CHEM. LIB.
Q D
341
A8
C975
B 453698
The University of Chicago.
Founded by JOHN D. ROCKEFELLER.
KENERAL LE
JUL 1 103
The Thermal Decomposition of Symmetrical
Diarylhydrazines A Reaction on
the First Order
!
A DISSERTATION
SUBMITTED TO THE FACULTIES OF THE GRADUATE SCHOOLS OF
ARTS, LITERATURE, AND SCIENCE, IN CANDIDACY FOR
THE DEGREE OF DOCTOR OP PHILOSOPHY.
DEPARTMENT OF CHEMISTRY
BY GEO. O. CURME, JR.
i
EASTON, PA.:
PRESS OF THE ESCHENBACH PRINTING CO.
1913
ļ
}
The University of Chicago.
Founded by JOHN D. ROCKEFELLER.
The Thermal Decomposition of Symmetrical
Diarylhydrazines
A Reaction on
the First Order
A DISSERTATION
SUBMITTED TO THE FACULTIES OF THE GRADUATE SCHOOLS OF
ARTS, LITERATURE, AND SCIENCE, IN CANDIDACY FOR
THE DEGREE OF DOCTOR OP PHILOSOPHY.
DEPARTMENT OF CHEMISTRY
BY GEO. O. CURME, JR.
EASTON, PA.:
PRESS OF THE ESCHENBACH PRINTING CO.
1913
The Thermal Decomposition of Symmetrical
Diarylhydrazines A Reaction on
the First Order
Many hydrazines, as is well known, show a tendency to undergo, with
great facility, a variety of molecular changes, involving a redistribution
of the atoms composing their molecules. The two chief types of such re-
actions are the following:
(1) Many hydrazines, when heated, decompose into compounds of a
lower and a higher state of oxidation. Thus hydrazobenzene decomposes
into aniline and azobenzene:
2C6H5NH.NHCH¸ → С¸H5N: NC6H5 + 2C6H5NH2 (1)
(2) Symmetrical arylhydrazines and their derivatives undergo the
benzidine and related rearrangements, under the influence of acids. Thus
hydrazobenzene changes into benzidine and diphenyline:¹
1
CHNH.NHCH + 2HC1 — ClH_N.CH4.CHẠNH,C1.
→
(2)
A study of the reactions of class (1) constitutes the work to be taken
up in this paper. The hydrazines used in the experimental work were
hydrazobenzene, p-methylhydrazobenzene, and p-hydrazotoluene. The
¹ Other products are theoretically possible, but have never been described as
occurring as a result of this rearrangement.
4
investigation was suggested by Professor Stieglitz, and was carried out by
the author under his supervision.¹
A series of quantitative determinations of the velocities of certain re-
actions of class (2) have also been carried out by us, with the object of
making a more critical analysis of the mechanism of the reaction, than that
reached by van Loon in a similar investigation.2 A different method of
measurement from that used by van Loon has been followed, and it is
hoped that within a short time the conclusions reached can be reported.
4
The oldest observation on the simultaneous formation of azo and amino
compounds from hydrazo compounds, under the influence of heat, is due
to A. W. Hofmann,³ who noticed that hydrazobenzene, when heated
considerably above its melting point, changes into a mixture of azoben-
zene and aniline. Biehringer and Busch found that roughly equivalent
quantities of azobenzene and of aniline are formed when an alcoholic
solution of hydrazobenzene is heated above 130° in a sealed tube. This
establishes the equation for the reaction, which the above-mentioned
investigators found to proceed irreversibly.5
2 Ph.NH.NH.Ph
PhN: NPh + 2PhNH.
(3)
While their work establishes the relative quantities involved in the re-
action, by which at least the greater part of the hydrazine is decomposed,
it gives no clue to the mechanism by which the reaction is effected.
There are, a priori, several possible, plausible explanations of the de-
composition, each of great interest from the point of view of the theory of
organic chemistry. For this reason, the decision between the possibili-
ties by the results of critical experimental work is of unusual interest and
importance. We will discuss the three most important of these schemes,
distinguishing them, for the sake of convenience, as A, B and C. In the
following discussion, the consideration will be limited to that of the de-
composition of hydrazobenzene into aniline and azobenzene. However,
the theoretical conclusions reached apply equally well to p-methylhydrazo-
benzene and p-hydrazotoluene, and probably to a large number of chem-
ically related substances.
A. In the first place, the formation of azobenzene and aniline might be
the result of an intermolecular oxidation and reduction, one molecule of
hydrazobenzene being oxidized to azobenzene by a second molecular
which itself would be reduced to aniline. The high reducing tension in
¹ See a preliminary report by Stieglitz and Curme, on the decomposition of hydrazo-
benzene, Ber., 46, 911 (1913).
2 Rec. trav. chim. Pays.-Bas., 13, 63 (1904).
3 Jahresb., 1863, p. 425.
4 Ber., 36, 339 (1902).
5 The symbol Ph represents the phenyl group.
5
the hydrogen attached to the nitrogen in the hydrazobenzene could be the
cause of such an intermolecular oxidation and reduction. This is the
conception of the reaction which has been advocated in particular by
Wieland.¹ At a constant temperature, such a reaction would be one of
the second order, and its velocity would be expressed by the dimolecular
law
dx
dt
=
Kdimol. (A — x)²,
(4)
where A is the original concentration of the reacting substance, x the con-
centration transformed at any time t, and Kdimol. the characteristic constant
of the reaction at the given temperature.
B. In the second place the formation of azobenzene and aniline from the
hydrazobenzene may be the result of a primary, measurably slow decom-
position of the hydrazine into aniline and phenyl imide, PhN:, a deriv-
ative of univalent nitrogen, followed by the condensation of two molecules
of phenyl imide to azobenzene, as expressed in the following equations:
PhNH.NHPh → PhNH2 + PhN:
2 PhN:
PhN NPh
(5)
(6)
The consideration of this possibility grew out of Stieglitz's theories2 on
the formation of univalent nitrogen derivatives as intermediate products
in the Hofmann-Curtius-Lossen-Beckmann rearrangements, in the benzi-
dine³ rearrangement, and in other reactions. The decomposition of
symmetrical diarylhydrazines into aromatic amines and azo compounds
held out the hope of a very simple experimental test, by velocity measure-
ments, of the applicability of such views to the reaction in question.
At a constant temperature such a decomposition would be one of the first
order, and its velocity would be subject to the monomolecular law:
dx
dt
= Kmono, (A-x),
(7)
where the symbols have the same significance as in equation (4).
C. Very closely related to this view and growing out of it is the third
possibility, namely, that the formation of the azobenzene and aniline from
¹ Ann. Chem., 392, 133 (1912).
2 Stieglitz, Am. Chem. J., 18, 751 (1896); 29, 49 (1903); Stieglitz and Earle, Ibid.,
30, 399 and 412 (1903); Slosson, Ibid., 29, 289 (1903); Hilpert, Ibid., 40,
155 (1908); Stieglitz and Peterson, Ber., 43, 782 (1910); G. Schroeter, Ibid., 42, 2336
and 3356 (1909), etc.; Wieland, Ibid., 42, 4207 (1909). Stieglitz and Curme, Ibid., 46,
911 (1913).
³ The application of this theory to the benzidine rearrangement, Am. Chem. J., 29,
62, foot-note, has now been abandoned as a result of Wieland's work, Ber., 39, 150
(1906), on the rearrangement of fully substituted hydrazines, and also of our measure-
ments of the velocity of the benzidine rearrangement (see Stieglitz and Curme, loc. cit.).
4 Preliminary report, Ber., 46, 911 (1913).
6
hydrazobenzene is preceded by a primary, measurably slow decomposi-
tion of hydrazobenzene into phenyl amide, PhNH, a derivative of bivalent
nitrogen. Such a decomposition would be entirely analogous to that as-
sumed by Wieland for fully substituted hydrazines¹ in a series of reac-
tions investigated preparatively:
Ph,N.NPh* 2Ph,N.
(8)
Wieland, however, explicitly stated that he considers the thermal de-
composition of hydrazobenzene not to be of this type, but rather, as ex-
plained above, an intermolecular oxidation and reduction reaction. We
can express this third view, a variation of the second one, in the following
equations:
PhNH.NHPh
> 2PhNH,
(9) 3
followed either by the action:
4PhNH → 2PhNH, + PhN: NPh
(10)
or by the actions:
2PhNH
PhN: + PhNH,
(11)
and
2 PhN:
PhN NPh.
(12)
At a constant temperature the primary reaction expressed in equation
(9) would also be one of the first order and its velocity would obey the
law for a monomolecular reaction.
The experimental results obtained showed that at a given temperature
the velocity of the reaction is independent of the concentration of the re-
acting substance and consequently the decomposition cannot be a di-
molecular reaction. When the proper substitutions are made in the equa-
tion for the velocity of a monomolecular reaction, concordant "constants'
are obtained both when the analyses are made for unchanged hydrazo-
benzene at the end of measured time intervals, and when the amount of
1 Ann. Chem., 381, 202 (1911).
2 Loc. cit.
""
3 We may consider the nitrogen atoms in hydrazobenzene to be united by a positive
charge on the one and a negative charge on the other (cf. W. A. Noyes, J. Am. Chem. Soc.,
23, 460 (1901); 35, 767 (1913); Stieglitz, Ibid., 30, 1798 (1908) and Stieglitz and Curme,
loc. cit.). In the dissociation the fragments might retain their charges (loc. cit.) but it is
better to assume that the electron passes from the negative to the positive nitrogen
atom and thus disrupts the molecule and that the dissociation product contains radicals
with only two valences of the nitrogen atom satisfied, i. e., that the latter has gained
+-
+−− +
only two electrons forming Ph.N.H. This is the view expressed in equation (9). The
alternative hypothesis that the nitrogen atoms retain their charges represents a simple
ionization, which the character of the compound renders much less likely than the de-
composition expressed in equation (9). J. S. and G. C.
7
aniline¹ formed in the same time intervals is determined. On the other
hand, widely discordant values for the "constants" are obtained when the
equation for a dimolecular reaction is used, both in the values of a single
experiment and in the values obtained for Kdimol. for solutions of varied
concentration (see Tables I and III, p. 18). The assumption that an
intermolecular oxidation and reduction represents the course of the action
is therefore disposed of.
2
Close inspection of the velocity determinations (see Tables I, II, III,
p. 18, and IV, V and VI, p. 20) shows that there is a slowing down of
the action, amounting to a falling off in the constant for a monomolecular
reaction of about 10% during the decomposition of about 75% of the
substance. A falling off in the value of the constant calculated for a
pure monomolecular reaction would result if the action really consisted
of two parallel reactions, one monomolecular (equation 5 or 9) and the
other dimolecular³ (equation 34), as expressed in the equation:
dx
Kmono. X (A-x) + Kdimol. X (A — x)².
•
dt
(13)
Inspection of Tables I, II, III, p. 18, and IV, p. 20, shows, how-
ever, that three solutions, varying in concentration as 4 2 1 all start
out with the same velocity, when referred to a monomolecular reaction,
and show exactly the same characteristic droop in the same time. If a
parallel dimolecular reaction were the cause of the droop, then the con-
stant found for a monomolecular reaction should start out with smaller
values in the more dilute solutions, in which the parallel dimolecular re-
1 The separate analysis of both substances was performed only for the decomposi-
tion of hydrazobenzene.
2 As calculated for the decomposition for long time intervals. The sag in the con-
stant would be still greater if calculated for small time intervals.
3 Such a parallel dimolecular reaction is suggested by the behavior of hydrogen
peroxide, which is closely related in structure and behavior to the hydrazines. Hydrogen
peroxide, according to Nernst [Z. Elektroch., 11, 710 (1905)], decomposes as a gas at
270° in a dimolecular reaction (2H2O2 → 2H2O + O2), but exposed to ultraviolet
light at low temperatures [Tian, Compt. rend., 151, 1040 (1910)] or to a series of catalytic
agents (see references to the literature by Tian, loc. cit.), it decomposes in a mono-
molecular reaction, H2O2 > H2O + O and 20 O2. The latter monomolecular
reaction is strictly comparable with that proved by us for the thermal decomposition
of hydrazobenzene. Tian considers that Wolffenstein's study of the decomposition
of hydrogen peroxide at 80° (Ber., 27, 3310 (1894)] shows that the thermal decomposi-
tion of the peroxide in solution must also be dimolecular. But Wolffenstein's data,
obtained for preparative purposes simply, are neither rigorous nor consistent and ac-
cording to the choice of experiments, one can consider a dimolecular, a monomolecular,
or parallel dimolecular and monomolecular reactions to be indicated. Mr. Denton
J. Brown has taken up this problem for rigorous study, as well as the problem of the
thermal decomposition of peroxide derivatives. J. S.
4 Such a dimolecular reaction could be due to the reduction of phenyl imide to
aniline by hydrazobenzene.
8
action would consume the substance at a slower rate than in the more
concentrated solution. It is apparent, therefore, that a parallel dimolecu-
lar reaction cannot be considered to account for the sag in the constant
for a monomolecular reaction. Mathematical analysis of the data fully
confirm this conclusion (see p. 20).
The experimental data therefore do not agree with the assumption
that the decomposition of hydrazobenzene is either wholly or in measurable
part the result of a dimolecular reaction of oxidation and reduction.
The data agree, on the other hand, with the conclusion that we are deal-
ing primarily with a monomolecular reaction, with some secondary action
causing a small droop in the constant during the course of the decomposi
tion, but not in the initial rates of decomposition, when solutions of concen-
trations in the ratio 4 2 1 are used. This result raises the question
whether a product of the action is not a retarding factor by virtue of a re-
versal of the action. Aniline and azobenzene are the products and it was
thought, in particular, that aniline might retard the decomposition by
combining reversibly with phenylimide to form hydrazobenzene (theory
B, equation 5). Such a reactive compound as phenyl imide must in fact
tend to combine with aniline, just as isocyanides RN: C combine with
amines. Velocity measurements were therefore made with mixtures of
hydrazobenzene and aniline; it was found that the action followed ex-
actly the same course as in the absence of aniline, starting at the same initial
rate and showing the same slight droop (see Table VIII, p. 24). While
the decomposition of hydrazobenzene into phenylimide and aniline may
be reversible, it is evidently not measurably so¹ and thus the droop in the
constant cannot be due to a reversed action of this nature.2 Exactly
the same result was obtained when an initial addition of azobenzene was
made to the hydrazobenzene (see Table IX, p. 24). Neither of the recog-
nized products of the action can therefore be held to account for the fall-
ing off of the velocity constant.
The possibility that the solvent, alcohol, is the cause of the secondary
action indicated, producing some reducible substance³ other than hydrazo-
benzene, was excluded by the fact that the same characteristic result was
obtained when the velocity was measured by analysis for aniline formed
and still more conclusively by the fact that the velocity of decomposi-
tion in benzene solution, although much slower than in alcohol, at a given
temperature, shows exactly the same character.
¹ The concentration of phenyl imide at any moment might be too minute and the
change too slow to bring the reverse action into the range of our measurements.
2 The question of reversibility was tested only in the case of hydrazobenzene.
While the result is apparently conclusive, the importance of the conclusion is such that
in the continuance of these studies, special attention will be paid to this question. J. S.
3 The reducing power of hydrazobenzene toward iodine is used in measuring the
velocity of the action.
9
While we have thus been able to exclude a parallel dimolecular reac-
tion, a reversed reaction of aniline or azobenzene, and an action of the sol-
vent¹ as the sources of the falling off in the constant for a monomolecular
reaction, we have not succeeded in finding the actual nature of the con-
flicting secondary reaction. It is probable that some derivative is formed,
with reducing power, that is tolerably stable and that has not been recog-
nized as yet as a product either by Biehringer and Busch,2 the first investiga-
tors of the decomposition, or by ourselves. In view of the great reac-
tivity of the substances involved and the many transformations hydrazo-
benzene is known to undergo, such a secondary reaction would not be
unnatural. An observation, made during the course of the analysis, that
some substance accumulates that forms with iodine a dark-colored prod-
uct, interfering with the delicacy of the titrations, is the only direct clue
we have as to the presence of such a product.
:
While the nature of the scondary reaction has not been discovered, we
consider that the perfect agreement in the initial rates of change of hydrazo-
benzene in solutions of concentrations in the ratio 4 2 1 as well as in
solutions containing an excess of aniline or of azobenzene proves conclu-
sively the monomolecular character of the thermal decomposition of hy-
drazobenzene. This result demonstrates that a primary, measurably
slow decomposition of hydrazobenzene into free unsaturated radicals,
phenyl imide, CH¿N, or phenyl amide, CHÁNH, must occur and that these
radicals must have at least a temporary existence, however brief it may be.
This quantitative proof of the formation and existence of such free radi-
cals, as intermediate products of reaction, we consider the most important
conclusion arrived at through this investigation.
A monomolecular reaction, in itself, would agree with the assumptions.
made either in the secondary theory presented (B, p. 5) or in the third,
(C, p. 5). The fact that the reaction velocity is not retarded by the
presence of an excess of aniline might, at a first glance, be considered to
favor theory C as against B, inasmuch as phenyl imide, a necessary prod-
uct according to B (equation 5), would be expected to act reversibly with
aniline. However, as stated above, the reversibility need not be measur-
ably great. But theory C would not demand any reversible action of
aniline at all, if aniline and azobenzene are formed directly (equations
9 and 10) from phenyl amide. A very careful test of the reduction of
azobenzene to hydrazobenzene (by a quantitative test for the formation
of a reducing agent active towards iodine, when a mixture of aniline and azo-
¹ Tests for benzidine showed that no discoverable quantities of benzidine are formed
and therefore excluded also the benzidine rearrangement as the secondary reaction
(a benzidine rearrangement would proceed faster at first than later in the more strongly
alkaline solutions).
2 Loc. cit.
IO
benzene was heated) showed in fact that no hydrazobenzene is formed
from these compounds under the conditions of the experiment. The ob-
servation that the thermal decomposition of hydrazobenzene is not re-
tarded by aniline would therefore agree directly with theory C, postulating
the primary formation of phenyl amide, provided the further course of the
action were that indicated in equation 10. But this course would involve
a tetramolecular reaction for the formation of aniline and azobenzene (see
equation 10). Reactions of such high order are exceedingly rare and they
should not be assumed without strong confirmatory evidence. For this
reason it is much more probable that if phenyl amide is the primary dis-
sociation product, it forms aniline and azobenzene by way of phenyl
imide¹ (see equations 11 and 12), which, as in theory B, would be expected
to act reversibly with aniline (although not necessarily measurably so).
It is clear then that the basis for a choice between theories B and C,
postulating, respectively, a primary decomposition into phenyl imide and
into phenyl amide, is too slight for any conclusive decision at present.
The vital fact remains established, however, that there is a measurably
slow, monomolecular decomposition into free radicals, which may be either
phenyl imide or phenyl amide.
The determination of the velocity of decomposition of hydrazobenzene²
at different temperatures shows that the rate of decomposition varies
very rapidly with small change of temperature. Van't Hoff's³ equation
for the change of velocity with change of temperature is
d ln K A
dT
T2'
(15)
where K is the velocity constant of the reaction, T the absolute tem-
perature, and A a characteristic constant where only small ranges of
temperature are being considered. By integration of equation 15 and
solving for A, an expression is obtained,
A
T.T.
X In
T₁ — To
Κι
Ko'
(16)
¹ The assumed intermolecular formation of phenyl imide and aniline from phenyl
amide corresponds, step by step, with what presumably occurs when alkalin liquids
are electrolyzed and hydroxide ions discharged: 2HO H2O + O and 20 O2
(14). Bivalent radicals are much more stable and perhaps more likely to be formed in
the free state than univalent ones. Ammonia and amines, oxygen and sulfur compounds
and isocyanides are instances of well-known compounds whose molecules contain
atoms with two free valences and which are known in the free condition, although ex-
tremely reactive. Triphenylmethyl, through Gomberg's classical researches, has be-
come known as one of the best studied of the free univalent radicals but, in a character-
istic fashion, it polymerizes and exists primarily as a polymer, [(C6H5)3C]2.
2 The influence of temperature on the velocity of decomposition was investigated
only with the one substance, hydrazobenzene.
3 Études de Dynamique Chimique, 1884, p. 114.
II
which should give a constant value for A at different temperatures, when
the corresponding values of T and K are substituted. Using the three
velocity constants¹ obtained experimentally with hydrazobenzene (see
Tables IV, V and VI, p. 20) at the temperatures 140° (= 413° Abs.),
145° (= 418° Abs.), and 150° (= 423° Abs.), respectively, we can calcu-
late three values for A.
(1) To
= 413°; Ti
'. A
=
418°; K。 = 0.00145; Kı = 0.00253.
19,210.
(2). T'% = 413°; T'₁ =
To
1
I
423°; K'o 0.00145; K' = 0.00436.
=
.'. A = 19,230.
(3) T″。 = 418°; T"1 = 423°; K"。
. A 19,240.
0.00253; K″1 = 0.00436.
A here is a function of the heat of reaction, but as it is safe to assume that
this is very close to constant over such a short range of temperature, the
excellent agreement of the different values of A shows that the rate of de-
composition changes regularly with the temperature in accord with the
thermodynamic principles on which van't Hoff's deduction is based.
It was not convenient experimentally to examin the behavior of the
three hydrazines, i. e., hydrazobenzene, p-methyl hydrazobenzene, and
p-hydrazotoluene, at the same temperature, and so to get a direct com-
parison of their velocities of decomposition. However, a comparison of
the temperatures at which they show approximately the same rate of change
shows clearly that the introduction of methyl groups has a marked weaken-
ing effect upon the force which holds the two nitrogen atoms together.
If the temperature coefficient of the velocity change were determined for
all of the substances, this would afford a quantitative measure of the change
of this force as well. In the tetra-arylhydrazine series Wieland² showed
qualitatively this same tendency, but here for the first time we have
an accurate method of detecting the change in stability of the substance
with the change in substituent groups. It is further noticed that, con-
trary to the experience of Wieland, those hydrazines most sensitive to
the influence of heat are also most sensitive to rearrangement by acids.
The fuller study of this detail will be published later, but the qualitative
fact is very well shown by the behavior of the different hydrazines on oxida-
tion in neutral (initially) solution, as was observed incidental to the analysis
of the pure substances. Here an alcoholic solution of the hydrazine was
added to an alcoholic solution, containing a known amount of iodine,
which was more than sufficient to oxidize the hydrazine to the correspond-
ing azo body. In the case of all hydrazines studied, under these conditions
1 The average value of K at each temperature is used for the purpose of this cal-
culation.
2 Ann. Chem., 392, 128 (1912).
3 Loc. cit., p. 132.
3
12
the amount of iodine consumed did not correspond rigorously to the amount
of hydrazine added. The acid formed in the earlier part of the oxida-
tion,
¹ArNH.NHAr + I2 → ArN NAr + 2HI,
:
(17)
caused a rearrangement of a part of the hydrazine, and so prevented its
reaction with iodine, according to this equation. In the case of hydrazo-
benzene, with certain conditions fixed, about 1% is rearranged by the acid
formed in the earlier part of the reaction, before the oxidation is com-
plete; in the case of p-methylhydrazobenzene this amount rearranged
is about 3% under the same conditions; and with p-hydrazotoluene it is
about 10%. The three hydrazines are sensitive to thermal decomposi-
tion in this same order. This shows qualitatively then that they are
probably² sensitive to rearrangement by acid in the same order that they
are sensitive to decomposition by heat.
In the case of p-methylhydrazobenzene there exist three ways in which
the thermal decomposition may take place. Aniline and p-azotoluene
may be formed:
2CH3C6H4NH.NHC6H5 →→
CH3CHAN NC6H4CH3 + 2C6H5NH2, (18)
p-toluidine and azobenzene may be formed:
2CH3C6H4NH.NHCH5 → C6H5N: NC6H5 + 2CH3C6H4NH2, (19)
or p-toluidine, aniline, and p-methylazobenzene may be the products:
2CHCHẠNH.NHCH →
3
CH3CHẠN : NCH + CHNH, + CH3CHẠNH,. (20)
Qualitatively both aromatic amines were identified in the reaction
mixture, and by an indirect method of analysis it was found possible to
estimate with rough accuracy the relative amounts of each formed. The
results showed that of the total amount of amine formed 33% was ani-
line and 67% was p-toluidine. This seems qualitatively in accord with
the observed fact that p-hydrazotoluene has the greater tendency to de-
compose, the diminution in attraction of the two halves of the hydra-
zine molecule for each other, caused by the introduction of methyl, ex-
presses itself here in the increased tendency of the tolyl group to exist
in the simpler molecule toluidine, rather than as a part of an azomole-
cule.
In the case of p-methylhydrazobenzene, an opportunity offered itself
to attempt to prepare two electromeric substances. p-Methylazobenzene
1 The symbol Ar here represents a simple or substituted phenyl radical.
2 The relative speeds of oxidation were not determined.
3 An effort will be made as soon as time allows to identify the three possible azo
compounds.
4 Falk and Nelson, School of Mines, Quarterly 30, p. 179; J. Am. Chem. Soc.,
32, 1637 (1910); W. A. Noyes, Ibid., 35, 767 (1913).
13
4
was formed from nitrosobenzene and p-toluidine (see experimental part):
C6H5.N++O+ CH3C¸H₁N¯ ¯H₂++
C¿H¿N++: N˜¯‚C6H4CH3 + H₂O. (21)
The substance thus prepared should have the indicated distribution of
charges on the nitrogen atoms, if no readjustment takes pace. On re-
duction, then, it should give the corresponding hydrazine. In a like man-
ner p-nitrosotoluene was treated with aniline:¹
CH3.C6H4.N++0˜¯ + C6H5.N˜¯H₂++
2
CH3.C6H4.N++: N.CH + H2O, (22)
and the substance thus prepared should have the indicated distribution
of charges on the nitrogen atoms, if no readjustment takes place. On re-
duction, it too might be expected to give the corresponding hydrazine.
However, the two preparations of p-methylhydrazobenzene proved iden-
tical in melting point, velocity of rearrangement, and percentage of p-tol-
uidine and aniline formed. The conclusion must be, then, if we cling to
the electronic theory, that only one of the two possible electromers is stable
under the conditions used.
EXPERIMENTAL PART
I. Decomposition of Hydrazobenzene
The hydrazobenzene used in this work was prepared from nitroben-
zene by the well-known method of reduction with zinc dust in an aqueous-
alcohol medium, strongly alkalin with caustic soda. The crude product
was twice recrystallized from 95% alcohol to which a few drops of strong
ammonia and several grams of zinc dust had been added.2 This pro-
cedure gave a snow-white product, melting sharply between 126° and 127°,
which is oxidized slowly in dry air to the yellow azobenzene. The sub-
stance thus prepared was dried for one hour in an oven at 50°, then in vacuo
over calcium chloride for two hours, and then it was sealed up in glass
bulbs under a pressure of 10-20 mm.
The samples were uniformly of the proper melting point, although very
slightly tinted with yellow through oxidation during the drying process.
Once sealed in the bulbs, no deterioration of the samples was noticed,
and they could be kept indefinitly, until used. Assays of the purity of
the samples by the iodine method (described below) consistently gave
values ranging between 99.7%-100.3%, which deviations represent
the experimental inaccuracies of the determination.
Analytical Method. Determination of Hydrazobenzene
Essentially the same method of analysis as was adopted in determining
¹ In aniline C6H5 probably carries a positive charge; in nitrosobenzene it is also prob-
ably positive, but may be negative. Since the distribution of charges between phenyl
and nitrogen is not material to the discussion, it will be omitted.
2 Van Loon, Rec. Trav. Chim. Pays.-Bas., 13, 63 (1904).
14
the purity of the hydrazobenzene was used in the velocity measurements,
for the determination of the quantity of unchanged hydrazobenzene left
at the end of any time interval, t. Accuracy demands certain precautions,
and so the method will be given in detail.
The method of analysis was based on the fact that iodine in alcoholic
solution oxidizes hydrazobenzene to azobenzene, with the formation of
two molecules of hydriodic acid per molecule of hydrazobenzene used
(equation 17).
Further on (p. 22), a second method of analysis, used to follow the
velocity of decomposition of hydrazobenzene by the determination of
the aniline formed, will be described.
The solutions used for the first method were:
(1) 0.1 N iodine made from 13. 1 g. of resublimed iodine, 18 g. of potas-
sium iodide, 1,000 g. of 50% alcohol. The solution was allowed to stand
one week before standardization, and after that time it did not change
appreciably.
(2) 0.1 N sodium thiosulfate, made from 25.7 g. of Na2S2O3.5H2O, 2
g. ammonium carbonate, 1,000 g. of water. This solution was also allowed
to stand for one week before standardization, and after that time it did
not change appreciably.
As the iodine solution was very volatil, weight burets, with ground
glass stoppers, were used in all titrations with both solutions. With the
aid of a weight buret, a weighed quantity of thiosulfate solution was
titrated against a measured volume of a standard potassium permanganate
solution by means of acidified potassium iodide and starch indicator.
The potassium permanganate solution was standardized against iron wire
of known composition. In this way the value of one gram of the
sodium thiosulfate solution was fixed in terms of cubic centimeters of
o. I normal solution. For the purpose of this investigation, then, a o.1
N solution of the standard solutions used was considered as one contain-
ing o. I gram-equivalent per 1,000 grams of solution.
In the assay of hydrazobenzene, first a known quantity of the standard
iodine solution, amounting to about 10 grams excess over the theoretically
required quantity, was run into a 500 cc. glass-stoppered bottle. To this
was added an amount of o.1 N alcoholic aniline solution, slightly more
than sufficient to neutralize the hydriodic acid which would be liberated
in the subsequent oxidation. The weighed sample of hydrazobenzene was
then dissolved in the least necessary amount of absolute alcohol and the
solution was added at once, with rinsing, to the alcoholic iodine and ani-
line solution. After standing for ten minutes the solution was acidified
with 10 cc. of 6 N sulfuric acid and the standard thiosulfate solution was
added from a weight buret until, judged by the color, the excess of
iodine was nearly consumed. 400 cc. of water was then added, which
15
precipitated out the yellow azobenzene in fine yellow crystals and formed
nearly 500 cc. of a solution entirely opaque, and colored bright yellow by
the suspended azobenzene. To this were then added 10 cc. of starch in-
dicator solution, which produced a greenish black coloration. Then more
thiosulfate was added until the bright yellow color of the solution was
just restored. Duplicate experiments agreed within 0.3% of the mean
value.
Wt. sample.
Gram.
0.2004
Grams
0.1 N I.
Per cent.
found.
99.7
100.3
1..
2.
0.2002
21.71
21.82
It was found that no loss of accuracy was occasioned by the presence
of the suspended azobenzene, although some of the iodine was temporarily
retained by the suspended solid, which caused an apparent end point to
come just before the true end point was reached. If the solution was
allowed to stand for one minute after the starch color first cleared, the
color darkened again, and one or two drops of thiosulfate solution per-
manently cleared the color (until it discolored from oxidation from the
air). Separate trial experiments showed that the aniline had no reducing
effect on the iodine unless allowed to stand with it for some time in rather
great concentration, when, owing to the alkalinity, the alcohol present
was attacked slowly. Independent experiments showed that the solu-
tion must be ultimately diluted so that the alcohol content was under
10%, in order to get a sharp color change with the starch indicator.
If no aniline was added and the hydriodic acid formed in the oxida-
tion was allowed to remain in the solution, a slight loss of hydrazobenzene
was occasioned, since the acid caused a portion of the hydrazobenzene
to undergo the benzidine rearrangement before the oxidation was com-
pleted. This loss is somewhat variable, depending on (1) the quantity
of hydrazobenzene oxidized, (2) the volume of the solution, and (3) the
excess of iodine used. However, under the conditions of the analyses,
as made, the loss does not exceed 1%. In the calculation of the velocity
constant from analytical data, obtained in this way, this loss may be en-
tirely neglected, since the decompositions show itself to be one of the first
order. For it may be assumed in all cases to amount to a constant per
cent. of the hydrazobenzene present, the correction factor will cancel out
of the expression for the velocity constant (see equation (7)).
Preparation for a Velocity Determination
The alcohol used as solvent for the substance was prepared by distilla-
tion of the commercial “absolute" alcohol from sodium, the middle third
only being used. For making the 50% alcohol used as wash-liquid, the
commercial “absolute" alcohol was used directly. In all cases it was
tested for reducing power toward iodine, and was found to have none.
16
Since the temperature at which the decomposition was allowed to take
place lies above the boiling point of the solvent, sealed tubes were used in
all determinations, after the manner described by Rosanoff.¹ To pre-
pare for a velocity determination, a weighed quantity of hydrazobenzene
was dissolved in a known volume of absolute alcohol. With the aid of a
5 cc. pipet, equal portions of this solution were transferred to a number
of clean tubes. The tubes were then immersed in a freezing mixture
and the constrictions sealed off at a pressure of 10-20 mm. The low tem-
perature prevented the loss of any appreciable quantity of the alcohol,
as was determined by weighing a large number of the tubes before and
after sealing. The portions of the solution thus prepared were always
colored with azobenzene, and, on standing, colored still further from the
oxygen dissolved in the alcohol. However, since the solution, before
taking the samples, was homogeneous, this slight loss of hydrazobenzene
was perfectly uniform. Furthermore, since the reaction proved to be
monomolecular beyond any doubt, the initial concentration is of no
significance, and separate experiments showed that azobenzene in large
excess does not affect the velocity of the reaction in any way.
The Thermostat
2
The thermostat used in this investigation consisted of a jacketed, seven-
gallon iron can filled with a non-volatil oil. It was heated electrically
with a primary circuit sufficient to maintain the temperature at a point.
about ten degrees below the desired temperature. A secondary circuit,
controlled by a mercury-in-glass electric thermo-regulator, operated an
electric lamp immersed in the oil, and gave the exact temperature regula-
tion. Stirring was furnished by an electrically driven rotary screw stirrer,
which kept the liquid of the bath in rapid motion. Under the conditions
of the experiment, the maximum variation was within 0.02° of a given
temperature. All points of the bath registered a temperature within 0.02°
of the mean. In setting the regulator, the accurate measure of the tem-
perature was obtained from the P. T. R. thermometer, No. 39,519, which
registers in degrees centigrade, hydrogen scale, accurate to 0.05°. Suit-
able correction was made for the slight length of exposed mercury column.
Determination of the Velocity of Decomposition
In the determination of the velocity of decomposition, the tubes prepared
as above described, were each wound with copper gauze (as a precaution
against explosion) and fastened to a metal deflagrating spoon, which
made a very suitable holder. The whole number of tubes was then heated
to 70-80° in a small oil bath and immersed simultaneously in the thermo-
1 Rosanoff, J. Am. Chem. Soc., 30, 1899 (1908).
2 Cottonseed oil was used at first, but with continued use it oxidized so badly
that it was replaced by a heavy mineral oil for which we are very much indebted to
Dr. Humphreys of the Standard Oil Co., Whiting, Indiana.
17
stat and the exact time recorded as "zero time." After a definit interval
(usually one hour) a tube was withdrawn and the time accurately noted
as t₁. On removal, each tube was first plunged into oil heated to 70-80°,
and then, after thirty seconds, into cold water. At regular intervals, tubes
were withdrawn until the reaction was about two-thirds completed. As
soon as convenient, after removal, the tubes were opened and the contents
analyzed for hydrazobenzene, exactly as described above for the assay of
the pure sample, except that no aniline solution was used. As the reac-
tion proceeded, the reaction mixture became more and more alkalin,
due to aniline formation. It was found necessary to add an amount of
0.1 N hydrochloric acid to the titration mixture to compensate for the
aniline formed, for in the alkalin, or very faintly acid solution, a dark
coloration appeared which was not noticed either in the analysis of the
pure substance or in a faintly acid solution. This coloration must have
been caused by the action of iodine on some substance other than the
hydrazobenzene or the normal products of the decomposition (see p. 9).
Proceeding in the manner just described, an examination was made
of the rate of decomposition of hydrazobenzene in alcoholic solution of
initial concentration roundly o.2, 0.1 and 0.05 molar, respectively. In
the first two concentrations, the amount of solution used was that deliv-
ered by a 5 cc. D. R. pipet, on draining fifteen seconds. In the 0.05.
molar concentration the sample put into each tube was that delivered.
by a 10 cc. D. R. pipet on draining fifteen seconds. On the duplicate
determination of this concentration, a serious explosion of two of the tubes.
wrecked the apparatus. It is probable that the cause of this accident.
was the leaving of insufficient vapor space over the liquid in the tube.
However, as the explosion was so violent, in all cases in the subsequent
work the sample put into each tube was that delivered by a 5 cc.
pipet. In the following tabulated results, apparent discrepancies in corre-
sponding tubes of supposedly the same initial concentration are due to
slightly different conditions in making up the solutions, which caused
different amounts of oxidation of the hydrazobenzene before the tubes were
sealed (see p. 16), and to irregularities in the heating between the times,
to and t₁, occasioned by the plunging of the cold tubes into the hot thermo-
stat. However, always, all of the tubes of one series were treated in an
identical manner until successively removed from the bath.
From the quantities of hydrazobenzene remaining after the different
time intervals a calculation of the velocity constant was made according
to the law for a monomolecular reaction:
1
A-X1
Kmono.
log. nat.
t₂ — tz
A
'x2'
X2
(7a)
In these and subsequent calculations the grams of iodine solution con-
sumed for a definit volume of the solution (given under I in the tables)
18
were used in the logarithmic term in place of the concentrations (A —x1)
and (A-x2), since the concentrations are directly proportional to these
titration results.
For the purpose of comparison a calculation was also made according
to the law for a dimolecular reaction:
Kdimol.
I
X2
X₂-X1
tz - tx (A — x½) (A — xx)'
(4α)
Since the latter formula is so patently inapplicable to this reaction, it
was not thought necessary to make the minor correction for the expansion
of alcohol between 15° and 140°.
In the following tables (I-III) is stated, first, the weight of hydrazo-
benzene used; then under V the total volume of its solution in absolute
alcohol at 15°; under v the volume of the solution that was put into each
tube; and finally the temperature. The time t in minutes is given in the
first column under T; the amount of iodine, expressed in grams of o. 1 N
solution,¹ which was consumed by the unchanged hydrazobenzene, is
given in the second column under I, and the velocity "constants," Kmono.
and Kdimol., respectively, for a monomolecular and a dimolecular reac-
tion, are given in the third and fourth columns. In the fifth and sixth
columns (Tables I and II) are given the results of duplicate experiments
carried out under the same conditions, with the use of the same time in-
tervals; and in the last column is given an average of the values of
Kmono..
2
The data of a later similar series of experiments (see Table IV, further
on), carried out at a slightly lower temperature, are analyzed to determin
whether a case of parallel monomolecular and dimolecular reactions is in-
dicated (see p. 20).
TABLE I
= 50 cc., v = 5 cc. Temp. 140.55° ± 0.05°.
2 grams hydrazobenzene. V
A.
B.
T.
1.
105 Kmono.
104 Kdimol.
I.
105 Kmono. Av. 105. Kmono.
60
19.67
19.17
•
•
180
16.35
154
86
15.81
160
157
240
14.83
157
92
14.34
161
159
300
13.35
161
100
13.12
158
159
420
II.24
156
106
10.91
156
156
540
9.42
153
115
9.18
153
153
720
7.34
149
129
7.18
149
149
900
5.93
143
140
5.71
144
144
154
1 See the exact definition given on p. 14.
2 In the preliminary report, loc. cit., the temperature at which this series was run
was given as 140.2° ± 0.05°. A recalculation of the correction gave this temperature
as 140.55° ± 0.05°.
19
TABLE II
= 5 cc. Temp. = 140.55° ± 0.05°.
I gram hydrazobenzene. V = 50 cc., v
A.
B.
T.
I.
105 Kmono.
I.
105 Kmono.
Av. 105 Kmono.
60
9.45
9.51
180
7.78
162
7.83
162
162
240
7.04
163
6.94
(175)
163
300
6.44
160
6.49
159
159
420
5.31
160
5.37
159
159
540
4.47
156
4.51
155
155
720
3.52
150
3.52
151
150
900
3.14
(131)
2.80
145
145
Av. 156
TABLE III
I gram hydrazobenzene. V
=
100 CC., V = IO CC.
Temp.
T.
I.
105 Kmono.
140.55° ± 0.05°.
104 Kdimol.
60
9.32
•
•
180
7.71
158
370
240
7. ΟΙ
158
400
300
6.39
157
410
420
5.31
156
450
540
4.51
151
480
720
3.50
148
540
900
2.76
145
610
Av. 153
The Effect of Temperature on the Velocity of Decomposition
Proceeding in a manner similar to that described, series of velocity de-
terminations were run at each of three temperatures, differing by 5.00° ±
0.05°, in order to investigate the temperature coefficient of the action.
Each series consisted of determinations made with different initial concen-
trations, namely approximately 0.2, 0.1 and 0.05 molar. At each
temperature the regulator was set by comparison with the P. T. R. ther- •
mometer, and a Beckmann thermometer was set so that the slightest
variation of temperature could be noticed, and corrected for. In all cases
the temperature given is the corrected temperature. In the following
tables (IV-VI), each table consists of the results of three separate velocity
determinations, each carried out with a hydrazobenzene solution of differ-
ent concentration. In each case experiment A was carried out with a
0.2 molar solution of hydrazobenzene, experiment B with a 0.1, and C
with a 0.5 molar solution [2, 1 and 0.50 grams of hydrazobenzene
in 50 cc. solution, respectively (V 50 cc.)]. In all cases each tube was
charged with 5 cc. of solution (v = 5 cc.) The data given and the
symbols used have the same significance as in Tables I-III. The calcu-
lation of Kdimol, has been omitted.
20
Temp.
Experiment A.
TABLE IV
140.35° ± 0.05°.
Experiment B.
Experiment C.
T.
I.
105 Kmono.
I.
105 Kmono.
I.
105 Kmono. Av. 105 Kmono.
60
19.15
9.47
4.68
19.17
9.47
4.66
180
15.97
151
7.94
147
3.87
156
151
240
14.68
148
7.24
149
3.55
152
150
300
13.51
146
6.65
147
147
420
11.36
145
5.59
146
2.71
151
147
540
9.69
142
4.76
143
2.33
145
143
720
7.57
141
3.70
142
1.83
142
142
900
6.06
137
2.98
138
I.45 139
138
Av. 145
Table V
Temp.
=
Experiment A.
T.
I.
105 Kmono.
I.
105 Kmono.
145.35° ± 0.05°.
Experiment B.
Experiment C.
I. 105 Kmono. Av. 105 Kmono.
19.44
9.60
4.71
30
19.39
9.60
4.73
90
16.69
252
8.26
251
4.06
251
251
120
15.37
259
7.61
258
3.74
258
258
150
14.34
253
7.01
262
3.45
261
259
210
12.34
252
6.07
255
2.96 259
255
270
10.73
247
5.20
255
2.52
261
253
360
8.66
245
4.12
256
2.05
252
251
450
7.11
239
3.32
253
1.69
244
245
Av. 253
TABLE VI
Temp.
-
Experiment A.
150.35° ± 0.05°.
Experiment B.
Experiment C.
T.
I.
105 Kmono.
I.
105 Kmono.
I. 105 Kmono. Av. 105 Kmono.
17.12
8.52
4.49
45
17.09
8.54
85
14.30
448
7.11
454
3.75
450
451
105
13.14
440
6.50
453
3.39 468
454
125
12.03
440
6.06 (426)
3.13 45I
446
165
10.19
432
5.04
438
2.62
449
440
205
8.66
425
4.27
432
2.21 443
433
265
6.89
413
3.41
417
1.79 418
416
325
5.93 (378)
2.75
404
1.39 419
412
Av. 436
Analysis of Data for Parallel Monomolecular and Dimolecular Reactions
For a parallel monomolecular and dimolecular reaction (see p. 7)
we have:
21
dx
dt
K₁ (A − x) + K₂ (A — x)².
(13)
Integrating, we find:
K₁ (t₂ — t、)
In
—
(A − x,)(K₁/K₂ + A − x₂)
(A — x2) (K₂/K₂ + A − x、)'
·
—
(13a)
2
By using the analytical data for two equal time intervals (t₂-t₁) we
obtain simultaneous equations from which the value of K₁/K2 is found.
From the data for the time intervals 60 to 300 minutes and 300 to 450
minutes in experiment A, of Table IV, we find K₁/K₂ = 0.8 and K₁ =
0.00121, K2 = 0.00151.
For the critical consideration of the data of
experiments A, B and C, where the concentrations of hydrazobenzene were
¹/½ and ¹/4 as great, respectively, as in A, we may properly assume that the
velocity constant of the monomolecular reaction must remain the same as
in A and calculate the value of the dimolecular constant, in order to see
whether there is really a parallel dimolecular reaction. Using the value
K₁ = 0.00121, accordingly, in analyzing the data for the time interval
60 to 300 minutes in experiment B and the interval 60 to 240 minutes in
experiment C, we obtain the values for K2 given in the first line of the fol-
lowing Table, IVa. If we use the data of each table for the calculation
of the values of K₁ and K2, the ratio K₁/K₂ = 0.8 being assumed to
hold in each of the three series, then the values for K₂ as given in
the second line of the table are found. The earliest time intervals were
selected because a pure case of a parallel monomolecular and dimolecular
reaction must show itself at once in the results. The later measurements
could have in them a retarding influence of some accumulating product
of the main action or of a secondary action (see p. 9).
105 K2...
105 K2...
TABLE IVa
A.
151
151
B.
C.
331
799
168
181
It is clear that no parallel dimolecular reaction is indicated. The
change in the values for K₂ shown in line 1, where the same constant for
the monomolecular reaction was used for the three experiments, A, B and
C, must be given most weight. In judging the results given in line 2,
one must recall that with the increase in the dimolecular constant K2,
the monomolecular constant K₁ also rapidly increases, these calculations
assuming that their ratio remains the same. We may safely conclude,
therefore, that the thermal decomposition of hydrazobenzene is not a case
of a parallel monomolecular and dimolecular reaction, but represents a
monomolecular reaction, with some other secondary reaction, causing
some retardation of the action (see p. 9).
22
Determination of Velocity by Analysis for Aniline Formed
As a check on the preceding work, an effort was made to analyze the
contents of the tubes for aniline in a velocity determination. By a modi-
fication of the well-known bromination analysis for aniline, a suitable
method was worked out, which gave very consistent results. The reac-
tion in this case runs quantitatively according to the equation:
C6H5NH2 + 3Br2 → BrзC6H2NH2 + 3HBr.
The following aqueous solutions were used:
(1) 10% potassium iodide; (2) 10% potassium bromide; (3) 6 N sul-
furic acid; (4) 10 N sodium hydroxide; (5) standard o.1 N sodium thio-
sulfate (described before); (6) standard o. 1 N potassium bromate. Also
o. 1 N iodine solution in 50% alcohol (described before).
The potassium bromate solution was standardized against the standard
sodium thiosulfate solution with acidified potassium iodide and starch
indicator.
In performing the analysis in the velocity determinations, the sample
was oxidized exactly as before with an excess of iodine solution to remove
any reducing substance present (e. g., hydrazobenzene). After 10 minutes,
10 cc. of 6 N sulfuric acid was added, and steam was passed through the
solution (in a distilling flask) until all of the azobenzene, excess iodine,
alcohol and other possible non-basic volatil substances were removed,
10 cc. of 10 N sodium hydroxide were then added, and the volatile bases
distilled with steam, in an ammonia distillation apparatus, into dilute
sulfuric acid, until 400 cc. of distillate had passed over.
A flame was kept
under the distillation flask, and toward the end of the distillation the vol-
ume was such that salts began to crystallize out. To the acid distillate
were added 10 cc. of potassium bromide solution, and a definite amount of
the standard potassium bromate solution. The flask containing this solu-
tion was then closed with a stopper, holding a calcium chloride tube of
beads, wet with potassium iodide solution. The solution was then warmed
to 50° and allowed to stand for fifteen minutes. The white tribromo-
aniline precipitated out rapidly, and at the end of the time the solution
was slightly yellow with free bromine. Potassium iodide solution (10)
cc.) was then rinsed through the tube of beads, the solution was cooled,
and the end point determined with the standard thiosulfate, with starch
indicator. In this case
I g. of o. 1 N KBrO3
0.00155 g. aniline.
In preliminary trials, pure aniline being worked through the entire process,
0.0500 g. could be determined with an accuracy of 0.5%.
In the velocity determination by this method, two tubes at time, tı,
were taken, one of which was analyzed for hydrazobenzene in the usual
manner, the other of which was analyzed for the aniline formed during
23
I
1
the heating preliminary to the first measurement made at the moment t₁.
If we represent by B₁ and B, the grams of o. 1 N solution¹ of bromine con-
sumed to convert aniline into tribromaniline2 at the first measurement
at tɩ and at any subsequent time t, respectively, then we have for N₁ and
N₁, the corresponding concentrations of aniline formed, N₁ = B1/6 and
N₁ = B₁/6. We will call A₁ the equivalent concentration of hydrazo-
benzene at the moment t₁, when the first measurement was made and the
hydrazobenzene determined directly, as explained above. Calling A,
the equivalent concentration of hydrazobenzene at any subsequent
moment t, we have the relation:4
-
A₁ = A₁ − 2(N₁ — N₁) or A₁ = A,
B₁ - B₁
3
3
By the insertion of the values for A, thus obtained from the experi-
mental data, into the logarithmic term of the equation for a monomolec-
ular reaction, a series of velocity constants was computed similar to those
found in the previous work at 140°. Duplicate determinations were made
with initial concentrations of hydrazobenzene approximately 0.2 normal
(0.1 molar).
In the following table (VII) is stated first the weight of hydrazobenzene
TABLE VII
1 gram hydrazobenzene. V = 50 cc., v
I
=
+
= 5 cc. Temp. 140.35° 0.05 °.
Second experiment.
At. 105 Kmono. Av. 105 Kmono.
First experiment.
T.
B.
60
2.13
At. 105 Kmono.
9.725
B.
•
2.03
9.795
180
6.94
8.12
149
6.94
8.16
154
151
240 8.98
7.44
149
9.07
7.45
152
150
300 10.95
6.78
150
10.94
6.82
151
150
420 14.33
5.65
151
14.35
5.68
151
151
540 16.84
4.82
146
16.94
4.82
148
147
720
19.98
3.77
144
19.79
3.87
14I
142
900 22.52
2.92
143
21.84
3.19
134
138
Av. 147
¹ See the definition of normal solution, p. 14.
2 An examination of the white crystalline precipitate obtained in these analyses
showed it to be practically pure 1,3,5-tribromaniline (m. p. 119° observed; m. p.,
Beilstein, 121°).
3 For the calculation of the constant of a monomolecular reaction molar concen-
trations need not be used and equivalent concentrations were used therefore in this
calculation as the analytical data presented these directly. Two equivalents of hydrazo-
benzene = one mole.
4 A molecule of aniline is formed from one molecule or two equivalents of hydrazo-
benzene. The aniline formed decreases the equivalent concentration of hydrazobenzene
therefore by twice the number of equivalents.
5 These values, A1, were determined by direct titration with iodine, the remainder
of the values in the columns At calculated as indicated in the text.
24
used; then under V the total volume of its solution in absolute alcohol at
15°; and under v the volume of solution that was put into each tube.
The time t is given in minutes in the first column under T; the grams of
0.1 N bromine solution consumed by the aniline formed is given in the
second column under B; the equivalent concentration of hydrazobenzene,
calculated as indicated in the above equation, is given in the third column
under A,; in the fourth column is the velocity constant, Kmono., for a mono-
molecular reaction; in the fifth, sixth and seventh columns are the results
of a duplicate determination carried out under the same conditions, and
with the same time intervals; and in the last column is the average value
of the velocity constant Kmono..
Experiments to Determin the Cause of the Falling Off of the Constant
In order to investigate the unexpected retardation of the reaction, as
it nears completion (see theoretical part), velocity determinations were run
with hydrazobenzene solutions, to which had been added amounts of ani-
line and azobenzene. It is noticed that there is no effect of these sub-
stances on the velocity, within the limits of observational error.
Three solutions of initial concentrations of hydrazobenzene, approxi-
mately 0.1 molar, were made. To two of these solutions was added suffi-
cient aniline to make them 0.05 and 0.1 molar, with respect to aniline.
TABLE VIII
1 gram hydrazobenzene. V = 50 cc., v
A (0.233 g. aniline).
=
5 cc. Temp. 140.35° ± 0.05°.
B (0.466 g. aniline).
105 Kmono. Av. 105 Kmono.
T.
I.
105 Kmono.
I.
60
S 9.67
9.69
9.67
9.71
180
8.02
156
8.04
156
156
240
7.37
151
7.35
154
152
300
6.73
151
6.72
153
152
420
5.71
146
146
540
4.94
140
140
720
3.84
140
140
Av. 148
TABLE IX
V = 50 cc., v = 5 cc.
0.450 gram azobenzene.
5 cc. Temp.
=
140.35° ± 0.05°.
105 Kmono.
I gram hydrazobenzene.
T.
I.
60
$ 9.90
9.88
180
8.26
150
240
7.51
153
300
6.91
149
Av. 151
25
To the third solution was added sufficient azobenzene to make it 0.05
molar with respect to that substance. From each of these solutions a
series of tubes was filled. The course of the reaction was followed in the
usual manner by analyses of the contents of the tubes after known time
intervals. In the preceding tables (VIII-IX) the data given and the sym-
bols have the same significance as described for Tables I-III. In addi-
tion, the amount of aniline (or azobenzene, as the case may be) added is
stated. The calculation of Kdimol, has been omitted:
In order to investigate the effect of the solvent as a possible cause of the
disturbing reaction, a duplicate series of velocity determinations was run
with hydrazobenzene, dissolved in pure thiophene-free benzene, dried over
sodium. The analytical method was the same as usual, for the benzene
caused no serious interference with the analysis. However, the end point
with this method was not as sharp as when alcohol was used. Just as was
always noticed before, toward the end of the reaction a coloration of the
titration mixture occurred, when an excess of acid was not present, thereby
indicating the presence of an, as yet, unidentified product of the reaction.
In the benzene medium, the velocity of decomposition is much slower,
as it requires a temperature of 155° to give approximately the same velocity
that obtains in an alcohol medium at 140°.
Aside from the effect on the speed of decomposition, the benzene medium
seems to have no further influence on the nature of the reaction, for the
same retardation of the reaction occurs as it nears completion.
TABLE X
I gram hydrazobenzene.
V = 50 cc., v
I.
= 5 cc. Temp.
155.35° ± 0.05 °.
T.
A.
B.
Av. 105 Kmono.
60
9.82
9.68
180
8.29
8.23
138
240
7.57
7.62
139
300
7.01
7.05
136
420
6.08
5.89
135
540
5.18
5.18
131
720
4.II
4. II
127
900
3.43
3.40
125
Av. 133
To investigate further into the nature of the apparent decrease in velocity
of the reaction as it nears completion, a search was made for benzidine
among the products of the reaction, on the theory that water, carbon di-
oxide, or some such faintly acidic substance might cause such a rearrange-
ment (see footnote, p. 7). To make this test, a portion of the solution
was heated at 140° for several hours, and the excess of hydrazobenzene
removed by oxidation with iodine in presence of an excess of aniline. The
alcohol, azobenzene and iodine were removed by steam distillation in acid
26
solution, and the volume reduced to about 15 cc. One gram of solid am-
monium sulfate was added, and the solution allowed to stand. A slight,
flocculent, dark-colored precipitate was noticed; it was filtered out and
treated with chlorine water, but gave no color change. As little as 0.0005
g. of benzidine gave a considerable precipitate under these conditions,
and on treatment with chlorine water gave a distinct red coloration to
the liquid. The experiment was repeated four times with the heated
hydrazobenzene solution, but in no case was any trace of benzidine noticed.
To determin whether hydrazobenzene can be formed reversibly from
aniline and azobenzene (p. 8), a series of tubes, containing each a por-
tion of an alcoholic solution, 0.2 molar with respect to aniline and o. I
molar in respect to azobenzene was heated at 140° for various intervals of
time, and on removal from the thermostat was analyzed as usual for
hydrazobenzene. In no case could any absorption of iodine be noticed
above a minimal amount consumed, owing to the standing of the alcoholic
iodine solution in the presence of the not inconsiderable amount of aniline.
This absorption, moreover, was as great in a tube that had not been heated
as in one that had been heated for twenty-four hours.
II. Decomposition of p-Methyl Hydrazobenzene
1
The p-methylhydrazobenzene used in this work was prepared by the
reduction of p-methylazobenzene with zinc dust and sodium hydroxide
in alcoholic solution. The p-methyl azobenzene was prepared according
to the method of Mills² from nitrosobenzene and p-toluidine. The crude
p-methylhydrazobenzene, after two recrystallizations from 95% alcohol,
gave a product, perfectly colorless, having a constant melting point of
91°.
It was dried and preserved in sealed bulbs, as was the hydrazo-
benzene. The pure product was assayed as was the hydrazobenzene,
and showed a reducing power very closely approximating that demanded by
theory:
I
2
•
Wt. sample.
Gram.
0.2007
•
0.2014
0.1 N I.
20.26
20.22
Per cent.
found.
100.0
99.6
It was noticed that this substance is much more sensitive to acid than
is the hydrazobenzene. When 0.2 g. of it is oxidized with iodine in a vol-
ume of about 50 cc. the acid formed in the oxidation causes 3 to 4% of
the substance to undergo the semidine (or similar) rearrangement before
the oxidation is completed. It is necessary then to add aniline to the
sample being analyzed, which prevents this loss entirely, if a slight excess
is used.
However, in the velocity determination it was found that the titra-
tions could not be carried out in alkalin solution, for also in the decom-
¹ Jacobson and Lischke, Ann., 303, 369 (1898).
2 J. Chem. Soc., 67, 929 (1895).
27
position of p-methylhydrazobenzene some disturbing reaction goes on,
which forms a compound, giving a black coloration with iodine in alkalin
solution. To avoid, as far as possible, both difficulties, an amount of
0.1 N aniline solution, in absolute alcohol, was added to the titration
mixture, which would be insufficient by 2.0 cc. to neutralize the hydriodic
acid formed in the oxidation. The volume of the titration mixture being
considered constant, this enables the titrations to be carried on under a
condition of constant minimal acidity. In this way, with the pure sub-
stance, consistently 98.5% of the p-methylhydrazobenzene could be ac-
counted for. If then in all analyses 98.5% of the undecomposed hydrazo
compound is found, the correction factor will cancel out of the expression
for the monomolecular velocity constant, and so the numerical value of
the constant is not changed. That the reaction is monomolecular cannot
be doubted, for this slight correction could not possibly change the ap-
parent order of the reaction.
From a preliminary experiment, an approximate value of the velocity
constant was obtained, and by using this value the proportion of the
hydrazo compound and of aniline present in a tube at any time, t, could be
calculated. Thus from a consideration of the amount of hydriodic acid
which would be formed, and the aniline (or p-toluidine) which already
was present in the solution, the proper amount of o. 1 N aniline solution
was added to the titration mixture, and the analysis performed as usual.
In the manner described for hydrazobenzene, with the above mentioned
modifications, an examination was made of the rate of decomposition of
p-methylhydrazobenzene in alcoholic solutions for initial concentration ap-
proximately 0.2, 0.1 and 0.05 molar, respectively. From the quantities
of the hydrazo compound left after the various time intervals the calcula-
tion of the velocity constant¹ was made as before.
2.10 grams p-methylhydrazobenzene.
TABLE XI
V =
50 cc., v = 5 cc. Temp. 125.30° ± 0.05°.
B.
105 Kmono. Av. 105 K mono.
A.
T.
I.
105 Kmono.
I.
60
18.77
18.83
180
15.71
148
15.73
150
149
240
14.40
147
14.46
147
147
300
13.18
147
13.12
150
148
420
II. 13
145
11.18
145
145
540
9.44
143
9.50
143
143
720
7.60
137
7.59
138
138
900
6.18
132
6.21
132-
132
Av. 143
¹ We have here two parallel decompositions of methyl-hydrazobenzene, aniline be-
ing formed in the one, toluidine in the other. The velocity equations show that
Kmono., as calculated, represents the sum of the corresponding velocity coefficients
Kanil, and Ktol..
28
}
TABLE XII
1.05 grams p-methylhydrazobenzene. V = 50 cc., v = 5 cc.
5 cc.
Temp.
=
Temp. 125.30 ± 0.05°. :
A.
B.
T.
I.
105 Kmono.
I.
105 Kmono.
Av. 105 Kmono.
60
9.27
9.22
180
7.66
159
7.71
148
153
240
7. ΟΙ
155
6.96
156
155
300
6.45
151
6.49
146
148
420
5.40
150
5.41
148
149
540
4.59
147
4.60
145
146
720
3.70
139
3.69
141
140
900
3.00
134
2.97
135
134
TABLE XIII
0.53 gram p-methylhydrazobenzene. V = 50 cc., v = 5 cc. Temp.
Av. 146
=
125.30° ± 0.05.°
A.
B.
T.
1.
105 Kmono.
I.
105 Kmono. Av. 105 Kmono.
60
4.54
4.5I
180
3.69
(173)
3.74
156
156
240
3.41
159
3.42
154
156
300
3.12
156
3.16
148
152
420
2.65
149
2.64
149
149
540
2.25
146
2.23
147
146
720
.1.79
141
1.81
138
139
900
1.43
138
1.43
134
136
Av. 148
An attempt was made to prepare the theoretically possible electromer
of p-methylhydrazobenzene (see p. 12) by reduction of the azo compound
prepared from nitrosotoluene and aniline by the Mills' synthesis. The
product thus obtained was then closely. compared with the p-methyl
TABLE XIV
1.05 grams p-methylhydrazobenzene.2 V = 50 cc., v = = 5 cc. Temp. 125.30° ± 0.05.°
105 Kmono.
T.
I.
60
9.24
180
7.64
158
240
6.97
157
300
6.37
155
420
5.36
151
540
4.59
146
720
3.65
141
900
3.01
133
Av. 149
1 Loc. cit.
2 The sample prepared from p-nitrosotoluene and aniline was used in this determina-
tion-designated as the ß-form.
29
hydrazobenzene, previously described. For the sake of convenience, these
two forms may be designated as "a" and "B," respectively, in the order
of preparation, although no difference was found between them. The
melting points of the two forms are the same, 91°,¹ and a melting point
of a mixture of the two shows no depression. A velocity determination
was also run with the "B" form, which shows that the rate of decomposi-
tion of the two forms is the same. The preparation of tubes and the method
of analysis were the same as described before.
Decomposition Products of p-Methyl Hydrazobenzene
A qualitative test was first made for the two aromatic amines, which
theoretically could be formed from the breaking down of the unsym-
metrical p-methylhydrazobenzene molecule. A portion of the reaction
mixture was oxidized with an excess of 0.1 N iodine solution, the solu-
tion acidified, and distilled with steam to remove iodine, alcohol and azo
bodies. The residue was made alkalin and again distilled with steam,
and the distillate tested for (1) aniline and (2) p-toluidine. Aniline²
was identified by the formation of a purple coloration, when calcium
hypochlorite solution was added to some of the distillate, made faintly
alkalin with sodium hydroxide. p-Toluidine³ was identified by the forma-
tion of a red coloration, when a portion of the distillate was boiled with a
few drops of ferric chloride in dilute hydrochloric acid solution.
By a modification of the method used to determin aniline (see p. 22),
it was found possible to obtain an indirect analysis for the proportion of
aniline and p-toluidine formed. It was found that in warm solutions
bromine destroys p-toluidine, with consumption of an indefinit amount
of the former. However, at o° bromine reacts quantitatively with p-tolui-
dine forming 1,5-brom-3-methylaniline according to the equation:
CH3C6H4NH₂ + 2Br₂ →→ CH3C6H₂Br₂NH2 + 2HBr.
From pure samples of p-toluidine a product with the correct melting point,
73°, is precipitated, and the proper amount of bromine is consumed, when
the reaction mixture is kept cooled to oº.
Since potassium bromate will not liberate bromine freely from acidu-
lated bromide solutions at o°, it was necessary to use an aqueous bromine
solution. This solution was made up as follows: 8.5 g. bromine, 50 g.
potassium bromide, 1000 g. water. It was standardized against o.1 N
thiosulfate solution with acid potassium iodide and starch indicator.
When a glass-stoppered weight buret was used in the titration, no great
loss from volatilization of the bromine was noticed. The other solutions
1 The melting point given by Jacobson and Lischke, Ann., 303, 369 (1898), is
86-87°.
2 Compt. rend., 67, 398 (1868).
³ Biehringer and Busch, Chem. Ztg., 26, 1128 (1902).
30
t
used were the same as described in the analysis for aniline. In the analysis
of a sample of the velocity mixture, the unchanged hydrazo compound
was first oxidized with iodine. However, to prevent the formation of cer-
tain amounts of semidine, a o. 1 N ammonium carbonate solution was used
to destroy the acid formed in the oxidation, just as aniline had been used
before. This was found to have the same effect and introduced no sub-
stance which would absorb bromine on the later stage of the analysis.
After the oxidation the alcohol, iodine and azo compounds were removed
by distillation in acid solution, and then the two amines were distilled into
an acid solution. The acid distillate was then cooled in a freezing mix-
ture until ice separated out; a known quantity of standard bromine water
was then added and the mixture allowed to stand for ten minutes. Po-
tassium iodide solution (10 cc.) was then added and the solution was
titrated with 0.1 N sodium thiosulfate solution and starch indicator.
Alone, as little as 0.05 g. of aniline or p-toluidine could be estimated by
this method within 1.0%. With a known mixed sample, however, the
accuracy was not greater than 5%.
To secure the necessary data, a number of tubes, each containing the
same amount of the p-methylhydrazobenzene solution, was placed in
the bath simultaneously. After one hour two tubes were withdrawn, one
of which was analyzed (as described above) for bromine-absorbing power,
the other of which was analyzed for p-methylhydrazobenzene, with ex-
cess of anline present. After a suitable interval, two other tubes were
withdrawn, and again analyzed for bromine-absorbing power. If we use
the known velocity constant, Kmono., the amount of p-methylhydrazoben-
zene that has suffered decomposition at this time can be calculated.
We will now let A1 and A2 represent the equivalents of hydrazine present
at any two times, t₁ and të, in a tube-the amount decomposed in the time
interval is then (A1-A2). From equations 18, 19 and 20 (p. 12) it is seen
that one molecule or two equivalents of methylhydrazobenzene produce one
molecule of the aromatic amine (i. e., either aniline or paratoluidine, as
the case may be). (A1-A2) equivalents of the hydrazo compound will
therefore produce (A1-A2)/2 molecules of amine. A molecule of aniline
consuming six equivalents of bromine, if all the amine produced were ani-
line, the total bromine consumed would be 3 (A1-A2) equivalents; and, a
molecule of toluidine consuming only four equivalents of bromine, if only
toluidine were formed, 2(A₁— A2) equivalents of bromine would be con-
sumed. The total possible excess of bromine used over this last amount
or (A₁— A2) equivalents of bromine is therefore proportional to 100% of
aniline and the actual excess of bromine used beyond what would be needed
for toluidine is a measure of the proportion of aniline found in the titra-
tion of the mixed amine. If we call B1 and B2 the grams of o. 1 N bromine
solution (equivalents of bromine) used in the titration of two samples
3I
taken at times t₁ and t₂, then (B2 -B₁) is the bromine actually consumed
for the amine obtained from (A₁-A2) equivalents of hydrazobenzene,
and we have:
(B₂ — B,) —2(A, — A2)
Per cent. aniline in the amine
A1
A₁- A₂
In
Per cent. toluidine in the amine 100%-% aniline found.
One series each of the so-called "a" and "B" forms of p-methyl hydrazo-
benzene was examined for p-toluidine and aniline in this manner.
both cases the initial concentration of the solution was approximately
0.2 molar. The whole number of tubes was heated one hour before the
first pair was withdrawn, and analysis was made for bromine-absorbing
power and for unchanged p-methylhydrazobenzene. After being heated
for four hours further, two more tubes were withdrawn, and a duplicate
analysis for bromine-absorbing power was made.
In the following summary of results is given in the first column, under
A₁, the grams of o. 1 N iodine solution consumed by the p-methylhydrazo-
benzene present at the time t₁; in the second column, under B₁, is given
the grams of o.1 N bromine solution consumed by the combined amines
present at time t₁; in the third column under (A1-A2) is given, in equiva-
lents, the calculated amount of p-methylhydrazobenzene which should
have decomposed by the time t2, when the final analyses for "bromine
consumed" were made; in the fourth column under B₂ is given the weight
of 0.1 N bromine solution consumed at the time t2; and in the fifth and
sixth columns are given the percentages of aniline and of p-toluidine,
respectively, which were calculated from the above formula:
2
A1.
B1. (A1-A2). B2.
Per cent. Per cent.
anil.
Tol.
I.
19.33
4.55
5.78
S 18.05
33.5
66.5
17.96
32.0
68.0
S 17.96
33.5
66.5
2.
19.17
4.58
5.73
18.06
35.0
65.0
Series (1) was carried out with the sample of p-methylhydrazobenzene
designated as a, and series (2) with that designated as B. It is seen here
again that no difference between the two samples is noticeable.
III. Decomposition of p-Hydrazotoluene
The p-hydrazotoluene used in this work was prepared by the reduction
of p-nitrotoluene¹ by the same process, by which hydrazobenzene was pre-
pared from nitrobenzene. The crude substance obtained gave, after two
recrystallizations, a product only very slightly tinged with yellow, having
a constant melting point of 134°. It was dried and preserved in sealed bulbs
and assayed exactly as was the hydrazobenzene and showed a reducing
power very close to that demanded by theory:
1 Janovsky, Monatsh. Chem., 9, 829 (1888).
32
Wt. sample.
Gram.
0.1985
0.1982
0.1 N I.
18.64
18.70
Per cent.
found.
99.6
100. I
This substance is much more sensitive to acid than either of the other
two hydrazo compounds studied. When 0.2 g. of it is oxidized with iodine
in a volume of about 50 cc., the acid formed in the oxidation causes 8-10%
of the substance to undergo the semidine or a similar rearrangement be-
fore the oxidation is complete. Here, too, an addition of an excess of the
O.IN aniline solution is necessary and sufficient to prevent loss in the
analysis.
Fortunately, in this case, it was found that in the velocity determina-
tions the titrations could be carried out in alkalin solutions (i. e., with
slight excess of aniline), as no interfering color reaction occurred. Only
in the tubes where the reaction had run nearly to completion was a slight
pink coloration noticed. So in all analyses the excess of acid theoretically
formed in the oxidation was calculated. Then an amount of 0.1 N
aniline solution, more than sufficient, by 2.0 cc., to neutralize this acid,
was added to the titration mixture, and the titration performed as usual.
TABLE XV
1.10 grams p-hydrazotoluene. V = 50 cc., v = 5 cc.
= 5 cc. Temp. 110.25° ± 0.05°.
B.
A.
T.
I.
105 Kmono.
I.
105 Kmono. Av. 105 Kmono.
50.
9.03
9.04
150
7.58
(175)
7.57
177
(176)
200
7.08
163
7.06
165
164
250
6.42
171
6.41
172
171
350
5.49
166
5.44
169
167
450
4.74
161
4.74
161
161
600
3.93
151
3.88
154
152
750
3.26
145
3.27
145
145
Av. 162
TABLE XVI
0.55 gram p-hydrazotoluene. V
50 cc., v = 5 cc.
Temp.
110.25° ± 0.05°.
A.
B.
T.
I.
105 Kmono.
I.
105 Kmono. Av. 105 Kmono.
50
4.33
4.44
150
3.67
165
3.75
169
167
200
3.38
165
3.45
170
167
250
3.06
173
3.16
170
171
350
2.67
161
2.69
167
164
450
2.31
157
2.35
159
158
600
1.99
14I
1.98
147
144
750
1.64
139
1.58
148
143
Av. 159
33
The limited solubility of the p-hydrazotoluene in absolute alcohol pre-
vented the use of solutions more concentrated than 0.1 molar.
In the manner described for hydrazobenzene, with the above-mentioned
modifications, an examination was made of the rate of decomposition of
p-hydrazotoluene in alcoholic solutions of initial concentration approxi-
mately 0.1 and 0.05 molar, respectively.
In conclusion, the author wishes to express his most sincere thanks to
Professor Julius Stieglitz, whose fruitful suggestions and inspiring counsel
alone have made the successful completion of this investigation possible.

UNIVERSITY OF MICHIGAN
Gaylord Bros.
Makers
Syracuse, N. Y.
PAT. JAN. 21, 1908
3 9015 02624 3439
1
Medion Library