Temperature-programmed Reaction Spectra Research Articles

Dual decomposition pathways for L-aspartic acid on Ni(100)

The thermal decomposition of aspartic acid (Asp, HO2CCH(NH2)CH2CO2H) on Ni surfaces occurs by a more complex pathway than on Cu surfaces where it decomposes via an aspartate species to yield CO2, H2, and CH3C≡N stoichiometrically. In addition to these three products observed on Cu, the decomposition mechanism on Ni yields CO, H2O and HC≡N. Herein, we probe the Asp decomposition mechanism on Ni(100) by X-ray photoelectron spectroscopy and by temperature programmed reaction spectra of a variety of adsorbed aspartic acid isotopomers (L-Asp, 1–13C-L-Asp, 4–13C-L-Asp, 1,4–13C2-L-Asp, 3,4–13C2-L-Asp, 1,2,3,4–13C4–15N-L-Asp, 15N-L-Asp, and D7-L-Asp). These isotope labeling experiments provide insight into the regiospecific origin within adsorbed aspartic acid of the carbon and nitrogen atoms found in the desorbing products. At low coverage, the adsorbed aspartic acid primarily undergoes inter- and intra-molecular condensation/dehydration to ketone and/or anhydride intermediates that ultimately yield CO. At higher coverages, the decomposition process is similar to that observed on Cu surfaces yielding mostly CO2 from the terminal carboxyl groups and CH3C≡N from the interior of the adsorbed aspartic acid. The continued presence of CO and H2O products at high coverages suggests the persistence of the condensation/dehydration reactions across all coverages.

Kinetics of low-temperature methane activation on IrO2(1 1 0): Role of local surface hydroxide species

The ability of the IrO2(1 1 0) surface to promote CH4 activation at low temperatures (~150 K) suggests possibilities for developing IrO2-based catalysts to selectively convert light alkanes to high-value chemicals. In this study, we present experimental results and microkinetic simulations, based on density functional theory (DFT) calculations, of temperature programmed reaction spectra (TPRS) obtained for CH4 σ-complexes adsorbed on IrO2(1 1 0), and focus on clarifying how surface OH groups modify the branching between CH4 desorption, activation and subsequent reactions during TPRS. Our DFT results predict that surface OH groups strongly destabilize CH4 σ-complexes on IrO2(1 1 0) in addition to deactivating surface O-atoms that are needed to achieve CH4 activation at low temperature. We demonstrate that a microkinetic model that incorporates the influence of surface OH groups on the CH4 binding and reactivity reproduces experimental TPRS results which show that adsorbed CH4 σ-complexes preferentially dissociate at low CH4 coverage, but that an increasing fraction of the adsorbed CH4 desorbs at low temperature (~120 K) with increasing initial CH4 or H coverage. The simulations reveal that low-temperature CH4 desorption during TPRS arises primarily from CH4 σ-complexes that are kinetically-trapped between adjacent surface OH groups, and are thus destabilized and unable to access reactive O-atoms. In addition, when the effect of adjacent OH groups is incorporated we find that the PBE-D3 functional incorporating dispersion provides CH4 binding energies that agree more closely with that obtained from the TPRS experiments. Our results demonstrate that the local effect of adjacent OH groups must be incorporated into any microkinetic models to properly capture the selectivity between extensive and partial oxidation in alkane conversion on IrO2(1 1 0) under reaction conditions.

Dual Lewis site creation for activation of methanol on Fe3O4(111) thin films.

Despite a wide application in heterogeneous catalysis, the surface termination of Fe3O4(111) remains controversial. Herein, a surface with both Lewis acid and base sites is created through formation of an Fe3O4(111) film on α-Fe2O3(0001). The dual functionality is generated from a locally nonuniform surface layer of O adatoms and Fetet1 sites. This reactive layer is reproducibly formed even in oxygen-free environments because of the high mobility of ions in the underlying α-Fe2O3(0001). The atomic structure of the Fe3O4(111) surface was identified by scanning tunneling microscopy (STM) and density functional theory (DFT) using the registry of the overlayers with the surface and the distinct electronic structure of oxygen adatom (Oad) and uncovered lattice Fetet1. The surface is dominated by the interface of Oad and Fetet1, a Lewis acid–base pair, which favors methanol dissociation at room temperature to form methoxy. Methoxy is further oxidized to yield formaldehyde at 700 K in temperature programmed reaction spectra, corresponding to an approximate activation barrier of 179 kJ mol−1. The surface termination of Fe3O4(111) is fully recovered by rapid heating to 720 K in vacuum, demonstrating the high mobility of ions in this material. The work establishes a clear fundamental understanding of a unique magnetite surface and provides insights into the origin of selective oxidation of alcohols on magnetite-terminated catalysts.

Effects of phosphorus and alkyl substituents on C–H, C–C, and C–O bond rupture within carboxylic acids on Ru(0001)

Quantitative analysis of temperature programmed reaction (TPR) spectra of formic acid, acetic acid, propionic acid, and butyric acid decomposition on Ru(0001) and phosphorus (P)-modified Ru(0001) surfaces show that both the length of carboxylic acids alkyl substituents (i.e., R=H, CH3, CH2CH3, and CH2CH2CH3) and the presence of P atoms alter the intrinsic activation energy for bond rupture. Inversion analysis of TPR spectra reveal that the intrinsic activation energy barriers on Ru(0001) for C–H bond cleavage in formate is 94 kJ mol−1, while C–C bond cleavage (decarboxylation) barriers for acetate, propionate, and butyrate are 67, 49, and 55 kJ mol−1, respectively. The activation energies to decarboxylate all four of these species correlate linearly with differences between the homolytic dissociation energies of the relevant bonds (e.g., H-COOH, CH3-COOH, CH3CH2-COOH, and CH3CH2CH2-COOH), which suggests that longer alkyl substituents (e.g., propyl and butyl) stabilize bond rupture transition states by donating electron density to the R–COOH bond via inductive effects. Simultaneously, longer alkyl chains also promote self-stabilizing lateral interactions between carboxylates via van der Waals forces that increase the barriers for R–COOH bond rupture slightly (by ∼1–5 kJ mol−1) at high surface coverages. The addition of P atoms to Ru(0001) increases the intrinsic activation energies for the rupture of all bonds (i.e., C–O, C–H, and C–C bonds), specifically, by 5–50 kJ mol−1 for C–H and C–C bonds. P atoms change the Ru(0001) surface likely via an electronic effect by decreasing the extent of electron back donation from Ru atoms to the antibonding orbitals of the carboxylate intermediate. These results provide useful information for transition metal phosphide catalyst design and carboxylic acid alkyl substituent selection to tailor selectivity toward C–O, C–H, and C–C bond rupture.

Surface reaction network of CO oxidation on CeO2/Au(110) inverse model catalysts.

CeO2/Au(110) inverse model catalysts were prepared and their activity toward the adsorption and co-adsorption of O2, CO, CO2 and water was studied by means of X-ray photoelectron spectroscopy, low energy electron diffraction, thermal desorption spectra and temperature-programmed reaction spectra. The Au surface of CeO2/Au(110) inverse model catalysts molecularly adsorbs CO, CO2 and water, and the polycrystalline CeO2 surface of CeO2/Au(110) inverse model catalysts molecularly adsorbs O2, and molecularly and reactively adsorbs CO, CO2 and water. By controllably preparing co-adsorbed surface species on CeO2/Au(110) inverse model catalysts, we successfully identified various surface reaction pathways of CO oxidation to produce CO2 with different barriers both on the CeO2 surface and at the Au-CeO2 interface, including CO oxidation by various oxygen species, and water/hydroxyl group-involved CO oxidation. These results establish a surface reaction network of CO oxidation catalyzed by Au/CeO2 catalysts, greatly advancing the fundamental understandings of catalytic CO oxidation reactions.

Molecular Chemisorption of O2 on a PdO(101) Thin Film on Pd(111)

We used temperature programmed desorption (TPD) to investigate the molecular chemisorption of O2 on a (101)-oriented PdO thin film grown on Pd(111) in ultrahigh vacuum (UHV) using an oxygen atom beam. Our results show that the molecular chemisorption of O2 is facile on the PdO film at 85 K, producing a saturation O2 coverage of 0.27 ML (monolayers). Experiments with co-adsorbed 16O2 and 18O2 further reveal that molecularly chemisorbed O2 dissociates negligibly on the PdO(101) surface under the conditions examined. The O2 TPD spectrum from the PdO(101) surface at O2 saturation exhibits two main features centered at 117 K and 227 K, as well as smaller features at 275 K and 315 K, associated with the desorption of molecularly chemisorbed O2. Comparison with O2 TPD obtained from clean Pd(111) demonstrates that a large fraction of the O2 molecules on the PdO(101) surface are more strongly bound than O2 chemisorbed on the metallic surface at 85 K and saturation of the respective O2 layers. We find that O2 molecules chemisorb only in small quantities (< 0.03 ML) on the p(2 × 2) and two-dimensional Pd5O4 phases of atomic oxygen on Pd(111), indicating that these phases have much weaker binding affinities toward O2 than the PdO(101) surface generated in our experiments. Finally, temperature programmed reaction spectra with co-adsorbed 18O2 and CO demonstrate that both PdO and molecularly adsorbed O2 actively participate in the oxidation of CO, with the atomic and molecular species exhibiting similar activities for the conditions studied. The results of this study may have implications for understanding Pd oxidation catalysis at high pressures given that we find relatively strong binding states of O2 on PdO and observe that these molecularly adsorbed species are active in CO oxidation.

CO Oxidation on Rh(100): Multisite Atomistic Lattice-Gas Modeling

A detailed multisite atomistic lattice-gas model is applied for CO oxidation on single-crystal Rh(100) surfaces under low-pressure conditions. Its behavior is analyzed by kinetic Monte Carlo simulation accounting for high adspecies surface mobility. The model input includes binding energies and interactions for CO and oxygen adspecies as well as a prescription of the adsorption−desorption dynamics and of the reaction pathway and energetics. Density functional theory (DFT) provides some guidance as to these energies, but significant uncertainties remain, particularly regarding the CO binding site on Rh(100). Thus, most modeling is performed with a judiciously chosen set of energetics in light of existing experimental evidence. We examine (i) the ordering and temperature-programmed desorption kinetics of the individual reactant adspecies, CO and oxygen, on Rh(100) and (ii) ordering in the mixed reactant adlayer and CO oxidation kinetics as reflected in temperature-programmed reaction spectra.

Atomistic lattice-gas modeling of CO oxidation on Pd(100): Temperature-programed spectroscopy and steady-state behavior

We have developed an atomistic lattice-gas model for the catalytic oxidation of CO on single-crystal Pd(100) surfaces under ultrahigh vacuum conditions. This model necessarily incorporates an detailed description of adlayer ordering and adsorption-desorption kinetics both for CO on Pd(100), and for oxygen on Pd(100). Relevant energetic parameters are determined by comparing model predictions with experiment, together with some guidance from density functional theory calculations. The latter also facilitates description of the interaction and reaction of adsorbed CO and oxygen. Kinetic Monte Carlo simulations of this reaction model are performed to predict temperature-programmed reaction spectra, as well as steady-state bifurcation behavior.

The Study of Nitric Oxide Adsorption and the Mechanism of Surface “Explosions” in the Reaction of CO + NO on Pt(100) and Pd(110) Single Crystal Surfaces

High resolution electron energy loss spectroscopy (HREELS), temperature-programmed desorption (TPD) and temperature-programmed reaction (TPR) were used to study NO adsorption and the reactivity of COads and NOads molecules on Pd(110) and Pt(100) single crystal surfaces. Compared to the Pt(100)-(1 × 1) surface, the unreconstructed Pt(100)-hex surface is chemically inert toward NO dissociation into Nads and Oads atoms. When a mixed adsorbed COads + NOads layer is heated, a so-called surface “explosion” is observed when the reaction products (N2, CO2, and N2O) synchronously desorb in the form of sharp peaks with a half-width of 7-20 K. The shape specificity of TPR spectra suggests that the “vacancy” mechanism consists of the autocatalytic character of the reaction initiated by the formation an initial concentration of active sites due to partial desorption of molecules from the COads + NOads layer upon heating to high temperatures. Kinetic experiments carried out on the Pd(110) surface at a constant reaction pressure and a linear increase in the temperature confirm the explosive mechanism of the reaction NO + CO.

Old Technique for New Materials: Characterization of Solid State Hydrogen Storage Materials by Temperature Programmed Reaction

ABSTRACTThe most common diagnostic tools in studying reversible solid-state hydrogen storage materials involve volumetric or gravimetric approaches. In this study, a mature gas solid analysis technique called Temperature Programmed Reaction (TPR) is presented as an alternative and complementary approach in investigating solid-state hydrogen storage materials. TPR is fast, simple and offers valuable information from a wide range of solid materials. This technique is routinely used to characterize the reduction behavior of metal oxide and sulfide catalysts. In a typical TPR experiment, a sample powder is heated at a constant rate under a non steady-state environment of constant pressure and flow of a hydrogen/inert gas mixture, while monitoring the rate of hydrogen absorption and desorption by a thermal conductivity detector. The principle behind TPR is the high thermal conductivity coefficient of hydrogen when compared to other gases. The detector is sensitive enough that a slight change in hydrogen concentration is reflected by significant changes in the thermal conductivity signal. Hydrogen storage materials exhibit fingerprint type TPR spectra, unique to their composition and structure. Specific rates of hydrogen intake, heats of absorption and desorption as well as extent of activation can be extracted from TPR curves. In addition, multiple absorption and desorption states can be identified from the number of absorption or desorption peaks. Material stability and potential structural or phase changes can also be inferred indirectly. The simplicity and the highly sensitive thermal conductivity measurement provide a number of advantages over the traditional volumetric or gravimetric methods. Metal hydride alloys of AB5-type are used to demonstrate the effectiveness of TPR technique in characterizing hydrogen storage materials.

The surface chemistry of alkyl and arylphosphate vapor phase lubricants on Fe foil

The surface chemistry of tributylphosphate (TBP) and tricresylphosphate (TCP) on a polycrystalline Fe surface was studied using temperature programmed reaction spectroscopy and Auger electron spectroscopy to illustrate some of the initial steps in the reaction mechanisms of alkyl and arylphosphate vapor phase lubricants. During heating, TBP [(C 4H 9O) 3P=O] adsorbed on the Fe surface decomposes via C–O bond scission to give butyl surface intermediates [C 4H 9–] that react via β-hydride elimination to desorb as 1-butene [CH 3CH 2CH=CH 2] and H 2 without appreciable carbon deposition onto the surface. The thermal decomposition of 1-iodobutane [I-C 4H 9] on Fe was observed to proceed via the same β-hydride elimination mechanism. In contrast to tributylphosphate, meta-tricresylphosphate ( m-TCP) [(CH 3–C 6H 4O) 3P=O] decomposes on Fe via P–O bond scission to produce methylphenoxy intermediates [CH 3–C 6H 4O–]. During heating to 800 K, methylphenoxy intermediates either desorb as m-cresol [CH 3–C 6H 4–OH] via hydrogenation or decompose further to generate tolyl intermediates [CH 3–C 6H 4–]. Some of the tolyl intermediates desorb as toluene [CH 3–C 6H 5] via hydrogenation but the majority decompose resulting in H 2 and CO desorption and carbon deposition onto the Fe surface. The P–O bond scission mechanism of m-TCP was verified by showing that the temperature programmed reaction spectra of m-cresol yield products that are almost identical to those of m-TCP. These results provide insight into the origin of the differences in the performance of alkyl and arylphosphates as vapor phase lubricants. The alkylphosphates decompose via alkyl intermediates that readily undergo β-hydride elimination and desorb into the gas phase as olefins, thus removing carbon from the surface. In contrast, the arylphosphates generate aryloxy intermediates by P–O bond scission and aryl intermediates by further C–O bond scission. Neither of these intermediates can undergo β-hydride elimination and thus they decompose to deposit carbon onto the Fe surface. The higher efficiency for carbon deposition may be the primary reason for the superior performance of the arylphosphates over alkylphosphates as vapor phase lubricants.

Propene Oxidation on Ag(111): Spectroscopic Evidence of Facile Abstraction of Methyl Hydrogen

Reactions of propene on an oxygen-modified Ag(111) surface were investigated by means of temperature-programmed reaction spectra and reflection--adsorption infrared spectroscopy (RAIRS) at 100 K. In the presence of oxygen adatoms, interactions of propene with the substrate are strengthened. The reaction between propene and oxygen adatoms (0.086 monolayer) produces not only total oxidation products (CO2 and H2O), but also partial oxidation products (CO and acetone). The formation of chemisorbed hydroxyl groups is identified by RAIRS when coadsorbed propene and oxygen adatoms are annealed to 200 K. The formation of hydroxyl is ascribed to the abstraction of methyl hydrogen by oxygen adatoms.

CO oxidation on Pt( [formula omitted]) promoted by coadsorbed H 2O

It is well established that CO 2 is produced at temperatures ∼300 K through the well known Langmuir–Hinshelwood mechanism, e.g. when CO and oxygen are coadsorbed on Pt( 1 1 1 ) at temperatures 150–170 K and then heated. We demonstrate that the same coadsorbates after the addition of minute amounts (⩽0.1 ML) of H 2O, causes CO 2 production at 180–230 K. By analysis of temperature programmed reaction (TPR) spectra of desorbing H 2O and CO 2 for different O/CO/H 2O coverage combinations, we suggest an autocatalytic reaction path where OH molecules, produced via the 2H 2O+O ad→3OH ad+H ad reaction, induce low temperature CO oxidation. This result suggests a reinterpretation of earlier reported TPR spectra of low temperature CO 2 production from CO+O coadsorbates. It also suggests that H 2O may reduce the self poisoning by CO of supported Pt catalysts in e.g. vehicle emission cleaning.

Annealing ultra thin Ta2O5 films deposited on bare and nitrogen passivated Si(100)

Ta2O5 films, deposited on bare and nitrogen passivated Si(100) surfaces, were annealed in the presence of three different oxidizing agents, NO, N2O, and O2, and then examined using X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (TOFSIMS). These films were thin enough to allow spectroscopic examination of the interface between the silicon and the oxide. During deposition of Ta2O5 on bare Si(100), the substrate is oxidized and the amount of oxide increases during subsequent annealing. No nitrogen accumulates when annealing in NO or N2O. Angle-resolved XPS and TOFSIMS depth profiles reveal, after annealing, oxidized silicon at the vacuum–solid interface and throughout the oxidized tantalum layer. For Ta2O5 on nitrided Si(100), annealing at 820°C depletes nitrogen in the silicon nitride and increases silicon oxide, regardless of the oxidant used. Temperature programmed reaction spectra of Ta2O5 deposited on previously oxidized Si(100) show SiO desorption above ∽950°C. The annealing results are accounted for in terms of chemical reactions at the buried interface that form SiO and, for the nitrided samples, NO species that migrate toward the vacuum interface. Oxidizing species from the gas phase also migrate to the buried interface to produce additional silicon oxide.

Gas-Phase Hydroxyl Radical Emission in the Thermal Decomposition of Lithium Hydroxide

The gas-phase hydroxyl (OH) radical emission was observed in the thermal decomposition of lithium hydroxide (LiOH). This phenomenon was investigated in a vacuum flow tube reactor at around 2 Torr by the temperature programmed reaction (TPR) experiments at 500−1300 K. The production of OH and other gaseous products was quantitatively investigated by the laser induced fluorescence method and quadrupole mass spectrometry, respectively. The TPR spectra of OH had a peak at 1100−1200 K, which largely exceeded the gas-phase thermodynamic equilibrium. The origin of OH was supposed to be either the surface OH groups on Li2O or the residual LiOH in the LiOH/Li2O solid solution. OH production exceeding the thermodynamic equilibrium was explained by means of the partial equilibrium in the reaction: LiOH + 1/4O2 ⇔ 1/2Li2O + OH. This phenomenon can be a new route for the OH production from H2O and O2 in cyclic reactions of lithium compounds.

Study of the preparation of bulk tungsten carbide catalysts with C2H6 /H2 and C2H4/H2 carburizing mixtures

The influence of the preparation procedure of tungsten carbide on the mechanism of carburization is discussed. This work is focused on the reduction and the carburization of tungsten trioxide by a mixture of hydrocarbon and H2 to form WC. Temperature-programmed reaction spectra obtained with CH4, C2H6 and C2H4 have been measured. In presence of the CH4-H2 mixture, H2 is the reducing agent and the hydrocarbon is consumed for the carburization whereas C2H6 or C2H4 participates in the reduction of the tungsten oxide. The temperatures of reduction and carburization are lower by about 150 K using C2H6 or C2H4 instead of CH4. Such a decrease of the temperature of reduction of tungsten oxide is needed to avoid the formation of poorly reducible compounds that can occur during the preparation of supported tungsten carbide. Furthermore, the surface area of the resulting carbide is 25 m2/g with C2H6 and C2H4 and 10 m2/g with CH4. During the carburization, the deposit of excess carbon on the WC surface is larger with the C2 hydrocarbons than with CH4, but it protects the carbide and can be removed by hydrogen treatment.

Quantitative adsorbate orientation from vibrational spectra: Ethoxides on Cu(111)

Fourier-transform infrared reflection absorption spectra have been obtained for a set of deuterium and isotopic carbon (13C) labeled ethoxides adsorbed on the Cu(111) surface at saturation coverages. The frequency shifts upon 13C isotopic replacement indicate that the proper description of the skeletal stretches is in terms of CCO modes and not local modes of CO and CC bonds. Comparison of the spectra of the deuterated ethoxides with gas phase and multilayer spectra of the corresponding ethanols allowed the determination of the orientation of the adsorbed ethoxide. A solution for the average structure of ethoxide shows that the plane of the species is tilted 17±2° from the plane of the surface normal and that the C–C bond, is tilted 70±2° from the surface normal. The tilt of the C–O bond from the surface normal is 17±2°. At elevated temperatures the ethoxides undergo β-hydride elimination and desorb as acetaldehydes. Temperature programmed reaction spectra for β-hydride elimination of the selectively deuterated ethoxides reveal both a primary kinetic isotope effect and a measurable secondary isotope effect.

CO Adsorption and Oxidation on Pd(110)-c(2 × 4)−O by Reflection−Absorption Infrared Spectroscopy

Adsorption and oxidation of CO on Pd(110)-c(2 × 4)−O were investigated by reflection−absorption infrared spectroscopy (RAIRS), low-energy electron diffraction (LEED), and temperature-programmed reaction (TPR). High-frequency CO at 2140 cm-1, which was greatly influenced by adsorbed oxygen atoms, was observed by RAIRS even at low CO coverage when the Pd(110)-c(2 × 4)−O surface was exposed to CO at 100 K. This peak developed along with a broad band around 2000 cm-1 from low CO coverage to the saturation. The reaction of CO and an oxygen atom to form CO2 on the c(2 × 4)−O surface was observed in the range 100−450 K. Below 250 K, two desorption peaks of CO2 were observed in TPR spectra and the corresponding RAIR spectra showed that linear CO species (2140−2070 cm-1) disappeared by desorption as CO or reaction with an oxygen atom to form CO2, while there was little decrease of bridge CO species (2020−1880 cm-1). It is likely that linear CO species mainly reacted to form CO2 on Pd(110)-c(2 × 4)−O below 250 K. We successfully followed the constant-flow reaction of CO and O2 on the Pd(110)-c(2 × 4)−O at 220 K; however, the reaction ceased with an increase of bridge CO at 1992 cm-1. RAIR spectra during the reaction of CO and O above 250 K suggest that CO began to form domains separated from oxygen atoms and the substrate structure of (1 × 2) was gradually converted into (1 × 1) according to the local coverage of CO.

Site-blocking effects in temperature-programmed reaction studies of surface decomposition reactions

The starting point for most analyses of temperature-programmed reaction (TPR) spectra has been the Wigner-Polanyi expression in the form: rate = v(θ) exp(− E a(θ) kT )θ n , where θ is the surface coverage. This paper discusses the b approximation for the case of decomposition reactions where the dissociation rate is inhibited by blocking sites on the surface. It is shown that coverage dependences other than θ n that arise from site-blocking have dramatic effects on TPR spectra. These effects are illustrated both with simulations and with experimental results for the decomposition of butyl groups bound to a Cu(110) surface. While the surface coverages of filled and vacant sites must, in general, be carefully distinguished, it is shown that if adsorbate coverage is defined relative to saturation of the monolayer then regardless of the number of sites blocked per adsorbate the initial rate varies as θ 0(1 − θ 0) where θ 0 is the initial coverage. The effects of interactions between adsorbates are also discussed.

Temperature-programmed hydrogenation of carbon monoxide on supported nickel catalyst, multiple reaction sites and their interactions on Ni/TiO 2

The catalytic behaviors of Ni/TiO 2 and Ni/SiO 2 for CO hydrogenation were investigated using the temperature-programmed reaction (TPR) of pre-adsorbed carbon monoxide with hydrogen. The TPR spectra of Ni/SiO 2 exhibits a single peak with the characteristic shape of first-order kinetics. In contrast, two distinct peaks are observed for Ni/TiO 2, which indicates that there are two different reaction sites. The high temperature peak shows behavior similar to that of Ni/SiO 2, and it is believed that the corresponding sites are composed of nickel atoms. The low temperature peak is attributed to the hydrogenation of CO adsorbed on Ni-TiO x sites. Some interrupted experiments indicate the formation of a common intermediate species from both reaction sites. Thus, a mechanistic model is proposed on the basis of experimental observation. Computer simulation of the TPR spectra shows that the model fits the system well. By comparison with carbon hydrogenation experiments, the intermediate is believed to be the active surface carbon. This mechanism satisfactorily explains the high activity of CO hydrogenation on Ni/TiO 2 catalyst.