THE PREPARATION AND REACTIONS
OF SOME FLUOROCARBON
WILLIAM SANDFORD DURRELL
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
The author wishes to express his deep appreciation to
Dr. John A. Young, his research director, for his constant
interest and encouragement throughout this investigation.
The members of the entire staff of Reed Laboratory are
also due thanks for their many timely and helpful suggestions.
The success of this work is due in no small part to the
author's wife who was a constant source of inspiration and
TABLE OF CONTENTS
ACKNOWLEDGMENTS ........... ii
LIST OF TABLES ...................................... v
LIST OF FIGURES ...................................... vi
I. INTRODUCTION ............................... 1
II. DISCUSSION ....... ..... .......... ........... 4
A. Synthesis of Diamides .................. 4
B. Some Reactions of Amides, Diamides,
and Their Derivatives .................. 22
1. The thermal decomposition of
their metal salts .................. 22
2. Reactions with perfluoropropene .... 24
3. Reactions with acyl halides ........ 26
4. Reactions of metal salts of di-
amides with bromine and iodine ....* 29
5. Indirect fluorination with
silver difluoride .................. 31
6. Some reactions of N-bromoper-
fluoroglutarimide .................. 33
C. The Reactions of Dis-(perfluorodi-
methylamino)mercury .................... 40
III. EXPERIMENTAL ............................... 43
A. General .................................. 43
B. Synthesis of Diamides .................. 54
TABLE OF CONTENTS (continued)
1. Reactions of acids and nitriles.... 54
2. Reaction of anhydrides and
amides ............................ 61
C. Some Reactions of Amides, Diamides,
and Their Derivatives .................. 62
1. The thermal decomposition of some
of their metal salts ..o............ 62
2. Reactions with perfluoropropene .... 64
3. Reactions with acyl halides ........ 66
4. Reactions of metal salts of diamides
with bromine and iodine ............ 71
5. Indirect fluorinations of amides
and diamides ....................... 72
6. Some reactions of N-bromoperfluoro-
glutarimide ........................ 74
D. The Reactions of Bis-(perfluero-
dimethylamino)mercury .................. 80
IV. SUMMARY ....... ......................... .... 82
BIBLIOGRAPHY ..................................... 85
BIOGRAPHICAL SKETCH .................................. 89
LIST OF TABLES
1. Relative Rates and Equilibria in
the Reaction, RCOOH + R'CN RCONHCOR' ...... 6
2. Nuclear Magnetic Resonance Spectra ........... 46
3. Infrared Spectra Used in this Work ........... 48
4. Reactions of Acids and Nitriles .............. 56
LIST OF FIGURES
1. Yield of diacetamide at 1500. ............ 7
2. Yield of N-acetyltrifluoroacetamide
at 1500. *... *. .... .... .... ......... 8
3. Yield of bis-(trifluoroacet)amide
at 150o ................................ 9
4. Yield of N-acetyltrifluoroacetamide
from the reaction of acetic acid
and trifluoroacetonitrile at 1500. ...... 10
Even though Moissan39 first isolated elemental fluorine
in 1886, it is hardly likely that fluorine chemistry could
have developed into a significant area of chemistry had it
not been for the discovery by Swarts of the usefulness of
antimony fluorides in replacing halogens with fluorine.26
The later work of Henne led ultimately to the commercial
production of fluorine compounds for the first time, and
still accounts for the greatest proportion of fluorocarbon
derivatives commercially produced.
The discovery of the electrochemical process27 by Simons
has been of great significance in producing classes of fluoro-
carbons generally obtainable only with difficulty, if at all,
from the Swarts process. Indeed, most of the fluorocarbon
derivatives used in this work are derived from products of
War-time development of nuclear weapons created the need
for materials stable to the volatile, corrosive, highly toxic
uranium hexafluoride used in the gaseous diffusion process in
the separation of the desired uranium235 from the much more
common isotope of mass number 238. The ensuing crash program
developed a tremendous amount of data concerning fluorocar-
bons and their derivatives. Probably most important, the
direct use of elemental fluorine, in conjunction with silver8
or copperl3 as catalysts, gave practical yields of usable
products for the first time. In addition, the indirect method
of fluorination, using a powerful fluorinating agent such as
cobaltic fluoride, was found to be useful.
Besides their chemical and oxidative stability, fluoro-
carbons have demonstrated great thermal stability. The
combination of these factors has aroused interest in the
practical applications of these unusual materials. The
theoretical implications of these properties have no less
interest among theorists.
Chemical development of an area requires that the
materials under study undergo convenient chemical reactions.
These properties which arouse great interest in fluorocarbons
therefore hamper development of knowledge concerning them.
In fluorocarbon chemistry a "handle" or functional group
present in the molecule is much more important than in hydro-
carbon chemistry because of the much lesser reactivity of the
C-F bond. Of the various possibilities acyl derivatives are
of great utility and of these fluorocarbon acylamine compounds
seem particularly worthy of study since relatively little
work has been described utilizing these materials. For
instance, almost the only interest in fluorocarbon amides
has been as intermediates in the synthesis of nitriles32
and/or N-bromo amides.31 The N-bromo amides have been of
interest largely in regard to their ability to brominate
or as intermediates in the Ilofmann reaction.3
This study was, therefore, undertaken to elucidate
further the chemistry of these species.
The long-known reaction of acids and nitriles to give
diamides7 was extended to fluorocarbon acids and nitriles
and the reaction studied in considerable detail.
Acylation of fluorocarbon amides is reported only for
the trifluoroacetylation of trifluoroacetamide2 and the
base-catalyzed addition to fluoroolefins is unknown, although
the addition of other amidos to fluoroolefins11 was described
after the work reported herein was complete.
The indirect fluorination of fluorocarbon amides and
their derivatives is not described in the literature
although Bigelow and coworkers192 have carried out the direct
fluorination of several hydrocarbon amides.
The syntheses and several interesting reactions of
fluorocarbon N-bromo amides were studied, and new methods
for preparing amides substituted on the nitrogen by por-
fluoralkyl groups were sought.
A. Synthesis of Diamides
Reactions of acids and nitriles
In the early stages of this work it was necessary to
prepare bis-(trifluoroacet)amide, (CF3CO)2NH, in quantity.
Earlier work in this laboratory had shown that trifluoro-
acetic acid and trifluoroacetonitrile react quantitatively
at 1500 to give the desired compound in an essentially pure
state. Attempts to scale the reaction up from several gram
quantities to larger amounts, in an autoclave, resulted in
much poorer yields and extensive decomposition. Smith2 had
reported the synthesis of this compound in good yields from
the trifluoroacetylation of trifluoroacetamide with trifluoro-
acetic anhydride. Sufficient quantities were obtained in
this manner to achieve the desired synthetic objectives.
Nevertheless, it seemed desirable to study the acid-nitrile
reaction because of its potential synthetic utility and the
intrinsic value of a more detailed knowledge of its mechanism.
The reaction of nitriles with organic acids was first
studied before the turn of the century37 and, in the case of
acetic acid and acetonitrile at 2000, gave principally the
CH COOH + CI3 CN --'CH CONHCOCH
This early study led to the following generalizations:
(1) Fatty nitriles react with fatty acids to give
(2) Fatty nitriles react with aromatic acids to exchange
their cyano and carboxyl groups to give fatty acids and aro-
(3) Aromatic nitriles and fatty acids give mixed
(4) Aromatic nitriles and aromatic acids give secondary
There are numerous exceptions to the above as might be
expected from a reaction which is carried out in the 150-200o
No further work was reported until Wiley and Guerrant47
reported that the reactions of phenylacetic acid and nitrile,
and p-nitrophenylacetic acid and nitrile, give an equilibrium
yield of about 40% of the diamides at 175-1900.
In the present work acetic and trifluoroacetic acids
were reacted with aceto- and trifluoroacetonitriles.
Equimolar quantities of the appropriate acid and nitrile
were heated together in glass ampoules at 150 and 2000. The
reactions of acetonitrile were followed by removing the
ampoule after the desired period and evaporating under vacuum
the unreacted acid and nitrile. The nonvolatile residue was
shown to be fairly pure diamide by comparison of its infrared
spectrum with that of a pure sample. However, after reaction
at 2000 and long periods at 1500 decomposition had occurred
to varying extents.
The reactions of trifluoroacetonitrile were followed
by measuring the change in vapor pressure of the volatile
nitrile before and after reaction. The diamide and unre-
acted acid were separated by distillation in large scale
runs and in small scale runs the nonvolatile residue was
examined spectroscopically. It was determined in this manner
that the reaction was basically straightforward since there
were no absorptions of consequence in the infrared spectra
which were not due to the acid or diamide. To test the
reverse reaction pure samples of the appropriate diamide
were treated in the same manner as the acid and nitrile and
the conversions calculated as above.
The experimental results are summarized in Table 1.
Relative Rates and Equilibria in the Reaction,
RCOOH + R'CN = RCONHCOR'
System: Equil. Hours for rxn. Effect of Catalysis
Conv. to reach 50%
__completion Acid Base
R=CH3,R'=CF3 97%* 2 ++ ++
R=CF3,R'=CH3 72% 3 + 0
R=CF3,R'=CF3 96% 8 +++ +++
R=CH3,R'=C1H3 60% 96 ++ 0
*Apparent equilibrium, see text.
0 o 4-
S E E E
cI C3 -0.-3 d
4.) Q> u
ITeT J o o
/ 0 0 0D
I I I I1
\o V H
\ "0 N P
ep ^ *p Jo (fl
0o o o o
0 0 1-
0 O O r4
0 0 0- (<
W t l W E
/ O O 0a+
a*4 -H *c
-H r1 P-4 0
0 0 0
0 0 c
I 0 P 0
4 .- I4'
oo C ~
i 0 W. 0
o 0 0
0- 0 -0
K 0 PTI-'
\PTW~P J PTfl
\ E I ^
\ --i (i -i-
\ (-1 i -: i
\ 4- 4-
\ ^ E
\ __ 0 0 0
0 0 0 0
OPTLUITp JO PTOTA
While Wiley and Guerrant47 had established that the
extent of reaction in the aromatic acid-nitrile systems
are controlled by equilibrium, it remained to be shown that
an equilibrium would obtain in the present cases. In this
work equilibrium was shown by approaching it from either
direction. That is, the pure diamides were heated under
the same conditions as an equimolar solution of the start-
ing acid and nitrile. In the slow reacting acetic acid-
acetonitrile system equilibrium was somewhat difficult to
establish since tar formation and other side reactions
became important after long reaction periods. This problem
was of less consequence in the much faster reacting tri-
fluoroacetic acid-acetonitrile system, and of no consequence
in the trifluoroacetonitrile systems. In the systems where
acetonitrile was present the unroacted, recovered acid-nitrile
mixture was shown by its infrared spectrum to contain a
smaller proportion of nitrile than the starting mixture.
This indicates, as is well known, that nitriles undergo
various reactions, in the presence of proton acids, which
would account for the tar formation in these cases. Tri-
fluoroacetonitrile with its lower basicity, due to the
electron withdrawing fluorine substituents, is much less
susceptible to attack by proton acids and, therefore, is
less likely to be destroyed in such side reactions.
The position of equilibrium in the systems studied
presents an interesting situation. Wiley and Guerrant4
noted that electronegative substituents cause greater con-
versions to diamido, but offer no explanation. Our data
substantiate this fact, increasing substitution of fluorine
for hydrogen bringing about increasingly greater equilibrium
conversions, and suggest that these facts might be explained
on the basis of relative resonance stabilization of the
various reactants and products in their ground states.
An organic acid is a hybrid of the following structures:
0 0o 0-
II I I +
RC-OH RC-OH -- RC=OH
IA IB IC
For a nitrile we may write:
RC;N .- RC=N-
And for a secondary amide:
0 HO O0H 0 O-H 0
II I II I I I 1 I I II
RC-N-CR' *-- RC-N-CR' *-- RC=N-CR' 9- etc.
IIIA IIIB IIIC
If R is an alkyl group, the energies of the B structures
may be lowered by either inductive or hyperconjugative effects,
while if it were a trifluoromethyl group the B structures
would be of higher energy relative to A or C structures, in
which there is no positive charge adjacent to the trifluoro-
methyl group. Since a nitrile has no available form of the
C type, it should follow that electronegative substituents
destabilize a nitrile to greater extent than either an acid
or a diamide, assuming resonance structures of the C type
make a real contribution.
A concise illustration of the situation just described:
CH3 COOH-CF CN __
CF COOH-CrCN ,. CH CONHCOCF Energy
CH3COOH-CH 3_CN _- (CH 3CO)2NH
The difference in energies of products and reactants is
a rough estimate for the value of H for the reaction, and
since entropy change should be about constant for any of the
acid-nitrile systems, we may use the familiar relationship:
AH A &F = -RTInK
Thus, the greater the value of -aH, the greater should be
the conversion to diamide. The order of increasing conver-
sions can be estimated from the preceding discussion to be
in the order:
CH COOH-CH CN CF COOH-CH CN < F COOH=CFCN CH COOH-CF3CN
This corresponds approximately to the order of increasing con-
version found experimentally.
Care must be taken to differentiate between apparent
equilibrium and actual equilibrium in the CH COOI-CF CN system.
The equation involved in this case is:
CH3COOH + CF3CN -CF3COCNHCOCH3' CF3COOH + CH3CN
The reaction proceeds rapidly to give an apparent equilibrium
mixture composed of 97 of the mixed diamide. As time passes
the actual equilibrium mixture consisting of mixed diamide,
acetonitrile, trifluoroacetic acid, and small but unknown
amounts of acetic acid and trifluoroacetonitrile will obtain.
In one experiment the CF COOH CH CN mixture was
allowed to react at 2000 for eight hours. Fractionation of
the product gave a small amount of bis-(trifluoroacet)amide
in the most volatile fraction of the diamide cut, and because
of its considerably higher boiling point, diacetamide was
easily isolated from the high boiling residue. Finding these
diamides is to be expected from the total equilibrium in this
CF COOH+CH3CN = CF CONBCOCH3-- CH COOH+CF CN
CF 3CN JICH, CN
(CF3CO) 2NH (CH3CO) 2NH
There is another reaction which can occur but which is
only of consequence in hydrocarbon systems. That is the
reaction of the diamide with acid:
RCONHCOR + RCOOH -= (RCO)20 + RCON}I2
Davidson showed10 that an equilibrium of this type exists in
the case of dipropionamide and propionic acid and the reaction
of diacetamide and acetic acid was reported19 even earlier.
Even if this reaction were important in the CH COOH-CH CN
system, it would not change any of the conclusions reached
but would only serve to emphasize them since the correction
of the results for the formation of any acetamide, boiling
point 2220, would decrease the yield of nonvolatile materials
and therefore decrease the equilibrium conversion even more.
In none of the fluorocarbon systems was any of the amide
found except as discussed later in the thermal decomposition
of the sodium salt of the mixed diamide, CH3CONHCOCF3.
To explain the results of the earlier workers37 in the
same manner is tempting. Thus, in a typical fatty acid-
aromatic nitrile system the following equilibrium would exist:
ArCOOH + RCN --ArCONHCOR. RCOOH + ArCN
Since the equilibrium lies to the right, it is obvious that
the fatty acid-aromatic nitrile pair are the more stable.
If it is assumed that, as in the previous cases, the relative
stabilities of the nitriles are the most important single
factor, then the obvious rationalization is that the aromatic
ring's pi electrons are more available for lowering the energy
of the resonance structure IIB than these of an alkyl group,
which is wholly expected on the basis of modern electronic
An explanation for the relative rates of reaction of
the various systems studied is less obvious. At first sight
the most obvious sequence of steps is:
RC;N + R'C-OHI R =N-CR' (1)
0 -OH- 0 +OH- OH 0
R=tN-CR =R=N-CR' -- RC=K-CR' (2)
Tautomerization would then give the product.
If the first step were slow and the subsequent steps
fast, the most important factor should be the nucleophili-
city of the nitrile and the eleetophilicity of the acid or,
in other words, the ability of the pair of unshared electrons
on the nitrogen to attack the carbonyl carbon of the acid.
Thus, the fastest reaction might be expected to be that
between the nitrile with the most available electrons and
the acid with the most electron deficient carbonyl carbon.
Using these criteria, the relative rates of the systems
studied should be:
CF3COOH- C-CHC CN CH COOH-CHC CCOH-CFCN> CH COOII-CF3CN
Instead it was found that the reaction between acetic acid
and trifluoroacotonitrile is the most rapid and that between
acetic acid and acetonitrile the slowest. Therefore, either
the simple mechanism given is not followed, or an alternative
pathway is available.
It should also be noted that acids should catalyze the
reaction by enhancing the reactivity of the carbonyl group
of the acid unless the unshared nitrile electrons effectively
scavenge the system of acid.
The major requirements of an alternative mechanism are
that it must explain:
(1) The higher than anticipated reactivity of tri-
(2) The enhancement of rate in all cases by the addition
of small amounts of acid.
(3) The enhancement in rate, in systems containing tri-
fluoroacetonitrile, by addition of small amounts of base.
The above requirements are met if it is assumed that the
oxygen atom of the carbonyl group might participate by simul-
taneously attacking the nitrile carbon in the following manner:
R R R
C C C
0 N = 0 0 N -- product
C C C
R' OH R' OH R' OH
Using this mechanism as an alternative pathway in systems
containing trifluoroacetonitrile, all of the experimental
results may be rationalized. Considering only these systems,
it can be seen that acetic acid would react more rapidly than
trifluoroacetic acid because of the greater availability of
the carbonyl oxygen for participation, overcoming the lower
reactivity of its carbonyl carbon in comparison with tri-
fluoroacetic acid. In view of this it is not surprising that
these systems are catalyzed by base. The carboxylate anion
is less electrophilic than the unionized species but, more
important, has an increased capability for attack on the
nitrile carbon, thus resulting in a higher reaction rate in
the presence of base:
R R R
I I I
0 N, 0 N product
/ \ //
C c C
R' 0 R' 0- R' 0"
This mechanism might be expected to be followed only
in systems in which the nitrile has an abnormally electron
deficient carbon atom and is particularly susceptible to
nucleophilic attack such as in trifluoroacetonitrile, as
shown above. It is, therefore, likely that in the systems
containing acetonitrile the simple mechanism given originally
is followed. There is no catalysis by base in these systems
and trifluoroacetic acid reacts much more rapidly than acetic
acid by virtue of its more electron deficient carbonyl carbon.
All the systems studied are catalyzed to varying degrees
by sulfuric acid. Apparently, then in both mechanisms the
rate is increased as the concentration of the protonated,
highly electrephilic species increases:
0 H+ +OH
II I +
R'C-OH = R'C-OH *-- R'C-(OH)2
This result is compatible with the simple mechanism but is
not wholly expected in the participation mechanism since the
effect here should be to lower the availability of the elec-
trons on the oxygens adjacent to a formal charge. Apparently
the effect of increasing the olectrophilicity of the carbonyl
carbon more than makes up for the former effect. When acetoni-
trile was the nitrile constituent acid catalysis was less
effective. In these cases the nitrile probably competes with
acid for protons. As mentioned previously, acetonitrile is
known to react with proton acids.
A third possible mechanism, which Davidson10 assumes,
is the preliminary formation of the isoimide, IV, which then
decomposes to the imide or diamide in the following manner:
N N N
R-C H R-C H R-C H
I I II
0 0 0 0 0 0
\\ / (1) \ // (2) //
c C c
I I I
R' R' R'
Using this mechanism it is difficult to account for the
experimental results. First, since the reactions involving
acetonitrile are not base catalyzed, (2) must be the rate-
determining step in these cases. Conversely, (1) must be
the rate-dteermining step in systems containing trifluoro-
acetonitrile since they are base catalyzed. Since the over-
all rates are about the same in the reactions of these two
nitriles with trifluoroacetic acid, we are saying, in effect,
that acetonitrile is more olectrophilic than trifluoroaceto-
nitrile in their reactions with trifluoroacotic acid. This
is obviously false; trifluoroacetonitrile is well known to
be more susceptible to reactions with nucleophiles, as has
been previously discussed.
There is some possibility that the isoimido mechanism
is a special case, such as postulated for the participation
mechanism. There is a resemblance in these two mechanisms,
particularly in the decomposition of the isoimide to diamide.
The important difference is that in the isoimide case the
attack of oxygen on carbon occurs prior to, rather than
simultaneously with, attack of nitrogen on carbon. Thus, in
the trifluoroacetonitrile systems the fact that they are base
catalyzed might be explained but a rationalization of the
enhancement of rate by acid becomes more difficult.
The case for any of the mechanisms is certainly neither
proved nor disproved, but on the basis of present information
a combination of the simple and participation mechanisms seems
Reactions of anhydrides and amides
Smith's"' method of synthesis was used to prepare suffi-
cient quantities of bis-(trifluoroacet)amide for use as an
intermediate. This reaction may be written
CF CONH2 + (CF CO) 0 t (CF CO) NH
3 2 2' (cF300)2N
Attempted trifluoroacetylation of N-methyltrifluoroacet-
amide in the same manner gave the desired N-methyl-bis-(tri-
fluoroacet)amide only in low yield. It was found that the
N-methyl diamide reacted with trifluoroacetic acid to give
trifluoroacetic anhydride and the starting amide. Again we
have an equilibrium situation:
(CF3CO)20 + CF3CONIICH3 C (CFCO) 2NCH3 + CF3COOH
In this case the equilibrium lies to the left, whereas in the
trifluoroacetamide-trifluoroacetic anhydride reaction it lay
far to the right (since it was not detectable under the con-
ditions used). It seems doubtful that any arguments on an
electronic basis can be brought forward which would explain
the large effect of substituting a methyl group for hydrogen.
A methyl group in place of the hydrogen in structure IIIC on
page 12 would, if anything, lower its energy due to the
greater electron donating capacity of methyl over hydrogen.
The actual explanation would seem, therefore, to be involved
with steric or entropic considerations.
It is interesting to contrast the absence of any forma-
tion of nitrile in the trifluoroacetamide-trifluoroacetic
anhydride system with results reported by Davidson10 in the
benzamide-benzoic anhydride system in which an equilibrium
mixture containing 90% benzonitrile and only 2.7% dibenz-
amide is reached by refluxing at 224-243 for an hour. The
reason for the difference in ease of formation of nitrile in
the two systems may be related to the relative thermodynamic
stabilities of the nitriles (in this case trifluoroacetonit-
rile and benzonitrile) or may be strictly kinetic since we
have no good evidence that the fluorocarbon system is at
equilibrium with respect to the formation of nitrile.
B. Some Reactions of Amides, Diamides,
and Their Derivatives
1. The thermal decomposition of their metal salts
The thermal decomposition of the sodium salt of bis-
(trifluoroacet)amide gave a fair yield of 2,4,6-tris-(tri-
fluoromethyl)-1,3,5-triazine, presumably by the reaction:
CF3 N CF3
3 (CF3 O)2N -- N N + 3 cF3co00
Although the mechanism for this transformation was not
studied, it is certain that it is not due to the base cata-
lyzed trimerization of trifluoroacetonitrile,5 because none
of the highly volatile trifluoroacetonitrile was evolved
into a cold trap, which certainly would be expected if it
were an intermediate. It might seem possible that an "acti-
vated" form trimerizes before it can proceed to nitrile.
N 0 N O" N 0 triazine
// \ // 1 \ / . /. .
CF3-C C CF3-C C CF3-C C
3-\ C3\ 3\ 0 c .z- \c
0 CF3 0 CFJ O CF CF CN+CF COO-
3- 3F 3
This mechanism demands that in the base catalyzed
reaction of acid and nitrile, which seems to go through
the same intermediate, the trimer should be formed to the
exclusion of diamide. Since this was shown earlier not to
be the case, the mechanism must involve either a series of
additions or some other form of simultaneous reaction of
two or more imide anions.
Similar treatment of the mixed diamide resulted in
the following transformation:
N- CF3CONHCOCH3I CF CONH2 + tars
A 70% yield of trifluoroacetamide was obtained based on
diamide, and 140% based on sodium. The exact nature of
this reaction remains unknown, however, in particular the
relationship with the related reaction of bis-(trifluoro-
Sodium perfluoroglutarimide decomposed to a carbon-
aceous mass from which no identifiable products could be
The thermal decomposition of the mercury salt of bis-
(trifluoroacet)amide did not give any identifiable products
except metallic mercury, and traces of expected pyrolysis
products, trifluoroacetic acid, trifluoroacetonitrile, and
fluoroform. The mercury salt of perfluoroglutarimide was
considerably more thermally stable than that of bis-(tri-
fluoroacet)amide, a portion subliming at 4700 rather than
2. Reactions with perfluoropropene
Additions of nucleophiles to fluoroolefins have been
known since 1946 when Hanford and Rigby15 reported the re-
action of alcoholates with several fluoroolefins. The re-
action has been of considerable synthetic utility. Among
the nucleophiles used, as well as various alcohols,29 have
been amines,30 mercaptans,33 organometallic compounds29
and, while this work was in progress, amides.1
Addition has not been reported, however, of fluorine
substituted amides or diamides to fluoroolefins. Additions
of four of these materials to perfluoropropene were attempted,
trifluoroacetamide, N-methyltrifluoroacotamide, perfluoro-
glutarimide, and bis-(trifluoroacet)amide. Of these only
N-methyltrifluoroacetamide gave the desired addition product:
The anion of the amide was made by adding sodium to the
molten amide and heating gently until evolution of hydrogen
was complete. The addition was carried out in an autoclave
and gave a 75% yield after 40 hours at 800. The distilled
product was shown to be 98-990 pure by vapor phase
chromatography. The structure of the adduct was shown une-
quivocally by means of the proton magnetic resonance spectrum
which showed, besides an absorption associated with the
N-methyl protons, a proton peak which exhibited a complicated
fine structure, a complete interpretation of which is incom-
plete but which is undoubtedly due to a proton in the
position, X-CF2CFHCF3. The product slowly evolved HF, but
an attempt to isolate the dibromide of the olefinic product
Attempted additions using identical conditions were
totally unsuccessful with the other amides mentioned pre-
viously. The sodium salts of all these materials were rather
unstable compared to the sodium salt of N-methyltrifluoro-
acetamide. The sodium salt of perfluoroglutarimide was most
readily prepared by an exchange reaction with sodium methyl-
ate, methanol being removed under vacuum until the dry imide
salt remained. The sodium salt prepared in this manner was
totally inert to perfluoropropene at 100o and decomposed at
slightly higher temperatures. When sodamids was added to
perfluoroglutarimide the mixture sparked and burned. The
attempted reaction of the sodium salt of bis-(trifluoroacet)
amide with perfluoropropene resulted only in decomposition to
give the triazine discussed previously.
*A complete description of the nuclear magnetic resonance
spectra of this compound will be reported on by Dr. W. S.
Brey and coworkers.
3. Reactions of acyl halides
As with perfluoropropene, only N-methyltrifluoroacet-
amide underwent reaction with acyl halides. Attempted
reactions of the sodium, mercury, and silver salts of per-
fluoroglutarimide; and of the sodium and mercury salts of
bis-(trifluoroacet)amide with trifluoreacetyl and/or per-
fluoroglutaryl chloride were unsuccessful. With perfluoro-
glutaryl chloride and the mercury salt of perfluoroglutar-
imide the desired reaction was:
[CF2 (CF2CO)22NHg + CF2(CF2COCl)2-- CF2 (CF2CO)2NCOCF2CF2CF2-
CON(COCF2)2CF2 + HgC12
While none of the desired N,N,N',N'-bis(perfluoroglut-
aryl) perfluoroglutardiamide was formed in this reaction, a
successful and interesting synthesis was eventually found
which is described later*
The sodium salt of N-trifluoromethylacetamide, prepared
from sodium sand in ether or tetrahydrofuran, reacted with
trifluoroacetyl chloride to give the desired N-methyl-bis-
(trifluoroacet)amide. This compound could also be prepared
from the amide and acyl chloride in the presence of pyridine,
as described by Thompson. 4
The sodium salts of both the imide and the diamide gave
precipitates with trifluoreacetyl chloride but no triacyl
nitrogen compounds were obtained. The only volatile products
were the diethyl etherate (when diethyl ether was used as the
solvent) and the tetrahydrofuranate (when tetrahydrofuran
was used as the solvent) of trifluoroacetic acid. Addition
compounds of this type were originally reported by
Hauptschein and Grosse. The diethyl etherate had the same
physical constants given by these workers and measurement of
the relative intensities of the proton absorptions in the
proton resonance spectrum confirmed the formula proposed
originally for the etherate, 3 CF3COOI2 (C2H5)20. The
formula of the unreported tetrahydrofuranate was shown by
the same means to be 2 CF COOH*C4Hg0.
Trifluoroacetyl chloride must have been the source of
the acid found in the product since it was the common re-
actant in both systems. Since neither of the etherates were
found in the reaction with the N-methyl amide, the imide and
the diamide must play a role in the formation of acid, but
there is insufficient information to draw any conclusions.
It was thought that the mercury salts of the imide and
the diamide might be more likely to react with acyl chlorides
in the desired manner since the mercury salt of N,N'-bis-
(trifluoroacetyl)hydrazine reacts fairly readily with tri-
fluoroacetyl and perfluoroglutaryl chlorides.
The mercury salts of the imide and the diamide were
synthesized by means of their exchange reaction with mer-
curic acetate in anhydrous media. The mercury salts were
completely unreactive toward the acyl halides in any system
tried at temperatures up to 1000. Use of trifluoroacetic
acid or FC 102 (an inert fluorocarbon ether mixture) as
solvents was also ineffective.
It is well known41 that the silver salt of perfluoro-
glutarimide is reactive toward bromine in trifluoroacetic
acid. Trifluoroacetyl chloride was reacted with silver salt
under identical conditions but instead of N-trifluoroacetyl-
perfluoroglutarimide, trifluoroacetic anhydride was formed,
evidently from the reaction of the acyl halide with the
silver salt of trifluoroacetic acid. It is to be expected
that the trifluoroacetate and perflueroglutarimide anions
exist in an equilibrium:
CF3COO- + CF2(CF2CO)2NH CF3COOH + CF2(CF2CO)2N-
but whether either would actually exist as the free, uncem-
plexed anion is difficult to say. It is likely that silver
plays a more important role than the above equilibrium would
suggest. At any rate, trifluoroacetyl chloride reacts
preferentially with trifluoroacetate anion (or a complex
silver salt), and bromine reacts preferentially with the
anion of perfluoroglutarimide (or its complex silver salt).
The silver salt of perfluoroglutarimide was prepared free
of trifluoreacetic acid but, using excess imide as solvent,
no reaction occurred with the acyl chloride.
The nonreactivity of the imide and diamide anions toward
electrophiles, such as acyl halides and fluoroolefins, is
probably due to resonance stabilization of the anion by the
electronegative substituents. Thus, the anion of N-methyl-
trifluoroacetamide, being less stabilized by the electron
donating methyl group, undergoes reactions with both tri-
fluoroacetyl chloride and perfluoropropene.
4. Reactions of metal salts of diamides with bromine and
Either the silver 8 or the mercury salts of perfluoro-
glutarimide could be.brominated readily in trifluoroacetic
acid to give N-bromoperfluoroglutarimide. Iodine reacted
slowly with the silver salt of perfluoroglutarimide but the
product was too unstable to isolate.
Bis-(trifluoroacet)amide did not react with bromine
under the above conditions, however. The bromine which was
absorbed when added to the silver salt of the diamide in
trifluoroacetic acid reacted according to the equations
below to give N-brometrifluoroacetamide.41
Ag20 + 2 CF COOH --2 Ag + CF COO" + H20
H20 + (CF3CO)2NH CF3CONH2+ CF CO011
CF3CONH2 CF CONHBr
When the residue which remained after evaporating the tri-
fluoroacetic acid from the brominated mixture was treated
with dry HBr the immediate formation of bromine was noted.
Both trifluoroacetamide and bis-(trifluoroacet)amide were
present in this product. Under anhydrous conditions,
achieved by adding a slight excess of trifluoroacetic
anhydride to the reaction mixture, neither the silver,
sodium, nor the mercury salts reacted with bromine. It
may be concluded, therefore, that any reaction in the
original bromination was due to trifluoroacetamide being
The difference in reactivities of perfluoroglutarimide
and bis-(trifluoroacet)amide toward bromine is remarkable.
Since both compounds should have similar electronic environ-
ments about their nitrogen atoms, the reason for this great
difference must lie elsewhere. This might be attributed to
the stabilities of the metal complexes. It could well be
that perfluoroglutarimide forms a stable silver complex in
solution and that the first step in bromination is complex-
ing of bromine with silver followed by intramolecular trans-
fer of bromine. Such sequences for complex ion reactions
have been demonstrated in numerous cases, This conject-
ural mechanism might be written:
CF2CO CF2CO CF2CO
CF2 N-Ag ---- CF2 N-AgBr2-- CF2 NBr + AgBr
\ /\ / \ /
CF2CO CF2CO CF2CO
In the bis-(trifluoroacet)amide case a lower stability
of the complex, compared to the free anion and silver ion,
would explain its lesser reactivity toward bromine. Further
evidence for this point will be given in the next section.
5. Indirect fluorination with silver difluoride
Bis-(trifluoroacet)amide reacts with silver difluoride
at 600 and atmospheric pressure to give trifluoroacetyl
chloride and trifluoromethyl isocyanate. A plausible mech-
(CF CO)2NH- Ag2 CF3CO)2N A--- CFCOF + CF CON: -- CF NCO
The parenthesized radical may or may not have an independent
existence; that is, it may actually be a complex of divalent
silver which is rapidly attacked by nascent fluorine or de-
composes to give the transitory radical which is then attacked.
In either case the attack occurs to break the carbon-nitrogen
bond to form trifluoroacetyl fluoride and the electron defi-
cient nitrogen compound which rearranges, as in the Hofmann
reaction, to the isocyanate. Both the acyl fluoride and
isocyanate volatilize as formed out of the reaction vessel
into a cold trap.
Perfluoroglutarimide might be expected to react similar-
ly but under the same conditions gives no volatile products
after an initial evolution of HF. Even at much higher temper-
atures in an inert diluent only a silver salt can be isolated.
Formation of a complex is indicated in both cases, in
the diamide case as an intermediate and in the imide reaction
as a stable product. Earlier work49 has shown that silver
difluoride gives more specific reactions with compounds
containing carbon-nitrogen unsaturation than other strong
fluorinating agents. This specificity might well be attri-
buted to preliminary formation of a complex, thus lowering
the activation energy and allowing the fluorination to be
carried out under relatively mild conditions:
NH ---- N AgF + HF (1)
N AgF = N- + AgF (2)
In the case of bis-(trifluoroacet)amide the equilibrium
in (2) lies far to the right and further fluorination occurs
rapidly. With perfluoroglutarimide the equilibrium lies to
the left and further fluorination does not occur under the
conditions used. As in the attempted bromination, the
relative stabilities of the complexes allow a reasonable
mechanism to be offered and rationalizes seemingly anomalous
behavior. Further, it was previously noted that the mercury
salt of the diamide is considerably less stable thermally
than that of the imide.
Attempted fluorination of perfluoroglutarimide with
iodine pentafluoride failed to give any fluorinated product,
only perfluoroglutarimide being identifiable in the product.
Bis-(trifluoromethyl)carbamyl fluoride and perfluoro-
N,N -dimethylacetamide react similarly to one another with
(CF3)2NCOX Ag2 (CF3)2N. + COFX
a (CF ) 2NF
X = F,CF3
In both cases the products were the same but the pro-
portions of each varied. When X = F, the yield of the
nitrogen fluoride was 60% and hydrazine only 13%, whereas
when X = CF3 the yield of the nitrogen fluoride was only
30% and the yield of hydrazine up to 45%. The reason for
this difference may be due to the replacement of fluorine
with the trifluoromethyl group, or may be due to differences
in the condition of the silver difluoride used. In the
heterogeneous reactions of silver difluoride the effect of
the surface area on yields can be tremendous.
6. Some reactions of N-bromoperfluoroglutarimide
It has been reported that both N-bromotrifluoroacet-
amide40o21 and N-bromosuccinimide52 form addition compounds
with cyclohexene, of the type :NHHCHBr(CIH2)4, and Young50
reported the facile addition of bis-(trifluoromethyl)nitro-
gen bromide to perfluoropropene:
(CF3)2NBr + CF3CF=CF2--- (CF3)2NCF2CFBrCF3
It seemed feasible to carry out the analogous addition of
X-bromoperfluoroglutarimide to perfluoropropene. When these
two were allowed to stand together in the dark no reaction
occurred, but if exposed to sunlight er heated in the presence
of peroxide a rearrangement of the X-bromo compound occurred
rather than bromination of, or addition to, olefin. It was
shown that the reaction occurred in the presence of perfluoro-
propene or chlorotriflueroethylene, but not in the presence
of trifluoroaceticanhydride or in the compound by itself.
The product was found to be o-bromoperfluorobutyryl
isocyanate, Br(CF2)3CONCO. The following evidence is offered:
(a) The product gave a band in the infrared spectrum
at 4.39 microns, indicating either nitrile or isocyanate, and
a negatively substituted carbonyl band at 5.5 microns. It
reacted violently with water, but gave no precipitate with
AgNO3 and did not oxidize iodide ion.
(b) It had a nuclear magnetic resonance spectrum, as
did its hydrolysis product, which conclusively showed an open
chain structure of the type X-CF2CF2CF2-Y.
(c) Hydrolysis gave '-bromoperfluorobutyramide, which
was identified by its infrared and nuclear magnetic resonance
spectra, as well as elemental analyses. Unfortunately, as
noted in the experimental section, the analytical result
for bromine on the amide was slightly low.
(d) The reaction with benzyl alcohol gave a product
which, according to the infrared spectrum and elemental
analyses, was Br(CF2)3 CONHCOOCH2C6H5.
Because of the cleanness of the rearrangement reaction,
good yields of a relatively pure product were obtained. Per-
fluoroglutarimide, present as a contaminant in the N-bromo
compound, was the only detectable impurity, along with the
unreacted olefin. It was immediately obvious that all the
elements of the starting material were present in the product
and that it actually was a rearrangement that was being dealt
with. There are only three possible products which contain
the -(CF2)3- chain:
CF2(CF2CO)2NBr --- BrCOO(CF2)3CN, or
Br(CF ) 3CONCO
The first product may be eliminated since its hydrolysis
should give a succinic acid derivative and bromide ion. The
second possible product should give a glutaric acid deriva-
tive and positive bromine which would oxidize iodide to free
iodine. It did neither of these. The third possibility
must, therefore, be the actual product.
Also important, a similar rearrangement has been re-
ported for N-bromosuccinimide34*22923 to give 9-bromoprop-
Martin and Bartlett propose a mechanism which simply
involves rearrangement of the succinimide radical:
CH2CO CH2C CH2* C
2\ 2\ Br*
I NBr -- N -- --- BrCH2CH2CONCO
/ / /
CII2CO CH2C CH}2C
Johnson and Bublitz,2 in a systematic study, showed that
the mechanism was more complex. In addition to chloroform
or bromoform as solvent, a terminal olefin containing a 20
or 30 allylic hydrogen was required to be present. Under
these conditions addition rather than abstraction of allylic
hydrogen, the usual reaction of the succinimide radical,14,46
might be expected to be favored. Since the rearrangement of
N-bromoperfluoroglutarimide occurs under somewhat similar
conditions, it might be surmised that the first step is
addition, but rearrangement of either the adduct radical
or a subsequent intermediate occurs to give the observed
product. Unfortunately, no reactions of this type have been
reported to this writer's knowledge. The rearrangement bears
a formal resemblance to the reactions of trifluoroacetamide34
and c-nitroacetamide7 with aqueous hypobromite which Barr
and Haszeldine3'P formulate as follows:
CF3 C= CF3 + NCO-
In anhydrous media the sodium salt of N-bromotrifluoro-
acetamide decomposes in a normal Hofmann fashion to give
sodium bromide and trifluoromethyl isocyanate. The analogy
is almost certainly only formal since in the trifluoromethyl-
acetamide case an ionic mechanism is followed, whereas the
light and peroxide catalyzed decomposition of N-bromoper-
fluoroglatarimide is certainly free radical in nature.
The exact mechanism must remain in doubt until a more
thorough study of both systems is possible.
The most unusual finding in this study was that N-bromo-
perfluoroglutarimide reacts, with increasing difficulty, with
hydrogen bromide, benzoyl bromide, and perfluoroglutaryl
chloride to give bromine and the corresponding N-H or N-acyl
CF2(CF2CO)2NBr + HBr----CF2(CF2CO)2NH + Br2
CF2(CF2CO)2NBr + C6H5COBr ---CF2(CF2CO)2NCOC6H5 + Br2
CF2(CF2CO)2NBr + C1CO(CF2)3COC1 --CF2(CF2CO)2NCOCF2CF2CF2CON-
The reactions of hydrogen and benzoyl bromides were
rapid and essentially quantitative. The N-benzoyl compound
was identified by its infrared and nuclear magnetic resonance
spectra which agreed with the proposed structure in every
detail. The nuclear magnetic resonance spectrum was partic-
ularly useful, showing absorptions similar to those of the
imide itself, but slightly shifted, and since only two types
of fluorine were present the benzoyl group must be substi-
tuted on the nitrogen, instead of on the carbonyl oxygen,
which could be considered a possibility.
It was noted that the N-benzoyl compound picked up
moisture from the atmosphere extremely rapidly.
The reaction with perfluoroglutaryl chloride gave a
25% yield of N,N,N ,N'-bis-(perfluoroglutaryl)perfluoro-
glutardiamide. It appeared that bromine was formed but the
fate of the chlorine is not known. About half the N-bromo
compound was recovered. Once again nuclear magnetic reson-
ance was important in the proof of structure, showing four
types of fluorine in the area ratios of4:2:8:4, as expected
from the structure:
/ 2 / \
CF2(c) NCOCF2CF2CF2CON (c) CF2
\ / \ /
(d) CF2CO (a)(b)(a) COCF2(d)
In addition, the fine structures of the assigned peaks
were identical to those in the parent systems, perfluoro-
glutarimide and perfluoroglutaryl chloride, and the chemical
shifts wore similar but, as expected, not identical. Al-
though this compound is probably sensitive to water, it
was not tested because of the small amount on hand.
The reaction failed with trifluoroacetyl bromide and
perfluorooctanoyl bromide, perhaps due to the impossibility
of removing bromine from the reaction mixture in the first
case and the difficulty of separating perfluorooctanoic
acid from any of the expected product which might have been
formed in the second case. In both reactions an initial,
rapid formation of bromine occurred which was undoubtedly
due to reaction of the N-bromo compound with dissolved HBr.
As mentioned before, this is a very rapid reaction. After
this initial surge, evolution of bromine was very slow,
evidently slower than the inevitable picking up of moisture,
which occurred even though every normal precaution was taken.
In less unusual reactions, the N-bromo compound oxidized
mercury to mercuric bromide and formed the mercuric salt of
the imide; it also reacted with bis-(1,2,2,2-tetrafluoro-
ethyl)mercury to give the previously unknown 1,1,1,2-tetra-
fluoro-2-bromoethane in 50% yield.
It appears that all of the reactions of the N-bromo
compound which were studied, with the possible exception
of its rearrangement to w-bromoperfluorobutyryl isocyanate,
depend in some way on the highly positive nature of the
bromine and its oxidizing power.
The driving force behind the reaction with acyl halides
is not clear unless the positive bromine functions as a
Lewis acid, similarly to a Friedel-Crafts catalyst, to form
an acylium ion which may then react with the weakly nucleo-
philic imide anion:
CF2(CF2CO)2NBr = CF2(CF2CO)2N- + Br+
Br+ + RCOBr -- RCO + Br2
CF2(CF2CO)2N- + RCO+ -- CF2(CF2CO)2NCOR
The fact that hydrogen and benzoyl bromides react much
more rapidly than triflueroacetyl and perfluorooctanoyl
bromides favors this mechanism since the positively charged
ions formed in the first and second cases should be much
Thermodynamically speaking, there is a net gain of
stability of about 20 kcal due almost wholly to the very
weak N-Br bond.45
C4 The Reactions of Bis-(per-
Bis-(perfluorodimethylamino)mercury has been found to
react readily with several acyl halides to give the fluoro-
carbon amides. This is the first reported general synthesis
of compounds of the type: RCON(CF3)2. The compounds where
R = CF CH and CH5 were made from the respective acyl
chloride in essentially quantitative yields:
(CF3)2N-Hg-N(CF3)2 + RCOC1 --- RCON(CF3)2 + HgC12
The trifluoromethyl derivative had already been prepared
in low yield from the electrochemical process.51
While the properties of these materials were not studied
in detail, except for the previously described indirect
fluorination of perfluoro-N,N-dimethylacetamide, the benzoyl
compound appeared to be quite stable to water while the per-
fluoro derivative seemed to be very easily hydrolyzed.
The mercurial, surprisingly, failed to react with phos-
gene, except slowly to form perfluoro-2-azapropene. Although
the other reactions occurred at or below room temperature,
this one had not proceeded to a substantial extent even after
five days at 1000.
The mercurial reacted readily with trimethylchlorosilane
to form, not the expected silazane, but the azomethine, as
above, and trimethylfluorosilane. It would seem likely that
the silazane is formed as an intermediate:
(CF )N-Hg-N(CF )2 + (CH ) SiCl --- CF )2NSi(CH )3
CF N=CF2 + (CH3)3SiF
The decomposition of the silazane is analogous to the
following pyrolysis, used to synthesize the azomethine:51
(CF3)2NCOF CF N=CF2 + COF2
Mao33a observed the formation of azomethine in an
attempted reaction of the mercurial with phosphonitrilic
All temperatures are given in degrees centigrade and
are uncorrected. The elemental analyses are by Schwarzkopf
Microanalytical Laboratory, Woodside, New York.
The fractionations described herein were carried out
on standard columns, the size and type of packing depending
upon the degree of difficulty of the desired separation and
the amount of material available.
The vapor phase chromatography was carried out on a
Perkin-Elmer Vapor Fractometer, Model 154-B. The column
was packed with the ethyl ester of Kel-F Acid 8114 supported
on Celite. The infrared spectra were taken on a Perkin-
Elmer Model 137B double beam "Infracord" Spectrophotometer,
equipped with sodium chloride optics. Nuclear magnetic
resonance spectra were obtained with a 60 mc Varian NMR
Spectrophotometer, using CF3COOH as a reference standard.
The following chemicals were commercial products and
were used as received without further purification:
Ethyl trifluoroacetate Peninsular ChemRosearch
Perfluoropropene Peninsular ChemResearch
Trifluoroacetonitrile Peninsular ChomResearch
Trimethylchlorosilane Peninsular ChemResearch
Perfluoroglutaryl chloride Hooker Electrochemical Co.
The fluorocarbon solvents, FC 102 (a mixture of fluoro-
carbon ether) and FCN-43,(C4F )3N, are products of the Minne-
sota Mining and Manufacturing Company.
The following compounds were synthesized for this work
by the literature method given:
CF3COC1 This compound was prepared by the method of
Henne17 from CF3COOH and benzoyl chloride.
CF COBr As above, this compound was prepared from the
appropriate benzoyl halide, as described by Tinker.
(CF3CO) 0 This compound was prepared in the manner of
Clark and Simons,9 from CF3COOH and P205.
(CF-) N-Hg-N(CF2 This mercurial was prepared by the
method described by Young50 from the reaction of HgF2 and
CF2(CF2CO)2NH Perfluoroglutarimide was synthesized
by the method of McBee35 from perfluoroglutardiamide.
The following compounds were commercial products that
were further purified prior to use by the means given:
Acetonitrile was dried over calcium chloride and frac-
tionated, the fraction boiling from 81.6-82.00 was used.
Acetic acid was fractionated, the cut boiling from
118-118.50 was taken.
Trifluoroacetic acid was fractionated from a small
amount of P205, the cut boiling from 72-72.5o was used.
Most of the materials used and made in this work dis-
played varying degrees of ease of hydrolysis by atmospheric
moisture. In general, all the materials were protected
from moisture by carrying out the appropriate manipulations
in a dry-box, or rapidly in the open room. Some of the
materials were so hygroscopic, or reactive toward moisture,
that extreme care had to be taken to insure good results.
Among these were: (CF3)2N-Hg-(CF3)2, (CF3CO)2NH,
BrCF2CF2CF2CONC0, CF2(CF2CO)2NCOC6H5, and probably
[CF2(CF2CO) 2N]2 (COCF2)2CF2.
Volatile materials were handled as a matter of course
in a glass vacuum system, using standard techniques.
1. CF2(a) NH
6. CF2(c) NCOCF2 CF2
L(d)CF2CO (a) 2(b)
magnetic Resonance Spectra
Nuclei Peak Area Chemical Fine
a 4 47.0 ppm triplet
F b 2 60.6 pentet
TABLE 2 (continued)
Nuclei Peak Area Chemical Fine
7. (CF CO)2NCH3
10. CF3CON(CH3)FCF HCF3
11. 2 CF3COOH-C4Hg0
12. 3 CF COOH*2
* Two slightly nonequivalent quartets.
** Complex spectrum; a full interpretation is not yet complete but
all evidence presently available fits structure given.
Infrared Spectra Used in this Work
Principal Absorptions (Microns)
5. (CF3C0)2 CH3*
2.86(m), 3.1-3.2(m), 4.28(w),
5.64(s), 5.74(s), 6.25(m), 6.75(w),
8.15(s), 8.25(s), 10.-10.l(w),
3.0-3.1(m), 3.33(w), 5.60(s),
5.79(s), 6.6(m), 7.05(w), 7.27(m),
7.60(m), 8.22(s), 8.65(s), 9.79(m),
10.85(w), 11.16(w), 13.62(m).
3.00(s), 3.38(m), 5.79(s), 5.95(s),
6.7(s), 7.05(m), 7.30(s), 8.38(s),
3.02(s), 3.22(w), 3.40(w), 5.78(s),
6.35(m), 7.05(m), 7.40(w), 7.60(w),
8.50(s), 11.36(w), 13.33(w),
3.4-3.5(w), 5.68(s), 5.74(s),
6.83(m), 7.30(s), 7.55(m), 8.60(s),
9.38(s), 9.92(m), 11.80(w),
13.05(m), 13.29(w), 13.93(s).
TABLE 3 (continued)
Principal Absorptions (Microns)
5.73(s), 6.73(w), 7.18(m),
7.41(m), 7.72(s), 7.93(s),
8.60(s), 8.93(m), 9.28(s),
, 11.72(m), 12.60(w),
, 13.90(s), 14.35(s).
8. [CF2(cF2o)2N] 2 Hg*
9. CF2(CF2CO) 2NAg*
3.02(m), 3.50(w), 5.59(m), 5.70(s),
7.07(w), 7.50(m), 8.00(s), 8.13(s),
8.60(s), 8.95(s), 9.55(s), 10.l1(m),
6.0(s), 6.38(w), 7.20(m), 7.30(w),
7.55(w), 7.75(m), 8.05(w), 8.30(m),
8.70(s), 9.18(w), 9.54(s), 10.22(s),
10.58(w), 10.00(w), 10.23(w),
5.90(s), 6.20(s), 6.55(m), 7.50(m),
7.80(s), 8.2-8.7(s), 9.1(s),
9.65(s), 10.2(s), 11.13(w),
11. 2 CF3COOH'C4H80*
12. 3 CF3COOH.
rABLE 3 (continued)
Principal Absorptions (Microns)
4.80(m), 5.8(m), 6.00(s), 6.87(m),
7.29(m), 8.3-8.7(s), 11.8(w),
2.8o(w), 3.35-3.45(s), 3.7(m),
4.0(s), 5.1(m), 5.64(s), 6.89(w),
7.35(m), 7.55(m), 8.25-8.6(s),
9.62(s), 10.90(w), 11.40(m),
12.3(m), 12.82(m), 14.30(s).
3.4(m), 3.72(w), 4.0(m), 5.1-
5.2(w), 5.60(s), 6.92(w), 7.21(w),
7.7(w), 8.22(s), 8.58(s), 9.14(m),
9.4(m), 10.3-10.4(w), 12.0(w),
12.3(w), 12.7(w), 14.32(m).
6.35(s), 7.06(s), 7.55(m), 7.70(m),
7.80(m), 8.00(s), 8.39(s),
7.60(s), 7.87(s), 8.02(s), 8.15(s),
7.46(m), 7.70(s), 8.10(s), 8.40(s),
11.26(m), 13.45(m), 13.85(s).
22. Br(CF2)3CONCO *
TABLE 3 (continued)
Principal Absorptions (Microns)
2.85(m), 6.63(s), 7.40(s), 7.90(s),
8.30(s), 8.72(s), 10.5(m),
11.32(w), 13.50(w), 14.7(m).
5.30(s), 7.27(s), 7.60(s), 8.10(s),
8.20(m), 9.68(m), 10.05(s),
5.55(m), 7.35(s), 7.65(s), 7.90(s),
8.10(s), 8.35(m), 8.55(m), 8.80(m),
9.96(m), 11.25(m), 13.6(m).
5.51(m), 7.50(s), 7.71(s), 7.90(s),
8.31(s), 9.9-10.0(s), 13.9(m).
4.39(m), 6.89(m), 8.30(s), 8.70(s),
5.90(s), 6.09(s), 7.19(m), 7.53(m),
7.78(s), 8.16(m), 8.27(m), 8.65(s),
9.6(m), 10.13(s), 12.11(m),
4.39(s), 5.50(s), 6.5-6.6(m),
6.98(m), 8.35(s), 8.85(m), 9.35(m),
10.1(m), 12.4-12.6(m), 14.25(m).
23. Br (CF2) 3CONHC00-
26. CF2(CF2CO)2NCOCF2 -
E 3 (continued)
Principal Absorptions (Microns)
2.85(w), 2.96(m), 3.23(w), 3.32(w),
5.50(s), 5.69(s), 6.28(w), 6.05(s),
6.33(m), 6.71(m), 7.6-9.0(s),
9.55(m), 9.70(m), 9.92(m), 10.3(m),
10.8(w), 11.2(m), 11.73(m),
12.01(m), 12.43(s), 12.93(m),
2.97(m), 3.12(m), 5.90(s), 6.20(w),
7.87(m), 8.5(s), 8.75(s), 9.03(s),
9.86(w), 12.21(w), 13.08(w),
3.22(w), 5.56(s), 5.83(s), 6.25(m),
6.70(m), 6.90(m), 7.5(s), 7.9-
8.8(s), 9.6(s), 10.05(s), 10.5(s),
11.67(m), 12.1(m), 12.95(S),
13.4(m), 14.2(m), 14.6-14.8(s).
5.45(s), 5.78(s), 7.5(s), 8.1-
9.0(s), 9.5-9.6(s), 10.0(s),
10.44(m), 9.4-9.6(m), 12.9-14.0(m).
28. I3c CON(CF3)2*
TABLE 3 (continued)
Principal Absorptions (Microns)
7.32(m), 7.79(s), 8.30(s), 8.70(s),
9.10(s), 11.5(m), 13.2(m), 14.4-
3.22(w), 5.60(s), 7.3-7.7(s),
8.0-8.7(s), 9.9-10.0(s), 9.67(m),
3.23(w), 5.50(s), 5.77(s), 6.25(m),
6.90(m), 7.5-7.7(s), 7.93(s),
8.3-8.5(s), 9.33(w), 9.64(m),
9.86(m), 10.03(s), 10.7(w),
11.32(s), 12.53(m), 12.99(m),
3.34(w), 7.70(w), 7.90(s), 10.92(s),
*Compound previously unknown.
B. Synthesis of Diamides
1. Reactions of acids and nitriles
Both CH COOH and CF COOH were reacted with CH CN and
CF CN at 150 and 2000. The reactions involving the higher
boiling CH3CN were carried out in ordinary glass ampoules,
while heavy-walled ampoules were used for the low-boiling
(-64) CF3CN. The volume of the ampoules was 7-10 ml.
For the reactions with CH CN a solution of equimolar
quantities of the nitrile and the appropriate acid was
prepared. One gram (+ 0.02) quantities of this solution
were weighed into the ampoule, the ampoule frozen in liquid
air, evacuated, degassed, and sealed. The ampoule was
allowed to come to room temperature and then completely
immersed in a constant temperature bath maintained at
150+0.50 for the desired period. Upon removal the tubes
were cooled to room temperature, the tip of the tube broken
open, and the tube weighed. The volatile unreacted starting
materials were removed at 1 mm. Hg and the contents reweighed.
The method was checked by repeating one run using 19 g. of
the CF3COOH-CH3CN solution. The conversion of the small
scale run was 68% and that obtained by distillation of the
19 g. run was 69%.
CF3CONHCOCH3, b.p. 900 at 60 mm. Hg
Anal. calcd. for C4H4F3NO: C, 31.0%; H, 2.58%;
N, 9.03%. Found: C, 31.25; H, 2.76%; N, 8.58%
In the reactions involving CF CN care was taken that
the ampoules used were strain free. Whether this precaution
was important is not known, although there were no losses of
samples due to failure of the ampoules. The acid, or a solu-
tion of catalyst in the acid, was weighed into the ampoule
and then a known amount of nitrile condensed into the ampoule.
The ampoule was degassed, evacuated, and sealed. While still
cold the ampoule was placed in a wire screen cage and after
warming to room temperature the ampoule, in its caGe, was
placed in a furnace maintained at 150+10 for the desired
reaction period. Upon removal the ampoule was frozen in
liquid air, removed from its cage, and opened into the
vacuum system. The change in pressure from the starting
amount of nitrile was used to calculate conversion.
The measurements of the reversion of diamide to nitrile
and acid in experiments starting with pure diamide were made
in similar fashion.
Reactions of Acids and Nitriles
Reaction of acetic acid and acetonitrilea at 150+0.50
time(hrs.) conversionb catalyst
24 30 5% H2S04 by
54 45 5% H2SO4 by
70.5 52 5% H2SO4 by
54 22 5% NaOAc by
132 38 5% NaOAc by
Reaction of diacetamide at 150+0.5o
time (hrs.) % unreactedb
5% H2S04 by wt.
5% H2S04 by wt.
5% XaOAc by wt.
5% NaOAc by wt.
TABLE 4 (continued)
Reaction of trifluoroacetic acid and acetonitrile at 150+0.50
time (hrs.) conversionb catalyst
0.92 33 5% H2SO4 by wt.
3.2 60 5% H2SO4 by wt.
0.92 30 5% KOOCCF3 by wt.
3.2 48 5% KOOCCF3 by wt.
Reaction of N-acetyltrifluoroacetamide at
time (hrs.) % unreactedb
5% I-HSO4 by wt.
5% H2SO4 by wt.
5% KOOCCF3 by wt.
TABLE 4 (continued)
Reaction of trifluoroacetio acid and trifluoroacetonitrile
5% H2SO4 by wt.
5% H2S04 by wt.
5% KOOCCF3 by wt.
5% KOOCCF3 by wt.
Reaction of bis-(trifluoroacet)amide at 150+0.50
time(hrs.) % unreactedf catalyst
5% 2S04 by wt.
5% H2SO4 by wt.
TABLE 4 (continued)
Reaction of acetic acid and trifluoroacetonitrile at 150+0.50
time (hrs.) % conversionf catalyst
1 73 10% H2SO4 by wt.
2 87c 10% H2S04 by wt.
1 71 10% NaOAc by wt.
2 85 10% NaOAc by wt.
Acid:nitrile reactions at 2000
system time (hrs.) % conversionO
CH COOH:CH CN 24 45
CF3COOII:CHI3CN 8 50h
CF COOH:CF3CN 8 50
a. The isolated diacetamide fractions were combined and re-
crystallized from hexane to give fluffy white crystals,
m.p. 69.5-70.50 (lit. m.p. 590).38
b. Corresponds to per cent nonvolatile at 1 mm. Hg.
c. Product partially decomposed and insoluble in acetone.
d. Product largely decomposed and insoluble in acetone.
e. Infrared spectrum of recovered, unreacted starting mater-
ial showed greater concentration of acid than in starting
TABLE 4 (continued)
f. Based on unrecovered trifluoroacetonitrile.
g. Conversions are only approximate as decomposition was
extensive at this temperature.
h. Fractionation of the product obtained by reacting 38 g.
of an equimolar mixture under these conditions gave 16
g. of recovered starting materials, 18 g. of a cut boil-
ing from 80-90o at 60 mm. Hg, which, from its infrared
spectrum, was shown to contain some (CF CO) 2H boiling
at about 1000 at 1-2 mm. Hg; and several grams of tarry
2. Reaction of anhydrides and amides
Reaction of (CF3CO)20 and CF3COINJ2. Essentially the
method described by Smith 2 was used. Careful fractionation
of the mixture of CF3COOH and (CF3CO)2NH gave a 76% yield of
the diamide boiling at 1450 and having a malting point of
around 700. The imide when pure was a very hygroscopic
solid melting at 84.5-85.50.
Reaction of (CF3CO)20 and CF3CONHCH3. Using the method
of Smith,2 as above, 25 g. (0.2 moles) of CF3CONHCH3 and 42
g. (0.2 moles) of anhydride were heated together under a
fractionation column on total reflux. After several days
the pot temperature had risen only from 51 to 75. The
contents were cooled and sealed together in a glass ampoule.
After heating at 1200 for two days the ampoule was cooled
and opened. Careful fractionation gave largely recovered
starting materials, but 8 g. of impure material boiling
around 1000 was obtained which was shown to be largely the
desired product, (CF3CO)2NCH13, by comparison with an authen-
tic sample prepared by a different method described later.
Two grams (0.009 moles) of the pure material and one gram
(0.009 moles) of CF3COOH were sealed together in a glass
ampoule. After heating at 1500 for 36 hours the ampoule was
cooled and opened into the vacuum system. An infrared
spectrum showed almost complete reversion to anhydride and
C. Some Reactions of Amides, Diamides,
and their Derivatives
1. The thermal decomposition of some of their metal salts
Thermal decomposition of salts of (CF3CO)gNH. The salts
were made and decomposed in the following manner:
(a) A slurry of sodium sand was prepared by shaking
molten sodium vigorously in xylene in a stoppered Erlen-
meyer flask. When the finely divided sodium had cooled it
was washed repeatedly with dry tetrahydrofuran. To such a
mixture, containing 2.3 g. (0.1 g. atom) of sodium in 300
ml. of tetrahydrofuran, was added 21 g. (0.1 moles) of
(CH3CO)2NH in 50 ml. of tetrahydrofuran. Evaporation of
the solvent left a clear, yellowish-brown syrup. This
material was gradually heated at 1500 over a 12-hour period.
Charring began at 800. After three days at 1300 there was
obtained 2 g. of CF3H condensed in a cold trap. Distillation
of the charred, partially solid residue gave 5 g. of the
triazine, (CF3CN)3. b.p. 950, nD25 1.3195. Known values20
for (CF3CN)3are b.p. 950, nD25 1.3161. Comparison of the
infrared spectrum with that of a known sample confirmed the
(b) The Hg salt of (CF3CO)2NI was prepared by reflux-
ing under a fractionation column a mixti:..~ of 166 g. (0.88
moles) of mercuric acetate at 120 mm.Hg. The CH COOH was
slowly taken off, eventually 27 g. (theory, 24 g.) including
an intercut being removed. After taking off the excess
(CF3CO)2NH and finally drying at 1000 at 0.5 mm.Hgl the
pure salt remained.
Anal. Calcd. for C F12HgN204: Hg, 32.4. Found: Hg,31.4
The dry Hg salt was heated at 200-3400. The only volatile
materials were small amounts of CF3COOH, CF3CN, and some
CF H. The residue consisted of metallic Hg and carbonaceous
Thermal decomposition of salts of perfluoroglutarimide.
(a) The sodium salt of perfluoroglutarimide was pre-
pared by mixing, with cooling, equimolar quantities of the
imide and sodium methylate, in methanol solution, then re-
moving the methanol by evaporation under reduced pressure.
It was stable at 1200 but decomposed rapidly at 140o to a
(b) The Hg salt of perfluoroglutarimide was prepared
in the same manner as described for the synthesis of the
mercury salt of (CF3CO)2NH.
Anal. Calcd. for C10F12HgN204: Hg, 31.2. Found: Hg,31.06
The dry mercury salt was heated for six days at 4700. Of 5 g.,
3 g. remained as residue, a portion of which was metallic Hg,
the remainder a charred mass. One g. of imide had distilled
into a receiver and 1 g. of the Hg salt had sublimed out of
the heated portion of the apparatus.
Thermal decomposition of the sodium salt of CF3CON-
HCOCH 3 A solution of 1.2 g. (0.05 g. atom) of Na in 15.5
g. (0.1 moles) of the diamide was heated at 130-1350 for 72
hours with little effect. After 36 hours at 1600 the solu-
tion had darkened considerably and a small amount of CF 31
evolved into a cold trap. Distillation gave 8 g. of
CF3CONH2. This represents a 70% yield based on diamide,
or 140% based on sodium. The residue was an intractable
material having the consistency of asphalt.
2. Reactions with perfluoropropene
CF3CONH(CH3). An excess of CH3NH2 was bubbled into
CF3CO2C2H5 at room temperature. The C2HO5H which was formed
was fractionated off and the product distilled over at 157-
160. Yields in the neighborhood of 80% were easily obtained.
The product solidified in the flask as it distilled, m.p.
Anal. Calcd. for C3H4F3NO: C, 28.4%; H, 3.15%;
N, 11.03%. Found: C, 28.6%; H, 3.38%; N, 10.86%.
C3CON(CH3)CF2CFHCF One hundred and sixty-eight grams
(1.3 moles) of CF3CONH(CH3) was heated with 4 g. (0.17 g.atom)
of metallic sodium until it had dissolved. The solution of
sodium salt was reacted with 40 g. (0.25 moles) of CF3CF=CF2
in a rocking autoclave at 800 for 40 hours. Upon cooling
and opening the bomb into the vacuum system essentially no
olefin was recovered. Fractionation gave 52 g. (75%) of
adduct boiling at 115-1160, n 28 1.3170, 98-99% pure by
vapor phase chromatography. Upon standing HF was evolved
and both the infrared and nuclear magnetic resonance spectra
indicated the formation of an unsaturated material. The
adduct was allowed to stand six months, then a small amount
of Br2 in CC14 was added but no decolorization occurred.
Attempted reaction with CF CONH2. Twenty-eight grams
(0.25 moles) of CF3CONHg, 1.2 g. (0.05 g. atom) of sodium
were heated and rocked at 800 in an autoclave overnight.
After cooling, 25 g. (0.16 moles) of CF3C=CF2 was added
and the mixture rocked and heated at 800 for 18 hours.
Upon cooling and opening the autoclave into the vacuum
system, all the olefin was recovered unchanged. The residue
in the autoclave was considerably decomposed. Repeating the
reaction using 6 g. (0.25 g. atom) of sodium made no differ-
ence in the results.
Attempted reaction with perfluoroglutarimide. Forty
grams (0.18 moles) of perfluoroglutarimide and 2 g. (0.09
g. atom) of sodium were heated together at 1200 but before
solution was complete extensive decomposition and carbonation
occurred. When.a gram or so of sodamide was added to the
imide the mixture sparked and burned. The sodium salt was
best prepared by reacting the imide with sodium methylate
and removing methyl alcohol until only the dry salt remained.
In this way 8.8 g. (0.04 moles) of imide was mixed with 0.01
moles of methylate in 2.5 cc of methyl alcohol. After re-
moval of the alcohol by evaporating under reduced pressure
the sodium salt was sealed in a glass ampoule with 3 e.
(0.02 moles) of CF CF=CF2. The ampoule was heated and
rocked at 85-1000 for 12 hours. On cooling and opening
the ampoule into the vacuum system the olefin was recovered
unchanged. The clear, syrupy residue in the ampoule was
soluble in water#
Attempted reaction with (CF CO)2NH. Thirty grams
(0414 moles) of (CF3CO)2NH, 1.5 g. (0.06 g. atom) of sodium,
and 6 g. (0.04 moles) of CF3CF=CF2 were reacted in an auto-
clave in the same manner as in the synthesis of CF3CON(CH3)
CF2CFHCF3 The CF2CF=CF2 was recovered unchanged and frac-
tionation of the residue gave 7 g. of the triazine, (CF3CN)3.
3. Reactions with acyl halides
(a) In tetrahydrofuran- A slurry of sodium sand was
prepared by shaking molten sodium vigorously in xylene in
a stoppered Erlenmeyer flask. When the finely divided sodium
had cooled it was washed repeatedly with dry tetrahydrofuran.
To such a mixture, containing 2.3 g. (0.1 g. atom) of sodium
in 300 ml& of tetrahydrofuran, was added, with stirring and
cooling, 13.7 g. (0~1 moles) of CF3CONII(CH3) dissolved in
a small amount of tetrahydrofuran. Then 13*2 g. of CF3COC1
(0.1 moles) was condensed into the cooled solution which was
stirred overnight and then brought to reflux. Filtration
and fractionation gave, after the solvent was removed, 5 g.
(25%) of (CF3CO)2NCI3, b.p. 118.
Anal. Calcd. for C511'F6NO2: C, 26.90; H, 1.34%;
N, 6.28%. Found: C, 27.2%; H, 1.52%; N, 5.99%.
(b) In excess CF3CONH(CH3) -
Sixty-eight grams (0.5 moles) of CF3CONH(CH3), 2.3 g.
(0.1 g.atom) of sodium, and 15 g. (0.11 moles) of CF3COCI
were reacted in an autoclave in the same manner as in the
synthesis of CF3CON(CH3)CF2CFHCF3 except that CF3COC1 re-
placed the CF CF=CF2, Fractionation of the semi-solid
product gave 7 g. (25%) of (CF3CO)2NCH3, b.p. 116-1180.
Fifty grams of CF3CONH(CH3) were recovered. Since no CF3COC1
was recovered, it is likely that it had leaked from the auto-
clave during the reaction period.
(c) In pyridine- Six and six-tenths grams (Oi05 moles)
of CF3CONH(CH3), 4.0 g. (0.05 moles) of pyridine, and 6.6 g.
(0.05 moles) of CF3COC1 in 25 ml. of tetrahydrofuran were
sealed in a glass ampoule. Upon warming to room temperature
a precipitate formed* After standing overnight the ampoule
was cooled and opened, then filtered and fractionated. After
the solvent was removed 3 g. (30%) b.p. 115-125o was obtained
which was largely (CF3CO)2NCH3, as shown by comparison of the
infrared spectrum with that of an authentic sample.
Attempted reactions of CF3COC1 with metal salts of
(a) Using the procedure that was successful previously
in the synthesis of (CF3CO)2NCH3, 4.6 g. (0.2 g.atom) of
sodium sand, 500 ml. of tetrahydrofuran, 42 g. (0.2 moles)
of (CF3CO)2NH, and 26.4 g. (0.2 moles) of CF3COC1 were re-
acted. Subsequent filtration and fractionation gave 30 g.
b.p. 123.5-124.5. As shown by infrared and nuclear magnetic
resonance, this was the tetrahydrofuranoate of CF3COOH. The
empirical formula is 2 CF3COOH.C4HgO, according to relative
absorption of the protons in the nuclear magnetic resonance
(b) The mercury salt was finely powdered and 30 g.
(0.05 moles) was heated with 6.6 g. (0.05 meles) of CF3COC1
along with 75 g. of (CF3CO)20 in a rocking autoclave at 800
for 24 hours. After cooling the autoclave and opening it
into the vacuum system, the volatile product was removed.
The bomb was evacuated for 12 hours, any additional volatile
material being trapped in a liquid air-cooled trap. Distil-
lation of the combined volatiles gave only recovered starting
Attempted reactions of CF3COC1 with perfluoroglutarimide
and several of its metal salts.
(a) Forty-four grams (0.2 moles) of the sodium salt of
perfluoroglutarimide, prepared by evaporation of methyl
alcohol under vacuum from a solution of equimolar amounts
of the imide and sodium methylatewas taken up in dry ether.
This solution was cooled while reacted with 26.4 g. (0.2
moles) of CF3COC1. A gelatinous precipitate of NaCl formed
rapidly. Stirring was continued for three hours and after
filtration the solution was fractionated. A product boiling
at 105-1080, 6 g., was obtained and after 10 g. of an inter-
cut 30 g. of recovered imide was taken over at 900 mm. Hg.
The 105-1080 product was subsequently shown to be the ether-
ate of CF3COOH, reported b. p. 102. The empirical formula,
also previously reported, 3 CF COOH'2 (C2H5)20 was confirmed
according to the relative proton absorption in the nuclear
magnetic resonance spectrum.
(b) Repetition of the above reaction using tetrahydro-
furan as solvent resulted in the formation of the previously
described tetrahydrofuranate of CF3COOH.
(c) Eleven grams (0.05 moles) of perfluoroglutarimide
was dissolved in a solution of 5.8 g. of Ag20 in 100 cc of
CF COOH. After condensing 6.6 g. (0.05 moles) of CF3 COC
into the above solution it was allowed to stand for several
days at 10-200. The solution was filtered from the voluminous
precipitate which had formed and distilled to give only
several grams of (CF3CO)20, CF COOH, and most of the starting
The above quantities of Ag20, CF3COOH, and imide were
mixed as above and then the CF3COOH fractionated off, the
last traces under reduced pressure. Twenty-two grams of
additional imide were added to its dry silver salt to form
a thick slurry which was reacted with 6.6 g. (0.05 moles)
of CF3COC1 in a rocking autoclave for two days at 500.
Distillation of the product gave only starting materials
in addition to some CF3COOH.
Attempted reactions of perfluoroglutaryl chloride with
the mercury salt of perfluoroglutarimide.
(a) Twenty-five and six-tenths grams (0.04 moles) of
the mercury salt, 9.1 g. (0.04 moles) of perfluoroglutaryl
chloride, and 50 ml. of FC 102 were stirred together vigor-
ously in a creased flask at reflux (ca. 1000). After three
days the solvent was fractionated off, along with unreacted
chloride from which separation was impractical, and the
residue sublimed onto a cold finger at 1 mm. HIg. This sub-
limate consisted only of the mercury salt and some perfluoro-
(b) In trifluoroacetic acid- Sixteen grams (0.025 moles)
of the mercury salt and 6 g. (0.025.moles) of perfluoroglut-
aryl chloride were heated at reflux temperature in 75 g. of
CF3COOH. The mercury salt was slightly soluble at this
temperature and, upon cooling, the salt recrystallized.
After three days of refluxing the mixture was cooled and
filtered. Fractionation gave recovered CF3COOH, 5 g. of per-
fluoroglutarimide and 1 g. of dark residue which was discarded.
4. Reactions of metal salts of diamides with bromine and
N-bromoperfluoroglutarimide. Bromination of the silver
salt of perfluoroglutarimide in CF3COOU7 by the method of
Henne,.8 gave yields of 70% when carried out in the cold.
The product could be distilled at 500 at 2 mm. Hg into a
cooled receiver in which it would solidify. The purity was
determined by titrating with thiosulfate the 12 formed from
addition of the N-bromo compound to aqueous KI. The purity
varied from 85-95%, the only identifiable impurity being
unreacted perfluoroglutarimide. The mercury salt could be
added to CF3COOH and brominated to give comparable results.
Attempted brominations of metal salts of(CF3CO)2NH.
Using the procedure of IIenne, two volatile products were
obtained. The first was an easily sublimed solid and the
second, secured in much lesser amounts, was a liquid boil-
ing at 50-600 at 2 mm. Hg. Titration of the iodine liberated
from iodide gave the following results for positive Br:
Calod. for (CF3CO)2NBr: 27.8%
Calcd. for CF CONHBr: 42.1%
Found: solid, 33.1%; liquid, 22.8%
The solid product gave CF3CONH2, along with a small amount
of (CF3CO)2NH, on reaction with anhydrous HBr. A slurry of
the dry Hg salt of (CF3CO)2NH in CF3COOH did not absorb Br2,
nor was any absorbed in the above method if sufficient
(CF3CO)20 was added to insure that no water was present.
Substitution of metallic sodium for Ag20 had no effect either.
Attempted synthesis of N-iodoporfluoroglutarimido. Sub-
stitution of I2 for Br2 in the above reaction resulted in a
much slower reaction, as evidenced by the lesser rate of
formation of silver halide as the reaction proceeded. Even
after four days of stirring in the dark only 50o of the
theoretical amount of AgI had formed. Filtration and evapor-
ation under vacuum resulted in the removal of CF3COOH and 12,
which seemed to be constantly formed. Finally, 10% of theory
of material which boiled at 80-90O at 2 mm. Hg was obtained.
Even in the dark this material rapidly decomposed, 12 being
formed. The only identifiable substance left after removal
of the 12 was perfluoroglutarimide.
5. Indirect fluorinations of amides and diarnides
Fluorination of (CF3CO)2NH. Thirty grams (0.21 moles)
of AgF2 was suspended in 60 ml. of FC 102 and stirred while
10 g. (0.05 moles) of (CF3CO)2NH was added slowly, out of
contact with the atmosphere, gaseous products being collected
in a liquid air-cooled trap. The reaction temperature was
maintained at 600 by adjusting the rate of addition, then
raised to reflux (ca. 1000) for a short period of time.
The condensate in the trap amounted to 5 g. and had a mol.
wt. of 103-122. Its infrared spectrum showed all the lines
for CF COF and CF3 CO, plus traces of unidentified products.
Attempted fluorination of perfluoroglutarimide. Sixty-
five grams (0.45 moles) of AgF2 was suspended in 100 ml. of
(C4F9)3N and stirred while 35 g. (0.15 moles) of imide was
added at reflux (ca. 1600). Extensive etching of the glass
vessel indicated the formation of HF. The reaction of the
black AgF2 to form a tan product was also evident, as was
the evolution of heat. No volatile products were obtained.
Since there was excess AgF2 present, no attempt was made to
isolate the solid product except for enough to obtain an
infrared spectrum. Two grams (0.009 moles) of perfluoro-
glutarimide and 1 g. (0.0045 moles) of IF5 were heated at
reflux for six hours. Iodine was formed on the walls of the
vessel but no volatiles were obtained. The rather viscous
residue was quite clear and only slightly colored. An infra-
red spectrum showed no bands not attributable to the starting
Fluorination of tF3)2NCOF. Thirty grams (0.2 moles) of
AgF2 and 20 grams (0.1 moles) of (CF3)2NCOF were rocked in
an autoclave at 100o for 18 hours. The volatile contents
were then fractionated to give 7.5 g. of overhead at Dry Ice
temperature, mol. wt. 68.5-81.6, largely COF2, mel. wt.
66;10 g. (60%) of (CF3)2NF, b.p. -370; and 4 g. of column
holdup which was shown from its infrared spectrum to be a
mixture of (CF3)2NCOF and (CF3)2NN(CF3)2. Passing through
aqueous alkali gave 2 g. of relatively pure hydrazine.
Fluorination of (CF3)2NCOCF3. Fifteen grams (0.1 moles)
of AgF2 and 10 g. (0.04 moles) of (CF3)2NCOCF3 were rocked
in an autoclave at 1000 for 24 hours. Fractionation of the
volatile product was only partially successful because of
the small amount, but the following amounts of products
were estimated from the infrared spectra of the various
cuts: 3.5 g. of CF3COF, 2 g. (30%) of (CF3)2NF, 3 g. (45%)
of (CF3)2NN(CF )2, and 1.5 g. of (CF3)2NH which was probably
present as an impurity in the starting material.
6. Some reactions of N-bromoperfluoroglutarimide
The free-radical rearrangement of N-bromoperfluoro-
glutarimide. Fifteen grams (0.05 moles) of N-bromo com-
pound and 5 g. (0.033 moles) of CF3CF=CF2 were sealed to-
gether in a heavy-walled ampoule and left in direct sunlight
for four days. The original yellowish color disappeared
after the first day and there was no further change. Dis-
tillation of the product resulted in recovery of all the
olefin and a quantitative yield of BrCF2CF2CF2CONCO, b.p.
420 at 30 mm. Hg. This material reacted almost explosively
with water and a drop placed on a watchglass exposed to the
atmosphere turned solid in just a few seconds. The aqueous
solution gave no precipitate with AgNO3.
The above quantities were reacted in the presence of
1 g. of benzoyl peroxide in an autoclave at 1000 to give
When 5 g. (0.011 moles) of the N-bromo compound and
2.5 g. (0.02 moles) of CF2=CFCl were sealed in a heavy-walled
ampoule and placed in direct sunlight for several hours the
yellow color disappeared. The ampoule was cooled and opened
into the vacuum system to recover the olefin quantitatively
and give 5 g. of the acyl isocyanate obtained previously.
When (CF3CO)20 was substituted for the CF2=CFC1 in the
above experiment no reaction occurred even after two days in
direct sunlight except for the formation of some Br2. The
only other identifiable material was the unchanged N-bromo
Fifteen grams of the N-bromo compound was sealed by
itself in an ampoule and left in direct sunlight for three
days, after which time what seemed a considerable amount of
Br2 had formed, but fractionation after removal of the Br2
gave only the N-bromo compound, 80% being recovered along
with some of the perfluoroglutarimide.
BrCF CF2CF CONHCOCH C 2p The isocyanato from above was
allowed to stand two weeks in an ether solution of benzyl
alcohol. Evaporation of the solvent left a solid residue
which gave, upon recrystallization from cyclohexane, fluffy
crystals, m.p. 76.5-78.50.
Anal. Calcd. for C12HgBrF6NO3: C, 35.2%; H, 2.0%;
N, 3.-4. Found: C, 35.7%; H, 2.3%; N, 3.0%.
BrCF2CF CF2CONH2. The isocyanate, 3 g., was added
dropwise to an excess of water with which it reacted vio-
lently. The white solid which formed was filtered, washed
thoroughly with dilute NaHC03, then water. After drying
and recrystallizing from benzene, 1.5 g. of solid melting
at 98-1000 was obtained.
Anal. Calcd. for C4H2BrF6NO: N, 5.11%; Br, 29.2%.
Found: N, 5.00%; Br, 26.90, 26.78%.*
Reaction with HBr. Two grams of the N-bromo compound
and 1 g. of HBr were sealed together in a glass ampoule and
allowed to warm to room temperature. Bromine was formed
immediately and after an hour the ampoule was opened into
the vacuum system. Perfluoroglutarimide remained after the
volatiles were removed.
C6H5CON(COCF21_CF9. Thirteen grams (0.043 moles) of
the N-bromo compound and 7.5 g. (0.043 moles) of C65cCOBr
were mixed together in a distilling flask. The mixture
warmed immediately and Br2 was formed. After heating at
900 under about 45 mm. Hg for four hours, the evolution
of Br2 seemed complete and the resulting clear solution was
fractionated. After 5 g. of forerun there was obtained 8 g.
(60%) of product boiling from 85-86 at 2.5 mm. Hg. This
* While the reason for the low bromine analysis is not known,
sometimes complete decomposition of fluorocarbon bromides
is quite difficult. The NMR evidence, and all else, is so
conclusive that there can be little doubt that the structure
is that given.
material eventually solidified in the receiver and after
recrystallization from n-pentane melted at 39-410. It was
quite hygroscopic and upon solution in water gave no pre-
cipitate with aqueous AgNO3.
Addition of small amounts of AlBr3 had no effect upon
the course of the reaction, nor did the N-bromo compound
react appreciably with AlBr3 alone.
Anal. Calcd. for C12H5F6NNO3 N, 4.30%.
Reaction with CF3COBr. Fifteen grams (0.05 moles) of
the N-bromo compound and 9 g. (0.052 moles) of CF3COBr were
sealed together in an ampoule and allowed to warm to room
temperature. Even before the contents had completely melted
a considerable amount of Br2 had formed. After standing
overnight the ampoule was opened into the vacuum system and
the volatiles removed. Since an obvious color of Br2 re-
mained, a few ml. of Hg were added but the mixture immediately
became quite hot and additional Br2 was formed which sub-
sequently reacted with the Hg. After several hours the
solid contents were extracted with about 50 ml. of tetra-
hydrofuran. After filtering off the greyish solid and
excess Hg, the solution was evaporated. An infrared spec-
trum of the non-volatile residue showed only the presence
of perfluoroglutarimide and its Hg salt. This residue
weighed 18 g. and appeared to be about 80% Hg salt. .The
volatile product previously obtained was distilled but only
Br2 and recovered CF COBr were obtained.
Reaction with CyF15COBr. The bromide was prepared by
the reaction of C7F15COOH with C6115COBr in 55% yield, b.p.
Anal. Calcd. for C8Fl50Br: Br, 16.8o. Found: 16.39.
Ten grams (0.35 moles) of the N-bromo compound and 17 g.
(0.035 moles) of C7F15COBr were mixed together. Some Br2 .
was formed immediately but even on heating under reduced
pressure little more was formed. Subsequent distillation
gave starting materials, along with some perfluoroglutarimide,
as the only identifiable products.
CF (CF2CO)-NCO(CF2)3CO (COCF2J2CF2, Twenty-one grams
(0.07 moles) of the N-bromo compound and 9 g. (0.033 moles)
of CICO(CF2)3COC1 were heated together at 1000. When the
color of Br2 became noticeable, a slight vacuum was applied
so as to remove that which had formed. After three days Br2
was no longer being formed and the slightly yellow solution
was distilled under vacuum. There was obtained a cut boil-
ing at 35-45P at 60 mm. Hg. which was largely recovered
N-bromo compound. The product was taken at 95-100o at 1.5
mm. Ig, The yield was 6 g., or 28e.
Anal. Calcd. for C15 18i 206: N 4.34%.
Reaction with Hg. Thirteen grams (0.043 moles) of the
N-bromo compound and 50 g. of Hg were sealed together in
a glass ampoule and allowed to warm to room temperature.
As Br2 was formed initially, considerable heat was evolved.
Eventually the Br2 had all reacted with the Hg and the
ampoule had cooled. It was then opened into the vacuum
system and heated gently. About a half a gram of perfluoro-
glutarimide was obtained. The residue was extracted with
tetrahydrofuran, filtered and evaporated to obtain a nearly
quantitative yield of the Hg salt, identified by means of
its infrared spectrum&
Reaction with Hg(CHFCF3j2. Four grams (0.013 moles)
of the N-bromo compound and 3 g. (0.0075 moles) of the mer-
curial* were heated at 1100 for 24 hours, at which time about
1 g. had collected as overhead in a cold trap. This material
had a mol. wt. of 171 (calcd. for CF CHFBr, 181) and an
infrared spectrum consistent with this structure. Repetition
of the reaction in a sealed tube exposed to sunlight gave
similar results. Sunlight had no effect on the reaction
residue, even after several weeks.
t Furnished by D. E. O'Conner, Department of Chemistry,
University of Florida.
D. The Reactions of Bis-(perfluoro-
CF CON(CF 2. Fourteen grams (0.026 moles) of (CF3)2N-HC-
N(CF3)2 were sealed in a glass ampoule with 8 g. (0.06 moles)
of CF COC1 and allowed to warm to room temperature. At this
time a considerable amount of a white precipitate had formed.
Heating the ampoule at 400 for an additional hour caused no
further noticeable change so the ampoule was cooled and
opened into the vacuum system. Approximately 10 g. (80%)
of CF3CON(CF3) was obtained in a relatively pure state out
of 13.5 g. of condensate. The infrared spectrum matched
that of a sample obtained previously from the electrochemical
process.51 The remainder, which was more volatile on dis-
tillation, was mainly unreacted CF3COCI and (CF3)2NH.
CH3CON(CF32 Twelve grams (0.022 moles) of (CF3)2 Ng-
N(CF3)2 and 3.1 g. (0.04 moles) of acetyl chloride were
reacted and distilled as in the synthesis of CF3CON(CF3)2.
There was obtained 5 g. (62%) of product boiling at 66,
mol. wt. 195, found 194.
Anal. Calcd. for C4H3S6NO: N, 7.22%
C26hCON(CFq2g. Ten grams (0.02 moles) of (CF3)2N-Hg-
N(CF3)2 and 5.4 g. (0.04 moles) of benzoyl chloride were
allowed to warm to room temperature together in a distilling
flask. Subsequent distillation gave 10 g. (95%) of product
boiling at 950 at 90 mm. Hg.
Anal. Calcd. for C9H5F6NO: N, 5.40%.
Reaction of (CF312N-Hg-N(CF3 1 with COC12. Twelve grams
(0.022 moles) of (CF3)2N-Hg-N(CF3)2 and 2 g. of COC12 were
sealed together in a glass ampoule and allowed to warm to
room temperature. There was no change in appearance of the
contents and after several hours without a precipitate being
formed, the ampoule was heated at 1000. After five days the
ampoule was allowed to cool to room temperature and then
opened into the vacuum system. About 6 g. of volatile prod-
ucts were obtained but these were shown, from the infrared
spectrum, to be a mixture of CF3N=CF2, COC12, and some (CF3)2NH.
The relatively non-volatile residue was unreacted mercurial.
Reaction of (CF3)2N-Hg-N(Crj)p with (CH3g3SiCl. Twenty-
one grams (0.041 moles) of (CF3)2N-Hg-N(CF3)2 and 8.7 g.
(0.08 moles) of (CH3)3SiCl were allowed to warm to room
temperature together in a distilling flask. A solid pre-
cipitate formed, as in previous reactions, but the products
were gaseous, passing through the distilling column and
being condensed in a trap cooled in dry ice-acetone. Eighteen
grams were obtained that were shown by their infrared spectrum
to be a mixture of CF3N=CF2 and (CH3)3SiF.
The syntheses and reactions of some fluorocarbon acyl-
amino compounds and their derivatives have-been studied.
The reactions of acetic and trifluoroacetic acid with
aceto- and trifluoroacetonitrile to give the diamides,
N-acetyltrifluoroacetarnido and bis-(trifluoroacet)amide,
were carried out and the relative rates of reaction of the
various systems determined at 1500C. Two mechanisms appear
to be followed, the actual choice being dependent upon the
nitrile used. The reactions of the acids with acetonitrile
were catalyzed by acid and these with trifluoroacetonitrile
were both acid and base catalyzed. These reactions were
shown to reach equilibria which were studied by approaching
them from the opposite direction, ie., reversion of the
diamides to the appropriate acid and nitrile. The effect
of substituting methyl for trifluoromethyl groups on the
position of equilibrium is discussed.
The reactions of trifluoroacetamide and N-methyltri-
fluoroacetamide with trifluoroacetic anhydride were studied
but only in the former case were good yields obtained of
the trifluoroacetylated product.
The reactions of metal salts of trifluoroacetamide,
N-methyltrifluoroacetamide, bis-(trifluoroacet)amide, and
perfluoroglutarimide with perfluoropropene were studied;
only N-methyltrifluoroacetamide giving an addition product,
CF3CON(CH3)CF2CFHCF3. The metal salts of bis-(trifluoro-
acet)amide and perfluoroglutarimide failed to react with
acyl chlorides, but again the N-methyl compound reacted in
the desired manner to give N-methyl- bis-(trifluoroacet)-
amide. The sodium salt of bis-(trifluoroacet)amide decom-
posed at 100I C to give 2,4,6-tris(trifluoromethyl)-l,3,5-
N-bromo-bis-(trifluoroacet)amide could not be made by
means of the bromination of the silver salt in trifluoro-
acetic acid, although N-bromoperfluoroglutarimide could be
Fluorination of fluorocarbon acylamino compounds with
silver difluoride proceeds in such a manner as to:
(1) Abstract the hydrogen atom(s) from the nitrogen.
(2) Cleave the carbonyl-nitrogen bond to give the acyl
fluoride and a nitrogen species which may either rearrange,
dimerize, or react with fluorine to form an K-F compound.
Thus, bis-(trifluoroacet)amide, p*rfluoro-N,N -dimethyl-
carbamyl fluoride, and perfluoro-N,N -dimethylacetamide
were fluorinated. Attempted fluorination of perfluoroglut-
arimide stopped after evolution of hydrogen fluoride and
the formation of a silver salt.
N-bromoperfluoroglutarimide rearranges under free
radical conditions, in the presence of perfluoropropene
or chlorotrifluoroethylene, to give -bromoperfluorobuty-
The first fluorocarbon triacyl-nitrogen compound,
NN,N ,N -bis(perfluoreglutaryl)perfluoroglutardiamide,
has been made from the reaction of perfluoroglutaryl
chloride and N-bromoperfluoroglutarimide. The N-bremo
compound reacts readily with benzoyl bromide to give
N-benzoylperfluoroglutarimide, and with hydrogen bromide
to give perfluoroglutarimide. No triacyl compounds were
isolated from the reactions with trifluoroacetyl bromide
or perfluorooctanoyl bromide.
A new synthesis of N,N-bis-(trifluoromethyl) aides
by means of the reaction of bis-(perfluorodimethylamino)-
mercury with acyl chlorides is reported. The acyl chlorides
used were trifluoroacetyl, acetyl, and benzoyl. Neither
phosgene nor trimethylchlorosilane gave the desired amino
1. Attaway, J. A., Groth, R. H., and Bigelow, L. A., J.
Am. Chem. Soc., 81, 3599 (1959).
2. Avonda, F. P., Gervasi, J. A., and Bigelow, L. A.,
ibid., 78, 2798 (1956).
3. Barr, D. A., and Haszeldine, R. N., Chem. and Ind.,
4. Barr, D. A., and Haszeldine, R. N., J. Chem. Soc., 30,
5. Brown, H. C., and Reilly, W. L., 128th Meeting A.C.S.,
Minneapolis, Minn., September 11-16, 1955.
6. Brown, H. C., and Reilly, W. L., J. Am. Chem. Soc.,
78, 6032 (1956).
7. Brownstein, S. K., J. Org. Chem., 23, 113 (1958).
8. Cady, G. H., et al., Ind. Eng. Chem., 39, 290 (1947).
9. Clark, R. F., and Simons, J. II., J. Am. Chem. Soc., 75,
10. Davidson, D., and Skovronek, H., ibid., 80, 376 (1958).
11. England, D. C., et al., ibid., 82, 5116 (1960).
12. Fowler, R. D., et al., Ind. Eng. Chem., j9, 292 (1947).
13. Fukuhara, N. and Bigelow, L. A., J. Am. Chem. Soc., 63,
14. Gould, E. S., "Mechanism and Structure in Organic Chem-
istry," Henry Holt and Company, New York, 1959,
15. Hanford, W. E., and Rigby, G. W., U.S. Patent 2,409,274
16. Hauptschein, M., and Grosse, A. V., J. Am. Chem. Soc.,
2I, 5139 (1951).
17. Henne, A. L., Aim, R. M., and Smook, M., ibid., 70,
18. Henne, A. L., and Zimmer, W. F., ibid., 73, 1103 (1951).
19. Hentschel, W., Ber., 2395 (1890).
20. Hine, Jack, "Physical Organic Chemistry," McGraw Hill
Book Company, New York, 1956, p. 317.
21. Homer, L., and Windelmann, E. H., Angew. Chem., 71,
22. Johnson, H. W., and Bublitz, D. E., ibid., 79, 753
23. Johnson, H. W., and Bublitz, D. E., ibid., 80, 3150
24. Kriegsmann, H., Z. anorg. u. allgem. Chem., 294, 113
25. Lewis, J., and Wilkins, R. G., "Modern Coordination
Chemistry," Interscience Publishers, New York, 1960.
26. Lovelace, A. M., Rausch, D. A., and Postelnek, W.,
"Aliphatic Fluorine Compounds," Reinhold Publishing
Corp., New York, 1958, p. 7.
27. Lovelace, A. M., Rausch, D. A., and Postelnek, W., ibid.,
28. Lovelace, A. M., Rauseh, D. A., and Postelnek, W., ibid.,
29. Lovelace, A. M., Rausch, D. A., and Postelnek, W., ibid.,
30. Lovelace, A. M., Rausch, D. A., and Postelnek, W., ibid.,
31. Lovelace, A. M., Rausch, D. A., and Postelnek, W., ibid.,
32. Lovelace, A. M., Rausch, D. A., and Postelnek, W., ibid.,
33. Lovelace, A. M., Rausch, D. A., and Postelnek, W., ibid.,
33a. Mao, T. J., Ph.D. Dissertation, University of Florida,
34. Martin, J. C., and Bartlett, P. D., J. Am. Chem. Soc.,
78, 2553 (1957).
35. McBee, E. T., Ind. Eng. Chem., 39, 415 (1947).
36. Migridichian, V., "The Chemistry of the Organic Cyanogen
Compounds," Reinhold Publishing Corp., New York,
1947, p. 56.
37. Migridichian, V., ibid., p. 61.
38. M1igridichian, V., ibid., p. 62.
39. Moissan, H., Compt. rend., 102, 1543 (1886).
40. Park, J. D., et al., J. Am. Chem. Soc., 74, 2189 (1952).
41. Simons, J. H., Editor, "Fluorine Chemistry," Vol. II,
Academic Press, New York, 1955, p. 389.
42. Smith, G. H., U.S. Patent 2,701,814 (1955).
43. Thompson, Q. E., J. Am. Chem. Soc., p2, 5841 (1951).
44. Tinker, J. M., U.S. Patent 2,255,868 (1941).
45. Walling, C., "Free Radicals in Solution," John Wiley
and Sons, Inc., New York, 1957, p. 48.
46. Walling, C., ibid., p. 381.
47. Wiley, R. H., and Guerrant, W. B., J. Am. Chem. Soc.,
71, 981 (1949).
48. Young, J. A., unpublished results.
49. Young, J. A., Durrell, W. S., and Dresdner, R. D., J.
Am. Chem. Soc., 81, 1587 (1959).
50. Young, J. A., Tsoukalas, S. N., and Dresdner, R. D.,
ibid., 80, 3604 (1958).
51. Young, J. A., Simmons, T. C., and Hoffmann, F. W., ibid.,
78, 5637 (1956).
52. Ziegler, K., et al., Ann., 551, 80 (1942).
William Sandford Durrell was born October 14, 1931,
in Miami, Florida. He was graduated from Ft. Lauderdale
High School in June, 1949. In August, 1953, he received
the degree of Bachelor of Science in Chemistry from the
University of Florida. From that time until February,
1956, he worked as a research chemist at Peninsular Chem-
Research, Gainesville, Florida, and at the Ethyl Corpor-
ation in Baton Rouge, Louisiana. He was then called into
active duty in the United States Air Force and was stationed
at Wright Air Development Center, Wright-Patterson Air Force
Base, Dayton, Ohio. In January, 1958, he enrolled in the
Graduate School of the University of Florida and has held
the position of Research Assistant until the present time,
while pursuing his work toward the degree Doctor of Phil-
William Sandford Durrell is married to the former Mary
Jane Kusbel and is the father of twin daughters. He is a
Captain in the United States Air Force Reserve and is a
member of the American Chemical Society, Alpha Chi Sigma,
Phi Beta Kappa, and Sigma Xi.
This dissertation was prepared under the direction 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 re-
quirements for the degree of Doctor of Philosophy.
June 5, 1961
Dean, College of Arts and Sciences
Dean, Graduate School
Cd- C rman