Mechanism of electroreduction of nitrobenzene in alkaline media

Abstract

Nitrobenzene is shown to give a stepwise reduction in certain media. With a view to examine the factors which influence the scheme of reduction, the effect of different solvents and ionic strengths are examined using tast polarography. In all, ten different organic solvents are used as 50% solution in aqueous media and the ionic strength is varied from 0.1 to 1.1 N. It is observed that in every solvent medium, lower ionic strength tends to make the wave more separated. In the aprotic solvents (THF, dioxane, acetonitrile, acetone and HMPA), the wave clearly splits into two steps in media of low ionic strengths and the wave heights are approximately in the ratio 1:3. This separation into two steps is a result of a small shift of the first step in the positive direction and a large shift of the second in the negative direction. The positive shift of the first step is found to be related to the donor number. The negative shift of the second step is related to the acceptor number of the solvent, except in the case of DMF and DMSO which fall apart from this trend. In the case of these two solvents and in the case of hydroxyl solvents (ethanol, ethoxyethanol and chloroethanol), the wave does become extended in media of low ionic strength but the tast polarograms do not show a clear two-step reduction even in media of low ionic strength. The shift of the steps with lowering ionic strength can be attributed only partly to the drop in the potential of the diffuse double layer. So this shift has its origin elsewhere. In all the media, whether we get a single step or two-step reduction, the initial step appears to be a one-electron addition to give a free radical anion PhNO??. This free radical anion being very reactive, reacts with the constituents of the media. It can be a solvent molecule or the ions of the supporting electrolyte; the solvent molecule may be that of either the organic component or of water. When the free radical anion is solvated with water or a hydroxyl solvent, the result may be the extraction of a proton from the solvent by the anion radical and the resulting neutral species PhNO?H could add on another electron more easily; further reduction is thus easy in hydroxyl solvents. Though DMSO and DMF are solvents of high acceptor numbers (DMSO = 19.3; DMF = 16.0), it is known that their ability to solvate an anion is much less than to solvate a cation. Thus in these mixed media, it may be the water molecule which solvates PhNO?? and the result is similar to that in hydroxyl solvents. In media containing the other aprotic solvents (acetonitrile, acetone, dioxane, HMPA and THF) the solvation may preferentially be by the organic solvent molecule and these solvated radical anion species are stabilised to a certain extent and thus resist reduction. The second step thus moves to more negative potential and two-step reduction is seen. This potential separation moves the first step slightly to positive values. A component in the medium which competes with the solvent species to react with PhNO?? is the cation of the supporting electrolyte. The radical anion is known to form ion-pair with cation and the ion-pair is known to be more easily reduced than the free radical anion. In media of high ionic strength, the concentration of cation being more, the probability for the formation of ion-pair is higher and further reduction of the free radical anion is favoured. So, in higher ionic-strength medium, the two steps come closer and merge together. Hence, the scheme of reduction can be given as follows: PhNO? + e? ? PhNO?? (8.1) PhNO?? + Solvent ? [solvated species] In hydroxylic media: PhNO?? ? PhNO?H ? further reduction ? products (8.3) Further reduction: PhNO?? + Cation ? ion-pair ? products (8.4) Step (8.4) is favoured in high ionic-strength medium in all solvents. In media in which the hydroxyl solvent molecules solvate the radical anion, step (8.3) is dominant. In other media, step (8.2) favours stabilisation of PhNO?? and stepwise reduction is seen. In cases where the reduction occurs in two steps, linear sweep voltammetric study shows clearly that the first step is.. Reversible and the second step irreversible. Cyclic voltammetry shows an anodic step corresponding to the first one-electron reduction in many experiments. With increase in scan rates, the second step shifts to more negative values and the separation of the two steps becomes better. This same effect brings about a separation of steps in some of the media which show only a single four-electron step in tast polarogram. Repetitive cycling of potentials also facilitates such separation of steps and separate steps are seen in third or later cycling in some media. The first step obviously corresponds to the reduction represented by equation (8.1). This step has been examined closely in cases where it occurs as a separate step. A small oxidation step corresponding to the oxidation of PhNO?? is seen in the reverse scan. Quantitative comparison of the wave heights of the oxidation step with the initial reduction step has not been possible in the present study, because of the practical difficulty of measuring the anodic wave height. Before the current dies down after the peak to the Cottrell region, the second reduction step starts and a real baseline for anodic step cannot be identified. However, it can be seen from the general behaviour that the oxidation step of PhNO?? is quite small at lower scan rates but improves as the time gap between its formation during the cathodic scan and its detection in the anodic scan decreases. As long as the potential is maintained within the first step or as long as the potential is not allowed to go to the potential of the second step, no other product is identifiable in the cyclic voltammograms. This is true even after repetitive scans extending over a few minutes. But if the potential is allowed to reach the second step or cross this step, phenylhydroxylamine can be identified in the medium by its oxidation step to nitrosobenzene during the anodic scan. The height of this oxidation peak is dependent on the potential of reversal. It is higher when the potential has reached values past the second step. This step is identified as due to the oxidation of PhNHOH by its potential and also the appearance of the corresponding reduction step of nitrosobenzene in the second and subsequent scan cycles. The ratio of the oxidation peak of PhNHOH to that of the reduction of PhNO? is >1, especially at slower scan rates, indicating the likely loss of PhNO?. Azoxybenzene, the product of the reaction of PhNHOH and PhNO?, cannot be observed clearly as a separate peak in cyclic voltammetry because of closeness of the reduction step of azoxybenzene and the second reduction step of PhNO?. However, the changes in the position and height of the second step indicate the formation of azoxybenzene. Corroborating evidence for the sequence of reactions is obtained by analysing the reaction mixture after different conditions and stages of electroreduction using HPLC. These results show that whatever be the appearance of the polarograms, the reduction of PhNO? is mainly one electron followed by 3 electron steps. The direct product of total reduction must be PhNHOH only, but this gets converted to azoxybenzene in alkaline medium. This conversion is facilitated by the presence of oxygen or intermittent oxidation and reduction. However, azoxybenzene formation occurs slowly even in oxygen-deficient media. Hydrazobenzene is only the reduction product of azoxybenzene. Azobenzene could be the oxidation product of hydrazobenzene. There are certain media (ethanol, chloroethanol, DMSO and DMF) which do not give stepwise reduction under most of the observed experimental conditions studied here. They give single four-electron irreversible step in cyclic voltammetry also. These also show the presence of phenylhydroxylamine in the reduced medium, thereby showing that the end product of reduction remains the same whether the reduction occurs stepwise or in a single step. Analysing the four-electron irreversible reduction steps of PhNO? in these media using Nicholson and Shain model of irreversible reduction gives ?n values which are scan-rate dependent. This is because, in these media, ? is potential dependent; ? at potentials Ep are different. It appears that though the cathodic step is not split up, the scheme of reduction remains the same as in the other cases with the two reduction steps being close by in these media. Both steps occur at potentials on the rising portion of the cathodic peak, by the contribution of two steps varying over the range of potentials. The ?n values would be potential dependent and the ? at Ep is not equal to the value at E?. The calculated E values at a given potential also may have contributions from the two steps. Study of the stepwise reduction by chronopotentiometry shows that the ratio of the step heights of the second to the first step is varying with concentration. In acetonitrile and acetone, the ratio of n?:n? (the number of electrons involved in the two steps) is less than 3:1 at low concentration of PhNO? and approaches the value 3:1 at high concentration. This indicates that there are some coupled chemical reactions between the two steps. To get a further insight into the mechanism of reduction, the first step has been closely examined in these media. The first step of reduction is represented by equation (8.1) and PhNO?? formed is stabilised to some extent as represented by equation (8.2). Current reversal chronopotentiometry shows that the first step is electrochemically reversible. This has also been observed in cyclic voltammetry. However, the wave height ratio of cathodic to anodic steps is not decisive in CV. In current reversal chronopotentiometry, the ratio of the wave heights of reduction and reverse oxidation steps clearly indicates that the product PhNO?? is not stable during the time scale of these experiments. This is being lost in a chemical step. Cyclic chronopotentiometry also supports this. Analysis of the results on the basis of well-known models of coupled chemical reactions like regeneration of PhNO? or slow decomposition of PhNO?? does not give constant values for the kinetic constant, thus indicating that there are more than one mode of reaction for PhNO??. The parallel modes of reaction contribute to different extents depending upon conditions. The transition time constants in a given solvent, calculated from experimental data, are found to be not really constant. They vary with the concentration of PhNO? and current density. A correction has been applied for the quantity of current used for processes other than the reduction of depolariser diffusing to the electrode, on the model given by Bard. This correction gives transition time constant which is independent of current density, but the correction shows that current consumed is less than the quantity required to reduce all the depolariser molecules diffusing to the electrode. There is something hindering part of the reduction that is normally expected of the diffusing depolariser molecules. It appears that some product of reduction is practically blocking part of the normal reduction of nitrobenzene. If this were due to the adsorption of the immediate product of reduction, in current reversal experiments, the adsorbed product should have given longer transition time for the oxidation step. That ratio then should be longer than 0.33 and should approach 1. This blocking could be due to adsorption of some secondary product of reduction which itself is not oxidisable at these conditions and hence cannot contribute to oxidation. However, it is likely that the reduction product PhNO?? is not getting adsorbed but it interacts with some of the incoming PhNO? molecules and causes hindrance to the reduction at the electrode. Hence one of the possible coupled chemical reactions is: PhNO?? + PhNO? ? Products (8.5) The concentration dependence of the current density corrected transition time constant shows that in the case of solvents with higher acceptor number (acetonitrile and acetone), the constant at low concentration is higher than expected, thereby suggesting that part of the PhNO?? participates in a reaction regenerating PhNO? at the electrode, the reduction of which gives higher current than expected. The reaction could be: PhNO?? + Solvent ? PhNO? + Solvent? (8.6) Solvent being one of high acceptor number, this is quite probable. In the five solvent media (aqueous solutions of acetonitrile, acetone, dioxane, THF and HMPA) the current density correction shows that reaction (8.5) is occurring. The contribution of this is minimum for dioxane followed by acetone and acetonitrile. The acceptor numbers of these solvents are: acetonitrile = 18.9, acetone = 12.3, dioxane = 10.8, HMPA = 10.6 and THF = 8.0. This contribution is larger in the case of HMPA and THF where reaction (8.6) is not dominant. Concentration dependence of the transition. Which is also aprotic in nature gave no such indication of the colour. In addition to all these observations, during cyclic voltammetry with repetitive scanning, in aqueous DMSO, indications for the existence of differently solvated species for PhNHOH and PhNO existing in equilibrium have been observed. This is quite possible in view of the fact that DMSO in about aqueous 60% solutions exists in hydrated form with no free water being left available for solvation. Hence in aqueous 40% solutions, there should be free water existing along with the hydrated DMSO. These two kinds of solvent species can solvate PhNHOH and PhNO and the differently solvated species exist in equilibrium and get oxidised/reduced at different potentials causing the appearance of two oxidation peaks and two reduction peaks. From an analysis of the variations in the peak heights with changes in scan rates and in concentration of nitrobenzene, the following schema of solvation equilibria coupled to reversible electron transfers are postulated, the differently solvated species being formulated as (PhNO) and (PhNO) for nitrosobenzene and (PhNHOH) and (PhNHOH) for phenylhydroxylamine. In the solvation equilibria III and IV, from the variation in the peak heights, it is found that k? < k? and k? > k?.