Transcriber’s Notes:

  Typographical and punctuation errors have been silently corrected.




                    A Further Investigation of the
                        Symmetrical Chloride of
                   Paranitroorthosulphobenzoic Acid.

                             Dissertation.

                 Submitted to the Board of University
                Studies of the Johns Hopkins University
                for the Degree of Doctor of Philosophy.

                                — by —
                         William E. Henderson.

                                 1897




Acknowledgment.


The author esteems it a privilege as well as a pleasure to give
expression to his sincere sense of gratitude to Prof. Remsen,under
whose guidance this work was carried on not only for instruction
received in the lecture room, but for his frequent suggestion, and his
constant and friendly interest in the work as it progressed. These have
at all times been an encouragement and an incentive.

He wishes also to express his appreciation of the instruction and
kindly guidance in the laboratory, of Drs. Morse and Renouf, as well as
of Dr. Ames of the Physical Laboratory.




Contents.


       I. Introduction.                                    Page 1

      II. Preparation of the Acid Potassium Salt
          of Paranitroorthosulphobenzoic Acid.                  6

     III. Preparation of the Symmetrical Chloride
          of Paranitroorthosulphobenzoic Acid.                 12

      IV. Properties of the Symmetrical Chloride
          of Paranitroorthosulphobenzoic Acid.                 19

       V. The Action of Benzene and Aluminium
          Chloride on the Symmetrical Chloride
          of Paranitroorthosulphobenzoic Acid.                 22

          The Barium Salts of
          Paranitroorthobenzoybbenzenesulphonic Acid.          24

      VI. The Action of Alcohols on the Symmetrical Chloride
          of Paranitroorthosulphobenzoic Acid.                 30
             1. Methyl Alcohol.                                31
             2. Ethyl Alcohol.                                 32
                Action of Ethyl Alcohol on the Unsymmetrical
                Chloride.                                      36

     VII. The Action of Phenols on the Symmetrical Chloride
          of Paranitroorthosulphobenzoic Acid.                 38
             1. Phenol.                                        40
             2. Orthocresol.                                   48
             3. Paracresol.                                    51
             4. Hydroquinone.                                  53
             5. Resorcin.                                      56
             6. Pyrogallol.                                    59
             7. β-naphthol.                                    61

    VIII. The Action of Aniline on the Symmetrical Chloride
          of Paranitroorthosulphobenzoic Acid.                 62

      IX. The Action of Phosphorus Oxychloride on the Fusible
          Anilid of Paranitroorthosulphobenzoic Acid.          71

       X. The Action of Reagents on the Dianil of
          Paranitroorthosulphobenzoic Acid.                    77
             1. Of Hydrochloric Acid.                          77
             2. Of Alcoholic Potash.                           78
             3. Of Glacial Acetic Acid.                        79

      XI. Conclusions.                                         82

          Biographical.                                        85




I. Introduction.


The sulphobenzoic acids have been the subject of investigation in this
laboratory for a number of years past. Among the many interesting
facts that have been brought to light in the course of this study,
perhaps no others have been attended with more interest than the
discovery of well characterized isomerism in the case of the chlorides
of orthosulphobenzoic acid, and its paranitro derivative; together
with the preparation of a series of isomeric derivatives of these
substances. The chlorides themselves have been isolated in the
crystalline condition, and have been found to differ markedly, not
only in chemical, but in physical properties as well.

The first evidence that such isomerism existed, was obtained by Remsen
and Coates[1] who, in the course of an investigation of the action of
aniline upon the chloride of orthosulphobenzoic acid, obtained two
isomeric anilids quite different in properties, which they designated
as fusible and infusible respectively. The following year, Remsen and
Kohler[2] obtained one of the chlorides in crystalline form, together
with an oil which they did not succeed in crystallizing.

This however was accomplished the succeeding year by Remsen and
Saunders[3], and a still more satisfactory result was obtained by
Remsen and McKee[4] in 1895. The chloride melting at 79° was found to
yield only the fusible anilid, together with an anil, while from the
lower melting chloride, in addition to these, the infusible anilid was
also formed.

[1] Am. Chem. Journ. XVII, 311.

[2] Ibid XVII, 230.

[3] Ibid XVII, 354.

[4] Ibid XVIII, 794.

In 1895, Gray[5] isolated the two corresponding isomeric chlorides of
paranitroorthosulphobenzoic acid, the lower melting chloride being
obtained in small quantity only. The succeeding year Hollis[6] made a
more careful study of this lower melting chloride, and prepared it in
considerable quantity.

From evidence drawn from the action of ammonia upon these chlorides,
taken in connection with a number of other facts, the higher melting
chloride is identified as the one possessing a symmetrical structure,
while the lower melting chloride possesses an unsymmetrical structure.
The first one, when treated with ammonia is slowly transformed into the
ammonium salt of paranitrobenzoic sulphinide:

[5] Inaug. Diss. J. H. Univ. 1895.

[6] Inaug. Diss. J. H. Univ. 1896.

                              CO
                              / \
        COCl                 /   N.NH₄
       /                    /   /
    C₆H₃——SO₂Cl + 4NH₃ = C₆H₃——SO₂     + 2NH₄Cl.
       \                        \
        NO₂                      NO₂

while the lower melting chloride is quickly transformed into the
ammonium salt of paranitroorthocyanbenzenesulphonic acid:

         CCl₂
         / \
        /   O               CN
       /   /               /
    C₆H₃——SO₂   + 4NH₃ = C₆H₃——SO₂ONH₄ + 2NH₄Cl.
       \                   \
        NO₂                 NO₂

Gray’s study of the symmetrical chloride was confined for the most part
to the preparation of a series of salts of this latter acid, and to
an investigation of the action of aniline upon the chloride itself.
It was thought to be of interest to extend this study to a wider
range of reactions, as well as to improve, if possible, the method of
preparing the chloride in pure condition. At the suggestion of Prof.
Remsen this work was accordingly undertaken.




II. Preparation of Material.


The method employed in the preparation of paranitroorthosulphobenzoic
acid was essentially that described by Hart,[7] Kastle,[8] Gray[9]
and Hollis.[10] The details of it are repeated here for the purpose of
calling attention to certain facts that came under the author’s notice.

[7] Am. Chem. Journ. I, 350.

[8] Ibid XI, 177.

[9] Inaug. Diss. J. H. Univ. 1845.

[10] Inaug. Diss. J. H. Univ. 1896.

100 grams of paranitrotoluene are added to 400 grams of fuming sulfuric
acid, and the mixture heated in a balloon flask at 100° on a water
bath. The heating is continued until a few drops of the mixture, added
to cold water, dissolves completely to a clear solution. The time
required for this operation varies much with the conditions. Continued
stirring very considerably hastens the reaction, as paranitrotoluene
forms a layer on the acid, which presents a small surface to its
action. With constant stirring the reaction is complete in a few hours,
whereas if no stirring is resorted to, as much as several days may be
required, especially when large quantities are employed at one time.

When the reaction is complete, the mixture is poured into a
large volume of water, and neutralized with calcium carbonate.
In the filtrate from calcium sulphate, the calcium salt of
paranitroorthotoluene sulphonic acid is found, and this is converted
into the potassium salt in the usual way.

The oxidation of the potassium salt is effected as follows. 50 grams of
the salt are dissolved in 2½ litres of water, and to this is added a
solution of 15 grams of potassium hydroxide. The mixture is heated to
100° on a water-bath, and when this temperature is reached, 110 grams
of potassium permanganate are added. Heating is continued until the
solution is decolorized, care being taken to prevent the evolution of
free oxygen.

The oxides of manganese are then filtered off, the filtrate neutralized
with hydrochloric acid, and evaporated to about one fifth of its
original volume. Strong hydrochloric acid is them added in excess, and
on cooling the acid potassium salt of paranitroorthosulphobenzoic acid
separates in very slender colorless needles completely filling the
liquid.

For the success of this operation it is important that the potassium
salt of paranitroorthotoluenesulphonic acid and the potassium hydroxide
should both be perfectly dissolved before they are heated together.
If the two substances lie together in solid form at the bottom of the
flask, a very slight elevation of temperature leads to the formation
of an extremely troublesome red substance, which is very difficult
to remove. It is almost impossible to remove it from the oxidation
product by recrystallization, since any considerable amount of it has
a marked influence on the solubility of the salt, rendering it much
more soluble. It persists throughout all subsequent transformations of
paranitroorthosulphobenzoic acid, and should therefore be carefully
avoided.

Otto Fischer[11] has shown that in concentrated solution, potassium
hydroxide acts on nitro derivatives of toluene, with the formation
of various colored substances derived from stilbene. In the case
of paranitroorthotoluenesulphonic acid, he describes the substance
formed as possessing a cherry red color. The reactions involved in its
formation are:

         CH₃          HC ============ CH
        /            /                 \
    2C₆H₃——SO₂OK = C₆H₃——SO₂OK  KO.O₂S——C₆H₃ + 2H₂O
        \            \                 /
         NO₂          \               /
                       \  /-- O --\  /
                         N         N
                          \-- O --/

By oxidation this passes to a nitro compound of the composition

       HC============ CH
       /               \
    C₆H₃——SO₂OK  KOO₂S——C₆H₃
       \               /
        NO₂         O₂N

It was no doubt the formation of substances of this nature that
occasioned the color observed in some of the oxidations.

[11] Ber. XXVI-2231; XXVIII-2281

The only effective method of separating this colored substance was
found to be to pass to the neutral salt of paranitroorthosulphobenzoic
acid, by making the solution slightly alkaline. The salt of this
colored substance is also formed and the two can be separated by a few
recrystallizations in a fairly satisfactory manner.

The yield in both of the transformations involved in the preparation
of paranitroorthosulphobenzoic acid does not fall far short of the
theoretical.




III. Preparation of the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.


This chloride was first separated from its unsymmetrical isomer
by Gray[12]. It was obtained by allowing a chloroform solution of
the mixed chlorides to evaporate until the chloroform had almost
entirely disappeared. In the thick liquid so obtained, crystals of the
symmetrical chloride were formed. It was also obtained by applying the
method devised by Bucher in connection with the corresponding chloride
of orthosulphobenzoic acid—i.e. by the action of dilute ammonia on the
mixed chlorides. Gray also found that the best conditions for securing
a relatively large proportion of the symmetrical chloride were, the
employment of as low a temperature as possible in the formation of the
chlorides, and of as small an excess of phosphorus pentachloride as
would suffice for the reaction.

[12] Inaug. Diss. J. H. Univ. 1895.

After many experiments, under widely differing conditions, the
following method of procedure, embodying the results of Gray’s work,
was adopted.

Dehydrated acid potassium salt of paranitroorthosulphobenzoic acid,
and phosphorus pentachloride, in the ratio of 40: 55 grams, are
brought together in a mortar and intimately mixed. The mixture is
put into an evaporating dish, and placed on a sulphuric acid bath,
previously heated to 150°. As soon as the action has been well started,
the dish is removed, and the reaction allowed to proceed without
further heating. When it is complete, and the contents of the dish
has cooled down to the temperature of the room, the oily product is
poured slowly into a salts bottle containing ice water, the bottle
being frequently shaken during the process. The shaking is continued
with renewed portions of water, as long as the wash water is cloudy.
The water is then poured off, the brownish gummy chloride dissolved
in chloroform, and the solution placed in a good-sized separating
funnel. Ice water is then added, and the contents of the funnel treated
with successive portions of ammonia (desk ammonia diluted one half).
Shaking is continued after each addition until the odor of ammonia has
disappeared, and ice is added from time to time as may be required.

When it is found that the odor of ammonia persists after several
minutes’ shaking, the chloroform layer, which is usually filled with a
solid substance that has separated during the process, is drawn off,
filtered, and dried with calcium chloride.

By this process all of the unsymmetrical chloride is converted into the
ammonium salt of paranitroorthocyanbenzenesulphonic acid, according to
the equation:

          CCl₂
         / \
        /   O              CN
       /   /              /
    C₆H₃——SO₂  + 4NH₃ = C₆H₃——SO₃NH₄ + 2NH₄Cl.
       \                  \
        NO₂                NO₂

while the symmetrical chloride remains for the most part unchanged,
though some of it is converted into the ammonium salt of
paranitrobenzoic sulphinide:

                              CO
                             /  \
        COCl                /    N.NH₄
       /                   /    /
    C₆H₃——SO₂Cl + 4NH₃ = C₆H₃——SO₂ + 2NH₄Cl.
       \                   \
        NO₂                 NO₂

It was found that working in this way the symmetrical chloride could
be prepared in pure condition, free from its isomer. The chloroform
completely evaporates in a short time leaving fine crystals of the
symmetrical chloride. In case the evaporation is slow and incomplete,
it may be concluded that not all of the unsymmetrical chloride has been
removed. The yield was uniformly about 40 per cent of the theoretical.

From the water used to wash the chlorides a considerable amount of the
original salt can be recovered, as the reaction under the conditions
employed, is never complete.

An examination was made of the substance mentioned as separating in
the chloroform solution of the chlorides, during the treatment with
ammonia, and it was found to possess the following properties. It
is insoluble in benzene, chloroform, acetone, ether and ligroin;
soluble in glacial acetic acid, from which it separates on cooling
in colorless, crystalline condition; insoluble in the cold in
water, alcohol and ammonia, but by boiling with these reagents, or
by long standing in the cold, it is dissolved with decomposition.
It was dissolved in hot water and the solution, which was acid
in reaction, was neutralized with potassium carbonate. On adding
an excess of hydrochloric acid to the solution, and allowing
it to cool, characteristic crystals of acid potassium salt of
paranitroorthosulphobenzoic acid separated. These properties identify
the substance as the anhydride of this acid.

The formation of the corresponding anhydride of orthosulphobenzoic acid
by the action of phosphorus pentachloride upon its acid potassium salt
was observed by Sohon[13], who made use of the reaction to prepare this
anhydride in quantity.

[13] Inaug. Diss. J. H. Univ. 1896.




IV. Properties of the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.


As first obtained, the crystals of the symmetrical chloride resemble
irregularly shaped pieces of amber, both in color, and in lustre.
On recrystallization from chloroform or ether, they may be obtained
perfectly colorless, and are often of very simple crystallographic
form. The chloride crystallizes in the monoclinic system, and possesses
a very remarkable crystallizing power, in which respect it differs
noticeably form its isomer. Even in chloroform solution that is far
from dry, crystals appear with the greatest ease.

The habit of the crystals differs very much according to the
conditions of crystallization. Not infrequently almost perfectly
formed crystals of the simplest form—the oblique octahedron—were
obtained though for the most part the form was much more complicated,
pinacoid and dome faces, together with basal planes being prominent.
As a rule, the crystals were not suitable for crystallographic work,
as the faces are usually uneven and the edges rounded. By proper
precautions however, good ones were obtained, and measurements of these
will be found in this dissertation when it appears in print.

The size of some of the crystals obtained was unusual for substances
of this class. One crystal obtained with no special precautions, save
letting a solution of the chloride stand undisturbed for several days,
in a rather cool place measured 3 × 2.5 × 1.5 cm., and weighed 11.2
grams. The crystals are quite compact, their density being abut 1.85.
They melt at 98° (uncorr.) The chloride is quite stable in crystalline
condition. Even in moist air the crystals were unchanged, and retained
their lustre as long as they were in my possession.

An analysis for chlorine gave the following results.

        .2200 gram gave      .2212 gram AgCl.

                 COCl
                /
    Cal. for C₆H₃——SO₂Cl
                \
                 NO₂
                              Found.
             Cl = 24.94       24.83




V. The Action of Benzene and Aluminium Chloride on the Symmetrical
Chloride of Paranitroorthosulphobenzoic Acid.


Hollis[14] in his study of the action of these reagents upon the
unsymmetrical chloride, tested their action upon one portion of the
symmetrical chloride, and found the products to be identical in the two
cases. A few experiments were made in confirmation of these results,
and the same products, in general, were obtained. It was observed
however that the reactions differ in the relative ease with which
they are brought about. In the case of the symmetrical chloride, the
reaction is a much more vigorous one. On adding aluminium chloride to
a solution of the symmetrical chloride, in benzene, action begins
at once the temperature of the hand, and very little external heat,
and that only in the latter stages of the operation, is needful for
the completion of the reaction. The application of much heat converts
all of the product into thick tarry substances from which nothing
satisfactory could be obtained.

[14] Inaug. Diss. J. H. Univ. 1896.

When the reaction was complete, the resulting product was isolated and
purified in accordance with the directions given by Hollis. Repeated
trials showed that, as in the case of the unsymmetrical chloride, only
one phenyl group could be introduced by this method. The resulting
compound, paranitroorthobenzoylbenzenesulphon chloride, was identical
with that derived from the unsymmetrical chloride. Owing, however,
to the fact that so much more decomposition occurs in the reaction
with the symmetrical chloride, in paranitroorthobenzoylbenzene sulphon
chloride could not be obtained in perfectly pure condition. In
appearance it agreed closely with that described by Hollis, forming
very characteristic greenish, rhombic crystals. These melted, not very
sharply, at 174° instead of 177° as observed by Hollis.

Accordingly, to establish the identity of the two compounds beyond
any doubt, the material on hand was converted into the barium salt of
paranitroorthobenzoylbenzene sulphonic acid. This was done by boiling
the sulphon chloride with dilute hydrochloric acid until complete
solution had been effected; evaporating to dryness on a water-bath;
dissolving the residue in hot water, and neutralizing with barium
carbonate. On filtering the hot solution from the excess of carbonate,
and allowing it to cool, the barium salt separated.

The solution was somewhat colored by impurities, and the long needles
in which the salt crystallized were also somewhat colored. They were
analysed with the expectation that they would prove to be specimens
of the salt described by Hollis as having three, or three and a half
molecules of water of crystallization, in as much as the conditions
under which they were formed were favorable to the formation of salts
with these ratios of water of crystallization. Hollis found that this
salt could be obtained with at least four different ratios of water
of crystallization viz. three, three and a half, six and seven
molecules respectively. The analysis was as follows, the amount of
barium being calculated on the basis of the anhydrous salt.

0.3087 gram lost 0.064 gram at 210°, and gave 0.0759 gram BaSO₄.

    Cal. for (C₁₃H₈O₆NS)₂Ba + 11H₂O         Found.
                        H₂O = 20.90         20.73
                         Ba = 18.29         18.23

The mother-liquor, in which the crystals remaining from analysis were
redissolved, was warmed, but not boiled, with boneblack, to remove
impurities. When filtered, the solution was perfectly colorless, and
on standing for some time, well formed colorless, rhombic crystals
appeared. On analysis they gave results as follows.

0.2804 gram lost 0.0405 gram at 210°, and gave 0.0759 gram BaSO₄.

    Cal. for (C₁₃H₈O₆NS)₂Ba + 7H₂O.         Found.
                        H₂O = 14.40         14.44.
                         Ba = 18.29         18.03.

In making a further supply of the salt it was found that if the
solution, after filtering from the barium carbonate, was diluted to
such an extent that no crystals separated on cooling, then on slow
evaporation under a bell-jar the first crystals to appear were very
long slender needles. As evaporation proceeded, these needles became
much thicker assuming prismatic proportions, and corresponded in
appearance to the salt described by Hollis as having six molecules of
crystal water.

As growth proceeded, the crystals became dark in color, and the
mother-liquor correspondingly clearer, the crystals evidently absorbing
the impurity in their growth.

When the solution had become quite colorless, rhombic crystals of the
salt containing seven molecules of water of crystallization appeared.
The larger prismatic crystals were carefully removed, and redissolved
in water in order to see if the same phenomena would repeat themselves.
This in fact was the case, crystals of both types appearing in the
same way as described. Without separating the crystals in this second
experiment, water was added, and the crystals dissolved. The solution
was then warmed briskly with boneblack, and filtered. From the
filtrate, which was colorless, nothing but rhombic crystals having
seven molecules of water of crystallization could be obtained,
although a great many variations in the conditions were tried. Analysis
of these last crystals was as follows:

0.2400 gram lost 0.035 gram at 210°, and gave 0.0637 gram BaSO₄.

    Cal. for (C₁₃H₈O₆NS)₂Ba + 7H₂O.         Found.
                        H₂O = 14.40         14.58
                         Ba = 18.29         18.27

Hollis states that treatment with boneblack decomposes this salt,
and hence he did not purify it prior to crystallization. From the
experiments just described it seems probable that the impurities
present affect the crystalline habit, and the degree of hydration of
this salt in a very striking manner. By careful warming with boneblack
no decomposition was observed, and the crystals so obtained have
constantly seven molecules of crystal water.




VI. The Action of Alcohols upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.


Kastle[15] found that when the chlorides of paranitroorthosulphobenzoic
acid (which he supposed to be an individual) were dissolved in alcohol,
and the solution boiled for some time, the acid etherial salt of
paranitroorthosulphobenzoic acid was the final product. The reactions
were shown to be:

            COCl                   COOC₂H₅
           /                      /
      I. C₆H₃——SO₂Cl + C₂H₅OH  = C₆H₃——SO₂Cl  + HCl.
           \                      \
            NO₂                    NO₂

            COOC₂H₅                COOC₂H₅
           /                      /
     II. C₆H₃——SO₂Cl + C₂H₅OH  = C₆H₃——SO₂OC₂H₅ + HCl
           \                      \
            NO₂                    NO₂

            COOC₂H₅                   COOC₂H₅
           /                         /
    III. C₆H₃——SO₂OC₂H₅ + C₂H₅OH  = C₆H₃——SO₂OH + (C₂H₅)₂O
           \                         \
            NO₂                       NO₂

[15] Am. Ch. Journ. XI--281.

Kastle, it will be observed, gave the symmetrical formula to this
mixture of chlorides. Several acid etherial salts were made, and a
series of the neutral salts of various metals described by him.

The action of pure symmetrical chloride was studied in the same general
manner to see if the resulting products would be the same as those
formed from the mixed chlorides.


1. Action of Methyl Alcohol upon the Symmetrical chloride.

A portion of the chloride was dissolved in methyl alcohol, and the
solution boiled until a drop added to cold water gave no precipitate,
of unchanged chloride. The alcohol was then distilled off, and the
thick syrup remaining, diluted with water. This solution was
neutralized with barium carbonate and filtered. On cooling, the barium
salt crystallized in shining mica-like plates, or in yellowish needles
corresponding accurately with those described by Kastle. They gave the
following analytical results.

0.2664 gram lost 0.0211 gram at 150°, and gave 0.0870 gram BaSO₄.

             [    COOCH₃  ]
             [   /        ]
    Cal. for [C₆H₃——SO₂O  ]   Ba + 3H₂O
             [   \        ]
             [    NO₂     ]2
                                                Found
                             H₂O =  7.79         7.88
             [anhydrous salt] Ba = 20.85        20.85


2. In like manner the barium ethyl salt was made. It also agreed
perfectly with Kastle’s description, crystallizing in fine, colorless
needles, forming in tufts from a not too concentrated solution. In case
it is necessary to concentrate these solutions, it is of advantage
to add a small quantity of alcohol to the solution as this prevents
any great amount of saponification, which otherwise takes place to a
noticeable extent.

Analysis.

    I. 0.2824 gram lost 0.0276 gram at 180°, and gave 0.0860 gram BaSO₄.

   II. 0.2655 gram lost 0.0262 gram at 190°, and gave 0.0815 gram BaSO₄.

              [    COOC₂H₃ ]
              [   /        ]
     Cal. for [C₆H₃——SO₂O  ] Ba + 4H₂O.
              [   \        ]
              [    NO₂     ]2
                                                 Found.
                                               I       II
                         H₂O =  9.51          9.77    9.86
                         Ba  = 20.00         19.84   20.02

Kastle also found that by dissolving the mixed chlorides in alcohol
in the cold, and allowing the solution to evaporate, there separated
after a time, crystals of the chloride of the acid etherial salt of
paranitroorthosulphobenzoic acid whose formation and composition are
represented in equation I.

This same product was sought for when pure symmetrical chloride was
employed, but without success. In every case, crystals of unchanged
chloride separated, or else it was found that it had been completely
converted into the acid etherial salt. In another trial cold water
was carefully added in small portions, since Kastle found that such
treatment facilitated the separation of the substance; the chloride
alone appeared. Still other attempts were made to obtain the substance
by adding a large amount of water to the solution of the chloride
in alcohol, after it had stood for some time. In this way, quite a
precipitate was thrown down, and this was filtered off and crystallized
from ether. It always proved to be the symmetrical chloride, and none
of the other substance was obtained.

Karslake[16] in working with the symmetrical chloride of
orthosulphobenzoic acid, was unable to isolate the analogous compound,
although from the mixed chlorides, by the action of alcohols, Remsen
and Dohme[17] had obtained chloro-etherial salts.

[16] Inaug. Diss. J. H. Univ. 1895.

[17] Am. Ch. Journ. XI, 341.

In as much as the pure symmetrical chloride is relatively stable in
cold alcohol (it can be crystallized from warm alcohol with very
little loss), it is possible that it is more stable than the chloro
etherial salt, and that in consequence the latter, when formed, yields
more readily to the further action of alcohol than does the unacted
on chloride. Hence when the action begins, it at once proceeds to the
limit. The fact that the symmetrical chloride is rather sparingly
soluble in cold alcohol, making the use of concentrated solutions
impossible, may also be a factor in the case. Whatever may be the
cause, this substance could not be obtained under any conditions that
were devised.

Having in my possession a very small specimen of crystallized
unsymmetrical chloride, it was submitted to the action of ethyl
alcohol, under as nearly as possible the conditions employed by Kastle.
Crystals of a colorless substance were obtained, which in every respect
agreed with Kastle’s description of the chloride of the acid ethyl
etherial salt of paranitroorthosulphobenzoic acid. Crystallized from
ether they melted at 68°.

The conditions employed by Kastle in preparing the chloride would
undoubtedly lead to a relatively large proportion of unsymmetrical
chloride, and it is to this chloride that the formation of the chloro
etherial salt is apparently due.




VII. The Action of Phenols upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.


Remsen and Saunders[18] in their investigation of the chlorides of
orthosulphobenzoic acid, studied the action of phenol upon these
substances, and from both the symmetrical chloride and the mixed
chlorides, they obtained a normal diphenyl ether together with a
red substance which was not further studied. It was formed in small
quantity and was probably the corresponding sulphonphthalein. Later
McKee[19] obtained these same substances from both the symmetrical
and the unsymmetrical chlorides. R. Meyer[20] obtained analogous
substances by the action of various phenols upon phthalyl
chloride. It seemed probable, therefore, that the chlorides of
paranitroorthosulphobenzoic acid would yield similar derivatives,
and a study was accordingly made of the reaction of the symmetrical
chloride with a series of phenols. The products in some instances were
exceedingly difficult to deal with, possessing properties that made it
impossible to prepare them for analysis, but even in such cases there
could be little doubt as to the general nature of the reactions which
had occurred.

[18] Am. Chem. Journ. XVII, 347.

[19] Ibid. XVIII, 798.

[20] Ber. XXVI, 204.


1. The Action of Phenol upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

A portion of the symmetrical chloride was brought together with
somewhat more than double the molecular amount of phenol. The mixture
was placed in a good-sized test-tube and the temperature gradually
raised by means of a sulphuric acid bath.

As soon as the phenol melts, some slight action occurs, as is indicated
by the fact that the mixture assumes a bright red color. No appreciable
amount of hydrochloric acid gas is evolved however, until the liquid
mixture has reached a temperature of about 115°. At this point the
gas is freely evolved, and the action is complete at a temperature
of 125°. The temperature observations were made by means of a
thermometer used as a stirring rod in the mixture. During the heating,
the color of the liquid becomes a much more intense red, growing darker
in shade, and the liquid itself becomes somewhat viscous but does not
solidify while hot.

When cool, the melt was repeatedly extracted with boiling water,
the aqueous solution being very deep purple in color. The colored
matter was removed very slowly in this manner, and so the process
was continued with dilute alkali. A solid insoluble residue was thus
obtained, of a light-brownish color. This was dissolved in alcohol,
boiled with boneblack and filtered. On cooling, needles of a straw
yellow color were deposited from the alcoholic solution.

This proved to be the normal diphenyl etherial salt of
paranitroorthosulphobenzoic acid, the formation of this substance being
expressed by the equation:

        COCl                   COOC₆H₅
       /                      /
    C₆H₃——SO₂Cl + 2C₆H₅OH = C₆H₃——SO₂OC₆H₅ + 2HCl.
       \                      \
        NO₂                    NO₂

Analysis of the substance gave the following results:

      I. 0.1627 gram gave 0.3398 gram CO₂ and 0.0510 gram H₂O.

     II. 0.1999 gram gave 0.4180 gram CO₂ and 0.0600 gram H₂O.

    III. 0.2649 gram gave 0.1561 gram BaSO₄.

                 COOC₆H₅
                /
    Cal. for C₆H₃——SO₂OC₆H₅
                \
                 NO₂                 Found.
                               I       II     III

              C = 57.14      56.97    57.03    ——
              H =  3.26       3.47     3.33    ——
              S =  8.02       ——       ——    8.09

This substance melts at 119° (uncorr).

It possesses properties similar to those of the diphenyl etherial
salt of orthosulphobenzoic acid described by Saunders. It is
insoluble in water, and is unaffected by hydrochloric acid or aqueous
alkali. On heating for a short time with alcoholic potash, the
needles were transformed into a voluminous precipitate. This was
filtered off, dissolved in water, and hydrochloric acid was added.
On cooling, characteristic crystals of the acid potassium salt of
paranitroorthosulphobenzoic separated.

Analysis.

0.1392 gram lost 0.009 gram at 150° and gave 0.0385 gram K₂SO₄.

                 COOH
                /
    Cal. for C₆H₃——SO₂OK + H₂O
                \
                 NO₂
                                     Found.
                  H₂O =  5.95     H₂O =  6.51
                    K = 13.65       K = 13.35

No attempt was made to isolate the corresponding intermediate
chlor-etherial salt of the composition

        COOC₆H₅
       /
    C₆H₃——SO₂Cl
       \
        NO₂

or its acid as was done by McKee[21] in his work on the analogous
etherial salt of orthosulphobenzoic acid.

[21] Am. Ch. Journ. XVIII-799

On evaporating the aqueous extract from the original melt almost to
dryness on the water-bath, there was a deposit on the sides of thedish
of scales possessing a beautiful bronze-green metallic lustre They
formed a deep purple solution in alkalis, or magenta, if the solution
was very dilute, and orange-yellow in acids. On acidifying the alkaline
extract with hydrochloric acid, this same substance was precipitated
as a brownish flocculent precipitate. It was, however, found to be
impossible to obtain this substance in pure condition. The amount
formed in the reaction is small, and its properties were such as to
render work with it very difficult. The method of precipitation is not
satisfactory because, owing to the fact that the substance is soluble
in acid solutions to an unusual extent for substances of this class,
the solution had to be concentrated to such a degree as to render the
precipitated substance very impure from acids and alkali salts. These
could not be removed by washing, obviously, without again dissolving
the substance. From its properties however, and its color reactions,
there can be little doubt that the substance is a sulphonphthaleïn,
and that it is always formed in considerable quantities in the
reaction of phenol upon the symmetrical chloride.

It was noticed that the aqueous extract of the mass left after fusion
was almost always decidedly acid in reaction, and it was thought that
this might be due to the formation of an acid etherial salt, whose
formation would be expressed by the equations:

        COCl                   COOC₆H₅
       /                      /
    C₆H₃——SO₂Cl + C₆H₅OH = C₆H₃——SO₂Cl + HCl.
       \                      \
        NO₂                    NO₂

        COOC₆H₅            COOC₆H₅
       /                  /
    C₆H₃——SO₂Cl + H₂O = C₆H₃——SO₂OH + HCl.
       \                  \
        NO₂                NO₂

Accordingly, the solution was saturated with barium carbonate, the
excess of carbonate removed by filtration, the filtrate concentrated,
and allowed to cool. Crystals in the form of pearly scales separated,
which upon analysis proved to be the neutral barium salt of
paranitroorthosulphobenzoic acid.

0.2291 gram anhydrous salt gave 0.1386 gram BaSO₄.

                   COO
                  /   \
                 /     Ba
                /     /
    Cal. for C₆H₃——SO₂O
                \
                 NO₂
                                   Found
            Ba = 35.85             35.57

This would seen to indicate that the reaction is an incomplete one even
in the presence of excess of phenol. No indications of the formation of
an acid etherial salt was observed.


2. The Action of Orthocresol upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

With orthocresol the reaction proceeds with more difficulty. A higher
temperature was required (135°-145°), and quite an amount of tarry
material was obtained from which very little could be extracted. The
product was warmed repeatedly with dilute alkali, the solution so
obtained neutralized with hydrochloric acid, and distilled with steam
for several hours to free it from cresol. The resulting solution was
then evaporated to small volume, and acidified with hydrochloric acid.
A considerable precipitate was thrown down, which was easily filtered
off and dried. In this condition it is a dark purple-red powder, lumps
of which possessed a yellowish-bronze metallic lustre. In dilute
alkaline solution it forms a deep-bluish purple solution, while in
acids it is crimson, or light yellow if the solution is dilute. It is a
excellent indicator, especially with ammonia.

In the insoluble tarry substance the etherial salt was sought for
and obtained in small quantity only. As this substance is soluble in
alcohol, and separates again on cooling in much the same condition,
the etherial salt could not be isolated be crystallization from this
solvent. By boiling the substance with benzene, purifying the filtrate
with boneblack, and allowing the benzene to evaporate, an almost
colorless gummy substance was obtained, which when dissolved in
alcohol, crystallizes in small colorless needles which melt at 89°-90°.
They were not obtained in quantity sufficient for analysis, but there
was little doubt that they were crystals of the diorthocresol etherial
salt.

Apparently much more decomposition occurred in this reaction than
when paracresol was employed, probably in consequence of the higher
temperature required for the reaction.


3. The Action of Paracresol upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

This reaction was conducted in the same manner as with phenol. No
hydrochloric acid was evolved until a temperature of about 110° was
reached, although after melting, the solution had steadily darkened to
a deep reddish-brown color. At 130°, after heating for several hours,
hydrochloric acid ceased to be evolved. The product was treated as in
the last experiment. The alkaline extract did not exhibit any marked
color reactions, such as were observed in most of these experiments,
being dull reddish-brown in both acid and alkaline solution.

The insoluble residue crystallized from alcohol in light brown
transparent crystals, which did not lose their color by repeated
crystallization, and boiling with boneblack, and melted sharply at
117°. From benzene they crystallized in colorless needles or flat,
narrow plates. These become opaque on exposure to the air, apparently
through loss of benzene of crystallization.

Analysis of the needles from alcohol gave the following results:

     I. 0.2372 gram of substance gave 0.5137 gram CO₂ and
        0.0965 gram H₂O.

    II. 0.2223 gram gave 0.1203 gram BaSO₄.

                COOC₆H₄.CH₃
               /
    Cal. for C₆H₃——SO₂OC₆H₄.CH₃
               \
                NO₂
                                   Found.
                                 I       II
            C = 59.08          59.06
            H =  3.98           4.52
            S =  7.49                   7.43


4. The Action of Hydroquinone upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

Action with hydroquinone occurs at 120°-135°, the mixture at the same
time becoming dark colored and viscous.

On cooling, the product was powdered and treated with dilute alkali.
It readily dissolved, without residue, forming a dark red solution.
In concentrated solution the addition of acid produces a voluminous
precipitate, dark brown in color, which when washed, and dried in paper
forms an almost black powder. A dilute solution of this powder is dark
red when alkaline, orange-yellow when acid.

From the way in which this powder was obtained, and owing to the fact
that its solubility prevented repeated washing, it was evident that it
would not give close analytical results for a calculated formula. It
was thought, however, that analysis would give a general idea of the
composition.

Analysis of different specimens gave results for sulphur which averaged
about 5.5%. The percentage required for the formula

          C[C₆H₃(OH)₂]₂
         / \
        /   O
       /   /
    C₆H₃——SO₂
       \
        NO₂

which represents the simplest sulphonfluoresceïn, is 7.43.

The compound could hardly have been so far from pure as to occasion
such a discrepancy in results as this. It would appear, therefore, that
more than two molecules of hydroquinone enter into the reaction with
one molecule of the chloride. Should four molecules be involved in the
reaction, leading to a compound of some such formula as

           C[C₆H₃(OH)₂]
          / \
         /   O
        /   /
    C₆H₃——SO[C₆H₃(OH)₂]₂
        \
         NO₂

the theoretical percentage of sulphur would be 6.00 which corresponds
much more closely with the results obtained.

This is in accord with the observations of a number of workers in
this laboratory—Lyman, Gilpin, Linn and others—who have worked on
various sulphonfluoresceïns, and have found that in many cases four,
six and even eight phenol residues condense with one molecule of the
anhydrous acid. Lyman[22] especially describes a tetra hydroquinone
sulphonfluoresceïn derived from orthosulphoparatoluic acid. No etherial
salt was observed.

[22] Am. Chem. Journ. XVI-525

5. The Action of Resorcin upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

The reaction of resorcin with the chloride is a much cleaner one
and proceeds more easily than in the case just described, leading
apparently to an individual compound which is well characterized.

During the reaction, which is complete at 125°, the mixture becomes
almost perfectly solid, and when cool, it is quite brittle. It was
reduced to a reddish powder in a mortar and dissolved in sodium
hydroxide, there being no insoluble residue. By the addition of
hydrochloric acid, the sulphonfluoresceïn was thrown down as a
chocolate-brown precipitate, which was filtered off, washed to
neutral reaction on a filter, and dried on paper. In this condition
it is a light chocolate-brown powder. In dilute alkaline solution it
possesses a slight fluorescence being pink by transmitted and yellow be
reflected light, suggesting eosin in a general way. It is interesting
to note that the sulphonfluoresceïn of orthosulphobenzoic acid
possesses a fluorescence that can hardly be distinguished from ordinary
fluoresceïn and that the introduction of a nitro group into the acid
residue produces much of the same effect as do the four bromine atoms
in eosin. In acid solution the color is reddish-orange.

Analysis of the compound, prepared as above described, gave the
following results.

      I. 0.1745 gram gave 0.3339 gram of CO₂ and
         0.059 gram H₂O.

     II. 0.1467 gram gave 0.2820 gram CO₂ and
         0.0432 gram H₂O.

    III. 0.1732 gram gave 0.3345 gram CO₂ and
         0.0571 gram H₂O.

     IV. 0.2000 gram gave 0.1104 gram BaSO₄.

      V. 0.1505 gram gave 0.0820 gram BaSO₄.

                        OH ]
                       /   ]
                   C[C₆H₃  ]
                  / \  \   ]
                 /   O  OH ]2
                /   /
    Cal. for C₆H₃——SO₂
                \
                 NO₂
                                      Found
                           I      II     III     IV      V
            C = 52.66    52.18   52.42   52.67    ——     ——
            H = 3.46      3.76    3.27    3.66    ——     ——
            S = 7.39        ——      ——      ——   7.57   7.48

An effort to obtain the anhydride was unsuccessful. Some loss of weight
was observed, but the compound underwent decomposition before this loss
amounted to much.


6. The Action of Pyrogallol upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

The product of this action dissolves readily in dilute sodium hydroxide
without residue, producing a very deep purple-black color when
concentrated, passing to grayish-violet as the solution is diluted. On
adding hydrochloric acid, precipitation occurs, as in most of these
reactions. On attempting to filter off this precipitate, it forms a
sticky, black mass on the filter with which little can be done. It is
best to evaporate to dryness before filtration and powder the residue.
This powder can then be washed fairly clean from alkali salts and
acid.

Nothing to suggest the formation of an etherial salt was observed.

Analysis of this product for sulphur showed that in this galleïn, as
in the case of the hydroquinone phthaleïn more than two pyrogallol
residues had entered the acid residue. The indications were that
six had entered into one of the chloride. This also agrees with the
observation of Lyman[23], who describes a hexapyrogallol galleïn of
orthosulphoparatoluic acid.

Probably a mixture of varying composition was obtained, and little
importance was attached to the results save as they showed that no
etherial salt is formed in the reaction.

[23] Am. Ch. Journ. XVI-527.


7. The Action of β-Naphthol upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.

It was hoped that here, as in the case of the monohydroxy phenols an
etherial salt would be obtained. It was found, however, that very
little action occurred, save such as was indicated by the development
of a bright carmine color in the melted mixture, until a temperature of
about 160° was reached. At this point hydrochloric acid was evolved,
but the chloride itself undergoes decomposition. Nothing definite could
be isolated among the reaction products, save unchanged β-Naphthol.




VIII. The Action of Aniline upon the Symmetrical Chloride of
Paranitroorthosulphobenzoic Acid.


As has been pointed out in the Introduction, it was in connection
with the aniline derivatives of orthosulphobenzoic acid, that the
isomerism of the chlorides was first noticed, two anilids being
obtained. Accordingly, when Gray began his study of the chlorides
of paranitroorthosulphobenzoic acid, his first effort was to obtain
evidence of the existence of two anilids. These were not obtained,
however, until after the chlorides themselves had been isolated, as
their properties made their isolation and preparation a matter of
difficulty.

Some points still remained in doubt after Gray’s study, and a further
investigation was thought to be desirable to clear these up.

Some time was spent in an endeavor to obtain a method by which a good
yield of fusible, or symmetrical, anilid could be obtained. The yield
in all cases tried, is not a good one. The presence of the nitro group
appears to complicate the reaction, leading to secondary reactions
whose course could not be followed. Upon bringing aniline and the
chloride together, a very vivid red color was always observed, and the
same was true when it was necessary to employ alkali. The fact that
such colors develop when nitro compounds are treated with alkali has
been noticed in many instances and some progress has been made in the
study of these compounds. Jackson and Ittner[24] have lately reviewed
this subject.

If a solution of the symmetrical chloride in ether is slowly added to a
similar solution of aniline, no appreciable amount of heat is evolved.
If the resulting solution is allowed to stand at ordinary temperatures,
action proceeds very slowly, aniline hydrochloride being precipitated
as the reaction proceeds. This can be filtered off from time to time
and the rate of action so observed. In such a way it was found that
five grams of chloride required about fifty hours time to react
completely with an excess of aniline. Similar results were obtained
with chloroform as the solvent. By boiling the solution for an hour or
more the reaction is complete.

[24] Am. Chem. Journ. XIX-199

The method employed was to bring the chloride and an excess of
aniline—somewhat more than four molecules—together in chloroform
solution. The flask was then boiled for about an hour, when the
chloroform was distilled off. During the boiling as well as the
distillation more or less bumping occurs in consequence of the aniline
hydrochloride which separates, and constant shaking of the flask is
sometimes necessary. The residue which is in a thick, gummy condition
in consequence of the presence of an excess of aniline, was digested
with water acidulated with hydrochloric acid. The excess of aniline
is thus removed, and the reaction product obtained as a reddish-brown
solid substance. This was treated with dilute sodium hydroxide, all
lumps being broken up with a stirring rod. The undissolved substance
is largely anil, which was filtered off. The anilid was then regained
by acidifying the alkaline solution, in which it was dissolved. It
separates immediately as a curdy colorless precipitate, though it is
frequently colored pink by impurity. It was found that this color could
be removed, in case not much was present, by redissolving the anilid in
alkali, and slowly pouring the solution into an excess of dilute acid.

In all cases a considerable amount of anil was obtained, even when the
substances were employed in the molecular ratios of 1:10. The reactions
involved, so far as the formation of anilid and anil are concerned are,

        COCl                    CO.NH.C₆H₅
       /                       /
    C₆H₃——SO₂Cl + 4C₆H₅NH₂ = C₆H₃——SO₂.NH.C₆H₅ + C₆H₅NH₃Cl
       \                       \
        NO₂                     NO₂

                                  CO
                                 /  \
        COCl                    /    N.C₆H₅
       /                       /    /
    C₆H₃——SO₂Cl + 3C₆H₅NH₂ = C₆H₃——SO₂   +  2C₆H₅NH₃Cl
       \                       \
        NO₂                     NO₂

On the whole the reaction seemed to be the most satisfactory in
chloroform solution, the main objection being, that, owing to the
simultaneous presence of chloroform, alkali, an a trace of aniline,
phenyl isocyanide is always formed, and renders the work more or less
unpleasant.

A number of experiments were also made to see if the yield could
be increased be employing a modification of the “Schotten-Baumann
Reaction”[25] for the formation of anilids. For this purpose an etherial
solution of the chloride was added to a like solution of aniline in
which was suspended finely powdered anhydrous potassium carbonate. The
proportions of the substances were those demanded by the equation

        COCl                             CO.NH.C₆H₅
       /                                /
    C₆H₃——SO₂Cl + 2C₆H₅NH₂ + 2K₂Cl₃ = C₆H₃——SO₂NH.C₆H₅ + 2KCl + 2KHCO₃
       \                                \
        NO₂                              NO₂

Very little anilid was, however obtained, but in its place a
substance soluble in water, of acid reaction capable of forming salts
and yielding several well characterized derivatives. I hope to
investigate this reaction more fully at some future time.

[25] Ber. XVII-2545; XXIII, 3430.

       *       *       *       *       *

The anilid is rather sparingly soluble in alcohol, from which it is
deposited on cooling in very small needles. These melt, as stated by
Gray, at 222°. It is also soluble in chloroform and glacial acetic
acid, but does not form well defined crystals from any solvent. It
dissolves in dilute alkali from which solution acids precipitate it
unchanged.

       *       *       *       *       *

The anil is also soluble in alcohol, glacial acetic acid etc. It
crystallizes in much better-formed crystals than does the anilid. These
melt at 188°.

On boiling the anil with aniline for a time, it is converted into the
anilid

          CO
         /  \
        /    N.C₆H₅              CO.NH.C₆H₅
       /    /                   /
    C₆H₃——SO₂   +   C₆H₅NH₂ = C₆H₃——SO₂NH.C₆H₅
       \                        \
        NO₂                      NO₂

In none of these reactions was any infusible anilid observed.




IX. The Action of Phosphorus Oxychloride upon the Fusible Anilid.


Hunter[26] found that when either of the anilids of orthosulphobenzoic
acid were treated with phosphorus oxychloride, or similar dehydrating
agents, a molecule of water was abstracted with the formation of a new
substance. A careful study of the compound led to the belief that it
was a dianil, and that its formation and structure could be represented
by the equation

                            C=N.C₆H₅
                           / \
        CO.NH.C₆H₅        /   \
       /                 /    .N.C₆H₅
    C₆H₄            = C₆H₄    /         + H₂O.
       \                 \  /
        SO₂NH.C₆H₅        SO₂

A corresponding study of the fusible anilid of
paranitroorthosulphobenzoic acid was undertaken.

The method employed in this study was as follows. A tubulated retort
of convenient size was fused onto the inner tube of a small condenser.
This was done to avoid connections, which are nearly always attacked by
the oxychloride. Another satisfactory plan is to have the neck of the
retort of the same size as the inner tube of the condenser. The ends
are placed in contact, and the tubes bound in position by wrapping with
asbestos paper. Over the joint so made, a tight rubber tube is drawn.

[26] Am. Ch. Journ. XVIII-810.

A convenient amount of phosphorus oxychloride (50 c.c.) was placed
in the retort and the anilid (5 gr.) added through the tubulus. On
boiling, with the condenser inverted, the anilid soon dissolved,
with evolution of hydrochloric acid gas, and the solution became
bright yellow in color, sometimes inclining to orange. The boiling was
continued as long as hydrochloric acid was given off. The oxychloride
was then distilled off under diminished pressure, care being taken to
shake the retort constantly during the distillation as violent bumping
is almost sure to occur especially towards the end of the operation.
The product remaining, spattered over the walls of the retort, was a
greenish yellow solid.

Water was then added, and the whole allowed to stand for an hour or
so to thoroughly dissolve the phosphoric acid formed in the reaction.

In case the anilid is not perfectly dry, a much more energetic reaction
occurs, and on distilling off the oxychloride, the product remains as a
dark, gummy mass. This should be spread out on the sides of the retort
while still liquid. On cooling and adding water, this gum gradually
disappears, and in its place is found the yellow solid product just
described. The gum appears to be a solution of this substance in
phosphoric acid.

After the substance is filtered off and dried, it can be crystallized
from acetone, benzene, glacial acetic acid or alcohol. From these
solvents it crystallizes in small yellow needles resembling quinone in
appearance.

The crystals obtained form acetone are rather larger than those from
the other solvents, and are more nearly orange in color, apparently
because of their greater compactness. When glacial acetic acid is used,
care must be taken to avoid any unnecessary heating, as continued
heating produces a change that will presently be described. The
substance melts at 208°.

Analysis of the substance resulted as follows:

      I. 0.3822 gram gave 0.8334 gram CO₂ and 0.1272 gram H₂O.

     II. 0.2645 gram gave 0.5812 gram CO₂ and 0.0910 gram H₂O.

    III. 0.2023 gram gave 0.1283 gram BaSO₄.

     IV. 0.2061 gram gave 0.1280 gram BaSO₄.

      V. 0.1853 gram gave 16.73 C.C.N (Standard).

                    C=N.C₆H₅
                   / \
                  /   .N.C₆H₅
                 /   /
    Cal. for C₆H₃——SO₂
                 \
                  NO₂
                                         Found.
                            I      II      III     IV       V
          C = 60.11       59.47   59.93     ——      ——      ——
          H =  3.44        3.69    3.82     ——      ——      ——
          S =  8.45         ——      ——     8.70    8.52     ——
          N = 11.08         ——      ——      ——      ——    11.35

For analyses I &amp; II I am indebted to Mr. Nakaseko, who kindly made them
for me.




X. The Action of Reagents upon the Dianil of
Paranitroorthosulphobenzoic Acid.


1. The Action of Hydrochloric Acid on the Dianil

When the dianil is boiled for some time with concentrated hydrochloric
acid, the yellow color of the substance disappears, and the dianil is
converted into a colorless substance without, however, passing into
solution. The substance so obtained was filtered off, and crystallized
from alcohol. It crystallized in small colorless needles, which melted
at 183°, and possessed all the properties of the anil, which, in fact,
it proved to be. The reaction was therefore

           C=N.C₆H₅                    CO
          / \                         /  \
         /   N.C₆H₅                  /    N.C₆H₅
        /   /                       /    /
    C₆H₃——SO₂      + HCl + H₂O = C₆H₃——SO₂      + C₆H₅NH₃Cl
        \                           \
         NO₂                         NO₂

This reaction also explains the fact that some anil was always obtained
in making the dianil from the anilid. Hydrochloric acid is formed
in the reaction, and in turn acts on the dianil in the sense of the
equation just given.


2. The Action of Alcoholic Potash on the Dianil.

On boiling the dianil with alcoholic potash for a time, the solution
turned red, and nothing but tarry products were obtained. In this
respect the dianil differs from the dianil of orthosulphobenzoic acid,
which under similar conditions, is transformed into infusible anilid.
This observation is, however, in keeping with the fact that the nitro
derivative, is in general much less stable in the presence of alkali.


3. The Action of Glacial Acetic Acid on the Dianil.

When the dianil is boiled with glacial acetic acid for some time, the
color of the solution changes to a much lighter shade of yellow, or
becomes colorless. On evaporating the solution to small volume, and
allowing it to cool, a colorless substance separates. This is infusible
anilid. It could not be obtained in crystals from any solvent, but
always separated in flakes. It does not melt or undergo change at 340°.

Like the fusible anilid it dissolves in dilute alkali, but on
acidifying the solution it does not immediately reappear. After
standing for some time, however, it gradually separates in perfectly
pure form. In this particular my observation differs from that of
Gray,[27] who states that this anilid is decomposed by solution in
alkali.

[27] Inaug. Diss. J. H. Unis. 1895.

A specimen that had been repeatedly precipitated gave the following
results on analysis.

      I. 0.1607 gram gave 13.88 C.C.N. (standard).

     II. 0.2195 gram gave 0.1285 gram BaSO₄.

    III. 0.1357 gram gave 0.0807 gram BaSO₄.

                   C[NH.C₆H₅]₂
                  /  \
                 /    O
                /    /
    Cal. for C₆H₃——SO₂
                \
                 NO₂
                                    Found.
                              I.     II.     III.

          N = 10.58         10.85     ——      ——
          S =  8.06           ——     8.00    8.16

By this series of transformations it is possible to pass from one
anilid to the other, the steps being:

          CO.NH.C₆H₅         C=N.C₆H₅            C[NH.C₆H₅]₂
         /                  / \                / \
        /                  /   N.C₆H₅         /   O
       /                  /   /              /   /
    C₆H₃——SO₂NH.C₆H₅ ➡ C₆H₃——SO₂      ➡ C₆H₃——SO₂
       \                  \                  \
         NO₂               NO₂                NO₂

This is of special interest as affording a means of passing from a
derivative of one of the chlorides, to a substance derived from the
other, by steps that can be clearly followed.




Conclusions.


In the course of this investigation several facts have been established.

1. By the methods described, the symmetrical chloride of
paranitroorthosulphobenzoic acid can be obtained in fine crystalline
form, perfectly free from its isomer, with an average yield of forty
percent.

2. By treatment of the chloride with benzene and aluminium chloride,
only one chlorine atom can be replaced by a phenyl group.

3. The barium salt of paranitroorthobenzoyl benzenesulphonic acid, when
perfectly pure, crystallizes constantly with seven molecules of water
of crystallization.

4. With alcohols, the symmetrical chloride yields directly the acid
etherial salt of paranitroorthosulphobenzoic acid, no evidence
having been obtained of an intermediate chloro-etherial salt. The
unsymmetrical chloride on the other hand yields the intermediate
product.

5. With phenols, two series of derivatives are obtained.

    (1) With monohydroxy phenols, both etherial salts and
        sulphonphthaleïns are formed, the former predominating.

    (2) With polyhydroxy phenols no etherial salts were obtained,
        but compounds of the unsymmetrical type, usually containing
        more than two phenol residues.

6. With aniline an anil and an anilid of symmetrical constitution are
formed.

7. With phosphorus oxychloride, the anilid, by loss of water, forms a
dianil.

8. This dianil undergoes transformation with

    (1) Glacial acetic acid, forming an anilid of unsymmetrical
        constitution.

    (2) Hydrochloric acid forming the anil.

    (3) Alcoholic potash, with the formation of colored decomposition
        products.




Biographical.


The author of the foregoing dissertation was born at Wilkinsburg, Pa.,
Jan. 29., 1870. Owing to prolonged sickness in childhood his education,
prior to entering college, was much interrupted, and was largely
confined to instruction received at home.

In the fall of 1887 he entered Wooster University (Ohio), from which
institution he received the degree of Bachelor of Arts in 1891. The two
following years were spent as a teacher of Sciences in the College of
Emporia (Kansas). In 1893 he entered the Johns Hopkins University where
he has since been a student of chemistry, with physics and mathematics
as subordinate studies.

In 1895 he was appointed University Scholar in Chemistry. During 1895-6
he served as lecture assistant to Prof. Remsen and Dr. Renouf in the
undergraduate courses. In the spring of 1896 he was appointed Fellow
for the present year.