Effect of Pr in CO2 Methanation Ru/CeO2 Catalysts

CO2 methanation has been studied with Pr-doped Ru/CeO2 catalysts, and a dual effect of Pr has been observed. For low Pr content (i.e., 3 wt %) a positive effect in oxygen mobility prevails, while for high Pr doping (i.e., 25 wt %) a negative effect in the Ru–CeO2 interaction is more relevant. Isotopic experiments evidenced that Pr hinders the dissociation of CO2, which takes place at the Ru–CeO2 interface. However, once the temperature is high enough (200 °C), Pr improves the oxygen mobility in the CeO2 support, and this enhances CO2 dissociation because the oxygen atoms left are delivered faster to the support sink and the dissociation sites at the interface are cleaned up faster. In situ Raman spectroscopy experiments confirmed that Pr improves the creation of oxygen vacancies on the ceria lattice but hinders their reoxidation by CO2, and both opposite effects reach an optimum balance for 3 wt % Pr doping. In addition, in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments showed that Pr doping, regardless of the amount, decreases the population of surface carbon species created on the catalysts surface upon CO2 chemisorption under methanation reaction conditions, affecting both productive reaction intermediates (formates and carbonyls) and unproductive carbonates.


INTRODUCTION
Cerium oxide materials have relevant utility in heterogeneous catalysts and have been the center of a significant battery of studies devoted to gaining insight into their high performance. The unique properties of ceria rely on its oxygen storage/ release capacity (OSC) and strong synergistic metal−support interactions when a metal phase is dispersed on a ceria matrix. 1−3 Not only promoting their catalytic activity but also understanding the effect on these active sites will lay the foundations of a rational design of heterogeneous catalysts. CeO 2 doping with other cations usually improves the catalytic features due to the creation of oxygen vacancies, 4−14 but the beneficial effect is often obtained only when the content is optimized. It is well-known that doping ceria with Zr 4+ , Ti 4+ , Eu 3+ , La 3+ , or Tb 3+ cations, among others, improves the catalytic activity in different reactions, such as volatile organic compound (VOC) oxidation, CO oxidation, soot combustion, and NOx reduction. [5][6][7]14,15 Ceria−praseodymia catalysts have demonstrated superior redox properties and improved catalytic performance compared to other ceria-based mixed oxides in soot combustion and CO oxidation reactions, and these Ce−Pr formulations are currently under investigation in other catalytic applications. 16−20 CO 2 conversion to methane by H 2 is a practical strategy for clean energy utilization with the dual benefit of reducing CO 2 emissions while meeting the increasing energy demand forecasted for the next few decades. 21,22 Metals supported on ceria can activate CO 2 molecules at lower temperatures compared to analogue catalysts with other supports. Ceria improves the efficiency of the reaction because oxygen vacancies serve as active sites for CO 2 dissociation. 23−27 Up to now, the research on this reaction has shed light on the catalytic role of each active center by means of in situ spectroscopies and other advanced techniques. 28−34 It is known that CO 2 methanation requires a bifunctional catalyst able to undertake both H 2 and CO 2 activation processes on one or two different type of sites. In heterogeneous catalysts based on inert supports (such as Al 2 O 3 or SiO 2 ), both events take place on reduced metal sites. 35−40 However, ceria-based CO 2 methanation catalysts are proven to be more efficient because ceria contributes to CO 2 chemisorption and dissociation in cooperation with the metal. That is, H 2 dissociation takes place preferentially on reduced metal sites while the most efficient sites for CO 2 dissociation are located at the metal−ceria interface. 28,41 This study focuses on the investigation of the role of Pr in Ru/CeO 2 catalyst for CO 2 methanation by analyzing the effect of Pr loading. Pr is able to adopt +3 and +4 oxidation states, like Ce cations, and Pr cations can be introduced into the fluorite CeO 2 lattice, forming a solid solution until ∼30% Pr (with regard to Ce + Pr), with segregated Pr 6 O 11 being expected above this threshold. 42 In this study, in situ Raman and in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments have been performed for the assessment of the role of oxygen vacancies and the identification of reaction intermediates under reaction conditions. Steady-state isotopic exchange (SSIE) experiments have been performed to study the oxygen exchange between CO 2 and the catalysts by evaluating the activation of the CO 2 molecules on the Prdoped Ru/CeO 2 catalysts.

EXPERIMENTAL DETAILS
2.1. Catalyst Preparation. Ruthenium catalysts with Prdoped CeO 2 supports, referred to as Ru/Ce%PrO x onward, were prepared and used in this study. The "%" corresponds to the weight percent of Pr with regard to Ce + Pr, which ranges from 0 to 25%. A Ru/PrO x catalyst was also prepared.
The necessary amounts of Ce(NO 3 ) 3 ·6H 2 O (99%, Sigma-Aldrich) and Pr(NO 3 ) 3 ·6H 2 O (99.9%, Sigma-Aldrich) were physically mixed in a mortar and calcined at 600°C during 6 h. Nitrates decompose to oxides during calcination and Ce 3+ and Pr 3+ cations were oxidized to a 4+ oxidation state, forming a Ce x Pr 1−x O 2 solid solution. Ruthenium was loaded afterward by incipient wetness impregnation of the corresponding amount of ruthenium(III) acetylacetonate (97%, Sigma-Aldrich) to achieve 3 wt % Ru on the final catalysts. Finally, the impregnated samples were thermally treated at 350°C during 3 h in N 2 atmosphere, using a heating ramp of 5°C/min.

Catalysts
Characterization. The ruthenium content was determined by inductively coupled plasma−optical emission spectrometry (ICP-OES) in a PerkinElmer device (Optima model 4300 DV) after digestion of the catalysts in a HCl/HNO 3 (3:1 volume) mixture assisted by microwaves.
The crystalline structure of each catalyst was analyzed by Xray diffraction in a Rigaku Miniflex II diffractometer. The diffractograms were recorded in a range of 2θ from 10°to 90°, with a step of 0.025°. The wavelength used was λ = 0.155418 nm, corresponding to the Cu Kα radiation. The average crystal size of ceria was determined using the Scherrer equation. N 2 physisorption isotherms were measured at −196°C in an Autosorb-6 device (Quantachrome) after outgassing the catalysts during 4 h at 250°C under vacuum conditions. The reducibility of the catalysts was examined by H 2 temperature-programmed reduction (H 2 -TPR) in a Micromeritics Pulse Chemisorb 2705 device. For such measurements, 25 mg of catalyst was loaded in a tubular quartz reactor coupled with a thermal conductivity detector (TCD) while flowing a mixture consisting of 40 mL/min of 5% H 2 /Ar. The temperature was increased at a 10°C/min pace from room temperature up to 950°C.
2.3. Catalytic Tests. Catalytic tests were performed in a fixed-bed tubular reactor (10 mm inner diameter) containing 100 mg of catalyst mixed with SiC particles (1.00−1.25 mm) to reach a bed volume of 1 cm 3 . The catalyst was pretreated in situ at 500°C for 1 h under 100 mL/min of a 50% vol H 2 /He mixture. After cooling down to room temperature, the reaction mixture was fed to the reactor. The feed consisted of 100 mL/ min of 10% CO 2 , 40% H 2 , and He balance at 1 atm. The GHSV was 9 000 h −1 . The outlet gases were monitored under steady-state conditions at each temperature with a gas chromatograph (Agilent 8860 GC System) equipped with two packed columns (Porapak Q 80/100 for CO 2 and Molecular Sieve 13× for O 2 and CO separation) coupled to a TCD.
2.4. Isotopic Experiments. Steady-state isotopic experiments were performed with 13 C 18 O 2 (Aldrich; 99% 13 C, 95% 18 O) pulses in a fixed-bed cylindrical reactor with 4 mm inner diameter coupled to a Pfeiffer Vacuum mass spectrometer (model OmniStar) operating at 1 ms frequency. The catalytic bed (50 mg) was pretreated under 20 mL/min of 50% H 2 /He mixture at 500°C for 1 h and then cooled down to room temperature under He atmosphere. The reaction mixture (10% CO 2 , 40% H 2 , and He balance) was then continuously fed to the reactor (20 mL/min), and different gas pulses were fed to this main stream. A six-way valve with a loop of 100 μL was used and was filled with 9 psi of the gas to be pulsed. This volume of gas was dragged by the main gas stream once the position of the valve was changed. One Ar pulse followed by a pulse of isotopic 13 C 18 O 2 was fed at 25, 100, 200, 250, and 300°C .
2.5. In Situ Raman Spectroscopy Experiments. In situ Raman spectra were recorded in a LabRam Jobin Ivon Horiba instrument with a laser excitation source of He/Ne (632.8 nm). Experiments were performed in a high-temperature chamber fed with a gas flow of 60 mL/min. The catalysts were reduced with 50% H 2 /N 2 for 1 h at 450°C, and then the catalysts were cooled down to room temperature under N 2 flow. The methanation gas mixture (10% CO 2 , 40% H 2 , and N 2 balance) was fed to the cell, and spectra were recorded in steady state at 25, 100, 200, 300, and 400°C. A monocrystalline Si reference (521 cm −1 ) was used to calibrate the position of the bands.
2.6. In Situ DRIFTS Experiments. In situ DRIFTS experiments were performed in a Jasco infrared spectrometer, model FT/IR-4000, using a reaction cell with temperature and gas flow control. The gas composition was monitored during the experiments with a Pfeiffer Vacuum mass spectrometer (model OmniStar). The catalytic bed consisted of 90 mg of catalyst, which was pretreated in 50% H 2 /He at 450°C for 1 h and then cooled down to room temperature under He atmosphere. A background spectrum was recorded in these conditions, and then 100 mL/min of the methanation mixture (10% CO 2 , 40% H 2 , and He balance) was fed. Spectra were recorded from 4000 to 1000 cm −1 in steps of 1 cm −1 at 25, 100, 200, 300, and 350°C once steady-state conditions were achieved at each temperature.

RESULTS AND DISCUSSION
3.1. Catalytic Tests. Figure 1a shows the CO 2 conversion to methane as a function of the temperature obtained in the catalytic experiments, and Figure 1b presents the conversion CO 2 at a selected temperature (270°C) as a function of the Pr loading of the catalysts. All catalysts displayed 100% selectivity to methane, which is a feature that can be attributed to ruthenium catalysts, 43 48 The thermal treatment performed to decompose the Ru precursor salt was carried out under inert atmosphere in order to minimize this effect, but the formation of RuO x volatile species with the oxygen available in the metal precursors cannot be ruled out. It is known that ceria interacts strongly with Ru species and diminishes the release of volatile RuO x oxides with regard to other supports, such as SiO 2 or Al 2 O 3 . 40−49 This effect would explain the lower Ru content determined for the Ru/Ce25PrO x catalyst (2.1% of Ru) with regard to that measured for the catalyst without Pr (2.5% for Ru/CeO 2 ) and with little Pr (2.6% for Ru/Ce3PrO x ) because Pr doping partially hinders the RuO x −CeO 2 interaction, as will be demonstrated later. The effect of Pr doping in the Ru content is relevant for 25% Pr doping but not for 3% Pr doping.
The catalyst porosity was studied by N 2 adsorption− desorption, and the isotherms at −196°C are shown in Figure 2.
All isotherms show certain adsorption at low partial pressure due to the presence of narrow porosity followed by a hysteresis loop at higher partial pressures attributed to meso-and/or macropores. Pr doping changes the shape of the hysteresis loop, whose type indicates, for the Ru/CeO 2 catalyst, the presence of mesopores, while the absence of a plateau in the isotherms of Ru/Ce3PrO x and Ru/Ce25PrO x evidences the formation of macropores. The Brunauer−Emmett−Teller (BET) specific surface areas, which are included in Table 1, range between 90 and 40 m 2 /g, and these values are consistent with those reported in the literature for similar materials. As shown in Table 1, the Pr-containing catalysts present a lower surface area compared to Ru/CeO 2 due to the promoting effect of the Pr cations in CeO 2 sintering. Usually, CeO 2 doping with foreign cations favors sintering during calcination for moderate temperatures (∼500−700°C) because dopants create defects on the ceria lattice that decrease the energy barrier that must be overcome for crystals to grow. On the contrary, dopants partially avoid sintering for high calcination temperatures where the energy barrier for sintering is already overcome.
The crystalline structure of the catalysts was characterized by XRD, and the diffractograms are compiled in Figure 3. As expected, all catalysts show diffractograms characteristic of the fluorite structure of ceria (JCPDS file 34-0394). Evidence of   The Journal of Physical Chemistry C pubs.acs.org/JPCC Article ruthenium species was not found in the diffractograms, which indicates that ruthenium species are highly dispersed on the ceria support. Identification of PrO x phases in CeO 2 -rich oxides is difficult by XRD because their diffraction patterns are analogous. Pr can adopt several PrO x stoichiometries with x ≤ 2, and their diffraction patterns are only slightly different from that of ceria (JCPDS file 24-1006 for PrO 2 and 42-1121 for Pr 6 O 11 , for instance). In addition, Pr cations can be easily inserted into the CeO 2 matrix. Evidence of segregated PrO x phases can be observed in some cases as shoulders in the CeO 2 peaks, but these are not obvious in Figure 3, suggesting the formation of solid solutions with Pr cations inserted in the CeO 2 framework and/or the presence of highly dispersed PrO x phases spread on the ceria surface.
The lattice parameter and crystallite size of the CeO 2 phase were determined, and the calculated values are compiled in Table 1. The lattice parameter is consistent with that reported for pure ceria (ca. 0.540 nm) for all catalysts, and Pr doping does not significantly affect this value. 18,37 This does not rule out the insertion of Pr cations into the ceria lattice because the sizes of Ce and Pr cations are similar and neither expansion nor contraction of the ceria lattice is expected upon Pr doping. The calculated crystallite sizes are also similar for the three catalysts (13−16 nm), evidencing that Pr doping has only a moderate effect on the size of the primary crystals.
3.3. H 2 Temperature-Programmed Reduction. The reducibility of the catalysts was studied by H 2 -TPR, and the reduction profiles are shown in Figure 4. Three reduction peaks can be distinguished in all reduction curves, but the temperature of each event depends on the nature of the catalyst. The consumption of H 2 on these peaks has been quantified, and the values are compiled in Table 2.
The lowest-temperature peaks at 67, 70, and 167°C for Ru/ CeO 2 , Ru/Ce3PrO x , and Ru/Ce25PrO x , respectively, are mainly attributed to RuO x reduction to Ru. According to the literature, 43,45 ruthenium is expected to mainly form RuO 2 in these types of materials, and the mol of H 2 consumed in the lowest-temperature peak for the Ru/CeO 2 catalyst is half of that required for the total reduction of RuO 2 to Ru. On the contrary, the amount of H 2 consumed in the lowesttemperature peak by the Ru/Ce3PrO x catalyst exceeds the amount required for 100% RuO 2 reduction, suggesting that Ce 4+ and/or Pr 4+ cations are being reduced together with those of ruthenium at this low temperature. This result evidences the enhanced reducibility of the Ru/CeO 2 catalyst upon 3% Pr doping. However, the opposite effect of Pr is observed for 25% doping because the lowest-temperature reduction peak of Ru/ Ce25PrO x is delayed to 167°C, and the consumption of H 2 decreases significantly with regard to Ru/Ce3PrO x .
The intermediate-and highest-temperature peaks must be mainly assigned to the reduction of Ce 4+ and Pr 4+ cations together with the reduction of the remaining RuO x species, the former to surface reduction and the latter to bulk reduction. The presence of shoulders in these peaks evidences the overlapping of several events, which can be attributed to the slightly different reducibilities of the Ce 4+ and Pr 4+ cations and to the presence of different surface species on the catalysts, including hydroxyl groups, carbonates/bicarbonates, etc., that can be reduced or desorbed and methanated during the heating reduction.
Surface reduction of bare ceria usually starts around 400°C, but the surface reduction peak is shifted to 250°C in the Ru/ CeO 2 catalyst due to the catalytic effect of Ru in the reduction of ceria. Ruthenium species, once reduced to a metal state, are very efficient for H 2 dissociation, accelerating the reduction of   The improved reducibility of the support upon Pr doping is also observed in the highest-temperature peak. The temperature of this bulk reduction peak is shifted from 757°C for Ru/CeO 2 to 742°C for Ru/Ce3PrO x as a consequence of the improved reducibility of the support in the latter catalyst, but it roughly keeps the same reduction degree (0.08−0.07 mmol H 2 /mmol Ln). On the contrary, the bulk reduction peak almost disappears for Ru/Ce25PrO x , while the area of the intermediate-temperature peak increases drastically. This behavior occurs because of the high oxygen mobility into the 25% Pr-doped CeO 2 lattice because bulk oxygen is pumped out to the surface with a low energy barrier once the surface oxygen is depleted.
As a summary, these H 2 -TPR experiments allow the conclusion that Pr doping significantly improves the reducibility and oxygen mobility of the ceria support. In addition, RuO x reduction is improved for low Pr loading (3%), while it is impeded for high Pr loading (25%).

Steady-State Isotopic Exchange Experiments.
To analyze the effect of Pr in the interaction of CO 2 with the Ru/ CeO 2 catalysts, steady-state isotopic exchange (SSIE) experiments were carried out using isotopic CO 2 . Table 3 compiles the carbon species and mass spectroscopy signals involved in these changes to facilitate the interpretation of the experiments.
Isotopic CO 2 ( 13 C 18 O 18 O; m/z 49) was pulsed at different temperatures to a methanation mixture with 10% CO 2 ( 12 C 16 O 16 O; m/z 44) + 40% H 2 in He balance, which was continuously fed to the reactor. The oxygen and carbon species present in the catalysts before the isotopic pulses only include 16 O and 12 C atoms, and therefore, the release of species with 18 O and 13 C atoms provides information about the reaction pathway followed by the CO 2 ( 13 C 18 O 18 O) molecules pulsed. Pulse experiments performed without catalyst confirmed that no gas-phase reactions take place in the experiment conditions. Note also that CO species were not detected in the isotopic experiments, in accordance with the high selectivity of these catalysts toward CH 4 Figure 5 allows for the conclusion that the qualitative behaviors of the three catalysts are similar, that is, the types of species evolved are the same for all catalysts. Nevertheless, quantitative differences can be distinguished that are attributed to Pr doping. The quantification of the areas under the different signals in the pulse experiments let us determine the  The Journal of Physical Chemistry C pubs.acs.org/JPCC Article mass balance of carbon in each pulse ( Figure 6). It was considered for those calculations that the system is in steady state, and therefore, there is no net accumulation of carbon species on the catalyst surface after a single pulse. As expected, reaction products (CH 4 and H 2 O) were not detected at 30°C for any catalyst, in accordance with the catalytic experiments shown in Figure 1, and the species released after the CO 2 ( 13 C 18 O 18 O; m/z 49) pulse at this temperature were CO 2 molecules in a diverse isotopic combination. This means that the catalysts are able to chemisorb isotopic CO 2 ( 13 C 18 O 18 O; m/z 49) gas molecules, to break part of the CO bonds even at room temperature, and to release other nonisotopic CO 2 molecules with 12 C and/ or 16 (Figure 6a), and single exchange (Figure 6b and c) is more relevant in Pr-containing catalysts; the higher the Pr loading, the worse is the double exchange. This difference between Pr-free and Pr-containing catalysts suggests that the CO bond-breaking rate is the slowest step of the process that involves: (i) CO 2 adsorption, (ii) CO bond breaking, (iii) reoxidation of surface carbon species with catalyst oxygen, and (iv) CO 2 desorption. The slowest step (CO bond breaking) is faster for Ru/CeO 2 than for Ru/CeXPrO x catalysts, and for this reason the Pr-free catalyst has more time to exchange both oxygen atoms of the CO 2 molecule than the Pr-containing catalysts before the CO 2 molecule is desorbed. It is known that the most active sites for CO 2 dissociation are oxygen vacancies located at the Ru− CeO 2 interface, 41,46 and the negative effect of Pr in the exchange of oxygen atoms between catalyst and CO 2 is in accordance with the negative effect of Pr in the Ru−CeO 2 interaction.
Minor differences in the distribution of the carbon species are noticed between 30 and 100°C, but significant changes are observed at 200, 250, and 300°C. The main change occurring at 200°C is that the percentages of evolved nonisotopic CO 2 (m/z 44) drop while the percentages of double oxygen exchanged CO 2 (m/z 45; 13 C 16 O 16 O) increase, that is, the exchange of oxygen atoms of the chemisorbed CO 2 molecules is enhanced. This means that the slow step of the oxygen exchange process (CO bond breaking) at 30 and 100°C is significantly enhanced at 200°C. As mentioned, the most active sites for CO 2 chemisorption and dissociation are vacant sites located at the Ru−CeO 2 interface, 41 and an increase in the reduction degree of the catalyst surface would promote a higher concentration of oxygen vacancies, which would explain the enhanced CO 2 18 O bonds are broken. The distribution of carbon products changes at 300°C, and the percentages of carbon species are the same for Ru/Ce3PrO x and Ru/Ce25PrO x at this temperature.
As a summary of the isotopic experiments, a dynamic equilibrium has been observed between gas-phase CO 2 molecules and carbon species on the catalyst surface. At low temperature (30 and 100°C), the adsorption−desorption of the CO 2 molecules occurs both without breaking the CO bonds and with breaking these bonds and exchanging oxygen atoms between the CO 2 molecules and the catalysts. At 200°C , the dissociation and exchange of oxygens is significantly favored with regard to the associative adsorption−desorption, and at 250 and 300°C, surface intermediates created upon CO 2 dissociation are hydrogenated to CH 4 .
Pr-doping on the CeO 2 support has different effects in these processes. Pr hinders the Ru−CeO 2 interaction and therefore hinders the dissociation of the CO 2 molecules at low temperature, which takes place at the Ru−CeO 2 interface. However, once the temperature is high enough (200°C), Pr improves the oxygen mobility in CeO 2 , and this enhances the dissociation of the CO bonds because the remaining oxygen atoms are delivered faster to the support sink, cleaning up the vacant sites.
3.5. In Situ Raman Spectroscopy Experiments. The important role of catalyst vacant sites in the CO 2 methanation reaction pathway has been inferred from isotopic experiments because these sites are related with CO 2 dissociation and with oxygen mobility in the ceria support. Raman spectroscopy is a powerful tool to identify vacant sites on ceria, and in situ experiments have been performed to monitor the behavior of such vacant sites under reaction conditions. Figure 7 shows, as a representative example, the Raman spectra recorded for the three catalysts at 25°C under the methanation gas mixture (10% CO 2 , 40% H 2 , N 2 balance) in steady state after reduction at 450°C in 50% H 2 /N 2 for 1 h.
All of the spectra show the same characteristic bands, but the area and intensity of these bands differ in each catalyst and temperature. The main band around 465 cm −1 is assigned to the F 2g mode of the fluorite structure of ceria, and it is produced by oxygen anions breathing around the equilibrium position in the tetrahedral sites of the cubic unit cell. Another relevant band, denoted as D, appears around 565 cm −1 , which is assigned to oxygen vacancies created upon removal of part of the ceria oxygens due to the partial reduction of Ce 4+ and Pr 4+ cations. 50−53 The bands at 523 and 640 cm −1 are assigned to monocrystalline RuO x , and they strongly overlap with the isolate surface oxygen vacancy bands from the CeO 2 lattice and the shoulder of 2TO overtone that results in a flattened band. 54−56 To monitor changes in the oxygen vacancies population during the methanation reaction, all spectra have been deconvoluted, and the relative concentration of vacancies has been calculated for each spectrum as the ratio between the area of the vacant sites D band and the total area of the spectrum. The results of this quantification have been plotted in Figure 8 for the three catalysts during the reduction pretreatment and during further methanation experiments.
As expected, the number of vacant sites increases during the reduction pretreatment with H 2 , and Pr favors the creation of vacancies; the higher the Pr loading, the higher is the amount of oxygen vacancies. Partial reoxidation of these oxygen vacancies takes place by CO 2 once the methanation mixture is fed to the reactor, and the reoxidation rate and degree are different for each catalyst. For Ru/CeO 2 , reoxidation occurs in a relevant extent at 100°C, while reoxidation of oxygen vacancies increases progressively from room temperature until 200°C for Ru/Ce3PrO x and Ru/Ce25PrO x . This filling of oxygen vacancies observed by Raman spectroscopy between room temperature and 200°C is consistent with the dynamic equilibrium between the gas phase and chemisorbed CO 2 molecules deduced from isotopic experiments. and CO 2 + H 2 gas mixture), each catalyst reaches a balance for each temperature between the H 2 reduction and CO 2 oxidation processes, and the amount of oxygen vacancies is the result of this balance. The highest concentration of oxygen vacancies is achieved by the Ru/Ce3PrO x catalyst under CO 2 methanation conditions, and this is also the most active catalyst (Figure 1) above 250°C. This correlation suggests that Ru/Ce3PrO x is able to keep the ceria support more reduced under methanation conditions than Ru/CeO 2 and Ru/Ce25PrO x , and this has a positive effect on the catalytic activity because the chemisorption and dissociation of CO 2 are very effective on reduced sites at the Ru−CeO 2 interface.
These Raman spectroscopy experiments confirm the double role of Pr doping. On the one hand, Pr improves the reducibility of ceria, which is evidenced during the reduction pretreatment by creation of oxygen vacancies on the ceria support. On the other hand, Pr hinders the Ru−CeO 2 interaction, which negatively affects the dissociation of CO 2 and the reoxidation of oxygen vacancies, which is the most relevant event under the methanation gas mixture at temperatures below 200°C. Once the temperature is high enough (T ≥ 200°C), the two effects of Pr doping contribute in opposite ways to ceria reduction by H 2 and reoxidation by CO 2 , reaching an optimum for the Ru/Ce3PrO x catalyst.
3.6. In Situ DRIFTS Experiments. Finally, the CO 2 methanation reaction intermediates were studied by in situ DRIFTS measurements under the CO 2 methanation reaction mixture at steady-state conditions, after reduction of the catalysts at 450°C in H 2 /He for 1 h and cooling down to room temperature in He flow. A background spectrum was recorded at room temperature in He and subtracted from further spectra; therefore, bands shown in the spectra ( Figure  9) only belong to surface species created (or depleted) under the CO 2 methanation mixture. As shown in Figure 9, three relevant wavelength ranges were distinguished with characteristic bands assigned to C−O and CO vibration modes (at 1800−1200 cm −1 range), ruthenium carbonyls (at 2200−1800 cm −1 range), and C−H vibration modes (at 2600−3200 cm −1 range), providing evidence of the formation of bidentate carbonates (at 1580 and 1280 cm −1 ), formates (2825−2950, 1615, and 1380 cm −1 ), and ruthenium carbonyls (1920 and 2017−2040 cm −1 ).
Pr affects the intensity of the signals of the different intermediates. The Ru/CeO 2 catalyst shows bands compatible with the formation of bidentate carbonates, formates, and carbonyls, even at room temperature. These species are the same ones previously detected in similar conditions for Ru/ CeO 2 catalysts, and the changes introduced by praseodymium are related with the intensity of the signals, which are lower for the Ru/CeXPrO x catalysts with regard to Ru/CeO 2 .
The presence of carbonates, formates, and carbonyls on the catalysts surface at room temperature is consistent with the conclusions of the isotopic exchange experiments, where evidence about the dynamic equilibrium between gas-phase CO 2 and surface carbon species was found, and is also in agreement with the reoxidation of the catalysts by CO 2 deduced from Raman spectroscopy experiments. As a general trend, the intensity of the signals increases with temperature until a maximum value, and above this maximum the signals remain stable or decrease. For easier analysis of changes with temperature, Figure 10 shows the intensity of selected bands for each species identified in the spectra after baseline subtraction.
It is reasonable to assume, according to Figure 10, that carbonates are hardly hydrogenated to methane while formates and carbonyls are the most efficient reaction intermediates toward total hydrogenation. Ru/CeO 2 support doping with Pr diminishes the formation of carbonates, as observed in Figure  10a, therefore keeping the catalyst surface with a lower coverage of carbon species with poor relevance for CH 4 production. This effect is expected to have a positive contribution to the catalytic behavior and could be related with the improved ceria reduction by Pr doping, as deduced in H 2 -TPR characterization. It is known that carbonates are created on ceria after chemisorption of CO 2 on surface oxygens, and the improved reducibility of the support will decrease the amount of these types of chemisorption sites under the methanation reaction conditions. The surface concentrations of carbonyls ( Figure 10b) and formates (Figure 10c) also decrease for Pr-containing catalysts with regard to Ru/CeO 2 ; that is, Pr not only hinders the formation of unproductive carbonates but also hinders the formation of the productive reaction intermediates, and this is expected to have a negative effect on the catalytic activity. Once again, DRIFTS experiments show a double role of Pr affecting the formation of carbon surface species that are both productive and unproductive for methane formation.
It is important to compare the trend of the carbonyl ( Figure  10b) and formate ( Figure 10c) signals with temperature. The surface coverage of formates increases from room temperature until 200°C for all catalysts and decreases above this temperature once the production of methane starts to be relevant. This is consistent with the participation of formates as reaction intermediates. On the contrary, carbonyl concentrations increase with temperature until reaching a quite stable value, and this could indicate that the slowest reaction step is the hydrogenation of carbonyls to methane.
In conclusion, the DRIFTS experiments show that Pr doping decreases the population of surface carbon species created on the catalysts upon CO 2 chemisorption under methanation reaction conditions, affecting both productive reaction intermediates (formates and carbonyls) and unproductive carbonates.

CONCLUSIONS
The effect of Pr in Ru/CeO 2 catalyst for CO 2 methanation has been analyzed in this study, and the main conclusions can be summarized as follows: Pr doping in low concentration is beneficial for CO 2 methanation, and it has been observed that a 3%-doped Ru/ CeO 2 catalyst is moderately more active than bare Ru/CeO 2 from 250°C onward. On the contrary, high Pr concentration has a negative effect on the catalytic activity.
H 2 -TPR experiments showed that Pr doping has a double role in the reducibility of Ru/CeO 2 . Pr hinders the reduction of ruthenium species because it partially impedes the Ru-CeO 2 Pulse experiments with isotopic CO 2 evidenced a dynamic equilibrium between gas-phase CO 2 molecules and carbon species on the catalyst surface. At low temperature (30 and 100°C ), the adsorption−desorption of the CO 2 molecules occurs both without breaking the CO bonds and with breaking these bonds and exchanging oxygen atoms between the CO 2 molecules and the catalysts. At 200°C, the dissociation and exchange of oxygens is significantly favored with regard to the associative adsorption−desorption, and at 250 and 300°C the surface intermediates created upon CO 2 dissociation are hydrogenated to CH 4 . Pr doping on the CeO 2 support has different effects in these processes. Pr hinders the Ru−CeO 2 interaction and therefore hinders the dissociation of the CO 2 molecules at low temperature, which takes place at the Ru− CeO 2 interface. However, once the temperature is high enough (200°C), Pr improves the oxygen mobility in CeO 2 , and this enhances the dissociation of the CO 2 molecules because the remaining oxygen atoms are delivered faster to the support sink and the dissociation sites are cleaned up faster. In addition, once the methanation temperatures are achieved (T > 200°C), hydrogenation of the previously chemisorbed surface carbon species is more favorable for Ru/CeO 2 , while Pr doping favors the hydrogenation of just chemisorbed CO 2 molecules.
In situ Raman spectroscopy experiments confirmed the double role of Pr doping and showed that Pr improves the reduction of ceria by H 2 , with creation of more oxygen vacancies on the Pr-doped ceria supports than in bare ceria. Nevertheless, Pr hinders the Ru−CeO 2 interaction, and this negatively affects the dissociation of CO 2 and the reoxidation of oxygen vacancies, which is the most relevant event under the methanation gas mixture at temperatures below 200°C. Once the temperature is high enough (T ≥ 200°C), the two effects of Pr doping contribute in opposite ways to ceria reduction by H 2 and reoxidation by CO 2 , reaching an optimum for 3% Pr doping.
Finally, in situ DRIFTS experiments evidenced that Pr doping decreases the population of surface carbon species created on the Ru/CeO 2 catalysts surface upon CO 2 chemisorption under methanation reaction conditions, affecting both productive reaction intermediates (formates and carbonyls) and unproductive carbonates.