Inorganica Chimica Actu, 16 (1976) 181-184 181

0 Elsevier Sequoia S.A., Lausanne - Printed in Switzerland

Kinetic and Mechanistic Studies of the Rearrangement of Benzaldoxime to

Benzamide Catalysed by Nickel Acetate

A. K. JOHNSON and J. D. MILLER

Department of Chemistry, The University of Aston in Birmingham, Birmingham 84 7ET, U.K.
Received June 25, 1975

When solutions of benzaldoxime in digol are heated initially added in our kinetic experiments. We have not
with nickel acetate, the solute undergoes the Beckmann been able to identify, unequivocally, the form of the
rearrangement yielding benzamide as the major product. oxime which is actually involved in reaction. We believe
Kinetic studies of this reaction show that the reaction it to be the anti-isomer. Some ‘H n.m.r. studies cast
proceeds via two distinct slow steps. The data are light on the feasibility of this postulate. The spectra of
interpreted in terms of an initial attack of solvent on the syn- and anti-isomers are sufficiently distinct for
benzaldoxime, co-ordinated to nickel in the anti- their relative concentrations in a mixture to be deter-
isomeric form. This reaction yields a benzimidate which mined by n.m.r. We find that solutions of either in digol
subsequently reacts to yield either benzamide or a side (diethyleneglycol) show a pure syn-spectrum after
product. Numerical data are cited, together with infor- being heated for five minutes at 100°C. However
mation on the effect of other complexes as catalysts, solutions containing 0.3M oxime, in either form ini-
and on the requirements imposed by different solvents. tially, and 0.3M zinc chloride show a syn :anti ratio of
approximately 10: 1 after being heated to 100” C for
eighty minutes. We therefore infer that metal ions are
Introduction capable of isomerizing one form of benzaldoxime to the
other, and of stabilizing the anti-form to some extent.
Recently we have become interested in ligand reac- When solutions of benzaldoxime in hydrocarbon sol-
tivity, especially in complexes where the central metal vents are refluxed in the presence of nickel acetate tetra-
ion is exhibiting a normal oxidation state. Here we report hydrate, benzamide is produced. The yield decreases as
the results of experiments aimed at elucidating the the boiling point of the solvent increases. Xylene (b.p. =
role of simple nickel compounds in the conversion of 140” C) yields circa 8.5% benzamide after several hours,
benzaldoxime to benzamide. The Beckmann rearrange- whereas higher boiling solvents such as tetrahydronaph-
ment of oximes to amides’ can be effected by a variety thalene (b.p. = 207” C) yield less product (40%) but
of reagents, including some metals and metal salts, e.g. more rapidly. Clearly a side reaction of high activation
cuprous chloride’. The first detailed study of such energy is involved. In these solvents nickel formate,
reactions under homogeneous conditions was that of nickel chloride and stable nickel complexes do not be-
Field et aL3, who studied the effect of nickel acetate have as catalysts, while the amide produced precipitates
tetrahydrate on solutions of oximes in xylene. Although from solution on cooling. Zinc acetate is effective as a
the synthetic possibilities of metal salts as catalysts for catalyst whereas manganese and cobalt acetates are in-
this rearrangement are documented, the mechanistic effective. The most convenient experimental system for
aspects of such isomerizations have, as yet, remained a kinetic studies under homogeneous conditions uses
matter for speculatior?. digol as solvent, and the benzamide production is mon-
The ability of oximes to act as monodentate hgands itored in the infra-red region. The band at 1675 cm-’
is documented. Normally, aldoximes co-ordinate to is suitable, and solutions obey Beer’s law. Any benz-
metals in the/l- or anti-form using the nitrogen atom as aldehyde produced could also be detected in the re-
dono?. Occasionally co-ordination can be achieved gion, at 1700 cm-‘. Calibration shows that 10% pro-
using oxygen as the donor atom-. duction of benzaldehyde would be readily detected.
We never observed evidence of benzaldehyde produc-
tion, which therefore represents a minor factor, at most,
Results in the overall disappearance of benzaldoxime’.
Plots of benzamide concentration against time for in-
The more stable isomeric form of benzaldoxime, the dividual runs with added nickel acetate are reproducible,
a - or syn -form, is used throughout as the form of reagent and show the characteristic profile for a consecutive
182 A. K. Johnson and J. D. Miller

reaction sequence. The data were successfully fitted, TABLE I. Computed Pseudo-first Order Rate Constants, k,
using an iterative least squares computer programme, to and kt,; and the Fraction of Oxime Converted to Amide, F;
equation (1) which occurs when two consecutive for the Rearrangement of Benzaldoxime to Benzamide in the
pseudo-first order reactions lead to the formation of Presence of Nickel Acetate Tetrahydrate in Digol.

benzamide.
Initial [Ni Temp/ 104k,/s-’ 104k&-’ F
[amide] = [amide], {l-- k, exp (-k,t) + -&- [oxime] (OAc),] ‘C
ki,-k, b a /M lM

eXP (-k&j (I) 0.1 0.012 184 3.93 6.59 0.535

The values of the constants k, and kb are presented 0.2 0.012 184 3.83 6.34 0.569
in Table I, together with data for the fractional conver- 0.3 0.012 184 4.24 6.76 0.553
sion, F, to benzamide. In Table II, we present data for 0.2 0.004 184 1.57 6.99 0.580
0.2 0.008 184 2.65 6.52 0.573
the analogous reaction using deuterated benzaldoxime,
0.2” 0.016 184 4.70 5.09 0.595
C6H,CDNOH, as reagent. The sample of this material
0.2 0.020 184 6.08 4.69 0.606
prepared by us contained 60% of the deuterated form 0.2 0.024 184 7.45 5.23 0.594
and 40% of the undeuterated oxime. The data for the 0.2 0.012 160.5 1.38 0.87 0.749
pure deuterated reagent were determined from a com- 0.2” 0.012 168.5 2.20 2.07 0.685
parison of otherwise identical reactions using the mixed 0.2” 0.012 176.5 3.18 3.20 0.639
reagent and undeuterated reagent. The C-D absorption 0.2 0.012 191.5 5.05 10.04 0.537
of the reagent is found at 2205 cm-‘, a region clear of 0.2 0.012 200.0 7.25 20.66 0.483
solvent absorption, and so the rate of disappearance of
the reagent can be monitored when a higher oxime con- “Under conditions where the rate constants are very similar
the accuracy of the individual values becomes suspect.
centration (2M) is used. The diminution of the peak
at 2205 cm-’ obeys first order kinetics, and the ob-
served rate constant is also quoted in Table II. TABLE II. Computed Rate Data in Digol at 189.O”C in the
The data summarised in Tables I and II represent the Presence of 0.012M Nickel Acetate.
bulk of our experimental effort. But there are further
results of relevance to the discussion. Under the con- Initial Reagent Peak Followed 104k,/ 104k,/ F
ditions of our kinetic studies benzonitrile is not con- Cont. SC' S-'

verted to benzamide. We have prepared imidates as

0.2M Amide
their hydrochlorides from the reactions of benzonitrile C,H,CHNOH 1675 cm-’ 4.98 9.30 0.538
with digol and with 2-methoxyethanol. These com- 0.2M Amide
pounds yield benzamide on heating in digol. Table III CsHsCDNOH 1675 cm-’ 3.15 8.95 0.525
contains data for catalysts, other than nickel acetate, 2.OM C-D
in digol. Some of these compounds are ineffective in C,H,CDNOH 2205 cm-’ 2.93 - -
hydrocarbon solvents. We also observed qualitatively
that zinc acetate dihydrate, zinc chloride and dichloro-
bis(triphenylphosphine)platinum(II) show slight TABLE III. Pseudo-first Order Rate Constants for the Re-
catalytic activity whereas the analogous palladium(II) arrangement of Benzaldoxime to Benzamide in Digol at
complex does not. 184” C. Initial concentrations being 0.2M oxime and 0.012M
catalyst.

Catalyst 104k,/s-’ 104k,/s-i F

Discussion
Ni(PPh,)& 1.15 5.60 0.646
From an examination of the data of Tables I and II, Ni(PPha)212 19.0 7.40 0.612
the following results emerge. The first step in the reac- Ni(PhCHNOH)4C12 1.73 3.47 0.675”
tion sequence is that described by k,. This step shows Ni(PhCHNOH)J2 9.88 6.74 0.670”
a strong dependance on the nickel concentration, and NiC&. 6H20 2.57 3.52 0.587
involves the cleavage of the side chain C-H bond. Ad- Ni(OAc)z.4H20 3.83 6.34 0.569
ditionally, since the formation of benzamide is accu-
aThese data take into account the additional quantity of
rately described by two consecutive pseudo-first order
oxime added as a ligand.
rate constants, the concentration of the active nickel
catalyst must remain essentially constant throughout
the course of a reaction. Since the initial oxime concen- concerning reactions in solvents other than digol, and
tration is much more than the total nickel concentration, the non-conversion of benzonitrile to benzamide imply
it follows that any nickel oxime complex must be formed that the overall constant k, describes a series of reac-
only at low concentrations. The experimental results tions involving the attack of solvent on co-ordinated
Benzaldoxime to Benzamide Couversion by Ni(l1) 183

oxime yielding an imidate. The variation of the frac- ment. The values of k, in both Tables I and III show
tional conversion to benzamide with temperature shows that the nickel salts are involved in the conversion to
the existence of a high activation energy side reaction. the intermediate imidate, but not directly beyond that
The linear plot of log(F/(l-F)) against ‘/T enables stage. The variation of k, with added catalyst appears
the difference in activation energies of main to side re- to be one in which k, increases as the availability of co-
actions to be evaluated as-l 1.7 + 0.6 kcal mol-‘. Within ordination sites on the nickel ion increases. Since the
experimental error an Arrhenius plot k, is linear, giving normal mode”’ of co-ordination of oximes is through
an overall activation energy of 15.8 + 0.8 kcal mar’. N-bonding with the less stable l2 syn-form; this mode
The side reaction cannot therefore arise before the for- should mean that only small fractions of the total nickel
mation of imidate since, in that case, the numerical val- are co-ordinated to oxime, as we find. Therefore we
ues would result in curvature of the Arrhenius plot. postulate the role of the nickel(I1) species to be that
Further support for this conclusion comes from Table II. of a Lewis acid, which co-ordinates to the nitrogen atom
While k, shows a strong isotope effect, F does not vary of anti-benzaldoxime, stabilizing this form and causing
significantly. The constant k,, shows a small but real polarization within it. Attack of nucleophilic solvent
variation with the nickel concentration. The rate con- molecules, or acetate in xylene, on the side chain carbon
stant decreases slightly as the catalyst concentration atom is thus made easier. Since the intermediate imidate
increases. is a poor ligand, the equilibrium between free and co-
Our observations lead us to deduce reactions (2) to ordinated imidate, reaction (4) is very much in favour
(6) as representing the overall reaction sequence of the free ligand. Imidates are known13 to undergo ther-
mal rearrangement to amides. These reactions are not
K,
Ni” + syn-oxime & Ni”(anti-oxime) (2) generally quantitative, as shown in our reactions (5)
and (6). Since we were unable to isolate and charac-
kz
Ni’*(unti-oxime) ----+ Ni”(imidate) (3) terize a pure specimen of the side product, we cannot
comment in more detail on reaction (6). Other wor-
K,
Ni” + imidate T Ni”(imidate) (4) kers have studied such side products6,13. In accord
with the scheme outlined by us, values of kb vary little
k4
imidate --+ benzamide (5) from one catalyst to another. Only when the catalyst
contains chloride does kb change significantly from the
k,
imidate - side product (6) value with nickel acetate. This observation may reflect
a change in K,, or be related to the Pinner fission re-
If the concentration of free nickel species, denoted by action between chloride ions and protonated imidates14.
Ni”, is approximately equal to the total nickel con-
centration, [Nil, then
Experimental
k, = K,k,[Ni] (7)
k, = (k4 + k&(1 + K,[Ni]) (8) The rearrangement of benzaldoxime to benzamide is
F = k4/(b + k,) (9) strongly exothermic, and is studied at a high tempera-
ture. The most satisfactory arrangement for kinetic
The calculated entropy of activation for the constant studies requires the use of a heating bath of silicone oil
as the heating medium. The reaction vessel is cylin-
k, is
drical with a flat base, to allow for magnetic stirring. A
AS’, = dS0i + dSIZ = Cu. -34 cal deg-’ mol-’
further cylinder of slightly smaller cross section is sus-
An examination of known” entropy changes for pended within and connected to a reflux condenser. A
equilibria comparable to reaction (2) suggests thatAS*, liquid, whose boiling point is just below the working
must be large and negative. We feel that this supports temperature, is placed in the inner compartment to act
the idea of reaction (3) being a bimolecular process as a heat sink. The annular space containing the reac-
involving attack by solvent. However we note that a tion solution is serviced via ground cones. Weighed
reaction believed” to proceed by a similar mechanism amounts of finely divided catalyst are placed in a small
shows a different value. glass cup, within which is sealed a small magnet. This
The second composite rate constant, ki,, varies little container is held at the top of the vessel, and later drop-
with [Nil. This means that individual activation ener- ped into the oxime solution to initiate reaction. In the
gies can be assigned to k4 and kg; 27.7 ? 1.3 kcal mol-’ case of nickel acetate dissolution of the solid was rapid,
and 39.1 + 2.2 kcal mol-’ respectively. Unfortunately although some other added solids required longer times
K, cannot be determined accurately : a value of 19M-’ to dissolve. Samples of the reaction mixture were ab-
at 184” C appears to be a reasonable estimate. stracted at known times using a syringe and quenched
The aim of this work was to determine the role of to room temperature. Since stainless steel can catalyse
metal salts in activating aldoximes towards rearrange- the reaction under study, the syringe needle was kept in
184 A. K. Johnson and J. D. Miller

contact with the reaction solution for as short a time as References

possible, and also cleaned after each sampling. Benz-
amide concentrations were determined by infra-red 1 L. G. Donaruma and W. 2. Heldt, Organic Reactions, 1960,
spectroscopy using matched solution cells with silver 11, 1.
chloride windows and a Perkin-Elmer 457 spectro- 2 W. .I. Comstock, Amer. Chem. J., 1897,19, 485.
3 L. Field, P. B. Hughmark, S. H. Shumaker and W. S. Mar-
meter. Essentially identical rate data were obtained
shall,J. Am. Chem. Sot., 1961,83, 1983.
from runs using unpurified and redistilled” solvent.
4 M.E. Stone and K. E. Johnson, Canad. J. Chem., 1973,51,
Syn- and anti-benzaldoxime were prepared by the 1260.
method of Vogel”j while the deuterated syn-ben- 5 A. Bryson and F.P. Dwyer, J. Proc. Roy. Sot. N.S. Wales,
zaldoxime, C6H5CDNOH, was prepared from the ylide 1940,74, 471.
benzal-bispyridiniumbromide’7 and its hydrolysis’* 6 idem, ibid., p. 107.
to benzaldehyde. 7 W. Hieber and F. Leutart, Chem. Ber., 1929,62, 1839.
Benzimidate hydrochlorides were prepared from 8 A. Chakravorty, Co-ord. Chem. Rev., 1974,13, 4.
benzonitrile and digol or 2-methoxyethanol by the liter- 9 R. Paul, Compt. Rend., 1937,204, 363.
10 S.J. Ashcroft and C.T. Mortimer, “Thermochemistry of
ature route”, as were the various transition metal com-
Transition Metal Complexes”, Academic Press, London,
plexes; the tetrakis-benzaldoxime nickel(B) com-
1970.
plexes7,20, bisdimethylglyoxinatonickel(II)Z’, and the 11 R. Breslow, R. Fairweather and J. Keana, J. Am. Chem.
dichlorobis(triphenylphosphine) complexes of Sot., 1967.89, 2135.
nickel(II)“, palladium(II)23 and platinum(II)23. 12 “Dictionary of Organic Compounds”, Eyre and Spoltis-
With the exception of the product derived from ben- woode, London, 1965.
zonitrile and digol, all reagents prepared gave melting 13 J. W. Schulenberg and S. Archer, Organic Reacrions, 1965,
or boiling point and analytical data in agreement with 14, 1.
the expected values. The various attempts to prepare 14 S. M. McElvain and B. E. Tate, J. Am. Chem. Sot., 1951,
an imidate hydrochloride from digol, HOCH,CH, 73, 2233.
15 W.H. Rinkerbach, Ind. Eng. Chem., 1927,19, 474.
OCH,CH20H, yielded products rich in chlorine. This
16 A. I. Vogel, “Practical Organic Chemistry”, 2nd Ed.,
probably indicates a significant degree of disubstitution Longmans Green, London, 1951.
on digol. 17 F. Krohnke and H. Leister, Chem. Ber., 1958, 91, 1295.
Microanalytical data were determined within this 18 R.A. Olofson, D.M. Zimmerman, J. Am. Chem. SOL,
department, or commercially by Dr. F.B. Strauss of 1967,89, 5057.
Oxford. Proton n.m.r. spectra were recorded on a Va- 19 G.D. Lander, J. Chem. Sot., 1900, 736.
rian Associates HA 1OOD spectrometer. 20 W. Hieber and F. Leutart, Chem. Ber., 1927,60, 2296.
21 B. E. Dawson, “Practical Inorganic Chemistry”, Methuen,
London, 1963.
22 L.M. Venanzi,J. Chem. Sot., 1958,719.
Acknowledgments 23 L. Malatesta and C. Cariello, J. Chem. Sot., 1958. 2323.

One of us (A.K.J.) thanks the Science Research

Council for a research studentship.