Yi Wu, Pengfei Song,*, Ningyan Li, Yanan Jiang, Yuan Liu*

Tianjin Key Laboratory of Applied Catalysis Science and Technology, School of Chemical Engineering, Tianjin University, Tianjin 300350, China

Keywords:

ABSTRACT

1. Introduction

Ethanol, as a renewable carbon resource, has attained great attraction in recent years [1–3]. It can be used to produce hydrogen, ethylene, ethyl acetate and other industrial chemicals [4,5].At present,the synthesis of ethanol is mainly through the fermentation of biomass, hydration of ethylene and direct thermochemical conversion of syngas [6–8]. Among these approaches, ethanol synthesis from syngas is a promising route due to its most environmentally friendly and more economical.

Direct ethanol synthesis(DES)from syngas mainly include four catalyst systems: modified methanol synthesis system [9,10];modified Fischer-Tropsch (F-T) synthesis system [11–14]; Mobased synthesis system[15–17]and noble metal synthesis system[8,18]. It is well known that Co-based FTS catalysts have a greater alcohols selectivity and low price compared to other catalysts[15,19]. Tan et al. [20] fabricated a series of CuCo/SiO2catalyst assisted with surfactant, and found that a proper amount of CTAB can promote the dispersion of metal particles and be beneficial to the CO hydrogenation.Song et al.[21]prepared self-optimized and renewable Ni-Co alloy@Co-Co2C catalyst,and found that the loss of Ni was suppressed due to the structure of core–shell and active sites can be regenerated.The CuCo active sites are in close contact to promote the production of ethanol, contributing to the catalyst structure changed during the reaction process [22]. Ethanol production from syngas requires two types of active sites in close contact, which are dissociated adsorption sites and non-dissociated adsorption sites.More importantly,these two different active sites(Coδ+-Co0) can be generated on Co in Co-based catalysts [23]. Co0is responsible for the dissociation and adsorption of Co to generate*CHx, and Coδ+is responsible for the non-dissociation and adsorption of Co to generate *CHxO.

In addition to the need for different active sites,the ratio of the two active sites is also an important factor in the generation of ethanol. Earlier studies proposed that the ratio of metal Co0/Co2+is the pivotal to HAS and selectivity can be adjusted through synergistic interactions among different Co species [24]. Chen et al.[23]prepared Co/CeO2catalysts,where Co2+material can be stabilized under CeO2due to the strong interaction of metal carriers.Ga element can play the same role,and the close contact of Ga with Co helped to separate Co sites and made Co get electrons[25].Besides,the isolated Co sites are responsible for the non-dissociated CO adsorption, facilitating the CO insertion. In recent years, cobalt metal carbide (Co-Co2C) catalysts have been applied to the synthesis of higher alcohols. Song et al. [26] found that Ca1-xLaxTi1-x-CoxO3catalysts were prepared and the ratio of Co0to Co2+was adjusted to change the amount of dissociative adsorbed and nondissociative adsorbed CO. Wang et al. [27] found that the valence state of Co can be changed through the addition of Ni to Cobased catalysts. According to these studies, regulating the Co0/Co2+ratio is a difficult task and promising strategy to improve the catalytic performance of Co-based catalysts for syngas to ethanol.

Perovskite oxides with a range of stoichiometries have a great potential in catalytic material, which can produce a large of catalysts combined with the periodic table of elements. In the ABO3structure,the transition metal ions at the A and B sites are confined in the PTO lattice, and these metal ions can be uniformly mixed and highly dispersed at the atomic level [28]. The B-site cation(Fe,Co,Ni)can be substituted by other cations with different state,thus the activity sites and oxygen vacancy can be control by this way[29–32].Some metal cations(Fe,Co,Ni)can be reduced under high temperature and reductive atmosphere, playing the role of active site.Moreover,the structure of perovskite has hydrothermal stability in high-temperature reactions [33–35]. In our previous work, Ca1-xLaxTi1-xCoxO3(x = 0–0.5) were used to the direct conversion of syngas to higher alcohols, and the structure of perovskite was retained after reduction and reaction [26].

Based on the above considerations, a series of CaTi0.9-xCox-Mo0.1O3(x=0,0.1–0.4)and CaTi0.7Co0.3O3catalysts were prepared to improve the selectivity of ethanol via CO hydrogenation. With perovskite composite oxide as the precursor,Ca,Ti,Co and Mo ions can be uniformly dispersed among perovskite macromolecules,leading to Co and Mo can be closely combined in the reduced catalyst. Co0-Co2+active pairs were adjusted through the change of Co/Mo ratio to increase ethanol selectivity. The doping of Mo can also stabilize the Co element in the ionic state and produce more non-dissociated active sites, leading to regulate the ratio of Co0and Co2+and synergistically producing more advanced alcohols.

2. Experimental

2.1. Catalyst preparation

A series of CaTi0.9-xCoxMo0.1O3(x = 0, 0.1–0.4) and CaTi0.7Co0.3O3catalysts were prepared by citric acid complexation.Firstly, calcium nitrate, cobalt nitrate, ammonium molybdate and citric acid were added to deionized water and dissolved with sufficient stirring, then tetrabutyl titanate and citric acid were dissolved in ethanol and added to the above mixed solution. Where Ca : Ti : Co : Mo : citric acid : glycol = 1:(1-x):x:0.1:1.26:0.252.The resulting solution was stirred at 80 °C until a sol was formed.Next, the resulting solid was dried at 120 °C for 12 h and then calcined at N2atmosphere 500 °C for 3 h and air atmosphere 700 °C for 4 h with a ramp rate of 2 °C?min-1. The synthesized CaTi0.9-xCoxMo0.1O3(x = 0, 0.1–0.4) and CaTi0.7Co0.3O3catalysts were abbreviated as CT0.9-xCxM0.1O3and CT0.7C0.3O3.

2.2. Catalyst characterization

The N2adsorption–desorption test was performed on a Micromeritics Trwastar 3000 physisorption instrument at liquid nitrogen temperature(-196°C).Prior to the test,the samples were degassed under vacuum at 300 °C for 4 h to remove the adsorbed material from the sample surface.The specific surface area was calculated from the BET method and the desorption branch data were used to calculate the pore volume and pore size.

X-ray powder diffraction(XRD)tests were done on a Bruker D8-Focus X-ray diffractometer with an angular range of 20°-80°and a scanning speed of 2(°)?min-1,and the radiation source was nickelfiltered Cu Kα (λ = 0.15406 nm) rays.

Programmed temperature rise reduction (TPR) tests were performed on a Thermo-Finnigan instrument. A sample (50 mg) with a particle size of 250–380 μm was placed in a quartz tube (D:0.95 cm,L:23 cm)and purged with 5%H2/Ar for 30 min to remove the air from the tube before the test, after which the temperature of the tube was heated from room temperature to 900°C at a rate of 10°C?min-1.The change in H2before and after the reaction was detected by a TCD detector.

Transmission electron microscopy (TEM) photographs and energy dispersive spectroscopy(EDS)analysis were taken on a field emission scanning electron microscope model JEOL JEM-F200.The samples were ground, sonicated in anhydrous ethanol, and the resulting suspensions were deposited on copper grids with porous carbon film carriers.

X-ray photoelectron spectroscopy (XPS) was obtained on an Axis Supra X-ray photoelectron spectrometer.The radiation source was Al Kα (hυ = 1486.6 eV). Correction was performed using the C 1s (284.6 eV) binding energy of the sample.

The CO programmed temperature desorption(CO-TPD)test was also performed on a Thermo-Finnigan instrument.In order to avoid the suspicion of oxidation on the surface of the reduction catalyst exposed to air, 100 mg of the sample was first treated with H2at 200 °C and He at 300 °C for half an hour. After the temperature was lowered to 25 °C, pure CO was passed through for 0.5 h to adsorb on the catalyst. After that, the gas was switched to He at a temperature from 25 °C to 900 °C at a rate of 10 °C?min-1and a flow rate of 30 ml?min-1.

The analyzer used for the thermogravimetric (TG) analysis test was the DTG-50/50H. The test atmosphere was air, and the temperature was heated from 25 °C to 900 °C with a heating rate of 10 °C?min-1.

2.3. Catalytic performance measurement

The performance of the catalyst was tested on a stainless steel fixed-bed continuous flow microreactor and a combined gas chromatography (GC) system. Before, the catalyst was reduced at 670 °C for 150 min under pure hydrogen (H2) with a heating rate of 30 ml?min-1. After cooling to room temperature, the reduced catalyst was loaded into the stainless steel reactor after diluting 500 mg of particle size 40–60 mesh (0.25-0.425 mm) with 500 mg of SiO2. Subsequently, a gas mixture with a molar ratio of H2/CO/N2= 8/4/1 was passed into the reactor, where N2was used as the internal standard gas for chromatography, and the reaction pressure was increased to 3 MPa with a reaction air velocity(GHSV)of 3900 ml?g-1?h-1and a reaction temperature range of 260–300°C.The reaction effluent gases(N2,CO,CO2and CH4)were separated by a TDX-01 packed column (2 m) and analyzed by a connected TCD detector. The collected liquid products and hydrocarbons were detected by a Porapak-Q column (3 m) and an attached FID detector.

The CO conversion(XCO),selectivity(Si)and mass fraction of an alcohol (Wi) in the collected liquid product were calculated using the following equations.

where,(CO)inand(CO)outrepresent the moles of CO in the feed gas and exhaust gas, respectively; n is the number of carbon atoms; Ciand mirepresent the moles of carbon-containing products and the mass of a certain alcohol in the liquid product, respectively.

3. Catalyst Characterization

3.1. N2 Adsorption and Desorption Analysis

The N2physical adsorption and desorption isotherms of all catalysts were shown in Fig.1(a).All catalysts exhibit typical langmuir type IV isotherms with H2-type hysteresis returns, indicating the formation of mesoporous structure [36]. The adsorption–desorption isotherms of all samples vary slowly at relative pressures (P/P0) of 0.7–1.0, indicating a relatively wide pore size distribution[27].The corresponding pore size distribution was shown in Fig. 1(b). The pore sizes are concentrated in the range of 5–17 nm.

Table S1 (in Supplementary Material) showed the BET specific surface area, pore size and pore volume of all catalysts, it can be seen that the BET surface areas of the CT0.9M0.1O3and CT0.7C0.3O3catalysts are very different.It shows that Mo is enriched on the catalyst surface and the barrier layer formed prevents the growth of catalyst particles, resulting in a high specific surface area. With the increase of Co/Mo molar ratio, the specific surface area and pore volume gradually decreased.

3.2. XRD

The XRD results of the calcined catalysts were shown in Fig. 2(a). All catalysts showed typical peaks at 2θ of 23.23°, 33.10°,47.56°, 59.29° and 69.50°, which are attributed to the (1 0 1),(1 2 1), (0 4 0), (0 4 2) and (2 4 2) crystal planes of CaTiO3(JSPDS card No.42-0423). For the CT0.7C0.3O3catalyst, a CaCO3diffraction peak appears at 2θ of 29.4°, After doping with Mo, the CaCO3diffraction peak disappeared, but when the Co/Mo molar ratio was 4, the peak appeared again. The results indicated that a small amount of C was involved in the reaction to form CaCO3during calcination.

After reduction of CT0.9-xCxM0.1O3(x=0,0.1–0.4)and CT0.7C0.3-O3catalysts (Fig. 2(b)), a characteristic diffraction peak attributed to Co (1 1 1) was detected at 2θ of 44.3°, indicating the successful reduction of metallic Co [37] and the intensity of Co diffraction peak increases gradually with the increase of Co/Mo molar ratio.Furthermore, the diffraction peak of the perovskite structure remained, proving that the perovskite structure was still stability after the precipitation of Co ions. The sizes of the reduced metal Co particles were shown in Table S1,which gradually increase with the increase of Co/Mo molar ratio.

The XRD results of the catalyst after the reaction were shown in Fig. 2(c). The diffraction peak of metallic Co was still present and the perovskite structure remains stable throughout the reaction.While, the diffraction peak of CaCO3reappears at 29.4° increasing with the Co/Mo molar ratio, indicating that a small portion of Ca2+flows out of the perovskite structure and combines with the CO2generated during the reaction to form CaCO3[31]. The sizes of the metal Co particles after the reaction were shown in Table S1, which gradually increase with the increase of Co/Mo molar ratio.

3.3. H2-TPR

The H2-TPR results for the calcined samples were shown in Fig.3.There is only one hydrogen consumption peak for the CT0.9-M0.1O3catalyst, which appears around 400–700 °C and is attributed to the reduction of Mo6+to Mo5+and Mo5+to Mo4+[38,39].For the Co-containing samples, there are two different hydrogen consumption peaks in the range of 300–750°C.the hydrogen consumption peaks (I and II) of the CT0.7C0.3O3catalyst are attributed to the reduction of Co3+to Co2+and the partial reduction of Co2+to Co0in the perovskite lattice, respectively [14]. CT0.9-xCxM0.1O3catalysts attributed to the reduction of Co3+to Co2+and Mo6+to Mo5+in 300–550 °C, and to the partial reduction of Co2+to Co0and Mo5+to Mo4+in 550–750 °C.

Furthermore, the reduction peak of CT0.9-xCxM0.1O3catalyst shifted toward higher temperature compared to CT0.7C0.3O3catalyst, indicating that the Mo doping is not favorable for Co reduction. It is worth noting that the hydrogen consumption peaks are small when the doping amount of Co is 0.1 and 0.2, which is due to the low doping amount of Co and the Mo entering the lattice hinders the reduction of Co, making Co stable in the ionic state.And when Co and Mo reached a certain ratio, for example,x = 0.3, the intensity of the hydrogen consumption peak increased significantly,but it was still lower than that of CT0.7C0.3O3catalyst.While the difference in the mass fraction of Co between the two was vary small, indicating that the CT0.6C0.3M0.1O3catalyst produced more Co2+and the Mo species in the perovskite played a role in stabilizing the Co ions.Thus,the reduced CT0.7C0.3O3mainly produced Co0, leading to high activity but low alcohol selectivity,while the catalyst with Co/Mo molar ratio of 3 contains moderate amounts of Co0and Co2+, promoting ethanol production.

Fig.1. Nitrogen adsorption and desorption isotherms of after calcination of CT0.9-xCxM0.1O3(x =0, 0.1–0.4) and CT0.7C0.3O3 catalyst: (a) nitrogen adsorption and desorption isotherms and (b) the distributions of pore size over.

Fig.2. XRD patterns of CT0.9-xCxM0.1O3(x=0,0.1–0.4)and CT0.7C0.3O3 catalyst:(a)after calcination, (b) after reduction, (c) after reaction.

Fig. 3. H2-TPR profiles of CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) and CT0.7C0.3O3 catalyst.

In order to study the reduction degree of the catalyst, H2-TPR analysis was performed on the reduced catalyst. As shown in the Fig. S1, the reduction peaks of CT0.9M0.1O3catalysts around 400–700 °C were similar to the TPR after calcination. on the one hand might be due to the barrier layer formed by the Mo-rich catalyst surface hindering the contact of Mo species with H2and thus its reduction,on the other hand,the catalyst after contact with air oxidation. The hydrogen consumption peaks of the remaining catalysts all showed trailing peaks at 670 °C, indicating the presence of some Co2+not reduced to Co0at 670 °C and under pure hydrogen.

3.4. TEM

The TEM images of the reduced CT0.6C0.3M0.1O3catalyst are shown in Fig.4.The Fig.4(a)showed that metallic Co is uniformly distributed over the catalyst with an average particle size of 9 nm.Lattice streaks with D-spacing of 0.205 nm are observed in Fig. 4(b), attributed to the Co (1 1 1) crystal plane [40]. As seen in the EDS images of Fig. 4(c)-(g), all elements are uniformly dispersed in the catalyst, indicating a uniform and tight distribution of elements in the catalyst,which results in a strong interaction between Co and Mo.

Furthermore, The Fig. 5 showed that the TEM results of the reacted CT0.6C0.3M0.1O3catalyst. The metal Co nanoparticles are still uniformly distributed on the catalyst surface and have an average particle size of 10 nm,which is due to the weak sintering that occurred during the reaction.HRTEM images showed that there are still D-spacing of 0.205 nm for Co(1 1 1)lattice streaks on the crystal plane,consistent with XRD.Importantly,the EDS images of the reacted CT0.6C0.3M0.1O3catalyst showed that the elements remained uniformly distributed in the perovskite structure, indicating that the catalyst has good stability during the reaction process.

3.5. XPS

Fig. 4. CT0.6C0.3M0.1O3 catalyst after reduction: (a) TEM images with particle size distribution, (b) high-resolution TEM image, (c-g) EDS elemental mappings images.

Fig. 5. CT0.6C0.3M0.1O3 catalyst after reaction: (a) TEM images with particle size distribution, (b) high-resolution TEM image, (c-g) EDS elemental mappings images.

To investigate the surface properties of the catalysts, XPS analysis of elemens in the reduced catalysts was carried out.As shown in the Fig.6(a),the peaks of Co,Mo,O,Ti and Ca all appeared in the full-scan spectra as expected.The Co 2p spectrum of the catalyst in Fig. 6(b) was shown that the peaks at 778.0–779.2 eV are attributed to Co0, 779.3–780.8 eV are attributed to Co2+, and the content increases with increasing Co incorporation. As seen in Table 1,the binding energy of Co0is higher than that reported earlier in the literature (778.5 eV) [41], and the binding energy increases with the increasing of Co/Mo molar ratio, indicating that Co is in the electron-donating state.In addition,the relative content of the two were analyzed by the ratio of Co0/Co2+(Table 2). It was observed that for CT0.7C0.3O3and CT0.6C0.3M0.1O3catalysts with similar Co mass fractions, the former had significantly more Co0and higher activity than the latter, while the latter had more relative Co2+and higher selectivity of ethanol.In contrast to the CT0.7-C0.3O3catalyst,the Co0/Co2+ratios of the other catalysts were lower than that of CT0.7C0.3O3and increased with increasing Co molar ratio, indicating that Mo would prevent the over-reduction of Co.When the Co incorporation amounts was 0.1 and 0.2, the ratios were 0.23 and 0.39, respectively, indicating that only a small amount of Co was reduced to Co0and most of it existed in the ionic state. When the Co incorporation amount was further increased,the ratios of Co0and Co2+reached the optimum ratio of 1,resulting in the generation of more ethanol.

The binding energies of Mo 3d at 231.6–235.6 eV in Fig. 6(c)were attributed to the 3d5/2and 3d3/2orbitals of Mo6+,the binding energies at 230.9–234.5 eV were attributed to the 3d5/2and 3d3/2orbitals of Mo5+,and the binding energies at 229.3-233.4 eV were attributed to the 3d5/2and 3d3/2orbitals of Mo4+. As shown in the Table 3, the binding energy of Mo6+were lower than the binding energy of (233.1) reported in the literature [42], indicating the existence of electron transfer from Co to Mo and a strong interaction between Co and Mo. The total contents of Mo5+and Mo4+of the CT0.9-xCxM0.1O3catalysts in the table are close to each other,but the selectivity of total alcohols and ethanol in Fig. 8 (b) and Table 4 differ greatly, indicating that Mo5+and Mo4+do not contribute much to the catalytic performance.

Fig. 6. XPS spectra of CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) and CT0.7C0.3O3 catalysts after reduction: (a) survey spectra, (b) Co 2p, (c) Mo 3d, (d) O 1s, (e) Ti 2p, (f) Ca 2p.

The high-resolution O 1s spectrum of the catalyst in Fig. 6(d)exhibits four characteristic peaks, with 528.2 eV attributed to lattice oxygen O2-, 530.3 eV to highly oxidative oxygen species,531.5 eV to hydroxyl or surface adsorbed oxygen, and 532.5 eV to adsorbed water molecules [43–45]. According to the results of peak splitting, the percentages of highly oxidizing oxygen species are listed in the Table 3, and it can be seen that the content grad-ually increases with the increase of Co/Mo molar ratio, indicating that increasing the Co content can promote the generation of oxygen vacancies.Oxygen vacancies can provide additional adsorption sites for CO and assist in the production of alcohols [46].

Table 1 Binding energy (BEs) of Ca, Ti and Co for CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) and CT0.7-C0.3O3 catalysts after reduction from XPS

The Ti 2p of the reduced catalyst was shown in Fig. 6(e). The binding energy at 456.7–457.3 eV is attributed to Ti3+and at 457.5–458.1 eV is attributed to Ti4+, and it can be seen from the Table 1 that the binding energy increases with the increase of Co/Mo molar ratio. Ti3+/Ti4+is listed in Table 2. It is well known that the radius of the elements in the perovskite structure affects the presence of the elements in the structure. the radius of Ti3+is larger than that of Ti4+, which gradually decreases with the increase of the Co/Mo molar ratio.This has an effect on the amount of Mo entering the lattice.The binding energies of Ti 2p and Ca 2p in the CT0.9-xCxM0.1O3(x = 0, 0.1–0.4)and CT0.7C0.3O3catalysts are lower than those of pure TiO2and pure CTO, respectively, indicating that Co provides electrons for Ti and Ca.

The corresponding surface compositions of the reduced catalysts are listed in the Table S2. In all samples, the surface atomic ratios of Ca are higher than the theoretical values due to the spread of the generated CaO on the catalyst. Meanwhile, the surface atomic ratios of Co are all lower than the theoretical values due to the fact that only part of the cobalt is reduced to metallic cobalt and is highly dispersed on the surface, while the ionic cobalt isinside the lattice and has a low contribution to the surface composition,which improves as the Co/Mo molar ratio increases.For the CT0.9M0.1O3catalyst, the surface atomic ratio of Mo is higher than the theoretical value, indicating that Mo is enriched at the surface after reduction. The surface atomic ratio of Mo is lower than the theoretical value after doping this catalyst with Co,further demonstrating that Co facilitates Mo into the lattice.

Table 2 Characteristic data of CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) and CT0.7C0.3O3 catalysts based on CO-TPD and XPS profiles

Table 3 Binding energy (BEs) of Mo 3d for CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) catalysts after reduction from XPS

3.6. CO-TPD

Fig.7 shows the CO-TPD results for the reduced CT0.9-xCxM0.1O3(x = 0, 0.1–0.4) and CT0.7C0.3O3catalysts to investigate the CO surface adsorption properties.The peaks at 150°C for all catalysts are attributed to the resolution of physically adsorbed CO(I),at 390°C to the resolution of non-dissociative adsorbed CO(II),and at 600°C to the resolution of dissociative adsorbed CO(III) [46].

The CT0.7C0.3O3catalyst had only a weak non-dissociative adsorption peak, while the dissociative adsorption peak was high,so the catalyst mainly produced hydrocarbons with low alcohol selectivity.Meanwhile,the dissociative adsorption peak was weak for CT0.9M0.1O3and Co/Mo molar ratio of 1, resulting in low activity. As the Co/Mo molar ratio continues to increase, the peak area increases(Table 2),indicating that the amount of both dissociative and non-dissociative adsorbed CO increases,and the ratio between the content of non-dissociative and dissociative adsorption sites is the largest when the Co incorporation increases to 0.3.It is known that the dissociative adsorption CO on the active site of Co0can be hydrogenated to form*CnHmspecies,then non-dissociative adsorption CO on the active site of Co2+can insert into*CnHmto form alcohol [26], so when x = 0.3 there is a better interaction between Co and Mo, which synergistically catalyzes the hydrogenation of CO to produce more alcohols.

Fig. 7. CO-TPD profiles of CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) and CT0.7C0.3O3 catalysts after reduction.

Moreover, it was observed that as the Co/Mo molar ratio increased, the dissociative adsorption peak shifted toward higher temperatures, indicating that dissociative adsorbed CO was more likely to generate long-chain hydrocarbons, while the nondissociative adsorption peak shifted toward lower temperatures,indicating that non-dissociative adsorbed CO was more likely to desorb from the surface,while the appropriate adsorption intensity was therefore more favorable for ethanol generation [26].

4. Catalytic Performance

As shown in Fig.8,The catalytic performance of reduced CT0.9--xCxM0.1O3(x = 0, 0.1–0.4) and CT0.7C0.3O3catalysts was tested at reaction conditions of GHSV = 3900 ml?g-1?h-1, P = 3.0 MPa, and temperature range of 260–300 °C. The very low CO conversion of CT0.9M0.1O3under the reaction conditions indicates that the catalyst has a weak ability to active CO and H2. It is well known that metal Co has dissociative adsorption of CO, carbon chain growth and hydrogenation, the CO conversion of CT0.9-xCxM0.1O3(x = 0,0.1–0.4) catalysts increased first and then decreases as increases of Co ratio, indicating that the number of metal Co active centers increased first and then decreased. While the CO conversion of CT0.7C0.3O3catalyst with similar Co mass fraction as CT0.6C0.3M0.1-O3catalyst reached the maximum, indicating that the latter reduced more Co0. As shown in Fig. 8(b), almost no alcohols were produced for the catalysts with CT0.9M0.1O3and x = 0.1. The total alcohol selectivity increased and then decreased with increasing Co/Mo molar ratio. The highest selectivity of total alcohol was observed when the Co/Mo molar ratio was 3. As shown in Fig. 8(c), the CT0.7C0.3O3catalyst mainly produced hydrocarbons. When the catalyst was doped with Mo, the selectivity of hydrocarbons decreased significantly. With the increase of Co/Mo molar ratio,the selectivity of hydrocarbons first decreases and then increases.Furthermore,it is evident that the CO2selectivity of the CT0.9M0.1-O3catalyst is high, which is consistent with the sensitivity of Mo species to the water–gas conversion reaction mentioned in the literature[47].When the Co content increased,the selectivity of CO2decreased significantly,indicating that the doping of Co can inhibit the occurrence of the water gas conversion reaction.

As shown in the Table 4, the performance of CT0.9-xCxM0.1O3(x = 0, 0.1–0.4) and CT0.7C0.3O3catalysts at T=270 °C, P = 3.0 MPa,GHSV=3900 ml?g-1?h-1is demonstrated, and it can be seen that both C2+OH reaches above 70 and ethanol predominates in the alcohol distribution. Among them, the highest ethanol selectivity was observed when the Co/Mo molar ratio was 3. A large amount of literature reports that Co0-Co2+is the active pair for ethanol generation [48,49], so the appropriate ratio of Co0/Co2+is more favorable for ethanol generation.In comparison to the high quality catalysts in the literature listed in Table 5,the CT0.9-xCxM0.1O3catalyst is very promising.

Fig.8. CO conversions(a),total alcohol selectivity(b),CxHy selectivity(c)and CO2 selectivity(d)of CT0.9-xCxM0.1O3(x=0,0.1–0.4)and CT0.7C0.3O3 catalysts at P=3 MPa,T=260 to 300 °C, GHSV=3900 ml?g-1?h-1 and H2/CO/N2=8/4/1.

Table 4 Catalytic performance of CO hydrogenation over CT0.9-xCxM0.1O3 (x = 0, 0.1–0.4) and CT0.7C0.3O3 catalysts

Table 5 Catalytic performance of some excellent catalysts reported in the literature

Fig. 9. Stability test of CT0.6C0.3M0.1O3 catalysts after reduction at a GHSV of 3900 ml?g-1?h-1 in the syngas mixture of H2/CO/N2 = 8/4/1 and T = 270 °C, P = 3 MPa: (a)conversion and selectivity; (b) distribution.

Fig. 10. TG curves of the CT0.6C0.3M0.1O3 catalyst after reduction, reaction, and 200 h stability test.

5. Stability

The stability of CT0.6C0.3M0.1O3catalyst was tested for 200 h. The reaction conditions were 270 °C, P = 3.0 MPa, and GHSV = 3900 ml?g-1?h-1. As shown in the Fig. 9, the catalyst showed relatively large fluctuations in performance during the first 40 h of the reaction, and then maintained a relatively stable state.After 130 h of reaction, CO conversion and alcohol selectivity slightly decreased and alkane selectivity slightly increased it can be seen from Fig. 9(b) that methanol gradually increases within 120 h. This may be due to the weakening of the adsorption strength of the active site with increasing stabilization time,which makes it easier to desorb to form methanol. The product distribution maintained a high stability after 120 h.

The XRD results after stability testing were shown in Fig.S2.The CT0.6C0.3M0.1O3remains almost unchanged compared to the reduced catalyst, indicating that the structure of perovskite is not destroyed after 200 h of reaction.The TEM diagram after the stability test is shown in Fig. S3. From the diagram, it can be seen that the catalyst particles are still uniformly distributed and there are still obvious d = 0.205 nm lattice stripes attributed to the Co (1 1 1) crystal plane. The size distribution plot shows that the average particle size of Co increases to 11 nm,indicating partial sintering of the catalyst. The EDS energy spectrum of the Fig. S3 (c)-(g) indicated that the elements are still uniformly distributed and maintain a tight contact.

Fig. 11. Schematic diagram of the catalytic mechanism of the catalyst.

The TG curves after reduction, after reaction and after stability testing were shown in Fig. 10. Three types of weight losses were observed in the ranges of 100–250 °C, 300–380 °C and 400–700°C for post-reaction and post-stabilization catalysts.They were attributed to volatilization of adsorbed moisture,paraffin combustion by formation of high-carbon hydrocarbon compounds and coke combustion, respectively [27], which was accompanied by decomposition of CaCO3in the range of 400–700 °C [55]. Furthermore,the comparison revealed almost no change in the weight loss of high-carbon hydrocarbons and coke after reduction and reaction, indicating the good stability of the catalysts.

6. Reaction Mechanism for HAS

In this work, CO is adsorbed in dissociated form on the active site of metallic cobalt and in non-dissociated form on Co2+, and the two adsorbed CO interact and hydrogenate to produce ethanol.The BET, XRD, and TEM confirmed that the Co and Mo species can promote each other to enter the perovskite lattice and stabilize the perovskite structure,making the two closely bound and uniformly dispersed in the perovskite lattice. The TPR, XPS and TPD results indicated that the Mo species in the perovskite lattice can modulate the degree of Co reduction,preventing the excessive reduction of Co to metal, which can assist Co2+to balance the active sites of CO dissociative adsorption and non-dissociative adsorption. As shown in the CO-TPD results in Table 2, the ratio of II/III reflects the amount of active sites for Co0and Co2+.In general,the dissociation of H2and CO is carried out at the active site of Co0, followed by the conversion of the two combined to*CHxor coupled to*CnHmintermediates. The non-dissociative adsorption of CO at the active site of Co2+to form*CO,and the insertion of*CO into*CHxor*CnHmfollowed by hydrogenation to form ethanol.The conversion of syngas to alcohol depends on the balance between the degree of tight binding of active sites and the number of different intermediates.Combining Tables 2 and 3 confirmed that the introduction of Mo species adjust the number of both active sites, and the ratio of II/III is optimal when Co/Mo is 3. The corresponding catalytic mechanism is shown in Fig. 11. Furthermore, too many dissociative adsorption sites lead to the formation of alkanes, and too many non-dissociative adsorption sites lead to the lack of *CHxor *CnHmand the combination of*CO with OH*to form CO2.Only the appropriate ratio of dissociative and non-dissociative adsorption sites is favorable for the formation of ethanol.

7. Conclusions

A series of CaTi0.9-xCoxMo0.1O3(x = 0, 0.1–0.4)and CaTi0.7Co0.3-O3catalysts were prepared by citric acid complexation and used for EDS from syngas. The active species were highly distributed on the catalyst and the perovskite structure exhibited great stability in the reaction process. Among these investigated catalysts,CaTi0.6Co0.3Mo0.1O3catalyst exhibited the highest ethanol selectivity. The high activity is mainly attributed to the domain-limiting effect of perovskite that allows Mo to balance the ionic valence of Co and regulate the ratio of Co0/Co2+.Besides,the higher ethanol selectivity of the CaTi0.6Co0.3Mo0.1O3catalyst compared to the other catalysts can also be attributed to the balance and close combination of the number of the two active centers, which synergistically catalyzed the CO hydrogenation to produce more ethanol.

CRediT Authorship Contribution Statement

Yi Wu:Conceptualization,Methodology,Formal analysis,Investigation,Data curation,Writing–original draft,Writing–review&editing, Visualization. Pengfei Song: Conceptualization, Writing –review & editing, Supervision. Ningyan Li: Investigation, Writing– review & editing. Yanan Jiang: Investigation, Writing – review& editing. Yuan Liu: Conceptualization, Resources, Writing –review & editing, Supervision, Project administration, Funding acquisition.

Data Availability

No data was used for the research described in the article.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The financial support of this work from National Natural Science Foundation of China (21872101, 21962014) and Science and Technology Program of Zungeer County, Inner Mongolia(2020YY-12).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2023.01.018.