Electrogenerated Hydrogen Peroxide -
From History to New Opportunities
by Derek Pletcher
Hydrogen peroxide is probably a unique chemical, ideally suited to the present age where en-vironmental considerations are always to the fore.
Why unique? Firstly, it is capable of very diverse chemistry. Hydrogen peroxide may act as either an oxi-dizing agent or a reducing agent. As an oxidizing agent, its application ranges from highly selective oxida-tion chemistries applicable to the manufacture of many organic com-pounds, through the bleaching of pulp, to the total oxidation of large organic compounds to carbon dioxide. Its reactivity as an oxidizing agent is determined largely by the ratio of the concentrations of H2O2 to substrate and the reaction conditions, particularly the choice of catalyst and factors such as UV irradiation. Secondly, it is a strong oxidizing agent that may be formed by cathodic reduction under mild and varied conditions, opening up the possibility of producing the same product at both anode and cathode. Thirdly, the feedstock for electrogenerated hydrogen peroxide may be air (an unusually cheap and available feedstock!) while its reac-tions lead only to oxygen and/or water.
Why environmentally friendly? Again, because its reactions leave no residuals in the reaction stream and it may be manufactured with only air and electricity as feedstocks. In ad-dition, when used in dilute solutions such as those produced in elec-trolysis cells, its reactions are non-hazardous and carried out in very moderate conditions. In this respect, the chemistry of electrogenerated hydrogen peroxide should be distin-guished from that employing very concentrated hydrogen peroxide and peracids which can indeed be hazardous.
The markets for hydrogen peroxide are well established and include or-ganic oxidations, bleaching in the pulp and paper industry, applications in fabrication technology, and water and effluent treatment [1-4]. Indeed, it has been predicted that demand will grow by 10% per annum for the next few years [5]. The opportuni-ties for electrogenerated hydrogen peroxide arise now because of (i) interest in on-site manufacture of hydrogen peroxide (ii) the recent availability of technology, namely gas diffusion electrodes and three dimensional electrodes, which allow the reduction of oxygen to hydrogen peroxide at a practical current den-sity (iii) the recognition that the yield of hydrogen peroxide may be high in acidic and neutral solutions as well as strongly alkaline media (iv) the identification of new cata-lysts which expand the chemistry of hydrogen peroxide. All of these factors will be illustrated below.
History and Background
Electrolysis has an extensive history in the manufacture of hydrogen per-oxide. For many years, all hydrogen peroxide was manufactured by elec-trolysis using a route where persul-fate is formed at the anode and then hydrolysed [1]. Although there have been recent proposals to reinstate this technology [6], it has generally been considered to have an excessive energy requirement for large-scale production. Hence, at the present time almost all hydrogen peroxide is manufactured by a route involving the reduction of oxygen by hydrogen with anthraquinone as the catalyst [1]. On the other hand, the anthraquinone route requires the availability of hydrogen feedstock and the large-scale use of non-aque-ous solvents for catalyst recycle. Hence, it is unsuited to small-scale manufacture or to on-site production of hydrogen peroxide.
Small scale, on-site processes are, however, of increasing interest because of the cost and hazards associated with the transport and handling of concentrated hydrogen peroxide. Particular attention has been focused on processes for the pulp and paper industry which require ( 2 % HO2- in 1.5 M NaOH. Interest has concentrated on the manufacture of hydrogen peroxide by the cathodic reduction of oxygen:
O2 + 2H2O + 2e- ( H2O2 + 2OH-
or
O2 + 2H+ + 2e- ( H2O2
These reactions occur in high yield at certain cath-odes including mercury, gold and carbon. Practical utility of these reactions demands the choice of a carbon cathode. In addition, com-mercial exploitation has required the development of technology to over-come the problems resulting from the low solubility of oxygen in aqueous solutions; at a plate elec-trode, the mass transport limited current density is << 1 mA cm-2 (equivalent to << 20 (moles cm-2 hour-1). As a result, processes have been developed based on both various types of three dimensional electrodes such as beds of carbon particles or reticulated vitreous carbon and gas diffusion cathodes fabricated from carbon powders without metal catalysts [7-13]. Both types of electrode permit an increase by a factor of 30 - 1000 in the rate of hydrogen peroxide production. Dow [11] and Huron Chemicals [6] have described processes which em-ploy cathodes fabricated from a bed of carbon chips, 0.5 - 2 mm in size and coated with a high surface area carbon powder/PTFE composite through which an oxygen saturated NaOH solution is allowed to trickle so as to give a film of catholyte over the surface of the cathode within an oxygen atmosphere. Such cells give a current efficiency 60-70 % at 15 mA cm-2. E-Tek [12,13] have pioneered the use of gas diffusion electrodes and demonstrated that HO2- in NaOH solutions can be pro-duced at a current efficiency of > 90 % at current densities > 100 mA cm-2. McIntyre [14] has also reported the development of cathode catalysts for an O2/H2 fuel cell so that the reaction of O2 and H2 leads to hy-drogen peroxide with an output of electrical energy from the cell.
Figure 1: Reaction of 2,4,6 trimethylphenol with electrogenerated H2O2 Catalyzed by HRP
While these technologies appear to be moving towards commercial success, published data on pilot and production scale developments are, however, limited to the manufacture of hydrogen peroxide in concen-trated alkaline solutions. Such media are not of great interest in synthesis or effluent treatment. On the other hand, they demonstrate the way forward and recent papers have con-firmed that both three dimensional electrodes made from reticulated vitreous carbon [15,16] and gas dif-fusion electrodes [16-20] may also be used to reduce oxygen to hydro-gen peroxide in both neutral and acid solutions at rates which are ap-propriate to the needs of synthesis and effluent treatment. Some exam-ples of the use of electrogenerated hydrogen peroxide discussed below already use such electrodes although most authors have chosen to demon-strate the concept of using electro-generated hydrogen peroxide using plate electrodes and accepting the very low current density. It is, however, the introduction of three-dimensional electrodes and gas dif-fusion cathodes that will transform these concepts into useful laboratory procedures and candidates for industrial exploitation.
Applications in Electrosynthesis
Much of the early work emphasized the conversion of aromatic hydro-carbons to phenols and/or aldehydes using an acidic catholyte containing Fe(III) and O2 [21-24]. The cathode was used to produce both con-stituents of Fentonís reagent, Fe(II) and H2O2, and typical reactions were the oxidation of benzene to phenol and of toluene to benzaldehyde. The reactions gave good yields (60 - 100 %) provided over-oxidation was avoided by stopping the reactions at very low conversions or con-tinuously extracting the products from the catholyte (often using a suspension of the hydrocarbon substrate in the aqueous catholyte). Some studies employed mercury cathodes but the more practical procedures used graphite plate cathodes; in consequence, an acceptable current efficiency could only be achieved at very low current densities. It would be timely to re-investigate these oxidations with three-dimensional and/or gas diffusion cathodes. Otsuka [25,26] has described preliminary but unconvincing attempts to carry out such oxidations at the cathode of an O2/H2 fuel cell with an energy output.
More interesting, however, are studies which use more selective catalysts than Fe(II) since Fentonís reagent is known to lead to the for-mation of hydroxyl radicals. The most selective catalysts for the reactions of hydrogen peroxide are the peroxidase enzymes; these en-zymes have a diverse chemistry, which often leads selectively to unusual products as well as the introduction of chirality [27,28]. Examples of reported reactions in-clude the conversion of sulfides to sulfoxides, of anilines to hydroxylamines or nitrosobenzenes, olefins to olefin oxides, alkyl ben-zenes to alcohols or aldehydes and indoles to oxindoles. The Southampton Group [16] have demonstrated the use of horseradish peroxidase as a catalyst for the reac-tion of electrogenerated hydrogen peroxide and 2,4,6-trimethylphenol in a phosphate buffer. Electrolysis is shown to be a convenient and controlled way to carry out such enzyme catalyzed reactions. In ad-dition, it appears to introduce the possibility of new reaction pathways and products from these biosynthetic reactions. The major products from the electrolysis can be, as in the classical chemical procedure, formed via the pathway shown in Figure 1, with either II, yield 68% after 3F of charge or IV, yield > 70%, after 8F of charge as the major products. These yields are significantly higher than those reported for classical chemical procedures for this reaction. In a cell with a very high cathode area to catholyte volume (e.g. a cell where the reticu-lated vitreous carbon cathode fills the catholyte compartment), a fur-ther product was identified as shown in Figure 2.
Figure 2: HRP-catalyzed reaction of electrogenerated H2O2 with trimethylphenol in a cell with high area to volume ratio.
The yield of the interesting coupled product V could be > 50 % and its formation has not been reported for chemical reactions. This implies that its for-mation results from the cathodic reduction of an intermediate in the enzyme reaction cycle. The use of a reticulated vitreous carbon electrode also allows the reactions to be carried out at a significant current density.
Another recently reported example of a selective cata-lyst for electrogenerated hy-drogen peroxide is tungstate [29,30]. Nonaka et al have shown that when hydrogen peroxide is formed at a graphite plate elec-trode in a phosphate buffer, pH 5, at 333 K, it reacts with tungstate
Figure 3: Tungstate - catalyzed reactions of electrogenerated H2O2.
HWO4- + H2O2 ( HWO5- + H2O
and describe the in situ use of the pertungstate for the oxidation of sulfoxides to sulfones and of di-butylamine to N-butylidene-butylamine N-oxide as in Figure 3. Both reactions give good selectivity and an acceptable current efficiency and neither occurs in the absence of tungstate in solution. In addition, the literature provides many pointers to other suitable catalysts for reactions of electrogenerated hydrogen per-oxide. These would include MoO4-, polyoxometalates containing transi-tion metals such as Ru, Mn and Cr, SeO2, RuO2, CH3ReO3 and Sharpless Os(VI) reagents.
As noted in the introduction, the formation of an oxidizing agent, hydrogen peroxide, at the cathode introduces the possibility of carrying out oxidations at both electrodes and, indeed, carrying out the same reaction at both electrodes, thereby (a) doubling the rate of chemical conversion (b) halving the number of cells required for the wanted annual tonnage as well as (c) halving the process energy con-sumption. A group in Taiwan has demonstrated this approach. [31-33]. The earliest example was the oxida-tion of toluene to benzaldehyde. A divided cell was used and the anode was used to oxidize Mn(II) to Mn(III) in sulfuric acid while the graphite cathode was used to reduce oxygen to hydrogen peroxide. The oxidation of toluene was carried out in situ in the catholyte in the presence of V(IV) as a catalyst while the Mn(III) was reacted with toluene external to the cell after the electrolysis. The overall current efficiency is reported as 171% (84% at the anode and 87% at the cathode) and the selectivity is also excellent.
It was also possible to employ a similar approach to the oxidation of anthracene to anthraquinone and this could even be carried out in an un-divided cell using a concentrated sulfuric acid electrolyte. Here, the mediator at the anode was the V(IV)/V(V) couple and the V(IV) also acted as the catalyst for the reaction of hydrogen peroxide formed at the graphite plate cathode. The current efficiency for anthra-quinone was 151% and the selectivity 98%.
The third system was the oxidation of n-butanol to n-butyric acid in basic solution and again it was possible to use an undivided cell. Here, the anodic oxidation was a direct process at a nickel anode and the hydrogen peroxide formed at the cathode reacted with the alcohol without an added catalyst. The cur-rent efficiency was > 170 % and no by-products were identified. Nonaka et al [29] have also used this concept in the oxidation of a NñOH group to a N(O group; the oxidation was carried out in an undivided cell where bromine was formed at the anode and hydrogen peroxide at the cathode (in an electrolyte containing tungstate). The total current effi-ciency for the formation of N-butylidene-butylamine N-oxide was > 180 %.
Effluent Treatment
In effluent treatment technology, the objective may be the complete oxi-dation of all organics in the stream to carbon dioxide, the removal of particular toxic compounds or just decolorization of the stream. Fentonís reagent is not a specific oxidant. It reacts with many molecules and in situations where the ratio of [H2O2]:[organic] is high, complete oxidation of organics to CO2 would be expected. Hence the electrogeneration of hydrogen per-oxide in the presence of Fe(II) should be an effective method of effluent treatment. In commercial application it must be envisaged that the source of oxygen is air and the Fe catalyst as well as any proton requirement results from the anode chemistry in the cell (i.e. the anode has a component of iron capable of dissolution). Several papers have demonstrated that electrogenerated H2O2 may be used successfully for effluent treatment. The emphasis in such studies is on the treatment of solutions containing < 1000 ppm COD (chemical oxygen demand). The reason is associated with energy consumption. The energy consump-tion for an effluent treatment process (usually quoted as kWh m-3) is proportional to the concentration of organic in solution and the number of electrons involved in the conver-sion of the organics to CO2; commonly, the oxidation of an or-ganic molecule may involve 20 - 150e- and for higher concentrations, the energy consumption becomes prohibitive. Many chemical effluent streams, however, contain relatively low COD.
Two of the earliest papers [34,35], using hydrogen peroxide generated at a graphite plate, examined the destruction of phenol and formalde-hyde as a function of pH. At pH 1 - 4, it was shown that in the presence of Fe(II), phenol (COD 260 - 2600 ppm) could be effectively oxidized and largely converted to CO2; final COD was < 40 ppm. The authors also followed the intermediates formed during the oxidation and were able to identify at least six compounds; it is clear that the oxidation pathway is complex but also that the H2O2/Fe(II) combina-tion is able to decompose a wide range of organic structures. Formaldehyde was also readily oxidized. At pH 13, the oxidation of formaldehyde is efficient but the product of oxidation was formic acid. In acidic solutions where it was possible to add Fe(II), complete oxidation to CO2 took place. These papers confirmed that hydrogen peroxide could be formed at a graphite cathode in conditions appropriate for effluent treatment and also that the chemistry could be favorable for complete oxidation. The current densities were, however, typically 0.4 mA cm-2 and clearly this is not sufficient for practical effluent treatment technology.
To overcome this difficulty, Pletcher et al [15,16] used a reticulated vitre-ous carbon cathode. The first sub-strate to be investigated was for-maldehyde and it was confirmed that the three dimensional electrode greatly accelerated the rate of for-maldehyde removal; with oxygen saturated solutions, current densities could reach > 20 mA cm-2 (higher values could also be reached using a thicker cathode). Formaldehyde (5 - 200 ppm) oxidation could be achieved in alkaline, neutral and acidic media but Fe(II) as a catalyst was essential to achieve complete oxidation to CO2; in its absence, the reaction stopped at formic acid. More recently [16], the study has been extended to sulfate media, pH 2, containing phenol, cresol, catechol, quinone, hydroquinone, aniline, oxalic acid or the azo-dye, amaranth. With each compound, COD could be reduced from 50 - 500 ppm to below 10 ppm, generally with a current efficiency > 50% and using a current density of ( 20 mA cm-2. Much of the carbon in the solutions could be identified as carbon dioxide after the electrolyses. If the objec-tive were to decolorize a dye effluent stream, this could be achieved rapidly and with only a low energy consumption (the re-moval of color requires only a rela-tively small change in chemical structure).
Brillas and co-workers [17-20] were the first to introduce gas diffusion electrodes into these effluent treat-ment processes. They have described the oxidation of aniline and 4-chloroaniline in aqueous solu-tions with a pH 10.1 - 12.7 and they were able to reduce the COD of such solutions from 100 ppm to < 5 ppm with a current density as high as 200 mA cm-2. They also consider sulfate solutions with pH ( 3 where complete oxidation of the aniline was again observed provided Fe(II) was present. It was also found that UV irradiation accelerated oxidation. In the acid solutions, the current densities for the reduction of oxygen to hydrogen peroxide were lower, ( 30 mA cm-2. The final paper of the series confirmed that the Fe(II) could, indeed, be formed in situ in an undivided cell by using an iron anode and the Fe(II) was sufficiently active to allow the rapid degradation of both 100 ppm and 1000 ppm aniline solutions.
Harrington and Pletcher [16] have also used an oxygen diffusion cath-ode at pH 2 - 3. They investigated the voltammetry of the solutions and concluded that the limitation on the oxygen reduction current densities to 50 - 100 mA cmñ---2 (cf. 200 - 500 mA cm-2 in strongly acidic or alkaline solutions) resulted from IR drop within the pores of the electrode structure. Even so these current densities are of a practical level and it was shown that the gas diffusion electrode could be used to remove phenol, aniline, acetic acid, for-maldehyde and three azo-dyes (amaranth, fat brown RR and methyl orange) provided Fe(II) were present in the solution; the COD of solutions containing such organics may be reduced by > 90 % with a current efficiency > 50 %, leading to acceptable energy consumption.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.