Low temperature methanation of CO2 over an amorphous cobalt-based catalyst†

CO2 methanation is an important reaction in CO2 valorization. Because of the high kinetic barriers, the reaction usually needs to proceed at higher temperature (>300 °C). High-efficiency CO2 methanation at low temperature (<200 °C) is an interesting topic, and only several noble metal catalysts were reported to achieve this goal. Currently, design of cheap metal catalysts that can effectively accelerate this reaction at low temperature is still a challenge. In this work, we found that the amorphous Co–Zr0.1–B–O catalyst could catalyze the reaction at above 140 °C. The activity of the catalyst at 180 °C reached 10.7 mmolCO2 gcat−1 h−1, which is comparable to or even higher than that of some noble metal catalysts under similar conditions. The Zr promoter in this work had the highest promoting factor to date among the catalysts for CO2 methanation. As far as we know, this is the first report of an amorphous transition metal catalyst that could effectively accelerate CO2 methanation. The outstanding performance of the catalyst could be ascribed to two aspects. The amorphous nature of the catalyst offered abundant surface defects and intrinsic active sites. On the other hand, the Zr promoter could enlarge the surface area of the catalyst, enrich the Co atoms on the catalyst surface, and tune the valence state of the atoms at the catalyst surface. The reaction mechanism was proposed based on the control experiments.

Introduction CO 2 is a well known greenhouse gas, while it is also a cheap, nontoxic, and renewable carbon resource. Chemical transformation of CO 2 into fuels or useful chemicals has attracted great attention all over the world. 1-10 Methane is one of the major energy sources in human life that can be easily fed into the existing infrastructures. In addition, methane is also a basic feedstock to produce other value added chemicals. [11][12][13] Hydrogenation of CO 2 into methane, i.e., CO 2 methanation, is among the most important topics of CO 2 valorization. 14,15 CO 2 methanation is a reversible and strong exothermic reaction, and is thermodynamically favorable. However, it is difficult to achieve because of the high kinetic barriers of the eight-electron reduction process. Many transition metals such as Ni, Fe, Co, Ru, Rh, and Pd have been investigated as catalysts to accelerate this reaction. [16][17][18][19] To obtain satisfactory catalytic results, the reaction usually needed to proceed at higher temperature (>300 C), where the undesired endothermic reverse water gas shi (RWGS) reaction tended to occur. For example, nickel based catalysts with various supports were the extensively studied catalysts for CO 2 methanation, which usually operated at 300-350 C. [20][21][22][23] Cobalt or iron based catalysts have also been widely investigated for CO 2 methanation, and satisfactory performances were generally obtained at 400 C or higher. [24][25][26][27][28] Low temperature catalysis is still one of the major challenges in methanation of CO 2 . Design of catalysts that can work effectively at lower temperature has received considerable attention. 29 Although many efforts have been made, the progress was restricted to several noble metal catalysts, especially at a temperature below 200 C. [30][31][32][33][34][35][36] Obviously, low temperature methanation of CO 2 over cheap metal catalysts is highly desirable. Herein we show that the amorphous Zrdoped Co-B-O catalyst can effectively accelerate the CO 2 methanation at above 140 C. Excellent activity was obtained at 180 C, which is comparable to or even higher than those of some noble metal catalysts (Table S1 †). Moreover, the addition of Zr promoter results in the highest promoting factor to date of the catalysts for CO 2 methanation. No CO was observed under all conditions and the reaction was not via the RWGS pathway. To our knowledge, this is the rst report of an amorphous transition metal catalyst that can effectively accelerate CO 2 methanation.

Results and discussion
The catalyst The Co-Zr 0.1 -B-O catalyst was prepared by a liquid phase reduction method using NaBH 4 as the reductant in the presence of ammonia. The reaction results over different catalysts at 180 C are shown in Fig. 1. The selectivities of all the catalysts were high (>97%), while the catalytic activities of different catalysts varied signicantly. The activity of the Co-Zr 0.1 -B-O catalyst was as high as 10.7 mmol CO2 g cat À1 h À1 . The yield of methane was 78.1% and the methane selectivity was 97.8%, with minor C 2+ hydrocarbons as byproducts. In contrast, the activity of the Co-B-O catalyst was merely 0.87 mmol CO2 g cat À1 h À1 , and the activity of the Zr-B-O catalyst was negligible. These indicated that remarkable synergy existed between Co and Zr in the Co-Zr 0.1 -B-O catalyst. The promoting factor, which indicates the ratio of the catalytic activity of the promoted catalyst to that of the non-promoted catalyst, was usually adopted to compare the impact of the promoter on the catalytic performance. In the previous reports of CO 2 methanation, the average promoting factor was about 3.0. 29 The addition of noble Pt to the Co nanocatalyst could enhance the catalytic activity by a factor of 6. 37 In our work, the promoting factor was as high as 12.3, which is remarkably higher than those of the reported catalysts. The Co/Zr 0.1 O x catalyst was prepared by the commonly used method, i.e., coprecipitation, calcination followed by reduction with H 2 at high temperature (400 C). But its catalytic activity was much lower than that of the Co-Zr 0.1 -B-O catalyst. Using the liquid phase reduction method, we also prepared the catalysts with other promoters (Cr, Zn, Al, and Ce). Most of them (Co-Zn 0.1 -B-O, Co-Al 0.1 -B-O, and Co-Ce 0.1 -B-O) were also more effective than the Co/Zr 0.1 O x catalyst, but they were markedly less efficient than the Co-Zr 0.1 -B-O catalyst. In short, Co-Zr 0.1 -B-O was an outstanding catalyst for low temperature CO 2 methanation.

The catalyst characterization
The TEM images of the Co-Zr 0.1 -B-O catalyst are given in Fig. 2a and b. The catalyst was mainly composed of 5-15 nm spherical like particles, the outer layers of which seemed different from the cores. No crystal lattice was observed in the TEM images, indicating that the catalyst had an amorphous structure. No diffraction ring was observed in the selected area electron diffraction (SAED) pattern of the catalyst either, which agrees with the TEM images (Fig. 2c). The XRD pattern showed that the Co-Zr 0.1 -B-O catalyst had no discernible diffraction peak, which further conrmed that it was amorphous (Fig. 3). Actually, all the catalysts prepared by the liquid phase reduction method were amorphous. In contrast, the XRD curve of the Co/ Zr 0.1 O x catalyst displayed remarkable diffraction peaks. The    Fig. S1. † The adsorption isotherm of the Co-Zr 0.1 -B-O catalyst can be classied as a type III curve, suggesting that multilayer adsorption occurred on the lyophobic surface. The Brunauer-Emmett-Teller (BET) surface area was 92.4 m 2 g À1 , indicating that it is not a porous material. The results of the EDS elemental mapping revealed that the Co, Zr, B, and O atoms were well dispersed in the catalyst (Fig. S2 †). The X-ray photoelectron spectroscopy (XPS) characterization suggests that different Co species, i.e., Co 0 , Co 2+ and Co-OH, existed on the catalyst surface ( Fig. S3 †). The FTIR spectra demonstrated the presence of the OH group on the catalyst, which coincides with the XPS result ( Fig. S4 †).

Impact of reaction conditions
Fig. 4 depicts the catalytic results at different temperatures. The reaction could occur at 140 C. The catalytic activity increased quickly with increasing temperature until 180 C. When the temperature was further increased the reaction rate can not be effectively improved. This could be explained by the partial transformation of the amorphous catalyst structure. The XRD analysis demonstrated that obvious crystals of cobalt were formed when the catalyst was treated at higher temperature ( Fig. S5 †). The peaks at 44.2 and 47.3 are ascribed to Co (111) and Co (101), respectively. The methane selectivity was very high and remained nearly constant at different temperatures. Without the precharged CO 2 and/or H 2 no product was detected aer the reaction, demonstrating that both CO 2 and H 2 took part in the reaction. The reaction started by adsorption of CO 2 and H 2 at the catalyst surface. The ratio of CO 2 and H 2 pressures remarkably affected the reaction results, and the suitable ratio was 1/1 (Fig. S6 †). We xed this ratio and conducted the reaction at different total pressures. As expected, the reaction rate increased with increasing pressure, and the increase became slow when pressure was high enough (Fig. S7 †). The adsorption of the reactants CO 2 and H 2 on the catalyst surface was remarkably enhanced by increasing their pressure, which agreed with the results of the N 2 adsorption test. The catalytic performance could also be tuned by the solvent effect. We conducted the reaction in different solvents and cyclohexane was proved to be an appropriate solvent (Fig. S8 †). At the optimized temperature and pressure, we carried out the time course study. It was shown that the reaction rate was very quick at the beginning and it gradually slowed down with the consumption of H 2 (Fig. S9 †).

Effect of the Zr promoter
As revealed in Fig. 1, the Zr-B-O catalyst could not promote the reaction, and addition of Zr promoter to the Co-B-O catalyst could signicantly accelerate the reaction activity. The Zr content in the catalyst obviously affected the catalytic activity, as depicted in Fig. 5. With increasing molar ratio of the Zr promoter (n Zr /n Co : 0.05, 0.1, and 0.3), the reaction rate increased markedly and reached a maximum at 0.1. When the n Zr /n Co was further increased the reaction activity decreased. This suggests that n Zr /n Co ¼ 0.1 was an appropriate ratio. The role of the Zr promoter in modulating the catalyst structure may be ascribed mainly to three aspects, i.e., changing the surface area of the catalyst, enriching the Co atoms on the catalyst surface, and    Table S2. † The results indicated that the surface area of the Co-B-O catalyst (18.8 m 2 g À1 ) could be signicantly enhanced by the Zr promoter. The surface area of the Co-Zr 0.1 -B-O catalyst was 92.4 m 2 g À1 , which was the highest surface area in the Zr doped catalysts. It is equally important that the Zr promoter could greatly enrich the active Co atoms on the catalyst surface. The Co content in total surface atoms of Co-B-O was 7.7%, while it reached above 30% in Co-Zr 0.05 -B-O and Co-Zr 0.1 -B-O catalysts. The synthesis of the Co-B-O catalyst is an exothermic process, which involves high surface energy and tends to cause agglomeration. 38 The addition of the Zr promoter could affect the fabrication of the catalyst. Higher surface area and Co enrichment on the surface may cooperatively increase the active sites for the reaction. However, when excess Zr promoter was added, both the surface area and surface Co content of the catalyst decreased. This may explain partially why the performance of the Co-Zr 0.3 -B-O catalyst was not as good as that of the Co-Zr 0.1 -B-O catalyst. Besides the impact on the structure, the Zr promoter could also alter the valence of the surface atoms (Fig. S3 †). The Zr atoms mostly existed as Zr 4+ in the catalyst, which did not change during the fabrication of the catalyst. With increasing Zr promoter, the Co atoms shied to higher oxidation states. It is noteworthy that the lattice O atoms doped in the catalyst were greatly increased by adding the Zr promoter (Table S3 †). It was reported that the oxygen atoms doped in cobalt metal may create surface defects, which acted as overactive sites and enhanced the rate of the catalytic reactions, including CO 2 methanation. 39,40 Besides the doped O atoms, the amorphous structure may further increase the surface defects. 41 In short, the Zr in the catalyst acted as a structural promoter and an electronic promoter simultaneously, and the Co-Zr 0.1 -B-O was the optimal catalyst.

The reusability of the catalyst
The recycling test of the Co-Zr 0.1 -B-O catalyst was conducted to appraise its reusability. Aer the reaction, the residual gases were analyzed and released, and the catalyst was used directly for the next run. The results of the recycling test indicated that the catalytic performance had no obvious decrease aer ve cycles (Fig. S10 †). The elements of the catalyst were still well dispersed aer the reaction (Fig. S11 †).

Mechanistic discussion
To understand the impact of the peculiar structure on the reaction, we conducted the temperature programmed desorption (TPD) analysis, i.e., CO 2 -TPD and H 2 -TPD (Fig. 6). The results demonstrated that the major desorption peaks of CO 2 and H 2 appeared at closely below 180 C, and the peaks of the Co-Zr 0.1 -B-O catalyst were signicantly larger than those of the Co-B-O catalyst. This may account for the better performance of the Co-Zr 0.1 -B-O catalyst than the Co-B-O catalyst. The peaks of CO 2 and H 2 of the Co/Zr 0.1 O x catalyst were observed at much higher temperature (nearly 300 C), and were much smaller than those of the Co-Zr 0.1 -B-O catalyst. This also helps to explain why the catalytic activity of the amorphous Co-Zr 0.1 -B-O catalyst was markedly higher than that of the Co/Zr 0.1 O x catalyst fabricated by the commonly reported methods.
In most cases of CO 2 methanation, the undesired RWGS reaction also occurred, especially at relatively high temperature (>300 C). The reaction route of CO 2 methanation also depends on the composition and structure of the catalyst. 42,43 In this work, no CO was observed under all conditions. Moreover, CO hydrogenation could hardly take place over the Co-Zr 0.1 -B-O catalyst, and the very small amount of CO consumed in the reaction was mostly converted to CO 2 (Table S4 †). This suggested that the CO disproportionation (CO / CO 2 + C) occurred in the reaction. The carbon deposit generated in situ blocked the active sites and inhibited further reaction. 44 To study the impact of the reactants on the valence of the catalyst, we conducted XPS characterization of the Co-Zr 0.1 -B-O catalysts aer H 2 adsorption and subsequent CO 2 adsorption at reaction temperature (Fig. S12 †). The results revealed that under the reaction conditions Co 0 and Co 2+ coexisted on the catalyst surface and the ratio of Co 0 /Co 2+ uctuated with the sequential introduction of H 2 and CO 2 . Moreover, the valencies of Zr, B, O uctuated synchronously with that of Co. The synergy of these elements in the catalytic cycles also helps to explain the excellent catalytic performance.
To detect the intermediates of the reaction, we conducted the in situ Fourier transform infrared (FTIR) analysis of CO 2 adsorption on the catalyst pretreated with H 2 at 180 C (Fig. S13 †). The result revealed that CO 2 was quickly reduced by the surface H atoms into intermediates, such as formic acid (HCOOH, 1080) and methoxyl (CH 3 O, 1051 cm À1 ) adsorbed on the catalyst surface. 45 Several remarkable peaks between 2000 and 2150 cm À1 were also observed, which could be ascribed to the intermediates H x CO y (x ¼ 1-2; y ¼ 2-3) formed by H 2 , CO 2 and/or hydroxyl groups on the catalyst surface. 46 All these intermediates adsorbed weakly on the catalyst surface and disappeared immediately when the CO 2 ow was stopped and H 2 was introduced again. The in situ FTIR spectrum of the catalyst with simultaneous introduction of CO 2 and H 2 is given in Fig. S14. † The peaks of the above intermediates became very small or could not be observed at all, which also suggested their quick transformation under CO 2 methanation conditions. The major peaks (1210, 1330 and 1573 cm À1 ) are ascribed to the intrinsic feature of the catalyst evolved at the reaction temperature (Fig. S15 †). The quick formation and conversion of the reactive intermediates at the reaction temperature may account for the very high activity of the catalyst. Some intermediates (HCO 3 À , -COO À , and CH 3 O-) were also observed by XPS characterization of the catalyst aer CO 2 methanation (Fig. S16 †). These intermediates may be formed and preserved during cooling of the reactor. Based on above discussion, we proposed the possible reaction mechanism. The CO 2 and H 2 adsorbed on the surface of the Co-Zr 0.1 -B-O catalyst, where the Zr promoter greatly and simultaneously enhanced their adsorption capabilities. The CO 2 and H 2 reacted on the catalyst surface, where the Co, Zr, B and O atoms worked cooperatively. The CO 2 was reduced by the H atoms to methane via a series of intermediates, such as HCO 3 À , HCOO À , HCOOH, CH 3 O-, and other possible species H x CO y (x ¼ 1-2; y ¼ 2-3) containing carbonyl. Under optimized conditions, reactive intermediates were generated and transformed very quickly, accounting for the very high catalytic activity. The outstanding performance of the catalyst may originate from its strong ability to adsorb both reactants and the synergy of the atoms on the catalyst surface in converting them.

Conclusions
In summary, we report an amorphous Co-Zr 0.1 -B-O catalyst for CO 2 methanation. The catalyst was very active and selective, and the activity of the catalyst reached 10.7 mmol CO 2 g cat À1 h À1 at 180 C with a methane selectivity of 97.8%. The promoting factor of the Co-Zr 0.1 -B-O catalyst was as high as 12.3, which is remarkably higher than those of the reported catalysts. It is noteworthy that the catalytic performance is comparable to or even higher than that of some noble metal catalysts under similar conditions. The outstanding performance of the catalyst originated from two aspects. Firstly, the amorphous nature of the catalyst may lead to abundant surface defects and intrinsic active sites. Secondly, the Zr promoter could increase the surface area of the catalyst, enrich the Co atoms on the catalyst surface, and tune the valence state of the atoms at the catalyst surface. All these factors may enhance the activity of the catalyst. In the reaction, CO 2 was reduced by the H atoms to methane via a series of intermediates, such as HCO 3 À , HCOO À , HCOOH, CH 3 O À , and other possible species H x CO y (x ¼ 1-2; y ¼ 2-3) containing carbonyl. We believe that this cheap and highly efficient catalyst has promising potential applications, and the protocol to design amorphous catalysts with promoters is useful to prepare other efficient catalysts using cheap metals. was synthesized by a coprecipitation method. 2 mmol cobalt acetate tetrahydrate and 0.2 mmol zirconium nitrate were dissolved in a mixed solution of 20 mL distilled water and 5 mL ethanol, then 10 mL of sodium hydroxide solution (0.8 mol mL À1 ) was added dropwise into the above solution under stirring, and aer stirring for another 20 min a dark blue precipitate was obtained by ltration, followed by washing with 200 mL distilled water and drying overnight at 80 C. The Co/Zr 0.1 O x was obtained by calcination of the precipitate in air at 400 C for 3 h followed by reduction with 5% H 2 in Ar at 400 C for 1 h.

Catalyst characterization
The N 2 adsorption-desorption isotherms were recorded at 77 K using an ASAP2020 (Micromeritics, USA). The catalysts were treated under vacuum at 150 C for 4 h before the test. The specic surface area was obtained by the Brunauer-Emmett-Teller (BET) method. XRD patterns of the catalysts were obtained using a Rigaku Ultima IV X-ray diffractometer with Cu Ka radiation (l ¼ 1.5418Å) at a rate of 20.0 min À1 over the range of 20-90 . The X-ray photoelectron spectroscopy (XPS) experiment was conducted on an AXIS SUPRA (Kratos Corp.) with an Al Ka excitation source and all binding energies were referenced to the C 1s at 284.8 eV. High resolution transmission electron microscopy (HRTEM) was conducted on a Tecnai G2 F30 FETEM (FEI Corp.). HAADF-STEM images and EDS mapping were obtained using a Tecnai G2 F20 (FEI Corp). FT-IR spectra were collected using an NEXUS670 Fourier transform infrared spectrometer (ScanStantion C5, USA).
CO 2 -temperature programmed desorption (TPD) was conducted using a Micromeritics AutoChem II chemisorption analyzer with He (30 mL min À1 ) as the carrier gas. About 0.1 g catalyst was charged into the quartz tube and heated to 100 C at the rate of 20 C min À1 . Aer 1 h the temperature was cooled down to 50 C, and CO 2 adsorption proceeded with 10% CO 2 -90% He (v/v) mixed gas for 30 min with a ow rate of 50 mL min À1 . Then the sample was purged with He of 50 mL min À1 for 1 h. Finally, CO 2 desorption proceeded from 50 to 600 C at 10 C min À1 . The H 2 -TPD measurement was performed using the same equipment with a similar procedure, except that 10% H 2 -90% He (v/v) was used.
In situ FTIR spectra were recorded with a NICOLET iS50 FT-IR spectrometer (Thermo SCIENTIFIC, USA) equipped with a high-temperature reaction chamber and a mercury cadmium telluride (MCT) detector at a resolution of 4 cm À1 and 32 scans per spectrum. The Co-Zr 0.1 -B-O catalyst and KBr were ground together and put into the sample cup. Then two tests were carried out using fresh catalyst, respectively. The rst test was as follows: at 180 C, the sample was rst purged with N 2 (100 mL min À1 ) for 2 h, and then it was treated with H 2 (50 mL min À1 ) for 2 h. Aer further purging the sample with N 2 (100 mL min À1 ) for 1 h, CO 2 (100 mL min À1 ) was introduced and the CO 2 adsorption began. Then the sample was further purged with N 2 (100 mL min À1 ) for 2 h. In the end, it was treated with H 2 (50 mL min À1 ) for 3 h. The background spectrum was recorded before CO 2 adsorption. The second test was as follows: the chamber was rst purged at 180 C with N 2 for 2 h and cooled down to 20 C. Aer the CO 2 was introduced for 5 min, both CO 2 and H 2 were charged and the sample was heated from 20 to 180 C at the rate of 5 C min À1 and kept at 180 C for 0.5 h. The background spectrum was scanned before CO 2 was introduced.

Catalytic reaction
The methanation of CO 2 was conducted in a 16 mL stainlesssteel autoclave. Typically, 40 mg catalyst and 2 mL cyclohexane were added in the reactor. The reactor was sealed and the air in it was substituted with CO 2 of 1 MPa three times, and then 4 MPa CO 2 and 4 MPa H 2 were charged successively at room temperature. The reactor was heated to 180 C under stirring and was kept for 12 h. Aer the reaction, the reactor was cooled in an ice-water bath, and the residual gas was released slowly and collected for GC analysis (Agilent 7890A) equipped with a thermal conductivity detector (TCD) and ame ionization detector (FID). The amount of hydrocarbons such as methane was determined on the FID using a HP-AL/S column. The amount of carbon dioxide and hydrogen was obtained on the TCD using a HP-PLOT/Q column. The liquid phase was analyzed using an Agilent Technologies 7890B GC system with a ame ionization detector using a HP-5 column. The conversion of CO 2 was the percentage of the CO 2 charged into the reactor that was converted to hydrocarbon products, as is given below.
CO 2 conversion ð%Þ ¼ P nCH 2nþ2ðgeneratedÞ CO 2ðchargedÞ Â 100 Because H 2 was the limiting reactant and it could more effectively reect the proceeding of the reaction, calculation of the methane yield was based on the H 2 charged into the reactor, as is shown below.

CH 4 yield ð%Þ ¼
CH 4ðgeneratedÞ H 2ðchargedÞ 4 Â 100 The CH 4 selectivity was the percentage of the C atoms in CH 4 over the C atoms in total products, as is given below. CH 4 selectivity ð%Þ ¼ CH 4ðgeneratedÞ P nCH 2nþ2ðgeneratedÞ Â 100 The recycling test Aer the reaction, the residual gas was released slowly. The gaseous and liquid samples were analyzed respectively. Then the reactor was sealed and fresh reactants (CO 2 and H 2 ) were charged to start the next run.

Conflicts of interest
There are no conicts to declare.