Reduction of Acetone and its Hydrocoupling with Acrylonitrile in Aqueous Sulphuric Acid Solution at a Mercury Cathode

BY 0. R. BROWN AND K. LISTER Dept. of Physical Chemistry, University of Newcastle upon Tyne, Newcastle upon

Tyne, NEl 7RU

Received 28th December, 1967

The mechanism of the reduction of acetone and its hydrocoupling with acrylonitrile in aqueous sulphuric acid at a mercury cathode has been investigated using steady and non-steady state controlled potential techniques under experimental conditions employed in synthesis. The results indicate that a primary electron transfer occurs to an adsorbed acetone molecule, with protonat ion, yielding an adsorbed radical which is the common precursor of the diverse products of acetone reduction alone and of the coupled products obtained in the presence of acrylonitrile.

Cathodic hydrocoupling proceeds by radical and by carbanion mechanisms. This latter process permits the syntheses of useful bifunctional materials in high yields from suitable substituted unsaturated substances. However, the combination of cathodically-generated radicals inevitably leads to poorer yields of crossed hydro- coupled product^.^

Sugino and Nonaka reported the hydrocoupling of acetone and aciylonitrile in good yield at a mercury cathode. They suggested that acetone reduces to a carbanion which adds to an acrylonitrile molecule with subsequent prctonation to y-hydroxy y -met hyl valeroni trile.

CH2:CHCN H +

(CH3)2C0 + H+ +2e-,(CH3),~OH-+(CH3)2COHCH,~HCN-+product.

Such a carbanion mechanism seems unlikely under these conditions (20 % by weight, sulphuric acid) and it is in conflict with a mechanism proposed by Sugino et al. to explain the cathodic reduction of acetone under similar conditions. However, the absence of appreciable amounts of polyacrylonitrile precludes the possibility of the formation of free radical intermediates.

Usually, examinations of the mechanisms of electrode reactions of organic mater- ials by electrochemical methods have been confined to low concentrations of the depolarizer ; this restriction is often imposed by the use of two electrode cells and by the need to minimize migration currents. In this work, using steady and non- steady state potentiostatic techniques, we have studied the reduction of acetone and its hydrocoupling with acrylonitrile under the experimental conditions employed in synthesis.

EXPERIMENTAL

Solutions were prepared from thrice distilled water, A.R. acetone, B.D.H. laboratory reagent grade acrylonitrile and A.R. conc. sulphuric acid. Prior distillation of the acetone and of the acrylonitrile caused no detectable change in the measurements.

106

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online / Journal Homepage / Table of Contents for this issue

0. R. BROWN AND K. LISTER 107 Preparative electrolyses were conducted in an all-glass divided cell containing a mercury

pool cathode. Kinetic experiments were performed in an all-glass undivided cell containing a platinum wire counter electrode wound as a spiral around the sitting mercury drop cathode (area 0.1 cm") which was renewed immediately before each measurement. Each cell was fitted with a Luggin capillary leading to a saturated KCl aq./calomel reference electrode through a liquid junction formed in a closed tap. The cell temperature, recorded but uncontrolled, was 21 rt2"C. Solutions were deoxygenated by displacement with purified nitrogen.

A fast-response potential regulator with a maximum output of 4A was used in conjunc- tion with a function generator for most measurements. Records taken at times less than 50 psec were checked using a potentiostat with an ultra-fast response.8 Currents were observed and photographed on a Tektronix 503 oscilloscope, preceded by a preamplifier (common-mode rejection of 80 dB) as voltages developed across a 10 ohm resistor.

The products of the reduction of acetone alone were analyzed as follows. A green oil which separated during the constant potential electrolysis was removed and identified by infra-red and mass spectrometry as di-isopropyl mercury. The catholyte was neutralized with sodium carbonate and was injected directly into a g.1.c. column (20% Tween 80 stationary phase, acid washed 60/80 mesh Diatoport support). isoPropanol and pinacol were identified and estimated by comparison with authentic samples using an internal standard of n-propanol.

After the crossed hydrocoupling reaction had been conducted at constant potential, the neutralized catholyte was extracted with ether and a sample chromatographed on a SE 52 column. Of the five peaks obtained, the first three were removed when the ether extract was evaporated under vacuum. Samples of the remaining syrup were injected on to a preparative g.1.c. column (8 ft x 0-5 in. diam. containing 10 % GE. SE 52 on 80/100 mesh Diatoport S). Two peaks appeared; the column effluent of the former was collected in a small tube and examined by infra-red analysis. In addition to a water peak, gem-dimethyl and carbonyl were detected and the spectrum corresponded with that of yy-dimethylbutyrolactone (kindly supplied by Prof. Sugino). Direct infra-red analysis of the syrup, after drying, gave peaks for hydroxyl, gem-dimethyl, carbonyl and nitrile so that the second peak of the chromatogram probably corresponds to y-methyl y-hydroxy valeronitrile as isolated by Sugino.

RESULTS

Unless otherwise stated, the solutions used were 0-5 M in H2S04. All potentials are quoted with respect to the saturated KCl/calornel electrode.

CONTROLLED POTENTIAL ELECTROLYSIS

The reduction products of acetone are described in table 1. The current level remained constant during the course of the electrolysis. A colourless gas was steadily evolved from the mercury surface and the catholyte solution became cloudy, apparently caused by colloidal mercury. A green oil formed on the cathode surface.

TABLE CONTROLLED POTENTIAL ELECTROLYSES OF 4.0 M ACETONE, 0.5 M H2S04 SOLUTION

product analysis : current efficiencies, % electrode c.d.

- 1 -270 1.6 22 2.5 17 - 1.300 5.0 21 2.1 28 - 1 -375 8.4 27 1.1 11

potential, V mA cm-2 isopropanol pinacol di-isopropyl mercury

Several controlled potential reductions of acetone were performed in the presence of acrylonitrile. As the initial concentration of acrylonitrile was increased from 0.3 to 3.4 M an increasing yield of a white polymeric product was formed. During the electrolysis the

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

108 CATHODIC H Y D R O C O U P L I N G

current decreased more rapidly than could be explained simply by the consumption of the reactants. Current yie!ds of the coupled product, obtained by weighing the syrup from the ether extraction, exceeded 50 %.

KINETIC MEASUREMENTS

POLARIZATION CURVES

Linear potential sweeps were applied to the mercury cathode in solutions containing acetone. The measured currents were independent of the sweep rate and direction at values below 3 V/sec provided that the diffusion limited currents were not approached. In this way the steady-state polarization curves were recorded. These results were confirmed using a potential-switching method to obtain points individually. Correction was made for potential errors arising from solution resistance. These experiments were extended to solutions con- taining, in addition, acrylonitrile (fig. 1).

took

/ i I - 1.20 -1.25

electrode potential, V

FIG. 1.-Steady state polatkition curves. 0, 0-5 M solution; 0, 0-5 M H2S04, 04 M acrylonitrile solution; -+, 0-5 M HnS04, 3.5 M acetone solution; A, 0.5 M HzSO4, 0.4 M acrylo-

nitrile, 3.5 M acetone soluticn.

The diversity of products in the reduction of acetone in acid solutions indicates a complex overall reaction scheme. Therefore a kinetic analysis of the steady-state polarization curves for that system was not attempted. However, the high yields of coupled product from the reduction of acetone in the presence of acrylonitrile do justify a steady-state kinetic study. Slopes of the Tafel plots (fig. 1) for 4 M acetone solutions containing 0.1 to 0.6 M acrylo- nitrile were (57 f5 mv)-l. The effects of independently varying the acetone and acrylonitrile concentrations are shown in fig. 2 and 3 respectively for several values of electrode potential.

FAST LINEAR POTENTIAL SWEEP EXPERIMENTS

In the presence or absence of acrylonitrile, cyclic voltammetry demonstrates the re- oxidation of material formed during the reduction of acetone (fig. 4 and 5).

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

0. R. B R O W N AND K. LISTER 109

POTENTIAL SWITCHING EXPERIMENTS

In order to investigate quantitatively the non-steady state phenomena revealed by the fast potential sweep experiments, the electrode was switched rapidly from a region of electro- inactivity to more cathodic potentials. Steeply falling transient currents were observed at the

4 0

3 0

N 1

E, 3 2 r

-z I C

C 1 I I I I I I

, 0 . 2 0.4 0 . 6 0 . 8 1.0 1.2 1.4

acetone concentration, M FIG. 2.-Coupling reaction : dependence of steady currents upon acetone concentration in 0.3 M

acrylonitrile, 0.5 M HzS04 solutions. 0 , -1.259 V ; 0, -1.238 V ; 0 , -1.195 V.

31

21 N

4 E

E ,

acrylonitrile concentration, M FIQ. 3.--Coupling reaction : dependence of steady currents upon acrylonitrile concentration in 3.5 M

acetone, 0.5 M HzS04 solutions. 0, -1.251V; 0, -1.231 V ; +, -1-211 V ; A, -1.191 V.

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

110 CATHODIC HYDROCOUPLING

N I

3 s

100 I2O t , /

/ /

\ I -4ot \ /.

- 6 / , ,. ‘,\, , ,;/; ,. /

- a 0 ..%. .;’ -100

- 1 2 0

- 1.10 - 1.15 - 1.20 - 1.25 I

electrode potential, V 0

FIG. 4.-Fast potential sweeps using single symmetrical triangular waves. The solution is 1-4 M in acetone and 0.5 M in HzSO4. - * - 960 V/sec ; - - - 150 V / S ~ C ; - 10 V/sec.

-loot I I I t

electrode potential, V - 1.10 - 1.15 -1.20 -1.25 3 0

FIG. 5.Dotential sweeps at 960 V/sec using single symmetrical triangular waves. The solutions are 1.4 M in acetone and 0.5 M in H2S04. Acrylonitrile concentrations are - - - zero ; -y 0.3 M.

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

0. R. B R O W N AND K. LISTER 111

beginning and end of the cathodic pulse (fig. 6). An exception to this behaviour was noticed on switching to high negative potentials in the presence of acrylonitrile, when rising transients occurred.

60-

4 0

2 0

ri

INTERFACIAL CAPACITANCE MEASUREMENTS

Information was sought concerning the adsorption of acrylonitrile and acetone ; in potential regions free from faradaic phenomena the capacitance charging current in response to a linear potential sweep was recorded. The presence of acrylonitrile in concentrations as

-

-

time, psec FIG. 6.-Response to a potential pulse from - 1.150 to - 1.250 V. The solution is 1.4 M in acetone

and 1.0 M in H2S04.

high as 0-3 M caused no appreciable change in the capacitance values. However, the gradual addition of acetone to the base electrolyte solution decreased the capacitance in the region of the capacitance minimum until a limiting value was reached. Also, acetone caused an increase in the height of the anodic adsorption capacitance peak (table 2).

TABLE 2.-vARIATION OF THE INTERFACIAL CAPACITANCE WITH THE ACETONE CONTENT OF 0.5 M HzSO4 SOLUTION

acetone potential of the capacitance at the capacitance at - 1-10 V, concentration, M capacitance peak, V peak value, pF cm-2 pF cm-2

0~000 - 0.380 54 23 0-035 - 0.380 54 23 0.350 -0.190 82 20 0-700 -0.160 94 15 2.soo - 0,080 145 15

DISCUSSION The cyclic voltammetric experiments indicate that acetone undergoes a reversible faradaic

process at potentials more cathodic than - 1.2 V. At high sweep rates almost all of the reduc- tion charge can be re-oxidized on the anodic sweep. Although, in the steady state, the inter- mediate can yield several products, it seems that non-steady state kinetic data should afford

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

112 CATHODIC HYDROCOUPLING

details of the nature and formation of this intermediate. The isolation of organomercury products indicates that an intermediate possesses a radical nature.

The absence of copious amounts of polymer, the possibility of re-oxidizing all of the inter- mediate, and the form of the current-time relation in response to a potential step all indicate that the radical is confined to the electrode surface. A solution-free intermediate formed by a potential pulse would cause a transient current which would show a smaller time dependence than that observed (fig. 6) ; a compIetely diffusion-controlled process would give a linear i against t 4 relation whereas a process partly controlled by the dimerization of free radicals would be characterized by a more slowly falling current.

We postulate a simple mechanism based on a Henry's law isotherm for the intermediate species :

(CHs)2CO+H++e+(CH3)2k0Hd.

dB/dt = kf-kbB = i/Fy

where kf is the reduction rate, a potential-dependent constant when the activities of acetone and the hydrogen ion do not change. This integrates to

Designating the surface concentration of the radical as B, then at constant potential

- kb t = [(kf - kbB) /(kf - kb&)l Y

where Bo is the initial value of B, which, for a sufficiently positive base potential must approach zero. In that case,

so that B = (kf lkdll- exP (-&)I,

i = FdB/dt = Fkf exp (- kbt).

We test this model by plotting the experimental current transients in response to cathodic potential steps applied in the absence of acrylonitrile in the form log,, i against t (fig. 7). The reIationship is linear, small deviations at low currents resulting from the contribution from the steady current. Values for kf and kb are obtained from the intercept and slope respectively. The variation of kfand kb with potential (fig. 8) takes the expected logarithmic form, the slopes corresponding to a cathodic transfer coefficient of a = 0.5. The surface concentration of radicals is obtained from the reduction charge ; at - 1-3 V, B is found to be approximately 10-lo mole cm-2 which corresponds to a coverage fraction of less than 0.1, assuming that an adsorbed molecule occupies some 10 A2.

Fig. 9 shows the effect of the acetone concentration on the initial currents recorded on the application of a cathodic pulse to a fixed potential (- 1.250 V). The fact that the reaction order is one at low concentrations and changes to zero above 1 M suggests that the acetone molecule or its protonated form, whichever is involved in the charge transfer reaction, is adsorbed at the electrode. This conclusion is supported by the facts that, in the same region of acetone concentration, the limiting value of interfacial capacitance is reached (table 2) and that the steady currents observed in the presence of acrylonitrile (fig. 2) completely reflect this change in reaction order with respect to acetone.

The high yields of coupled products formed in the reduction of acetone in the presence of acrylonitrile reported by Sugino et aL4 were essentially confirmed. The Tafel slope (fig. 1) indicates that the rate-determining step is a chemical process following a one-electron transfer electrochemical pre-equilibrium step. The reaction order with respect to acrylonitrile is one at concentrations below 0.05 M but falls at higher values although zero order behaviour is not reached (fig. 3). On the basis of these results we propose the mechanism :

rapid

e,H * (CH3)2d0Had+ CH2=CHCN 3 (CHs)zCOHCH;?dHCNad-+ (CH3)2COHCH2CHaCN

(2) (CH&dOHad + CH2=CHCNad

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

0. R. B R O W N AND K. LISTER 113

We support this scheme by reference to the results of the potential switching experiments to cathodic potentials in the presence of acrylonitrile. Whereas, at low values of potential, falling current transients were seen as for acetone alone, switching to potentials progressively more cathodic caused the transient component of the current to fall to zero and then to change sign ; i.e., a rising transient appeared. This behaviour is explained by the proposed

time, p e c FIG. 7.-Response to potential pulses from - 1-150 V to various potentials using a solution 1.4 M in acetone and 1.0M in H2S04. 0, -1.280V; 0 , -1-26OV; 0, -1.250V; m, -1-24OV;

A, -1.230V.

electrode potential, V FIG. 8.-Potential dependence of the rate parameters of the initial stage of acetone reduction.

kf refers to a solution of 0-5 M in H2S04 and 3.4 M in acetone.

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

114 CATHODIC H Y D R O C O U P L I N G

mechanism ; the cathodic pulse (from a base potential of - 1.150 V) causes an increase in the concentration of adsorbed radicals. These can undergo principally two reactions : (2) and the reverse of ( I ) . At sufficiently cathodic potentials the rate of (2) will exceed that of the competing process so that as the new steady-state concentration of adsorbed radicals is approached the cathodic current will rise to the steady value. No transient is observed at the potential at which the rates of these competing reactions are equal. Experimentally, this was found at - 1.300 V. This value is not accurate ; the c.d. was high so that a considerable ohmic contribution is included in this potential. The same experimental restriction precluded a measurement of the polarization curve in this potential region where, according to the suggested mechanism, the Tafel slope would be expected to change as reaction (1) ceases to be in equilibrium. For the same reason the exact form of the transients was not analyzed.

acetone concentration, M FJG. 9.Initial currents obtained on switching the electrode potential from -1.150 to -1.250 V in

0.5 M H2S04 solutions containing various amounts of acetone.

We are able to test the correspondence of the data as follows. At - 1.300 V, B was mole cm-2 from the charge involved in the sweep experiments when 1.2 M acetone

solution was used. Fig. 8 shows that kb has the value lo4 sec-l at that potential. Thus, if the presence of acrylonitrile affects the value of B at - 1.300 V only to a small extent, then the rate of (2) at that potential should be - mole cm-2 sec-l (or 2 x 10-1 A cm-2 for a two- electron transfer reaction). This can be compared with the c.d. (2.5 x 10-1 A cm-2) obtained by extrapolating the uppermost Tafel plot in fig. I to - 1.300 V. Although this latter result refers to a different acetone concentration (3.4 M) the argument is not affected ; the reaction is zero order with respect to acetone in this range (fig. 2).

The suggested mechanism involves a rapid reduction of the radical species

(CH3)2COHCH2dHCNad.

In order to prove that this should occur in this potential range, the polarography of a-bromo- propionitrile was studied. This material was prepared from the corresponding acid.g In the reaction medium this nitrile, in M solution, gave a wave at a half-wave potential of -0.10 V. Its electrolysis at a potential on the plateau of the wave indicated a consumption of 1.6 electrons per molecule. The products of this c.p.e. were analyzed gas chromato- graphically ; the major product was propionitrile. The mechanism of the reduction of niono- halogen substituted organic compounds proceeds normally through the radical intermediate although this is usually reducible at the potential of its formation.'O We conclude that the radical CH3dHCN and therefore its homologue (CH3)2COHCH2eHCN are readily reduc- ible in the potential range in which the hydrocoupling reaction occurs.

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online

0. R . BROWN A N D K . LISTER 115

In order to determine the sequence of the protonation, adsorption and electron transfer steps in the initial stage of acetone reduction in acid solution, some additional experiments have been performed. Using the drop time method, electrocapillary curves were determined in solutions containing acetone and sulphuric acid. The surface excess, determined from the relation

is of the order mole cm-2 in concentrated acetone solutions, essentially confxming that saturated coverage does occur, as inferred by the kinetic and capacitance measurements. The equilibrium constant for the protonation ofacetone is given as aH+aA/am+ = antiloglo 1-58.

In the solutions used in this work, the concentration of protonated acetone is far exceeded by that of the neutral species but the ratio is approximately constant. That the surface excess determined above refers to uncharged acetone was proved by varying the acid concentration, the value for acetone being fixed. The resulting variation in concentration of the protonated species failed to affect the drop times significantly.

The dependence of the rate of the first step of acetone reduction upon the acid concentra- tion was examined in both regions of acetone concentration. At low acetone activities, first-order behaviour was found with respect to the hydrogen ion. In the region where the reaction order with respect to acetone is zero, the effect of changing the acid concentration was diminished. Similarly, in this latter region, the coupling reaction showed an order of approx- imately 0.5 with respect to hydrogen ion. We conclude that adsorption of acetone is the initial step but we are unable to determine whether protonation precedes or accompanies electron transfer.

This work is part of a programme supported by the Science Research Council. The authors thank the S.R.C. for the award of a research studentship to one of them (K. L.).

P. J. Elving and J. T. Leone, J . Anzer. Chem. SOC., 1958, 80, 1021. M. M. Baizer, J. Org. Chem., 1965,29, 1670. M. J. Allen, W. G. Pierson and J. A. Siragusa, J. Chem. SOC., 1961, 2081. K. Sugino and T. Nonaka, J. Electrochem. SOC., 1965, 112, 1241. T. Sekine, A. Yamura and K. Sugino, J. Electrochem. SOC., 1965, 112,439. A. Bewick and 0. R. Brown, J. Electroanal. Chem., 1967, 15, 129. ' 0. R. Brown, Electrochim. Acta, 1968, 13, 317. * A. Bewick and M. Fleischniann, Electrochim. Acta, 1963, 8, 89.

K. L. Barry and J. M. Sturtevant, J. Amer. Chem. SOC., 1941, 63, 2679.

S. Nagakura, A. Minegishi and K. Stanfield, J. Amer. G e m . SOC., 1957, 79, 1033. lo L. G. Feoktistov, Usp. Elektrokhim. Org. Soed. Akad. Nauk S.S.S.R., 1966. 135.

Publ

ishe

d on

01

Janu

ary

1968

. Dow

nloa

ded

by U

nive

rsity

of

Bir

min

gham

on

31/1

0/20

14 1

5:16

:20.

View Article Online