HYDROGEN ISOTOPE EFFECT IN T H E AMINATION O F

CHLOROBENZENE BY SODAMIDE1

ABSTRACT
WIixt~rresof chlorobenzene and chloroben~ene-2-~Hhave been subjected t o partial amination
by sodamide in liquid ammonia and both the unreacted starting material a n d the product
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

aniline have been analyzed for deuterium. Deuterium in the aniline is distributed approxi-
mately equally between the ortho a n d lneta positions. The results give strong support t o
the niecha~lisrnproposed by Roberts and co-workers in which the slow step is the formation
of an intermediate, such as benzyne, which is symnietrical with respect t o carbon atoms 1
and 2.

The amination of aromatic halides by alkali metal amides in liquid ammonia was
discovered in 1938 by Bergstrom and co-worlcers ( I ) , who were able to convert chloro-,
bromo-, and iodo-benzene into aniline and a mixture of by-products containing diphenyl-
amine, triphenylainine, and p-aminobiphenyl. Fluorobenzene did not react. In 1945
Gilman and Avalciail reported that a number of ortho-halogenated ethers on treatment
by Bergstro~n'sprocedure gave ineta-aminoethers (2) as illustrated by the equation
Can. J. Chem. 1961.39:180-191.

Bunnett and Zahler reviewed the reaction in 1951 (3) by which time aminations with
rearrangement under the Bergstrom conditions had been observed with halogen sub-
stituents located ortho and para to ether, sulphide, sulphone, and trifluoroinethyl
groups. I t had also been clemonstrated, largely by Uenlceser and co-worl<ers (4), that the
entering amide displaces a hydrogen atom ortho to the halogen atom. Bunnett and
Zahler proposed a mechanism for the rearrangement in which attaclc by amide ion on
carbon atom 2 of a halobenzene forces the migration of a hydride ion from carbon 2 to
carbon I and the displacement of a halide ion by this hydride ion. T h e sequence of
reactions in equation I below illustrates this mechanisin.

In 1953 Roberts aiicl co-worlters ( 5 ) showecl that rearrangement occurs even in u11-
substitutecl halobenzenes by demonstrating that cl~lorobenzene-1-'-'Cwith potassium
amide in liquid ammonia gives aniline-1-14C ancl aniline-2-14C in alillost equal amounts.
The near equality of rearranged and unrearranged product suggested the participation
of an intermediate symmetrical with respect to carbons 1 and 2. They coilsidered the
most lilrely intermediate to be benzyne, which would enter into the mechanism as shoivn
in sequence 11.
IManuscript receGed Jzily 11, 1960.-
Contribzttion front the Cltelnistry Departn~entof the University of iManitoba, Winnipeg, ilIanitoba.
?Holder of a National Research Coz~ncilbursary, 1955-56.
Can. J. Cbem. Vol. 39 (1961)
 DUNK ET AL.: AMINATION O F CI-ILOROBENZENE BY SOD-AMIDE
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

The existence of a symmetrical internlediate was further indicated when Roberts and
co-worlters (6) reported in 1956 that chlorobenzene-1-'"C, bromobenzene-1-"C, and
iodobenzene-1-IT with potassium ainide all give aniline-1-'T and anilii1e-2-~T in
approximately equal quantities. They also studied the reaction of ortlzo-deuterated
fluoro-, chloro-, and bronzo-benzene by anlinating inixtures of deuterated and undeuter-
ated halobenzenes and comparing the percentage of deuterium before reaction with that
after partial reaction. I11 deuterated broillobenzene the percentage of deuterium increased
with the percentage of reaction indicating that the undeuterated material reacts faster,
Can. J. Chem. 1961.39:180-191.

and the isotope effect, i = k,/k,, was found to be about 5.5. Fluorobenzene-2-2H was
found to exchange deuterium with the solvent very rapidly. Dilute solutions of chloro-
benzei1e-2-~H in chlorobenzene showed a small increase in deuterium content after
partial reaction while concentrated solutions showed a decrease. This was talten to nlean
that cl1lorobei~zei~e-2-~H exchanges cleuteriunz with the solvent a t a rate comparable
with the amination. On the basis of these and other observations the mechanisn~ I11
was postulated for the amination and exchange of chlorobei1zene-2-~H.

Before the appearance of Roberts' second paper the authors had performed some
experiments intended to distinguish between n~echanisnlsI and TI by finding out whether
or not deuterium is lost from the aromatic ring when chlorobenzene-2-W is aminated
by sodainide in liquid amn~onia.Mechanism I predicts that ~hlorobenzene-2-~H should
 182 CANADIAN JOURNAL OF CHEMISTRY. VOL. 39. 1961

give aniline-2H with the same total deuterium content, while inechanism I1 predicts
ai~iline-~H\\Tit11 a smaller deuterium content than the starting material. T h e results of
these preliininary experiments are shown in Table I.

TABLE I
Atom per cent deuterium in the labelled positions of reactants and products

Chlorobenzene-2-?H Aniline-2H
2,4,6-Tribro-
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

Experiment Initial* Final* Initial* Final* moanilinet

1. Aniline recovered by acid extraction - - 84.7Zk2.01 81.9Zk0.31 -
2. Aniline recovered b y distillation - - 84.7Zk2.01 83.6Zk1.21 -
3. Chlorobenzene and 1.7 moles
sodamide 64.3Zk1.9 - - 35.4Zk0.1 -
4. Chlorobenzene and 1.7 moles
sodamide 64.3h1.9 - - 30.8Zk1.2 17.3
5. Chlorobenzelle and 1.7 ~noles
sodamide 64.3h1.9 - - 30.4Zk0.1 16.3
6. Chlorobenzene and 0.5 mole
sodamide 4 4 . 1 31~. 5 4 4 . 1 31~. 8 - - -
*Mean of three analyses&maximum deviation.
tone analysis.
tT11is material is probably a mixture of 2-. 2 . 4 , and 2,4.6-deuterated aniline but the analyses are calculated o n t h e assumption
that it is aniline-2H.
Can. J. Chem. 1961.39:180-191.

Experiments 1 and 2, Table I , were intended to find out whether or not ring-deuterated
aniline exchanges hydrogen isotopes with the reaction inixture or during recovery and
purification. In experiment 1 a sample of aniline partially deuterated in the ortho and
para positions was treated with sodainide ill liquid aininonia and recovered by Berg-
strom's method (1) which involves extraction of aniline from the reaction inixture with
aqueous hydrochloric acid. I t is seen t h a t soine deuterium was lost by this procedure. In
experiment 2 the use of acid was avoided by extracting the reaction inixture with ether
and recovering aniline by fractional distillation. T h e recovery of aniline was reduced
by almost half b u t Table I shows t h a t the loss of deuterium by this method was
negligible.
A inixture of chlorobenzene and ~hlorobenzene-2-~H was then treated with 1.7 times
the equivalent quantity of sodamide in liquid ammonia and the aniline was recovered.
Experiments 3, 4, and 5 , Table I , show t h a t the aniline produced contained about half
the deuteriuin originally present in the starting material. This aniline was then brominated
and the 2,4,6-tribroinoaniline produced was analyzed for d e u t e r i ~ ~ m
Experiments
. 4 and
5, Table I , show that almost half the deuterium in the aniline was displaced by bromine
so t h a t a little inore than half the deuterated aniline must have had its deuterium in
the ineta position.
These results were talten to favor mechanism I1 since it predicts a loss of deuterium
and the forination of equal quantities of ortho- and ineta-deuterated aniline. However,
if the reaction did take place by nlechanisnl 11, the fact t h a t half the deuterium was
lost would lead to the surprising conclusion t h a t protiuin and deuterium are lost at
equal rates. H further test of this point was inade by reacting a inixture of chlorobenzene
and chloroben~ene-2-~H with half the equivalent quantity of sodainide and recovering
the unreacted chlorobenzene. T h e d a t a from experiinent 6, Table I , show t h a t there
was no change in the deuterium content of the chlorobenzene, which seemed t o confirin
that protiuin and d e ~ ~ t e r i ureact
m a t equal rates.
With the appearance of the second paper by Roberts and co-workers (6) i t was evident
 DUNK ET AL.: A M I N A T I O N O F CHLOROBENZENE BY SODAMIDE 183

that our results could be accounted for by mechanism I11 in a very particular corn-
billation of circumstances. AS the reaction proceeds, the usual isotope effect tends to
increase the deuterium in the unreacted chlorobenzene and the product aniline, but
exchange would tend to decrease it, so that a suitable combination of rate of amination,
rate of exchange, initial concentration of deuterium, and percentage of reaction could
result in a soy0loss of deuterium and no change in the deuterium content of the chloro-
benzene. Since Roberts and co-worlters did not measure the deuteriunl in their aniline
and we did not measure the deuterium in our unreacted chlorobenzene it \bras not possible
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

to determine whether or not the conlbination of factors necessary to fit our data to
mechanism I11 actually existed. Since the fit of Roberts' own data to mechanism I11
was far from precise, it seelned worth while to repeat and extend our experinlents to
include in one set of reactions all the measurements necessary for a test of the mechanism.

EXPERIMENTAL
Chloro benzene-2-TI
Grignard reagent was prepared from 194 g of Eastlnan I<odak white label o-chloro-
iodobenzene (0.814 mole) in a flame-dried flask under an atmosphere of dry nitrogen.
Acid titration of the solution indicated 90-970jo yields. In the preliminary experiments
the Grignard was decomposed by adding 1.63 g of deuterium oxide* (0.814 mole) very
slowly with vigorous stirring. Gilman's color test I (7) was negative a t the end of the
Can. J. Chem. 1961.39:180-191.

addition. T h e solution was then washed with water and dried over anhydrous potassium
carbonate. T h e product was recovered from the ether layer by fractionation through
a 10-cm Vigreaux column. Yields ranged from 45% to 650j0, b.p. 128-132" C a t 742 mm,
D
21.8 = 1.5261. Deuterium analyses averaged about 800jo cl1lorobenzene-2-~H.
Since Roberts and co-workers reported the preparation of pure ~hlorobenzene-2-~H
(without stating the method of analysis) by decoinposing o-chlorophenylmagnesiun~
iodide with deuteroacetic acid, this method was tried. T o the Grignard reagent prepared
as described above was added 112.4 g (1.84 mole) of deuteroacetic acicl slowly with
vigorous stirring. At the end of the addition Gilman's color test I was negative (7). Water
was added to dissolve the magnesium salts and the ether layer was separated. T h e
ethereal solution was washed with sodiuln hydroxide to remove excess acetic acid, dried
over anhydrous calciuin sulphate, and distilled through a n approximately 20-plate
column. The yield of chlorobeilzene was 49.6 g (53.8701, b.p. 130.0-130.5" C a t 745 mm,
nD25 = 1.5209. This material analyzed 81.GOjo ~hlorobenzene-2-~H. Two samples of deutero-
acetic acid were used with identical results; one sample was prepared by hydrolyzing
freshly distilled acetic anhydride with deuterium oxide, the other was obtained from
Merck and was certified by the supplier to contain more than 98% deuteroacetic acid.
Deuterated Aniline
The preparation of aniline-2-2H was attempted by the decarboxylation of deuterated
ailthranilic acid as shown in the following sequence of reactions.

Decarboxylation of aronlatic acids with calcium deuteroxide a t 500" C has been show11
*Obtained from the Comnzercial Products Division of Atonzic Energy of Calzada Ltd., and reported by the
supplier to contain 99.83 atom% deuterium.
 184 CANADIAN JOURNAL O F CHEMISTRY. VOL. 39, 1961

to give a mixture of isomerically deuterated products (9, but it was thought that the
mild conditions possible in the decarboxylation of anthranilic acid might perinit success.
Thirty grams of ailthranilic acid (0.219 mole) was dissolved in 70 ml of dioxane in a
distilling flask and 6.6 g of deuterium oxide (0.330 mole) was added. After the solution
had stood for 160 hours a t room temperature the solvents were reinoved under vacuum.
A condeilser and receiver were attached to the distilling flask and the latter was heated
in an oil bath a t 200-210" C until no more aniline distilled over. This crude aniline was
redistilled to give 16.2 g of product (85% yield), b.p. 179-183" C a t 742 mm, nD21.8
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

= 1.5826. Deuterium analyses of the aniline obtained from two runs, A and B, gave
the atom per cent deuteriuin in product A = 2 1 . 2 f 0.5% and in product B = 12.0fO.l%.
(These analyses would correspond to an aniline-"3 content of 49.41t1.27~ in A and
27.0f 0.2y0 in B.) In order to remove deuterium froin the ailliilo group of the aniline,
it was twice dissolved in about 500 1111 of liquid ainillonia and allowed to stand until
the ammonia had evaporated. Distillatioil gave aniline, b.p. 60-65" C a t 2.5 mm,
nD?'.8= 1.5834, in which the intensity of the N H stretching overtoile a t 1408 inp matched
that of ordinary aniline.* Analysis showed 1 2 . l f 0.3 at.yo deuteriuin in A and 6.66f0.02
at.yo in B. (This would correspoild to an ai1iline-2-~Hcontent of 8 4 . 7 ~ k 2 . 0 7in~ A and
46.6f0.170 in B.)
These results show that the deuterated aililiile obtained by decarboxylatiilg partially
Can. J. Chem. 1961.39:180-191.

deuterated anthranilic acid has a little inore than half its deuteriunl in the ring instead
of one-third as predicted for a statistical distribution, or less than one-third as would
be expected if there \i7ere a normal isotope effect in the decarboxylation. This requires
that the decarboxylation take place by a mechanism such that either deuteriuill migrates
in preference to protium or deuteriuin enters inore than one positioil in the ring. T h e
latter possibility seems much the inore likely, especially in view of the observation (8)
that this actually happens with simple aroinatic acids on decarboxylation with Ca(OD)?.
In order to find out how general the redistribution of deuteriuin is, some of the aniline
froin run B was coilverted to 2,4,6-tribrornoaniline as described below. This tribroino-
aniline contailled no deuteriuin so the most likely hypothesis is that decarboxylation
gives a mixture of aniline-2-TI, aniline-2,4-"2, and ai1iline-2,4,6-~H~.
2,4,6-Tribromoaniline
T o determine the fraction of deuteriunl in the meta positioil of the deuterated aniline
obtained by decarboxylation or by ainination the sample in question was brominated
and the 2,4,6-tribrornoaniline was analyzed for deuteriuin. This inethod has been used
previously by Lauer and Errede (9). About 1.0-1.5 g of deuterated aniline was dissolved
in 20 1111 of water. The solution was cooled in ice and a solutioil of bromine in ethanol
(25% by weight) was added from a burette until the first appearailce of a permanent
yellow tinge. The precipitate was filtered, washed with li~ater,dissolved in alcohol, and
reprecipitated by the addition of water. The inaterial obtained in this way melted
sharply a t 118" C (uncorr.) and was analyzed for deuteriuin as described below.
Amination of Chloroben~ene-2-~I1
In the prelinlinary experiments the procedure of Bergstroin was used. Sodamide was
prepared by the illethod of Vaughan, Vogt, and Nieulvland (10) from 5.2 g of sodium
(0.226 g-atom) in about 600 in1 of liquid ammonia. \Vhen the blue color of the metal
solution had turned to the grey of sodamide, 15 g of ~hlorobenzene-2-~H (0.133 mole)
*Spectral measzirenzents were made on a Beckman spectropkotometer model DK-1.
 DUNN ET AL.: AMINATION O F CHLOROBENZENE BY SODAMIDE 185

was added dropwise over a period of 20 minutes. After another 40 minutes, 11 g of

alnmonium chloride (0.202 mole) was added and the solvent was allowed to evaporate
overnight. Ether was added to the residue and refluxed to drive off residual ammonia.
T h e ether-insoluble salt was dissolved in water and the aniline was extracted from the
aqueous layer with ether. T h e ether solution was dried over anhydrous potassium car-
bonate, and the aniline was recovered by repeated fractionations through a 10-crn
Vigreaux column a t 60-65' C a t 20 inm. Purity of the aniline was checked by com-
parisoil of the intensity of the 1498 rnp peak with that of pure aniline. Yields of pure
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

product averaged about 25% by this inethod of recovery. Bergstroin's method of acid
extraction gave yields of 40% but could not be used because of the danger of exchange
between the acid and ring hydrogen. Deuterium ailalyses on the product were carried
out as described below.
111 later experiments, where it was desired to isolate uilreacted chlorobenzene as well
as aniline, Roberts' procedure was used. T h e chlor0benzene-2-~Hin 150 in1 of liquid
amlnonia was added to the sodamide solutioil over a period of 1 minute. Three ininutes
after the addition was complete the reaction was stopped by the addition of 20 g of
ammonium nitrate (0.250 inole) dissolved in 100 1111 of liquid ammonia. Anhydrous
ether, 150 1111, was added to the reaction mixture and the ammonia mas allowed to
evaporate overnight. Then 200 1111 of water was added and stirred for 10 minutes. T h e
Can. J. Chem. 1961.39:180-191.

ether layer was separated and washed with water. T h e aqueous layer and washings
were boiled to reinove excess ain~noniaand ether, then analyzed for chloride by the
Volharcl method. T h e ethereal solutioil was dried over anhj,drous calcium sulphate,
the11 distilled. Fractionation through a 13-cm Vigreaux column gave aniline, b.p. 55-65' C
a t 4-5 111111, ?zD2j = 1.5836, and chlorobenzene, b.p. 130.0-130.5' C a t 754 mm,
1.3207. The initial separation of aniline and chlorobenzene was most convenient a t
reduced pressure but the final purification of chlorobenzene mas done a t atmospheric
pressure in a 20-plate column. Each substance was shown to be free of the other by its
infrared spectrum. Each was analyzed for deuterium as described below.

Deuterated -4niline under Amination Conditions

In order to determine whether or not aniline loses deuterium after it is produced in
the reaction or during recovery, the first amination procedure was repeated exactly
except that the deuterated aniline obtained from the decarboxylation of anthranilic
acid was used instead of chlorobenzene. T h e results are shown as experiments 1 and 2,
Table I.
Deuterium Analyses
The samples were burned in a standard inacroconlbustion apparatus using a Vycor
tube and a plug of silver wire to trap halogens. T h e water of combustion was collected
i l l a Drj, Ice trap and treated with clean copper mire t o remove halogens, or barium
carbonate, to remove soluble oxides of nitrogen. I t was then purified by the inethod
of Keston, Rittenberg, and Schoenheimer (11) except that calcium oxide and potassiuin
permangallate mere used instead of potassiuill hydroxide and chromium trioxide. T h e
density of the water of combustion was determined by the gradient density tube method
described by Ailfinsen (12). This gives the weight per cent, w, of deuterium oxide in
the water of combustion. Froin this the atom per cent deuteriuin in the labelled positioils
of the molecule was calculated by the equation
atoll1 per cent deuterium in n' positions = (n/nl) . 900w/(1000-w),
 186 CANADIAN JOURNAL OF CHEMISTRY. VOL. 39, 1961

where n' = number of labelled positions in the nlolecule and n = total iluillber of
hydrogen atoms in the unlabelled molecule.
I11 the case of chlorobenzene, combustion analyses were suppleinented by spectro-
photometric analyses using a Perkin-Elmer model 21 spectrophotometer. R4easurernents
were made a t 10.115 p with slit width IGO p and a t 10.555 p with slit width 248 p. These
correspond to spectral slit widths of 0.020 p and 0.028 p respectively. All measureineilts
were made by the base-line technique in 0.025-mm cells. Mixtures of chlorobenzene and
the 81.6% ~hlorobenzene-2-~H described above gave linear Beer's law plots which were
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

used to analyze the chlorobeilzene obtained froin the amination reactions. Table I1 shows
the data for standard mixtures and Table I11 shows those for reaction products.

'TABLE I1
Infrared analyses of known mixtures of chlorobenzene and ~hlorobenzene-2-~I-I

Mole %
chlorobenze1le-2-~H Wavelength lot I$ log Io/I
Can. J. Chem. 1961.39:180-191.

*Determined b y combustion.
t%Transmission at the base line.
i% Transmission at the peak.

Infrared analyses of chlorobenzene - chlorobenzene-2-W mixtures obtained from amination rc;ictions

Mole yo Average nlole %

Run Wavelength lo* It log Io/I o-DC6H4C1I o-DCsH4C1

*yoTransmission at the base line.

t%Transmission at the peak.
$Calculated from log Io/I and the data of Table 11.

RESULTS A N D DISCUSSION
Roberts and co-workers (6) derived a kinetic expression from mechanism I11 to account
for the change in deuterium content of chlorobenzene during ainination. I t will be
necessary to reproduce their derivation in order to extend it to cover the deuterium
content of the aniliile produced in the reaction.
 DUNN ET AL.: AMINATION OF CHLOROBENZENE BY SODAMIDE 187

If k2 is large compared with kH' and kD', a s seems likely, the rate of disappearance of
deuterated chlorobenzene, D , is given by the expression

-d[D1
- dt
-
- (kHf+kDf)[D][NH?-]- kL1[anion B].

Application of the steady-state approximation to anion B gives

kH'
[anion B]
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

= ----
k-i+k? [Dl [NHF I,
which converts the preceding expression to

By a si~nilarprocess the rate of disappearance of chlorobenzene, H, is found to be

Letting the isotope effect, kr%'/knf = ,i, and the fraction of anion returning to starting
Can. J. Chem. 1961.39:180-191.

material, k-~/(k-l+k?) = F, division of one rate expression by the other gives

Integration of this expression yields

Roberts' equation [I] can be Inore conveniently applied by setting the fraction of chloro-
benzene deuterated a t time t = d, and the fraction of starting material which has reacted
a t time t = P. Equation [I] then becomes

d 1 - 1 (l/d)-1-[F/{,i(l-F)-111
log -(l-P) = -.
do i(1-F)-1 l o g ( l / d ~ ) - l - [ ~ / ( ( l - F ) - l } ]
~ ~ u a t i o n - ' or
[ l ][2] concerns only the change in concentration of deuterium in unreacted
chlorobenzene. In order to calculate the aniline co~npositionpredicted by lnechanis~nI11
it is necessary to extend the liinetic treatment as follows. Let AD and AH represent
deuterated and undeuterated aniline, respectively,
 188 CANADIAN JOURNAL O F CIIEMISTRY. VOL. 39. 1961

Dividing one rate by the other gives

Substituting the steady-state expression for B and introducing i and F, where appropriate,
gives
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

Integrating with the condition that [AD] = 0 when [Dl = [Dlo gives

If it is assumed that ([AD]+[AR])/([D]o+[H]0) = P and that ([D]+[HI)/([D]o+[H]o)

= 1 - P the above expression becomes

This expression is simplified by letting the fraction of aniline which is deuterated a,

Can. J. Chem. 1961.39:180-191.

=
and the fraction of chlorobenzene which is deuterated = d , which gives

Equations [2] and [:3] were used to calculate a and d according to mechanism 111 and
the results are compared with the experimental results in Table IV. The F values shown
in Table IV were calculated from equation [2] using Roberts' value of i = 5.7. The
constancy of F through four different runs is strong support for the Roberts' mechanism
of concurrent amination and exchange. The average value of 0.864f 0.020 is much
more precise than the 0 . 5 6 4 0.20
~ reported by Roberts and co-workers (6). The different
nlagllitudes of F i n the two investigations may be due to some difference in experimental
conditions, such as the fact that potassiulll amide was used in one case and sodamide
in the other, or it lnay be the result of a difference in the method of analyzing chloro-
benzene for deuterium. A spectrophotometric method was used in both cases and the
standard sample on which all the analyses rest was prepared in the same way in both
investigations. In our case the standard sample analyzed S1.Gyo ~hlorobenzene-2-~H by
combustion, while in Roberts' case the standard sample was reported to be 100yo chloro-
benzene-2-2H without mention of the method of analysis. If the two standard samples
are assumed to be both about SOYo deuterated then the two ranges of F values overlap,
although the precision of the set by Roberts and co-worlters is not improved.
The a values in Table IV were calculated from equation [3] using i = 5.7 and the
calculated values of F. Comparison of the calculated and observed values of a shows
that the former is lower by about lOyo in each case. Although the a values calculated
for Roberts' mechanisnl are low, they parallel the observed values so closely as to leave
little room for doubt that the mechanism is essentially correct.
The 10yo discrepancy is probably due largely to the method of determining P by
measuring the moles of chloride ion produced. Bergstrom and co-workers (1) showed
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.
Can. J. Chem. 1961.39:180-191.

0
c
z
z
m
4

TABLE 1V
ALllninatio~~
of mixtures of chlorobenzene and chloroben~ene-2-~1-1
-g*
z
Run do't d*f P* Fcx~cll aea~olT** sobs* ** =!
0
1 0.407Zk.003 0.350It.008 0.4521t.006 0.844f.011 - -tt z
2 0.816f.019 0.414f.014 0.718It.006 0.870f.005 0 . 4 1 6 3 ,004
~ 0.441f.000 o
3 0.594It.014 0.290It.005 0.819&.012 0.868Zk.001 0.2831t.001 0.326f.005 o
4 0.396It.009 0.227It ,009 0.783It.007 0.874It.005 0.185It.001 z
0.204f.003
0'P
*Each datum is tlie mean of tliree analyses + the ~ n a s i m u mdeviation from the mean. 0
?do = mole fraction chlorobenzene-2-21-1 a t zero time. M
i d = mole fraction clilorobenzene-2-2H a t time 1. r-:
$ P = mole fraction starting material converted t o chloride ion a t time 1.
IIThe maximum value was calculated from equation [I] using tlie minimum value of d aiid tlie maximum values of do and P: the re\.erse combination gave tlie mioimum value.
5
1iTIie maximum value was calculated from equation I31 using the minimum value of d and tlie rnasiniurn values of do, P, and 1.'.
**a = mole fraction aniline-?H at time 1.
2!m
t t N o t enough product for analysis.
2
?
5
:
 190 CANADIAN JOURNAL O F CHEMISTRY. VOL. 39, 1961

that the allilille produced by amination of chlorobenzene is always acconlpanied by

by-products, alnong which they identified diphenylamine, triphenylamine, and $-amino-
biphenyl. They suggested that these are formed fro111aniline as illustrated in the following
equations :
CGHsNHz+ NaNHz d C8HsNHNa + NHa,
CGH5NHNa + CBHsCl+ (C6Hs)zNH + NaCl.

1<.'is seen that these side reactions produce chloride ion while consuming both chloro-
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

benzene and aniline in a inanner not included in the Roberts' mechanism. Since the side
reactions should consume ring-deuterated and undeuterated reactants a t equal specific
rates, they will not alter the mole fraction of deuterated chlorobenzene, d, or aniline, a ,
but they will make the apparent percentage of reaction, IOOP, too large. Insertioil of
the experilnental values of a , d, do, and i = 5.7 into equations [2] and [3] give calculated
P values of 0.646, 0.647, and 0.648 in runs 2, 3, and 4, Table IV, respectively. While
the precision of the calculated P values anlong runs is probably largely fortuitous, they
show that a reasonable amount of side reaction of the type proposed by Bergstrom could
account for the low calculated a values. A more attractive mechanism for the formation
of by-products within the Roberts' mechanism would be the following:
Can. J. Chem. 1961.39:180-191.

but the effect on a would be the same as that suggested for the Bergstrom mechanism.
In summary, our data and the predictions of Roberts' mechanism of concurrent
amination and exchange are in excellent agreement on the deuterium content of partially
reacted chlorobenzene, on the ortho-to-meta ratio of the deuterated aniline produced,
and in reasonable agreement on the total deuterium content of the aniline. There can
be little doubt that the mechanism is essentially correct.
A referee has pointed out that the excess of aniline-3-TI over aniline-2-W (56% and
54y0 of the former in experilnents 4 and 5, Table I) suggests a secondary isotope effect
on the rates of addition of ammoilia to ben~yne-2-~H. Roberts and co-workers (13)
have observed that inductive electron withdrawal by the substituent in a 2-substituted
benzyne favors the formatioil of 3-substituted addition products. This would indicate
that deuterium has a - I effect, as has been suggested by Shiner on other grounds (14).
The agreement of our data with Shiner's prediction is interesting, but we doubt whether
the accuracy of the analysis warrants a definite conclusion.

ACKNOWLEDGMENTS
The authors wish to thanlc Dr. H. R. Coish for suggestions regarding the mathenlatical
derivations, Miss A. Hetherington and Dr. B. L. Funt for assistance with the Bendix
G-15 computer, and the Natiollal Res2arch Council of Canada for a grant in aid of
this research.

REFERENCES
1. F. W. BERGSTROJI,
R. E. WRIGHT,C. CHANDLER,and W. A. GILKEY. J. Org. Chem. 1, 170 (1936).
R. E. WRIGHTand F. W. BERGSTROM.J. Org. Chem. 1, 179 (1936).
2. H. GILMANand S. AVAKIAN.J. Am. Chem. Soc. 67, 349 (1945).
3. J. F. BUNNETTand R. E. ZAHLER. Chem. Revs. 49, 382 (1951).
 DUNN ET AL.: AMINATION OF CHLOROBBNZENE BY SODAMIDE 191
R. A. BENKESERand W. E. BUTING. J. Am. Chem. Soc. 74, 3011 (1952). R. A. BENI~ESER and
G. SCHROLE.J. Am. Chem. Soc. 75, 3196 (1953). R. A. BENKESERand R. G. SEVERSON.J.
Am. Chern. Soc. 71, 3838 (1949).
J. D. ROBERTS,H. E. S I ~ I ~ I O NJR.,
S , L. A. CARLSAIITH,and C. W. VAUGHAN.J. Am. Chem. Soc.
75, 3290 (1953).
J. D. ROBEIV~S, D. 11. SE>IENOW, H. E. SI~IXIONS,
JR., and L. A. CARLS~IITH. J. Am. Chem. Soc.
78. 601 (1956).
H. GILJIANand I;. SCHULZE.J . Am. Chem. Soc. 47, 2002 (1925).
L. H. P. WELDONand C. L. \VILSOV. J. Chem. Soc. 244 (1946).
\V. M. LAUEKand L. A. ERREDE. J. Am. Chem. Soc. 76, 5162 (1954).
T. H. VAUGHAN, R. R. VOW, and J. A. NIEUWL.~ND.J. Am. Chem. Soc. 56, 2120 (1934).
A. S. I<ESTON,D. RITTENBEKG, and R. SCHOENHEIXIER. J. Bioi. Chem. 122, 227 (1937).
Downloaded from www.nrcresearchpress.com by 47.15.12.184 on 08/19/17. For personal use only.

C. AXFINSEN. In Preparation and measurement of isotop~ctracers. D. W. Wilson, A. 0. C. Nier,

and S. S. Reirnann. Ann Arbor, Mich. 1946. pp. 61-65.
J. D. ROBERTS, C. W. VAUGHAN, L. A. CARLSMITH,and D. A. SE~IENOW. J. Am. Chem. Soc. 78, 611
(1956).
V. J. SHINER,JR. J. Am. Chem. Soc. 82, 2655 (1960).
Can. J. Chem. 1961.39:180-191.