1 Introduction

Extensive research on catalysis by supported gold has been reported since the pioneering discoveries by Hutchings [1] and Haruta [2] demonstrating high catalytic activities of highly dispersed gold. CO oxidation and the water gas shift are among the best investigated of the reactions catalyzed by supported gold; most of the work has focused on the former [3], as it apparently offers the advantages of taking place at low temperatures combined with the simplicity of small reactant molecules and the value of CO as a sensitive probe of surface structure [4].

Notwithstanding the extensive research on supported gold catalysts for CO oxidation, the mechanism(s) of the reaction and the catalytically active species remain matters of debate, and the reports of quantitative kinetics of the reaction, although numerous, are largely incomplete. The lack of thorough kinetics data reflects the complexities of the catalyst performance, influenced by catalyst activation and deactivation, which are often rapid; it is sometimes difficult to determine from published reports whether the reaction rates or conversions characterize fresh or deactivated catalysts.

Our goal was to provide a summary facilitating access to the literature of the kinetics of CO oxidation catalyzed by supported gold. The literature is summarized here in tabular form; earlier, much less complete summaries were reported by Bond et al. [5], Deng et al. [6], and Kung et al. [7]. Some issues regarding the challenges of comparing supported gold catalysts on the basis of performance were addressed by Long et al. [8]. We have limited the content here by excluding catalysts with doped supports (except when they were part of a set including undoped supports) and results characterizing “preferential oxidation” of CO in the presence of excess H2. Otherwise, the compilation contains most of the literature that includes kinetics data for CO oxidation catalyzed by supported gold, although it is not exhaustive, with a number of examples of only partially documented kinetics data being omitted.

2 Tables of Data

The data are presented in three tables, with the entries linked by the entry number shown in the left-hand column of each table. Table 1 is a list of supported gold catalysts used for CO oxidation, how they were made and treated, their gold contents and surface areas, and the average gold particle sizes and methods used to determine them. Table 2 is a summary of the conditions under which the kinetics data were determined, with information about the degree of deactivation of the catalyst. Table 3 is a summary of the kinetics data, including values of TOF and how they were determined, reaction orders, and apparent activation energies.

Table 1 Characteristics of the supported gold CO oxidation catalysts
Table 2 Reaction conditions under which the supported gold CO oxidation catalysts were tested
Table 3 Kinetics data reported for the supported gold CO oxidation catalysts considered in this work

We believe that these tables provide the most complete available statement of kinetics of CO oxidation catalyzed by supported gold.

3 Generalizations Based on the Data

Table 2 is a summary of the catalysts tested for CO oxidation; the catalysts were investigated at temperatures in the range of 203–373 K. Haruta [35] referred to a low-temperature regime (typically, ~210 K) and a high-temperature regime (typically, >300 K). The O2 partial pressures were varied between 4 and 200 mbar, and the CO partial pressures between 10 and 40 mbar. The results indicate orders of reaction in CO and in O2 in the range 0.0–0.6. The reaction order in CO has been approximated as zero by some researchers [24]. Correspondingly, numerous researchers have postulated that CO is adsorbed on the gold; some [4] have suggested that CO is bonded to gold at the gold-support interface.

The roles of oxygen in the gold-catalyzed CO oxidation are evidently not fully elucidated. Some authors have postulated that oxygen adsorbed on the gold [4] or at the gold-support interface [36] may play a role. In contrast, Guzman et al. [37] reported evidence of the involvement of reactive oxygen species (such as superoxides) on their CeO2 support; the influence of the presence of reactive oxygen species on some supports but not on others (e.g., γ-Al2O3 [38]) would suggest that the form of kinetics would differ from one support to another, but there are too few data to test this statement.

A few reports of the influence of CO2 on the rate indicate that it inhibits the reaction; according to one report [16, 22], the desorption of CO2 from Au/TiΟ2 is rate limiting under some conditions. Others [39] have reported that CO2 (rather than O2) is the oxidizing agent of gold in supported gold catalysts, implying that the gold in the catalytic sites cycles between more than one oxidation state.

Haruta’s group [40] reported a detailed investigation of the influence of water in the reactant stream on CO oxidation catalyzed by TiO2 , Al2O3 , and SiO2 supported gold. Water in low concentrations increases the activity of the catalyst.

The most thorough investigation of the kinetics of CO oxidation catalyzed by supported gold was reported by Vannice’s group [9]; the catalyst support was TiO2. The authors tested several catalysts that had been subjected to various pretreatments, and kinetics parameters are reported for each (entry numbers 1–9 in Tables 1, 2, 3).

Many of the most active supported gold catalysts for CO oxidation are supported on TiO2 or on various oxides of iron or of cerium. Turnover frequencies (rates of reaction per accessible gold site; Table 3) span a wide range, between 10−6 and 10−1 s−1. There is one report of an intrinsic turnover frequency—that is, per active site [41] (entry numbers 55, 56, 64, 76, and 77, Tables 1, 2, 3)—determined in transient measurements with isotopically labeled reactant 13CO for Au/γ-Al2O3; the value is 1.6 × 10−1 s−1 at 296 K and CO and O2 partial pressures of 24.2 mbar each.

Only a few values of apparent activation energies of CO oxidation catalyzed by gold have been reported, and the information about the conditions under which they were determined is often lacking. The apparent activation energies range from values that are essentially indistinguishable from 0 to 138 kJ/mol (Table 3).

Most reports of catalyst deactivation and how it occurs (e.g., [11]) do not include kinetics data, but the work of Vannice’s group [9] is exceptional, providing kinetics data for various catalysts before and after deactivation (Table 2).

Supported gold catalysts typically undergo rapid deactivation during CO oxidation, and this complication has hindered the collection of kinetics data. For example, the initial conversion observed with a zeolite-supported gold catalysts was about 40%, and this decreased to <5% within 15 min of operation in a once-through flow reactor at 298 K [42]. An Au/TiO2 [27] catalyst, on the other hand, showed an initial conversion at 303 K of nearly 100%, and the conversion had declined to 10% after 2,000 min of operation in a flow reactor when O2 was present in stoichiometric excess; but the decline in activity was more rapid when the O2 was not present in stoichiometric excess. Other authors have also observed that the rate of catalyst deactivation was less when the reaction took place in an O2-rich atmosphere [25].

It is clear that the available data do not lend themselves to conclusive integration and that much work remains to be done to consolidate the literature and to represent CO oxidation catalyzed by supported gold quantitatively.

4 Conclusions

The results summarized here show that the literature of CO oxidation catalyzed by supported gold is extensive but fragmented and not easily generalized; it is not easy to make meaningful comparisons of various supported gold catalysts for this reaction, and much work remains to be done to consolidate the literature of CO oxidation catalyzed by supported gold.