Infrared reflection absorption spectroscopy and temperature-programmed desorption studies of CO adsorption on Ni/CeO₂(111) thin films: The role of the ceria support

Ni-based catalysts haven been widely used in catalysis. In particular, ceria-supported Ni has attracted great attention in recent years as ethanol and methane reforming catalysts.^1–7 Ni is active and economical for the reforming process, with a high propensity for water activation.^8,9 However, it can rapidly deactivate owing to carbon deposition and particle sintering.^10–13 The ceria support can provide a potential solution to improve the stability and catalytic performance of Ni. The catalytic reactivity of ceria-supported Ni nanoparticles can be influenced by the redox properties of ceria as well as the synergistic effect between the two. The strong metal-support interaction between Ni and ceria can modify the structure and electronic properties of Ni.^5,14–17 Dispersing Ni particles on ceria form small particles that can help reduce coke formation. The ability of ceria to readily transfer its oxidation states between Ce⁴⁺ and Ce³⁺ that facilitates the oxygen release, transfer, and storage can help remove the carbon deposit from the Ni-ceria surface.^18–21 

Due to the increasing interest in the Ni/ceria materials, a fundamental-level investigation of their structural and electronic properties with respect to the nature of the ceria support is very important to understand its catalytic performance. Previously, we investigated the Ni/ceria interaction by employing model Ni/CeO_x(111) (1.5 < x < 2) surfaces using scanning tunneling microscopy (STM) and x-ray photoelectron spectroscopy (XPS) under ultrahigh vacuum (UHV) conditions.^17,22 We have demonstrated that well-ordered CeO_x(111) thin films with controlled degrees of Ce reduction and number of surface defects can be prepared and characterized at the atomic scale.^17,23 Structures and sizes of Ni were examined in detail in our studies with respect to the nature of the ceria support, the Ni coverage, and the annealing temperatures.^17,22 Specifically, submonolayer coverages of Ni were deposited on CeO_x(111) as research has shown that Ni-ceria interfacial sites play a significant role in the reactivity of Ni and good activity was observed with low coverages of Ni on ceria.^5,14,16,18 Our STM results show that deposition of ∼0.25 monolayer (ML) Ni on a stoichiometric CeO₂ surface at room temperature forms small Ni particles that are on average two-atomic layers high (0.4 nm), suggesting a strong interaction between ceria and Ni.¹⁷ Significant particle sintering and development into three-dimensional clusters with an average particle height of 1.0 nm were observed upon heating to 700 K.¹⁷ Under the same Ni coverage and deposition/annealing temperature, smaller Ni particles with larger particle densities were observed on partially reduced ceria due to the presence of surface defects. For example, Ni particles on CeO_1.77 exhibit an average height of 0.3 nm at 300 K and 0.8 nm after heating to 700 K.¹⁷ In addition to the size difference, a change in the Ni particle distribution was also observed on CeO₂ and CeO_1.77 surfaces as a result of heating. Decoration of Ni particles at step edges and terrace sites was observed upon 0.25 ML Ni deposition on both ceria surfaces. However, the Ni particles clearly show a more preferential growth at the step edges on CeO₂ than on CeO_1.77 with heating to 700 K.¹⁷ Our XPS studies have further demonstrated that the nature of the ceria surface influences the oxidation state of Ni.²² Ni remains metallic upon deposition on reduced ceria at 300 K as well as after subsequent heating. Metallic Ni is the predominant species on CeO₂ upon deposition at 300 K. However, oxidation of metallic Ni to Ni²⁺ was observed on CeO₂ at the cost of the Ce⁴⁺ reduction to Ce³⁺. The study of the extent of Ni oxidation versus Ni coverage suggests that the oxidation and reduction process occurs at the Ni-ceria interface. Furthermore, the oxidation of Ni can be promoted with heating. Ni²⁺ was identified as the major species after heating above 700 K with low coverages of Ni on CeO₂. Computational studies have shown that deposition of Ni on a stoichiometric CeO₂ surface can induce the transfer of the electronic charge from the Ni atom to the ceria, and thus Ni can be oxidized to Ni²⁺.^16,24,25 The appearance of Ni²⁺ features could also be due to the formation of NiO and/or a Ce_1−yNi_yO_x (0 < y < 1) mixed metal oxide, particularly with heating.^9,22,26

We followed up the study of Ni/CeO_x(111) model systems and report in this paper the investigation of CO adsorption using infrared reflection absorption spectroscopy (IRRAS) and temperature-programmed desorption (TPD). CO is the most common molecule in surface science studies.^27–29 This is in part due to the interest of CO as the reactant to be converted into less environmentally polluting products or the reaction product to be formed during the reforming process over heterogeneous catalysts. CO is also an excellent probe molecule that can further give arise information about adsorption/reaction sites and Ni-ceria interactions to better help understand reaction mechanisms. Furthermore, IRRAS CO studies over Ni/CeO_x(111) model systems under UHV conditions can be bridged to high-pressure environments. To date, the structure and chemical reactions of Ni/CeO_x(111) have been investigated by a variety of experimental and theoretical approaches.^{4,22,24,30,31} CO IR studies have been carried out over low-index Ni single-crystal surfaces.^27,32–34 However, the application of IRRAS to well-characterized CeO₂ single-crystal surfaces is limited and a fundamental IRRAS investigation of CO adsorption over Ni supported on single-crystal ceria surfaces is still missing.^4,35–37 In our study, we were able to monitor the surface sites of Ni deposited on both fully oxidized CeO₂(111) and partially reduced CeO_1.75(111) surfaces by employing CO IRRAS and TPD. Our results demonstrate that different surface sites resulting from ceria and Ni can be probed with CO. CO IRRAS and TPD data are in agreement with our XPS results that the oxidation state of Ni is dependent on the nature of the ceria support.²² 

The experiments were carried out using two UHV surface analysis chambers. A surface analysis instrument from the University of Wyoming (UW) was used to examine the structure and morphology of the Ni/ceria surface with STM, XPS, and low energy electron diffraction (LEED). The instrument was manufactured by Omicron Technology with a base pressure of less than 5 × 10⁻¹¹ Torr. A detailed description of the system can be found elsewhere.^23,38 IRRAS and TPD experiments were conducted using an UHV instrument at the Chemistry Department, Brookhaven National Laboratory (BNL). This IRRAS system is equipped with a shielded and differentially pumped UTI 100C mass spectrometer, a Perkin Elmer Auger system, a sputter ion gun, a homemade sample stage capable of heating and cooling in a temperature range between 90 and 1200 K, a 4-pocket electron beam evaporator with flux monitors from Oxford Scientific, and an elevated-pressure reactor/IRRAS cell system. The elevated-pressure cell is coupled to a commercial Fourier transform infrared (FT-IR) spectrometer (Bruker, IFS 66v/S) as described in detail elsewhere.³⁶ 

Both groups at UW and BNL have established the growth procedure of ceria thin films and ceria-supported Ni particles.^{5,14,17,23,36} The same procedure was used for the studies in both STM and IRRAS UHV instruments. Briefly, a Ru(0001) single crystal was used as the substrate for the ceria film growth. It was mounted on a standard Omicron-style Ta plate for the STM, XPS, and LEED studies. The temperature was calibrated with a C-type thermocouple spot-welded on the plate. For the IRRAS and TPD studies, the Ru substrate was machined with a groove around the crystal edge where a Ta filament loop was tightly imbedded. The Ta loop was attached directly to the feedthrough that provides mechanical support and heating/cooling to the crystal. A K-type thermocouple was attached to the top edge of the Ru crystal for the sample temperature measurement. CeO_x(111) thin films were prepared by deposition of Ce onto Ru(0001) at 700 K from an e-beam evaporator in the presence of oxygen followed by subsequent heating.³⁹ In particular, fully oxidized CeO₂ can be prepared with an oxygen chamber pressure of 2 × 10⁻⁷ Torr or higher and partially reduced films (e.g., CeO_1.75) can be generated with a reduced oxygen pressure of 2 × 10⁻⁸ Torr.²³ Low coverages of Ni were deposited on ceria surfaces at 300 K from either a homemade evaporation source or an e-beam evaporator.

STM, LEED, XPS, and Auger electron spectroscopy (AES) studies were used to ensure the surface cleanness of the Ru crystal and monitor the growth of ceria thin films and Ni particles. LEED patterns were taken with the beam energy of 80 eV. All the STM images were recorded using an etched tungsten tip at room temperature in a constant current mode (0.05–0.1 nA, 1–3 V). XPS experiments were carried out using an Mg Kα anode (1253.6 eV, 15 kV, 20 mA) with a fixed electron passing energy of 50 eV and an entrance slit size of 6 × 12 mm² in a high resolution scan mode. The spectra were collected with a 0.020 eV step and averaged over two scans. AES was carried out using 2 keV electrons. In this study, the structure and chemical states of Ni on ceria thin films at 300 K as well after heating to 700 K were investigated in detail using CO TPD and IRRAS. IRRAS data were recorded at a resolution of 4 cm⁻¹ using a grazing angle of approximately 85° to the surface normal. TPD data were collected with a heating rate of 3 K/s using a UTI mass spectrometer. The crystal was placed in a “line-of-sight” configuration with a distance of ∼0.5 cm to the aperture of the mass spectrometer. CO gas was introduced to the sample surface by backfilling the UHV chamber via a high precision leak valve. A range of CO pressures from 5 × 10⁻⁸ to 1 mTorr was used to probe various surface sites on the Ni/ceria surface.

Both fully oxidized CeO₂(111) and partially reduced CeO_1.75(111) were grown as the oxide supports for Ni to examine the effect of the structure and the degree of Ce reduction of ceria. Shown as an example in Fig. 1(a), the CeO₂(111) film is well ordered and consists of flat smooth layers of O-Ce-O tri-layer structures that are separated by an atomic step of 3 Å high. The atomic spacing between Ce-Ce on the surface is confirmed to be about 3.8 Å.²³ Compared to the stoichiometric CeO₂(111) surface, surface defects are present on the terraces of reduced CeO_1.75(111) [Fig. 1(b)]. These defects can be attributed to oxygen vacancies and Ce³⁺ cations, and the number of the defects as well as surface roughness increases with the degree of Ce reduction in CeO_x.²³ O 1s and Ce 3d XPS spectra were collected after each film growth. The lattice O in CeO₂ displays the O1s XPS peak at 529.1 eV, which exhibits a positive shift as expected in reduced ceria (e.g., CeO_1.75) due to the band bending [Fig. 1(c)].^38,40–42 By fitting the Ce 3d XPS spectrum as elaborated in detail in our previous publications, percent contributions of Ce³⁺ and Ce⁴⁺ cations and thus the stoichiometric values of the CeO₂ and CeO_1.75 films were determined.^23,43 The same recipe was used for the growth of fully oxidized CeO₂ and partially reduced ceria CeO_1.75 in the IRRAS UHV instrument for the CO IRRAS and TPD studies. The films were prepared with the same deposition duration and Ce flux values. They are free of impurities and completely cover the Ru(0001) substrate surface as confirmed by the AES analysis (data not shown).

Figure 2 shows the adsorption of CO on both fully oxidized CeO₂(111) and reduced CeO_1.75(111). CO stretch bands were monitored while cooling the ceria surface from room temperature to ∼100 K by increasing the CO chamber pressure up to 1 mTorr. While no CO IR peak was observed in 5 × 10⁻⁸ Torr CO, a single sharp peak at 2158 cm⁻¹ was detected at 1 mTorr of CO. This peak can be assigned to the CO adsorption at the Ce⁴⁺ sites on the stoichiometric CeO₂(111) surface, which is consistent with the previous report of 2154 cm⁻¹ for CO that adsorbs to Ce⁴⁺ sites with the molecular axis nearly normal to the surface.^35,37 On a reduced CeO_1.75 surface, CO IR peak at 2165 cm⁻¹ was observed at 100 K, which shifts to 2162 cm⁻¹ with further cooling to ∼90 K. This is a characteristic CO absorption band related with a reduced CeO_x(111) surface, although it is still a subject of debate with respect to the exact nature of the adsorption sites. Wöll’s group revealed a dominating vibrational feature at 2163 cm⁻¹ in the IRRAS results that were recorded after exposing reduced ceria to CO at 74 K.³⁵ They attributed this band to CO adsorption at surface sites in the vicinity of defects on the reduced ceria surface. However, 2163 cm⁻¹ was assigned to CO adsorption directly over Ce³⁺ sites by Stacchiola and co-workers.³⁶ Nevertheless, our current study along with previous investigations clearly demonstrates that CO is a good molecule to probe the nature of the CeO_x(111) surface with distinct CO stretch frequencies between fully oxidized and partially reduced ceria.^35,36

Strong evidence in the reported literature shows that Ni supported over ceria exhibits unique catalytic activity that can be attributed to a strong metal-support interaction. Better activity was observed for Ni with a coverage of ∼0.2 ML or lower on CeO₂.^5,14,16 Single-crystal adsorption calorimetry and DFT studies suggest that Ni binds more strongly to the step edges than terrace sites on CeO₂ and optimum activity of Ni can be obtained when Ni saturates the step edge sites.¹⁶ Figure 3(a) shows the STM image of ∼0.2 ML Ni over CeO₂ at 300 K as an example. As indicated in the line profile, Ni forms small two-dimensional particles that are distributed at both step edges and terraces, which is consistent with our STM results reported previously.¹⁷ The Ni coverage reported in the unit of ML is based on the STM measurements of particle size and density.³⁸ In our CO IRRAS and TPD studies, similar low coverages of Ni (e.g., 0.1–0.3 ML) were deposited onto both fully oxidized CeO₂(111) and reduced CeO_1.75(111) at 300 K. The Ni-ceria surfaces were subsequently annealed to 700 K.

Figure 3(b) shows the CO adsorption data over both as-deposited and annealed Ni on CeO_1.75. Two CO bands are shown at 1950 and 2066 cm⁻¹ with an introduction of 5 × 10⁻⁸Torr of CO at 300 K onto as-deposited Ni over CeO_1.75(111). IR CO adsorption studies on low-index Ni surfaces have shown that the CO vibrational bands at 2020–2100 and 1910–1960 cm⁻¹ can be assigned to CO adsorption at atop and two-fold bridging Ni sites, respectively.^33,44–49 The C=O stretching vibration of CO on metallic Ni was further confirmed by Thiel and co-workers using infrared emission spectroscopy.⁵⁰ When the Ni-CeO_1.75 surface was cooled down to 100 K for CO adsorption, the position of the 2066 cm⁻¹ band shifts to a higher number of 2092 cm⁻¹. Such behavior was also observed in the CO adsorption study over Ni dispersed on a SiO₂ support at 100 K, which is attributed to the formation of compressed CO overlayers.⁴⁵ Bridging and atop CO absorption bands were observed upon an exposure of 5 × 10⁻⁸ Torr of CO onto the annealed Ni-CeO_1.75 surface. These bands are less intense for the annealed Ni/CeO_1.75 surface. This is consistent with the STM results suggesting the particle aggregation to form larger Ni particles after heating to 700 K, which causes the decrease in the Ni surface site for CO adsorption.¹⁷ When increasing the CO dose up to 1 mTorr at 100 K, the IR peak due to the CO adsorption over the ceria support can be detected at 2157 cm⁻¹.

Atop CO (2062 cm⁻¹) IR peak over Ni was shown upon a dose of 5 × 10⁻⁸Torr of CO onto the as-deposited Ni over CeO₂(111) at 300 K [Fig. 3(c)]. This peak also shifts to a higher wave number (2084 cm⁻¹) with cooling to 100 K. The IR band at 1950 cm⁻¹ due to CO adsorption at the bridging sites of Ni is not shown over the Ni/CeO₂ surface. This is likely due to a slightly lower coverage of Ni on CeO₂ than on CeO_1.75 as confirmed by the AES analysis (not shown), causing the formation of smaller particles on CeO₂. In contrast to the study over Ni/CeO_1.75, little signal was observed upon an exposure of 5 × 10⁻⁸Torr of CO onto the annealed Ni-CeO₂ surface at room temperature. By cooling the surface to 100 K in 5 × 10⁻⁸Torr of CO, a new feature appears at 2146 cm⁻¹ while the CO IR band (2084 cm⁻¹) due to CO adsorption at atop sites of metallic Ni becomes broader with a significantly reduced intensity. Based on the previous CO study over NiO(100), this new feature can be assigned to CO adsorption on Ni²⁺.^51,52 The IRRAS data suggest the oxidation of Ni to Ni²⁺ to form Ni²⁺ as the major species on CeO₂ occurs after heating the Ni-CeO₂ surface to 700 K. This is consistent with the CO TPD desorption data (Fig. 4). A broad CO desorption feature between 150 and 500 K with a main peak around ∼370 K was observed upon an exposure of 5 × 10⁻⁸ Torr CO for 10 min at 100 K over as-deposited Ni on CeO₂. This feature is consistent with molecular CO adsorption and desorption from metallic Ni sites.^33,53 The broad shoulder at lower desorption temperatures of the main peak arises from repulsive CO-CO interactions in compressed CO layer that are formed with high CO doses at 100 K.⁴⁶ The higher desorption feature at 550 K arises from dissociated CO, which was observed upon CO adsorption on Ni metal surfaces and oxide-supported Ni particles.^54,55 The desorption temperature can be dependent on the Ni particle size and the Ni-oxide support interaction. Furthermore, ceria is a reducible oxide and the formed C over Ni can also recombine with lattice O in ceria to desorb as CO at this high temperature range.⁵⁶ The additional feature around 100 K is due to CO desorption from the sample/filament mounting. Upon heating the Ni/CeO₂ surface to 700 K, a different desorption profile was observed. The main desorption peak at ∼125 K is consistent with desorption of CO from NiO.⁵⁴ CO TPD profiles over as-deposited and annealed Ni on CeO_1.75 (data not shown) only show the characteristics that is in agreement with molecular CO adsorption and desorption from metallic Ni sites. Our combined XPS, CO IRRAS, and TPD studies have demonstrated that the major fraction of deposited Ni with low coverages on CeO₂ is oxidized to Ni²⁺ with heating to 700 K, while metallic Ni is the only species that is present on reduced ceria.²² 

In summary, we examined the surface sites over Ni-ceria using CO IRRAS and TPD. Well-ordered CeO₂(111) and CeO_1.75(111) thin films were prepared on Ru(0001) under UHV conditions as oxide supports for Ni in order to investigate the effect of the degree of Ce reduction (Ce³⁺) and surface structure of ceria. CeO₂(111) thin film consists of smooth terraces and atomic steps on the surface while additional surface defects are present on the terraces of partially reduced CeO_1.75. The nature of the ceria surface influences the CO adsorption and interaction. 1 mTorr CO adsorption over fully oxidized CeO₂ and partially reduced CeO_1.75 at 100 K shows distinct IR bands at 2158 and 2165 cm⁻¹, respectively. Ni with low coverages grows two-atomic layer high clusters on ceria at 300 K, which agglomerates into three-dimensional structures with heating to 700 K. Atop and bridging CO IR peaks (1950 and 2066 cm⁻¹) over metallic Ni are shown over as-deposited and annealed Ni on CeO_1.75(111) at room temperature under UHV conditions. IR bands due to CO adsorption over metallic Ni were also observed upon Ni deposition on CeO₂ at 300 K. However, a dominant IR peak at 2146 cm⁻¹ was detected over the annealed Ni on CeO₂ that is consistent with CO adsorption on Ni²⁺ species. Our CO IRRAS and TPD results demonstrate that ceria plays a role in the oxidation state of deposited Ni. IRRAS and TPD studies over model Ni-CeO_x(111) surfaces using CO as the probe molecule can elucidate the surface sites that provide insight into their catalytic behavior.

The group at the University of Wyoming wants to acknowledge the support from National Science Foundation (NSF) (Grant No. CHE 1151846) and Wyoming NASA Space Grant Consortium (NASA Grant No. NNX15AI08H). J.Z. thanks J. A. Rodriguez for hosting her sabbatical leave at his group at BNL. The work carried out at BNL was supported by the U.S. Department of Energy under Contract No. DE-SC 0012704. S.D.S. is supported by a U.S. Department of Energy Early Career Award.

The data that support the findings of this study are available from the corresponding author upon reasonable request.