CHEM. LIB.
QD
341
K2
B 453707
P485
The University of Chicago
Founded by JOHN D. ROCKEFELLER
Stereoisomeric Chlorimido Ketones
A DISSERTATION
SUBMITTED TO THE FACULTIES OF THE GRADUATE
SCHOOLS OF ARTS, LITERATURE, AND SCIENCE,
IN CANDIDACY FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
DEPARTMENT OF CHEMISTRY.
By Peter P. Peterson
EASTON, PA.:
PRESS OF THE ESCHENBACH PRINTING CO.
1911.
The University of Chicago
Founded by JOHN D. ROCKEFELLER
Stereoisomeric Chlorimido Ketones
A DISSERTATION
SUBMITTED TO THE FACULTIES OF THE GRADUATE
SCHOOLS OF ARTS, LITERATURE, AND SCIENCE,
IN CANDIDACY FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
DEPARTMENT of CHEMISTRY.
By Peter P. Peterson
EASTON, PA.:
PRESS OF THE ESCHENBACH PRINTING CO.
1911.
Chemical Library
QI
341
.K2
P485
Stereoisomeric Chlorimido Ketones'
In 1803 Stieglitz and Earle², in the course of their work
on the theory of the "Beckmann Rearrangemedt," discovered
the first representatives of a new group of steriosomeric
nitrogen derivatives, in which the stereoisomerism depends
on the "syn" and "anti" positions of a chlorimide group in
chlorimido esters:
R-C-OR
||
NCI
4
R-C-OR
CIN
Later Stieglitz and Hale³ obtained a second pair of such
stereoisomeric chlorimido esters and showed under what con-
ditions the labile form can be converted into the stable variety.
Finally Hilpert, by preparing five other pairs of stereoiso-
mers and determining the conditions for the transformation
of the members of a pair reversibly into each other, estab-
lished their existence on a sufficiently broad basis. These
isomers are interesting, partly because of the simple charac-
ter of the radical (C1+) attached to the nitrogen atom, and
partly because of the light they shed on the question of the
"Beckmann Rearrangement.
115
The stereoisomers heretofore prepared, have in all cases been
chlorimido esters. They made probable the existence of
stereoisomerism also for chlorimido ketones,
R-C-R'
and
NCI
R-C-R'
11
CIN
comparable with the stereoisomeric ketoximes and hydrazones
of Hantzsch and Werner and of the former's collaborators.
The isolation of such stereoisomers seemed desirable, not only
as confirming the conclusions reached in the work on the
chlorimido esters, but also because it was thought their prepara-
¹ See a preliminary paper by J. Stieglitz and P. Peterson: Ber. d. chem. Ges., 43,
782 (1910).
2 Am. Chem. J., 30, 399 (1903).·
8 Unpublished results.
+ Am. Chem. J., 40, 155 (1908).
5 See a later paper, to be published by Stieglitz, covering unpublished results with
Hale and Eckstein.
4
tion would settle, experimentally, important questions connected
with the "Beckmann Rearrangement." In the rearrange-
ment of ketoximes, R-C(: NOH)R', under the influence of
phosphorus pentachloride, many chemists are inclined to as-
sume¹ the formation of such chlorimides, R—C( : NCI)—R',
as intermediate products, the first products isolated in the
reaction being arylimidoacyl chlorides. These are supposed
to be formed according to the scheme
R-C-R'
CIN
C1—C—R'
11
R-N
According to the theory advanced by Stieglitz, such chlor-
imides would not be intermediate products; the rearrangement
is supposed to be due to the anhydrizing power of phosphorus
pentachloride and the other reagents used to accomplish the
rearrangement, an intermediate product, containing uni-
valent nitrogen, being considered to be the actual substance
undergoing the rearrangement. For instance, we would have:2
R-C-R'
R'
do
+ HCI → R_C_C
NOH
and
R'
R-C
-C1
H₂O
HNOH
R'
da
R-C-CI
N
Λ
1 Cf. Hantzsch: Ber. d. chem. Ges., 35, 3579 (1902).
HNOH
R-C-CI
NR'
2 In confirmation of these views, Stieglitz and Reddick have recently found that
triphenylmethylhydroxylamine, (CH5)зC.NHOH, whose structure corresponds to
that of the assumed intermediate addition product with hydrogen chloride, undergoes
the "Beckmann rearrangement" with remarkable ease and smoothness under the
influence of a dehydrating agent (phosphorus pentachloride was used). This result is
the more noteworthy as it forms a striking contrast to the fact that hitherto all at-
tempts to effect a rearrangement of chlorimidobenzophenone, (CH5)2C : NC1, have
been unsuccessful (Stieglitz and Watkins). G. Schroeter has recently (Ber. d. chem.
Ges., 44, 1201 (1911)) brought valuable experimental evidence supporting the above
theory. Experiments have been undertaken in this laboratory toward establishing a
connection between the rearrangement and the electrical charges on the atoms involved.
It is thought that, in the final instance, the significance of the rôle of the univalent
nitrogen atom will be found to lie in such electrical relations. In a later paper these
views will be discussed more elaborately and due consideration given to the confirma-
tive experimental work of Schroeter, as well as to the criticisms of J. Stieglitz,
Montagne (Ibid., 43, 2014 (1910)) and others.
сл
At the suggestion and under the direction of Professor Stieg-
litz, I undertook the problem of isolating, if possible, such
stereoisomeric chlorimido ketones. I have found three pairs
of stereoisomers, viz., stereoisomers of:
(a) Chlorimido-p-chlorbenzophenone,
5
and
NCI
||
CIN
(b) Chlorimido-p-methoxybenzophenone,
5
CH¸—С—СН₁OCH₂
||
NCI
and
C₂H5 C—CH₂OCH¸
4
CIN
and (c) Chlorimido-p-chlor-p-methoxybenzophenone,
CICH-C-CH₂OCH,
||
NCI
and
CICH-C-CH₂OCH,
3
[]
CIN
In each case the structural identity of the a and ẞ forms was
established by converting each, by dry hydrogen chloride, into
the hydrochloride of an imidobenzophenone, which gave with
water the same ketone; a mixture of the ketone obtained
from the a and that obtained from the ẞ varieties melted at
the same point as either separately. The reactions are:
and
R-C-R'
11
+ HCl
NCI
R-C-R'
+ H₂O
NH,C
R-C-R'
||
+ Cl2
NH,C1
R-C-R'
+ NHẠC1
That the two forms of a pair are not crystal or physical
modifications is shown by the facts that each form has its
own characteristic solubility, each form is recovered un-
changed from its solutions, from its liquid form, when fused,
and from its vapors-even in the presence of added crystals
of the other form. Further, each form persists even up to a
temperature of 100°, and a mixture of approximately equal
quantities of the two forms invariably melts lower than either
form alone. One determination of the molecular weight
1 Vide Stieglitz and Earle and Hilpert: Loc. cit.
1
6
of the higher melting of a pair of stereoisomers was made,
namely of the a form of chlorimido-p-methoxybenzophenone;
the higher melting form should in case of polymerization of
either form, be the polymer. The molecular weight deter-
mination showed it to have the normal molecular weight for a
nonpolymerized body.¹
The change of one stereoisomer into another was observed
only once; the lower melting (8) form of chlorimido p-chlor-
benzophenone was found, after being kept for three months
during a very hot summer, to have become completely con-
verted, spontaneously, into the higher melting, stable or a form.
Efforts to induce the same change by artificial means, e. g., by
treatment of the two forms with chlorine, were not successful,
the stereoisomers being unusually persistent, as compared with
the chlorimido esters. It may be recalled here that Hantzsch
also found the stereoisomeric hydrazones of these ketones to
be extremely stable, resisting transformation into each other
by the agents which are usually successful.³
2
The stereoisomeric chlorimido ketones, thus far, have not
given the slightest indication of suffering the "Beckmann
rearrangement" which they were supposed to undergo so
readily by those who had assumed them as intermediate
products in the action of phosphorus pentachloride on ketox-
imes. They can be heated to 100° a short time without under-
going any such rearrangement. It is quite obvious, then, that
the view that they form such intermediate products is now
wholly untenable.
EXPERIMENTAL PART.
Chlorimidobenzophenone, (CH),C: NCI.-The first chlori-
¹ Stieglitz and Earle and Hilpert carried out molecular weight determinations for
both stereoisomers of a number of pairs of stereoisomeric chlorimido esters and in every
case found the molecular weight of each form normal.
2 Vide Stieglitz, Hale and Hilpert.
3 By way of the return to the imidohydrochloride of the chlorimido-p-chlorbenzo-
phenone, the a form was, by subsequent conversion back into a chlorimide, converted
into a mixture of the a and ẞ forms-this process being an aid in the preparation of
pure material (the hydrochloride of the imidobenzophenone) for the formation of the
chlorimides. As the higher melting a form is always obtained in large excess, the cir-
cle enables one to use the material over and over again to accumulate the lower melt-
ing ẞ modification. These actions, while proving the structural identity of the stere-
oisomers, do not accomplish any direct change of one stereoisomer into the other.
7
mide prepared was that of benzophenone. Here, of course, no
stereoisomers were expected, but the work was done for the
purpose of developing a good method for preparing chlor-
imido ketones. Benzophenone was prepared according to
the Friedel and Crafts¹ reaction and converted into benzo-
phenone dichloride by means of phosphorus pentachloride
according to Kekulé and Franchimont's² directions. Benzo-
phenone imidohydrochloride was prepared from the dichloride
with the aid of urethane by the method of Hantzsch and
Kraft,³ a yield of forty-two per cent. of the theoretical being
obtained. The best yield was obtained when three mole-
cules of urethane were heated with one molecule of benzophenone
dichloride. The reaction is a very slow one when carried
out at 140, but the product turns brown, indicating a decom-
position, if the temperature is raised much higher. At this
temperature the mass gives off gases (carbon dioxide and
alkyl chloride) and crystallizes very slowly, and at least three
hours are required for completion of the action. Longer heat-
ing does not increase the yield.
The chlorimide of benzophenone was prepared from the
imidohydrochloride by treatment of the base with hypo-
chlorous acid, much in the same way as the chlorimido esters
are prepared. The action is represented by the equation
5/2
4
(CH),C NH + HOCI → (C,H₂)₂C: NCI + H₂O
The hypochlorous acid was prepared from sodium carbonate
according to the directions of Erlenmeyer and Lipp. A solu-
tion of two equivalents of the acid was prepared and one
equivalent of potassium bicarbonate added to it. Then, while
the solution was at freezing temperature, the solid imidohydro-
chloride was added. For instance, 8.5 grams of the hydro-
chloride was added to a mixture of 4 grams of potas-
sium bicarbonate and of hypochlorous acid (prepared from 17
grams of dry sodium carbonate dissolved in 250 cc. of water
and 11.5 grams of chlorine). The solution became milky as
1 Ann. chim phys., [6] 31, 510.
2 Ber. d. chem. Ges., 5, 909 (1873).
3 Ibid., 24, 3516 (1891).
5
4 Ann. Chem. (Liebig), 219, 185 (1883).
5 Graebe: Ber. d. chem. Ges., 35, 2750 (1902).
8
the imidohydrochloride dissolved, and a solid gradually crys-
tallized out. After about ten minutes, during which the
mixture was shaken constantly in the freezing bath, chloro-
form was added to dissolve the crystals. The two layers were
separated by means of a separatory funnel, the stoppers of
which were carefully greased to prevent any scratching.
This precaution was deemed necessary because some nitrogen
trichloride was liable to be produced in the reaction and the
pungent odor of the mixture indicated its presence.
On evaporation of the chloroform, a yellowish crystalline
mass was left. This mass was purified by crystallization
from chloroform and ligroin and by recrystallization from warm
ligroin. A yield of chlorimide representing seventy per cent.
of the theoretical was thus obtained. The melting point of
the crystals was 37°. The chlorimidobenzophenone thus ob-
tained, when treated with potassium iodide in acid solution,
sets iodine free quantitatively according to the equation
(CH)₂C: NC1 + 2HI = (CH)½C : NH¿Cl + 2I
This was used as a means of analysis, the iodine set free being
titrated with sodium thiosulphate.
0.1872 gram substance gaye 0.0308 gram chlorine.
C1
Calculated for
C13H10NCI
16.44
Found
16.46
The chlorimide (0.4 gram) was dissolved in 25 cc. of ligroin,
and dry hydrogen chloride was passed into the solution; a fine
white precipitate resulted which analysis proved to be the
regenerated imidohydrochloride.¹ The yield was quantita-
tive. Analysis of it, by the silver nitrate method, gave the
following results:
0.0776 gram substance gave 0.0128 gram chlorine.
C1
Calculated for
C13H12NC1
16.30
Found
16.49
Treatment of the imidohydrochloride with hot water gave ben-
1 Vide the equation, page 5.
9
1
zophenone quantitatively. That the substance thus obtained
was the ketone was proved by the fact that it melted at 46°,
the melting point of synthetically prepared benzophenone,
and when it was mixed with the synthetic product its melting
point was not depressed.
Hydrochloride of Imido-p-chlorbenzophenone,
5
CH¿—C( : NH₂Cl)—C¸H¸C1.—Chlorimido-p-chlorbenzophenone
was prepared in the same way as chlorimidobenzophenone,
starting from p-chlorbenzophenone. This ketone was prepared
according to Hantzsch and Kraft's² directions, and when treated
with phosphorus pentachloride according to Overton's³ method
it gave p-chlorbenzophenone dichloride. Fractionation of the
dichloride was carried out in a Brühl apparatus, the pure sub-
stance distilling at 192° under 12 mm. pressure. A yield of 88
per cent. of the theoretical was thus obtained.
To prepare p-chlorbenzophenone imidohydrochloride one
molecule of p-chlorbenzophenone dichloride was mixed with
three molecules of urethane and the whole heated to 140
160° on the metal bath until the entire mass solidified. This
was then extracted four times with benzene to free it from
unchanged urethane and p-chlorbenzophenone which is formed.
Recrystallization, by dissolving the crude residue in chloro-
form and precipitating with ligroin, gave the pure white hydro-
chloride. Owing to its difficult solubility in chloroform, it
was found impractical to use this method for the purification
of large quantities of material. I found that the imidohydro-
chloride did not need to be purified before conversion into the
chlorimide. More nitrogen trichloride was formed, but this
did not interfere with the work, and is not dangerous if the
stopcocks are kept well greased with vaseline. The chlor-
imide could then be converted back into the imidohydrochlor-
ide quantitatively by the use of hydrogen chloride. This was
found to be an easier method of preparing the pure imido-
chloride than the method of precipitation spoken of above,
and it was used exclusively after the first preparation. A
yield of 53 per cent. of the theoretical was obtained of the
1 Vide the equation, page 5.
2 Loc. cit.
3 Ber. d. chem. Ges.; 26, 28, (1893).
ΙΟ
crude material, and this was reduced to 42 per cent. in pass-
ing through the chlorimide. Analysis of the pure substance
gave the following results:
0.1172 gram substance gave 0.0167 gram Cl, and 0.1224
gram substance gave 0.0175 gram Cl.
Calculated for
C13H₁₁NC12¹
Ionizable Cl
14.09
Stereoisomeric
5
Found
I
II
14.24
14.30
Forms of Chlorimido-p-chlorbenzophenone,
and
-a- and B-p-Chlor-
CIN
NCI
benzophenone chlorimide were prepared from p-chlorbenzo-
phenone imidohydrochloride by treatment with hypochlorous
acid. A solution of two equivalents of hypochlorous acid
was prepared as described above and the solid imidohydro-
chloride added to it. When the crude salt, containing am-
monium chloride, was used much effervescence occurred.
Great care had to be taken in handling the product, owing to
the formation of nitrogen trichloride. No explosions ever did
occur, but chemists trying the method must work very cau-
tiously. After the imidohydrochloride was added, the flask
was shaken vigorously in the freezing mixture for about ten
minutes. The mixture was then extracted with very little
chloroform and the chloroform solution drawn off in a
separatory funnel. The chloroform was evaporated by a blast
of air, a semisolid mass being left. This was extracted with
warm ligroin. A solution of a mixture of the two forms of
the chlorimide was thus obtained (about 80 per cent. of the
theoretical yield). The two forms were separated as follows:
Crystals obtained by cooling the ligroin extract were dissolved
in as little chloroform as possible and about four volumes of
ligroin added. This precipitated a mass of crystals which
melted at 101°-102°. Recrystallization from warm ligroin
raised their melting point to 104°. This form of the com-
pound is called the a form. It appears in thin plates.
1 Only the ionizable chlorine is determined, the method used being the titration
of the salt with silver nitrate.
I I
The chloroform-ligroin filtrate was then cooled to about
-10° for a few minutes. Crystals melting at 46°-48° sep-
arated out. On repeated crystallization of this product from
ligroin its melting point was finally raised to 55°. This form
of the compound is called the ẞ form. It appears as thin
prisms or needles. Analysis of the two varieties by the potas-
sium iodide method gave the following results:
0.1592 gram of the a form gave 0.0225 gram active chlorine
and 0.1190 gram gave 0.0168 gram active chlorine.
Active Cl
Calculated¹ for
C13H9NC12
14.18
Found
I
II
14.14
14.12
0.1274 gram of the ẞ form gave 0.01817 gram active chlorine
and 0.0940 gram gåve 0.0136 gram active chlorine.
Active Cl
Calculated for
C13H9NCI₂
14.18
Found
12
112
14.26
14.45
The identity of the structure of the two compounds was
proved by converting them both into the imidohydrochloride
and then into the ketone as follows:3
Three-tenths gram of the pure a form was dissolved in about
25 cc. of ligroin and dry hydrogen chloride passed into the
solution to saturation. The imidohydrochloride was thus
precipitated. The ligroin was decanted and the last traces
blown off by a blast of air. Then the salt was treated with
warm water. This gave 0.223 gram (89 per cent. of the
theoretical yield) of the ketone, melting at 75°.5-76°. Mix-
tures of this substance with the synthetically prepared ketone
and with the ketone obtained from the ẞ form melted at the
same temperature.
Three-tenths gram of the pure ẞ form was treated in the
same way and gave 0.230 gram of the same ketone as was
obtained from the a form.
1 Only one of the two chlorine atoms is determined, it alone being active toward
the hydroiodic acid which was used in the analysis.
2 In the preliminary report by Stieglitz and Peterson, as the result of errors in
calculation 14.19 and 13.97 per cent. was reported.
3 See page 5.
I 2
These experiments show conclusively that the substances
are not structural isomers and are the same except in the con-
figuration of the radical (NCI).
The following experiments were carried out to exclude the
possibility of crystal isomerism. One-tenth gram of the a
form was heated to 120° and cooled slowly. At 85° solidifica-
tion took place very rapidly. One-tenth gram of the ẞ form
was heated to 60°, then cooled to 45° and kept at that tem-
perature for some time. No crystals separated out until
inoculated with a or ẞ crystals. The melting point was then
found to be 54°; 55° is the melting point of the purest ẞ form.
A trace of the ẞ form probably was in the a material and
caused crystallization. The melting point shows that the
substance persisted in the form. One-tenth gram of each
form was then brought into solution, separately, so that each
solution was saturated at 25°. Cooling of the solutions then
gave only the crystals from which the solutions were originally
made, as melting point determinations proved. The crys-
tals were again brought into solution, the solutions super-
cooled and inoculated with crystals of the other variety.
Again, nothing but the crystals from which the solutions
were originally made were obtained, as proved by their melt-
ing points and habit.
ß
Hydrochloride of Chlorimido-p-methoxybenzophenone,
CH,C(: NH₂CI) CH₂OCH,.-p-Methoxybenzophenone was pre-
pared according to the method of Gattermann, Ehrhardt, and
Maisch.¹ The first preparation was made from anisyl chlor-
ide and benzene, but the yield was small. Benzoyl chloride
and anisole gave, by the same treatment, almost a quantitative
yield. This method was therefore used in subsequent prepara-
tions. It was found best to dissolve one equivalent of benzoyl
chloride in ten times its volume of carbon disulphide and to
add aluminium chloride to the mixture in about the same
weight as the benzoyl chloride. A little more than one equiv-
alent of anisole was then allowed to run slowly from a dropping
funnel into the mixture. The reaction was complete as soon
as all the anisole had been added. The carbon disulphide
1 Ber. d. chem. Ges., 23, 1204 (1890).
13
was then distilled off, the aluminium chloride decomposed
with water acidulated with hydrochloric acid, and the ketone
extracted with ether. Recrystallization from high-boiling
ligroin was the best method found for the purification. The
pure substance melted at 62°.
The ketone dichloride was obtained by the Hantzsch and
Kraft¹ method, except that it was not purified. It was found
that purification was not necessary before treatment with
urethane. The excess of phosphorus pentachloride and the
phosphorus oxychloride were distilled off in a vacuum at
about 100°. The residue was treated directly with urethane.
This gave an impure hydrochloride which could be readily
purified. For instance, 20 grams of p-methoxybenzophenone
and 20 grams of phosphorus pentachloride were heated in a
flask on a metal bath to 140°-160° for three hours. The
phosphorus oxychloride was then distilled off in a vacuum.
The residue was treated with 18 grams of urethane and the
temperature again raised to 160° for several hours. The
viscous mass thus obtained was extracted several times with
benzene to remove any urethane, ketone and dichloride.
white solid remained. The crude yield was 17 grams, or 73
per cent. of the theoretical. The substance melted above
170° with decomposition. When treated with warm water
it gave a substance melting at 55°-58°, which, when mixed
with synthetically prepared ketone (melting at 62°), melted
a little higher than before. These facts indicated that the
substance was impure imidohydrochloride of p-methoxybenzo-
phenone. The hydrochloride was obtained in a pure state
by converting the crude salt into the corresponding chlor-
imide, which by treatment with hydrogen chloride in ligroin
solution forms the imidohydrochloride again. The analysis
of the salt, purified in this way, gave the following results:
0.2390 gram substance gave 0.0339 gram chlorine.
Cl
Calculated for
C14H14ONCI
14.48
Found
14. 19
A
1 Loc. cit.
14
The salt is decomposed into the ketone (melting at 62°)
and ammonium chloride on treatment with water.
Stereoisomeric Chlorimides of p-Methoxybenzophenone,
5
CH₂-C-CHOCH,
3
CH-C-C,H,OCH,
3
and
-The chlor-
NCI
CIN
imides of p-methoxybenzophenone were obtained from the im-
pure imidohydrochloride by treatment with hypochlorous acid.
A solution of hypochlorous acid was prepared in the same way
as described for the preparation of benzophenone chlorimide,
and the solid imidohydrochloride of p-methoxybenzophenone
was added to the solution. When three equivalents of hypo-
chlorous acid were used the product was gummy and hard to
handle. When only two equivalents were used, the product
was crystalline and gave a higher yield of the low-melting
variety of the chlorimide, the one which is the more difficult
to isolate. Use of the pure imidohydrochloride gave a still
more satisfactory result. The rapid evaporation of the chloro-
form used for extracting the chlorimides from the hypochlorous
acid solution also seemed to give a larger yield of the low-
melting variety. In one preparation 10 grams of imidohydro-
chloride was treated with the equivalent of two molecules of
hypochlorous acid. The chlorimides were extracted with chloro-
form and the chloroform evaporated rapidly by a blast of
air. The mass was fractionated by dissolving it in as little
chloroform as possible and reprecipitating part of the solid
with about four volumes of ligroin. This gave a substance
consisting almost entirely of the a variety. After several
precipitations from chloroform by ligroin and recrystalliza-
tions from warm ligroin alone, the true melting point of the
a-variety was found to be 90°. These crystals were thin, six-
sided plates. The first chloroform-ligroin filtrate from the
precipitate of the a form was cooled to -10° and the beaker
scratched. Fine white crystals came out; they were found
to melt at 40°-42°. These crystals were dissolved in ligroin
in the proportion of 1 gram of crystals to 25 cc. of ligroin and
the solution set in the ice box for two days in a stoppered
bottle. One very large asymmetric crystal of the ẞ form and
15
several flat ones of the a variety appeared and could be
easily separated mechanically. The melting point of the
large crystal was found to be 54°, which was taken as the true
melting point of the ẞ form. A mixture of the two forms
melted at 40°, which is far below the melting point of either
form alone. Analysis gave the following results:
0.1249 gram of the a form gave 0.0177 gram chlorine and
0.1792 gram gave 0.0254 gram chlorine.
Cl
Calculated for
C14H12ONCI
14.44
Found
I
II
14.15
14.19
0.1363 gram of the ẞ variety gave 0.0198 gram of chlorine.
C1
Calculated for
C14H12ONCI
14.44
Found
14.52
Conversion of the chlorimides back into the imidohydro-
chloride and the ketone, similar to that carried out with the
stereoisomers of chlorimido-p-chlorobenzophenone, was effected
as follows: 0.335 gram of the a form was dissolved in chloro-
form and ligroin and dry hydrogen chloride passed into the
solution to saturation. A yellow, sticky mass which smelled
strongly of chlorine separated out. The filtrate was evapo-
rated and gave but a small residue. Treatment of the mass
with warm water dissolved it, and when the solution cooled a
white crystalline solid, p-methoxybenzophenone, separated
out. After about two hours the crystals were collected on a
filter and dried on a clay plate. A yield of o.263 gram of the
ketone, or 83 per cent. of the theoretical, was obtained. Its
melting point was found to be 59°.5-60°. Then 0.259 gram
of the ẞ form was subjected to the same treatment. A yield
of 0.200 gram of the same ketone, or 89 per cent. of the theo-
retical, was obtained. The ketone obtained from the ẞ form,
and mixtures of this ketone with the preparation obtained
from the a form and with synthetically prepared p-methoxy-
benzophenone, all melted at 59°-60°. There can be no ques-
tion, therefore, of the identity of the structure of the two
chlorimides.
16
Attempts to convert either variety of the chlorimide di-
rectly¹ into the other were made but without success. First
1.5 grams of the substance was heated to 100° for three hours.
Its melting point was thereby lowered from 90° to 85°-87°.
A mixture of the product with some pure a form melted at
88°-89°. Consequently very little transformation, if any,
took place even at 100°. Then 0.1 gram of the ẞ form was
heated in the same manner; the melting point of the substance
was lowered from 54° to 47°. A very small proportion of
a form mixed with the ẞ form lowers the melting point of the
latter to 40°. There could have been very little change,
therefore, if any, of the ẞ into the a form. Finally o. 1 gram
of the substance was heated for three hours in an atmosphere
of dry chlorine² in a sealed tube. The mass became gummy,
and after having been crystallized from chloroform, decom-
posed slightly at 175° but did not melt. Some transforma-
tion must have occurred, but not the transformation into the
a form. The investigation of this reaction will be continued.
The molecular weight of the a form was determined by the
cryoscopic method: 0.0903 gram substance dissolved in
11.58 grams benzene lowered the freezing point o°.1675; 0.4791
gram in 11.58 grams benzene lowered the freezing point o°. 835.
Mol. wt.
Calculated for
C14H12NOCI
242.5
I
246.8
Found
II
262.6
The theoretical constant, 53, for the lowering of the freezing
point of 100 grams of benzene by one gram molecule of solute
was used, as in previous determinations with chlorimido
esters. The experimental constant, 49, would give the weights
226.5 and 238.
4
p-Chlor-p-methoxybenzophenone, CICH-CO—СH₂OCH¸.·
The method used for the preparation of p-methoxybenzophe-
none was also used for the preparation of p-chlor-p-methoxy-
benzophenone. Thirty-two grams of dry aluminium chloride
1 Indirectly the two forms may be converted into each other by treatment with
hydrogen chloride and subsequent change of the regenerated hydrochloride of the
imidomethoxybenzophenone into the chlorimides by means of hypochlorous acid.
2 Chlorine effects the reversible change of the labile stereoisomers of the chlor-
imido esters into the stable forms.
17
was pulverized in a hot mortar and added to a mixture of
100 cc. of carbon disulphide and 32 grams of freshly distilled
p-chlorbenzoyl chloride.¹ The mixture was then cooled to
and kept at that temperature during the addition of twenty
grams of anisole. The anisole was added very slowly, only
one or two grams at a time, and the mixture was well shaken
between the additions. Hydrogen chloride escaped very
regularly. When the action ceased at this temperature, the
mixture was allowed to warm up to room temperature. The
carbon disulphide was distilled off and a mixture of ice and
water, acidulated with hydrochloric acid, was added to the resi-
due to decompose the aluminium chloride. During this opera-
tion the flask was kept in an ice-salt mixture. The gummy
mass resulting was extracted with benzene, and the solution
shaken five times with sodium hydroxide to remove any
p-chlorbenzoic acid. The benzene was then evaporated
and the residue crystallized from hot alcohol as follows: The
whole mass was dissolved in the least possible quantity of
boiling alcohol and then the solution was cooled. Thirty-two
and a half grams of crystals separated out. The filtrate was
added to an equal volume of cold water. A further precipita-
tion occurred. This fraction was then crystallized from
alcohol. The entire yield was 37 grams, or 82 per cent. of
the theoretical. The melting point is 125°. Analysis of the
substance gave the following results:
2
0.2323 gram substance gave 0.5797 gram CO₂ and 0.0980
gram H₂O.
4
C
H
2
Calculated for
C14H11O2CI
68.18
4.49
Found
68.06
4.69
Hydrochloride of Imido - p- chlor-p-methoxybenzophenone,
CICH₁-C(: NH₂C1)—CH₂OCH,.-This was prepared from
the ketone as follows: p-Chlor-p-methoxybenzophenone was
heated with one equivalent of phosphorus pentachloride on the
metal bath to 160° for three hours. The phosphorus oxy-
chloride was then distilled off in a vacuum at a temperature
1 This was prepared from Kahlbaum's p-chlorbenzoic acid, melting at 237°.
18
of about 100°. Without further purification this crude
product was dissolved in chloroform, and ammonia, dried over
sodium hydroxide, run into the solution to saturation.
fine white mass, which was mainly ammonium chloride, separa-
ted out. This was filtered off and dry hydrogen chloride
run into the filtrate to saturation. A yellowish precipitate
was thrown down. The chloroform was decanted and the
precipitate washed several times with chloroform and finally
with ligroin. It was then dried in a vacuum over paraffin.
The substance thus obtained gave low values for chlorine
content for the hydrochloride of imido-p-chlor-p-methoxyben-
zophenone, but its behavior leaves no doubt that it consists
chiefly of this salt, in an impure condition. From 10 grams
of p-chlor-p-methoxybenzophenone 0.5 gram of the crude
product was obtained. Water decomposes it into ammonium
chloride and p-chlor-p-methoxybenzophenone (melting point
122°, which was raised somewhat by the admixture of the
pure ketone, melting at 125°).
Treatment of the crude salt with hypochlorous acid con-
verts it into the corresponding chlorimide, which is described
below. It was expected that the latter would be converted
by hydrogen chloride back into the hydrochloride of the imide,
but contrary to expectations and to the behavior of other
chlorimides, the chlorimide of p-chlor-p-methoxybenzophenone
did not form a pure imidohydrochloride when treated with
hydrogen chloride. The following analysis was obtained for
the crude product prepared by the action of ammonia on the
ketone dichloride:
0.2816 gram substance gave 0.0296 gram chlorine.
C1
Calculated for
C14H13ONC12
12.57
Found
10.54
The behavior of the salt towards water agrees with the
structure assigned to it, and the preparation of pure chlorimides
from it (see below) confirms this structure. The preparation
of the hydrochloride by this method is considered interesting,
as it is the first instance of the formation of an imidohydro-
chloride from the aromatic ketone dichlorides by the direct
19
action of ammonia.
Hantzsch showed that under the most
varied conditions benzophenone dichloride would not form
the imido hydrochloride with ammonia, and had to have re-
course to the use of urethane with the dichloride in order to
prepare the salt. I obtained exactly the same results on re-
peating Hantzsch's work with benzophenone dichloride.
Stereoisomeric Chlorimides of p-Chlor-p-methoxybenzophenone,
CIC¸H₁-C-C¸H₂OCH, CICH₁-C-CH₂OCH,
3
and
4
11
CIN
3
-The crude
NCI
hydrochloride of imido-p-chlor-p-methoxybenzophenone (5
grams) obtained in the previous experiment was treated with
the equivalent of two molecules of hypochlorous acid in the
same way that the other chlorimides were prepared. This
gave 3.5 grams of a mixture of a- and ß-chlorimides, a yield of
70 per cent. of the theoretical. The product was extracted
with about 25 cc. of cold ligroin (40°-60°). The solution
was cooled to -20° and gave 1.4 grams of a substance melt-
ing at 50°-52°. This was again dissolved in ligroin and the
solution set in the ice box in such a manner as to allow a very
slow evaporation of the ligroin. Large crystals, mixed with
some flat plates of the a variety, came out. The large crys-
tals were recrystallized several times from ligroin before a
substance was obtained which gave a sharp melting point,
65°, the highest melting point obtained for this form, the ß
variety. The residue from the first ligroin extraction was
recrystallized from boiling ligroin (boiling at 70°-80°) several
times and finally gave the pure a form with a melting point
of 94°.5. Melting points of the two varieties and of mixtures
of them were taken simultaneously and gave the following
results: ẞ form, 64°; a form, 94°; mixture, 50°-52°.
Analysis of the two varieties showed the following: 0.1778
gram of the a form gave o.0227 gram active chlorine, and
0.1131 gram of the ẞ form gave 0.01442 gram active chlorine.
Active Cl
Calculated¹ for
C14H11ONC12
12.66
α
12.75
Found
B
12.75
1 Only the chlorine in the (: NC1) group is indicated by the analysis, which was
made with the aid of potassium iodide.
20
་
The same kind of evidence as was used with the other
stereoisomeric chlorimides was brought to exclude crystal and
structural isomerism. After melting, each substance crystal-
lizes again in its own characteristic form, even when it has
been heated several degrees above the melting point of the
higher-melting form. When the substances are mixed with
each other the melting point is decidedly depressed, as shown
above. And, finally, each substance crystallizes in its own
form from a solution of a mixture of the two. These facts
exclude crystal isomerism.
The stereoisomeric chlorimidoketones were converted back
into the ketone in the usual way in order to exclude the possi-
bility of structural isomerism. The results follow: 0.266
gram of the a form was dissolved in chloroform and the solu-
tion saturated with dry hydrogen chloride. This treatment
produced a white precipitate. Ligroin was then added to
complete the precipitation. The chloroform-ligroin mixture
was decanted and the last traces of the solvent blown off in
a blast of air. The residual salt was dissolved in hot
water and gave 0.216 gram of p-chlor-p-methoxybenzophe-
none (melting at 124°-124°.5). This is a yield of 92 per
cent. of the theoretical. A mixture of the product with the
ketone obtained in the same way from the ẞ form melted at
the same point and a mixture with synthetically prepared
ketone (melting at 125°) melted at 125°. When 0.173 gram
of the ẞ form was treated by the same process o. 134 gram,
or 90 per cent. of the theoretical weight of ketone (melting at
125°), was obtained.
Attempts to prepare chlorimido-p-nitrobenzophenone and
chlorimido-p-methoxy-o-chlorbenzophenone had to be given
up as a result of difficulties experienced in preparing the cor-
responding imidohydrochlorides of the two ketones. A
chlorimide of p-methylbenzophenone was prepared but not
investigated further when it was found that the lower melting
of the two stereoisomers, if present at all, formed an oil diffi-
cult to purify. A crystalline compound, presumably the
higher melting stereoisomer, was not specially purified and
melted between 35° and 45°; it contained 14.62 per cent.
21
3
of chlorine, while a compound CH,C,H,C( : NC1)CH¸ would
demand 15 44 per cent. There is no doubt that such a chlor-
imide may be prepared, but since the main object of this
investigation was to isolate well-defined stereoisomeric com-
pounds, the product was not further investigated. The oil,
according to the analytical results, was not a pure compound,
but contained some chlorimide.
o-Chlor-p-methoxybenzophenone, CIC,H,.CO.CH₂OCH. -In
the attempt to prepare o-chlor-p-methoxybenzophenone chlor-
imide, o-chlor-p-methoxybenzophenone was obtained from
o-chlorbenzoyl chloride and anisole. Since it has not been de-
scribed in the literature, its preparation will be briefly dis-
cussed here. Twenty-five grams of o-chlorbenzoyl chloride
was dissolved in 100 cc. of carbon disulphide and 25 grams
of dry powdered aluminium chloride was added to the solution.
The mixture was then cooled in a freezing mixture and the
equivalent of one molecule of anisole allowed to drop into it
slowly from a dropping funnel. Hydrogen chloride was given
off regularly and a crystalline body separated out. The car-
bon disulphide was then distilled off and the aluminium chlor-
ide decomposed by water acidulated with hydrochloric acid.
The residue was extracted with ether and the extract well
washed with sodium hydroxide to free it from any o-benzoyl
chloride. The ether was then distilled off and the residue
fractionated in a Brühl apparatus at reduced pressure. The
pure substance distilled at 250° under 50 mm. pressure.
Analysis of it shows the following results:
0.1553 gram substance gave 0.3865 gram CO₂ and 0.0680
gram H₂O.
C
H
Calculated for
C14H1102Cl
68.18
4.49
Found
67.90
4.90
In conclusion, I wish to express my gratitude to Professor
Stieglitz for his kind guidance and for the inspiration he has
given me.

UNIVERSITY OF MICHIGAN
3 9015 06789 0890
2
Medical Library
1