High-Pressure CO Adsorption on Cu-Based Catalysts: Zn-Induced

Mar 25, 2011 - CO adsorption at 1 MPa on Cu−Zn stearate colloids and supported Cu catalysts was studied in situ by attenuated total reflection infra...
1 downloads 4 Views 1016KB Size
ARTICLE pubs.acs.org/Langmuir

High-Pressure CO Adsorption on Cu-Based Catalysts: Zn-Induced Formation of Strongly Bound CO Monitored by ATR-IR Spectroscopy Zhimin Liu, Andre Rittermeier, Michael Becker, Kevin K€ahler, Elke L€offler, and Martin Muhler* Laboratory of Industrial Chemistry, Ruhr-University Bochum, Universit€atsstr. 150, D-44780 Bochum, Germany ABSTRACT: CO adsorption at 1 MPa on CuZn stearate colloids and supported Cu catalysts was studied in situ by attenuated total reflection infrared (ATR-IR) spectroscopy. Subsequent to thorough reduction by H2, the IR band at 21102070 cm1 due to linearly adsorbed CO on clean metallic Cu was always observed initially on all Cu catalysts. During the exposure of Zn-containing samples to CO at high pressure, a new IR band at ca. 1975 cm1 appeared in addition and increased in intensity even at room temperature. The detailed analysis of the IR spectra showed that the new IR band at ca. 1975 cm1 was not related to coadsorbed carbonate/formate-like species, but to the content of Zn in the samples. This IR band was found to be more stable than that at 21102070 cm1 during purging with inert gas. It disappeared quickly in synthetic air, pointing to a strongly reduced state of the Zn-containing Cu catalysts achieved during high-pressure CO exposure. It is suggested that CO can reduce ZnO to Zn in the presence of Cu, resulting in the formation of a CuZnx surface alloy. As the CO species with the characteristic IR band at ca. 1975 cm1 binds more strongly to this CuZnx alloy than the linearly adsorbed CO to pure Cu, it is suggested to be adsorbed on a bridge site.

1. INTRODUCTION Cu-based catalysts are widely applied in various redox reactions such as the hydrogenation and dehydrogenation of organic compounds.1 One of the main applications is the synthesis of methanol from a mixture of CO/CO2/H2 over Cu/ZnO-based catalysts. Due to its large and still increasing industrial importance, methanol synthesis over Cu-based catalysts has become one of the best studied reactions in heterogeneous catalysis.2 All the commercial Cu catalysts for methanol synthesis and the watergas shift reaction contain ZnO. The synergistic effects of Cu and ZnO have been studied for several decades focusing on the electronic properties of Cu or ZnO, mechanisms featuring spillover of adsorbed species, CuZn alloy formation,3 and specific interactions at the Cu/ZnO interface.4 However, the nature of the active sites,5 the roles of Cu and ZnO in the solid catalyst,6 and other issues still remain controversial.7,8 IR spectroscopy of adsorbed CO is one of the most frequently applied methods for the determination of the surface states of supported metals.9,10 Bands at 20002200 cm1 characterize linearly bound CO molecules.11 Vibrations of CO molecules bridging two metal atoms show IR bands between 1880 and 2000 cm1, whereas those of multiply coordinated molecules appear at 16501880 cm1.10,12,13 In recent IR studies, large differences in the stretching frequency of CO adsorbed on Cubased catalysts after different reducing conditions were found ranging from 2115 to 2060 cm1.14 In our previous studies on the synthesis and characterization of Cu colloids15 and CuZn r 2011 American Chemical Society

stearate colloids,16 two IR bands were observed at ca. 1980 and 1920 cm1 by attenuated total reflection infrared (ATR-IR) spectroscopy during CO adsorption at high pressure and high temperature. IR bands in the region of 20001900 cm1 are generally attributed to bridge-bound CO on metal surfaces, but bridged CO on Cu single crystals shows an IR band at ca. 1860 cm1, which can only be observed at low temperatures and high coverages.13 There are several reports in literature in which the IR bands of adsorbed CO on solid Cu catalysts were observed below 2000 cm1, but most of them paid little attention to the IR band in this region or did not provide a detailed discussion of its origin.1721 The aim of the present study is to investigate the formation and nature of the IR bands between 2000 and 1900 cm1 observed during high-pressure CO adsorption on CuZn stearate colloids. Because of the complexity of the Cu colloidal system, CO adsorption on Cu/ZnO/Al2O3 was also studied in detail by means of ATR-IR spectroscopy. The IR band at ca. 1975 cm1 can also be observed during high-pressure CO adsorption on the most commonly used solid ternary Cu/ZnO/Al2O3 catalyst. The IR studies show that the CO species with the IR band at ca. 1975 cm1 binds more strongly to Cu than the linearly bound CO. The ongoing reduction of oxidic species by CO is important for the formation of the strongly bound species. It is suggested that CO can reduce ZnO to Zn in the presence Received: January 7, 2011 Revised: March 9, 2011 Published: March 25, 2011 4728

dx.doi.org/10.1021/la2000766 | Langmuir 2011, 27, 4728–4733

Langmuir

ARTICLE

of Cu resulting in the formation of a CuZnx surface alloy, and that the IR band at ca. 1975 cm1 originates from bridge-bound CO on this CuZnx surface alloy.

2. EXPERIMENTAL SECTION 2.1. Materials. Cu stearate (Cu(CH3(CH2)16COO)2) was synthesized by precipitation in ethanol using cupric acetate and stearic acid according to methods in literature.22 Zn and Al stearates were obtained from Sigma-Aldrich and used without further purification. The following gases were used: H2 (99.9999%), CO (99.997%), CO2 (99.9995%), Ar (99.9999%), and synthetic air (20% O2 in N2) (99.999%, hydrocarbonfree). CO (>99.997% purity) was further purified by passing through a trap filled with molecular sieve (5 Å) at 513 K to remove carbonyl impurities. The Cu/Zn/Al2O3 sample (approximately 50 wt % Cu, 35 wt % ZnO, 15 wt % Al2O3) used in this study was a typical industrially applied ternary catalyst. 2.2. Catalyst Preparation. Cu-based colloids were prepared from Cu stearate and M stearate (M = Zn, Al). For the synthesis of the Cu colloids, Cu stearate and Zn stearate or Al stearate were suspended in n-hexadecane and reduced by H2 (0.5 MPa) at 473 K for 16 h with stirring (800 rpm).16 Supported Cu (10 wt %) on Al2O3 was prepared by chemical vapor deposition (CVD).23 2.3. Spectroscopy. Because the rotation-vibrations of gas-phase CO molecules have strong absorption in the region of 23002000 cm1, the investigation of CO adsorption at high pressure is difficult when applying diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) and transmission IR spectroscopy.24 Due to the limited rotation of the dissolved CO molecule, high pressure adsorption measurements using ATR-IR spectroscopy were possible in the colloidal and liquid/solid interfacial system. If a thin film of a solid material is deposited on the internal reflection element (IRE), ATR-IR spectroscopy is also highly surface sensitive, because the penetration depth of the IR radiation is usually a few micrometers. ATR-IR spectra of CO adsorption at the gas/solid interface mainly show the signal of adsorbed CO and are not dominated by the bands of gas-phase CO even at high pressure.25 ATR-IR spectra were recorded using a Nicolet Nexus spectrometer equipped with a Spectra-Tech ARK attenuated total reflectance kit and a liquid nitrogen-cooled MCT detector. The Si IRE (12 reflections) was mounted in a homemade stainless-steel flow-through cell. This homemade ATR cell enabled in situ measurements up to 473 K and 1.0 MPa. Spectra were recorded by averaging 120 scans at 4 cm1 resolution in the region of 4000650 cm1. For the UVvis measurements, an AvaSpec-2048 fiber optic spectrometer from Avantes was used. The spectrometer was connected to an in-line flow cell (optical path: 5 mm) with fiber optic cables. The in-line flow cell was connected to the ATR setup with Swagelok fittings directly behind the reactor. This configuration allowed us to perform in situ UV/vis measurements simultaneously with in situ ATR measurements. All the Cu colloids were prepared in the homemade steel reactor equipped with a glass inliner and were continuously circulated through the ATR-IR cell (20 mL min1) by a gear pump (ISMATEC Reglo-ZS). Background spectra were recorded after the reactor had been flushed by Ar. For the supported Cu catalysts, a thin film deposited on the Si IRE was used to investigate the adsorption of CO from the liquid phase or from the gas phase. The slurries of CuO/Al2O3 or CuO/ZnO/Al2O3 in water or ethanol were allowed to depositevaporate several times on the Si IRE at ambient conditions. The deposited CuO/Al2O3 and CuO/ ZnO/Al2O3 films were highly stable under the conditions applied. Because of the relatively high thickness of the resulting film, it is likely that only a small fraction of the layer was probed by the evanescent wave generated at the Sicatalyst interface by the IR beam. For CO adsorption on supported Cu in liquid phase, hexadecane was used as solvent and the solid films were pretreated by 0.5 MPa H2 at 473 K for 14 h. For CO adsorption in gas phase, the solid films were pretreated by

Figure 1. In situ UVvis spectra of Cu and Zn stearates in hexadecane after continuous treatments at 473 K: (a) 0.5 MPa H2 for 14 h; (b) 1 MPa CO for 10 min; (c) 1 MPa CO for 60 min; (d) 0.1 MPa synthetic air for 3 min, (e) 8 min, (f) 10 min, (g) 12 min, (h) 14 min. 5% H2/Ar at 453 K for 14 h and then by pure H2 at 473 K for 1 h. After the IR cell had been purged by Ar, reference spectra of the clean surfaces were collected.

3. RESULTS 3.1. CO Adsorption on Cu Colloids. Figure 1 shows the UVvis spectra of CuZn stearate (50:50) colloids in hexadecane at 473 K after different treatments. Initially, the Cu and Zn stearates containing solution in hexadecane showed a broad absorption band at 700 nm (not shown here), which can be attributed to the ligand-to-metal-charge-transfer (LMCT) band of Cu(II) stearate.26 After reduction by H2 for 14 h (Figure 1a), the color of the CuZn stearates solution turned from green to red, and a band around 570 nm appeared, which originates from the surface plasmon of Cu0 nanoparticles.26 CO adsorption on the reduced Cu colloid can slightly change the UVvis spectra, but it cannot change the surface plasmon of Cu0 nanoparticles (Figure 1b and c). Subsequently, the introduction of synthetic air at 473 K generated copper oxide on the surface of Cu particles as indicated by the disappearance of the surface plasmon band and the emergence of the band at 700 nm (Figure 1dh). CO is not only a probe molecule, but also a reducing agent at 473 K. After the CuZn stearate colloids were oxidized by synthetic air for 20 min, CO adsorption clearly showed the binding of CO to Cu2þ as indicated by the appearance of an IR band at 2183 cm1 (Figure 2a). After CO adsorption for 60 min (Figure 2c), the IR band at 2183 cm1 had disappeared, and another band at 2166 cm1 emerged. The later band is usually assigned to monocarbonyls of Cuþ (Cuþ[CO]),27 implying that Cu2þ on the surface was reduced to Cuþ by CO. Cuþ can also be reduced to Cu0 by CO as indicated by the increase of the band at ca. 2120 cm1 along with the decrease of the band at 2166 cm1. After CO treatment for 90 min, almost all the Cuþ was reduced to Cu0. The IR band at 2120 cm1 due to linear CO on Cu0 also begins to decrease correlated with the increase of the IR band at 4729

dx.doi.org/10.1021/la2000766 |Langmuir 2011, 27, 4728–4733

Langmuir

Figure 2. Time-resolved in situ ATR-IR spectra of CO adsorption in hexadecane at 473 K on CuZn stearate colloids (50:50) preoxidized for 20 min: (a) 10 min, (b) 30 min, (c) 60 min, (d) 90 min, (e) 100 min, (f) 110 min, (g) 120 min. Conditions: T = 473 K, PCO = 1 MPa.

Figure 3. ATR-IR spectra of CO desorption from CuZn stearate colloids in hexadecane in Ar at 453 K.

1980 cm1 after CO treatment for 100 min. It is suggested that another reduction process occurred generating some unknown new sites, on which adsorbed CO species show IR bands at ca. 1980 and 1910 cm1. Figure 3 shows the desorption process of CO from the Cu colloids in hexadecane at 453 K. Initially, the IR bands at 2118 and 1910 cm1 disappear very quickly, and at the same time the intensity of the IR band at ca. 1970 cm1 increased a little and then gradually decreased. The desorption study clearly showed that CO binds stronger to the unknown new sites than to the pure metallic Cu0 surface. 3.2. CO Adsorption on Solid Cu Catalysts at High Temperature. The CuZn stearate colloid is a complex system in which the Cu particles are surrounded by a Zn stearate shell and also by the solvent. Though the CuZn stearate colloids can show very high activities in methanol synthesis, the most commonly used and deeply studied methanol synthesis catalysts are Cu/ZnO/Al2O3 solid catalysts. It is therefore important to know whether the new adsorption sites can also be formed on

ARTICLE

Figure 4. Time-resolved ATR-IR spectra of CO adsorption on Cu/ ZnO/Al2O3 in hexadecane at 473 K and a CO pressure of 1 MPa.

Cu/ZnO/Al2O3 solid catalyst and to study the differences and similarities of surface states between CuZn stearate colloids and the Cu/ZnO/Al2O3 solid catalysts. Figure 4 shows the time-resolved IR spectra of CO adsorption on a typical industrially applied Cu/ZnO/Al2O3 catalyst in hexadecane at 473 K and a CO pressure of 1 MPa. The IR band at 2070 cm1 due to linearly adsorbed CO on reduced Cu0 was observed at first. This IR band frequency is lower than that on Cu colloid (2118 cm1), pointing to differences in the state of the surface of colloidal and solid Cu catalysts. After CO adsorption for 5 min, an IR band at 1945 cm1 emerged, which proved that the unknown new sites can also be formed during CO adsorption on solid ternary Cu catalysts. The IR band at 1945 cm1 increased and shifted gradually to 1977 cm1 associated with the decrease of the band at 2070 cm1. This phenomenon, which is similar to that on the CuZn stearate colloids, suggests that the metallic Cu surface sites gradually became the unknown new sites during CO adsorption on solid ternary Cu catalysts. Carbonate/formate-like species were also formed and showed an IR band at 1550 cm1 with a shoulder at 1585 cm1 during the CO adsorption. When H2 was added to the feed gas, the IR band in the region of 20001900 cm1 became broad and shifted to 1955 cm1, while the IR bands due to carbonate/formate-like species almost disappeared. It is clear that the IR bands in the region of 20001900 cm1 have no direct relation to the coadsorbed carbonate/formate species. CO adsorption on solid Cu catalysts has been usually studied in the absence of solvents at the gas/solid interface. These studies were usually performed at low pressure by transmission IR, because the rotation-vibration bands of gas-phase CO molecules strongly absorb in the region of 23002000 cm1. It is therefore necessary to study adsorption at high pressure not only at the liquid/solid interface but also at the gas/solid interface by ATR-IR spectroscopy. In order to clarify the specific role of every component of the ternary Cu/ZnO/Al2O3 catalyst in the generation of new surface sites, especially the role of Zn, a Cu/Al2O3 sample (10 wt % Cu, SCu = 7.7 m2/g) prepared by a high-purity CVD method was used as a reference here.23 Elemental analysis showed that this Cu/Al2O3 sample contained essentially no Zn, that is, less than 0.05 wt %. Only the IR bands of gaseous CO can be observed when pure ZnO is used as the adsorbent (not shown here). CO 4730

dx.doi.org/10.1021/la2000766 |Langmuir 2011, 27, 4728–4733

Langmuir

Figure 5. ATR-IR spectra of solvent-free CO adsorption on different Cu catalysts from the gas phase at 473 K: (a) 0.1 MPa CO on Cu/ZnO/ Al2O3 for 60 min; (b) 0.1 MPa on Cu/Al2O3 for 60 min; (c) 1 MPa CO on Cu/Al2O3 for 60 min; (d) 1 MPa CO on ZnOþCu/Al2O3 mixture (1:4 weight-based ratio) for 60 min; (e) in 0.1 MPa CO after (d).

adsorption on the Cu/ZnO/Al2O3 catalyst at 0.1 MPa (Figure 5a) clearly shows the IR band at 1960 cm1, while almost no IR band can be observed in this region of 20001900 cm1 for the Cu/Al2O3 sample (Figure 5b). Only a very weak IR band at ca. 1990 cm1 can be detected even at 1 MPa on this Cu/Al2O3 sample (Figure 5c). When ZnO was added to the Cu/Al2O3 sample even without grinding, the IR band of CO adsorption at 1990 cm1 became more prominent (Figure 5d and e). It is clear that the intensity of the IR band at ca. 1980 cm1 is not related to the coadsorbed carbonate/ formate-like species, but is induced by the Zn content of the samples. 3.3. CO Adsorption on Solid Cu Catalysts at Room Temperature. Jung et al.28 showed that ZnO can be reduced in the presence of Cu with CO or methanol as evidenced by X-ray diffraction (XRD) and the thermogravimetric (TG) measurements during reduction, but it can hardly be reduced with H2. It was suggested that two types of brass were formed resulting from the reduction of ZnO in the Cu/ZnO sample.28 After reduction with H2 at 473 K, the Cu/ZnO/Al2O3 was treated with methanol at 473 K for 2 h in order to generate a CuZnx alloy on the surface of the catalyst. After such a strongly reducing treatment, the IR band at ca. 1970 cm1 can also be observed during CO adsorption at 308 K (Figure 6a). When the IR cell was purged with Ar subsequently, the IR bands at 2066 and 2088 cm1 decreased very quickly within 2 min, but the intensity of the IR band at ca. 1970 cm1 increased (Figure 6b and c). The IR band at ca. 1970 cm1 decreased very slowly in the inert gas flow (Figure 6d), but it disappeared very quickly when synthetic air was introduced into the IR cell (Figure 6e). It is clear that the IR band at ca. 1970 cm1 is related to CO adsorbed on strongly reduced surface sites and that oxidation by air destroys these CuZnx sites. It can be fully ruled out that it is due to coadsorbed methoxy species. Figure 7 shows the IR spectra of CO adsorption on Cu/ZnO/ Al2O3 at room temperature and at different partial pressures of CO after reduction by H2. Only one IR band at 2100 cm1 was observed at low partial pressures of CO below 10%. The IR band at ca. 1970 cm1 emerged, and its intensity increased with the partial pressure of CO above 10%. When the IR cell was purged with Ar subsequently, the IR band at 2100 cm1 decreased very quickly within

ARTICLE

Figure 6. ATR-IR spectra of CO adsorption and desorption on CH3OH-pretreated (473 K for 2 h) Cu/ZnO/Al2O3 catalyst at 308 K: (a) in 0.1 MPa flowing CO for 30 min, then purged with Ar for (b) 1 min, (c) 2 min and (d) 40 min; (e) after changing to synthetic air for 1 min subsequent to (d).

Figure 7. ATR-IR spectra of CO adsorption on Cu/ZnO/Al2O3 at 308 K and P = 0.1 MPa: (ag) in 3%, 5%, 10%, 25%, 50%, 75%, 100% CO/Ar for 7 min, respectively. (hj) Purged with Ar for 3, 15, and 600 min after (g). (k) 5% CO/Ar for 7 min after (j).

a few minutes, but the intensity of the IR band at ca. 1970 cm1 became stronger at first and then deceased very slowly. After all the adsorbed CO species had desorbed from the Cu catalyst, 5% CO/Ar was introduced to the Cu catalyst again. The two IR bands at 2100 and 1970 cm1 were observed again at the same time during the second adsorption process. It is suggested that CO adsorption at high pressure can reduce ZnO in Cu/ZnO/Al2O3 even at room temperature. After reduction by H2, clean metallic Cu surfaces dominated on Cu/ZnO/ Al2O3. CO adsorption at high pressure induced a significant change of the Cu surface on Cu/ZnO/Al2O3, and a CuZnx surface alloy was formed during the first adsorption process.

4. DISCUSSION There are a few reports in literature in which the IR band of adsorbed CO on Cu below 2000 cm1 was observed. Most of 4731

dx.doi.org/10.1021/la2000766 |Langmuir 2011, 27, 4728–4733

Langmuir them paid little attention to this IR band or did not provide a detailed discussion about its origin.1721 In a CO temperature programmed desorption (TPD) study on Cu/ZnO/Al2O3 catalyst, Waugh and co-workers29 observed a CO desorption peak at 345 K corresponding to an adsorption energy of about 95 kJ/mol. After surface oxidation by CO2 decomposition at 213 K and then reduction by CO at 473 K, the desorption peak at ca. 345 K became much more prominent than that on samples reduced by H2. They attributed this peak to CO desorbing from Cu(211) facets. The desorption peak at ca. 370 K has also been observed in our previous study, when the adsorption of CO on Cu/ZnO/Al2O3 was carried out by dosing CO at 300 K and subsequent quenching of the catalyst by cooling the furnace with liquid N2.30 When the adsorption of CO on Cu/ZnO was carried out by dosing CO at 90 K, no CO desorption peaks are higher than 300 K in Kanai et al.’s CO TPD studies.31 Witte and coworkers32 studied the adsorption of CO on Cu single crystals and polycrystalline Cu samples and did not find any CO desorption peak higher than 250 K. The desorption peak of CO from Cu(211) was observed at ca. 240 K, and its adsorption energy was derived to amount to about 58 kJ/mol.32 The synergistic effects of Cu and ZnO have been widely studied. It has been considered that Cu can be alloyed with ZnO reduced in the vicinity of the Cu particles.33 The possible formation of dilute R-brass in the surface of Cu crystallites has been analyzed theoretically,34 and Nakamura and co-workers3,6,31 found experimental evidence of ZnOx migration and surface brass formation under relevant conditions. Jung et al.28 have also shown that ZnO can be reduced in the presence of Cu with CO (higher than ca. 420 K) or methanol (higher than ca. 520 K) as evidenced by XRD and TG measurements. CO adsorption at 473 K at 1 MPa can reduce not only CuO but also ZnO, resulting in the formation of a CuZnx alloy, or at least some surface alloy and Zn adatoms on the Cu surface. The isotherms of CO adsorption by Jung and Joo35 also showed that the treatment with methanol induced higher heats of CO adsorption on the Cu/ ZnO sample. The heats of adsorption at θ = 0.1 increased with the methanol treatment time within 30 min and leveled off afterward.35 The desorption monitored by ATR-IR spectroscopy showed that the CO species related to the IR band at ca. 1970 cm1 is bound more strongly to the Cu surface than the common linearly adsorbed CO on clean metallic Cu0 surface. It is therefore reasonable to assume that the IR band at ca. 1970 cm1, which appeared on ZnO-containing Cu catalysts during high-pressure CO exposure, is due to the CO adsorption on a CuZnx surface alloy. Greeley et al.36 predicted that linearly adsorbed CO on the Cu surface with Zn adatoms shows IR bands at ca. 1950 and 1920 cm1, which seems to be in good agreement with the observed IR bands at ca. 1970 and 1920 cm1 in our experiments. However, their density functional theory (DFT) calculations show that the binding energy of CO bound linearly to the Cu(111) surface in the presence of Zn adatoms is smaller than that to the clean Cu surface; that is, the IR band at ca. 2090 cm1 should be more stable than the IR bands at 20001900 cm1. This is not consistent with our experimental results suggesting that the observed IR bands at ca. 1970 cm1 are not due to linearly adsorbed CO on the Cu surface with neighboring Zn adatoms. IR bands at 20001900 cm1 are usually attributed to bridgebound CO on metal surfaces. However, bridge-bound CO on Cu single crystals can only be observed at low temperatures.13 Our experiments show that the IR bands at ca. 1975 cm1 are not

ARTICLE

Scheme 1. Surface Structures and CO Adsorption Geometries on ZnO-Supported Cu (top) and CuZn Stearate Colloid (bottom) Catalysts during Exposure to CO at High Pressure

related to coadsorbed methoxy and carbonyl/formate-like species, and they do not seem to result from linearly bound CO on Cu0 surfaces with Zn adatoms. It is therefore assumed that the IR bands at ca. 1975 and 1920 cm1 originate from bridge-bound CO on a CuZnx surface alloy. At the onset of the CO desorption, the IR band intensity at ca. 1975 cm1 increased, as the IR band of linearly bound CO decreased. This observation also supports the hypothesis that the IR band at ca. 1975 cm1 is due to multiply bound CO. DFT calculations are in progress which suggest that the binding energy of multiply bound CO on the CuZnx surface alloy is indeed higher than that of linearly bound CO on clean Cu0 pointing to a ligand effect of Zn. Scheme 1 illustrates the surface structure changes of Cu catalysts during high-pressure CO adsorption and the proposed CO adsorption geometries. ZnO can be partially reduced in the presence of Cu, and a CuZnx surface alloy is assumed to be formed during CO treatment even at room temperature. The strongly adsorbed CO species detected by the IR band at ca. 1970 cm1 is suggested to be multiply bound to the CuZnx surface alloy. It is interesting to note that bands in the range from 2050 to 1950 cm1 were observed on gold electrodes during the electrooxidation of CO at negative potentials.42 Boccuzzi et al.43 detected IR bands in the range from 2050 to 1950 cm1 when studying the interaction of CO with reduced Au/TiO2 and Au/Fe2O3 catalysts. Both Mihaylov et al.44 and Boccuzzi et al.43 assigned the bands in the 20501950 cm1 range to CO adsorbed on very small Au clusters, which are assumed to be negatively charged due to an electron transfer from the highly reduced supports to the small Au clusters. The 2110 and 2050 cm1 bands are assigned to CO adsorbed on top on uncharged and negatively charged Au clusters, respectively, whereas the band at 1990 cm1 band is assumed to originate from CO adsorbed in bridge position on negatively charged Au clusters.43 Metallic Cu, Pd, and Pt exhibit an fcc crystal structure with lattice parameters of 3.61, 3.89, and 3.92 Å, respectively. All three metals can form alloys with Zn under relevant conditions.37,38 Iwasa et al.39 showed that Pd/ZnO catalysts reduced at high temperatures (>573 K) were very active and selective to H2 and CO2 in steam reforming of methanol. The high catalytic performance of Pd/ZnO catalysts was attributed to the presence of a PdZn alloy, as revealed by the combination of XRD, X-ray 4732

dx.doi.org/10.1021/la2000766 |Langmuir 2011, 27, 4728–4733

Langmuir photoelectron spectroscopy, and temperature programmed reduction methods.39,40 Carbon monoxide was found to bind more weakly to the Zn/Pd(111) alloy surfaces compared to clean Pd(111).41 Zn addition was also found to alter the preferred adsorption sites for CO from 3-fold hollow to on-top sites.41 A similar behavior was observed for supported Pd/ZnO/Al2O3 catalysts. It is suggested that both the ensemble and electronic effects play a role in how Zn alters the interactions of CO with the surface. The results in our work suggest that there is an opposite effect for Zn addition on Cu compared with Zn on Pd: CO binds more strongly to the CuZnx surface alloy than to clean Cu surfaces, and Zn addition on Cu changes the preferred adsorption sites for CO from on-top to bridge sites.

5. CONCLUSIONS In situ ATR-IR results showed that high-pressure CO adsorption on Cu-based catalysts is a dynamic process. During CO adsorption on Zn-containing samples, a new IR band at ca. 1975 cm1 appeared and increased with adsorption time even at room temperature. It was found that the presence of this adsorbed CO species with the IR band at ca. 1975 cm1 is induced by the Zn content in the samples and that it is bound more strongly than linearly adsorbed CO on the clean Cu0 surface. It disappeared very quickly in synthetic air, pointing to a strongly reduced state of the Zn-containing Cu catalysts achieved during high-pressure CO exposure. CO acts not only as a probe molecule but also as a reducing agent: ZnO is reduced by CO under the applied high-pressure conditions, forming a CuZnx surface alloy, and it is assumed that CO is adsorbed on a bridge position binding more strongly to the CuZnx surface alloy than to the clean metallic Cu surface. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Fax: þ49 234 32 14115. Telephone: þ49 234 32 28754.

’ ACKNOWLEDGMENT Financial support by the Deutsche Forschungsgemeinschaft (DFG) within the Collaborative Research Center (SFB 558) “Metal-Substrate Interactions in Heterogeneous Catalysis” is gratefully acknowledged. ’ REFERENCES (1) Kochloefl, K. In Handbook of Heterogeneous Catalysis; Ertl, G., Kn€ozinger, H., Weitkamp, J., Eds.; Wiley VCH: Weinheim, 1997; Vol. 4, p 1831. (2) Schur, M.; Bems, B.; Dassenoy, A.; Kassatkine, I.; Urban, J.; Wilmer, H.; Hinrichsen, O.; Muhler, M.; Schl€ogl, R. Angew. Chem., Int. Ed. 2003, 42, 3815. (3) Sano, M.; Adaniya, T.; Fujitani, T.; Nakamura, J. J. Phys. Chem. B 2002, 106, 7627. (4) Naumann d’Alnoncourt, R.; Xia, X.; Strunk, J.; L€ offler, E.; Hinrichsen, O.; Muhler, M. Phys. Chem. Chem. Phys. 2006, 8, 1525. (5) Andreasen, J. W.; Rasmussen, F. B.; Helveg, S.; Molenbroek, A.; Stahl, K.; Nielsen, M. M.; Feidenhans’l, R. J. Appl. Crystallogr. 2006, 39, 209. (6) Nakamura, J.; Uchijima, T.; Kanai, Y.; Fujitani, T. Catal. Today 1996, 28, 223.

ARTICLE

(7) Vesborg, P. C. K.; Chorkendorff, I.; Knudsen, I.; Balmes, O.; Nerlov, J.; Molenbroek, A. M.; Clausen, B. S.; Helveg, S. J. Catal. 2009, 262, 65. (8) Kurtz, M.; Strunk, J.; Hinrichsen, O.; Muhler, M.; Fink, K.; Meyer, B.; W€oll, C. Angew. Chem., Int. Ed. 2005, 44, 2790. (9) Topsøe, N.-Y.; Topsøe, H. J. Mol. Catal. A: Chem. 1999, 141, 95. (10) Hadjiivanov, K. I.; Vayssilov, G. N. In Advances in Catalysis; Gates, B., Kn€ozinger, H., Eds.; Academic Press: Amsterdam, 2002; Vol. 47; p 307. (11) Eve, J. K.; McCash, E. M. Chem. Phys. Lett. 1999, 313, 575. (12) Raval, R.; Parker, S. F.; Pemble, M. E.; Hollins, P.; Pritchard, J.; Chesters, M. A. Surf. Sci. 1988, 203, 353. (13) Hayden, B. E.; Kretzschmar, K.; Bradshaw, A. M. Surf. Sci. 1985, 155, 553. (14) Topsøe, N. Y.; Topsøe, H. Top. Catal. 1999, 8, 267. (15) Schr€oter, M. K.; Khodeir, L.; Hambrock, J.; L€offler, E.; Muhler, M.; Fischer, R. A. Langmuir 2004, 20, 9453. (16) Rittermeier, A.; Miao, S.; Schr€oter, M. K.; Zhang, X.; Berg, M. W. E. v. d.; Kundu, S.; Wang, Y.; Schimpf, S.; L€offler, E.; Fischer, R. A.; Muhler, M. Phys. Chem. Chem. Phys. 2009, 11, 8358. (17) Edwards, J. F.; Schrader, G. L. J. Catal. 1985, 94, 175. (18) Clarke, D. B.; Bell, A. T. J. Catal. 1995, 154, 314. (19) Sakakini, B. H.; Tabatabaei, J.; Watson, M. J.; Waugh, K. C. J. Mol. Catal. A: Chem. 2000, 162, 297. (20) Bailey, S.; Froment, G. F.; Snoeck, J. W.; Waugh, K. C. Catal. Lett. 1995, 30, 99. (21) Sun, Q.; Liu, C. W.; Pan, W.; Zhu, Q. M.; Deng, J. F. Appl. Catal., A 1998, 171, 301. (22) Kimura, H.; Matsutani, K.; Tsutsumi, S.; Nomura, S.; Ishikawa, K.; Hattori, Y.; Itahashi, M.; Hoshino, H. Catal. Lett. 2005, 99, 119. (23) Naumann d’Alnoncourt, R.; Becker, M.; Sekulic, J.; Fischer, R. A.; Muhler, M. Surf. Coat. Technol. 2007, 201, 9035. (24) Weigel, J.; Koeppel, R. A.; Baiker, A.; Wokaun, A. Langmuir 1996, 12, 5319. (25) Grunwaldt, J. D.; Baiker, A. Phys. Chem. Chem. Phys. 2005, 7, 3526. (26) Creighton, J. A.; Eadon, D. G. J. Chem. Soc., Faraday Trans 1991, 87, 3881. (27) Iwamoto, M.; Hoshino, Y. Inorg. Chem. 1996, 35, 6918. (28) Jung, K. D.; Joo, O. S.; Han, S. H. Catal. Lett. 2000, 68, 49. (29) Hadden, R. A.; Sakakini, B.; Tabatabaei, J.; Waugh, K. C. Catal. Lett. 1997, 44, 145. (30) Strunk, J.; d’Alnoncourt, R. N.; Bergmann, M.; Litvinov, S.; Xia, X.; Hinrichsen, O.; Muhler, M. Phys. Chem. Chem. Phys. 2006, 8, 1556. (31) Kanai, Y.; Watanabe, T.; Fujitani, T.; Saito, M.; Nakamura, J.; Uchijima, T. Catal. Lett. 1994, 27, 67. (32) Vollmer, S.; Witte, G.; W€oll, C. Catal. Lett. 2001, 77, 97. (33) Spencer, M. S. Top. Catal. 1999, 8, 259. (34) Spencer, M. S. Surf. Sci. 1987, 192, 323. (35) Jung, K. D.; Joo, O. S. Bull. Korean Chem. Soc. 2002, 23, 1765. (36) Greeley, J.; Gokhale, A. A.; Kreuser, J.; Dumesic, J. A.; Topsoe, H.; Topsoe, N. Y.; Mavrikakis, M. J. Catal. 2003, 213, 63. (37) Wiame, F.; Salgin, B.; Swiatowska-Mrowiecka, J.; Maurice, V.; Marcus, P. J. Phys. Chem. C 2008, 112, 7540. (38) Boccuzzi, F.; Chiorino, A.; Guglielminotti, E. Surf. Sci. 1996, 368, 264. (39) Iwasa, N.; Mayanagi, T.; Nomura, W.; Arai, M.; Takezawa, N. Appl. Catal., A 2003, 248, 153. (40) Karim, A.; Conant, T.; Datye, A. J. Catal. 2006, 243, 420. (41) Jeroro, E.; Lebarbler, V.; Datye, A.; Wang, Y.; Vohs, J. M. Surf. Sci. 2007, 601, 5546. (42) Chang, S. C.; Hamelin, A.; Weaver, M. J. Surf. Sci. 1990, 239, L543. (43) Boccuzzi, F.; Chiorino, A.; Manzoli, M.; Andreeva, D.; Tabakova, T. J. Catal. 1999, 188, 176. (44) Mihaylov, M.; Kn€ozinger, H.; Hadjiivanov, K.; Gates, B. C. Chem. Ing. Tech. 2007, 79, 795. 4733

dx.doi.org/10.1021/la2000766 |Langmuir 2011, 27, 4728–4733