounted to 169 grams and were fractionated. The unused SF^ was stripped off, a major fraction was recovered between -9° and -6°, It amounted to i\-i^ grams and based on the starting amount of NaSCN, this represented a yield of 57 per cent. The infrared spectrum shows these absorptions in cm" , 1500 (w), 1385 (v.s.), doublet at 1205 and 1175 (v,v,s.), 817 (m), 760 (s).

PAGE 21

15 P-'-^ N. M. R. data: Peak a b Shift -126.5 -28.46 Area 2 3 Group N=SF2 CF^-N Preparation of pentaf luoroethyl iminosulfur dl. fluoride, C2FtN=SF2. — A 3OO ml steel autoolave was loaded at liquid nitrogen temperature with (76 g., O.8O mole) of CF^CN (b. p. -60°) and (108 g. , 1.00 mole) of SF^. The autoclave was heated for Z^ hours at 350°. The volatile products were distilled and (I06 g., O.52 mole) of C2F^N=SF2 was obtained at 23°. This represented a yield of 65 per cent based on the CFoCN. The following infrared absorptions (in cm" ) were observed for C2F^N=SF2. 1^10 (v. s.), 1350 (v. V. s.), 1240 (v. V. s.), 1165 (v. V. s.), 1050 (v. V. s.), 785 (s.), 758 (v. s.), 715 (v. s.). Preparation of heptaf luoropropyl iminosulfur difluoride, C2Fr,N=SF2. — This material was prepared by reacting C2F^CN (116 g., 0.80 mole) and SF^ (I08 g., 1.00 mole) in an autoclave. The vessel was heated at 200° for two hours and at 250° for four hours. The cooled vessel was attached to the vacuum system and the more volatile components, mainly unreacted starting materials, were stripped off. The desired product boiled at 51°. Its infrared spectrum showed

PAGE 22

16 tlie following absorptions! 1390 (v.s.), 13^6 (v.s.), 1287 (s), 1250 (v.v.s.), 1202 (s), 11^4 (s), 1094 (s), 979 (s), 9^0 (s), 775 (s), 750 (s) and 712 (s). General Procedure All the catalytic reactions were carried out in a reactor whicti was a round-bottomed stainless steel (Hoke) cylinder of 500 ml capacity, I3 inches tall, and 2.5 inches external diameter. Powdered cesium fluoride (100 g. ) was placed in the cylinder which was then fitted with a 15 inch long portion of a 1/4 inch pipe, a pressure gauge and an outlet valve. The reactor was heated to 350*'-400'' for two days under vacuum to remove any residual moisture from the CsF, It was absolutely essential to remove the moisture completely otherwise traces of blue products (9) are formed in the reactions. The upright furnace for heating the reactor was an 18 inch tall, 3-iiich diameter iron pipe which accommodated the reactor snugly. The iron pipe was heated and the temperature was controlled by varying the voltage of the heater. The temperature inside the furnace was measured using an iron-constantan thermocouple. All volatile materials were stored in weighed glass ampoules and kept cold in a dry ice-acetone mixture. These

PAGE 23

17 ampoules could be connected to a vacuum system and an appropriate amount could "be removed by condensation into a weighed graduated ampoule. The gases were then e2cpanded into a large bulb and an infrared spectrum and a vapor phase chromatogram of a representative sample recorded. The gases were then transferred into the reactor containing CsF at -196*'. The reactor was heated by placing it in the fiirnace. The temperature inside the furnace and reactor pressure were recorded periodically and the results plotted. Normally if there is no condensation or decomposition, as the temperature is raised, the pressure-temperature relation is linear. Any sharp break in the curve is indicative of a reaction. A rise in pressure indicates decomposition whereas the fall in pressure indicates the formation of less volatile products. A reaction was presumed to be complete when the rate of change of pressure became negligible. After the reaction, the reactor was connected to the vacuum system through a weighed trap which was cooled in liquid nitrogen. The products were vented and any gas which did not condense was separated for identification. The products were allowed to warm. The volatile compounds were expanded into a large flask and the liquids were left in the trap. The liquid and gaseous fractions were analyzed separately.

PAGE 24

18 Attempted reaction of G7-,'N=S'F^ with CsF. — Trifluoro? 2 methyl iminosulfur difluoride (W g., 0.26 mole) was loaded in the 500 ml reactor containing dry cesium fluoride. The temperature was raised to 300° and maintained at 300" for another three days. The recovered material (3^.2 g.) was expanded into a 22 1. bulb. An infrared spectrum and V.P.C. analysis of a representative sample did not show any new product. There was an apparent loss of 5.8 grams of starting material. When the reactor was heated above 200° under vacuum, 2 to 3 grsons of very reactive volatile material was collected which was not identified. Reaction of CF^N=SF2 with metal fluorides. — The fluorination of CF2N=SF2 was attempted using HgFp, AgF^, and CoF, at temperatures between 100° to 200° in an autoclave. In general, the metal fluoride was loaded into a 300 ml autoclave (except in a final reaction) and after evacuation, the GF,1T=SF2 was added. Table 1 summarizes the conditions and results of several reactions. Reaction of CF,N=SF2 and CF^GF=CF2 over CsF. — Trifluoromethyl iminosulfur difluoride (^1 g., 0,27 mole) and perfluoropropene (96,0 g., 0,6^ mole) were loaded into the reactor. At room temperature, the pressure in the vessel was 6,2 atm. The system was heated to a reaction temperature of 83° over a five hour period and maintained at this

PAGE 25

19 Table 1 Reaction of CP-2N=SF2 with Metal Fluorides Metal Fluoride

PAGE 26

20 temperature for 115 hours. During this Interval, the pressure dropped from 17 atm. to I3 atm. and heating was discontinued. The final pressure at room temperature was U.l atm. The volatile contents of the reactor were transferred to a vacuum system. This material amounted to I36 grams, of which 53 grams (O.35 mole) was CPoCF=CF2. From the fraction boiling above 25°, which amounted to 70 grams, only two products were isolated. The major product amounted to ^3.5 grams and had the following physical properties. Boiling point 65.5°. The infrared absorption spectrum (Fig. 1) showed a strong peak at 1298 cm~ which is associated with -N=S stretching. f19 n.

PAGE 27

21 o

PAGE 28

22 :

PAGE 29

23 0-| K oCX5 O o o o u O -1 o L -I lO 00 CN Os O z CX)IU LU > < •O o I ^<^ o_o n CVJ CM O o o u +3 o p< CO o •H •P P< fH o w o U «H o CM \ •O

PAGE 30

2^ o

PAGE 31

25 Nolecular weight: Calculated for C2^F-j^-j_NS, 303. Found (vapor density method), 306, 305* This product was assigned the structure CF_N = S CF — CF^ F CF3 The yield of C/^Fq^]_NS was 52 per cent, based on starting difluoride. It was observed that the new material became slightly discolored on standing in glass over a period of two weeks. Upon gas transfer, it became water white and had the same infrared spectrum as in Figure 1. The minor product (3.0 g.) had the following physical properties: Boiling point 75.5°. The infrared absorption spectrum (Fig. 4) showed a strong peak at 178O cm~^ which v/as associated with ring C = N stretching mode. f19 n.

PAGE 32

26 Analysis: Calculated for C„F NS, C, 20.24; N, 3.37' Found: C, 20.39; N, 3.34. riolecular weight: Calculated for C F KS 415. Found (Mass spectruin), 415. This product "was assigned the structure (CF^)„CF-C -SF-CF (CF-.)^. 3 2 j,^ 3 2 ir The reaction of C2F^N=SF2 and CF„CF=CF2 <^^^^ CsF.— Thlrty-slx grains (0.18 mole) of C2F^N=SF2 and 51 grams (0.34 mole) of CF^CF=CF2 were condensed In the reactor. The vessel was heated In an upright furnace to a maximum temperature of 89°. The heating was discontinued when pressure remained constant at 89°. The volatile contents were transferred to a vacuum system and the distillation gave the follov;lng mass distribution: CF^CP=CF2 300 g. A pure product 52.5 g« Residue 2.2 g. Apparent loss 2.0 g. The product (52.5 g) has the following physical properties : Boiling point 82.3°. Analysis: Calculated for C^^^F^ NS, F, 70.0; C, I7.I; N, 4.0; S, 9.1. Found: F, 69.6; C, 17.O; N, 3.8; S, 9.2. Tlolecular weight: Calculated for C^F-[_^NS, 353. Found (vapor density method), 352, 354.

PAGE 33

c

PAGE 34

28 At 250° a sliaj-p increase in tlie pressure was observed and heating was discontinued. The volatile contents of the vessel were removed and were found to consist mainly of Cp^cCN, SF^ and C;,F„N=SF2. The amounts of nitrile and sulfxir tetrafluoride recovered were equivalent. The reaction of C-,F„N=SFo with CF^CP=CFo over CsF. — 2_Z 2 ^ 2 The reactor, containing 100 g. of dry CsP v/as loaded v;ith 27 grams (0.11 mole) of C,FcN=SFp and 59 grams (0.40 mole) of CF,CF=CF2. The pressure in the vessel at 27* was 6.1 atm. The temperature was raised to 82* and a maximum pressure of 13.9 atm. was observed. The temperature was maintained between 75" to 85* for another 88 hours, v/hen the observed pressure v^as 10.7 atm. at 77.5°. On cooling the reaction vessel, the pressure was 5.5 atm. at 27''. The products were vented and distilled. The following fractions were obtained: CF2CF2CF^ + GF,CF=CF2 ^5 g. new products 58.7 g. No C,FoN=SFo was recovered. The products v/ere separated by of '^ V.P.G. and the tv;o new compounds were isolated and identified. The major product amounted to 27 grams and had the following physical properties:. Boiling point IO5.5*. Analysis: Calculated for G^F^^NS, F, 70.81; G, 17.86; N, 3.^7; S, 7.94. Found: F, 70.40; G, 17.90; N, 3-^0; s, 8.31.

PAGE 35

b

PAGE 36

30 Molecular weight: Calculated for C^F3_oNS, 365. Found (by vapor density method) 368, 370. Infrared absorption spectrum (Fig. 5) showed a strong absorption at I707 cm"-^ which was associated with ring C = N stretching. F^9 N. M. R. data: Peak a b c d e Shift -45.1 -i^.4 +7.7 42.7 87 Area 1 6 3 2 1 Group SP ^^^3^2 ^^3 ^^2 ^F Based on the above data, the minor product of the reaction between CoFr,N=SF2 and CPoCF=CF2 over CsF was assigned the structure CP-,-CPp-C — S — CF — CP^ ^ ^ ll/l I 3 N F CF3 The yield of C^F-|_^NS was 16 per cent, based on C^FoN=SF2. The attempted reaction of CFoCF=CFCF3 with CsF. — When Young (24) passed a mixture of cis and trans perfluorobutene-2 over CsF heated to 350*^ in a flow system he did not report any dlmerization or isomerization. A mixture of about 30 per cent cis and 70 per cent trans perf luorobutene-2 (80.0 g., 0.40 mole) was loaded into the reaction vessel containing dry CsF. The temperature was

PAGE 37

31 -

PAGE 38

32 raised to 355° in a period of two days and no irregular pressure changes v/ere observed. The temperature was maintained at 555° for another two days. After expanding the cooled products into a 22 liter bulb, the infrared spectrum and V.P.C. analyses of representative samples indicated that no chemical reaction had occurred. But V.P.C. analysis did indicate a slight increase in the trans isomer. Attempted reaction of CF,N=SF2 and CF;,CF=CFCF;, (without CsF) . — Perfluorobutene-2 (11.0 g, 0.52 mole) and trifluoromethyl iminosulfur difluoride ('^-.O g., 0.026 mole) were condensed in a 100 ml stainless steel autoclave and heated to 550'' for 2^ hours. Almost all, about 98 per cent by weight, of the starting material was recovered unchanged, with a small amount of black crystalline material. Thus, there was no appreciable reaction. The reaction of CF^N=SFo xvith CF-jGF=CFCF2 over CsF. — 5 ^ 2 2 Trifluoromethyl iminosulfur difluoride (26,0 g^ 0,17 mole) and perfluorobutene-2 (5'^.0 g., 0,17 mole) were condensed in the reaction vessel containing dry CsF. The temperature was raised slowly in intervals to allow for thermal equilibrium. No pressure drop was noticed at 225° for 12 hours and the temperature was raised, A small pressure drop was observed at 256**. The temperature was raised to between 280° to 290° to speed up the reaction. After heating for 96 hours, the

PAGE 39

53 pressure remained almost constant and heating was discontinued. As only 50 grams of volatile products were collected, the reactor was heated to ^00° under vacuum. A mixture containing 5.0 g. of a liquid and solid sulfur was collected in a liquid nitrogen cooled trap. The liquid was very volatile and very reactive toward glass. ITo attempt was made to identify it. The volatile products (50 g.) were expanded in 22 liter bul"b and V.P.C. analysis and infrared spectra indicated the following fractions: CF^N = SF2 5.9 g. . CF^CF = CFCF^ 17.0 g. new product 25.0 g. two unidentified products 3.0 g, The new product had the following physical properties: Boiling point the vapor pressure at different temperatures was determined as 2^'' ^59 mm -ll^ 8? mm 11° 267 mm -78° mm 0° 16^ mm From the plot of log P vs ~ the boiling point of the product was found to he 36.6° at 760 mm pressure. Analysis: Calculated for Cc-F,,N: F, 75.85; C, 21.20, ^, 4.95. Found: F, 7^.00; C, 20.84; N, 5.00.

PAGE 40

3^ Molecular weight: Calculated for C^F-l]_N 283. Found (vapor density Tnethod), 277.5, 278.8. (Mass spectrum), 283. Infrared spectrum (Fig. 6) showed a strong peak at 1725 cm~^ which was associated with C = N stretching. f19 n. M. R. data: Peak

PAGE 41

35 o o

PAGE 42

36 temperature further to nearly 300°, a rapid rise In pressure was noticed. The temperature was lowered to 293° and a maximum pressure of 29.1 atm . was observed. The pressure started dropping and within the next 72 hours it dropped to 24.5 atm. at 290°. The products (85 g.) were removed. The gases and the liquids were analyzed and the following fractions identified; CF-^CF = CF CF3 28 g. CF3CN 13 g. CF3CF2 CF2 CF3 10 g. Liquid product 32.2 g. Unidentified 2.8 g. Apparent loss 10.0 g. No C2F^N = SF2 was recovered. The following physical data on the product was obtained. Boiling point 135°. Refractive index at 20° 1.3554. Analysis: Calculated for Ci|FoS, : F, 62.44; C, 22.54; S, 15.02. Found: F, 62.14; C, 22.91; S, 14.97. Molecular weight: Calculated for (Cji^Fr,S;j_)2 426. Found (by mass spectrum), 426. Molecular formula CsF-,kS2. Infrared absorption spectrum (Fig. 7) did not show any peak in the C = C or C = N region. F ^, N. M. R. data: Peak a be d Shift -19.9 -4.7 +2.0 32.0 Area 6 3 3 2 Group ^^^^^2 CFo CF3 CFg

PAGE 43

37 lO CO CN O U") OJ CQ •dH CO o «M O & P O 0) Pi CO CO

PAGE 44

38 The product of the reaction between Cg F^N = SF2 and CF^ CF = CF CF3 over CsF at 293° was assigned the structure CFr, C = CCF^ ^ I I 3 S' s CF^ -CFg-^C CF^ which is consistent with the physical data. The reaction of a mixture of SF. and SOF2 with CF CF = CF CP^ (without CsF).--A rrlxture (25 g.) consisting of sulfur tetrafluoride and thlonyl fluoride whose composition was indefinite and perf luorobutene-2 (30 g., 0.15 mole) were condensed in a 50O ml. steel vessel. The pressure at room temperature was 7 atm. The reaction vessel was heated to a temperature of 259° and no abnormal rise or fall In pressure was observed. The P versus T relation was a straight line. After venting the gases, the Infrared spectrum and V. P. C. analyses of representative samples Indicated that no reaction had occurred. The reaction of SOF2 with CF-3 CF = CF CFr^ over CsF. -Thlonyl fluoride (41.0 g., O.535 mole) and perfluorobutene -2 (47.0 g., O.23 mole) was condensed in the reactor containing dry CsF. The temperature was raised to 202° over a period of 16 hours when a maximum pressure of 28.0 atm. was observed and then started dropping. The temperature was held between 200 and 205°. The pressure dropped to I3.2 atm. in a period of 120 hours. The following fractions were recovered:

PAGE 45

39 Volatile products 47 g. New product (liquid) 25.2 g. Unidentified liquid 2.8 g. Apparent loss 18 g. The volatile gases contained perf luorobutene-2 , thionyl fluoride and one unidentified substance which was very reactive to glass, Kel-P grease, the NaCl windows of Infrared cell, mercury and the packing in the V. P. C. @9lyf?in. On Slanging in a Z1 1U§f bulb, the rgaotive ga§ was converted to SOF2 and S1F/|. Although the material was never Isolated its chemical behavior suggested that it was probably SPoOF. The exact composition of the volatile gases could not be determined. No perf luorobutane was obtained. The liquid product was 95 psr cent pure and was further purified by V. P. C. The purified product had the following physical properties: Boiling point 149°. Refractive index at 20° I.33I8. Analysis: Calculated for C^ Fq S 0: F, 64.05; C, 18.00; S, 42.00. Found: F, 64.4?; C, 18.29; S, 12. 23. Molecular weight: Calculated for (C^^F^ 30)2 534. Found (by mass spectrum), 534. Molecular formula Cg F^g S2 O2' Infrared absorption spectrum (Fig. 8) showed a strong doublet at 1351 Cm" and I333 Cm"-*which Is associated with one oxygen bonded to sulfur as in SOFg.

PAGE 46

40 OJ o C\J CO 00 iH 00 o o +3 o
PAGE 47

41 F^9, N. M. R. data: Peak

PAGE 48

42 fell to 15.3 atm. in a period of 12 hours. The products were removed from the reactor and the following fractions were collected. Volatile gases 64 g. Liquid product 29 g. Apparent loss 4 g. The volatile fraction was expanded into a 221. bulb. There was evidence that the glass walls and Kel -F grease were being attacked. When it was passed through the V. P. C. column only perf luorobutane, perf luorobutene-2 and some SOF2 appeared. The failure of the reactive component (s) to exit from any V. P. C. column (Silicone oil, Kel -F ester, D. N. P.) made an analysis difficult. To find the amounts of each component, comparison Infrared spectra of known samples were used and compared to the correspondixig spectra of the mixture of volatile materials. The quantity of each in the volatile fraction were crudely estimated to be Reactive gas + SOF2 18 g. CFj CF = CF CF3 28 g. CFo CF2 CF2 CF3 18 g. The V. P. C. chromatogram of the crude liquid product had a major peak with a shoulder which did not resolve on any of V. P. C. columns used; silicone oil, silicone gum, Kel -F ester and D. N. P., even at temperature lower than 20° below the boiling point and at lower helium flow rates. When the material represented by this peak in the chromato-

PAGE 49

43 gram was separated from the mixture by preparative scale V. P. C, Its boiling point was determined to be 1^2°. From its Infrared spectrum (Fig. 9) It was presumed to be a mixture of C — CF, C I CF, I 3 CFI GFo CF, CF, II I ' S— 0— S — CF I CF, I CF CF, CF, (Infrared spectrum Fig. ?.) (Infrared spectrum (Fig. 8.) To be sure that this product was a mixture of above two components, the two compounds were mixed in the ratio of 1:3 and an exactly superlmposable infrared spectrum was obtained. The boiling point of this synthesized mixture was 1^3°. The retention time and position of the shoulder of its V. P. C. chromatogram were exactly the same as that for the separated product. Attempted reaction of CF3N = SFg with CF^C = C CF^ (without CsF) .-Trlfluoromethvl iminosulfur difluorlde (25. g., 0.16 mole) and hexaf luorobutyne-2 (15«0 g., 0.09 mole) were condensed in a 500 ml. Hoke cylinder fitted with a stem, pressure gauge and an outlet. The reaction vessel was heated to 100°. for 6 hours. The Infrared spectrum and V. P. C. chrom.atogram of the reactor contents Indicated that no reaction had occurred. The reaction of CFoN = SPg with CF^C = C CF3 over CsF. Trial A. — Trlfluoromethvl iminosulfur difluorlde

PAGE 50

44 rvj o (\j w 00 rH 00 o + CVJ CO CO o
PAGE 51

^5 (20.0 g., 0.13 mole) and hexaf luorobutyne-2 (I5.6 g., 0.10 mole) were condensed In a graduated glass ampoule and allowed to melt. A slightly milky mixture but no separate layers were formed. This mixture of CF^N = SFg and CFo-CSC-CFo was loaded into the reaction vessel containing CsF. The pressure in the reactor at 23° was 2.6 atm. A prior calibration of the furnace had shown that a 10 volt potential on the heater coil resulted in an equilibrium fe§fnp§?atur§ of 50 = 5^» In thlg oa§e the equilibrium temperature was 72° indicating an exothermic reaction, so heating vjas discontinued. Even then the temperature continued increasing and the pressure was observed to fall. After 3 hours at 80 the observed pressure was 3*33 atm. and then the temperature started to drop. Electrical heating was resumed and the temperature held at 80°. After another 3 hours the pressure was 1 atm. The reactor was bled and products condensed but the mass of recovered material was only a fraction of the mass of reactants. The reactor was heated to 150° with pumping and a less volatile liquid collected. To remove the last traces of material, the reactor was heated to 350° with pumping when the decomposition products of absorbed species came out. Following fractions were recovered: CF^ N = SF2 9 g. CF^ C = C CF3 3 g. Liquid product 20. 3 g. Residue 3-4 g.

PAGE 52

46 The V. P. C. chromatogram of the liquid product indicated it to be 90 per cent for one component and two more components each approximately 5 per cent. The major component was separated by preparative V. P. C. and had the following properties: Boiling Point 16?°. Refractive index at 20° I.33I5. Analysis: Calculated for C^F-^^^ NS: F, 66.35; C, 19.5; N, 4.44; S, 10. I5. Found: F, 67.2?; C, 19-45; N, 4.58; S, 10.10. Molecular weight: Calculated for C. F-^-^ NS 3I5. Found (by mass spectrum), 63O. Molecular formula C-^q F22 N2S2. Infrared absorption spectrum (Fig. 10) did not show any peak in the C = C or C = N region. F^^, N. M. R. data: Peak a b c d e f g Shift -16.6 -32 -2.4 +3.4 39.5 ^0.6 84.6 Area 36 7 4 2 Group CF^-N 2xCF^ 2XCF3+CF 2xCP2 2xCF The above data was consistent with the proposed structure CPo CF„ CF, CF S N S CF I It II I CFo CF N CF, I I ^ CFCP^ The yield of the product C-j^q F22 N2S2 from the reaction of CFoN = SF2 with CF^C £ C CF3 over CsF at 80° was 74 per cent based on hexaf luorobutyne-2.

PAGE 53

47 Trial 3. — In order to prepare the monomer CFoN = S C = CP CP^, CF^N = SFo (25 g., 0.l6 mole) and 3 I I 3 3 F CF3 hexaf luorobutyne-2 (I5.O g. , 0.09 mole) were condensed In the reactor vessel containing Cs? and maintained the temperature at 56*^ for 4 hours. The products indicated 85 per cent of C-j^Q F22 Ng Sg and 8 per cent of another compound, which after separation by V. P. C, showed a strong infrared absorption band at 1748 cm"-^ (Fig. 11). The total quantity of a second product that was separated amounted to only 2-3 drops and thus further Identification was abandoned. A third produce (7^0) showed a weak Infrared absorption, at I7I8 cm~^. Trial C. — In another attempt to prepare the expected monomer, CF3N = 3 C = CF CF-j , from CF^N = SF2 (I6.O g., F CF^ 0.11 mole) and CF^CSC CF3 (7.5 g., 0.047 mole), the substances were condensed into the reactor and allowed to stand at room temperature for 38 hours. In the interval the pressure in the vessel dropped from 3.27 atra. to 0.8 atm. Nine grams of a crude liquid product which consisted of 81 mole percent C-^q F22 ^2 ^2 ^"^ -^^ mole percent of the third product, in trial B, were prepared. Apparently no monomer was formed. The reaction of C2 F5 N = SF2 and CF^ C = CCFo over CsF. — Perf luorobutyne-2 (I7.2 g., 0.11 mole) and pentafluoroethyl imlnosulfur difluoride (28. g. , 0.14 mole) wiere condensed in the reactor containing CsF. The temperature was raised slowly and at 124° a pressure of 8.67 atm. was ob-

PAGE 54

48 CO OJ OJ OJ o o o -p o o p< w o -p o u a hO ^

PAGE 55

49 o I o o-o CO— Pq II o O o (i> P< w o •H o CO (i> IH nS (H «H Pi M I

PAGE 56

50 served before the pressure started dropping. The temperature was maintained between 12^° and 127° for a total of 52 hours and the pressure became constant at 3.2 atm. The products were removed and the following fractions were recovered: CP C = C CF^ 2 g. C2F3 N = SF2 7 g. New liquid product 33.7 g. Unidentified 0.5 g. Thr product was almost pure and needed no further purification. It had the following physical properties: Boiling point 98°. Refractive index at 20° I.3158. Analysis: Calculated for C^F-j_^ NS: F, 67.67; C, 19.73; N, J.Qk; S, 8.77. Found (I): F, 63.9; C, 20-59; N, 5.07; S, 10.48. Found (II): F, 66.9; C, 19-9; N, -; S, -. Molecular weight: Calculated for C^ F^^, NS 365. Found (by vapor density method), 367.4, 368. Infrared absorptions spectrum (Fig. 12) had a strong peak at 1724 cm" which was associated with C = C stretching mode. 19 F-, N. M. R. spectral data: Peak a b c d e f Shift -40.06 -3.92 -3.15 +3.59 41.79 87.5 Area 13 3 3 2 1 Group SF CF^ CF^ CF^ CFg CP

PAGE 57

51 O KN O-O UJ S CO— pq It CVJ o I o «H o P o 0) p< w d o •H -P ft U o H f CM •H

PAGE 58

52 The above data were consistent with the proposed structure CF^CFp N = S— C = CF CF^ 3 2 , J 3 F CF^ The yield of the product C-^^ F-^j NS from the reaction of C2 F^ N = SFg with CF-j C = C CF3 over CsF at 124° was 85.5 per cent based on the perf luorobutyne-2. The reaction of C^ F^ N = SF2 with CFo C = C CF^ over CsF.— H^ptafluoropropvl imlnosulfur difluorlde (32. g., 0.125 mole) and perf luorobutyne-2 (32.8 g., 0.2 mole) were condensed in a 500 ml. reaction vessel containing dry CsF, It was slowly heated and a maximum pressure of 7.8 atm. was recorded at 142°. The temperature was held between 150° to 155° and the pressure continued dropping. After a total of 72 hours, the pressure dropped to 6,2 atm. at 149°, when the products were taken out. Analysis indicated the following fractions: CF^ C = C CF^ 15.5 g. New liquid product 46,0 g. Unidentified 3.5 g. No Co Fr, N = SF2 was recovered. The V. P. C. separated product had the following properties: Boiling point 111°. Refractive index at 20° I.3142. Analysis: Calculated for C„ F^^ NS: C, 20.24; S, 7.71. Found: C, 20.68; S, 8.00. Molecular weight: Calculated for Cr, F-j^^ NS 415. Found (by vapor density method), 420, 423.

PAGE 59

53 Infrared absorption spectrum (Fig. I3 ) shovjed a strong peak at 1715 cra"-^ v/hich v/as associated with C = C stretching mode. F^^ , N. M. R. data: Peak abed e f g Shift -il-5-16 -3.1^ +3. 81 7.12 ill. 58 42.16 87.27 Area 1 3 3 3 2 2 1 Group SF CP^N CF^ CF CF2-N CF^ CP On the basis of the above data, the product of the reaction betv^een C^ P„ N = SF^ and CF^ C^C CF^ over CsF at 150° was £issigned the structure CF^ CFo CPp N = S-C = CF CF^ . J '^ ^ II J F CFo The yield of the product C„ F NS vjas 88 per cent based on Cr^ F N = SFp. The reaction of SOF2 a nd SFj>^ with CF^C ^ C Q.F^ over CsF. — A mixture (14 g.) of thionyl fluoride and sulfur tetraf luoride and perf luorobutyne-2 (10.8 g., O.O67 mole) was condensed in the reaction vessel containing CsF and allowed to stand at room temperature. A very small amount of reaction v;as observed to occur at room temperature, as indicated by drop of about 0.01 atm. over a 4 hour period. When the reaction vessel was heated a maximum pressure of 8.67 atm. was observed at 90°. The temperature was held at 80° for next 24 hours and the pressure dropped to 2.99 atm. The follovjing fractior.s were recovered:

PAGE 60

54 o

PAGE 61

55 Colorless gas 10 g. Colorless liquid 9*5 S« Apparent loss 6 g. In order to recover the material (6 g.) which remained in the vessel, it was heated to 200° under vacuum and an intensely blue-colored material (4.7 g*) condensed in the liquid nitrogen cooled trap. The colorless liquid product had a boiling range of 147°-148° and the V. P. C. analysis indicated that 95 per cent of the material was a mixture of two components which were later shown to be CoF-,q So 0^ and Co F-,k S2. The blue-colored volatile liquid which was stored at -78° was blood red in color after 24 hours. When the red-colored product was v;armed , the red color started disappearing, and as a colorless gas, it reacted with the Kel -F grease and glass. It also attacked the NaCl windows of the Infrared cell and the V. P. C. column packing. When this expanded gas was recondensed at -195° the solid was colorless. The infrared absorption spectrum of a representative sample indicated a strong doublet at 1260 cm~ and 1241 cm~-^ along with the characteristic frequencies of SOF2, SF^. The infrared and V. P. C. analyses of the colorless volatile product indicated that it contained O.025 moles of perf luorobutyne-2 and smaller amount of perf luorobutane.

PAGE 62

56 05 f-, 4-> o •H 03 »-> (U CO f-> t-i < -P u cr; 3 cr o 4-> 3 3 3 IH O O -P fciO'd -O O 5 •» ~ ;l, CO -P 4J +i 0)
PAGE 63

57 +> Oi a* JU 4J +J -P P rO fd
PAGE 64

58

PAGE 65

DISCUSSION The most fundamental basis for mechanistic speculation is the identification of the reaction products. Vapor phase chromatography is very valuable for separation of materials. Infrared and nuclear magnetic resonance spectroscopy are among the more important techniques that the chemist uses to gain information about a particular substance. But, then, there are instances when a highly reactive and corrosive mixture defies all the "routine" analytical tools. During the progress of this work, two such mixtures (A) and (3) were encountered which could not be separated and identified. "Thermal" Stability of Perf luoroalkyl Iminosulfur Difluorlde The perf luoroalkyl Iminosulfur difluorldes were synthesized (5) from nitriles and sulfur tetraf luorlde or sodium thiocynate and sulfur tetraf luorlde at temperatures ranging from 200° to 350° and pressures ranging from 120 to 200 atm. Obviously these sulfur tetraf luorlde derivatives are stable up to these temperatures under these pressures. However, most of the reactions in this v;ork were expected to Involve a maximum pressure of only 35 to 40 atm. Thus it was Important to determine the stability of R^ N=SF2 under the proposed experimental conditions. 59

PAGE 66

60 Trlfluoromethyl Iminosulfur difluorlde, CF^ N=SP2, and perf luorobutene-2, CFoCF=CF CPo, were heated In a steel vessel at 350° for 24 hours without any decomposition. When CF3N=SF2 was heated over CsF at 300° (maxlTnum pressure observed was 15 atm.) for a period of three days, no reaction was observed. This clearly demonstrated that CF3N=SF2 is stable over CsF up to 300° in a closed steel vessel. Nearly 15 per cent of the starting material could not be purrped out of the reaction vessel at room temperature. It remained in the vessel presumably as a complex with CsF, i.e. , CsP + CF3N = SF2 >Cs^ (CF^N = SF3"). This ionic solid complex decomposes on heating above 200° under vacuum. When pentafluoroethyl Iminosulfur difluorlde, C2FcN = SF2, was heated with CF^CF = CF CFo over CsF, no reaction was observed up to 282°. The maximum pressure at this temperature was 24.2 atm. Thus it is safe to conclude that C2 F^ N = SP2 is stable over CsF up to 282°. 3ut when the temperature was raised to 300°, a rapid rise in pressure was observed. This abnormal rise in pressure is indicative of some decomposition. When the products were analyzed, a quantitative yield of CPoCsN vias obtained, whereas no C2F5N = SF2 or SF^,. was recovered. The liquid recovered from this reaction vjas later found to be a product of the reaction involving SF^ and CFoCP = CF CF^ . Thus C2F^N = SF2 apparently decomposes to SF/^ and CF-CN, viz.

PAGE 67

61 C2F^N = S?2 above 282° CF CN + SP^^. Heptaf luoropropyl Imlnosulfur difluoride is even less stable than the ethyl derivative. When C^FyN = SF2 was heated over CsF, no decomposition was observed at 200°. At 250° a sudden rise in pressure occurred and hence decomposition. The resulting products were an equimolar mixture of C2F^CN and SF^, viz, C^F^N = SF2 ^^Q° ) C2F^CN + SP^ The decomposition data indicates that the pyrolytio stability of the fluorocarbon iminosulfur difluorides decreases with increasing chain length. For example, CF-N = SF2 is stable up to 300° over CsF; C2P15N = SF2 decomposes above 282° over CsF; and C^Fr;N = SF2 decomposes at 250° over CsF. Fluorination of R^N = SF2 with High Valency Metal Fluorides Attempts to form an N-F bond in the R^N = SF2 compounds with high valency metal fluorides tended to be frustrated by (a) rupture of the N = S bond and (b) reversible decomposition of difluorides to nitrile and SFi^, which in turn frequently underwent fluorination. The metal fluorides used were H Fg, A Fg and CoF^. The iminosulfur difluoride mainly used in these studies was CFoN = SF2 because (a) it is thermally most stable, (b) most readily prepared and (c) easiest to handle in a vacuum system. In a typical reaction 0.26 mole of CF-^N = SF2 and I.32 mole of cobalt (III) fluoride, CoFo, were loaded in a 100 ml. autoclave and heated

PAGE 68

62 to 200° for 24 hours. T'ne inaxlmurn pressure recorded was 130 atm. Seven compounds were Identified CF^, SF^, SOF2 1 CF^SP , (CF2)2NF, (CF3)2N -N(CF3)2 and N2. A reasonable mechanism to account for these products assumes decomposition of the starting material in presence of CoF^. CF N = SF2 ^'^^3^ CF^N + SF* CF oN + F (from CoF^) f — » CF^^ + N* ^ ^ L^ cf; + 3 (NF) CF* + SF* > CFj SF„ CF^-N T CFA > CF3 -N* 2CP3 -N* > ^^^2^2 N-N(CF3)2 CFo N* + NF > ^2 -^ F* CF3N* + (F*) > CF3 -N -F CF3 CF3 SF3 + F* > SF^ SF. + 2F (from C0F3) ^ ^^6 SFi^ + (H2O) > SOF2 + KP When SF^^ is passed through the V. ?. C. column, it is converted to SOF2. Thus the appearance of SOF2 in the products may result from reaction of SF^^ with V. ?. C. column packing. There appeared to be zero retention of C N = S backbone in the reaction of CFoN = SF2 and C0F3. It is concluded that the N = S bond is probably the most reactive bond in the system and is easily polarized and ruptured

PAGE 69

63 in the presence of CoFo. Slight decomposition of CFoN = SFg was observed as low as 100*^ when Ap-Fp and CoFo v;ere used as reagent. The compounds R^-N = SF2 have been fluorlnated (44) by a direct reaction with elemental fluorine to give CF^NF SFj-, CP^NFgjCCF^N = )2SF2 and SF.. The Preparation of R^N = S CF CFo F CF^ The Nperf luoroalkyl Sperf luorolsopropyl Imlnosulfur monof luorides, R^N = S(F) CF (CFo)2, were prepared by reaction of the corresponding iminosulfur dlf luorides with perf luoropropene over dry powdered cesium fluoride in a stainless steel vessel at about 85°. The reaction time was several days. The autogenous pressure varied with time but was in the order of 10 to 20 atm. The course of reaction R^N = SFg + CF^CF = CF2 ^^1 > RfN = S(F) CF (CF3)2 can be represented by the formation of the perf luorolsopropyl anion (CPo)2 CF which then attacks, as a Lewis base, an empty d orbital of the sulfur atom in the Iminosulfur dlfluorlde with subsequent elimination of a fluoride ion, viz. CF„CF = CF2 + F' R^ N = S— F \CP3.

PAGE 70

64 spectral absorption associated with the -N = SFg system generally found around I39O cra~^ (7.2^) seemed to have disappeared for the Isopropyl derivative. The shift appears to have occurred to longer wave length and was probably obscured by the strong C-F stretching bands (2?) which lie in the range of l^iOO to 1000 cm"-^. This hypothesis was checked using the compound C,H^N = S (OCH3 )2, which had no interfering CF absorptions and in which the N = S absorption (28) was observed at I283 cm*^. Absorption bands were observed for CF-^N = SF CF (CF3)2 ^^ '"^^^ cm"-'", for CgF^K = SF CF iCF.^)^ -at 1279 cm"^ and for C^Fr^N = SF CF (CFo)2 at 1272 cm"^ which are assigned to the N = S stretching mode, A second problem arose as a result of the n. m. r. spectral splittings attributed to the fluorine atoms on the carbon attached to nitrogen. In both the perf luorethyl group in C2F^N = SF-CF (CF3)2 and perf luoropropyl group in C^F^N = SF-CF (CF^)^, the spectral splitting for -CF2-Nis a non-equivalent quartet. Although CFo on N in perfluoromethyl CF^li = SF-CF (CFo)2 ^PP^^rs to be a quartet, it is actually two doublets so placed as to appear to be an equivalent quartet presumably as a result of spin-spin coupling with both the SF and CF fluorine atoms. It is believed that quartets in the ethyl and propyl derivatives are actually two doublets. The alkyl groups next to the -CF2 group have a tendency to bend so as to be els to one and trans to the other fluorine atom of the -CF2-N group.

PAGE 71

^5 Tbls causes both the fluorines to be non-equivalent which results in two doublets which appear as a non-equivalent quartet. The Syntheses of Several Perf luoromines A second product which was isolated in the reaction between trif luoromethyl iminosulfur difluoride and perfluoropropene over CsF had a molecular formula Cr,F, ^NS based on its molecular weight and elemental analysis. The Infrared absorption spectrum showed a strong peak at I78O cm~ which can be associated only with C = C or C = N stretching for the available elements. Its n. m. r. spectrum indicated five kinds of fluorine, one of them being attached to sulfur and none to nitrogen. Normal valency considerations ruled out the possibility of a -C = Cbond and indicated the possibility of a C = N bond In a ring. The absorption frequencies of the cyclic systems, nonaf luoropyridine at 175^ om" (29) and heptaf luoropyroline at I721 cm"-^ (30) have been associated with C = N stretching mode. On the basis of elemental analysis, molecular weight, infrared and n. m. r. spectra the compound C,-,F^^NS has been assigned the structure CP^ CF C S CF CFo ^ \ 11/ I I ^ CF^ N F CF^ A probable mechanism of the reaction is through the formation of CF3N = SF-CF i.C?'^)^' A fluoride ion attacks an empty d orbital of the sulfur atom with subsequent re-

PAGE 72

arrangement followed by elimination of fluoride ion, CF^ N = SF2 + CF^C? = CF2 -2sF^CF3 N = S CF CF^ £l—>^f^CF2 -'^= SF CF (CF3)2 ^ ^^3 F ^CF2 = N SF-CF (CFo)2 + F" F The intermediate has a'termin9.1 dif luoromethylene group (CF2 = '^) v;hich is very susceptible to attack (3I) by another Carbonion, ^V^CF = N SF CF (CFo)o CF CF CF3 > (CF-), CF CF = N S (F) CF (CF^), + F" F Abstraction of 2 fluorine atoms V (CFo), CF-C=N-SF CF (CF^)^ and ring closure ' ^ ^ \^_y The two fluorine atoms were taken up by CF^ CF = CF2 to form Co Fg. When C3F N = SF2 and CFo CF = CF2 were heated over CsF, the main product was C^F N = S(F) CF (CF3)2 along with a minor product (16 mole %) which had a molecular formula CgFNS based on elemental analysis and molecular weight. It has two fluorine atoms less than the major product C/.F-,tNS. The infrared absorption spectrum showed a bond at 170? cm" which can be associated with C = C or C = N stretching

PAGE 73

67 Tiode, but the normal valency considerations ruled out a G = C bond and indicated the possibility of a C = N bond in a rir^. Its n. m. r. spectrum indicated five kinds of fluorine, one of them being attached to sulfur. On the basis of these observations it is possible to propose two structures, viz, CFo CFo C S — CF CFo (b) ^ II ^1 1(e) ^ N^ F CF^ (c) (d) (a) ^ (b) (III) and CF3(c)

PAGE 74

68 no coupling between CF^ CF2in peaks c and d. Both are doublets. On the basis of evidence against structure (IV), coupled with the fact that the minor product of reaction betvjeen CF^N = SFg and CFoCF = CF2 has a three r.embered ring, the product C^F-j^-^KS is assigned the structure (III). A probable mechanism to (III) is via the formation of C^Fr,N = SF CF (CF^)2. A fluoride ion attacks on empty d orbital of the sulfur atom with subsequent rearrangement followed by elimination of fluoride ion, ViZj C^F N = SF2 + CP^ CF = GF2 -2sF_^ C2Fr;N = SF CF (CF^)^ F^ CF-. CFo CF N = SF CF (CF^), CF^— CFp — C — N = SF CF -(CF^)p -r F" P F ^^°gy''cf ^gg f Cr^" ^^"'^'^^'^' ^^''^"> CF3-CF2-C = rSF CF (CF3)2 The two fluorine atoms were taken up by perf luoropropene which was converted to perf luoropropane. Another fluoro imine was prepared when CF^N = SF^ was heated with CF„CF = CF CFo over CsF at 280° to 290°. It was expected that the reaction would follow the usual course; formation of carbanion which would attack the sulfur of iminosulfur difluoride with subsequent elimination of F~ ion, i.e. , CF3 CF = CF CF„ + F" > CF3 CF2 CF CF^ CF3 N =S— ^F > CF3N = SF CF CF2CF3 + P CF3 CF2 CF CF3 CF3

PAGE 75

69 but instead the Imlne , CF-.N = G CF2 CF^ vjas obtained. Perf luorobutere-2 is already in its most stable state with respect to CsF isomerisatlon and the absence of a terminal CF2group makes acceptance of a F~ more difficult. In that case formation of CF-^-CFg — CF — CF-^ is expected to require a high activation energy requiring temperatures in the order of 250° to 300°. It is possible that the expected compound, CF^N = SF CF CF2 CF^, even if CF^ formed, might be susceptible to decomposition at 3OO . It has been shown that -N = S bond is susceptible to rupture. It is postulated that the carbanlon first attacks the nitrogen atom of -N = S bond and SF2 is eliminated which reacts with CsF to form the complex Cs SF^, viz, CF^ N =^F, CF„ N SFp CF^ CF2 CF CF^ CF^ CF2 CF CF^ CF^ N— T-SFo CFo N = C CFoCF^ 3 N{ ^ 3 , 23 CF^ CFo C -'^F > CF 3 2 I . 3 CF ^ + SF2 + F" Cg^ + F" + SFg > Cg SF It would be expected that CsSFo is a solid ionic species. It appears to decompose on heating at 400° under vacuum. The decomposition products were sulfur and other unidentified volatile species some of which attacked glass, the sodium chloride windows of the Infrared cell, Kel-F grease and mercury.

PAGE 76

70 The perfluorolmine was identified on the basis of elemental analysis, molecular weight, n. tf. r. spectrum and by comparison with the infrared spectra of the several known imines (29, 32). The new inine shows infrared absorption at 1725 cm"-^ and has boiling point 36.6°. Haszeldine (29) reported the synthesis of the isomer CF„N = CP-CP^CFp CF^ 3 2-^3 (b. p. . 39O) and compared the infrared absorption frequencies of the double bond in the following compounds: (i) CF3 N = CF2 (ii) CF^-N^CF CF^ (iii) CFo-N=CF CF2CF2CF (iv) CF^ CF2 CF = CF2 (v) CF^— C = CP, 3 , 2 CF3 Comparison of (i), (ii) and (iii) indicates that as the chain length increases, the C = N absorption frequency decreases. Similarly comparison of (iv) and (v) indicates that as branching increases, absorption frequency decreases. A comparison of the properties of two isomers of C^ F -^2. ^ further confirms that the isomer prepared in this work is CFo -N = C — CFp — CF^ . 3,^3 CFo In order to determine whether the syntheses like that of CF3 N = C (CFo) G2 F. prepared from CF^N = SFg and CF_CF = CF CF over CsF is a general preparation of imines ; C2F5N = SF2 and CF3CF = CF CF3 were heated over CsF. No Absorption C = N

PAGE 77

71 C2F£rN = C (CF^) C2F^ was obtained. The reason Is that although CgF^N = SF2 and CF^CF = CF CF-3 were expected to react at temperature around 300°, the decomposition of C2F^N=SF2 to SF4 and CF3GN took priority at temperature 280° to 290°. The Reactions of SFi,, and SOF2 with Olefins The reaction between CgF^N = SF2 and CF^CF = CF CF^ over CsF v;as expected by analogy to result in either C2F^N = S(F) CF(CF3)C2F^ or C2F^N = C(CF3)C2F^ but a product with a molecular weight 426 and molecular formula CoFji^S2 resulted instead. If we consider that the imlnosulfur difluorides basically have the shape of the SOF2 molecule, the added steric hindrance of the C2P£' group over that of CFo group raight require a higher activation energy for a reaction between C2F^N = SF2 and CPoCF = CF CFo. A sudden rise in pressure at a temperature between 290° to 300° suggested that a decomposition was occurring. The analysis of the gaseous product indicated that CF3CN in quantitative yield, with respect to C2F^N = SF2, had been formed. C2F^N = SF2 ^CF^ C=N + SFi^ No SF^ was found. It had reacted with perf luorobutene-2. When sulfur tetraf luoride and perf luorobutene-2 were heated up to 300° without CsF, no reaction was observed. The experiment was repeated with SF/^ and CFoCF = CF CFo over CsF and a reaction occurred at 250°. The product {CqF-,i^S2)

PAGE 78

72 obtained was the same as that formed when C^Pi-N = SFp and CF CF = C? CFo were reacted. The F-^° n. m. r. spectrum of this new compound had a single peak located at -19'9 ppm. with respect to trifluoroacetic acid and is in the range of CF^ groups attached to a carbon atom which in turn is doubly bonded to another carbon atom. The presence of C = C bond is confirmed by a band in the infrared spectrum at I626 cm~l. Krespan (33) prepared CF^ C = C CFo by heating perf luorobutyne-2 S — S with sulfur in presence of iodine at 200°. The infrared spectrum showed a band at I629 cm"'^. The F " n. m. r. spectrum exhibited a single peak in the trif luoromethyl region at (-811 cps.) -14.4 ppm. On the basis of elemental analysis, molecular weight, infrared and n. m. r. spectra the nev/ compound was assigned the structure CFo C = C CF, s s \/ CF^ CF2 C — CF The reaction can be represented by these equations: 2 CF3CF = CF CF3 + 2 SFj;^ ^sP > C3F]_2^S2 + 5^2 CF3 CF = CF CF^ TP2 > CF3CF2CF2CF3 The compound CgF-j^2|S2 was also obtained when CF^C = C CFand SFi^ were heated over CsF. The reaction occurred at 150°. Perfluorobutane v;as also recovered.

PAGE 79

73 CF3 C = C CF3 + SF4 > CgF^i^S2 + F2 CFo C = C CF3 + F2 > CF3 CP2 CF2 CF3 The lov;er temperature required for perf luorobutyne-2 reaction (150°) than for the perf luorobutene-2 reaction (250°) Is a result of the smaller sterlc effect about the triple •bond. Moreover perf luorobutyne-2 might be a stronger reducing agent than perfluorobutene-2 and so less activation energy is required for the reaction with perf luorobutyne-2. In the gyntheseg of ^q^ii^^z* ^^^ sulfur of SP^ has been reduced from +4 to +2 oxidation state. Loth and Michaells (I5) treated thionyl chloride with anlsole in presence of zinc chloride and obtained 4,4 dimethyl dlphenyl sulfide in which sulfur has been reduced from the +k to the +2 oxidation state, viz, Cg H^,0CH3 + SOCI2 ^"^^2 ^ CH3O. C^Hi^ — S— C^H^^. OCH Recently Smith et, al. (5) treated phenyl Irainosulfur dlfluoride with anlsole in presence of zinc chloride and obtained 4,4 dimethyl dlphenyl sulfide. They did not report a mechanism. But when perf luoropropene was heated with SFi,, over CsF (43) at 150°, the products obtained were (CF3)2 CF SP3 and (CF3)2 CF — SF2 — CF(CF3)2. Though the authors did not report any mechanism, it is apparent, the following course of reaction takes place: CF3CP = GF2 + F" > CP3CF CF3 (CFo)^CF + SPo F > (CF-.)oCF SPo + F" 32 3 ^32(^)3 (CF3)2CF + (CF3)2CF SFg F > (CF3)2CF-SF2-CF (CF3 )2+ F (VI)

PAGE 80

7^ The sulfur of (VI) which is In +4 oxidation state can be reduced to +2 state by treatment with TiCl^^ (43). Perf luoropropene and perf luorobutene-2 react with SFi^ at 150° and 250° and produce different products. Probably the propene has a terminal =CP2 group which is very susceptible to fluoride ion attack and so lower reaction temperature is needed. But the lower temperature may not be high enough for the propene to act as a reducing agent. It is believed that in this work perf luorobutene-2 and perfluorobutyne-2 are acting as reducing agents and accept fluorine. The material balance of the reactants and products showed a loss of 10 grams which was SF4. The SP^^ reacted with CsP (34) to form CsS?< which is an ionic solid and .thus remained in the reactor when the volatile products were removed. The solid complex decomposes reversibly on strong heating under vacuum to give SF^^ and CsP. The reaction between CgP^N = SF2 and CP-^CP = CP CP^ over Cs? at 290° was in fact a catalysed reaction between SF^^ and CP CP = CP CP„. In order to justify this observation SPjij, and CF„CP = CP CP3 were heated over CsF and a reaction occurred at 250°. The V. P. C. analysis of the liquid product showed a shoulder that could not be resolved on any V. ?. C. column available, silicone oil, silicone gum, silicone elastomer, Kel-P ester or D. N. P., at temperatures below 70° and at very low helium flow rate. This mixture had a boiling point 142° whereas CgF-,^S2 bolls at

PAGE 81

75 135°' Comparison of infrared spectra of pure C3F-i_i^S2 (Fig. 7) and the mixture (Fig. 9) Indicates that the mixture 'did contain ^q^ii^^^' "''^ ^'^^^ assumed that one of the other products in the mixture was the product of a reaction between SOF2 and CF3CF = CP CPo, since commercial sulfur tetraf luoride (b. p. -38°) contains some thionyl fluoride (b. p. -iiS'^). In order to test this hypothesis, SOFg was reacted with CFoCF = CF CFo over CsF. At the reaction temperature of 200°, the main product obtained boiled at 149°. The molecular formula calculated from the elemental analysis and molecular weight was CgF-[_g S2O2. The Infrared absorption spectrum (Fig. 8) did not show any bond in C = C absorption region. The compounds CgP-j^^ S2 and '^q^iq^z^Z were mixed in the ratio 1:3 and the mixture boiled at IU3 . The Infrared spectrum of the synthetic mixture was similar to that in Figure 9. This proved that the mixture, obtained as a result of reacting SP/^ with CP^CF = CF CF3 over CsF contained CgF . S^ and CgP-,oS202 in the ratio of approximately 1:3. The boiling point of the mixture varied as a function of its composition. The n. m. r. spectrum of CgF-|_gS202 consisted of 4 pairs of peaks which were all multlplets. One pair of peaks (CF-i groups) was resolved more than another. A pair of peaks (CF2 groups) could not be resolved completely but the fourth pair of peaks (CF groups) were resolved. On the basis of above information it was estimated that there

PAGE 82

76 are two isobutyl groups, CF^-CF2 "^F — CF^ in slightly different magnetic environments in the compound. The two possible structures are CF« CFo I 3 II I ^ CF S-0S — CF I I CFp ^^2 and ^^2 ^ ^-^2 (VII) (VIII) structure (VIII) was ruled out on the basis of infrared spectral data. The infrared spectrum of sulfuryl fluoride SO2F2 (20) shows a strong = S = stretching frequency o at 1500 cm ^. Similarly = S = in || t| = S CF -CF-O-S-0 CFo 1 ^ 11

PAGE 83

11 characteristic of an Infrared absorption of a grouping consisting of one oxygen doubly bonded to sulfur. The mass spectral data for CoF-,gS202 shows a very strong peak at mass number 64 which indicates either 0-3=0 or -S-S. As the -S-3bond is much weaker than the -S=0 + bond, statistically an 0-S-O ion has better chance of survival under mass spectral conditions. M. Goehring prepared cyclic S^NgOg (40) (by heating SOCI2 with S^N^) having a -0-3=0 linkage which on reaction with moist nitrogen decomposed (41) to give S2^N^ and SOg. Based on these arguments, structure (VII) CFo CP-, I ^ II I ^ GF S — S — CF I I CFo CFo 1 ^ ,2 CF^ CP3 seems to be more plausible. A major by-product of the reaction between SOF2 and CF^CF = CF CF-^ over CsF was extremely reactive. This gaseous component could not be separated by V. P. C. as it reacted with the column packing. The infrared absorption spectrum was not reliable since it reacted with the NaCl windows and glass walls of the infrared cell. It could not be distilled because it reacted with glass and Kel-F grease. Thus, this reactive species was not identified and hence it is not possible to be certain about the mechanism of the reaction. 2 SOP2 * 2 CP3O? = CP CF3 |SP_> C3 P,8 S2 O2 H. (A)

PAGE 84

78 No perfluorobutane was detected among the products which indicates that CP^CF = CF CF3 did not act as a reducing agent as it did in case of the reaction between SF2^ and CF^CF = CF CF^ over CsF at 250°. The only other possible reducing species was SOF2 although no SOFi^ or F^-SOF was detected. Thus the reactive product (A) must be an oxidized species . The first step in the reaction between SOF2 and CF-CF = CF CF^ involves the nucleophilic attack of F~ ion on a carbon atom in the C = C bond of perf luorobutene-2 to form the carbanion: CF^CP = CF -CF3 + F" > CF^ CF^ CF -CP^ The carbanion attacks sulfur with subsequent elimination of P~ , viz, '' /^ 200° " F S -P — ^^ ^ P S CP^ CP CF2 CFj CF3 CF CP2 CP3 Rf = CFo CP2 CF CF F S = + F" >F-,S 0" 1 ^ F II ^ n R^ S— P > S^ S SPo -r F~ FtS 0" II ^ II ^ RfS SF2 P >Rf S S -R^ t_ ^2 ^^ + F '3

PAGE 85

79 RS SR^ + SOF2 >Rf S S Rf + (A) It is Just possible that (A) is F^S F. Another possible mechanism involves free radicals: CF3CF = CF CF-, + F~ > CF3CF2 CF CF^ *^ SOF2 ^ Bf S F -200^ R^ _ '^. + p. so; H R^ S . rearrangement , r^ ^^ p. • 0" u " R^ S. + *0 — S — R^ ^Rf -S-O-S-Rf F* + F' >F, Elementary fluorine is knovra to react with SOF2 in presence of CsF (42) to produce SF^OF in quantitative yield. The infrared absorption spectrum of SF^OF shows a band at 935 cm~^ (39), but no such band was detected in the spectra of the products. Fluorine also reacts with SOP2 without any catalyst (42) and the product is F/^S = which was not detected either. Though it is possible that SOF4 might have reacted with CsF (46) and the solid ionic product Cs-o-SF^ remained in the reactor when the volatile products were removed. Thus the free radical mechanism does not seem to be more plausible than the ionic mechanism. A material balance on the above reaction indicated a loss of 18 grams. This, probably, is a result of the reaction of CsF and SOF2 to give Cs-SF-. (42). This supposedly solid ionic compound remained in the reactor after

PAGE 86

80 removal of volatile conponents. The compound Cs-OSPo should decompose on heating to 300° to 400° to yield volatile species. Perfluorobutyne-2 and thionyl fluoride were heated over CsF and the liquid product was Cg F-,o $2 Og. The volatile products did not contain the extremely reactive byproduct (A) which was obtained from the reaction of SOP2 and CFoCP = CP CP^ over CsF. The reason is that CPoC = C CP^ requires three atoms of fluorine for conversion to the perf luorobutyl radical and hence no excess of fluorine is involved. 2CF^C =?CCFo + 3 SOF2 > CqF g S2O2 + (B) The by-product (3) remained in the reaction vessel. It m.ay have been absorbed on CsF or reacted to form a complex similar to that formed by SOF^^ (46), SOF2 (42) and S-V (3^) with CsF. When the reaction vessel was heated to 300° under vacuum, the effluent obtained had an intense blue color which changed to red on standing at -78° for 24 hours. On ivarming, the red-colored species changed to a colorless gas. Apparently the color transformation was associated with a free radical. But it is not clear whether the free radical obtained is the by-product (B) or a decomposition product of a complex (CsF + (3).). As the identity of (3) is not knovjn, it is not possible to speculate intelligently about the mechanism of the reaction between SOF2 and CFoC 5™ CCF^ over CsF. Since C^Fo N = SF2 was known to be less "thermally"

PAGE 87

81 stable than CgF^N = SF2 and since the extra sterlc effect of the C^Fr, group over that of C2F^ group would call for a higher activation energy, It seemed apparent that any reaction that occurred would be between SF^^ and CF^CF=CFCFo, the results of which have been established above. Thus, no attempt was made to react C^F^N = SF2 and CF^CF = CF CFo over CsF. The Reaction of R^ N = SFg with CFoC = C CFMiller (25) heated CFg CF-CF = CFg over CsF at 150° and obtained CFoC = CCF3 . The reaction clearly demonstrated that perf luorobutyne-2 Is stable over CsF at least up to 150°. The reaction between R^N = SF2 and CF^CsCCF^ over CsF was expected to yield a product having conjugated double bond, Rf.N = SF -C = CFCF^. It was true In case of ethyl and propyl Imlnosulfur dlfluorldes and the derivatives formed were CF^N » S C = CF -CFo and CF^CFoN = S -C = CF CF, . 3 , , 3 3 2,, 3 F CF^ F CF3 The course of the reaction appears to follow the same route as that postulated for the preparation of Rf.N = SF CF (CF3)2. The fluoride Ion attacks the olefin to produce a carbanlon, CF C = C CF+ F" > CF3CF = C" CFo The carbanlon adds to the sulfur atom of the Imlnosulfur dlfiuorlde with the elimination of fluoride ion.

PAGE 88

82 CF3 CFg N = SF -^ • CF3 > GP^CP2 N = F CF^ CF^ CF = C CF3 > GF0CP2 N = S G = CF CF, 3 The boiling point of ethyl and propyl derivatives, 98° and 111° respectively, were extrapolated to determine the expected boiling point for the methyl derivative CF -N=S-C=CP CF^ (approximate b. p. 85°). F CF3 However, the major product of the reaction between CF^N = SF2 and CF„C=C CF-^ over CsF boiled at 16?*^ and its infrared absorption spectrum (Pig. 10) did not show any absorption characteristic of the C = C or C = N bond. Obviously it was not CF-^N = SF-C (CF^) = CF CF^. The molecular weight and the elemental analysis of the pure product indicated that it had a molecular formula C-jQPgo ^o ^2 ^^^'^'^ ^^ a diraer of C^F^^ NS. When the reaction vessel containing CF^N = SF2, CF^CsC CF^ and CsF was heated to 80*^, the heat evolved was enough to maintain the system at that temperature until the reaction was over. The reaction to produce C2F5N = SF-C (CP3) = CF CF^ and C^Pr^K = SF C (CF3) = CF CP^ were not observed to be exothermic and the process to produce CP^N = SF C (CFo) = CF CPo was not expected to be exothermic Presumably the secondary reaction which resulted in the dimerisation accounts for the observed exothermicity . The infrared absorption spectrum of C-j_q P22 N2 S2 (Fig. 10) bears a strong resemblance to infrared spectrum of CgP-j^g S2O2

PAGE 89

83 (Fig. 8) because of the secondary butyl groups. The n. m. r, spectrum showed the compound C-,qF22N2S2 ^as a mixture of two isomers In the ratio 2:3. The spectrum was so complex and unresolved enough so that attempts to differentiate between the two Isom.ers els and trans, had to be abandoned. However the n. m. r. spectrum did shov; that both isomers had two Isobutyl groups in slightly different magnetic environments as in case of ^q^iqS>2'^2' The peak (a) in the n. m. r. spectrum (Table 2) appeared at -16.6 ppm . which was assigned to fluorines in CFo-Ngroup. This assignment is confirmed by the fact that fluorine resonances in the CF^N group of the compound CF^N = C (CF^) CF2 CF^ appear at -I6.88 ppm. A minor problem arose as a result of the low field value of +3.^+ ppm. for peak (d) in the new compound C-^q F^p NS which from its relative area was attributed to a single fluorine atom on a carbon atom. The reason that this CF resonance appears at the normal region for CF^ resonance, far down field with respect to a normal CF resonance (approximately 80 to 100 ppm.) (Table 2) is attributed to its position in the molecule. Based on elemental analysis, molecular weight, boiling point, infrared and n. m. r. spectrum the compound C^q F22 ^p ^2 ^^ assigned the structure CK3

PAGE 90

8i+ The anamolous resonance of CF in the ring is due to the decrease of the electron density around the fluorine atom; the lower the electron density around the fluorine, the lower the shielding and the lower the field at which the fluorine resonance is observed. The fluorine in the ring is deshielded as a result of the following factors: (1) Presence of the electronegative nitrogen atom next to carbon atoiTi drains off some electron density. (2) The cyclic part contains a large loop of TT electrons in which strong diamagnetic currents are induced by the magnetic field. This effect deshields the fluoride. (3) The relatively electronegative isobutyl groups could possibly further drain off the electron density from the ring resulting in further deshielding of the fluorine nucleus. The reaction of perf luorobutyne-2 and CF-^N = SF2 over CsF at room temperature yielded only two products, 81 mole per cent C^q F22 Ng S2 and 19 mole per cent of another product whose n. m. r. spectrum did not indicate an SF bond. Howeve:; in another reaction in which the temperature v.'as maintained at 56 for 4 hours, one of the products which was in 8 mole per cent showed a strong infrared spectral absorption at 1748 cm"^ (Fig. 11). Extrapolation of C = C bond absorption frequencies of the propyl and ethyl derivatives, Rf -N = SF C (CF^) = CF CF^, indicates that the methyl derivative should have a C = C bond spectral absorption between 1735 to I750 cm"-^.

PAGE 91

85 C^F N = SF G = CF CF3 1715 cm~^ CgF^N = SF C = CF CF3 172i+ cm""^ CP3 CF^N = SF C = CF CF, (1735-1750 cm"l) CF^ The spectrum (Fig. 11) of the product present In 8 mole per cent resembles that in Figures 12 and 13 In all essential features. The quantity obtained was so meager that further structural confirmation of the material was not possible. All subsequent attempts to increase Its yield were unsuccessful. It appears that although CtF-j_-,NS is formed, It largely dlmerlses In presence of CsF. Probably the presence of conjugated double bond system facilitates the dlmerisatlon process. The ethyl and propyl derivatives show little tendency to dlmerise under similar conditions as a result of the greater steric effect of bulkier groups. The dlmerisatlon of C^F^^^NS to C]_qF22N2S2 in presence of CsF can be regarded to occur as follows: CFo N = S— ^^^ CF CF^ -' I -, ^ >CFoN = S = C-GF,CFo + F" CF-. t ^ ^ C^ CP^ -N = S = C (CF3) CF2 CF3 ^ — ^CF^ N = S CF CF2CF3 F' CP3 R^ = CF3 CFg CF CF3 Two moles of anion condense with elimination of F~.

PAGE 92

R^ S" = N L CF-B. ci IRf S = N > CF. 86 "S Rf li N +F' CFr, I 3 Rf S T N, K S — R, Q CF I F CF^ Rf S = N S R^ + FII II ^ CF N

PAGE 93

SUMMARY Some reactions of perf luoroalkyl Imlnosulfur dlfluorldes, sulfur tetraf luoride and thlonyl fluoride with perf luoroolef ins in presence of dry powdered cesium fluoride have been studied. The thermal stability of perf luoroalkyl iminosulfur difluorioes over CsF was determined. The perf luoromethyl iminosulfur dif luoride, CF^N = SF2 , was most stable and did not decompose even at 300*^ for 72 hours. The perfluoroethyl homolog, C2F^ N = SF2, did not decompose below 282° whereas the perf luoropropyl homolog, C^FoN = SF2 1 decomposed to the corresponding nitrile and SF^, at as low as 250°. The thermal stability decreases with the increasing chain length. A new class of fluorocarbon compounds, N perf luoroalkyl S perfluoroisopropyl Iminosulfur monof luoride, involving a -C-N=S-C bo.nding arrangement has been prepared by the reaction of the analogous fluorocarbon Iminosulfur difluorides with perf luoropropene over solid powdered CsF at 80-90° in a closed vessel. A minor product of these reactions was isolated and identified as a unique cyclic Imine involving a three membered rings ,-C — S-. N The reaction betv;een CF-^N = SF2 and perf luorobutene-2, 87

PAGE 94

CF^ CF = CF CF^, at 300° did not follow the trend observed with perfluoropropene; Instead an Imlne CFoN = C(CF3)C2F^ Is produced. The reaction betv/een C2FS N = SF2 and CF3 CF = CF CF3 failed to occur up to 282° as the prior decomposition of iminosulfur compound to CFoCN and SF^ prevailed. The product formed was exactly the same as that obtained from the reaction of SF4 and CF3 CF = CF CF^ over CsF and was identified as the five membered ring compound CF3 C S. The reaction of perf luorobutyne-2 II C(CFr,) CoFc. CF3 — c — s^ -^ ^ and SF/^ over CsF at 150° also yielded the same substance. Thionyl fluoride reacted readily with CF3 CF = CF CF3 at 200° and with CF3 C = C CP^ at 150°. In both reactions the major liquid product was identified as ^2^5 (^^3^ CF S S CF (CF^) C2F^. The gaseous product was extremely reactive and was believed to be SF3OF. Another new class of compounds bearing a conjugated double bond system, -N=S-C=C, was prepared. The reaction of Cg F^ IM = SF2 and C3 F„ N = SF2 with CF3 C^CCp3 over CsF at 100 to 150° produced sulfur monof luorides of the structure Rf-N = S-C=CF CF^, , in almost quantitative F CF3 yield. But the reaction of CF3N = SF2 and CF3 C ^C CFo over CsF produced only a trace of sulfur monof luoride, CF3N = S C = CF CF3, the main product being its dimer F CF^

PAGE 95

89 C-, n '^oo No So identified as Co F. (CF^) CF S S CF (CF^) Co F,.. 253 ^ y 325 CF N The following table summarizes the pairs of reagents and products respectively involved in each of the CsF catalyzed reactions described in this dissertations. Table 3 The Reactants and Products of CsF Catalyzed Reactions Reactants

PAGE 96

90 Table 3 (continued) 3N=SF2 and 280° CF-^N=C-CF2-CF3 36.6° CF^CF=CF CF3 CF, ,0 " II ^C-CFoCF^ CF3CHC CF3 CF3-C-S^i ^ CFoCF— CF CF^ oro u '^C'3 CF3CSC CF3 CF^ CF3 1. I II I c CFr o CgF N=3F2 and 290 CF3-G-S ^ 11 .C-CF2CF0 135 CF3CF=CF CFo CFo-C-S^ | "^ ^ -^ CF3 SFi^ and 250° CF^-C-S ' 135° ^ II "^C-CFo CF, CFoCF=CF CFo CFo-C-S I ^ ^ ^ ^ CF3 SF^ and 150° CF^-C-S 135^ CF3 ^ P _ O —rilP—C _ A_<7 _r'D /-> _ Ct TlinO SOF2 and 200^ C2F^-CF-S-0-3-CF-C2F5 149 SOFp and 150° CoF^-CF-S-O-S-CF-CoF. 149 CF^N=SF, and 56 CnF^-CF S ^S-CF C,P^ 16?* 32 ^ ^ I II II I ^ ^ CF-^CSC CF CF3 CF N CF3 2. CF^-N=S-C = CF CFc> ^ I ( ^ F CF3 C2F^N=SF2 and 124° C2F^-N=S C=CF CF3 98^ CF3CSC CF3 F CF3 C3F„N=SF2 and 150° C3Fr;N=S C=CF CF3 111^ CF3CSC CF3 F CF3

PAGE 97

BIBLIOGRAPHY 1. 0. Glemser, Angew . Chem. Internat. Edit., 2_, 53O (I96O). 2. A. Mlchaells, Lleblg Ann. Chem., 22iL» 201 (1893). 3. 0. Glemser and H. Schroder, Z. anorg. allgem. Chem., 284, 97 (1956). 4. 0. Glemser and H. Rlchert, Z. anorg. allgem. Chem., 307 . 313 (1961). 5. W. C. Smith, C. W. Tullock, R. D. Smith and V. A. Engelhardt, J. Am. Chem. Soc . , 82., 551 (I96O). 6. W. T, Miller, J. H. Fried and H. Ooldwhite, J. Am. Chem. Soc, 82, 3091 (i960). 7. 3. Andreades, J. Am. Chem. Soc, 86 > 2003 (1964). 8. W. H. Christie, Doctoral Dissertation, University of Florida (1958). 9. R. D. Dresdner, F. N. Tlumac and J. A. Young, J. Am. Chem. Soc, 82, 583I (I96O). 10. E. R. Van Artsdalen, A. R. Brosl, T. A. Gens and J. A. Wethington, J. Am. Chem. Soc, 21, 1001 (1957 ) • 11. W. T. Miller, W. Frass and ?. R. Resnick, J. Am. Chem. Soc , 81, 1767 (1961). 12. R. D. Dresdner, F. N. Tlumac and J. A. Young, J. Org. Chem., 10, 3524 (1965). 13. F. S. Fawcett, C. W. Tullock and D. D. Coffman, J. Am. Chem. Soc, 84, 4275 (1962). 14. W. R. Hasek, W. C. Smith and V. A. Engelhardt, J. Am. Chem. Soc, 82., 543 (I960). 1 15. F. Loth and A. Miohaelis, Ber., 22., 2540 (1894). 16. A. Mlchaells, Ann., 224, 262 (1893). 91

PAGE 98

92 17. L. C. Yen, Master of Science Thesis, University of Florida (1957). 18. P. A. 3ond and D. A. Williams, J. Am. Chem. Soc, Xi> 3^ (1932). 19. A. M. Lovelace, D. A. Rausch and W. Pastelnek, Editors, "Aliphatic Fluorine Compounds." Relnhold Publishing Corporation, New York, 1958. 20. J. H. Simons, Editor, "Fluorine Chemistry," Vol. II. Academic Press, Inc., New York, 1950. 21. T. J. Brlce, J. D. LaZerte, L. J. Hals and W. H. Pearlson, J. Am. Chem. Soc, 21> 2698 (1953). 22. R. D. Dresdner, T. J. Mao and J. A. Young, J. Am. Chem. Soc. , 80., 3007 (1958). 23. R. N. Haszeldlne, J. Chem. Soc, 2504 (1953). 24. J. A. Young, Private Communication, Dunn Research Inst. 25. W. T. Miller, W. Frass and P. R. Resnlck, J. Am. Chem. Soc , 82, 1767 (1961). 26. A. L. Henne and W. G. Flnnegan, J. Am. Chem. Soc, 71 . 298 (19^9). 27. M. Stacey, J. C. Tatlow and A. G. Sharpe , Editors, "Advances in Fluorine Chemistry." Butterworths, Washington, 1965, Vol. k, pp. 253-313. 28. V. A. Engelhardt, Private Communication. 29. R. E. Banks, W. M. Cheng and R. N. Haszeldlne, J. Chem. Soc, 3407 (1962). 30. B. C. Bishop, J. B. Hynes and L. A. Blgelow, J. Am. Chem, Soc, 84, 3^09 (1962). 31. J. H. Fried and W. T. Miller, J. Am. Chem. Soc, 81, 2078 (1959). 32. J. 3. Hynes, 3. C. Bishop, P. Bandophay and L. A. Blgelow, J. Am. Chem. Soc, 8^, 83 (1963). 33. C. G. Krespan, J. Am. Chem. Soc, S^, 3434 (I96I). 34. C. W. Tullock, D. D. Coffman and E. L. Mutterties, J. Am. Chem. Soc, 86, 359 (1964).

PAGE 99

93 35* J. ^' Shreeve and G. H. Cady, J. Am. Chem. Soc, Q^., 4521 (1961). 36. J. D. Delfino and J. M. Shreeve, Inorg. Chem., 1, 308 (1966). 37. W. B. Fox and G. Franz, Inorg. Chem., ±, 9^6 (I966). 38. R. J. Gillespie and E. k. Robinson, Canad. J. Chem., li, 2171 (1961). 39. F. B. Dudley, G. H. Cady and D. F. Eggers , J. Am. Chem. Soc, 23.. 1553 (1956). 40. FI. Goehring and U. J. Helnke, Z. anorg. u. allgem. Chem., 22i, 297 (1953). 41. M. Goehring and U. J. Helnke, Z. anorg. u. allgem. Chem., 228, 53 (1955). 42. J. K. Ruff and Max Lustlg, Inorg. Chem., 2, 1422 (1964). 43. R. M. Rosenberg and E. L. Muettertles, Inorg. Chem., 1, 756 (1962). 44. M. Lustlg and G. K. Ruff, Inorg. Chem., 4, 1444 (I965). 45. M. Lustlg and J. K. Ruff, Inorg. Chem., 1, 287 (1964). 46. W. C. Smith and V. A. Engelhardt, J. Am. Chem. Soc, 82, 3838 (i960).

PAGE 100

BIOGRAPHICAL SKETCH Joglndar Singh Johar was born In Rawalpindi, India (now in West Pakistan) on January 1, 1935, and attended G. N. Khalsa A. V. Middle School, Dhamyal (near Rawalpindi) up to 19^7. Because of political disturbances in the Indian Subcontinent at the time of Independence, he was forced to leave his hone town In September, 19'^?, and spent one year In Refugee Camps. He joined school again in 1948 and passed the Matriculation Examination held by Panjab University in 1951* He obtained Bachelor of Science degree in 1955 and received B. Sc . (Hons. School) degree in Chemistry in 19^7. After completing all the requirements towards a Master's degree In Chemistry (by Thesis) in 1958 he Joined Arya College Ludhlana as a lecturer in Chemistry, though the M. Sc. (Hons.) degree was officially awarded to him in August, 1959. He was selected to Panjab Education Service in November, 1959, and was appointed as a Lecturer in Chemistry at the Government College Gurdaspur. He resigned that position when he came to the United States to Join the University of Florida in the fall of I962. He is married to the former Manjit Kaur Sodhia and is the father of one son. 94

PAGE 101

This dissertation was prepared under the direction of the ch?irman of the candidate's supervisory committee and has been approved by all members of that committee. It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1966 Dean, Colleygife of Ar/ts and Sciences Dean, Graduate School Supervisory Committee: Chairman J^//f/

PAGE 102

\^ 46 60 ':'*


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EVJEQMH7B_F7CYBH INGEST_TIME 2017-07-19T20:23:15Z PACKAGE UF00097860_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES