вход по аккаунту



код для вставкиСкачать
This material was produced from a microfilm copy of the original document. While
the most advanced technological means to photograph and reproduce this document
have been used, the quality is heavily dependent upon the quality of the original
The following explanation of techniques is provided to help you understand
markings or patterns which may appear on this reproduction.
1. The sign or "target" for pages apparently lacking from the document
photographed is "Missing Page(s)". If it was possible to obtain the missing
page(s) or section, they are spliced into the film along with adjacent pages.
This may have necessitated cutting thru an image and duplicating adjacent
pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it
is an indication that the photographer suspected that the copy may have
moved during exposure and thus cause a blurred image. You will find a
good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being
photographed the photographer followed a definite method in
"sectioning" the material. It is customary to begin photoing at the upper
left hand corner of a large sheet and to continue photoing from left to
right in equal sections with a small overlap. If necessary, sectioning is
continued again — beginning below the first row and continuing on until
4. The majority of users indicate that the textual content is of greatest value,
however, a somewhat higher quality reproduction could be made from
"photographs" if essential to the understanding of the dissertation. Silver
prints of "photographs" may be ordered at additional charge by writing
the Order Department, giving the catalog number, title, author and
specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as
Xerox University Microfilms
300 North Zeeb Road
Ann Arbor, Michigan 4B106
1 J
Kenkin, Hymn.
The reaction of atomic hydrogen and
Hew York, 1939.
4p.l.,62 typewritten leaves, diagr.
Thesis (ffc.D.) - New York university,
C-raduate school, 1940.
Bibliography: p.60-62.’
Shelf List
Xerox University Microfilms, Ann Arbor, Michigan 48106
N. Y. U*nv2!
Submitted in partial fulfillment of the require­
ments for the degree of Doctor of Philosophy at
New York University,
This research is an outgrowth of the work of Davis,
Burton, and Taylor on the azomethane decomposition, in which
the complexity of the reaction was demonstrated. It was felt
that a study of the azomethane-atomic hydrogen would help in
the determination of the mechanism of the azomethane decompos­
ition, particularly as regards the existence of free radicals
and the possibility of association reactions of free radicals
to azomethane.
The author wishes to express his gratitude
and appreciation to Professor H. Austin
Taylor for his many helpful suggestions
which contributed much to the progress
of this research.
Azomethane was first prepared and studied by Thiele,'*’ who
made it from symmetrical dimethyl hydrazine, using potassium
chromate as the oxidizing agent.
He reported its properties
ass boiling point, l,5°Cj readily soluble in water, with rapid
hydrolysis to yield formaldehyde and methyl hydrazine as the
principal products.
The gas could be easily combusted and ex­
ploded by an electric spark, but could be quieiflfcy decomposed
thermally if diluted with twice its volume of an inert gas, such
as carbon dioxide.
The products from the explosion were chiefly
ethane and nitrogen, along with some methane, ethylene, and
hydrogen, while the products of the thermal non-explosive de­
composition contained even more ethane and less hydrogen than
the explosion products.
In 1927, Ramsperger2 undertook an investigation of the
kinetics of the thermal decomposition of this compound in the
vapor phase, working at temperatures of 278-327° C and pressures
of 30-250mm, and following the progress of the reaction by means
of the pressure change.
The observed ratio of the final to the
initial pressure was 2,03-2,05,
f c y
) 2
If the reaction had gone entirely
which Thiele had shown to be the predominant reaction, the end
point ratio should have been exactly 2,00.
Ramsperger demonst­
rated that the reaction products contained 1,7# of an unsaturabed
compound, as determined by shaking with bromine water, which he
showed corresponded to an end point ratio of 2.034, If this un­
saturated compound is assumed to be ethylene.
The reaction was
shown to be homogeneous, as increasing the surface 6,5 times had
no measureable effect on the rate*
Good unimolecular constants
were obtained, assuming the extent of reaction to be proportionaleto the pressure change, with, however, a slight but definite
decrease in the constant at the lowest pressures studied.
results of this investigation gave the energy of activation as
51,200 calories per mole.
Ramsperger, later, also studied the reaction at still
lower pressures, down to 0.3 mm and found that the unimolecular
rate constant continued to fall, but did not reach a second order
Ramsperger showed that at these low pressures the decom­
position was still homogeneous.
This falling off of the rate
constant in unimolecular reactions is a phenomenon evidenced
not only by azomethane, but by a fairly large number of other
molecules which decompose in a unimolecular manner.
such reactions may be cited the decomposition of dimethyl and
diethyl ethers and of propionaldehyde, investigated by Hinshelwood^.
This falling off of the rate constant at low pressures
appears to be, not an isolated instance in the decomposition of
Azomethane, but a common and almost universal rule for all unmoleculazjlpeaetions •
Before continuing, therefore, it might be
advisable to briefly discuss the theory of unimolecular reactions
and the dependence of the velocity constant on the pressures
(No space need be given tojthe radiation theory of unimolecular
reactions, as it has failed to meet the experimental tests
gxmpaiBrt proposed.)
A suggestion by Lindemann
accepted viewpoint.
was the first step toward the now
He showed how to explain the falling off by
an examination of the processes of activation and deactivation
and the time lag between activation and decomposition, during
which time it is possible f<r the molecule to suffer a deactiv­
ating collision instead of decomposing.
The activation and
deactivation processes are assumed to be colllslonal and therefore
Now, at high pressures the rate of the deactivation
process is much greater that that of the decomposition, which is
a unimolecular step; the decomposition is therefore the rate con­
trolling step and the reaction is first order.
As the pressure
is lowered, the deactivation step, being a bimolecular one, is
decreased more rapidly than the unimolecular decomposition until
at quite low pressures the deactivation is the rate controlling
dtep and the reaction approaches second order.
Any constants
calculated in this region on the basis of a unimolecular rate
will appear to be lower than those for the high pressure region.
We should expect all first order reactions should approach the
second order type at low pressures; that we do[not always obtain
this result is due to the very low pressures needed in some cases
before the constant is observed to decrease, e»g. 0.05 mm for the
NgOg decomposition.
This theory has been put on a firmer basis by Rice and Ramsc o
who treated the problem from the Viewpoint of classical
statistical mechanics, considering the molecule to be composed of
coupled oscillators.
They supposed two possibilities could exist:
(1 ) an activated molecule has a certain chance of decomposing
irrespective of the amount of energy it has in excess of tin t re­
quired for activation.
This is Hinshelwood 1 s suggestion. In this
case plotting l/k against l/p gives a straight line.
(3 )
(2 ) that the chance of decomposing does depend on the energy over
and above that necessary far activation, as this excess would
termine the probability; of energy from other bonds reaching the
bond that will break.
Either of these suppositions will give unimolecular constants
at high pressures where there is an equilibrium concentration of
activated molecules, for then the rate depends only on the probab­
ility of something occurring within a given molecule. On the basis
of these two possibilities^ Rice and Ramsperger work out two
equations for these two cases to apply tcjthe low pressures region
where tl»
equilibrium concentration of activated molecules no
longer exists.
Both of these expressions give a dependence of the
rate constant on pressure, the first giving the linear relationship
between l/k and l/p and the second a non-linear relation.
The data
then existing did not permit of a choice between the two possib­
ilities in the case of azomethane.
At about this time Kassel
considered the matter from the
viewpoint of quantum statistics, treating the molecule as composed
of quantized oscillators.
He derived expressions which were ess­
entially quite similar to those of Rice and Ramsperger.
By this
time, sufficient data had been obtained to permit of a choice
between assumptions (1) and (8 ) in the case of azomethane.
showed that a straight line was not obtained by plotting l/k
against l/pj therefore, it must be that assumption (1 ) does not
hold and that for azomethane the probability of reaction depends
on the excess energy.
Kassel, in his jmririnrasi, treatment, de­
termines empirically the number of degrees of freedom which best
suits his equation.
For azomethane, the number which best
fits the experimental data is 18, leading to the|bonclusion
that the six carbonehydrogen bonds in the molecule are in­
That the latter do not contribute to the activation
is quite understandable in view of the fact that the frequency
associated with a C-H bond is high and so may be inactive at
the temperature of the decomposition (^300°C).
and Rice
have further amplified the theory of
unimolecular reactions by considering not only the low pressure,
but the high pressure, case as well.
They have shown that the
first order case applies only over a limited range of pressures,
leading to a second order type at either end of the pressure
At high pressures we may have the usual bimolecular
case, of collision immediately followed by decomposition.
we consider that there is a part of the molecule which is conn­
ected by a rather weak coupling to the remainder, so that energy
transfer between the two portions is considerably slowed up,
then we may have two kinds of collision.
The first would be
collisions involving that part of the molecule which contains
the bond tcjbe broken; here collision between molecules results
in immediate rupture, according to the second order equation.
The other would be a collision involving that part of the mol­
ecule on the other side of the barrier link; such collisions
would result in decomposition only after some time lag in which
the energy transfer would slowly take place and would follow
the first order law.
At high pressures the bimoleoular case
would predominate; at lower pressures where the time between
(5 )
collisions is greatly increased, the transfer of energy across the
weakly coupled link becomes a more important source of activation
for the decomposing part cf the moleculte than the rather rare coll­
ision?; and the reaction becomes first order.
At still lower press­
ures the reaction again becomes second order for the reasons cited
on the previous page.
It may be remarked that for no molecule
has this complete sequence been observed.
Ramsperger has studied tl® efficiency of ethane and nitrogen
as compared to azomethane in the aaomethane pyrolysis.
In his
earlier work Ramsperger showed that the first order rate constant
did not change during the cotuese of any given experiment, even at
low pressures.
This indicated that the products of the reaction
exerted about the same effect in maintaining the rate as their eq^~
ivalent of azomethane.
According to the Rice- Ramsperger-Kassel
theory of unimolecular reactions the rate is not maintained at
low pressures since collisions cannot supply activated molecules
sufficiently fast for reaction and still maintain the MaxwellBoltzmann fraction of activated molecules.
If now an inert gas
be added it will increase the rate of activation Insofar as it is
able to transfer internal energy on collision with azomethane mol­
ecules •
His results showed that ethane is practically as effi­
cient as azomethane while nitrogen is much less so.
The efficiency
apparently depends In some manner on the number of internal degrees
of freedom; ethane is therefore the more efficient.
In some cases
it has been stated that very simple molecules are quite efficient
as activating agents.
In none of these cases,however, has It been
definitely shown that tbe "activating" molecule does not react
(6 )
Chemically with the substance undergoing decomposition* (See p.9)
Ramsperger has studied not only the pyrolysis but the photolysis
of azomethane as well*
trum, and also tlB
chain reaction.
He determined the absorption spec­
quantum yield to see if the photolysis was a
The quantum yield was obtained using a thermo-
couple-galvanometer setup, the extent of the decomposition being
determined by the pressure change, using the end point ratio of
2.04 as determined in the pyrolysis.
The quantum yield so obtain­
ed was about two, with practically no temperature coefficient. Ramrsperger suggests that this can be explained
by assuming that a new­
ly formed ethane molecule activates another azomethane molecule by
collision, thus causing two azomethane molecules to decompose per
quantum absorbed.
It is difficult to see, however, why this pro­
cess could not be repeated again and why the yield should not rise
above two; also on such a basis, one would expect that the quantum
yield should depend on the pressure, since on this theory it invol­
ves a collision mechanism, whereas it has been shown by Forbes,
Heidt, and Sickman^-0 that this is not so*
These latter have also
studied the quantum yield of the photolysis, tising a uranyi oxal­
ate actinometer and also basing extent of decomposition on pressure
In contrast to Ramsperger1s work they report a quantum
yield of unity at pressures below
mm, the yield dropping off
at higher pressures due to collisional deactivation*
temperature coefficient is reported above 250°0.
A slight
They also found
a decrease in the quantum yield as the wave length of the incident
radiation was lowered.
Patat1‘L has checked the quantum yield of unity reported by
(V )
the above*
Too much reliance, however, cannot be placed In his
results as he based his extent of decomposition on the amount of
product not condensed at -140°C*
this assumption is not valid.
It will be shown later that
In none of this work was there
any evidence of a chain reaction*
It must be emphasized that in
none of these researches was any systematic gas analysis of the
products attempted; it was assumed that, asidd from a slight
side reaction to give ethylene, the only mechanism for the decom­
position was a rearrangement to give ethane and nitrogen.
was the first to demonstrate the presence of
free radicals in the pyrolysis of azomethane, although Thiele
makes mention of the possibility of their existence.
evidence was obtained by the mirror removal method u.sed for the
detection of alkyl free radicals.
He showed that at a temperfcfcure
of 475°C lead mirrors were quickly removed by the products of the
pyrolytic decomposition even when the mirror was placed 1 0 cm
from the furnace. In further work Leermakers attempted to use
this method to see if short reaction chains played any part In
the thermal decomposition.
Prom the work of Ramsperger on tie
negative effect of altering the surface-volume ratio it is known
that long chains must be excluded, but still the possibility of
short ehains had to be reckoned with.
To do this , he decomposed
mixtures of azomethane and tetraethyl lead at 250-275° C, at which
temperature lead tetraethyl readily decomposes to give ethyl rad­
icals, while azomethane decomposes only slowly.
It was argued
that If any chains were taking part in the decomposition of azo­
methane, that the ethyl radicals could initiate th chains and a
(8 )
more rapid decomposition could be observed.
However, the velocity
constants for the decomposition of the lead tetraethyl in the mix­
ture was not much different than for pure lead tetraethyl, being
actually a little lower after correcting for the simultaneous
slow decomposition of the azomethane.
this will be discussed later.
The full significance of
This shows that free ethyl radicals
have no appreciable effect on the azomethane decomposition and
since the reactions of ethyl and methyl radicals are very similar
it would indicate that there are no chains taking part in this de­
Prom this work an activation energy of 15 kcal. was
obtained fcr the reaction of azomethane and ethyl radicals.
While there is no evidence for chains involving methyl radical
there is considerable evidence that methyl radicals are produced
in the decomposition.
This has been demonstrated by Sickman and
, who studied the decomposition of acetaldehyde-azomethane
mixtures at temperatures around 300° C, at which temperatures azo­
methane readily decomposes but at which acetaldehyde is thermally
Under these conditions complete decomposition of the acet-
aldehyde took place.
It is known from the Rice-Herzfeld mechanism
for the decomposition of acetaldehyde that methyl radicals are an
essential part of the mechanism.
It is quite evident from this
that the azomethane is supplying the methyl radicals to initiate
the chains which acetaldehydecannot itself produce until about
Sickman and Allen give chain lengths for this induced de­
composition varying from 500 at 250°to 22 at 330°C.
The chain
breaking step is assumed to be the recombination of methyls to
give ethane, this process becoming more efficient with rise of
(9 )
Allen and Sickman have also shown
that azomethane causes
polymerization of ethylene as well as setting up chain decompos­
itions in propionaldehyde and isotmtane, all presumably Initiated
by the methyl radicals coming from the decomposing azomethane.
Patat has also shown that azomethane given methyl radicals.
He supports this conclusion by showing that dimethyl ether can be
decomposed by small amounts of azomethane if the system is irrad­
iated with light of a frequency which will decompose azomethane
and have no effect on the dimethyl ether by itself.
Sickman and O.K. Rice have made a very thorough study of
the rate of the thermal decomposition, covering a wide range of
pressures and again following the progress of the reaction manometrically.
Their results are in somewhat better accord with tl©
Rice- Ramsperger and Kassel theories than was the earlier data of
They also studied the decomposition in the presence
of supposedly foreign gases,He,COgjHgO^H^NgjOC^Dg, and H2 » Their
conclusions were that the only effect of the foreign gas was that
of collisional activation.
They calculate collision efficiencies
for th various gases, relative to azomethane.
With hydrogen a
greater acceleration was noted in the rate than with any of the
other gases; this would ordinarily not be expected as hydrogen is
a simple molecule and should not be as efficient as the more com­
plicated ones.
They ascribe this to a possible reaction of methyl
radicals with hydrogen.i It is unfortunate that this was not defin­
itely shown inasmuch no gas analyses were performedJvrom the known
fectivation energy of the reaction
CH3 + Hg -- > CH4 * H
E 11.1 kcal.
it seems as though this reaction must have occurred, as it goes
readily above about 170°C and therefore certainly at 30010.
An interesting aspect of the decomposition of azomethane is
furnished by the paper of Emmett and Harkness
, who studied its
catalytic decomposition over an iron (synthetic ammonia type)
catalyst at 250-280° C.
A rapid decomposition was obtained, with
a rapid pressure increase followed by a much more gradual change*
The ratio of the final to the initial pressures averaged 2.20 •
The products ifethined were quite different from those of the homo­
geneous decomposition*
It was found that nearly all of the nitro­
gen formed either a nitride with the iron catalyst or fcrmed a
water solu^ble gas, which was shown by vapor pressure determinat­
ions to be ammonia*
The carbon appeared as elementary carbon de­
posited on the catalyst.
In a few experiments sufficient hydrogen
was added initially to prevent nitride formation.
In these cases
an initial pressure decrease was noted In place of the former
initial rapid rise.
They attribute this to the reduction of the
nitride by the hydrogen to form ammonia, which reaction, of course,
would result in a pressure decrease.
An especially interesting
result from the standpoint of the data tcjbe presented later in this
thesis is that they definitely established the existence of methylamine as an intermediate product.
This will be discussed later.
Emmett and Harkness also studied the catalytic decomposition
of symmetrical dimethylhydrazine and found it to pass through mud*
the SLame stages as the above.
This would suggest that in the azo­
methane case It is quite possible that an intermediate previous
to the methylamine is the dimethylhydrazine.
The thermal decomO
position of the hydrazine was also studied at 520-350 C and shown
(1 1 )
to be a complex first order reaction, which goes about 50-75%
according tothe equation
CHgMHNHCHg— * 2 CH4 +•Ng + excess
ethane +■ some Unidentified water soluble material.
It must be emphasized again that since the early work of
Thiele no complete analysis of the products of the decomposition
had been made.
All subsequent investigators had assumed that the
course of the reaction could be represented by
7 c 2h6
The earliest indication that such was not the case is tojbe found in
a note by Forbes and Heidt ,who reported the possibility of meth­
ane being among the reaction products from consideration of the
amounts of product condensed by dity ice and by liquid air.
But it
was left for the work of Riblett and Rubin, and of Davis, Burton,
and Taylor (working in this laboratory) to follow the progress of
the reaction by means of gas analyses and not by pressure change
Burton, Davis, and Taylor
had started out to determine the
true chain length in the decomposition of acetaldehyde induced by
methyl radicals from azomethane, and to determine the-number of
methyl radicals formed by nitrogen determination.
In the course of
the investigation they analyzed the products from the photolysis
of pure azomethane and found that the results did not agree with
the idea that the products were solely ethane and nitrogen; the
amounts of nitrogen formed did not bear any simple ratio to the
yield of hydrocarbon product.
The most important features were
that 54.7^of the acid insoluble products was nitrogen, regardless
of the extent of decomposition.
At ten percent decomposition, at
20°C, there was found 3 % methane and 42% ethane.
Also propane
was found, the yield becoming larger as the percent decomposition
These analytical results show that azomethane must be
removed from the system by some means other than by decomposition
to yield nitrogen and ethane.
It was suggested that the mechan­
ism for the photolysis might run as follows}
Primary steps:
(CH3 )gNg
— 7
(CHg)gNg — 9
CH3 ^ Ng
followed by reaction of the methyl radicals.
Among the suggested
reactions was an addition of the methyl to azomethane, to yield
a more complicated radical, which on further reaction with methyl
would yield a completely methylated product, tetramethylhydrax&ne,
(CH3)gNg 4- CH3
-* (CH3)3Ng
(CH3 )3 N2 4- CH3
— > (CH3 )4Hg
Among other possibilities which were suggested ’was the possible
union of two of these (CH3)3Ng radicals to form a substituted
To explain the deficiency of carbon over nitrogen in JP
the products insoluble in acid, one of the acid soluble ones must
contain a higher carbon-nitrogen ratio than does azomethane itself.
The reaction above to yield tetramethylhydrazine presents such a
compound, the carbon-nitrogen ratio here being tv/ice tin t in azo­
If we assume that the tetramethyl^is formed from methyl
radicals,and the ethane by direct rearrangement, it follows from
the nitrogen yield that otofc of five molecules of azomethane decom­
posing only four give ethane and that therefore at least one of the
five must have decomposed tcjgive methyl radicals.
This is at best
a minimum estimate as the possible recombination of two methyls to
g±a give ethane was not considered.
It may
be remarked that no
product resembling tetramethyl hydrazine was actually found.
This compound is a liquid, boiling at 60°C under a pressure of
17mm (unpublished data of Mr. Leondud May) and therefore should ka'C
have been observed in the gas burette prior to analysis.
it was not so observed may be due to the very small amounts of
it which could have been formed.
Calculation shows tha t in none
of theBe experiments did the amount formed correspond to more than
about one drop of liquid, which could very easily have been over­
In contrast tcthe work of Sickman and Allen^ who found that
the pyrolysis product of azomethane caased complete decomposition
of acetaldehyde at 300°C, Davis, Burton, and Taylor found that the
products of photolytically decomposing azomethane at room temper­
ature caused, not a decomposition, but a polymerization of the
acetaldehyde. Their results show that for every two molecules of
acetaldehyde polymerized 3.3 molecules of nitrogen were produced.
If every methyl radical is assumed involved in the polymerization
of two aldehyde molecules, it would follow that one out of every
6 .6
molecules of azomethane decomposed via free radicals.
This f
figure is not a maximum, as the authors claim, since as mentioned
above they have neglected the possibility of two phsi methyls re­
combining to from ethane.
These results would indicate that in
the photolysis azomethane nay decompose in two ways, via free
radicals and by direct rearrangement.
These authors call attention to the facjt that the data of
on the effect of ethyl radicals on azomethane may
have another explanation than that offered by Leermakers.
found that there was no increase in the rate of the simultaneous
thermal decomposition of azomethane and lead tetraethyl as compared
to their decomposition separately and he therefore concluded that
there is no effect of alkyl radicals on azomethane.
Actually he
had obtained a 10-25$ decrease in the rate over tht of pure lead
The explanation now offered is that the ethyl radicals
could disappear by addition to the azomethane instead of all form­
ing butane, thus causing a lower apparent rate.
The results of
Leermakers experiments, as interpreted on this basis, indicate that
not all of the radicals formed react at the experimental temperature*
(260-275°C) with azomethane.
In a later paper, Button, Davis and Taylor
of temperature on the photolysis of azomethane.
studied the effect
The previous work
had shown that at room temperature the hydrocarbon fraction consists
chiefly of ethane and that there were two primary processes occurr­
ing, one via direct rearrangement and the other through a free rad­
ical mechanism.
These later results showed that as the temperature
increased, methane production increased while that of ethane dim­
inished until the methane became the predominant hydrocarbon. Above
200°C, propane began td(be formed in appreciable
above 163°C, a condensable liquid was observed in the gas burette
prior to analysis.
These results would indicate that the reaction
to give methyl radicals becomes more predominant as the temperature
The reactions following the primary split may be:
CHg t CgHg
—^ C 3 H 8 + H
CH3 +C 2 H 6 — '* CH4 + C 2 H 5
The ethyl radicals may disappear by CH3 •+CgH3
or by association with azomethane to form compounds of the type
discussed above (’tetramethylhydrazine) • It should be noted
that in none of this work was any hydrogen obtained, but this
is no serious criticism as the hydrogen atoms may be removed
much more readily by
A. reaction which has an activation energy of only 7*2 kcal.
The increased yield of methane at higher temperatures accom­
panied by a decrease in the ethane yield suggests that reaction
(2 ) may be the predominant one*
The liquid condensate appearing at 163°and higher is in­
teresting in tha^microanalysis of one sample of it gave its em­
pirical formula (assuming it to be a pure substance) as CgJfsHs,
Comparing this with azomethane, CgHgNg,it is
apparent that there is a deficiency of carbon atoms.
This does
not then explain the high nitrogen-carbon ratio in the acid insolublejgases.
It may be that other kinds of associated molecules
also occur,e.g. by hydrogen replacing some of the methyls in these
This would not, however, explain the deficiency of
carbon, unlessit is assumed that an error had been made in the
The carbon-nitrogen ratio in the gases rises above
200° C, which would, indicate that the associated molecules formed
by reaction of azomethane with radicals are themselves unstable
at higher temperatures.
A redetermination of the quantum yield of
the photolysis, based on nitrogen determinations, disagreed with
the conclusionjreached by previous workers that the temperature
coefficient of the photolysis is zero below 226°C.
Riblett and Rubin23have carried out a thorough investigation
of the thermal decomposition, doing detailed gas analysis at
various stages of completion.
(1 6 )
They were able to determine the
residual azomethane, and they based extent of reaction on azometh­
ane decomposed, which is,after all, the only true criterion*
results are qualitatively, although not quantitatively, in agreement
with those of Burton, Davis, and Taylor in the photolysis*
At 3102
340°C, they showed that in all cases methane was formed to a larger
extent than ethane.
Also a fairly large amount of some material,
liquid at room temperature, was observed.
They present their data
in an interesting manner, plotting percentage decomposition versus
number of moles of product per
moles of azomethane decomposed*
They show thbt in this manner straight lines are obtained for the
nitrogen and liquid formations, the nitrogen line sloping upward
with Increase In percentage decomposition and the liquid yifldui
yield decreasing.
Both lines meet at the zero percent axis at app­
roximately 50 (I.e., one half mole of each for every azomethane
mole disappearing)*
This would suggest a very interesting possib­
ility, I.e., that the initial step Is the bimolecular association
of two azomethanes with a simultaneous splitting out of one mole
of nitrogen.
This corresponds to the formation of one mole of
tetramethylhydrazine, and it would be very informative to see if
tetramethylhydrazine decomposes similarly to azomethane*
work is being conducted at present but no results are as yet
The other products could be accounted for as further
degradation of tbs heavy molecule.
Such a scheme is compatible
with the disappearance cf the liquid and the linear rise of meth­
ane production form zero at zero percent decomposition. It cannot
however account flor the fact th t the ethane yield is constant at
27$ and independent of the extent of decomposition.
(1 7 )
It may be
that the ethane is produced by a concurrent reaction, possibly
from singl^fazomethane molecules.
Davis, Jahn, and Burton
have studied the photolysis of
azomethane in the presence of nitric oxide, which is an efficient
agent fcr tl© removal of free radicals, such as methyl, by com­
bination with them.
is formed;
The results show that practically no ethane
this is taken to indicate that the reaction goes
almost entirely via the free radical route and not by rearrange­
ment directly into ethane and nitrogen.
No pressure increase isy^
for some while after the reaction is started.
This is attributed
to the sequence of reactions:
(CH3 )2 N2
for which no pressure change should occur if CHgNO Is volatile
(CH3 N0 may be an isomeric form of formaldoxime) • After all the
NO has been used up the reaction should then proceed and give a
pressure increase, as usual; this was observed.
Actually this
scheme cannot accourt for everything occurring, since it would
require tha t the amount of nitrogen formed up to^he point at
which the pressure starts to Increase should be equal to one half
of the nitric oxide originally present.
The experimental results
show that in all cases the nitrogen-nitric oxide ratio is greater
than 0.5. This must mean that nitric oxide adds on to one of the
complicated free radicals discussed by Burton, Davis,and Taylor
or reacts in some other manner different from the above causing
removal of some of the NO.
has also investigated tbepyrolysis of azomethane in
the presence of nitric oxide.
The amount of nitric oxide
consumed was about twice that of the azomethane disappearing.
This would show that in the pyrolysis, as well as in the photo­
lysis, the decomposition occuri's entirely by a free radical
mechanism. If sufficient nitric oxide is present no association
reactions take place; if not, association was shown to have
Further work by Jahn on pure azomethane gave the
energy of activation of the decomposition as 52,500 calories per
mole, somewhat higher than Ramsperger's old value of 51,200.
The progress of the reaction was followed by determination of the
residual azomethane.
The only other work on the hydrogen-azomethane reaction,
aside from that of Sickman and Rice^previously mentioned, has
been the work of Patat , who investigated the photolysis of
azomethane at 265°C in the presence of para-hydrgoen at wave­
lengths to which hydrogen is indifferent. He showed that a paraortho hydrogen conversion took place, which he explained as due
to the presence of hydrogen atoms in the photolytic mixture.
His sequence of steps are: the photolytic decomposition of the
methyl radicals and nitrogen; the reaction
of the methyls with the hyntanraigg hydrogen to yield atomic
CHg+ Hg
CH4 -+ H, which reaction we know occurs
readily above 165°C; subsequent reaction of the atomic hydrogen
with azomethane
— > CH^v Ng-» CH3 *
From the extent of the para-ortho conversion and from the known
value for the activation energy of the reaction of methyl with
hydrogen, he deduced a value of 5.1 kcal. for the energy of
Activation of the atomic hydrogen-azomethane step*
very similar reaction, CH3 + (CHg^Ng
4" %
assigns a minimum activation energy of 20 kcal.
For the
^ CEg;k©
No gas anal­
yses were done on the products to se®if they corresponded en­
tirely to msthane and nitrogen.
Evidence will be presented in
this thesis thfat at somewhat lower temperatures at least, this
reaction to yield methane, while it does occur, is by not
means the only one occurring.
Addition of hydrogen to^he azo
linkage must also be considered.
The material presented in this the sibs is an outgrowth of
the wcrk of Burton, Davis, and Taylor in which they showed the
complexity of th^azomethane decomposition.
It was thought
that by a study of the hydrogen-azomethane reaction that
further knowledge of this decomposition and of its mechanism
would be forth coming, particularly as regards the presence
of methyl radicals and the possibility of association reactions.
The azomethane used in this research was made from symm­
etrical dimethylhydrazine by oxidation with cupric chloride
The dimethylhydrazine was prepared using the
method of Hatt
, of benzo$rlating hydrazine, methylating
the product, and then hydrolyzing offjthe benzoyl groups with
strong HC1.
The yields obtained were in general as good as
those given by Hatt, except for the methylation step where a
yield of
6 5 %
was the best attainable, as compared to the 90%
which he quotes.
The dimethylhydrazine was oxidized following
the procedure of Jahn
, by adding concentrated cupric chloride
solution drop by drop to a well stirred solution of the hydra­
zine dihydrochloride which was well buffered with about six
times the number of moles of sodium acetate.
The brick-red
precipitate of cuprous chloride-azomethane complex, which is
formed was thoroughly washed and then dried by letting stand for
two weeks in a vacuum desiccator over concentrated sulfuric acid,
after which time constant weight was attained.
The precipitate was then finely ground and placed in a 500
c.c. flask connected by means of a ground glass joint to the
vacuum line.
Two traps, one cooled with toluene-solid carbon
dioxide and the other with liquid nitrogen were placed in the
line not far from the flask with soda-lime drying columns before
and between the traps.
After a good vacuum «0.01mm) was obO
tained, the flask was cautiously heated up to 135-140 C and was
kept at this temperature until all the complex had decomposed,
as evidenced by the change in color from red to a grayish white:
Cu2 C1 2 *(CH3 )2 N 2
Cu2 Cl2 +(CH 3 )2 N 2
In contrast to the work of Jahn, it was found that very
little product froze out in the dry ice trap and that most of it
was in the liquid nitrogen trap; this difference is probably due
to the differences in the rates of heating and pumping.
product so obtaimed was repeatedly fractionated by distilling
from solid carbon dioxide to liquid nitrogen; ten such fraction­
ations were performed in all.
The resulting liquid (at-78°C)
was about 7-8c.c. of a very pale yellow, almost colorless liquid;
this amounted tcja yield on the oxidation step of approximately
65$, in good agreement with Jahn* s value of 70$.
The liquid
had a vapor pressure of 7.2 mm at -78° C and 751 mm at 0°C. This
latter value is identical with that given by Ramsperger , and
close to that of Burton, Davis, and Taylor (unpublished, but
given orally as 754 mm), but considerably lower than the value of
765 mm reported by Jahn.
No reason can be offered for this
The product so obtained was found to be completely
soluble in dilute sulfuric acid.
No combustions were attempted
to determine its purity as it has been found
that combustions
of azomethane often give rise to disastrous explosions.
The reaction vessel was a 500 cc. flask, with two inlets,
one for azomethane and one for hydrogen, with an outlet tube
connected to the pumping system, which consisted of a mercury
diffusion pump backed by a Hyvac oil pump. The two inlet tubes
were so sealed in as to face each other and cause the two in­
coming streams of gas to meet and mix in the middle of the
reaction vessel. The tubes were tapered at the ends so as to
prevent back diffusion of gases. For the lower temperature work
the reaction flask was surrounded by a water thermostat, stirred
by compressed air, which was held at
which could be regulated to
0 , 2 °
2 7 ° C
by manual control and
C. At the higher temperatures
studied (110°and 195° C) an]plectrically heated, well insulated
furnace was used. Two thermometers, placed in different parts
of the furnace, agreed to within 1.5°C and it would seem from
this that the thermostatic control at these temperatures was
good to
-2 °C.
ThehA'ydrogen used came from a tank of electrolytic hydrogen.
It was at first attempted to prepare the hydrogen by electrolysis
o f 1 0 %
potassium hydroxide solution, but it proved too cumbersome
to Introduce the hydrogen so prepared into an evacuated system
without having siphoning of the electrolytic solution. This method
was soon abandoned in favor of the tank electrolytic hydrogen.
No attempt was made to remove the small amount of oxygen present
in tank hydrogen, inasmuch as it has been found that small amounts
of oxygen aid in the proper working of a Wood’s discharge tube for
making atomic hydrogen, presumably by the formation of traces of
water, which acts as a negative catalyst for the recombination.
The hydrogen was let out of the tank by a needle reducing valve
and its pressure further reduced by several partly opened stop­
cocks in the line. Past the last stopcock a long, small bore
capillary tube of about 60cm length was inserted. During the pre­
liminary work an automatic pressure regulator,operated by a relay,
was used, but it was soon found that this was unnecessary, as the
capillary itself functioned as an excellent regulator and kept
the pressure and rate of flow constant to .0 1 -.0 2 mm, as read on
a double Mcleod gauge^which served as a flowmeter. From the flow‘
P 7 a
meter the hydrogen^ the Wood’s discharge tube6 '*, which was a long
tube of about 135cm length and 1.3cm diameter. The inlet for the
hydrogen was near one electrode and the outlet in the center of
the tube, as it is at this point that we get the greatest con­
centration of atomic hydrogen, as here the gas is furthest re­
moved from thejcatalytic effect of the electrodes. These latter
consisted of hollow aluminum cylinders, about
1 0
cm in length and
one cm in diameter and of thickness about one mm. These were spot
welded to tungsten leads which were in turn sealed through the
glass at the ends of the discharge tube.
The walls of the Wood’s tube and connecting outlet tube were
poisoned against the recombination of the hydrogen atoms by let­
ting the system stand tinder acacuum in the presence of concentratedd
(1OO/0 sulfuric acid. The recombination of hydrogen atoms to form
molecular hydrogen is a termolecular process, occurring either
with the surrounding wall as the third body or homogeneously in
the gas phase by collision of either three hydrogen atoms or two
atoms and a molecule, with the atom being the more efficient third
. It cannot occur as a two body process because if it did
there would be no provision for the energy of formation and the
molecule would break up again.
A very efficient third body is
the wall of the surrounding vessel, and at the low pressures used
in this research most of the recombination would take place via
this route. It has been found that the dehydrating acids, HP0 3
and HgSO^, are excellent negative catalysts for the recombination
on the walls and prevent it so effectively that the only appreci­
able recombination is that occurring in the gas phase. The sul­
furic acid is used as described above, as first recommended by
Smallwood , while the phosphoric acid is usually applied directly
to the walls of the tube either as syrupy phosphoric acid or
melted on from sticks of HPOg30. It is probable that the ability
of these acids to prevent recombination is due to their dehydrat­
ing action and that it is really the water which they adsorb on
to the walls which is the real poison. Of the two, the phosphoric;
acid has been used the more frequently but it has the disadvant­
age that the hydrogen used must be
2 . 5 %
water vapor,
obtained by bubbling the hydrogen through water) to secure best
results. With the sulfuric acid dry hydrogen may be used; it may
be that the small amount of water present in sulfuric acid plus
the traces of water formed from the oxygen present
the hydrogen
are sufficient for the purpose. For this reason the sulfuric acid
method was the one chosen, as it was not known what effect the
water would have on the azomefRhane. It would certainly have complicated the gas analyses as azomethane is known to react rapidly
with water1,
From the discharge tube, the hydrogen was passed directly
into the reaction vessel, where it mixed with the azomethane.
The flatter was kept in dtoluene-solid COg trap, and its pressure
regulated by a stopcock over the reservoir. The rate of flow was
measured by a capillary flowmeter similar to that used for the
hydrogen. The gases coming out of the reaction flask were first
passed through a liquid nitrogen trap placed before the mercury
ppmp, then through a silica gel trap cooled by liquid nitrogen and
placed between the mercury and Hyvac pumps. It has been stated by
Steacie3-*-and others that silica gel at this low temperature is a
very efficient agent for separating ethane and methane from hydro­
gen; the vapor pressure of the former is too great to permit of
their being completely taken out by liquid nitrogen alone. As will
be seen in the later portion of this thesis, the silica gel is
quite efficient for this purpose, separating nitrogen as well as
hydrogen from the hydrocarbons. The silica gel was prepared according to the directions of McGavack and Patrick
solution of sodium silicate was added to
. A concentrated
N HCl and thoroughly
agitated. The gel so formed was allowed to stand for several hours,
until quite firm. It was then filtered off and washed until it
gave practically no test for chloride, after which it was dried in
an air furnace at 110°C. The gel was then placed into the trap
which was then affixed into the system, and then heated under vacuum
with a small Bunsen flame. This left a gel of high adsorptive power
Both of the condensation traps wBre connected to a liter
Toepler pump, from which the reaction products could be sent to a
gas analyzer of the Fisher type, consisting of:
(1 ) a gas burette
(2) potassium hydroxide (35f t ) absorption bulb for COg
(3) alkaline pyrogallate for oxygen
(4) sulfuric acid (0.5N) for azomethane and other basic gases
(5) CuO combustion tube for hydrogen
(6 ) bromine water (sat.) absorption bulb for unsaturat^ds
(7) catalytic combustion apparatus for hydrocarbons
The problem of accurately determining hydrogen atom concentrations, especially in dynamic systems, is one which has been
attended with considerable difficulties. In the static method
used by Patat
to investigate the photochemical hydrogen- azo­
methane reaction, the concentration of atomic hydrogen was deter­
mined using the para-ortho hydrogen conversion. In a dynamic system
such as the one employed of necessity in this research, this method
is not applicable. There have been other methods proposed for meas­
uring the extent of dissociation in dynamic systems, among v/hich
may be mentioned:
7 ‘T
(1) the method of 7/rede and Harteck
-making use of Graham's law
that the velocity of effusion is inversely proportional to the
square root of the molecular weight. Results were obtained using
a gauge devised especially for the purpose.
(2) the method of Bichowsky*^-making use of Graham's law to meas­
ure the number of hydrogen atoms passing through a tiny hole in a
given time, after having^calibrated with molecular hydrogen. From
the observed heat effect of the recombination, he was able to cal­
culate the heat of formation of hydrogen.
There is a serious criticism which may be leveled against
both of the above methods. In order to make certain that all the
gas passing throtigh the opening does so by effusion and not by
mass flow occasioned by a pressure difference, it is necessary
that the criterion of Knudsen
be obeyed, that the diameter of
the opening should be roughly of the order of one tenth the mean
free path. There is no evidence that any attention was paid to
this point and in all probability the holes were considerably
too large and the results obtained therefore not valid. It is true
that Bichowsky reported a value of 103,000 calories for the heat
of formation of hydrogen,cin good agreement with spectroscopic
data, but a check on their mathematics shows the presence of an
arithmetical error which brings the value down to 83,000 calories,
considerably away from the accepted value.
(3) the purely thermal or calorimetric methods, all based on the
Hgt Hi* M
—» H g + M + Q cal.
Among these may be mentioned the methods of Bonhoeffer36a, Bichowsky36^, v.Wartenburg and Schultze36c, Dixon36d, Smallwood®0,and
Poole36f. Leaving out of consideration temporarily the last named &
of these, none of the above methods was capable of high accnracy,
coupled with reasonably rapid operation and the possibility of
making successive readings. Of these methods, leaving out Poole*s,
Smallwood's was probably the best, but even in this case correction
for heat losses ran to 40-45$ and twenty minutes were required
for each determination; no successive readings were possible.
At about the time this research was started the method of
Poole made its appearance. In his calorimeter he claimed to have
overcome the disadvantages of the earlier methods and to have
developed an accurate method, capable of successive readings
and with heat losses ranging from 1-3$. Readings could be taken
after the first 30 seconds. A variation of this calorimeter was
used in this research; it was found that the heat losses were
-8 $ and that thermal equilibrium was reached in 2-3 minutes.
This is still a decided improvement over any of the other'methods.
The gauge consisted of a glass tube about 40cm long and 2.5
cm wide. By means of de Khotinsky vacuum cement a silver tube,
7mm in diameter and 3/4mm thick, was fastened coaxially into the
glass tube. The iqixture of atomic and molecular hydrogen enters
4 ^W. *-$. t1 'Z j- 'i
at the middle.ana is pumped out at one end. The silver is an ex­
cellent catalyst for the recombination and any atomic hydrogen is
r,Jc '
practically immediatelyp aused to recombine on the surface, giving
it s heat of recombination to fthe silver which in turn delivers it
to a stream of water flowing through the interior of the silver
tube at a fixed rate. The increase of temperature of the water
was measured with the aid of a pair of sensitive hundredth degree ^
thermometers, which could be read to ,0 0 2 °by use of a hand lens.
The thermometers were thin enough to fit right into the silver
tube and therefore right into the stream of flowing water. Twisted
strands of copper wire were placed within the silver tube to pro­
mote thorough mixing of the water. A McLeod gauge was attached to
measure the pressure, so as to permit calibration for hea# losses.
Practically all of the hydrogen recombines within one or two centi­
meters of the entrance to the gauge; the remaining length is to
make certain that none of the atoms pass through withAut combining*
Of course, the thermometers must he placed outside of the region
of heat production. To the heat as measured by the observed temp­
erature difference must be added the heat carried off by the hydro­
gen itself as determined by calibration at various pressures. This
heat loss was determined by passing water through the calorimeter,,
the entering water being several degrees below room temperature,
and observing the temperature change ah various presstires of hydro­
gen when the arc was not working. Since the heat loss depends on
the rate of flow of the water, this was standardized at
in all the calibrations and in the actual runs. Suich a rate allowed
a reasonable temperature difference during the runs and still not
too high a percentage correction' (6 -8 $. Poole has shown that the
rate of flow of the hydrogen has no effect on the heat loss, pre­
sumably due to the small
curve so obtained resembles the usual ones for this type of work,
rising rapidly at low pressures and then beginning to flatten out
at higher pressures.
Prior to the first run, the Wood's tube was operated for 24
hours at a high current (9-10&mps). This has been recommended by
Bonhoeffer*^a and others as a procedure to ”burn out" any active
spots on the surface of the tube which might aid in the recombin­
ation of the atoms.
The apparatus was let stand overnight under
vacuum before each run, except in the few cases where two runs a
day were made. This sufficed to allow enough sulfuric acid to be
adsorbed on the walls to poison them effectively. The discharge
tube was operated for thirty minutes before each run; this is nec(SO)
essary as the atomic hydrogen production is not steady at first
and only gradually becomes so. The normal operating conditions were
7-8 amps in the primary, corresponding to 350-400 milliamps in
the secondary at about 3000 volts.
After thirty minutes the azo­
methane was run in and the reaction allowed to proceed for forty
five minutes. The reaction products were condensed in the traps
previously described and analyzed. Just before the azomethane was
let in and just after it was shut off, readings were taken on the
atomic hydrogen gauge to determine the number of hydrogen atoms pass­
ing through. The gauge was so situated that it was exactly the same
distance from the outlet of the Wood's tube as was the reaction
vessel and connected to it by tubing of the same bore and having
the same number of bends. By means of two stopcocks it was possible
to direct the hydrogen either through the gauge or the reaction
vessel. No change was observed in either the pressure or the rate
of flow when th^stream of hydrogen was diverted from one channel
to the other. It is logical to assume that the concentration of atom­
ic hydrogen as read from the calorimeter was the same as that enter­
ing the reaction vessel. It might be argued that the flow of the azo­
methane might alter the rate of flow of the hydrogen sufficiently
so that this would no longer be true. However, it was noticed
that after a momentary fluctuation when the azomethane was intro­
duced that the flow readjusted itself to the previous value, as
did also the arc current, showing that the above assumption is
probably quite valid.
OP AZOMETHANE (J. CHEM. PHYS., 7, 829, (1939))
The flowmeter used for the hydrogen was not calibrated direct­
ly; the volume of hydrogen passing through was calculated from the
dimensions of the capillary and the known viscosity of hydrogen.
The small correction for slippage was applied. For azomethane, no
data on the viscosity has appeared up to the present; it was there­
fore necessary to calibrate the flowmeter directly in terms of the
volume passing through per unit time under various pressure diff­
erences. The azomethane was allowed to flow from the reservoir
through the flowmeter capillary (length, 6.97cm; diameter, 0.140hb&)
and was frozen out in a liquid air trap placed behind the reaction
vessel and before the mercury and Hyvac pumps. After each run the
liquid air was removed and the azomethane forced over to a gas
burette by a Toepler pump. All volumes were reduced to S.T.P. As
a check on the purity of the azomethane the material from a few of
the runs was passed through dilute sulfuric acid. The gas dissolved
All runs were for a period of fifteen minutes in order to
conserve azomethane. The experimental results are shown in the
following table, where P 2 anubcjpt and p 1 are the pressures in
millimeters on the high and low pressure sides of the capillary,
respectively; Vs is the volume of azomethane in cc. at S.T.P.;
the viscosity as determined directly from Poiseuille's law
(see below); and
is the viscosity corrected for slippage.
Q. 6 8
0 C74
P 2 -Px
Thw viscosity of a gas may be obtained by the use of Poiseuille’s law in the form:
- d 7f*l * P2-P-.
at " 128L^RT
where a n is the amount of gas in moles passing in time at through
a capillary of length, L, and diameter, a, with a viscosity,7 ,
under a pressure difference, Pg-Ri*
absolute temperature,T.
This equation is valid if the flow is viscous and isothermal and
if the gas is ideal. These conditions are met by azomethane under
the conditions employed. All runs were made at room temperature,
which remained constant during the time at 23-24°C. Ii^the calcul­
ations the temperature vms taken as 296<>K. Viscosities calculated
in thi£ way give only the apparent viscosity,^, as given in the
sixth column above. This must be corrected for slippage and for
the pressure effect due to the mass motion of the gas. The latter
may be entirely neglected at these small rates of flow, but at
the low pressures employed the correction for slip amounts to
several percent and must be taken into consideration.
To obtain the true v i s c o s i t y , t h e apparent viscosity, ,
must be corrected by use of the equation:
7. - 7^
1 / 2
where z is given by: Z * 1.70
(3 RT/M)
Applying the correction
we get the value of % given in the last
column of the tablB; the average of the eight runs is 0.754 10“^
poise, with an average deviation of approximately
1 % ,
Molecular diameters may be calculated from viscosity data,
provided such data are obtained at two different temperatures, so
as to permit a determination of the Sutherland constant. Since this
work was done all at one temperature It was not possible to obtain
directly a value for this constant. As the Sutherland constant is
really a measure of the attractive forces between the molecules, it
should have very nearly the same value for azo methane as for any
other molecule having a similar strusture and a molecular weight
not too far different. Such a molecule would be ^-butylene,which
has a nearly identical molecular weight, the only structural diff­
erence being the presence of a HC=-CH in place of a N=N. It^Ls to be
expected that the force fields surrounding these two molecules
should be nearly the same. That the two molecules do have similar
attractive forces is evidenced by the fact that their boiling points
are nearly identical, both boiling at approximately 1° C ,
The Sutler-
land constant for p-butylene has been determined by Titani
, who
gives it as 362. If we accept this as the Sutherland constant for
azomethane as well, we can calculate the molecular diameter from
the kinetic theory equation:
L '
. 0.499p’cLo, where 1, - 1 _______
where^densityof the gas; c""= average velocity of molecules at temp(34)
erature,T, and C = Sutherland’s constant; 3^ is the mean free path
on the basis of non-attracting spheres; and n is the number of
molecules per cc; e* is the molecular diameter. The above data gives
us a value for ^of 4.54 A. A value slightly less than for ^-butyl­
ene is to be expected. Titani gives the molecular diameter of this
compound as 4.58 A, showing that the value of 4.54 1 cannot be in
serious error. It is interesting to note that Kassel^kas used the
value 3.98 A° xxx as the molecular diameter for collisional deact­
ivation in azomethane decomposition. This, of course, need not be
the same as the diameter obtained from viscosity measurements.
The experimental results can best be studied by an examination
of the data at each temperature and the contrasting of the results
obtained at the different temperatures.
AT 27 ° C :
The data for the seven runs made at this temperature are tab­
ulated below. All runs were for a period of forty five minutes.
0. 0
0.0 0
r 2.18
VH» 2 /V az
Here t is the temperature of the thermostat; c is the current
going through the primary circuit of the transformer (to be divided
by twenty to get the arc current) in amperes;
is the cc. of azo­
methane passing through in the given time at the pressure, P
j Vaz
is the cc. of hydrogen (molectilar plus atomic) passing through at
the combined pressure,
the pressure of atomic hydrogen
entering the reaction vessel, as calculated from the gauge reading?;
VH 2 /V a 2 is the ratio of the amount of hydrogen (total) to that of
represents the cc. of prdduct which can be drawn off
from the first condensing trap at dry ice tenperatures, using a
Toepler pump; V_ is the cc. of product which was condensed out in
the second trap, that containing the silica gel at liquid nitrogen
Run #1 was made with the arc not in operation, i.e., with
molecular hydrogen only. The purpose of this run was to see if any
reaction occurs between azomethane and molecular hydrogen a£ room
temperature and also to determine whether the silica gel adsorbed
any appreciable amounts of hydrogen. As will be seen from the above
table, only about 0.3 cc. was so adsorbed. THe volume of product
in the first trap agrees almost exactly with the amount of azo­
methane originally passed through. When it was found that this gas
was completely soluble in dilute sulfuric acid, it was assumed that
no reaction had taken place and that this gas was all azomethane.
When the results of experiments 2 and 3 gave evidence of the form­
ation of dimethylhydrazine among the products, run
# 1
was duplic­
ated in #4. In this case the product was dissolved in EfeSO^ and
dimethylhydrazine tested for by the addition of cupric chloride
solution, which forms a characteristic red precipitate with secon­
dary hydrazines. No trace of such precipitate was found. This in­
dicates that in the presence of molecular hydrogen alone no reaction
takes place. In both runs practically no adsorption of hydrogen
had taken place in the silica gel trap, showing that a sharp sep­
aration of hydrogen from the hydrocarbons would be possible by
this means.
In all the other rims the arc was on; the pressure of atomic
hydrogen is given in the preceding table (roughly constant at
0,15mm). The most noticeable feature on examination of this table
is that very little of the product is gaseous. There is an almost
conplete absence of product volatile at dry ice temperature from
the first trap. The product observed was a colorless liquid, hav­
ing little if any vapor pressure at -78 °C. This material was defin­
itely basic as it reacted vigorously and exothermally with HC1 sol­
ution. Furthermore this liquid was a good reducing agent, precip­
itating silver from Tollens 1 reagent but giving a positive test
with Fehling's solution only upon heating. Under the conditions of
the experiment the only basic reducing material which could have
been formed was a hydrazine; since primary hydrazines reduce Fehling
solution easily in the cold, the hydrazine must have been a secon­
dary one, probably symmetrical dimethylhydrazine. The test with buff­
ered cupric chloride solution gave in this case the characteristic
precipitate of azo compound- cuprous chloride given by secondary
The slight amount of product which could be drawn off by the
Toepler pump at dry ice temperature might have been due to a slight
vapor pressure of dimethylhydrazine or to a slight amount of azo­
methane left undecomposed or possibly to a small amount of methylamine
formed by the rupture of the hydrazine. In experiment
the whole
volume of the liquid was distilled over to a trap cooled by liquid
nitrogen and removed for analysis without any of the product being
drawn off by the Toepler pump. Half of this product was analyzed
qualitatively for primary amines, using the sensitive Rimini test
which consists of adding one cc. of acetone and a few drops of a
dilute sodium nitroprusside solution to a few drops of the unknown
in five cc. of water. file presence of a primary amine would be
indicated by the formation of a pink color. No such color developed;
it was therefore assumed that no amine was present although it may
well be that the amount present was too small to be detected, as
the control gave a faint pink itself. In any case, if it were really
methylamlne, it represents a negligible portion of the reaction,
1 % ,
As will be seen from the bottom rov; of the table, practically
was isolated from the silica gel trap, indicating the ab­
sence of ethane or methane among the products. The volumes of the
last three runs, amounting to about
cc., were accumulated and
analyzed by passing over hot copper oxide, the results showed Jrhat
practically all of it was hydrogen, only about 0.4 cc. remaining.
The results of the first few runs had shown that the only
product which could be identified was a secondary hydrazine and
that practically complete conversion of the azomethane to this hyd­
razine took place. It was thought desirable to establish definitely
that the product was s-dimethylhydrazine. The amounts of material
available precluded any other method of investigation than micro­
analysis. It was attempted in run #5 to isolate the
hydrazine as
the hydrochloride, but this compound proved to be very hygroscopic
and too difficult to work with in these small quantities. Deter­
mination of the equivalent weight of the hydrazine was likewise
rendered impractical by the excessively hygrosiippic nature of
hydrazines. It was finally decided to prepare organic derivatives
of the hydrazine and to compare their percentage composition with
the theoretical for s-dimethylhydrazine. From experiment
hydrazine was precipitated as the picrate and from #7 as the oxalate*
Both of these compounds were recrystallized from alcohol. Melting
points were taken and
% > K
ddtermined by micro-Dumas analyses. THe
results are given below and compared with the data given by BeilS ', (
stein for the melting points of the s-dimethylhydrazine derivatives.
$N 2
Theoretical 147
m *p *
It will be seen that the above results show conclusively that the
product was s-dimethylhydrazine, uncontarainated by any appreciable
amounts of any other products.
The first thing to be noticed is that no break of the carbonnitrogen link took place. The maximum estimate for the extent of this
reaction is 0.1$. It must be considered whether this is to be attrib­
uted to the strength of the C-N bond or to the occurrence of a much
faster reaction removing the atomic hydrogen. We know that the C-N
bond ic_10-20 kcal. weaker than the C-C bond in ethane; yet ethane
is split to the extent of some
$ by atomic hydrogen at room temp-
erature, so that it is not probable that the absence of hydrocarbon
product is to be attributed to the strength of the carbon- nitrogen
bond but rather to the presence of a reaction of much lower activ­
ation energy. That this is the case is strengthened by the obser­
vation that practically all of the azomethane has reacted. If we
consider the small amount of product volatile at -78°C as azometh­
ane, the percentage reaction would be 97-98$; if raethylamine or
It is very interesting that the predominant, nearly the only,
reaction occurring is the hydrogen addition to the azo linkage.
Since two experiments have shown that azomethane does not react
with molecular hydrogen at room temperature, the reaction must go
by either a termolecular collision involving at least one hydrogen
h jt
atom or by a series of bimolecular steps involving free radicals.
Termolecular collisions are quite infrequent at these low pressures
and it is unlikely that such a large percentage conversion could be
accounted for by termolecular collisions. The probable mechanism
for the reaction is:
CH5 -N=N-CH3x + h
-> ch 3 -nh-n-ch3
followed by
CH3 -WH-N-CH3 + H — >CHg-KH-NH-CH3
(2 )
ch3 -nh-n-ce3 + H 2 -'>CH5 -NH-NE-CH3 «■H
The last equation must be considered as an examination of the
ratio of atomic hydrogen to azomethane shows it to be always less
than -unity (0.5, approximately); yet it would have been necessary
for the ratio to have been at least two for the reaction to occur
entirely by steps involving atomic hydrogen only. This last equation
gives us the possibility of a chain reaction and sufficient hydrogen
atoms to explain the extent of reaction. It is probable, considering
the relative concentrations of atomic and molecular hydrogen, that
the reaction occurs via step (3) to the practical exclusion of
step (2). Reaction (3) must be examined to see if it is energetic­
ally possible. It involves the rupture of an H-H bond, which is
about 102 kcal.; the new N-H bond must be of at least that magnitt evidence as to bond strength in hydrazines,
but Imanishi
that the continuum begins at ~'2200 A for
hydrazine itself and Wenner and Beckman
have shown that in this
region hydrazine is decomposed photochemically into NgHg and H. If
we consider 2200 A as the dissociation threshold, it gives a value
of ^ 1 2 8 kcal., sufficiently high to make the reaction exothermal.
Of course, this is a maximum and the bond strength may be less
than 128 kcal. but it at least shows that the reaction is possible.
Reaction (1), as well as (2), may occur as termolecular proc­
esses, although it is much more likely that they are bimolecular.
It has been shown recently by Kimball
that for complicated mole­
cules, bimolecular association reactions are quite possible, the
life time of the associated molecule being long enough for stab­
ilization to occur. This., is especially true of those cases involv­
ing what he calls an asymmetrical method of approach, such as we
have in these ease's. While Kassel
has shown that Kimball's cal­
culations are wrong by several powers of ten, still Bven on this
basis the lifetimes of the molecules formed in steps (1 ) and (2 )
would be long enough for stabilization.
It is interesting to compare this association of hydrogen
to azomethane with that of methyl radicals as postulated by Davis,
Burton, AhcL Taylor22. It is quite evident that the hydrogen reaction
has a much lower activation energy than the corresponding methyl
addition, inasmuch as the latter occurs only above 163°C, while
the hydrogen addition occurs rapidly at room temperature. For a
reaction occurring at room temperature to such an extent as this
does, the maximum activation energy which can be assigned is about
3-4 kcal. It seems that the likely assumption to make is that
step (1 ) req^iires practically no activation energy and that the
activation energy, if any, is to be assigned to step (3 ).
These results, showing the hydrogenation of azomethane to
dimethylhydrazine and the reaction stopping at this stage, contra46
diet to some extent the work of Dixon
on the reaction of atomic
hydrogen and hydrazine. He found that at room temperature hydrazine
was considerably (up to 80$) decomposed to give ammonia. Yet in
this research practically none of the corresponding methylamine
was fovind at 27°C. It may be that the difference is due to the
somewhat dissimilar conditions employed. Dixon used hydrogen atom
concentrations ranging from 3-10 times the concentration of the
hydrazine, while in this research the ratio of atomic hydrogen to
azomethane averaged 0.5. Also, he xxsed ’'moist’' hydrogen; it may
well be that his reaction was initiated by the dissociation products
of the water vapor.
Dixon discusses this possibility himself, but
does not think that such is the case; however, it is well known
that atomic hydrogen reactions often go quite differently according
to the presence or absence of water vapor. In conclusion, it ajee
may be stated that Steiner^, in his work on the formation of hyd­
razine from N and Hg, states that atomic hydrogen up to 10$ has no
effect on the hydrazine yield, indicating that atomic hydrogen does
not attack hydrazine unless the hydrogen is in great excess (agree­
ing with this research in the absence of reaction) or that, as
Dixon supposes, there are two opposing reactions in Steiner’s case
which result in a net constancy of the hydrazine yield. His reasons,
however, on this point do not carry much weight. Finally, what is
true for hydrazine need not be necessarily so for its derivatives.
AT 110°C:
The experimental results may be summarized as in the table
2 12.0
0.78 C:.00.80
0 .2 1
0 . 2 0
vC 2 H 6
VH 2
ph 2
VH 2 /Vaz 2 *35
0.2 2
Most of the symbols have already been explain; the remainder
are self-explanatory. In the runs at this temperature, contrary to
those at 27°C, appreciable amounts of hydrocarbons were isolated in
the silica gel trap. They were analyzed by the usual catalytic com­
bustion method of gas analysis, allowance being made for the nitro­
gen present in tank oxygen and for the 0.4 cc. of hydrogen which
previous results had shown to be the amount adsorbed during the
course of a run. The results showed that the hydrocarbon was chiefly
methane with small amounts of ethane also present, the ratio of the
two being about ten to one. Tests for the presence of dnsaturated
compounds proved negative.
At 27°C the first liquid nitrogen trap was found to contain
only s-dimethylhydrazine. At this higher temperature the hulk of
the product was still non-volatile at -78°C hut considerable amounts
of a gas could he drawn off hy a Toepler pump at this low temperature.
This gas on combustion gave nearly equal amounts of carbon and
nitrogen; from the volume of C02 formed and from the residual Ng,as
well as from the oxygen consumed, it was deduced that the formula
was close to CHgN. Experimentally the carbon was about 2-3 $ too
low for such a formula and the hydrogen a little too high. This gas
was completely soluble in dilute acid on the first pass through the
absorption pipette. It would seem that the only oompound answering
such a description is monomethylamine, possibly contaminated by a
little ammonia. Bubbling some of this gas through the appropriate
solution gave a positive test for a primary amine, as previously
described. No residual azomethane was present as this would have
caused the empirical formula to show more carbon and nitrogen than
is experimentally found to be the case; the same is true for dimethylhydrazine.
The experimental results are shown graphically
on the foll­
owing page (Pig. 1), where the cc. of product per cc. of azomethane
method of plotting has been
has been plotted
chosen as it gives almost directly the percent reacting in the
various different fashions. As might be expected, the amount of
methane and methylamine increases with increase in the hydrogen azomethane ratio; it is difficult to state the effect of changing
the ratio on the ethane yield as so little is produced but it seems
as though a slight decrease in the ratio takes place with increase
in the ratio.
The residual liquid in the first trap was analyzed in much the
same manner as in the runs at 27 C, by means of micro-Dumas deter­
minations on the purified oxalate derivative. Here again the anal­
yses checked within experimental error with the theoretical value
for dime thy lhydrazine, showing this still to be the only non-vol­
atile constituent at this temperature in this fraction.
The first and most outstanding difference between the results
at this temperature
and those at the lower temperature is that we
now find appreciable amounts of hydrocarbons among the products,
v/hereas at 27 C their presence was problematical. This indicates
that the C-N bond in azomethane is no longer stable to atomic hydrogen at 110°C and breaks giving methane. It is very unlikely that
the methane is formed from the reaction of dimethylhydrazine or
methylamine with atomic hydrogen, as this should give ammonia or
hydrazine as one of the products. Analysis, however, showed that
only about 2-3% of the gas could have been ammonia, and this small
amount is probably formed by a different reaction to be discussed
below. The methane is to be attributed, then, to a reaction of the
azomethane itself; indeed it is necessary to postulate this to ex­
plain the small amounts of ethane formed which must have come from
the other fragment of the azomethane molecule. We know that at 110 C
azomethane is thermally stable and we should expect no ethan^from a
purely thermal decomposition.
A mechanism which seems plausible and which can explain the
experimental results is the one now offered:
First, the bond splitting step to yield methane and a radical:
(ch3 )2 n2 + H
ch4 + ch3 -n*si-
followed by the practically immediate rupture of this radical:
ch3 -n=n-
ch3 + n 2
(2 )
in turn followed by either:
CH3 + CH3
+ H
-j? c2 h 6
f)f these last three reactions, (3) and (4) must occur as termolecular processes if they occur at all. While reaction (5} is also
apparently termolecular it probably occurs as a series of two body
collisions involving the
CH3 + H2
CHg intermediate postulated by Byring
CH3 -H-H +CH3 _ 7 2CH4
Taylor and Burton have recently shown that the reaction between
methyls and hydrogen must go according to step (5) above rather
CHg +Hg
C I ^ + H , as previously supposed. It
is this reaction, rather than step (4), which is important as re­
gards the large CH4 production.
If steps (3) and (5) occurred with equal speeds, then the ratio
of methane to ethaneshould have been three to one. That it is more
nearly ten to one must be due to the more frequent occ^crrence of
reaction (5) as compared to (3). This predominance may be explained
from several viewpoints: (a) the small concentration of methyl rad­
icals, giving btit a small probability of a ternary collision Involv­
ing two methyls; it Is quite possible that all the ethane is formed
only by a wall reactions involving the two methyls; (b) the smaller
activation energy of (5) as compared to (3); the activation energy
of (5) has been given^is 9 kcal. while that of the two methyl re­
combination is probably quite high; (c) the possibility that some
of the ethane formed might have been broken down to methane by
the atomic hydrogen, a reaction shown to go at room temperature
by Trenner, Morikawa, and Taylor
. All this might explain why
reaction (5) goes roiighly five times as fast as reaction (3).
It is improbable that the ethane is formed by the reaction:
CHg +(CH3jgH2
^ OgHg + Ng + OEg
This reaction has been shown to have a high activation energy, at
least 20 kcal.
, and therefore not likely to occur at 110 C. Also
the possibility of association of methyl radicals to azomethane
is to be excluded as the microanalyses gave theoretical results
for dimethylhydrazine uncontaminated by other material having a
different nitrogen content.
If we assume that this mechanism represents the sole means of
formation of ethane and methane, it is possiblejto calculate an
approximate value for the activation energy ofjstep (1). This assump­
tion involves the hypothesis that tfeeir the amount of azomethane
decomposing by this mechanism is given by the sum of the ethane
and one half ofjthe methane produced. We may employ for the calcuil•■T?/r t
ation the kinetic equation, v = sZe
, where v i s the specific
reaction rate, Z is the frequency of collision, s is the steric
factor, and E is the activation energy. Taking the data of experi­
ment (15) as typical, we have thejfollowing: CH ' 5 .65»10_^mole/cc.j
* 2.38»10“®mole/cc.; we take the collision diameter,
for H
as 2.14*10 cm ;for azomethane, there appears to be some doubt
as to what value to take for the collision diameter. K a s s e l ^ has
used the value, 3.98 A; the results of this research give it as
4 .54 A*5
, as obtained from viscosity data. Using Kassel’s value we
obtain a collision frequency of
value of 4.54
.2 2 ^mole/cc./sec. j using the
we get Z - 8.18*10 mole/cc./sec. From the amounts
of ethane and methane produced we calculate a reaction rate, v =?
2.74* 10”^mole/cc./sec. Assuming a steric factor of 0.1, as has
been done in similar reactions involving atomic hydrogen, we obtain
for the energy of activation the value of 7.8 kcal. using Kassel's
value for^j. and|b. 2 kcal using the value obtained from the viscosity
data. The difference between the two is quite within the experimental
Both of these values are several kilocalories highervthan the
5.1 kcal. reported b£ Patat
for this reaction, based on the ortho­
para hydrogen conversion. The difference is not too great when it
is remembered that the experimental techniques were so different
from those employed in this case. Eyrthermore, no gas analyses were
done to see if the reaction actually followed the mechanism proposed
■'Y-j i " A
by Patat, and it^well be that the value of 5.1 kcal. is really an
average activation energy for several reactions which might have
been occurring simultaneously and not necessarily for the reaction
to yield methane alone. Thus, a relatively small amount of the
hydrogenation reaction to give dime thylhydrazine, which we have
seen occurs with a very low activation energy, would have been
sufficient to have reduced the net activation energy to 5.1 kcal.
To return to the product volatile at -78°C from the first
liquid nitrogen trap, which was shown to be monomethylamine. This
compound must obviously have come from decomposition of the s-dimethylhydrazine. The problem is to determine the mechanism of the
formation of the amine. Dixon
has studied the reaction of hyd­
razine itself with atomic hydrogen; the mechanism which he favors
is :
^2 ^ 4 + ^
followed by
N2 H 3
^2 ^ 3 +
NE 3
+ Ng
This series of steps is favored by Dixon over the alternative
possibility suggested by gedye and Rideal s
+ H
2 NHg
- *
—7 NH3 + NHg
Ng + 2Hg
because the work of Wiig and Kistiakowsky
showed that the quantum
yield had no temperature coefficient, as step (9) would require.
Also their work would exclude such a reaction as
1®2 + H2
NH 3 + H
as this in connection with step (8 ) would give a chain mechanism,
in contradiction to their observed value of about unity and also
, *)
to the data of Dixon, who showed that only one mole of NH 3 was
formed per mole of N 2 H 4 decomposed.
It is doubtful, however, if one has the right to assume that
Dixon*s mechanism represents the course of the reaction for dimethylhydrazine at temperatures above 100°C. Indeed, Wiig and
Kistiakowsky state that the reactions (8 ) and (10) would probably
become more predominant as the temperature is raised. The data of
this research seem to show that the primary step for the dimethylhydrazine reaction is the analogue of (8 ) above,
+ H
— > CH3 NH 2 + CH3NH
and is practically the only one occurring. Had the reaction gone
it is
CH 3 -NH-NH-CH3 + H
difficult to see how the
(12) (p.f/
formation of appreciableamounts
of dimethylamine could have been avoided. However, no evidence at
all for the formation of this compound was found, the results of
the gas analyses pointing definitely to monomethylamine. The ab­
sence of dimethylamine shows that reaction (1 1 ) must have been the
initial step'. This is, indeed, not surprising since the N-N bond
strength is considerably less than the N-C or N-H bond strengths.
W h e th e r t h e
a n a lo g u e o f
s te p
(9 )
fo llo w s
s te p
(1 1 )
v e r y d o u b t-
ful. This reaction would also require the formation of dimethyl-
li ^
amine and also of ethane (of which only a few cc. were found). It
is probable that a reaction similar to reaction (1 0 ) is the step
following (1 1 ), i.e.,
CHjgNH + Hg
CH3NH2 + H
(1 3 )
This reaction Is possible as it is but very slightly endothermal,
the value for the N-H bond In ammonia being given as 96 kcal. by
Dixon and Steiner . On the other hand, the reaction,
— 7 CHgNHg
either as a bimolecular or termolecular process is quite possible
and may be the step following (1 1 ) instead of the reaction postulated
just above. Whether the above postulated mechanism Is correct,or
whether Dixon’s is could be decided by a determination of the Ng
yield and also by seeing if one or two moles of amine are formed
per mole of hydrazine decomposed. This was not possible with the
experimental setup employed in this research, but at least the
mechanism suggested above is in accord with the experimental facts
as presented above.
The small amounts of ammonia observed among the reaction prod­
ucts is probably formed by some side reaction involving one of the
radicals discussed above. It is useless, however, to hypothesize
further regarding this as there would be no means of justifying any
assumptions made, although it might have been formed by splitting
of the C-N bond In the dimethylhydrazine, yielding subsequently
NH . If this is so then the value of E for step (1) would be a
little too low as some methane would be formed In this manner.
As has been mentioned before, the mechanism which has been
suggested for the dimethylhydrazine-hygrogen reaction gives two
methylamines per dimethylhydrazine decomposed. On this basis it
is possible to make a rather rough calculation of the activation
energy of reaction (11). The calculation is at best approximate,
as no accurate value can be given for the concentration of the
dimethylhydrazine. If this is taken somewhat arbitrarily as 0.85
of the original azomethane concentration, the calculation is then
possible. Taking again the data of run #15 and using the same
value for
as for azomethane, we obtain in the same manner as
described previously for reaction (1) the value of E for step (11)
as 8.3 kcal, again using a steric factor of 0.1. This value of E
is of necessity approximate, on account of the assumptions made,
and is probably not more accurate than
f t o
AT 195°Ci
The experimental results are tabulated below using the same
symbols as previously ••
0 . 0
0 . 2 0
0 . 2 1
/v 2az* 5 6
H p'
H 2
P h 2
VC H 0.5
It will be seen that several differences exist between the
results at this temperature and those at 110°C. The yield of methylamine, VL, is nearly twice that obtained undersimilar conditions
1 1 0
; analysis showed that the gas was still practically entirely
CH^NHg. On the other hand, the yield of ethane has decreased to
practically zero and is almost within the experimental error for
the analyses. The amount of methane, although larger than at 110^
is not appreciably so, being only abo-pt 15-20$ greater.
The major difference is to be found in the fraction which is
non-volatile at -78°C. At the two lower temperatures this was found
to be only dimethy lhydrazine, within the experimental error, as
determined by melting point and micro-Dumas analyses. At this
temperature, however, such is not the case. Neither the melting
points nor the nitrogen determinations gave theoretical results
for the di-substituted hydrazine. As before, the determinations
made on the oxalate derivative on the fractions from mans
(18) through (25). Fractions from successive runs were combined
so that in all three sets of determinations were made. The melting
point of the recrystallized compound no longer was sharp but covered
a range of some five degrees, and the melting point itself was some
twelve degrees below the 132°given by Beilstein for the oxalate deriv­
ative. The nitrogen determinations were in all three cases too lpw
for dimethylhydrazine; the results ar^given below and compared with
the value for the dimethylhydrazine derivative:
18.67% Ng
Micro-Dumas 1
The difference between these results and the value for dimethylhydrazine is too great to be attributed to experimental error, which
should be no more than 0.1-u.2/o. When this is oaken in conjunction
with the lowered melting point the auppoaioion mxso'oe that# '.76 now
have something in the non-volatile fraction besides dimethylhydrazine.
Experiment (17) was made with molecular hydrAgen only, the
arc being off, the object being to see if molecular hydrogen reacts
with azomethane at 195°C. As will be seen from the table, as much
gas was obtained as the amount of azomethane originally passed throu^j
No hydrocarbons were found, nor was there any non-volatile constit­
uent at -78^ This shows that at 195"there is still no reaction of
molecular hydrogen and azomethane. Eests for unsaturat&rs proved
negative in all cases.
The data shows that the yield of methylamine is nearly
doubled over the temperature range from 110*to 195; this is to
be expected from the temperature coefficient of the reaction. That
the product was still almost entirely monomethylamine shows that
the reaction must still be going by the same mechanism as at the
lower temperature. This also shows the absence of any appreciable
reaction between methylamine and hydrogen atoms as otherwise N H^
would have occurred among the reaction products to a fairly large
extent, which is not the case.
This, however, cannot be the case for the hydrocarbon products*
Whereas it might have been expected that the yield of ethane would
also increase, the reverse was found to be the case.
If we assume
that the ethane is formed from two methyld via a wall reaction, then
another reaction must be occurring which prevents as many methyls
from reaching the walls as previously, i.e., another reaction using
up methyls in the gas phase. A decrease in the accommodation coeff­
icient of the methyl at the higher temperature might also account
to some extent for the decreased ethane yield, although the first
explanation must be considered as the main reason for the diminution.
The methane yield is also somewhat surprising. It might Jiave
been expected that the volume of methane would increase in the same
degree as did the methylamine, especially as the activation energies
for the reactions producing the two compounds are so similar. It
must be that more methyl is being produced at this temperature than
1 1 0
°but that the methyl is being removed by some other reaction
than the one to form methane. Such a reaction may be imagined by
considering addition reactions to azomethane, e.g.,
(0H3 )2 H 2 + 0HS -»■ JfflEfeJafe
followed by either:
(CH^JgNg + CHg
-?• (CHgJ^Hg, tetramethy lhydrazine
(CH^)3 N 2 + H
(CH3 )3 NgH, trimethylhydrazine
Each of these reactions would serve as a methyl removing step and
probably occur with a small activation energy.
Evidence for the
existence of such compounds may be found in the results of the
micro-Dumas analyses, which showed a lower nitrogen percentage than
that to be attributed to dimethylhydrazine alone, the inference
being that we now have some other compounds in the dimethylhydrazine
£xajE±±ffiHUckkxx fraction besides dimethylhydrazine. Both of the
compounds mentioned above have lower nitrogen contents than the
dimethylhydrazine, so that their presence would explain the low
results in the Dumas determinations.
As to whether the compound
present is the tri- or tetra-methylhydrazine it is impossible to
say; in all probability it is a mixture of all three hydrazines.
It is interesting to note that the existence of these com~
pounds had been postulated by Burton, Davis, and Taylor
21 22
* ,in
the photolysis of azomethane. They reported that the yield of these
compounds became noticeable only above 165 C, in agreement with
the results of this research. It may be, however, that small amounts
were formed at
but were lost in the subsequent recrystalliz-
Calculations based on the nitrogen percentages of the
various hydrazines and on the experimental data showed that if the
mixture were di- and tri-methyl only that the amount of trimethyl
' — • «!
would have to be about 3/8 of the total liqgtid fraction, corres­
ponding roughly to 20-25 cc.of gas, approximately equal to the
methane yield. If it were all di- and tetra-methyl, then only
about 1/5 of the liquid would have to tetramethylhydrazine to
explain the experimental data. This would correspond to about 10 cc.
of gas, about 1/2-1/3 the methane yield. If we assume that, like
the methylamine formation, ’the yield of methane at 195°would have
been twice that at
° in the absence of any other methyl removing
step, then it would seem from the above that the bulk of the assoc­
iation reaction mast have gone to form trimethylhydrazine.
The possibility of the formation of these compounds still does
not adequately explain why the methane yield did not increase more
than was observed. It does not seem logical that practically all of
the methyls produced in excess over those produced at HO^should
have gone to the association reactions and so little to increased
methane formation. It may well be that more methane is initially
produced but that some of it may be subsequently removed by some
reaction, such as
CH^ + H
-? CH 3 + Hg,
activation energy of about 15 kcal.
a feaction having an
In such a case the methyls
formed could either react again to form methane or they could add
on to azomethane to form compounds such as discussed above. This
might explain why such a large percentage of the methyls go towards
the association reactions, while the methane yield apparently remains
constant, as compared to 110°. If such a reaction as the one pictured
above actually is taking place there should be a temperature at
which the yield of methane should be a maximum, decreasing above
this temperature as this reaction becomes more predominant*
1- A new type of calorimeter has been applied for the first time
to the study of atomic hydrogen reactions in order to determine
atomic hydrogen concentrations.
2- The viscosity of azomethane has been determined to be 0.754*10
poises at 23°C and the molecular diameter to be 4.54*10
3- At 27°C the azomethane- atomic hydrogen reaction results in
addition to the N-N double bond to yield s-dimethylhydrazine. This
reaction has a maximum activation energy of 3-4 kcal.; it is the
only reaction occurring at this temperature.
4- At 110°C we have not only the dimethylhydrazine formation but
rupture of the C-N bond in azomethane to^rield methane plus a little
ethane; also the rupture of the
yield monomethylamine.
bond in dimethylhydrazine to
The activation energies of the last two
reactions has been estimated as 8.0 and 8.3 kcal., respectively.
5- At 195°C we have all the above taking place plus also an add­
ition of methyl radicals to azomethane to yield compounds of the
type of tri- and tetra-methylhydrazine.
Data of a typical run to show how P
Data of run #13
was obtained
P jj- = 0.19 mm
Room temperature
T-]_ at entrance to calorimeter
Tg at exit from calorimeter
Heat loss (from calibration data)
Total ^T
cc. of water passing through calorimeter
Calories given to water via calorimeter per min
Heat of formation of Hg per mole in kcal
100/15 min
cc. of H (atomic) passing through per 45 min
(calculated as 24.8 C)
cc. of total hydrogen (as Hg) passing through per
45 min (from flowmeter data)
Pressure total hydrogen
0.76 .
198.4 4'l/2(56'.vn'
0.76 mm
0.19 mm
1- Thiele, Berichte, 42, 2875, (1909)
2- Ramsperger, J.A.C.S., 49, 912, (1927)
5- Ramsperger, J.A.C.S., 49, 1495, (1927)
4- Hinshelwood, Proc. Roy. Soc., 115A, 230, (1926)
5- Lindemann, Trans* Par. Soe., 17, 598, (1922)
a) Rice
J.A.C.S., 49, 1617, (1927)
50, 617, (1928)
Kassel, J. Phys. Chem., 32, 225, (1928); 52, 1065, (1928)
Ramsperger, J. Phys. Chem.,
34, 669, (1930)
9- Ramspwrger, J.A.C.S., 50, 123, (1928)
10- Porbes, Heidt, and Sickman, J.A,C.S., 57, 1935, (1935)
11- Patat, Naturwissenschaft, 23, 801, (1935)
12- Leermakers, J.A,C.S., 55, 3499, (1933)
Leermakers, J.A.C.S., 55, 4504, (1933)
14- Sickman and Allen, J.A.C.S., 56, 2031, (1934)
15- Sickman and Allen, J.A.C.S., 57, 1385, (1935)
16- Patat, Nachr. Ges. Wiss. Gottingen, Math. Phys. Klasse,
Fachgruppe II (N.F.),
, 77, (1936)
17- Sickman and Rice, J. Chem. Phys., 4, 242, (1936); jsxx
4, 608, (1936)
See 41
19- Emmett and Harkness, J.A.C.S., 54, 538, (1932)
20- Porbes and Heidt, J.A.C.S., ,57, 2331, (1935)
21- Burton, Davis, and Taylor, J.A.C.S., 59, 1038, (1937)
22- Burton, Davis, and Taylor, J.A.C.S., 59, 1989, (1937)
23- Riblettqnd Rubin, 59, J.A.C.S., 1537, (1937)
24- Davis, Jahn, and Burton, J.A.C.S., 60, 10, (1938)
25- Jahn, Thesis, N.Y.XJ., 1939
26- "Organic Syntheses", J. Wiley and Sons, New York, Vol. 16,
1936, p. 18
27- Jahn, J.A.C.S., 59, 1761
27a- Wood, Phil. Mag., 42, 729, (1921)
28- Amdur, J.A.C.S., 57, 856, (1935)
29- Smallwood, J.A.C.S., 56, 1542, (1934)
30- Bonhoeffer and Harteck, Zeits. f. Phys. Chem.,A159, (64',:-Vi( 1928)
see also 36c
31- Steacie, J. Chem. Phys., 4, 461
32- McGavack and Patrick, J.A.C.S., 42, 946, (1920)
33- Harteck, Zeits. f. Phys. Chem., A159, 98, (1928)
34- Wrede, Zeits. Phys., 54, 53, (1929)
35- Knudsen, Arm. Phys. Leipzig, 28, 71, (1909)
36- a Bonhoeffer, Zeits. f. Phys. Chem., A113, 199, (1924)
Bichowsky and Copeland, J.A.C.S., 50, 1315, (1928)
c v. Wartenburg and Schultze, Zeits. f. Phijjte. Chem., B 6 , 261, (193
d see 46
e Smallwood, J.A.C.S., 51, 1985, (1929)
f Poole, Proc. Hoy. Soc., A165, 404, (1937)
37- Taylor’s Treatise on Physical Chemistry, p. 175
38- Chapman, Phil. Trans.,217, 115, (1917)
39- Titani, Bull. Chem. Soc. Japan., ^,310, (1930)
40- Kassel, "Kinetics of Homogeneous Gaseous Reactions", (A.C.S2
monograph) 1932, p. 195
41- Trenner, Morikawa, and Taylor, J. Chem. Phys., 5, 203, (1937)
42- Imanlshi, Sci. Papers, Inst. Phys. Chem. Research (Tokyo),
15, 166, (1931)
439 Wenner and Beckman, J.A.C.S., 54, 2787, (1932)
44- Kimball, J. Chem. Phys.,5, 310, (1937)
45- Kassel, ibid., j£, 922, (1937)
46- Dixon, J.A.C.S., 54, 4262, (1932)
47- Steiner, Zeits. Elektrochem., 36, 807, (1930)
48- Gorin, Kauzmann, Walter, and Eyring, J. Chem. Phys.,
49- Taylor and Burton, ibid., 8, 675, (1939)
50- Harteck, Zeits. f. Phys. Chem., A139, 98, (1928)
51- Gedye and Rideal, J. Chem. Soc. (London), 1160, (1932)
52- Wiig and Kistiakowsky, J.A.C.S., 54, 1806, (1932)
53- Dixon and Steiner, Zeits. f. Phys. Chem., B17, 327, (1932)
ly tB S A R Y
N. Y. UmtyJ
Без категории
Размер файла
3 153 Кб
Пожаловаться на содержимое документа