A CoFe Bimetallic Catalyst for the Direct Conversion of Syngas to Oleﬁns

: Syngas conversion is a useful technology for converting nonpetroleum carbon resources into chemicals such as oleﬁns. Iron-and cobalt-based catalysts, as two major categories, have been extensively studied in Fischer–Tropsch synthesis to oleﬁns (FTO) reactions. Although both iron and cobalt catalysts have shown distinct merits and shortcomings, they are also complementary in their properties and catalytic performances when combined with each other. Herein, Na-modiﬁed CoFe bimetallic catalysts were fabricated using a co-precipitation method. It was found that there was a synergistic effect between Co and Fe that promoted a CO dissociation rate and carburization, and an appropriate Co/Fe ratio was conducive to improvements in their catalytic performances. The desired oleﬁns selectivity reached 66.1 C% at a CO conversion of 37.5% for a Co2Fe1 catalyst, while the methane selectivity was only 4.3 C%. In addition, no obvious deactivation was found after nearly 160 h, indicating their potential industrial application.


Introduction
Fischer-Tropsch synthesis (FTS) can convert syngas (a mixture of carbon monoxide and hydrogen) into valuable chemicals, such as paraffins, olefins, oxygenates, and so on, depending on the catalyst types and reaction conditions used [1][2][3].Among these, olefins, including lower olefins (C 2-4 = ) and higher olefins (C 5+ = ), are very important products that can be used to synthesize various chemicals, such as lubricants, plastics, solvents, and drug intermediates [4][5][6][7].The Fischer-Tropsch to olefins (FTO) reaction has become a promising technology due to its simplified reaction process and low energy consumption compared to the traditional "syngas-methanol-olefins" indirect route.
Promoted Fe-based catalysts have been used in Fischer-Tropsch synthesis to olefins (FTO) processes for a long time [8,9].Torres Galvis et al. reported a supported Fe-based catalyst, and the selectivity of the lower olefins reached 61 C% at 340 • C and 20 bar [10][11][12].Zhai et al. prepared an iron carbide catalyst modified by Zn and Na, and it demonstrated a high level of selectivity for olefins (up to 79%) at 340 • C and 20 bar [13].Although Fe-based catalysts exhibit high olefin selectivity, the reaction temperatures are typically over 300 • C.Under these harsh conditions, particle sintering and carbon deposition on a catalyst's surface easily occurs, leading to rapid catalyst deactivation.
For Co-based FTO catalysts, Zhong et al. reported on a Co 2 C nanoprism catalyst exhibiting high olefin selectivity in an FTO reaction [14].The olefin selectivity reached 60.8 C%, while the CH 4 selectivity was <5 C% at a mild condition (250 • C).Further characterization and theoretical simulation confirmed the exposed facet of ( 101) and (020) on the obtained Co 2 C nanoprisms, which greatly favored olefin production while restraining CH 4 production.Subsequent works have displayed that Na and Mn could act as electronic and structural promoters, respectively, facilitating the formation of Co 2 C with special The Co-, CoFe-, and Fe-based samples were synthesized via a co-precipitation method.The concentrations of cobalt, iron, and sodium in the calcined samples are listed in Table 1.By controlling the washing times, all samples had a similar residual Na content (2.3~3.8 wt.%), and the molar ratio of (Co + Fe)/Mn/Al was approximately 1.0:1.0:1.0.The element concentrations, as determined by XRF, were close to the nominal concentrations.The pore structure properties of the samples are also presented in Table 1 and Figure 1.The as-prepared samples exhibited characteristic mesoporous structures, as evidenced by the similar hysteresis loops for the N 2 adsorption-desorption isotherms with average pore sizes centered at 6~20 nm.The pore size of the sole Co sample was 8.8 nm, and this increased to 12.4 nm for the Co2Fe1 catalyst.However, the pore sizes dropped to 7.2 and 8.0 nm for Co1Fe2 and Fe, respectively.The BET specific surface area showed an opposite trend.For example, a surface area of 222 m 2 /g was observed for the Co sample, and this dropped to 144 m 2 /g for Co2Fe1.Slight increases in the specific surface areas were observed for Co1Fe2 (184 m 2 /g) and Fe (178 m 2 /g).We assumed that this was mainly due to the formation of CoFe composite oxides, such as Co 2 FeO 4 (Co/Fe = 2) species.The combination of Fe species with small Co particles, reduced the number of micropores, resulting in a decrease in the specific surface area and an increase in the average pore size.This phenomenon was even more pronounced for the Co2Fe1 sample (Co/Fe = 2).However, excess Fe cannot combine with Co species.The additional Fe species contributed to the total specific surface area.As a result, the specific surface area increased, and average pore size decreased again.Apparently, the addition of Fe to the Co sample resulted in significant change in the textural property [26].
Catalysts 2023, 13, x FOR PEER REVIEW 3 of 14 to the formation of CoFe composite oxides, such as Co2FeO4 (Co/Fe = 2) species.The combination of Fe species with small Co particles, reduced the number of micropores, resulting in a decrease in the specific surface area and an increase in the average pore size.This phenomenon was even more pronounced for the Co2Fe1 sample (Co/Fe = 2).However, excess Fe cannot combine with Co species.The additional Fe species contributed to the total specific surface area.As a result, the specific surface area increased, and average pore size decreased again.Apparently, the addition of Fe to the Co sample resulted in significant change in the textural property [26].X-ray diffraction was employed to identify the phase structure of various catalysts.Figure 2a shows the XRD patterns of the calcined samples.The diffraction peaks for sole Co and Fe were Co3O4 (JCPDS no.43-1003) and Fe3O4 (JCPDS no.26-1136), respectively.Figure 2b shows the step-scanning patterns of the 2θ from 30 to 45°.As the Fe concentration increased in CoFe samples, the peak position shifted to a higher angle.This was mainly due to the increased intensity of the diffraction peak of MnO2 (JCPDS no.44-0141) with the addition of Fe content.After reduction in a H2 atmosphere at 300 °C for 5 h, diffraction peaks ascribed to CoO (JCPDS no.71-1178) and FeO (JCPDS no.89-0687) phases were found in the Co and Fe samples (Figure 2c).The diffraction peaks of reduced CoFe samples were located between CoO and FeO, indicating the formation of the CoFe composite oxide with a lower valence state.Since the atomic radius of Fe is slightly higher than that of Co (1.26 vs. 1.25 Å), the diffraction peaks shifted to a lower angle with increasing Fe concentration (Figure 2d).
The XRD patterns of spent CoFe catalysts are presented in Figure 2e.All samples showed the diffraction peaks of MnCO3 (JCPDS no. .Due to the weak interaction between Fe and Mn, the separated MnO2 in Figure 2b on a calcined sample would convert to MnCO3 with sharp diffraction peaks during the reaction.For samples containing both Co and Fe, the diffraction peaks of CoO and FeO disappeared due to the carburizing reaction, while diffraction peaks located at 37.0, 34.4,42.6, and 54.6, corresponding to Co2C (JCPDS no.65-1457), were observed.The crystallite size of Co2C calculated by Scherer's formula based on Co2C (111) was 19.4,16.3, 15.9, 10.1, and 9.0 nm for Co, Co5Fe1, Co2Fe1, Co1Fe1, and Co1Fe2, respectively, indicating the addition of Fe could restrain the growth of Co2C.As for the spent Fe catalyst, the peaks of the Fe2C (JCPDS no.17-0897) at 2θ of 41.3, 42.5, and 43.0 and FeO (JCPDS no.74-1886) were found.These results suggested that there was incomplete carbonization in the Fe sample.X-ray diffraction was employed to identify the phase structure of various catalysts.Figure 2a shows the XRD patterns of the calcined samples.The diffraction peaks for sole Co and Fe were Co 3 O 4 (JCPDS no.43-1003) and Fe 3 O 4 (JCPDS no.26-1136), respectively.Figure 2b shows the step-scanning patterns of the 2θ from 30 to 45 • .As the Fe concentration increased in CoFe samples, the peak position shifted to a higher angle.This was mainly due to the increased intensity of the diffraction peak of MnO 2 (JCPDS no.44-0141) with the addition of Fe content.After reduction in a H 2 atmosphere at 300 • C for 5 h, diffraction peaks ascribed to CoO (JCPDS no.71-1178) and FeO (JCPDS no.89-0687) phases were found in the Co and Fe samples (Figure 2c).The diffraction peaks of reduced CoFe samples were located between CoO and FeO, indicating the formation of the CoFe composite oxide with a lower valence state.Since the atomic radius of Fe is slightly higher than that of Co (1.26 vs. 1.25 Å), the diffraction peaks shifted to a lower angle with increasing Fe concentration (Figure 2d).
The XRD patterns of spent CoFe catalysts are presented in Figure 2e.All samples showed the diffraction peaks of MnCO 3 (JCPDS no. .Due to the weak interaction between Fe and Mn, the separated MnO 2 in Figure 2b   The H2-temperature-programmed reduction (TPR) of the calcined sample is presented in Figure 2f.For the sole Co and Fe samples, the first H2 consumption peak around 250 °C was ascribed to the reduction of Mn 3+ to Mn 2+ .For the Co sample, three H2 consumption peaks centered around 300, 336, and 423 °C were observed, which were assigned to the partial reduction of the Co3O4 and CoMn oxide to Co 2+ , and the further reduction of the Co 2+ to Co 0 [27][28][29].For the Fe sample, the reduction peaks around 370 °C and 425 °C were attributed to the reduction of Fe3O4 to FeO, and FeO to metallic Fe [30].The fourth peak centered at 550~620 °C was attributed to the reduction of CoAl and FeAl spinel [31,32].For the CoFe samples, the reduction temperature increased with the  The H 2 -temperature-programmed reduction (TPR) of the calcined sample is presented in Figure 2f.For the sole Co and Fe samples, the first H 2 consumption peak around 250 • C was ascribed to the reduction of Mn 3+ to Mn 2+ .For the Co sample, three H 2 consumption peaks centered around 300, 336, and 423 • C were observed, which were assigned to the partial reduction of the Co 3 O 4 and CoMn oxide to Co 2+ , and the further reduction of the Co 2+ to Co 0 [27][28][29].For the Fe sample, the reduction peaks around 370 • C and 425 • C were attributed to the reduction of Fe 3 O 4 to FeO, and FeO to metallic Fe [30].The fourth peak centered at 550~620 • C was attributed to the reduction of CoAl and FeAl spinel [31,32].
For the CoFe samples, the reduction temperature increased with the increase in the Fe/Co molar ratio, implying that the addition of Fe inhibited the reduction process.
High Resolution Transmission Electron Microscopy (HRTEM) images of the spent catalysts are presented in Figure 3

Catalytic Performance
The CoFe catalysts with different Co/Fe ratios (5/1, 2/1, 1/1, and 1/2) were evaluated for the CO hydrogenation.As for comparison, the catalytic performance of the sole Co and Fe catalysts were also investigated.As shown in Table 2 and Figure 4a, the sole Co or Fe catalysts show a much lower CO conversion of 13.5% and 13.6%, respectively, at 260 • C.Although the Fe x C has been reported as the active phase of the Fe-based FTS reaction, the much lower reaction temperature led to the observed low activity, as most of the Fe-based reported catalysts operated at around 300~350 • C [33 -35].For the CoFe catalysts, the catalytic activity gradually increased with the increase in the Fe content.For example, the CO conversion of Co5Fe1 was 24.5%, and it was raised to 37.5% for Co2Fe1, which is nearly 2.8 times-higher than that of the sole Co and Fe catalysts.However, the CO conversion decreased to 33.8% and 25.0% for Co1Fe1 and Co1Fe2, respectively.The high water-gas-shift (WGS) reaction activity for both carbide catalysts [36,37] often leads to a high CO 2 selectivity (~50 C%), and it is believed that the syngas with a low H 2 /CO ratio is more suitable for this type of catalyst [38][39][40][41].Figure 4b shows the product distribution of various catalysts.The sole Co catalyst showed a slightly higher olefin selectivity (67.3 C%) and a lower CH 4 selectivity (2.9 C%).Although the olefin selectivity reached 70.1 C% for the sole Fe catalyst, the fraction of CH 4 was as high as 7.2 C%.The Co2Fe catalyst also showed a much higher olefin selectivity (66.1 C%) with the lowest CH 4 selectivity (4.3 C%) among various CoFe catalysts.Benefiting from both high CO conversion and olefin selectivity, the Co2Fe1 exhibited the highest olefins yield (24.8 C%) (Figure 4c).The higher oxygenates selectivity was also observed for these catalysts.For example, a 25% oxygenates selectivity was obtained over the sole Co catalyst, and the sum selectivity of olefins and oxygenates was even as high as 92.3 C%.As the Co/Fe molar ratio decreased from 5/1 to 1/2, the oxygenate selectivity showed downward trend from 18.5 C% to 7.9 C%.Table 3 and Figure 4d-f show the effect of the reaction conditions on the FTO performance over the Co2Fe1 catalyst.When raising the reaction temperature from 250 to 260 • C, the CO conversion increased from 22.5 to 37.5%, while keeping the product selectivity almost unchanged, with about 66 C% olefin selectivity and 4 C% methane selectivity, respectively.As the reaction pressure increased from 5 to 20 bar, the CO conversion was promoted from 17.8 to 34.8%.Increasing pressure is beneficial to hydrogenation reaction, so the olefin fraction decreased from 63.8 to 57.7 C%, and the CH 4 fraction increased from 3.6 to 5.8 C%.The H 2 /CO ratio also shows a great influence on the catalytic performance [42].When the H 2 /CO ratio was increased from 0.5 to 1, the CO conversion increased from 28.9% to 41.9%, while the olefin selectivity decreased from 63.8 to 44.4 C%.The detailed hydrocarbons distribution was also investigated, as shown in Figure 5.The olefins fractions dominated the products for all catalysts, which were as high as 89.7% and 78.0% for the sole Co and Fe catalyst, respectively.As the Co/Fe ratio decreased from Co5Fe1 to Co1Fe2, the olefin fractions also decreased from 80.7% to 67.7%.For the sole Co catalyst, the CH 4 selectivity was significantly lower than the value predicted by the classical ASF model, which is in accordance with the previous point that Co 2 C nanoprisms can effectively suppress CH 4 formation [14].However, the hydrocarbon products distribution of the sole Fe catalyst followed the typical ASF distribution.The chain growth probability (α) of the sole Co and Fe catalyst was calculated to be 0.66 and 0.71, respectively.With the addition of Fe, the hydrocarbon products distribution of CoFe catalysts gradually showed two different chain growth probabilities, which was mainly due to the existence of two different catalytic active sites.
effectively suppress CH4 formation [14].However, the hydrocarbon products distribution of the sole Fe catalyst followed the typical ASF distribution.The chain growth probability (α) of the sole Co and Fe catalyst was calculated to be 0.66 and 0.71, respectively.With the addition of Fe, the hydrocarbon products distribution of CoFe catalysts gradually showed two different chain growth probabilities, which was mainly due to the existence of two different catalytic active sites.Figure 6 shows the stability performance of the Co2Fe1 catalyst.An induction period was observed during the first 50 h.This was mainly because the FTO active phase of Co 2 C and Fe 2 C was derived by carbonizing CoO and FeO in the start-up period of the FTO reaction [14].After an initial activation period, the CO conversion reached 18%, and no obvious deactivation was found in the 160 h test.The CO 2 selectivity remained at about 40 C%, and the selectivity toward CH 4 and C 2-4 = kept at around 2.6 C% and 18 C%, respectively, while the selectivity of C 5 + and oxygenates remained at about 75 C%.
Catalysts 2023, 13, x FOR PEER REVIEW 9 of 14 Figure 6 shows the stability performance of the Co2Fe1 catalyst.An induction period was observed during the first 50 h.This was mainly because the FTO active phase of Co2C and Fe2C was derived by carbonizing CoO and FeO in the start-up period of the FTO reaction [14].After an initial activation period, the CO conversion reached 18%, and no obvious deactivation was found in the 160 h test.The CO2 selectivity remained at about 40 C% , and the selectivity toward CH4 and C2-4 = kept at around 2.6 C% and 18 C%, respectively, while the selectivity of C5 + and oxygenates remained at about 75 C%.

Structural-Performance Discussion
To explore the influence of Fe addition into Co on the CO dissociation rate, the reduced catalysts of Co, Fe, and Co2Fe were exposed to a CO atmosphere at 260 °C.By comparing the fitted peak areas of the CO2 signal recorded by MS in the effluent gas, the formation rate of surface carbonaceous species was accessed.As shown in Figure 7, under

Structural-Performance Discussion
To explore the influence of Fe addition into Co on the CO dissociation rate, the reduced catalysts of Co, Fe, and Co2Fe were exposed to a CO atmosphere at 260 • C. By comparing the fitted peak areas of the CO 2 signal recorded by MS in the effluent gas, the formation rate of surface carbonaceous species was accessed.As shown in Figure 7, under the same treatment conditions, the Co2Fe1 can effectively catalyze a disproportionation reaction of CO to form CO 2 and surface carbonaceous species, indicating the higher CO dissociation rate of Co2Fe1 than that of the sole Co or Fe sample.The concentration of surface carbonaceous species commonly determines the carburization rate and Co 2 C formation.Therefore, it is easily to understand that the Co2Fe1 catalyst was more easily carbonized to form the carbide phase.This phenomenon also explains the observed highest activity of the Co2Fe1 catalyst.Based on the above results, the synergistic effect of Fe and Co could be confirmed.The addition of Fe to the Co catalysts results in the formation of the CoFe composite oxide, as confirmed by the X-ray diffraction results.The formation of a composite structure inhibited the reduction of the CoFe catalyst, and the H2 consumption peaks shifted to a higher temperature with the increasing Fe/Co molar ratio.Previous studies suggested that the reduction process possesses a significant effect on the Co2C-based FTO performance [14].The Co-based composite oxide with low valence states contributed to the formation of Co2C nanoprisms, while the metallic Co would lead to the formation of Co2C nanospheres during the carburization, whose catalytic activity was much lower that of Co2C nanoprisms.XRD characterization indicated that the addition of Fe also restrains the growth of Co2C.Our previous work has indicated that the Co2C catalyst shows remarkable particle size effects on syngas conversion [43].However, a further increase in the Fe/Co molar ratio would weaken the synergetic effect and decrease the number of Co active sites, resulting in a decrease in CO conversion.The presence of Co2C nanoprisms also inhibited methane formation, with only a 4.3 C% CH4 selectivity obtained.No obvious deactivation was observed within 160 h of the stability test, suggesting promising practical applications.

Catalyst Synthesis
The Na-modified CoFe catalysts with different Co/Fe mole ratios were prepared by Based on the above results, the synergistic effect of Fe and Co could be confirmed.The addition of Fe to the Co catalysts results in the formation of the CoFe composite oxide, as confirmed by the X-ray diffraction results.The formation of a composite structure inhibited the reduction of the CoFe catalyst, and the H 2 consumption peaks shifted to a higher temperature with the increasing Fe/Co molar ratio.Previous studies suggested that the reduction process possesses a significant effect on the Co 2 C-based FTO performance [14].The Co-based composite oxide with low valence states contributed to the formation of Co 2 C nanoprisms, while the metallic Co would lead to the formation of Co 2 C nanospheres during the carburization, whose catalytic activity was much lower that of Co 2 C nanoprisms.XRD characterization indicated that the addition of Fe also restrains the growth of Co 2 C. Our previous work has indicated that the Co 2 C catalyst shows remarkable particle size effects on syngas conversion [43].However, a further increase in the Fe/Co molar ratio would weaken the synergetic effect and decrease the number of Co active sites, resulting in a decrease in CO conversion.The presence of Co 2 C nanoprisms also inhibited methane formation, with only a 4.3 C% CH 4 selectivity obtained.No obvious deactivation was observed within 160 h of the stability test, suggesting promising practical applications.
CO conversion (X CO ) and products' selectivity (S CO2 , S i ) were calculated according to the following formula: in which, CO inlet and CO outlet mean the CO moles in the feed gas and outlet gas, respectively, CO 2 represents the CO 2 moles in the outlet gas, N i represents the moles of product i in the outlet gas, and n i is the carbon number of product i.The chain growth probability (α) of hydrocarbons was calculated according to the Ander-Schulz-Flory (ASF) model: where n is the carbon number of products.W n is the weight fraction of hydrocarbons (including paraffins and olefins) products with carbon atoms of n.

Conclusions
In summary, we present a Na-modified CoFe catalyst that facilitates the production of olefin through direct CO hydrogenation.It was observed that an appropriate Co/Fe ratio improved the catalyst performance.The addition of Fe restrained the growth of Co 2 C nanoparticles and the synergistic effect of Co and Fe also promoted the CO dissociation rate and carburization rate, thereby enhancing the catalytic activity.However, with the increase in Fe content, the number of Co 2 C active sites decreased, resulting in the deterioration of catalytic performance.For the Co2Fe1 catalyst, the desired olefins selectivity reached 66.1 C% at a CO conversion of 37.5%.The formation of methane was significantly suppressed, with only 4.3% undesired CH 4 selectivity.The CoFe catalyst also demonstrated good catalytic stability during a 160 h stability test.
on a calcined sample would convert to MnCO 3 with sharp diffraction peaks during the reaction.For samples containing both Co and Fe, the diffraction peaks of CoO and FeO disappeared due to the carburizing reaction, while diffraction peaks located at 37.0, 34.4,42.6, and 54.6, corresponding to Co 2 C (JCPDS no.65-1457), were observed.The crystallite size of Co 2 C calculated by Scherer's formula based on Co 2 C (111) was 19.4,16.3, 15.9, 10.1, and 9.0 nm for Co, Co5Fe1, Co2Fe1, Co1Fe1, and Co1Fe2, respectively, indicating the addition of Fe could restrain the growth of Co 2 C.As for the spent Fe catalyst, the peaks of the Fe 2 C (JCPDS no.17-0897) at 2θ of 41.3, 42.5, and 43.0 and FeO (JCPDS no.74-1886) were found.These results suggested that there was incomplete carbonization in the Fe sample.
. After the reaction, the CoO and FeO species were carbonized into Co 2 C and Fe 2 C with the aid of a Na promoter.For the Co sample, plenty of Co 2 C nanoprisms were observed, while the Fe sample showed the presence of Fe 2 C nanoparticles.For the CoFe samples, both Fe 2 C nanoparticles and Co 2 C nanoprisms were observed, which is in line with the results of XRD characterization.

Figure 5 .
Figure 5. Hydrocarbon products distribution and the corresponding chain growth probability (α) of various catalysts.The blue dots in (a-f) denotes the value of Ln(W n /n), where the W n is defined as the weight fraction of hydrocarbon product with carbon number of n.The pink dash line denotes the fitted trend.The carbon chains probability (α) is obtained by fitting the results using the ASF model.Reaction conditions: H 2 /CO = 0.5, 260 • C, 10 bar, 2000 mL/g/h.

Figure 5 .
Figure 5. Hydrocarbon products distribution and the corresponding chain growth probability (α) of various catalysts.The blue dots in (a-f) denotes the value of Ln(Wn/n), where the Wn is defined as the weight fraction of hydrocarbon product with carbon number of n.The pink dash line denotes the fitted trend.The carbon chains probability (α) is obtained by fitting the results using the ASF model.Reaction conditions: H2/CO = 0.5, 260 °C, 10 bar, 2000 mL/g/h.

Figure 7 .
Figure 7. MS signal of CO2 for the disproportionation reaction of CO to form CO2 and surface carbonaceous species.Experiments were conducted at 260 °C.

Figure 7 .
Figure 7. MS signal of CO 2 for the disproportionation reaction of CO to form CO 2 and surface carbonaceous species.Experiments were conducted at 260 • C.

Table 1 .
Composition and textural properties of the various catalysts.

Co a (wt.%) Fe a (wt.%) Na a (wt.%) Co/Fe b (Co + Fe)/Mn/Al b S BET (m 2 /g) Pore Volume (cm 3 /g) Pore Size (nm)
a The element composition was determined by XRF.bThe molar ratio of the element composition.

Table 2 .
Catalytic performance over various catalysts for the FTO reaction a .
bOxy.(aldehydes and alcohols).cThefraction of paraffin with two or more carbon numbers in all products.

Table 3 .
The effect of reaction conditions on FTO performance over the Co2Fe1 catalyst a .

Table 2 .
Catalytic performance over various catalysts for the FTO reaction a .