A STUDY OF THE INFRARED SPECTRA OF
SOME SUBSTITUTED ACETAMIDES
HARRY LETAW, JR.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
TABLE OF CONTENTS
LIST OF TABLES ........................... it
LIST OF FIGURES ....... .. .. ... ... .. .. ili
I INTRODUCTION...................... 1
II EXPERIMENTAL....... ..... .... ..... 20
III EXPERIMENTAL DATA ....,........... 4
IV DISCUSSION ................ ...... 61
V SUMMARY ........................ 93
BIBLIOGRAPHY ........................... 95
ACKNOWLEDGMENTS ...................... 100
BIOGRAPHICAL NOTE ............1......... 101
LIST OF TABLES
I Abbreviations of Compounds . ........ 25
II Three Micron Region, N-dibutyl Series . 26
III Three Micron Region, N-butyl Series ...... 27
IV 5.7 Micron Band, N-dibutyl Series ........ 29
V The Six Micron Region. ..... ........ 30
VI 6.4 Micron Region, N-butyl Series . . . 31
VII Fluorine Stretching Bands . . . .. 33
VIII A Band and B Band Dilution Shifts, N-butyl Series 65
LIST OF FIGURES
1. Noise Level and CC14 Compensation Pattern ....
Z. N-butyl, Pure . . . . . . . . . .
3. N-butyl, 0.004M ....... .. ........ ..
4. N-butylChloro, Pure .................,
Chloro, 0.006 M ..... ..........
Dichloro, Pure .... .. ..........
Dichloro, 0.006 M . . . . ... .
Chlorofluoro, Pure . ... .... .. .
Chlorofluoro# 0.005 M. ...........
Difluoro, Pure . . . . . .
Difluoro, 0.009M ..... .... .
Trifluoro, Pure . . .. . . . .
Trifluoro, 0.004 M .... . ...
N-dibutyl Pure . . . . . . . . . .
N-dibutyl Chloro, P.u .. . .. .. u. .
N-dibutyl Chloro. 0.007 ....... ......
N-dibutyl Dichloro, Pure
N-dibutyl Dichloro, Pure . . . . . .
N-dibutyl Dichloro, 0.008 M . ...M .,...
20. N-dibutyl Chlorofluoro, Pure .... .. 5
ZI. N-dflutyl Chlo2rofluoro. O 0. 007.. M 56
22-0 Xdibuty1E Iiluero. Pure- i #..... 57
23. X-dibutyl Difluoro, 0. 007 ...... 0.. 0 8$
25. N-dibutyl Trifluoro. O.o0 6 ip .o.0........0 60
Infrared spectrophotometry has taken its place among the more
useful tools at the disposal of the investigator studying inter- and intra-
molecular phenomena. This vast field has been thoroughly treated by
G. Herzberg in his two volume work "Molecular Spectra and Molecular
Structure", 3 Barnes and Bonner have briefly reviewed the history and
origin of infrared spectra in a reasonably elementary fashion.
The development of this technique has taken place from two points
of view which have only recently been fused in the literature. The em-
pirical approach is typified in a recent book, "Infrared Determination
of Organic Structures". 58 Theoretical aspects, on the other hand, are
discussed by Dennison in two papers22 23 which have appeared since
the discovery of the quantum mechanics. Barnes, Gorem Liddel and
Williams have assembled a bibliography of the literature of infrared
spectroscopy through the year 1943.9 In cooperation with the American
Petroleum Institute, the National Bureau of Standards is reproducing
reference spectra of a variety of compounds. Spectrograms are
periodically added to this publication
Because of the nature of the compounds dealt with in the present
study, much emphasis must be placed upon the concept of the hydrogen
bond or, as it is often termed, the hydrogen bridge. These two phases
will be considered to be interchangeable.
The phenomena which led to the postulation of the existence of
the hydrogen bond were observed primarily during cryoscopic and
ebullioscopic measurements. It was found that anomalous, or unex-
pected results were frequently obtained in the laboratory. Oddly
enough, the first postulation of the hydrogen bond was by Moore and
Winmill54 who, in 1912, found the concept necessary to explain the
weakness of the basic properties of trimethylammronium hydroxide.
Latimer and Rodebush39 are usually credited with recognizing
the generality of the hydrogen bond. In their development, however,
they erred in that they considered the hydrogen to be dicovalently
bound to its neighbors. Pauling56 found this to be incorrect after the
introduction of the quantum mechanics and placed the hydrogen bond on
a largely electrostatic basis. Though an occasional argument to the
contrary appears in the literature, it has become obvious that Pauling's
concept is basically correct.
Three rather thorough reviews on the general subject of hydrogen
bonding have appeared in fairly recent times. 19, 38, 60 The definitive
work on the subject is Pauling's "Nature of the Chemical Bond".55
Care must be taken in the study of the latter reference since the revision
published in the year 1948 is hardly complete up to that date.
Discussions of the amide group and its intermolecular linkages
have been centered largely about the problem of tautomerism in these
compounds. The origin of the proposal for a keto-enol tautomeric
shift in the amides is deeply bound in the macrochemistry of the group.
It has been found that the reactions of the amides occur in two different
categories. First, attack is possible upon the amide nitrogen itself.
This is typified by the reaction of the amides with rather strong bases.
One of the amine hydrogens can be directly replaced by a sodium atom.
This sodium salt may then be reacted with an alkyl iodide to yield the
N-substituted amide. Intermediately, the silver ion can react as the
sodium ion did or else demonstrate a second mode of attack by adding
to the carbonyl oxygen. When treated with an alkyl iodide, these com-
pounds react in such a way as to produce the N-substituted amide and
the imido ester, respectively. The extreme case appears on the treat-
ment of the free acid amides with dimethyl sulfate. Matsui53 reports
that the alkylation proceeds at temperatures below 1000 C. with the
recovery of the methyl hydrogen sulfates of the imido esters. The
present author believes that this reaction is not directly comparable
to the two previously described in that its mechanism is undoubtedly
The reaction schemes above have demonstrated essentially the
liability of the proton attached to the amide nitrogen. This fact is
startling in view of the apparent availability of the "electron pair" on
the electronegative nitrogen. Comparison of the amides with the
amines led to the rather ill-considered opinion that the former com-
pounds should be basic. The amides are actually feebly acidic except
in the most extremely acid solvents. It is this apparent contradiction
which has led to many studies of the amide linkage.
It became evident that purely chemical methods would hardly suf-
fice for the final resolution of the problem of the structure of the
amide group. Recognition of the fact that this linkage is intimately
related to the overall question of the nature of the proteins and poly-
amides served as a tremendous stimulus to the development and
improvement of means of attack upon this problem.
Kumler and Porter37 applied data obtained from dipole moment
measurements to the problem of the molecular configuration of acetam-
ide, N-ethyl acetamide, N-diethyl acetamide and N-dimethyl acetamide.
Utilizing the method and apparatus of Williams and \7eissberger,72
they calculated the proportion of the classical enolic tautomer present.
It was found that this fraction is vanishingly small even for acetamide
itself. They stated that the true configuration, at least. in benzene solu-
tion, is a resonance hybrid of the classical ketonic amide and an approx-
imately ten percent contribution from an excited form. This excited
form was drawn in such way that the nitrogen is doubly bonded to the
carbon and the oxygen is singly bonded to the carbon. The nitrogen and
oxygen possess formal positive and negative charges, respectively.
The first application of the infrared spectrophotometer in this
problem was by M. and R. Freymann in 1936. These researchers
noted that in a keto-enol type tautomeric mixture, the characteristic
hydroxyl valence bond stretch would appear in the absorption spectrum
and that its intensity would correspond to the proportion of enolic form
in the mixture. Since they were unable to find this band in the spectra
of the liquids, melts or vapor phases of several amides, they concluded
that the classical tautomer does not exist. The Freymanns postulated
an intramolecular association. Their final picture resulted from ascrib-
ing a partial double bond character to the carbonyl and carbon-nitrogen
bonds and attaching a partial no-bond character to the hydrogen bridge
between nitrogen and oxygen. The implication of their structure is that
the two sets of bonds are approximately symmetrized with respect to the
bond strengths. This conclusion is amplified slightly and reiterated by
R. Freymann27 and by M. Freymann in publications appearing very
early in 1939.
Raman spectral data have also been of use to workers studying
problems involving the amides. The general background material is
covered adequately by Herzberg. 32 In 1937, Ananthakrishnan3 inter-
preted the Raman spectra of acetamide and propionamide as indicating
the existence of the classical structure plus an excited state differing
from the classical tautomer in that the proton was not actually trans-
ferred from the nitrogen. This second structure was an approach to
that of the Freymann's26 mentioned above; however, in it was implied
a somewhat more sophisticated grasp of the pattern involved in reso-
During this time, Buswell, Rodebush and co-workers had been
studying molecular association phenomena by means of the infrared
spectrometer. In 1938, their first attack on the amide problem
appeared.14 It was concluded that the simple amides are associated
in indeterminately large polymers, the structure of which they were
unable to deduce. The N-disubstituted amides showed no evidence of
association. Because of the spectroscopically obvious perturbations
of the nitrogen-hydrogen bond in the N-monosubstituted amides, it was
deduced, following the suggestion of Copley, that these structures form
a cyclic dimer through two hydrogen bridges. Feeling the necessity of
showing the existence of some tautomeric form, these workers stated
that in the cyclic dimer there is an effective transfer of the protons to
the carbonyl oxygen with the simultaneous assumption by the carbon-
nitrogen linh of double bond character.
Sannie and Poremski61, apparently in ignorance of the work of
Buswell, Rodebush and Roy14 and that of Ananthakrishnan3, deduced
approximately the same structures for the amide linkage as did these
authors. They found that the amides associate to a great extent in
benzene solutions, but very little in water and in ethanol. For the polar
solvents, they extended the idea of M. and R. Freymann 26 concerning
an internally symmetrized form of the hydrogen bridge, by applying
independently the idea of Ananthakrishnan3 and justifying it with a
rather unsatisfactory steric argument. Sannie and Poremski believed
that the amides and N-monosubstituted amides in the non-polar solvent
form dimeric or trimeric rings through hydrogen bonding. In the case
of the N-disubstituted amides, an internal ring involving hydrogen bonding
of an alkyl hydrogen to the carbonyl oxygen was suggested. It is inter-
esting to note that each form presented by Sannie and Poremski was
essentially a resonance hybrid of the tautomeric form and a quasi-tauto-
meric form in which the nitrogen retains the hydrogen and a positive
charge, while the oxygen is singly bound and negatively charged.
The studies by Buswell's group continued for several years in the
more general aspects of association. During the year 1940, several
significant advances arose from the work of this group. Buswell,
Downing and Rodebush12 stated that they had eliminated the problem of
the tautomeric shift previously postulated for the dimeric form of the
N-monosubstituted amides. They observed that the characteristic
absorption of the carbonyl group in high dilutions in carbon tetrachloride
is essentially the same as that observed in the spectra of the pure
liquids. This led to the realization that, in the dimer, there is the pos-
sibility of cis-trans isomerization with respect to the orientation of the
aryl or alkyl substituents on the a mide nitrogen. They then proposed
that only the trans form of the dimer is stable, presumably for steric
reasons. Introducing a note of indeterminism, they stated that the
characteristic frequencies of the nitrogen-hydrogen valence stretch in
which the hydrogen was bridged either to another nitrogen or to an
o::ygen are identical.
Busvwell, Krebs and Rodebush, 13 on the basis of the preceding
work, stated that the modes of association in the proteins are through
amino hydrogen bridges to oxygens. This paper also contained the
heuristic idea that the characteristic perturbed bands of the amino
hydrogen link occurring at 3. 00 microns (3333 cm"1) and 3.22 microns
(3125 cm-1) are due to linear and cyclic dimers# respectively. It was
suggested that the difference in frequencies associated with these forms
lies either in resonance stabilization of the ring or in the cis-trans iso-
merization mentioned above.
Lecompte, collaborating with R. Freymann 40s41 extended infra-
red studies of the amides into the six to nineteen micron (1667-526cm 1)
range of the spectrum. Buswell, Downing and Rodebush12 had utilized
an isolated datum from that region, but had neglected the golden
opportunity which was at hand by continuing to concentrate upon the
three micron region. In the first paper by Lecompte and Freymann,
there is a fairly complete discussion of the theories of the amides
which had been advanced prior to that time. Once again, a case is
encountered in which the authors were unfamiliar with the literature.
The paper by Busvwell, Krebs and Rodebush 3 are unknown to them.
They reiterated that the excited dimer of Sannie and Poremskit1 is an
improvement upon the form proposed by Buswell, Rodebush and Roy;14
however, they rather vehemently denied the validity of the inner tau-
tomer of the former. Lecompte and Freymann foundthe cyclic tauto-
mer of Buswell, Rodebush and Roy14 to be satisfactory, being totally
unaware of the fact that this form had been discarded in a later pub-
In an X-ray diffraction study of the acetamide molecule# Senti'
and Harker62 found that the molecule is planar. This result is predic-
ted by either the ketonic or enolic form. It was found that the carbon-
oxygen bond distance is 1. 28 A., the carbon-methyl distance, 1. 51 A,
and the carbon-nitrogen distance, 1.38 A. The first measurement
agrees well with lata from formic acid, oxalic acid dehydrate, dilreto-
piperazine and glycine, while the second value is normal. The carbon-
nitrogen bond distance is somewhat greater than the distance reported
for diketopiperazine, but is approximately the same as that obtained
for urea and thiourea. The oxygen-carbon-methyl angle is 1290, the
nitrogen-carbon-methyl, 1090, and the nitrogen-carbon-oxygen, 122O.
It was found that a nitrogen-hydrogen-oxygen bridge exists in the mole-
cule. Its length is 2.86 A. Senti and Harker conclude that the nitro-
gen-hydrogen bonds are in the plane of the molecule and that only
the ketonic form exists in the crystal.
Little more than confirmatory work was done during the remain-
ing war years. A tremendous quantity of data pertaining to the amide
problem was collected in connection with the effort to synthesize peni-
cillin. The germane portion of this work is reported in a previously
cited reference. 58
Because of the immense productivity of Henri Lenormant, his
publications i4Z-51 will be described from his excellent review paper.
i'Infrared Spectra of the Peptide Linkage".51 In order to utilize this
review to its maximum extent, it is necessary to step out of chronol-
ogy for the examination of two papers appearing during this interval.
Richards and Thompson59 stated that in their opinion, the N-
monosubstituted amides could exist in two mesomeric forms com-
posed of four resonating structures, the classical amide and its
tautomer and the corresponding excited states. From these deductions,
Richards and Thompson felt that the characteristic amide vibrations
should be the valence stretches and deformations of the carbon-ox-ygen
double bond, carbon-nitrogcn double bond, hydrogen-oxygen bcnd and
In their study, emphasis is placed for the most part on the
absorption occurring in the five to seven micron (2000-1429 cm-1)
range; however, they did study the three micron region briefly. They
concluded that bands occurring in the latter range are actually caused
by various nitrogen-hydrogen stretching modes, but that the hydroxyl
stretch is not observed. This, of course, eliminated the possibility
of the classical enolic form and its resonance hybrid. Other bands
in the three micron region were attributed to hydrogen bridges be-
tween two atoms of nitrogen and between nitrogen and oxygen.
The most interesting phenomena reported were in the lower fre-
quency region. Two bands were observed in this spectral region of
the amides and the N-monosubstituted amides whether they were in the
solids fused or liquid state or in solution. The one appearing at the
higher frequency, about 1667 cm'" (6.0 microns), was denoted "A"
and that at 156Z cm"1 (6.4 microns), "B". These have been labeled
the "Amide I" and "Amide II" bands, respectively, by other authors.58
It was found that the A band occurs in all amides regardless of
the state of substitution upon the nitrogen. This band is displaced to
slightly lower frequencies when compared to the carbonyl band of other
substances, but it is sufficiently close to that absorption to be classi-
fied. The B band does not occur in the N-disubstituted amides. It is
shifted to rather lower frequencies in the case of the N-monosubstituted
amides than those observed in the non-substituted amides. It was
found that upon dilution, the A band shifts to higher frequencies and the
B band to lower. Substitution of an electrophilic group upon the nitro-
gen tends to reinforce the A band and weaken the B band. This is
further evidence that the A band can be identified as the carbonyl
They then proposed four possibilities for the origin of the B band.
First, the ketonic carbon-nitrogen bond with some double bond charac-
ter; second, the enolic carbon-nitrogen bond with somewhat less than
one-hundred percent double bond character; third, the amino hydrogen
deformation; fourth, an overtone or combination band. They argued that
if this band is assigned to the enolic form of the molecule, it is tacitly
assumed that both tautomers exist in the solid or pure state and in solu-
tions. This is contrary to the evidence presented by Senti and Harker.
That it is the somewhat doubly bound carbon and nitrogen in the ketonic
form was considered to be unlikely since this form would be expected to
exist to a higher degree in the N-disubstituted compounds than in the
N-monosubstituted. The B band is not observed in the former class of
compounds. The shape and intensity of the band tend to rule out the
possibility of its being an overtone or combination band. Thus it was
presumed that this band is caused by the nitrogen-hydrogen deformation.
Hartwell, Richards and Thompson, 30 in a general investigation
of the characteristic absorption frequency of the carbonyl group,
reported that the N-disubstituted aides absorb in the neighborhood
of 1710 cm* (5. 85 microns). In studies of the chloroacetic acids, they
observed that the substitution of such electrophilic groups shifted the
absorption of the carbonyl group to higher frequencies.
Returning now to the work of Lenormant, 5 it is found that in
previous investigations, he had observed the so-called B band in the
amides, but had not studied its sensitivity to dilution. In the course of
his researches, he came to the conclusion that the B band results from
a symmetrization of the carbon-nitrogen and carbon-oxygen bonds in
that the carbonyl bond has somewhat less than one-hundred percent
double bond character and the carbon-nitrogen link approaches that un-
known percentage double bond character. This conclusion is reinforced
to a great extent by the fact that only the A band occurs in N-.bromo-
acetamide, only the B band in sodium acetamide and both bands with
equal intensities appear in N-ethyl acetamide. It was noted earlier
that the silver salt of acetamide plays a dual role in substitution reac-
tions. Lenormant found that the substitution of silver upon the nitrogen
in acetamide yields a strong B band and a relatively weak A band which
is displaced to somewhat lower frequencies.
It was found that in the spectra of the lactams, cyclopeptides and
diketopiperazine, the B band does not appear. Observing that the
nitrogen-hydrogen deformation should be found in each of these com-
pounds, Lenormant felt that the explanation of the origin of this band
by Richards and Thompson is invalid. On basic hydrolysis, the cyclo-
peptides lose the A band and gain the B band, while diketopiperazine
presents a very intense B band and a weakened and shifted A.band. In
the polyamides, such as the nylons, the A and B bands both occur as
long as there is an amino hydrogen present. Replacement of this atom
by an alkyl group results in the disappearance of the B band.
On deuteration of the compounds possessing the B band, it is
found that the band shifts to lower frequencies by a ratio of 1/1. 05. In
other less ambiguous cases, replacement of a hydrogen atom by a deuter-
lum atom results in shifts of the frequency very nearly in a ratio of
1/1. 33. 3- The implication here is that the B band is not strictly a
function of the nitrogen-hydrogen defurmation as was stated by Richards
In view of the fact that only the A band appears in the N-disubsti-
tuted amides and that this band is shifted to lower frequencies with
respect to the normal absorption of the carbonyl group, Lenormant
concludes that those compounds exist in a mesomeric state to
which the excited form makes a contribution of approximately
fifteen percent. He arrives at a similar conclusion for the non-
With respect to the N-monosubstituted amides and the peptides,
he concludes that the B band exists only if the group substituted upon
the nitrogen can assume a position cis to the oxygen. This is exem-
plified by the N-monosubstituted amides, peptides and polyarnides.
This band fails to appear if this orientation cannot be attained as is
the case for the lactams and diketopiperazine.
Lenormant mentions the beta-di-ketoncs and beta-diketonic
esters. When the sodium, copper or manganous salt of any of these
compounds is formed, the carbonyl stretch shifts to markedly lower
frequencies. Sodium tends to shift the band to the greatest extent.
This is compared to the phenomena observed following the substitution
of bromine, sodium or an alkyl group for the amino hydrogen in
N-monosubstituted aides and is found to support his idea of the con-
figuration of the amide group.
Lenormant clarified the situation in the three micron region
to a great extent by obtaining the infrared spectra of N-deuterated
amides. He found that all of the bands in that region previously re-
ported to be nitrogen-hydrogen stretching modes shifted the proper ratio.
This, of course, could not serve as a basis for the elimination of
the tautomeric form. He agreed with the work of Buswell, Krebs
and Rodebush13 in the assignment of bands in this region to linear
and cyclic di-ners.
Cromwell, EMiller, Johnson, Frank and V-aIllpce prepared a
series of amino-substituted, alpha-beta unsaturated ketones. Numer -
ous anomaliess" appeared in the infrared spectra of these compounds.
Because of the complexity of these compounds, it is difficult to assign
many of the bands appearing. It is significant to note that a strong or
very strong band occurs in the region in which the previously de-
scribed B band appears. The characteristic carbonyl absorption is
shifted to lower frequencies than those observed in the simple dialkyl
ketones. The B band is found in the two cases in which the amino
group is substituted in the alpha position; however, the nitrogen is
either part of a morpholino group or else it is methyl. benzyl substi-
tuted. The carbonyl group appears at about the same frequency at
which it is found in the N-disubstituted amides. In the beta amino
compounds, the B band appears regardless of the state of substitution
of the nitrogen. The A band occurs at much lower frequencies than
would be expected in the corresponding amides. The authors were
unable to draw any definite conclusions regarding the nature of these
Darmon and Sutherlandl8 found that the presence of water in pro-
teins used in infrared spectroscopic work does not affect the nitrogen-
hydrogen stretching frequencies. They confirmed the previously
reported supposition13' 51 that the 3060 cm"I (3.27 micron) band is
characteristic of the cyclically dimeric form of the amides and that
the 3280 cm-1 (3.05 micron) band is characteristic of the linear dimer.
In the year 1951, the Discussions of the Faraday SocietyZ4 con-
tains a series of interesting comments regarding the general problem
of the amides. On page 274. Sutherland stated that the characteristic
absorptions of the nitrogen-hydrogen stretches perturbed by the car-
bonyl group are in the range 3320-3240 cm-1 (3.01-3.09 microns).
When bonded to another nitrogen, the absorption occurs from 3350-
3150 cm"1 (3.03-3.17 microns). Lenormant suggested on page 319
that the two bands appearing in the three micron range and the A and
B bands might be caused by two distinct forms of the amides.
In a thorough study of acetamide, Davies and Hallam0 found
that this compound exists almost wholly as a cyclic trimer in chloro-
form solution, a dimer in acetone and a monomer in methyl cyanide.
In the trimer, by virtue of the two distinct vibrational modes shown
for the amino hydrogen stretch, it is assumed that only one of the
hydrogens of each amino group is involved in ring formation. They
found the carbonyl stretch to be at 1700 cm"1 (5. 88 microns) in the
monomeric form and at 1678 cm-1 (5.96 microns) in the associated
form. A very strong band occurring at 1595 cm"1 (6.27 microns)
is assigned to the deformation of the unassociated amino group follow-
ing Richards and Thompson.59
Bates and Hobbs10 have recently investigated the dipole moments
and group structure of some acid amides. In an effort to explain the
nature of the group without utilizing the concept of resonances they
assumed the existence of a planar structure and proceeded to discuss
the observed moments on the basis of dipole interactions* Their deduc
tions are somewhat vitiated by comments by Kumler;36 however, the
conclusion that the amides exist in the ketonic form is in good agree-
ment with the numerous investigations presented heretofore.
A reasonably. consistent interpretation of the work which has been
presented above would lead one to believe that the amides exist in an
associated form in the solid state& melt, pure liquids and fairly con-
centrated solution in non-polar solvents. They are neither dimerized
in polar solvents nor associated in dilute solutions in non-polar solvents.
Further, the amide group must be thought of as being planar and ketonic
in form.; If one assumes that the enolic form may exist, it must be
only as a result of the symmetrization of a hydrogen bridge between the
amino hydrogen and the carbonyl oxygen. It is assumed that hydrogen
bridging to nitrogen does not tend to occur. The Bland is found in the
spectra of certain amides. This band has not been explained on the
basis, of the present information.
Previous work in this field has concentrated upon the variation
of the substituents upon the nitrogen. One finds references made to
benzamide and, occasionally, to amides of the higher aliphatic acids;
however, little work has been done using amides in which the hydrogen
of methyl or methylene groups alpha to the carbonyl group have been
systematically replaced by electrophilic groups. It is evident that this
substitution would inhibit the formation of An essentially single-bonded
configuration of the carbonyl group. It would encourage the deformation
of the electronic cloud about the nitrogen in such a way as to promote
the assumption of partial double bond character by the carbon-nitrogen
bond. Thus, it should be possible, these shifts being predictable, to
create an amide with sufficient abnormality that the questionable absorp-
tions of this group could be identified.
This study was carried out on a series of N-butyl-ethanamides and
one of NN-di-butyl ethanarmides. The alpha-methyl group in each series
is step-wvise halokcnatcd. The compounds investigated in this research
are listed in Table I.
Two infrared spectrophotometers were utilized in this research.
The first was a Perkin-Elmer Model 12-C single-beam instrument in
the Naval Stores Research Laboratories of the Glidden Corporation in
Jacksonville, Florida. A Perkin-Elmer Model 21 double-beam infra-
red spectrophotometer in this laboratory was used for the final phases
of this research. The frequencies of all bands are reported as
obtained from data gathered by use of the latter instrument because
of its superior calibration characteristics.
The electrical and optical details of both instruments are com-
pletely described in manuals and drawings prepared by the Perkin-
Elmer Corporation. Further data concerning the double-beam instru-
ment are available in a series of papers by White and Listron.69-71
The method of calibration of the instruments is mentioned in the
previously cited manuals. A table of wavelengths and wavelength
standards suitable for such a calibration was published by Plyler and
Experimental techniques are described in a number of sources.
Two previously mentioned books958 contain valuable information on
this subject. A rather general paper concerning both instrumentation
and techniques has been published by Williams. Specific applications
of the double-beam instrument are discussed in a later paper.
The :methods used in the present study are essentially conformist.
It was found that concentrations could be rapidly determined by prepa-
ration of several solutions of known concentration and obtaining their
spectra. Comparison of the carbonyl band absorption intensities of
the known solutions to those of the roughly prepared solutions of the
same order of magnitude of dilution permitted this determination.
The compounds used were obtained from two sources. All of
the acetamides en:cept for the trifluoro derivatives were prepared by
Fried28 under the direction of Tarrant in this laboratory. The latter
compounds were prepared by Tarrant and Letaw. 66 Reagent grade
carbon tetrachloride was used as the solvent in all dilution work. In
each case, the absence of extraneous bands in the infrared spectro-
gram was the criterion of purity.
The spectra used in this research were all obtained under iden-
tical conditions with respect to instrumentation. Scanning was at the
rate of one micron per minute. Automatic slit program IV was used.
This provided slit widths of 14, 37 and 111 microns at wavelengths of
1. 8, 5 and 10 microns (5556, 2000 and 1000 emr"l), respectively. In
the original work, the scale of recording was two inches per micron;
however, for the purpose of reproduction, this was reduced to one
inch per micron.
The spectral region studied was from two to fifteen microns.
These limits were imposed by the sodium chloride optics used. Work
was carried out at 250 C. in a relative humidity of less than fifty
percent. As a result of compensation for the solvent, carbon tetra-
chloride, in dilution work, the sensitivity of the instrument was
reduced considerably in the region beyond ten microns. For this
reason, discussion of the spectra studied will be limited to the 2-10
micron (5000-1000 cm1) interval.
Spectra were obtained for the pure materials as well as for
several dilutions of each compound. In general, further dilutions were
not prepared if there were no change in the spectrum for two consec-
utive dilutions. Because it is a solid at 250 C., N-butyl-2, 2-dichloro-
ethanamide was not run as the pure substance. A very small quantity
of carbon tetrachlorTde was added to it in order that a homogeneous,
non-scattering mull might be obtained.
For the pure compounds, a demountable cell consisting of two
salt plates pressed together over the liquid was used. The thickness
of the sample was adjusted by variation of the pressure upon the plates
in order to obtain a spectrograph of suitable intensity. Dilutions were
carried out in cells approximately 0.025, 0. 100 and 0.500 mm. thick.
All dilutions reproduced in this paper were obtained in a cell 517.8
microns thick compensated with a cell containing a layer of carbon
tetrachloride 486.6 microns in thickness. The degree of compensa-
tion obtained is shown in Figure I.
Reproductions of the spectrograms obtained in this research are
attached at the end of this chapter. Only the spectrum of the pure
compound and one dilution are reproduced. This concentration,
approximately 0.01 M, is of such a magnitude that no further shift
in the frequency of any band is observed upon further dilution. The
value of each dilution and the name of the compound is indicated on
each of the figures. The bands will be discussed in order of decreas-
ing frequency or, equivalently, increasing wavelength.
Because of the rather lengthy names of the compounds investi-
gated, abbreviations have been constructed for them. There is little
possibility of ambiguity in the abbreviated nomenclature as it appears
in Table I.
To facilitate the description of the bands in the spectra at hand,
the following notation will be utilized to indicate relative band inten-
sities: vs e very strong, s = strong, m medium, w = weak, and
vw a very weak.
In all of the compounds investigated, a broad, weak absorption
band was found in the 2.25-2.75 micron (4444-3636 cm" ) region. It
can be ascribed either to an overtone or combination band or to a
small amount of water existing as an impurity in the compounds.
Because of its contour, the former explanation is the more probable;
moreover, it will be recalled that Darnmon and Sutherland have estab-
lished the fact that water does not affect the characteristic amide
ABBREVIATIONS OF COMPOUNDS
Name of Compound Abbreviation
N-butyl-2-chloro-ethanamide N-butyl chloro
N-butyl-2, 2-dichloro-ethanamide N-butyl dichloro
N-butyl-2-chloro-2-fluoro-ethanamide N-butyl chlorofluoro
N-butyl-2 2-difluoro-ethanamide N-butyl difluoro
N-butyl-2, 2# 2-trifluoro-ethanamide N-butyl trifluoro
N, N-di-butyl-ethanamide N-dibutyl
NO N-di-butyl-2-chloro-ethanamide N-dibutyl chloro
N, N-di-butyl-2, 2 -dichloro-ethanamide N-dibutyl dichloro
N, N-di-butyl-2-chloro-Z -fluoro-ethanamide N-dibutyl chlorofluoro
N. N-di-butyl-2, 2-diflporo-ethanamide N-dibutyl difluoro
N, N-di-butyl-2, 2, Z-trifluoro-ethanamide N-dibutyl trifluoro
The N-dibutyl compounds present a rather puzzling set of ab-
sorptions in the region immediately around three microns. These
bands and their relative intensities are noted in Table II. On dilution,
there is no evidence of shifts in these bands. The bands appear to be
overtones or combinations. With respect to impurities, the corres-
ponding N-butyl compounds could be present; however, the rather
large differences in boiling points relative to the N-dibutyl compounds
makes this possibility rather unlikely.Z8
THREE MICRON REGION, N-DIBUTYL SERIES
Compound Wavelength Frequency
N-dibutyl 2.91 (w) 3436
3.07 (vw) 3257
N-dibutyl chloro 2.90 (w) 3448
3.04 (w) 3287
N-dibutyl dichloro 2.95 (vw) 3390
3.04 (w) 3289
N-dibutyl chlorofluoro 2.93 (vw) 3413
3.05 (w) 3279
N-dibutyl difluoro 2.88 (vw) 3472
3.03 (vw) 3300
N-dibutyl trifluoro 2.88 (vw) 3472
3.01 (w) 3322
The bands occurring in the neighborhood of 2.9 microns
(3448 cm-1) correspond in frequency to the unperturbed hydrogen-
nitrogen stretch found in the N-monosubstituted amides. Aside from
the fact that such impurities are unlikely, it must be recognized that
virtually none of the nitrogen-hydrogen links would exist in an un-
perturbed state under the concentration conditions existing in the
THREE MICRON REGION. N-BUTYL SERIES
3.05 (s) 2.91 (w)
3.07 (s) 2. 93 (w)
3.07(s) 2.92 (w)
3.06(s) 2.92 (w)
2. 92 (m) 2.92 (w)
3.04 (s) 2.92 (w)
As will be shown later, the carbonyl absorption frequencies of
the N-dibutyl series occur in the range 1650-1680 cm-1 (6. 06-5. 95
microns). The first overtone of this absorption should occur in the
3300-3360 cm-1 (3.03-2. 98 micron) region or at slightly lower fre-
quencles. It would seem logical to identify the bands occurring
around 3.00 microns (3300 cm-1) as the first overtone of the carbonyl
group. By the same reasoning, the much weaker absorptions at
slightly higher frequencies could be overtones of a band occurring at
only slightly shorter wavelengths than that of the carbonyl. This
band is described later.
The three micron region of the spectra of the N-butyl amides is
extremely rich in significant bands. The pure compounds all have
two well-defined bands which may be identified as perturbed nitrogen-
hydrogen stretches. In additions N-butyl difluoro presents a very
sharp absorption at 2i92 microns (3425 cm-1) This is doubtless due
to the unperturbed nitrogen-hydrogen stretch. As the solutions be -
come more dilute, the two perturbed bands are markedly weaklened
and the unperturbed band is strengthened, In every case, it is found
that the lower frequency band disappears before the higher one.
Observed frequencies and intensities are recorded in Table II.
In both series of compounds, the carbon-hydrogen stretch pre-
sents three absorptions in the neighborhood of three and one-half
microns. A shoulder of medium intensity occurs at 3.42 microns
(2924 cm-1), a strong band is found at 3.43 microns (2915 cm"1) and
a medium band appears at 3.51 microns (2849 cm-1). There is no
apparent change in the bands of the compounds into which halogen
substituents have been introduced. This is, of course, due to the
overwhelming intensities of the carbon-hydrogen valence stretches
attributable to the butyl group.
5.7 :.MICRON BAND, N-DIBUTYL SERIES
Compound -Wavelength Fr e qucncy
(microns) (cm- )
N-dibutyl 5. 79 (vw) 1727
N-dibutyl chloro 5.77 (w) 1733
N-dibutyl dichloro 5.70 (w) 1754
N-dibutyl chlorofluoro 5.65 (w) 1770
N-dibutyl difluoro 5.67 (w) 1764
N-dibutyl trifluoro 5.63 (vw) 1776
Another unusual band appears In the spectrograms of the N-
dibutyl series at approximately 5.70 microns (1754 cm1-). These
bands have not been previously reported. The frequency and
intensity data appear in Table IV. It is interesting to note that all
of the bands are sharp except for those of the N-dibutyl chloro and
N-dibutyl trifluoro amides. There is no shift of frequency upon dilu-
tion. It is possible that these are overtone or combination bends.
THE SIX MICRON REGION
N-butyl 6.04 (vs) 5.93 (vs)
N-butyl chloro 6. 02 (vs) 5.93 (vs)
N-butyl dichloro 5.98 (vs) 5.86 (vs)
N-butyl chlorofluoro 5.96 (vs) 5.84 (vs)
N-butyl difluoro 5.94 (vs) 5.82 (vs)
N-butyl trifluoro 5.86 (vs) 5.76 (vs)
N-dibutyl 6.07 (vs) 6.07 (vs)
N-dibutyl chloro 6.03 (vs) 6.04 (vs)
N-dibutyl dichloro 6.00 (vs) 5.94 (s)
N-dibutyl chlorofluoro 5.97 (vs) 5.93 (s)
N-dibutyl difluoro 5.97 (vs) 5.93 (m)
N-dibutyl trifluoro 5.92 (vs) 5.91 (vs)
Table V contains the data obtained for the absorptions occur-
ring in the immediate neighborhood of six microns. Each of the
compounds presents an extremely strong absorption in this region.
It has become rather obvious in the course of previous work that
these bands are characteristic of the carbonyl group.
6.4 MICRON REGION, N-BUTYL SERIES
6.41 (vs) 6.66 (ve)
6.40 (s) 6.60 (a)
6.40 (s) 6.61 (s)
6.40 (s) 6.60 (s)
6. 39 (s) 6.57 (s)
6.40 (vs) 6.57 (s)
By far the most intriguing band to be found in the infrared spec-
tra of the amides and amide-like compounds is that occurring in the
neighborhood of 6.40 microns (1562 cm-1) in the spectra of the
N-monosubstituted amides. This band has been discussed at length
in Chapter I. Deductions concerning the data obtained in this research
will be presented in the following chapter. The data are listed in
A fairly strong absorption band was found in the neighborhood of
6,85-7. 10 microns (1460-1408 cm-1) for all of the compounds inves-
tigated. On the basis of theoretical band assignments for hydrocar-
bons, 32 this band is assumed to be caused by the asymmetric defor-
mations of the methyl and methylene groups. It is observed that the
band is much more intense in the N-dibutyl compounds than it is in
the N-butyl derivatives
It was found that a band occurred at about 71 25 microns (1379
cm"1) in all cases, This band has been identified by Davies and
Hallam6 specifically for acetamide6 as the carbon-nitrogen stretch.
In view of the fact that there is no change in the frequency of this
band upon diiutions this assignment is not justified in the present
case* In subsidiary spectrograms obtained in the course of this
research it was found that this band occurs in n-butanol and in the
three n-butyl, n-hexyl and n-octyl amines with virtually no change
in form or in relative intensity. It is believed that this band and
those occurring in the neighborhood of 7.50 microns (1333 cm*l)
are probably due to symmetric modes of deformation of the various
carbon-hydrogen configurations in these compounds.
FLUORINE STRETCHING BANDS
8.3 5 (vs)
8. 60 (vs)
8. 30 (vs)
8. 8* (s)
A band found in the 8.00-8.20 micron (1250-1Z19 cm"1) range
is presumed to be attributable to rocking or wagging modes in the
carbon-hydrogen links. In some of the fluorinated compounds, this
band is partially masked by carbon-fluorine valence stretches.
In view of the pronounced hydrogen bonding possibilities
through fluorine atoms, it is not surprising that dilution produces
considerable changes in the absorption frequencies of the carbon-
fluorine bonds. In Table VII, the data for the absorption attribu-
table to the valence stretch of the carbon-fluorine bond are given.
The carbon-fluorine stretch assignments are in the 8.28-9. 77
micron (1208-1024 cm-1) range. It is evident that several of the
non-fluorinated. compounds absorb radiation in that band of frequen-
cies. This leads to some ambiguity in the selection of the bands;
however, it is seen that the bands cited in Table VII are, for the
most part, vastly more intense than are those found in this region
of the spectra of the other six compounds. The intensity of an infra-
red band is, to a very good first approximation, proportional to the
change of the dipole moment produced by the vibrational mode which
is excited. The stretching modes of the carbon-fluorine vibrations
are found in this region. It can be said that the intensities of these
bands should be! much greater than those of the less active deforma-
tional modes of the carbon-hydrogen bonds which also occur here.
Thus, it is reasonable to assume that these extremely intense bands
are due to carbon-fluorine valence stretches.
35 .5 6 T 8 j
I 6I -Ioo
/- I- r
I I I
I I -4--
0. 006 M
.. ...... ........
6 I TI
--- I --- -- I --- -- I-----i______
I I I
It --sq -
SItoo IVM Itoo I40 0ee
& ~ ~T4o''
rAfi i oo l tc
I3I Io Io I I I
I I~b'~ D100
6 7 9 q Tc
S'8 I i i I I I
1300 16" 1*-o0 stolo cc
-oo B*L* soo* 1io.
Iow I 8o-
3000 o~ 180.
ooo ,ooo ( goo itoo f *.i
' I I
o o Ioo
I o a t* IA o Io
I I I I
i I I
I I II I I
i I I I
ooo .. laQ+ VID
l I' lO Ia
1~0o 't~ 'a.. lope
The most interesting general features of this research are the
clearly defined frequency shifts found upon systematic substitution
of halogen atoms upon the number two carbon atom of the acetamides
studied. Further, these shifts are not in random order, but may be
significantly related to the electronegativity of the substituent.
Discussions of the relative electronegativities of chlorine and
fluorine must take into account two very general phenomena. The
most obvious of these is the ability of the halogen atom to attract
electrons inductively, the I effect. On the basis of the electronic
screening of the nucleus, it follows that fluorine should possess the
stronger I effect. This is borne out by experiment. The second
effect is the resonance or mesomeric shift of electrons from the
valence shell of the halogen concerned into a coordinate bond. It
has been determined on the basis of a large amount of experimental
work that fluorine has a greater tendency to operate this / M shift
than has chlorine. Although Ingold has stated34 that the energies
Involved in the I and -AL M effects may be separated as the results
of two distinct phenomena, it is well to recall to mind that the effect
of any one halogen is unique. In other, words, pragmatically, it
must be recognized that there is only one effect, that of the replace-
ment of a hydrogen atom by a halogen atom.
In, the compounds under consideration, there is little or no
possibility for the operation of the / M effect. In order for the
halogen to shift a pair of electrons into the halogen-carbon bond,
there must be some mechanism for the removal of an electron pair
from the carbon atom. This follows unless pentavalent carbon is to
be assumed. If, on the other hand, two halogens are attached to the
same carbon atom, the / M shift of the one may be enchanced by
the I effect of the other. This actually would vitiate the I effect
with respect to the remainder of'the molecule.
Perhaps the most satisfactory quantitative data available for
a comparison of the effects of the halogens are the ionization con-
stants of substituted acids. Deductions concerning the inductive
effect introduced into the compounds investigated in the present study
are of the highest validity if related to the halogenated acetic acids.
The ionization constants of chloro-, dichloro-, and trichloro-acetic
acids are 1.396 x 10 5 5.5 x 10"2 and 0. 13, respectively. 29
Henne31 reports the ionization constants of the analogous fluoroacetic
acid series to be 2.17 x 10-3, 5.7 x 10-2 and 0.588 or 0.533. This
clearly demonstrates that the net inductive effect attributable to
fluorine is greater than that of chlorine.
In order to facilitate the presentation of the interpretation of
the data, a step out of logical order. will be made by first postu-
lating a configuration of the amides. This configuration will then
be shown to account for the infrared absorption spectra of the
N-monosubstituted and N-disubstituted amides in the region studied.
The interpretative work of Lenormant Is accepted with a
few reservations. It will be recalled that he proposed the N-mono-
substituted amides to be in a mesomeric state of the normally
written ketonic form and the corresponding excited state. He then
stated that the 6.40 micron (1562 cm-1) band in these compounds
was due to the stretch of the nitrogen-carbon partial double bond.
The fact that this band does not occur in the N-disubstituted aides
was not explained.
From the present works it is concluded that this band does
occur in the N-disubstituted amides, but that it is masked by acciden-
tal degeneracy with the carbonyl band. It is concluded that this band
and the carbonyl band are both found in the immediate neighborhood
of 6.00 microns (1667 cm-1).
It has been stated above that an effort has been made to
explain the absence from the spectra of the N-disubstituted amides
of the band in question on a steric basis. This was necessary since
it was recognized that the replacement of an amino hydrogen atom
by an electron releasing group would definitely increase the likeli-
hood of the assumption by the carbon-nitrogen link of some double
bond character. Furthermore, from the bond distances found in
acetamide, 62 some double bond character must be attributed even
to that carbon-nitrogen bond.
Rather strenuous objection may be raised to the idea of acci-
dental degeneracy in these compounds because of its generality.
It is obvious that there are certain minimal energy limits involved
in the assumption by heteroatomic systems of completely symme-
trized configurations. It is this limit which has prevented the
realization of such a state by the N-monosubstituted amides. The
substitution of an additional electron releasing group upon the nitro-
gen allows the system to cross this energy threshold and attain
the symmetrical condition described above.
The lowest energy state should be# according to the resonance
theory, that which is the most symmetrical with respect to the
electronegativities of the participating atoms. It is believed that
this state can be attained by the N-disubstituted amides and that
its realization is attested to by the absence of the 6.40 micron
(1562 cm"1) band.
Table VIII lists the data of Tables V and VI which are per-
tinent to the discussion of the N-monosubstituted amides. The B
band shift is that of the most prominent peak in the band. It should
be recalled that the N-butyl chloro compound was not run in the pure
state, but was mulled with carbon tetrachloride. For this reason,
the shifts shown for it are possibly not entirely comparable to
those of the other compounds. At the most, however, they are only
slightly lower than the true value.
A BAND AND B BAND DILUTION SHIFTS, N-BUTYL SERIES
Compound A Band Shift B Band Shift
N-butyl 3SO 58
N-butyl chloro / 25 31
N-butyl dichloro 7 34 14
N-butyl chlorofluoro L 34 31
N-butyl difluoro / 34 29
N-butyl trifluoro / 30 41
The notable feature of Table VIII is the more or less central
position which the dichloro derivative assumes; that is, this com-
pound shows the greatest shift of the A band and the least shift of
the B band. There are no striking differences in the frequencies of
the A bands; however, Table V has shown that there is a steady drift
toward higher energies of both the pure and dilute A band absorp-
tions of the N-butyl amides. Assuming that this band is the carbonyl
stretch, the implication of these data is that the substituted halogens
exercise the expected I effect. This serves to inhibit the outward
shift of the electron cloud in the neighborhood of the oxygen atom.
The frequency of this absorption, in the case of the trifluoro deriv-
ative, is 1736 cm-1 (5.76 microns) in dilute solution. This is near
the high frequency limit, 1740 cm-1 (5. 75 microns), of the absorption
of the carbonyl band observed by Batuev 1 in dilute dioxane solutions
of the lower fatty acids. He found that the average shift of this band
is 70-84 cm"1 on dilution.
Table VI, on the other hand, shows that the net shift of the B
band is toward lower frequencies on dilution. It is seen that the
minimum shift is found in the case of the dichloro derivative; however,
the B band in the pure material occurs at lower energies than it
does in any other of the N-monosubstituted amides. The B bands of
the dilute, or unperturbed, compounds progressively increase in
frequency. Thus, there is a tendency on the part of the substituted
halogens to strengthen the bond with which the B band is associated,
but this does not hold true in the pure materials. Obviously, some
other effect must be sought in explanation of this observation.
The pure liquid N-monosubstituted amides are highly associ-
ated. This association involves the bonding of the amino hydrogen
to a donor atom such as the carbonyl oxygen, the amide nitrogen or
a fluorine. It is fairly evident, from the shift of the A band upon
dilution, that the oxygen atom has participated in a rather high
order of hydrogen bonding. A sensibly reverse phenomenon is
observed with the B band upon dilution. If this band is caused by
the absorption of a carbon-nitrogen bond of rather high double bond
character, it is obvious that a deformation of the electron shell
about the oxygen to a position away from the carbon-oxygen bond
would facilitate a shift of the electron density about the nitrogen
into the nitrogen-carbon bond. These tendencies are reinforced by
hydrogen bond formation. This is the situation which pertains in
the pure material. Upon dilution, the hydrogen bonding through
the oxygen is destroyed and is accompanied by a consequent shift
inward of the electron cloud about the oxygen. Thus, the observed
behavior of the B band upon dilution is that which would be expected
if the postulate above is correct.
The frequency of the B band in the pure materials is some -
what anomalous. Reference to Table VI shows that the most
prominent band tends to shift toward lower energies as the dichloro
compound is approached and then to higher energies upon further
substitution. Since the dilute solutions have been shown to behave
as expected, it is necessary to consider possible differences in
the nature of the association processes as substitution progresses.
Table V has shown that the oxygen atoms become less susceptible
to hydrogen bonding as the electronegativity of the adjacent group
increases. Examination of the frequencies of the unperturbed B
bands shows that this is also true with respect to the nitrogen, but
to a lesser degree. This fact calls attention to the possibility of
hydrogen bonding through the nitrogen atom. If this occurs, it
can be seen that the carbon-nitrogen bond is deprived of electrons,
thus becoming weaker.
The reasoning in the paragraph above is no longer applicable
upon encountering the chlorofluoro compound for there it is found
that the frequency of the B band in the pure compound is higher
than that in the pure dichloro amide, This trend continues as the
substituents become more electronegative. It will be noticed that
this tendency appears upon the introduction of fluorine into the series.
The extreme electronegativity of this element serves to increase the
double bond character of both the carbon-oxygen and the carbon-
nitrogen links. Simultaneously, the electronic density in the
neighborhood of the fluorine increases. These three effects pro-
mote hydrogen bonding through the fluorine rather than through
either the oxygen or the nitrogen. The occurrence of hydrogen
bonding of this type serves further to increase theprotonic nature
of the amino hydrogen, thus releasing more electrons for partici-
pation in the nitrogen-carbon band.
The scheme of association postulated above does not find
application in the discussion of the N-dibutyl amides. This is to
be expected in view of the absence of an amino hydrogen atom. In
Table V, it has been shown that the absorption frequency of the
carbonyl band shifts to higher values as the electronegativity of
the halogen substituent increases. These values are much lower
than those found for the unperturbed N-butyl derivatives. The be-
havior is that to be expected purely on the basis of the I effect.
On dilution, except for a general sharpening, there is no change
in this band for the first two members of the series.
An apparent anomaly is found upon examination of the dichloro
compound. The carbonyl band splits, one part shifting to higher
and the other part to lower frequencies. than that of the center of
the perturbed band. This effect is found to a smaller extent in the
chlorofluoro compound and even less in the difluoro. It is not
observed in the case of the N-dibutyl trifluoro amide. This behavior
must be explained in terms of the underlying concepts involved in
the hypothesis of accidental degeneracy in these compounds.
In the three compounds presenting this split, there exists a
rather highly activated hydrogen atom in the methyl group; There
is no question of the ability of this atom to participate in hydrogen
bonding through the oxygen or the nitrogen and, except in the first
case, through a fluorine Upon dilution, this intermolecular
effect is destroyed. This would strengthen the carbon-o:xygen bond
or the carbon-nittogen bond depending upon which atom was the
Sannie and Po remskii 61 as has been previously mentioned,
postulated the formation of an intramolecular hydrogen bond in the
disubstituted amides. This bond involves the carbonyl oxygen and
an alkyl hydrogen of one of the N-substituentsi This hydrogen bond
would be much weaker than one involving a hydrogen initially bound
to an atom rnore electronegative than carbon. An intramolecular
bridge would not be expected to be broken upon dilution. It symme-
trization initially exists in the presence of this intramolecular bond
and an intermolecular bond to the nitrogen, It will at least conceP-
tually, be destroyed upon dilution. The situation pertaining in
dilute solution is rather complex. The intramolecular bond still
exists, pulling electrons out of the carbon-oxygen bond and, through
the alkyl group, feeding the carbon-nitrogen bond. The electronega-
tive substituents are shifting the electronic cloud of the compound
toward themselves. If this occurs with symmetrization, there can
be no splitting of the 6.00 micron (1667 cm"^) band.
The net effect of dilution, then, is to strengthen the carbon-
nitrogen bond. Further, the perturbation on the nitrogen is removed.
This results in a sharpening of the band associated with the carbon-
nitrogen valence stretch. A symmetrical strengthening of the
oxygen-carbon and nitrogen-carbon bonds must involve a weakening
of the intramolecular hydrogen bond. On the other hand, a strength-
ening of the carbon-nitrogen bond alone tends to increase the
stability of the internal hydrogen bond. In view of the phenomenon
observed, it must be concluded that the least energy configuration
for these compounds demands the existence of this hydrogen bond;
thus, a non-symmetrized state. From the intensities of the portions
of the split band, it can be concluded that the higher energy contri-
bution is that of the carbon-nitrogen bond.
It must be inferred from the argument above that the internal
hydrogen bond exists in all six of the dibutyl derivatives. The
absence of the split band, even in dilute solution, from the spectra
of some of the dibutyl compounds tends to discount the preceding
argument. In the cases of the non-halogenated and rmonochloro
derivatives, this can be explained by the fact that insufficient elec-
tronegative stress is placed in the molecule to effect the separation.
That is, the hydrogen bond is stabilized internally at very nearly
the proper energy to effect complete overlap of the carbon-nitrogen
and carbon-oxygen Stretches. Consideration of the trifluoro deri-
vative shows that the extremely high electronegativity of the tri-
ftuoromethyl group should be more than sufficient to split this
band. Observation of the spectra shows that the split is observed
to its greatest extent in the case of the dichloro compound and that
it becomes smaller as the electronegativities of the substituents
increase. This implies that at lower electronegativities there is
a mean displacement of the electron cloud which stabilizes the
hydrogen bond. Further increase in the strength of the electroneg-
ative center forces the hydrogen bond to break in favor of the
greater electron availability introduced in the unbound oxygen.
This once again leads to the almost complete symmetrization of
the nitrogen-carbon and oxygen-carbon links in the case of the
Inspection of the spectra of the pure N-disubstituted amides
shows that these bands are amply wide for the concealment of two
peaks split on the order indicated by dilution. This natural breadth
is intensified by intermolecular interactions of a lower order of
energy than that of hydrogen bonding.
Reference has been made to hydrogen bonding possibilities
through the fluorine atoms in the compounds studied. Table VII
shows that there are definite shifts in the carbon-fluorine valence
stretch frequencies upon dilution. These shifts are of greater mag-
nitude in the N-butyl compounds. This implies that, even though
they are activated, the alpha hydrogen atoms are not as electroposi-
tive as is an amino hydrogen. Sutherland states that if the shift
in frequency is approximately three percent of the frequency of the
band in question, it may be assumed that hydrogen bonding has
occurred. Although the shifts being referred to here are all less
than this value, it is believed that they stand in evidence of hydrogen
bonding. Sutherland, in his discussion, placed emphasis upon
shifts of absorption frequencies of bonds of the acceptor atom with-
out reference to the donor atom.
The general weakness of the shifts in the chlorofluoro com-
pounds and the absence of a shift in the N-dibutyl trifluoro amide
are of particular significance. The latter, having no acidic hydro-
gens, would not be expected to participate in intermolecular bonds.
In the former pair, the degree of activation of both the halogens and
the methyl hydrogen are of rather low order.
Several references to hydrogen bonding through fluorine are
available. Though most of these involve the hydrogen fluoride case,
several are concerned with other molecules. Copley, Zellhoefer
and Marvel15 investigated bonding of fluorine-activated hydrogens
to amides. They found that these hydrogens are bonded to the nitro-
gen or oxygen of the disubstituted amides only. The conclusion
reached to that no matter how highly activated it may be. a hydrogen
bound to a carbon atom is not capable of forming a hydrogen bond in
preference to the bonds usually occurring in either the amides or in
their N-monosubstituted derivatives. This study was based upon
the solubility of dichlorofluoro methane in several amides. Marvel,
Copley and Ginsberg52 found that the heat of mixing of benzotrifluo-
ride in N-dimethyl acetamide is much greater than that of benzo-
trichloride in the same solvent. On the other hand, it was found
that the heats of mixing of the two solutes in acetone are the same.
This difference in energy was interpreted as being due to the fact
that hydrogen bonding initially exists in benzotrifluoride, the
p-hydrogen acting as the acceptor atom, From cryEtallographic
data, 64 it has long been know that a hydrogen bridge Involving flu-
orine as the donor atom exists in the ammonium fluoride crystal.
Table VII shows that the stretching modes of the carbon-
fluorine link have been found in the 8.28-9.68 micron (1208-1033
cm-") range. It is noted that characteristic bands of this bond are
found at higher frequencies as the number of fluorine atoms present
increases. This phenomenon is analogous to the Scanlan-W'Vrhurst
Effect63 which is described as the shortening of the carbon-halogen
bonds as the number of halogen atoms upon a given carbon atom
increases. This holds true in molecules in which resonance occurs
between covalent and ionic forms. It may be added that there is a
high probability of finding this type of resonance in any fluorine-
carbon bond because of the large difference in electronegativity
between the two elements..
The observed shifts of the carbon-fluorine frequencies are all
in a negative direction upon the breaking of the hydrogen bridges.
This must be explained by stating that the shift of electrons out of
the bond through the agency of hydrogen bonding is more than bal-
anced by the increased availability of electrons from the carbon-
An interesting generality mentioned by Buswell, RodeDuuL ..M.
Roy14 with reference to hydrogen bonding has been stated by
Venkateswaran. The Rule of Venkateswaran relates the acidity
of a hydrogen participating in a hydrogen bond to the frequency shift
observed upon the formation of that bond; that is, the more acid the
hydrogen, the greater the shift toward lower frequencies and the
broader the resulting absorption band. Reference to the spectra
reproduced in this dissertation shows that the bands of the pure com-
pounds listed in Table III definitely differ in broadness. In each
case, the band occurring in the 3.05 micron (3279 cm-1) region is
much broader then the neighboring band at 3.25 microns (3077 cm'l).
If the Rule of Ven-kateswaren is applicable, its implication is obvious.
It is interesting to observe that there is some progression in
the frequencies of the two perturbed bands of the nitrogen-hydrogen
stretch as was seen in the case of the perturbed B bands. Although
the differences in wavelengths are on the order of magnitude of the
error to be expected, it is assumed that they are significant because
of their regular trend. From the lower frequency of the perturbed
nitrogen-hydrogen valence stretch in the case of the dichloro com-
pound, one can deduce that the amino hydrogen is more protonic
in this instance than it is in the other compounds. Thus, the approach
of a hydrogen atom to the nitrogen results in an average increase
in the protonic character of the amino hydrogen atoms.
An important fact to be borne in mind in studying the bands
in this region is that, energy-wise, the wavelength plot utilized
presents a false band width when compared to the bands in the
lower frequency regions. In other words, the actual energy spread
of the bands corresponding to the nitrogen-hydrogen stretch is far
larger than that of the A and B bands. For this reason, the large
number of possible bonding modes which exist are masked by the
method of recording inherent in the instrument used.
There is one salient fact in the data of Table Ill. This is
that the shifts of the nitrogen-hydrogen stretch bands in the fluoro
compounds tend to be smaller than those of the other compounds.
The implication is that hydrogen bonds through fluorine are weaker
than those through oxygen cr nitrogen in the present cases. Exam-
ination of the cyclic dimers which would be formed through fluorine
as contrasted to those through the other two donor atoms yields a
clue to the reason for this. A dimer through amino hydrogen bridges
to fluorine involves a ten-membered ring, but the dimer between the
same acceptor and the carbonyl oxygen is only an eight-membered
ring. Aside from dubious remarks which might be made regarding
steric effects, it can be seen that a quasi-resonance stabilization
is more probable the smaller the ring. This comment, of course,
does not refer to conjugated double bond systems.
Several conflicting opinions are available regarding the possi-
bility of distinguishing between amino hydrogen bonds to nitrogen
and those to oxygen with the resolving power of the instrument
being used in this research. Stanford and Gordy reported ex -
tremely small shifts of the absorption frequency of the acetylenic
hydrogen in phenyl acetylene when it was dissolved in either di-
methyl acetamide, dimethyl formamide or acetone. Corresponding
to heat of mixing calculations, they found that a smaller perturba-
tion was introduced in acetone solution than was present in the
solutions of the two amides, From this, they concluded that the
hydrogen bond to nitrogen is stronger than that to oxygen. Of
course, these data might just as easily have been interpreted as a
measurement of the relative electronegativities of the various
carbonyl groups involved. Stanford and Gordy assumed that the
amide nitrogen is more basic than is the amide oxygen. That this
is not necessarily the case has been demonstrated in this research.
Anzilotti and Curran determined the percentage of o-fluorophenol
molecules possessing internal hydrogen bonds from spectral and
dielectric constant measurements. Although the calculated elec-
tron density upon both atoms is the same, Curran17 found that the
oxygens in catechol were more strongly hydrogen-bound than was
the fluorine in the compound mentioned above. Buswell, Rodebush
and Roy14 had previously concluded that the hydrogen bonds to
nitrogen and oxygen in the amides are indistinguishable.
Upon dilution, the nitrogen-hydrogen stretching modes become
almost completely unperturbed. This fact follows from the disap-
pearance of the two important bands discussed above and the
appearance of a band at 2. 92 microns (3425 cm-1). There is an
apparent tendency for the lower frequency band to disappear first
upon dilution. Initially, this band is less intense than is the higher
frequency band. This is due to the fact that, though the formation
of a ring tends to stabilize the system as the magnitudes of the
frequency shifts show, the probability of obtaining the orientation
necessary to the formation of a ring is only one-half that of form-
ing a linear dimer, An explanation which does not consider the
entropy change upon dimerization is that resonance stabilization in
a ring system reduces the effective dipole moments of the absorbing
bonds, thus decreasing the intensity of the absorption.
It is interesting to note that the unperturbed nitrogen-hydrogen
stretch band is present in the pure liquid N-butyl difluoro amide.
As has been mentioned, this compound is expected to have the most
acidic hydrogen of the series other than the amino hydrogen. The
presence of the 2.92 micron (3425 cm-1) band is an indication of the
fact that a fraction of the available donor atoms is blocked with
respect to the amino nitrogen because of bonding through the active
The question of keto-enol tautomerism has not been satisfac-
torily settled. Because of the broadness of the bands in the three
micron region, one can postulate that perturbed hydroxyl stretches
actually exist. On the other hand, in extremely dilute solutions,
there is no evidence whatsoever of the existence of a true hydroxyl
stretching mode. This limits the existence of the enolic form to
the cases in which there is actual association. It may be stated that
upon the formation of a hydrogen bridge through an oxygen, there
is some doubt as to the position of the proton; that is, that the proton
resonates between the ironically and covalently bound states. It is
certain that the classical tautomeric mixture does not exist in dilute
It has been mentioned before that the cyclopeptides do not
absorb radiation in the range of the B band.51 Diketopiperazine,
during basic hydrolysis, gradually begins to absorb in the 6.40
micron (1562 cm"1) region. The band at 5.95 microns (1681 cm"1)
disappears and another band appears at 6.08 microns (1645 cm-1).
It was mentioned in Chapter I that the carbon-nitrogen bond length
in diketopiperazine is shorter than that of the corresponding bond in
acetamide. This implies that the carbon-nitrogen bond in the former
compound is of greater double bond character than it is in the latter.
Several resonance structures may be drawn for this compound. The
ordinary structure may be coupled with an excited state in which
both nitrogens are doubly bound in the ring. There is an alternative
hybrid which may lie between the two equivalent structures in which
only one of the amide -like groups is excited at a time. This is
roughly analogous to the Kekule structures written for benzene.
Resonance between two such structures would be expected to be com-
plete, involving a high stabilization energy.
It must be concluded, after consideration of the pronounced
possibilities for resonance hybridization, that the diketopiperazine
molecule behaves in a fashion analogous to that of the N, N-disubsti-
tuted acetamides. Thus, the band observed at 5.95 microns (1681
cm^-) is characteristic of both the carbon-nitrogen and carbon-
oxygen bonds. Upon attack by sodium hydroxide, this resonance is
destroyed and the ring is split. The two bands then appearing may
be identified as corresponding to the A and B bands in the N-nmoro-
The linear peptides behave in a fashion identical to that of the
corresponding amides. Lenormant51 states that a linear polyamide
of the nylon type absorbs very strongly at 6.36 microns (1572 cm"1)
and 6.0 microns (1667 cm-1). Upon methylation of the nitrogen,
only the very strong absorption at 6.05 microns (1653 em"I) appears.
The same author, with reference to the lactams, states that
they do not absorb in the B band range. The lactam of 6-amino-
hexanoic acid absorbs at six microns. Dilution data are not avail-
able for this compound. The occurrence of a symmetrized system
is hardly to be expected in this compound. A possible explanation
for the absence of the B band is that the carbon-nitrogen bond is
virtually one-hundred percent single bond in character. A reason
for this is not advanced. Room temperature treatment with one-tenth
normal sodium hydroxide followed by frequent checks of the spectra
would hardly solve this problem. The hydrolysis product is the
salt of an amino acid. The carbonyl group is too far removed to
enter into conjugation with the 6-amino group.
Succinimide poses a problem similar to that of caprolactam.1L
The B band is not found in the spectrum of this compound. This can
be explained by observation of the improbability of the fomation of
a double bond from the nitrogen in the presence of the competing
Lenormant51 has reproduced the spectra of several N-monosub-
stituted acetamide compounds. These include the sodium, mercury.
bromine and ethyl derivatives. Sodium acetamide does not posses
a band absorbing in the A region, but does absorb 6.37 microns
(1570 cnim). The di-acetamide salt of mercury absorbs at 6.23
and 6.33 microns (1605 and 1580 cm'1). N-bromo acetamide does
not have a band in the B region, but absorbs strongly at 6.06 microns
(1650 cm'l), As would be expected, N-ethyl acetamide possess
both A and B bands, absorbing at 6.00 microns (1667 cm-1) and
6.39 microns (1565 cm-l1)#
It has already been stated that the reason for the presence of
the A and B bands in the spectra of the N-monosubstituted amides is
that there is insufficient electronic density in the neighborhood of
the nitrogen atom to allow the symmetrization of the carbon-nitrogen
and carbon-oxygen bonds. The absence of the A band from the
spectrum of the sodium salt of acetamide implies that the carbonyl
bond does not exist in this compound. Unfortunately, these salts
are not totally unambiguous in structure. The reaction of metallic
sodium with acetamide is accompanied by a liberation of hydrogen.
One would deduce from this fact alone that a negative charge
resides upon the nitrogen. Further, the reaction of the sodium
salt with an alkyl iodide yields the N-alkyl acetamide, A most
elementary consideration of the electronegativities of nitrogen and
oxygen, on the other hand, leads directly to the proposal that the
structure of the salt involves a double bond from the nitrogen to
the carbon with the negative charge of the ion virtually localized
upon the oxygen. This latter fact and the spectral data seem to be
in agreement. In view of the tendency of the compounds of nitrogen
to undergo rearrangements in their reactions, one must suppose
that such a mechanism pertains in the formation of the N-alkyl
amides by the method described.
It is established, then, to a reasonable degree of certainty,
that the B band is characteristic of the valence stretch of nitrogen
essentially doubly bound to carbon. LenormantSI has also repro-
duced the spectrum of ethyl imido acetate. It shows only one absorp-
tion in the region under discussion. This band occurs at 5.95
microns (1681 cm-1). This must be attributed to the
carbon-nitrogen double bond. It will be recalled that a similar
deduction was made concerning bands occurring in the neighborhood
of 5.93 microns (1686 cm"1) in the spectra of certain of the dilute
N-disubstituted amides studied in the present investigation. The
reason for the higher frequency of absorption of this bond in the Imido
acetate as contrasted to that of the amido-anion of the sodium salt
should be obvious. The electronic density in the neighborhood of the
oxygen is reduced considerably by the reaction of the latter with an
ethyl carbonium ion. This inductively strengthens the nitrogen-car-
bon double bond with the resulting higher absorption frequency.
There seems to be little doubt; as to the structure of the bromo
derivative of acetamide. The studies of the Hofmann degradation of
amides seem to indicate that the N-bromo configuration is the proper
one. This would appear to be the more logical in view of the electro-
negativities of the elements involved. In acetamide itself, Davies
and Hallam20 found that the carbonyl band occurs at 1695 cm-1 (5.90
microns) but that upon dilution it shifts to 1714 cm-1 (5.80 microns).
Since Lenormant did not specify the conditions under which the spec -
tra referred to above were taken, one must presume that they are of
the pure compounds. The absorption of N-bromo acetamide .s far
displaced from that of acetamide. This shift to lower energies
appears to be anomalous in the light of the I effect attributed to
bromine; however, the effect of association in the pure compound
has not been considered. In this instance, the amino hydrogen is
extremely acidic. It can form a stable hydrogen bridge with the
ox-ygen of another amide. Further stabilization will occur with
cyclization. This bridge will tend to weaken the absorption fre-
quency of the carbonyl bond. It can be surmised that this weakening
effect predominates over the I effect of the bromine because the
former is not spatially insulated. Further, the inductive transmis-
sion of charge to the bromine atom is probably more effective over
the oxygen-hydrogen-nitrogen path than it is over the oxygen-carbon-
nitrogen path. The absence of the B band is self-explanatory.
The absorption spectrum of the di-acetamide salt of mercury
is produced by a structure which is quite as mysterious as is that
of the sodium salt of acetamide. This former salt is prepared by
fusion of acetamide with mercuric oxide and is accompanied by a
splitting out of water. This salt can occur in one of two possible
configurations or a combination of these. The first involves ionic
bonding through the oxygen and the second, covalent bonding through
the nitrogen. This deduction is based on the nature of the nitrites of
silver and sodium.33 The former, in the solid state, appears to
exist with the metal bound to the nitrogen, and the latter, with the
metal adjacent to the oxygen. It is suspected that the silver salt
is actually covalent in nature. Consideration of the behavior of mer-
cury leads to the plausibility of the suggested modes of bonding in
the di-acetamide salt of that element. The existence of rather dis -
placed A and B bands implies that a covalent salt of mercury exists.
In this condition, symmetrization is approached, but is not attained
because of competition of the two amide groups for the electrons of
In contract to succinimide which was discussed previously,
Lenormant51 found that IT, N-dibenzoyl hydrazine absorbs radiation
at 6.06 and 6.48 microns (1650 and 1543 cm"1). In this compound,
there is a tendency for the phenyl rings to act as an electron source
in the formation of negatively charged oxygen atoms and as a sink
with respect to this excited state. Thus, the carbonyl bond would
be expected to absorb at rather low energies. Further, there is the
possibility that as it acts as a sink, the phenyl ring can induce the
nitrogen to shift a pair of electrons partially into its bond with the
adjacent carbon. This accounts for the presence of the B band and
its rather low frequency.
Following the suggestion of Richards and Thompson, 59
Lenormant51 N-deuterated certain alkyl acetamides. As was men-
tioned in Chapter I1 the B band was shifted, but not in the proper
ratio. It is true that the reduced mass of the deuterated system
differs from that of the hydrogenated system. From the simplest
two-body approximation in which the absorption frequency is inverse-
ly proportional to the square root of the reduced mass, it is evident
that the system behaves properly. This, of course. assumes that
there is no change in the force constant, as there rightfully should
not be. Because of the quantitative inapplicability of the approximate
equation, no calculation of the shift is made on this basis.
It will be interesting to determine if the slightly divergent
series of compounds prepared by Cromwell, Miller, Johnson, Frank
and Wallace16 behave as would be predicted by the present postu-
lates. It should be recalled from Chapter I that the B band is
found in certain of the alpha-amino, alpha-beta unsaturated ketones.
This occurs in spite of the fact that the amino groups are disubstituted.
In several cases, an absorption characteristic of the double bond is
not observed. This tends to throw suspicion upon the analogy which
has been drawn with respect to the B band. The probable explanation
is that, the carbonyl bond and the carbon-carbon double bond being
conjugated, both are shifted to lower frequencies than would be
expected. This series of compounds, then, stands largely as an
externally consistent explanation of the lowered absorption frequency
of both the carbonyl bond and the nitrogen-carbon double bond. Here,
symmetrization is not possible because of the intervention of a
carbon-carbon single bond.
Much discussion of the relationship between the shifts of ab-
sorption frequencies of bonds through association and the chemical
heats of formation of hydrogen bonds has appeared in the literature
Several excellent reviews exist; however, the most recent and com-
plete is that by Davies. 19 It is the consensus of opinion that no
simple linear relationship exists between these shifts and the heats
of formation of hydrogen bonds. Davies has mentioned that numerous
forces must operate in the formation of a hydrogen bond. Among
these are dispersion forces and dipole-dipole forces. It is seen that
neither of these effects would necessarily be reflected in absorption
The largest frequency shifts observed in the present research
were on the order of 360 cm-1. This is equivalent to I. 03 kcal./bond-
mole. Davies and Hallam, 20 who proposed a cyclic trimer for the
associated forms of acetamide liquid, state that such a structure
would involve a heat of formation of about 1.1 kcal./bond-mole.
They report, however, that a non-spectroscopic determination of
the energy involved in hydrogen bond formation in aniline-benzo-
phenone mixtures yielded a value of 2.0 kcal./bond-mole. The heat
of formation of the formanilide dimer is reported to be Z. 9 kcal./
A more nearly comparable case, that of the acetic acid dimer,
is reported by Pauling55 to involve an energy of 8. Z kcal./bond-
mole. Further discussion by Pauling indicates that the energy of
the hydrogen bond formed in ammonium fluoride crystals is about
5 kcal./bond-mole. It is seen that poor agreement is reached in the
There seems to have been no discussion of the part which the
shift of the absorption frequency of the bond involving the donor atom
must play in the calculation of bond energies. If one were to con-
clude that these shifts are to be added, calculation of the energies
of hydrogen bonding from the present data would range from 0.54
kcal./bond-mole in the linear diners to 1.1 kcal./bond-mole in the
cyclic dimers. One might accept these values except for the very
unfavorable comparison to that mentioned above in the case of
acetic acid. It would seem that cyclization would, in the present
case, introduce sufficient stabilization to double or treble the bond
energies observed in the aniline-benzophenone dimer. V'irtz has
calculated that the resonance stabilization of hydrogen bonds of
amides is about 1 kcal. /bond-mole.74
Davies and Sutherland21 have reported the difference in the
frequencies of the hydroxyl stretch and the carbonyl stretch between
monomeric and dimeric acetic acid to be 448 cm-1 and 67 em-,1
respectively. The average shifts observed in the present research
are on the order of 360 cm-1 and 35 cm 1 for analogous bands.
Using as a basis the heat of association of acetic acid reported above
and the ratio of the major shift reported for acetic acid to the aver-
age major shift of the N-monosubstituted acetamides, it is calcu-
lated that the average heat of association of the latter compounds is
6.5 kcal. /bond-mole.
Badger and Bauer and, later, Badger, utilized a similar
method to obtain spectrographically the heats of association of
several alchols. It was reported that a simple graphical interpola-
tion method yielded values with an uncertainty of 500 cal. /bond-
mole. Although there is partial justification of the use of this
method in the present case, it is not to be taken as an absolutely
reliable measurement. Unfortunately, too little is known of the
more subtle forces operating between molecules.
It is believed that the hypothesis of .ccidental degeneracy
stated at the bcginnin-' of this chapter has been maintained in the
presence of experimental data obtained both in the present research
and in investigations reported in the literature. Though the expla-
nations at times seem to be rather devious, it must be observed
that the phenomena dealt with here are by no means simple and
have for many years resisted reduction to mathematical terms.
1. It has been demonstrated on the basis of the present
research and in accord with previous conclusions that the 6.40
micron (1562 cm-1) band of the N-monosubstituted amides is attrib-
utable to the stretch of the carbon-nitrogen partial double bond.
2. Previously unreported absorptions have been found in the
spectra of the N, N-disubstituted acetamides studied. These bands
occur at 2.91 microns (3436 cm-1). 3.04 microns (3289 cm-"1) and
5.70 microns (1754 cm-1). The former bands are considered to be
overtone bands, but the latter is unclassified.
3. It is postulated that the absence of the 6.40 micron
(1562 em-1) band from the spectra of the N,N-disubstituted amides
is explained by the accidental degeneracy of this band and the car-
bonyl stretch band. This is the result of a nearly complete symme-
trization of the bond energies of the amide group.
4. By means of the shifts of absorption frequency of the
carbonyl bond upon halogenation of the N-substituted amides studied,
the fluorine atom has been shown to exert a greater electronegative
inductive effect when bound to carbon than does chlorine. This
is in agreement with previous determinations of the magnitude and
direction of the inductive effect.
5. It has been shown that hydrogen bonding exists in the
N, N-disubstituted aides in the case in which a halogen-activated
hydrogen atom is present. Furthermore, it is found that hydrogen
bonding in the compounds studied occurs through oxygen, nitrogen
and fluorine atorms.
6. A splitting of the 6.00 micron (1667 cm-1) band of N-dibutyl
dichloro, N-dibutyl chlorofluoro and N-dibutyl difluoro compounds
has been observed upon dilution.
1. Aelion and Lenormant, Compt. rend. 224, 904 (1947).
2. American Petroleum Institute Research Project 44, "Catalog of
Selected Infrared Absorption Spectrograms", National Bureau of
Standards, Washington, D. C.
3. Anauthakrishnan, Proc. Indian Acad. Sci. SA, 200 (1937).
4. Anzllotti and Curran, J. Am. Chem. Soc. 65, 607 (1943).
5. Badger, J. Chem. Phys. 8, 288 (1940).
6. and Bauer, ibid., 5, 839 (1937).
7. Barnes and Bonner, J. Chem. Education 14, 564 (1937).
8. ibid., 15. 25 (1938).
9. R. B., Gore, R. C., Liddel, U. and Williams, V. Z..
"Infrared Spectroscopy"', Reinhold Publishing Co., New York,
10. Bates and Hobbs, J. Arn. Chem. Soc. 73, 2151 (1951).
11. Batuev, Compt. rend. acad. sci. U. R. S. S. 53, 507 (1946).
12. Buswell, Downing and Rodebush, J. Am. Chem. Soc. 62, 2759
13. Krebs and Rodebush, J. Phys. Chem. 44, 1126 (1940).
14. Rodebush and Roy, J. Am. Chem. Soc. 60, 2444 (1938).
15. Copely, Zellhoefer and marvel, ibid., 60, 266 (1938).