A treatment of the Temkin model was developed for describing the adsorption of CO on gold. In this presentation, it is made clear that the Temkin thermodynamic model is an extension of the Langmuir model that incorporates a linear variation of the adsorption enthalpy. It is also stressed that the Temkin model can be interpreted in terms of two distinct physical situations: the first assumes equivalent binding sites and an adsorption enthalpy that varies with coverage due adsorbate interactions, while the second assumes a uniform distribution of heterogeneous binding sites and an adsorption enthalpy that varies due to the heterogeneity of sites. In addition, a midrange approximation, where the surface is roughly half covered with adsorbates, is commonly employed. While each of these situations is similar, they are not equivalent and yield slightly different analytical expressions that describe the adsorption coverage as a function of pressure and temperature. Fitting data with the different analytical expressions therefore produces slightly different thermodynamic values for the adsorption enthalpy and entropy. These different cases are explicitly defined and developed. For the adsorption of CO on gold, the adsorbate interaction case is shown to be most consistent with the current body of experimental and theoretical evidence. The various analytical expressions are used to apply the Temkin model to data for CO adsorption on 1% Au/TiO 2 real-world catalysts (from the World Gold Council) studied under catalytically relevant isothermal (T = 275À325 K) and isobaric (P CO = 0À10 Torr) experimental conditions. Infrared transmission spectroscopy was the analytical technique used for quantitatively measuring the adsorption coverage. The coverage as a function of pressure (θ,P) and as a function of temperature (θ,T) was fit with the Temkin adsorption models. The models yield the thermodynamic adsorption enthalpy at zero and full coverage as well as the adsorption entropy. The values of these thermodynamic parameters differ slightly depending upon the particular Temkin situation considered (cf. adsorbate interaction, heterogeneous surface, and midrange approximation cases). While all three Temkin cases produced excellent fits to both the isothermal and isobaric data sets, the adsorbate interaction case is most consistent with experimental and theoretical evidence describing the adsorption of CO on gold. The average enthalpy values from fitting the isothermal and isobaric data sets using the adsorbate interaction model are ÀΔH 0 = 59.2 kJ/mol and ÀΔH 1 = 54.6 kJ/mol for zero and full coverage, respectively. The adsorption entropy, ÀΔS = 142 J/(K mol), was determined by fitting the data sets from multiple isothermal experiments with the adsorbate interaction case. These thermodynamic adsorption values are in excellent agreement with previously reported values. The validity of the Temkin adsorbate interaction model was further supported by fitting isothermal and isobaric data for the adsorption of CO on a well-defined gold sur...

The adsorption of CO on three different gold nanoparticle catalysts supported on high surface area TiO 2 was studied using infrared transmission spectroscopy at room temperature and CO pressures typically used in CO oxidation reactions. The three, real-world catalysts were Au catalysts synthesized in our laboratory from thiol monolayer protected clusters (MPCs) and two commercial catalysts from the World Gold Council (WGC and AuTEK). Within experimental reproducibility, the adsorption data for the three catalysts are indistinguishable. While showing approximately Langmuir behavior, the adsorption data also show coverage dependence, as others have observed for many catalyst systems. Two approaches were used to fit the data, a two-site model and a variable binding constant model. The two-site Langmuir model yielded strong (36%) and weak (64%) binding constants of 2740 and 146 atm -1 , respectively. Alternatively, using a sliding-tangent Langmuir fit gave a variable binding constant of 2670-120 atm -1 at room temperature for coverage θ ) 0-0.8. The heat of adsorption was then extracted from the binding constants using a literature value for -T∆S. These values were determined as ∆H ) -64 and -56 kJ/mol for strong and weak binding according to the two-site model and ∆H ) -63 to -56 kJ/mol for coverage θ ) 0-0.8 for the variable binding constant model. These values agree well with literature values obtained (i) using supported catalysts under higher pressures and (ii) using model catalysts under higher pressures and ultrahigh vacuum conditions.

No abstract