In-situ Changes in Surface Activity

An important factor with regard to the operation of an electrocatalyst is that these materials usually operate in contact with an electrolyte. A solvent such as water is not inert, it frequently interacts not only with the other reactants or intermediates at the interface but also with the electrode (or the electrocatalyst) itself. It is well established for instance that with platinum (an efficient electrocatalyst for many reactions) in aqueous media (5) hydride (or H_ads) or oxide (OH_ads and O_ads) films are formed on the metal surface over certain ranges of potential, see Figure 1. It is generally accepted that, as is clear from Figure 1, oxygen gas evolution at a platinum anode surface (which usually occurs well above the reversible potential for this reaction, 1.229 V(RHE) at T = 298 K) occurs not at the metal but at the oxide-coated metal surface (1). The solvent is not the only species that can react with, and modify the reactivity of, the surface of the electrocatalyst – so also can electrolyte anions such as Cl, or , adventitious impurities and organic reactants. For example in the oxidation of methanol, at the anode of a direct methanol/air fuel cell, at low pH the process of interest is complete conversion according to the following reaction:

[ii]

However, this process tends to be severely inhibited due to the accumulation of a CO_ads poisoning species produced at the interface via the following type of side reaction (6):

[iii]

Other alterations in the electrocatalyst that may occur in practice include dissolution -resulting (with multicomponent systems) in surface enrichment phenomena – and sintering (which results in loss of true surface area and surface density of active sites).

The Gap between Ilieory and Practice

A rather disconcerting feature in the area of heterogeneous catalysis in general is the presence of a wide gap between theory and practice.

Thus, Bond has pointed out recently that “Reflections on the research carried out in the field of heterogeneous catalysis over the last four decades will show not so much a catalogue of problems solved as of problems shelved” (7). At approximately the same time Schlögl claimed that “All the catalytic processes vital to industry were developed by purely empirical methods and countless screening experiments” and “Our understanding of the relationship between structure and reactivity is inadequate” (8). The situation with regard to electrocatalysis is not different; the empirical nature of this area was pointed out earlier by Pletcher who also stated that “there remains a gulf between those who developed electrode materials and those who seek to understand the physical chemistry of electrocatalysis” (9).

A major factor contributing to the lack of understanding may well be insufficient attention devoted to defect centres or active sites at solid surfaces. A typical bulk metal such as platinum or gold has a well characterised 3-dimen-sional lattice structure – and the surface is frequently envisaged as a planar termination of such a lattice. Such a surface is easier to comprehend and model (as compared with its defect counterpart) and considerable attention has been devoted in recent years to producing, and investigating the activity of such single crystal plane electrode surfaces (10). They are obviously excellent substrates for investigation by newly developed high resolution microscopy (for example scanning tunnelling and atomic force (11)) techniques. The problem is the major gap (discussed by Schlögl (8)) between single crystal plane surfaces and those used in practice, for example platinum dispersed at quite low coverage, as an electrocatalyst, on carbon at the cathode of a fuel cell. The ill-defined surface of the metal in such a state is not easily modelled and as is clear, for instance from a recent conference proceedings on electrode/electrolyte interfaces, the role of defects, at a fun-damental level, is widely ignored at present (12). This is quite surprising as the “active site” concept, originally proposed by Taylor, has been well established in the empirical interpretation of heterogeneous catalysis for many years (13).

A further complication in electrocatalysis is the possibility that low co-ordination surface metal atoms exhibit significantly different redox behaviour as compared with their high co-ordination equivalents at the same surface. The basis of this approach, which was discussed in some detail in a recent controversial paper from this laboratory (14), is supported by the following statement by Adamson (15):

Low co-ordination, energy-rich, surface metal atoms are unusual in several respects. Firstly, because of the lack of crowding at the surface, these atoms react with solvent molecules (or other species) in an unusually rapid manner. Secondly, these less crowded atoms, on oxidation, can co-ordinate more oxygen or hydroxide species and hence may form hydrous, rather than anhydrous, oxides. Finally, because the reactive metal atoms (the reduced form of the couple) are energy (or activity)-rich, they generally oxidise at a lower potential than high coordination surface metal atoms in a process known as premonolayer oxidation. As already discussed earlier (14) anomalous oxidation is an independently established feature of the reaction of supported platinum clusters with gaseous oxygen (16). The formation of low coverage hydroxy species on gold in base at premonolayer potentials has already been established using Surface Enhanced Raman Spectroscopy (SERS) (17). Unfortunately, SERS does not yield precise information as to the identity of the product and the technique is applicable, to a large extent, only to the Group 1(b) metals (gold, silver and copper).

A Revised View of Noble Metal Electrochemistry

As pointed out in an earlier review, hydrous oxides are now known and have been investigated for most of the noble, and several non-noble, metals (18). They are best regarded as a distinct series of compounds (14) that differ from their regular, anhydrous, oxide or hydroxide equivalents in that they frequently exhibit poor structural regularity combined with high redox activity; the latter is particularly evident in systems such as iridium, rhodium, iron and nickel hydrous oxides (18) where intermediate redox transitions, for example those between the II/III or IlUTV oxidation states (of the same cation), are involved.

It is assumed that all real metal surfaces contain poorly stabilised, high energy, atoms; these may be concentrated at unusual sites, such as tips of asperities, or occur in a random manner (due to thermal vibrations) on the surface. Normally the coverage of such species is quite low and both the electrochemical and spectroscopic responses due to their oxidation is usually quite weak. As mentioned here earlier there is spectroscopic (SERS) evidence for the formation of hydroxy species on gold in base at low potentials. All noble metals are assumed to exhibit such premonolayer oxidation behaviour and – although the evidence is not yet conclusive – the correlation (as discussed here below) between hydrous oxide electrochemistry and the electrocatalytic behaviour of the noble metals, both with regard to oxidation and reduction processes, is impressive.

From an historical viewpoint the conventional view of noble metal surface electrochemistry -hydrogen, double layer and oxide region in the case of platinum or double layer and oxide region in the case of gold - received early acceptance. However, while such an interpretation is evidently valid for the relatively well stabilised surface metal atoms (the vast majority of such atoms at the interface), it does not take into account the redox behaviour of the very minor coverage of energy-rich, poorly lattice stabilised, surface metal atoms. The low responses for the latter (as they undergo anomalous oxidation – at potentials well below that for the onset of regular monolayer formation) is evident in certain instances, for example silver (19) and gold (20); the premonolayer response for the latter is sometimes attributed (21) – incorrectly in the view of the author (and especially in view of the SERS data) – to the oxidation of impurities.

The recent claim (14) that premonolayer oxidation of platinum occurs within the hydrogen adsorption region may be controversial, but the observation of enhanced premonolayer oxidation of both platinum (22) and gold (23), plus charge displacement (24) from the hydrogen adsorption to the double layer region of platinum in acid (following prolonged polarisation at low potentials), clearly shows that the basic electrochemistry of the noble metals is more complex than is generally assumed.

The Incipient Hydrous Oxide/Adatom Mediator Model of Electrocatalysis

The Incipient Hydrous Oxide/Adatom Mediator (IHOAM) model of electrocatalysis (14) is based on the assumption that the low coverage (or incipient) hydrous oxide, M(OH)_x, formed in the premonolayer oxidation process, is the effective oxidant (or mediator) in electrocatalytic oxidation processes, see Scheme I, whereas the low co-ordination surface metal atom, M*, is the effective reductant (or mediator) in electrocatalytic reduction processes, see Scheme II.

According to these schemes the mediator is restored to its initial state by externally supplied anodic (i_a) or cathodic (i_c) current, that is electrocatalysis involves an interfacial cyclic redox transition between two unusual states - the low co-ordination surface metal atom and the incipient hydrous oxide.

The above approach (which is controversial) has been outlined in considerable detail recently for both gold (25) and platinum (14). Significant support is provided by the data summarised in Figures 2 and Figures 3. A schematic outline of the oxide (including the hydrous oxide) electrochemistry of gold in aqueous media is given in Figure 2(a): a more detailed account (including an explanation of the shift in the hydrous oxide/metal transition to a lower potential on increasing the solution pH) of the behaviour of this system has been published (14, 25). It is clear from Figure 2(b) that numerous electrocatalytic oxidation and reduction processes commence and terminate in the region of the active metal/hydrous oxide transition. Similar data is shown in Figure 3 for platinum – an account of novel aspects of the oxide electrochemistry of this metal has been published (26, 27).

Some general points concerning the data in Figures 2 and Figures 3 are noteworthy:

Conclusions

The major role of active sites or surface irregularities in determining both the catalytic and electrocatalytic activity of surfaces is widely accepted (31, 32). Such sites are frequently regarded in terms of co-ordinatively unsaturated surface atoms (or groups (33)) but precise details with regard to structure and electronic properties of such species are generally not available (32). The novelty of the present approach is that it assumes (largely on the basis of the consid-erable knowledge now available on noble metal hydrous oxide behaviour, correlations between this and electrocatalysis – see Figures 2 and Figures 3, and SERS data for gold (17)) that low co-ordination surface metals or groups exhibit a significantly different type of redox behaviour as compared with that of similar but high co-ordination number surface species.

A major practical problem in this area is the difficulty in detecting, and investigating the properties of, the quite low coverage, relatively labile, species involved. In addition, there is a range of practical considerations, for example it would obviously be useful to investigate methods for the optimisation of the activity, effective coverage (which is not necessarily the maximum coverage) and long term stability of a particular mediator.

Finally, the concept of activated chemisorp-tion should not be discarded. In some instances, such as hydrogen gas evolution on platinum, this may well be the sole mechanism of reaction. In other cases, such as methanol oxidation at a platinum/ruthenium-activated anode (3), the electrocatalyst may operate in a Afunctional manner – the alcohol may be activated by chemisorption on platinum sites while the ruthenium functions as the oxygen transfer agent (to produce carbon dioxide). The active platinum atom may still undergo an intermediate redox transition in the course of the latter reaction, that is there may be an intermediate transfer of an oxygen species (derived from water) from a ruthenium to a platinum centre.