Ira Remsen.

American chemical journal online

. (page 4 of 77)
Online LibraryIra RemsenAmerican chemical journal → online text (page 4 of 77)
Font size
QR-code for this ebook

observed side by side with the slender white needles of the
bromdinitrotoluol. It seems, therefore, that the two bodies
do not react on each other, and that only one of the amido
groups of the metaphenylenediamine can be affected by brom-

Many attempts were made to replace the amido group in
amidophenylenimidodinitrotoluol with bromine by means of
the diazo reaction, but none of them were successful, black,
tany masses being obtained in each case. The experiment
was tried with strong hydrobromic acid and an excess of sodic
nitrite ; with just the calculated amount of sodic nitrite to form
the diazo compound, as we thought the bad result might be
due to the formation and decomposition of a nitroso compound
from the imido group ; in acetic acid solution ; by Sandmey-
er's reaction ; and finally an attempt was made to prepare the
diazo perbromide, but the bromine instead acted on the phen-
ylene ring forming a tribrom compound.


This substance was prepared by treating amidophenyleni-
midodinitrotoluol with bromine water, which converted it into
a yellow substance. When the red color of the original body
was entirely changed to yellow, the product was filtered out,
and purified by crystallization from boiling benzol, until it

Digitized by VjOOQ IC

28 Jackson and Itiner,

showed the constant melting-point 222^, when it was dried at
100^, and analyzed with the following result :

0.1445 gram of the substance gave by the method of Carius
0.1552 gram of argentic bromide.

Calculmted for t

NH,C«HBraNHC«H,(NOt),CHs. Pottnd.

Bromine 45«7i 45 -70

The amidophenylenimidodinitrotoluol has therefore taken
up three atoms of bromine in place of three of hydrogen, and
it is probable that these atoms of bromine are attached to the
benzol ring of the amidophenylenimido group. One gram of
the original substance gave 1.8 grams of the tribrom com-
pound instead o; the 1.82 grams required by theory. The
yield is, therefore, 99 per cent.

Properties of TrihromamidophenylenimidodinitrotoluoL — The
substance crystallizes from boiling benzol in rhombic plates
of a brownish-yellow color, which melt at 222*. It is easily
soluble in acetone; moderately soluble in benzol, chloroform,
carbonic disulphide, or ethyl acetate; slightly soluble in
methyl alcohol, ether, or glacial acetic acid; almost insoluble
in ethyl alcohol, or ligroin ; insoluble in water. Boiling ben-
zol is the best solvent for it. It is slightly soluble in strong
hydrochloric acid or nitric acid ; fuming nitric acid dissolves
it more easily, and water precipitates from this solution a
light yellow substance, perhaps a nitro-compound ; it is solu-
ble in strong sulphuric acid. Solutions of alkalies produce
no effect on it in the cold. When warmed with alkalies, a
reaction takes place producing a black solution, from which
we did not succeed in isolating a pure substance. We had
hoped by this decomposition of the body to determine the po-
sition of the atoms of bromine, but, as one of the products of
the reaction was bromide of sodium, this was impossible.
Sodic ethylate gives a bluish-violet color with it. Aniline
had no effect upon it.

THhromanUidodinitrotoluol, C.H,Br,NHC.H,(NO,),CH,.

This substance was prepared in the course of the work on
complex molecules mentioned in the introduction to this
paper. Bromine water does not act easily upon the anilido-

Digitized by VjOOQ IC

ParahromdimetanitroioluoL 29

dinitrotoluol, which, in this respect, differs from the corres-
ponding amidophenylenimido body, whose behavior was de-
scribed above. It was necessary, therefore, to use pure bro-
mine, some of which was poured into an evaporating dish,
and anilidodinitrotoluol added in small amounts at a time.
The action was violent, the substance dissolving, and fumes
of bydrobromic acid being given off. Enough bromine should
be taken to moisten thoroughly all the solid used. The ex-
cess of bromine was evaporated off, the solid residue broken
up and washed with cold alcohol, in which it is almost in-
soluble. The yellow powder thus obtained was recrystallized
bom boiling benzol until it showed the constant melting-
point, 238**, when it was dried at 100*, and analyzed with the
following result :

0.1 9 18 gram of the substance gave, by the method of
Carius, 0.2107 gi'&m of argentic bromide.

Calculated for
C«H,BrtNHC«H,(NOa)aCHt. Found.

Bromine 47-05 46-75

Properties of TribromanUidodiniirotoluoL — It crystallizes
from boiling benzol in yellow rhombic plates often with the
acute angles truncated, which melt at 238**. It is soluble in
chloroform or carbonic disulphide ; moderately soluble in
benzol, acetone, or ethyl acetate ; slightly soluble in ether or
glacial acetic acid ; almost insoluble in ethyl or methyl alco-
hol, or in ligroin ; insoluble in water. Benzol is the best
solvent for it. It is insoluble in either of the three strong
acids, but dissolves in fuming nitric acid. A solution
of sodic hydrate has no action on the substance, even if boiled
with it for several minutes.

In the course of the work on complex molecules outlined
in the introduction, an attempt was made to convert the sub-
stance just described into a nitro compound. For this pur-
pose the tribromanilidodinitrotoluol was mixed with nitric
acid of specific gravity 1.5, in which it dissolved without
evolution of nitrous fumes forming a red solution. Cold
water separated from this solution a precipitate resembling
red sulphide of antimony in appearance, which was washed
with water and dried. All our attempts to obtain a crystal-

Digitized by VjOOQ IC

30 Jackson dfid litner.

line body from this precipitate have failed. It dissolved in
several of the organic solvents, but these solutions left on
evaporation a red viscous or resinous substance, which
showed no signs of crystallization, even after long standing.

We next heated this product with aniline in the hope of
getting a more manageable anilido compound. An anilido
compound was produced, as shown by the formation of ani-
line bromide, and this proved that the uncrystallizable sub-
stance was a nitro compound, since the tribromanilidodini-
trotoluol, from which it was made, did not give up bromine
to aniline. This new anilido compound, however, behaved
with solvents in the same unsatisfactory manner as the nitro
compound, from which it was derived. Nor did a bromine de-
rivative which was formed from this new anilido compound by
the action of bromine, as indicated by the evolution of hydro-
bromic acid, show any signs of crystallization. As none of
these derivatives could be brought into a state fit for analy-
sis, we were obliged to give up this branch of the work.
Action of Pyridine on Bromdinitrobemoic Acid,

If bromdinitrobenzoic acid is added to pyridine it gives a
white crystalline precipitate, which almost immediately dis-
solves with evolution of heat, and from this new solution a
heavy yellow precipitate is thrown down. To isolate the
two products, which seemed to be formed by this action, we
dissolved some bromdinitrobenzoic acid in absolute alcohol,
and added to it an excess of pyridine, when white needles
separated in large quantity. These were filtered out, washed
with absolute alcohol, in which they were not very soluble,
and then analyzed, but the result obtained did not correspond
to any formula. It seemed to indicate rather that the sub-
stance was a mixture of the pyridine salt of bromdinitroben-
zoic acid and the free acid ; and this, as a matter of fact, is
what we should have obtained, if the original crystals had
been this pyridine salt, as they lost pyridine with great ease,
when treated with alcohol. Although the analysis did not
settle the composition of the crystals, the following facts
leave little doubt that they were the salt


Digitized by VjOOQ IC

Parabromdimeianitroioluol. 31

They showed the solubility of a salt, being slightly soluble in
cold and freely soluble in hot water ; sodic hydrate set free
P3rridine, and acids precipitated bromdinitrobenzoic acid from
a solution of them. The solution gave no test for a bromide.
Sodic ethylate gave the same reddish-purple color, which it
gives with salts of bromdinitrobenzoic acid. We may add
that tlie crystals were moderately soluble in cold, easily solu-
ble in hot alcohol, and that with an excess of pyridine they
passed easily into the yellow product obtained direct from
pyridine and the acid.

The yellow product was prepared by evaporating the
mother-liquor from the white salt, which still contained some
of the crystals, or even by allowing it to stand at ordinary
temperatures. We did not attempt to analyze this substance,
as it is decomposed even by crystallization from alcohol, but we
have been able to draw a safe inference in regard to its nature
from its decomposition-products and its properties. As first
prepared, it melted with decomposition at about 210", but, if
it was recrystallized from alcohol, some bromide of pyridine
was eliminated, and the crystals showed a higher melting-
point. By continuing the crystallization often enough all the
bromine and pyridine were removed as bromide of pyridine,
and the crystals melted constant at 245^-246'', and contained
no bromine. This same substance could be obtained from
the yellow substance by treating it with dilute sulphuric acid.
After recr>'Stallization from alcohol it appeared in elongated
hexagonal plates of a light straw color. To determine the
nature of this substance, it was converted into its calcium
salt by warming it with calcic carbonate and water. The
solution thus obtained was filtered, and upon concentration
deposited yellowish-orange, well defined octahedra of the
salt, which, after it had been crystallized, redissolved with
great difficulty. The crystals were filtered out, pressed be-
tween filter paper, dried in the air, and analyzed with the re-
sults given under I. The filtrate, on further concentration,
gave another crop of crystals, which were dried in the same
way, and analyzed with the results given under II. The
mother-liquor from these second crystals, when acidified, gave
the acid with its proper melting-point, 245^-246*.

Digitized by VjOOQ IC

32 Jackson and iHner,

I. 0.2552 gram of the air-dried salt lost 0.0304 gram of water
when heated to 157®.

II. 0.5226 gram of the air-dried salt lost 0.0602 gram of
water at 160**.

Calculated for Fonnd.

Ca0C00CcH,(N0,),.aH,0. I. H.

Water 11.92 11.92 11.52.

I. 0.2248 gram of the salt dried at 157** gave o.i 122 gram of
calcic sulphate.

II. 0.4624 gram of the salt dried at 160® gave 0.2269 gram
of calcic sulphate.

Calculated for Pound.

CaOCOOC.H,(NO,),. I. II.

Calcium 15. 04 14.69 i4*44

These results leave no doubt that the substance is the cal-
cium salt of oxydinitrobenzoic acid, but the melting-point of
this acid, 245**-246*', did not agree with that given by H. Sal-
kowski* for paroxydimetanitrobenzoic acid which was 235**-
237*. We accordingly made a careful study of the melting-
point of this acid, which is described in a later section, and
proved that its true melting-point is 245'*-246**.

The work just described shows that the yellow pyridine
compound is broken up by crystallization from alcohol, or by
dilute sulphuric acid, into oxydinitrobenzoic acid and bromide
of pyridine. That it was not a mixture of these two substances
was shown by the fact that it took several crystallizations to
remove all the pyridine and bromine, whereas bromide of p5rr-
idine is so soluble in alcohol that the whole of it would have
been removed by the first crystallization, if it had existed
already formed. That the substance is a salt is made proba-
ble by its solubilities, as it is slightly soluble in cold, freely
soluble in hot water ; slightly soluble in cold, moderately sol-
uble in hot alcohol ; insoluble in benzol or ether ; and this
view is confirmed by the facts that alkalies set free pyridine
from its solutions, and argentic nitrate forms a red salt when
added to the neutral solution. With an acidified solution it
gives a precipitate of argentic bromide, but this may come
from the bromide of pyridine formed by the acid, rather than

1 Ann. Chem. (Uebig;, 163, 36.

Digitized by VjOOQ IC

ParabromdimeianUrotoluol. 33

from the original yellow substance. The only formula which
we have found that explains all these observations is

derived from two molecules of pyridine and one of the brom-
dinitiobenzoic acid. On this theory one of the molecules
of pyridine is attached to the carboxyl group, while the other
is attached to the bromine atom and the benzol ring of the
substituted benzoic add, forming a salt analogous to bromide
o/pjridine, as shown by the following formulas :

(C.H,N)Br(C,H,(NO,),CCK)HC,H,N). (C.H.N) BrH.
The decomposition of the yellow substance would then be rep-
resented by the following reaction :

C.H.NBrC,H,(NO.),COOHC.H.N + H,0 = C.H.NBrH +
The MdHng'Point of Paroxydimeianiirobenzaic Add, — ^The
acid obtained from the yellow pyridine compound by the ac-
tion of dilute sulphuric acid or by crystallizing it from alcohol,
as described in the preceding paragraph, melted at 245*-246*,
which was also the melting-point of the oxydinitrobenzoic
add prepared by us by the action of sodic hydrate on brom-
dinitrobenzoic acid. All these acids must be paroxydimeta-
nitrobenzoic acid, and, as H. Salkowski' ascribes the lower
melting-point, 235*-237*, to this body, we have felt it neces-
sary to prepare this acid by the two methods given by him in
order to remove this discrepancy. Accordingly, some chrys-
anisic add was made by the method of Prederici,' the oxida-
tion of dinitrotoluidine mdting at 166*, with potassic dichro-
mate and sulphuric acid. This melted at 259^. It was treated
with sodic nitrite in a boiling, weak add solution, until all the
chrysanisic add had been transformed into the new body.
This took place very slowly. The product melted at the tem-
perature stated by Salkowski, 235*-237'', with decomposition
and blackening, but, after repeated crystallization from alco-
hol, the melting-point rose to 245^, and the signs of decompo-
sition during melting vanished.

In order to apply Salkowski's second method some chrys-
anisic add was heated with a dilute solution of sodic hydrate.

1 Ann. Chem. (liebic)* i^ 31^
SBcr. d. cbem. G«s., ii» 1975-

Digitized by VjOOQ IC

34 Jackson and Ittner.

The mixture became dark-colored, and ammonia was given
off. The liquid was evaporated to dryness on the steam-bath,
the residue taken up with a little hot water, filtered, and pre-
cipitated with hydrochloric acid ; this yielded a crude prod-
uct, melting a little above 240"*, which was raised to the con-
stant melting-point, 245^, by crystallization from alcohol. II
the heating of the chrysanisic acid was continued for a long
time with a strong alkaline solution, upon acidification of the
liquid no oxydinitrobenzoic acid was deposited, but there was
a vigorous evolution of hydrocyanic acid. This behavior is
analogous to that of the anilidodinitrobenzoic acid described
earlier, which gives aniline and oxydinitrobenzoic acid at
first with sodic hydrate, but by more ^vigorous treatment
phenyl isocyanide. It also recalls the work of Wohler* and of
Post and Hiibner* on the formation of cyanides by the action of
alkaline hydrates in aqueous solution on aromatic nitro com-

The work just described shows that paroxydimetanitroben-
zoic acid prepared in four different ways had the constant
melting-point 245^-246*, and, therefore, that the melting-point
235^-237** ascribed to it by H. Salkowski rests probably on an
error of observation. We should add that the thermometer
used in this work was tested by determining the boiling-point
of aniline and naphthalene with it. Aniline with the column
in the vapor and at a pressure of 760 mm. showed the boiling-
point 184*. Naphthalene under the same conditions gave a
temperature of 2i8*'-2i9*. There can be no doubt, therefore,
that the thermometer was accurate enough for this work.

Chlordinitrobenzoic Acid,

The following work on the chlor compound could not be
finished so thoroughly as we wished, because of the pressure
of other parts of the research, but we have decided to describe
our observations even in this imperfect state, since they seem
to leave no doubt in regard to the nature of the substance ob-

A quantity of chrysanisic acid was mixed with cold, strong
hydrochloric acid, and treated with an excess of sodic nitrite.

1 Poffg. Ann. 13, 488, (z8a8).

s Ber. d. chem. Oes.. 5. 408, (1873).

Digitized by VjOOQ IC

Parahromdimetanitrotoltiol, , 35

To the mixture was then added a solution of cuprous chloride
in strong hydrochloric acid, and the yirhole heated for about
half an hour with occasional addition of more sodic nitrite.
After the liquid had been cooled by the addition of water the
product, a reddish, viscous mass, was removed by winding it
about a rod, and was then sucked out as completely as possi-
ble with a filter pump. The residue was pressed between fil-
ter paper, and recrystallized from alcohol, and later from ben-
zol, ^which is much more efficient in removing the tarry mat-
ter. In this way a crystalline body was obtained, the melt-
ing-point of which we did not succeed in raising above 159^,
but we do not feel certain that the true melting-point may not
lie a few degrees above this temperature, as our work was
done with an insufficient amount of material, the peld being
exceedingly small. For this reason also we were unable to
analyze the substance, but the following observations show
without doubt that it is chlordinitrobenzoic acid. It is insol-
uble in water or acids, but dissolves easily in alkalies or alka-
line carbonates. It dissolves in aniline with the aid of heat,
forming aniline chloride, and an orange-red acid, melting
when not perfectly pure at 235^-238* ; anilidodinitrobenzoic
acid melts at 239"*. This product also gave the pansy-purple
color with sodic ethylate, which is characteristic of anilido-
dinitrobenzoic acid, whereas the chlor acid, melting at 159*,
gives with sodic ethylate a reddish-purple color, closely re-
sembling that obtained from bromdinitrobenzoic acid. With
ammonic hydrate the acid, melting at 159**, gave ammonic
chloride and a substance resembling chrysanisic acid in ap-
pearance, and melting in the crude state not far from the
melting-point of this acid.

Proptrties of Chlordiniirobemoic Acid, — It crystallizes in
white, rather ill-defined, prisms, which probably melt at 159**.
It is very soluble in ether or acetone ; easily soluble in
methyl alcohol, glacial acetic acid, or ethyl acetate ; moder-
ately soluble in ethyl alcohol, benzol, chloroform, or carbonic
disulphide ; almost insoluble in ligroin. Benzol is the best
solvent for it.

Digitized by VjOOQ IC

36 . Jackson and Ittner.

AcHon of Hydrochloric Acid on Bromdinitrobenzoic Acid.
These experiments, like those described in the preceding-
section, are unfinished, as it was intended to make this work
the starting-point for a new research and, therefore, the study
of the most interesting product of the reaction has been post-
poned until that could be undertaken. Bromdinitrobenzoic
acid in quantities of two grams at a time was heated with fif-
teen cc. of hydrochloric acid, containing 1 1 per cent, of the
acid, to 200^ for five hours in a sealed tube. At the end of
this time, it was found that the undissolved portion had col-
lected into a liquid globule. The whole contents of the tube
were poured into an evaporating dish, and allowed to stand,
until the globule had solidified, and crystals had separated
from the supernatant liquid. As these crystals gave with
sodic ethylate the characteristic reddish-purple color of brom-
dinitrobenzoic acid, they were heated with a dilute solution
of sodic carbonate on the steam-bath until no more dissolved.
The residue melted after one recrystallization from alcohol at
119*, and was probably vicinal bromdinitrobenzol, although
this supposition needs confirmation by analysis before it is
accepted. The dark-red solution obtained by sodic carbon-
ate gave, when acidified, a precipitate, which, by cr>'stalli-
zation from alcohol, was separated into two substances. One
of these melted at 245**, and was, therefore, oxydinitroben-
zoic acid, formed by the action of the solution of sodic carbon-
ate upon the bromdinitrobenzoic acid, although possibly some
of it was one of the direct products of the reaction. The
other substance melted at 62**, and we, therefore, inferred that
it was the vicinal dinitrophenol (OH.i.NO,.2.NO,.6.) as this
melts at 61**. 78 according to Mills,* whereas its melting-point
is given usually as 63®-64**. To confirm this view, we oxi-
dized the substance with nitric acid, which converted it into
picric acid, recognized by its melting-point 121*, its copper
and lead salts, and the deep red color produced when it was
heated with potassic cyanide. The formation of the dinitro-
phenol was to be expected, as Salkowski and Rudolph' have
found that oxydinitrobenzoic acid, if heated with water to

1 Phil. Mag. [5]. 14. a?.

s Ber. d. cbem. Ges., lo, 1355.

Digitized by VjOOQ IC

Aluminum Ethylate. 37

170* is converted into this vicinal dinitrophenol. The prod-
ucts of the action of hj'drochloric acid at 200"* on bromdini-
trobenzoic acid wotdd seem to be the following : Bromdinitro-
benzol (?) (Br.i.NO,.2.NO,.6.) about one-fourth of the prod-
uct ; dinitrophenol of the same series, about one-third of the
remainder; and unaltered bromdinitrobenzoic acid mixed
perhaps with some oxydinitrobenzoic acid.


Bt H. W. Hilltbr and O. B. Ckookbr.

In our preliminary notice, which appeared in this Journal
under the head of ** Aluminum Alcoholates,'** mention was
made of our investigation, as then incomplete, of the action
of alcoholic solutions of anhydrous chlorides upon aluminum.
That part of the work wherein ethyl alcohol was used has
been completed, and the results as obtained are given in this

In their work on aluminum alcohols,' Gladstone and Tribe
have prepared, by means of their aluminum-iodine reaction,
several aluminum alcoholates, among which was the ethylate.
Their method of procedure was as follows : by bringing to-
gether 2 grams of aluminum foil, i gram of iodine, and 20 cc.
of absolute alcohol the elements were found to combine
quickly and to produce the aluminum iodide necessary for
the reaction, whereupon the decomposition of the alcohol be-
gan before the whole of the free iodine had passed into com-
bination. The action proceeded with moderate rapidity until
the metal had disappeared and the evolution of hydrogen
ceased. The reaction was supposed to take place through
the intervention of intermediate bodies, as follows :

1 After sending this mrticle to the editor, our attention was first directed
to a note by Mr. W. Tistscbenko, in the Octot>er Jonmal. Although in many
respects oar results are identical with his. it still seems to us desirable to publish
them, since our method is not the same, and since related facts not hinted at in the
noteabore mentioned have been observed in the course of our research. It is our
p nrp o »e to complete the work already begun with methyl and the two propyl alco-

t This Journal, iS, 621.

• J. Chem. Soc.. 1881, 39, x ; 1883, 41, 5 ; x886, 49. 25.

Digitized by VjOOQ IC

38 Hillyer and Croaker,

(a) 3C.H.OH + A1.I. = A1.(C.H.0).I. + 3HI ;

{b) A1.(C,H.0).I, + 3C.H.OH = A1,(C.H,0). + 3HI ;

{c) 6HI + A1. = A1.I.+ 3H..

The contents of their flask was then heated in an oil-bath
to about 300^, and held at this temperature until the excess of
alcohol had been driven off.

The residue in their flask was next distilled under dimin-
ished pressure* and a yellowish white distillate obtained
which melted at 115* C.»* and boiled somewhere near the
limits of the ordinary mercurial thermometer. This sub-
stance, to which they gave the name aluminum ethylate, was
found to be quite free from iodine, and to contain considera-
ble aluminum. It dissolved slightly in absolute alcohol, and
from alcoholic solution water precipitated aluminum hydrox-
ide. Heat at atmospheric pressure decomposed it into alumi-
num oxide, alcohol, ethylene, and ether. »

Even though distilled in vacuo, they found considerable de-

Online LibraryIra RemsenAmerican chemical journal → online text (page 4 of 77)