High Selectivity to Aromatics by a Mg and Na Co-modified Catalyst in Direct Conversion of Syngas.

The demand for aromatics, especially benzene, toluene, and xylene, has been increased in recent years as the crucial feedstocks of coatings and pharmaceutical industry. In this work, a modified Fischer-Tropsch synthesis (FTS) catalyst FeNaMg was fabricated via a sol-precipitation method and integrated with an HZSM-5 aromatization catalyst for the aromatics synthesis from syngas by a one-step process. Syngas was first converted to lower olefins as intermediates on the active component of the FeNaMg catalyst followed by aromatization on zeolite. Different characterization approaches, such as BET, XRD, XPS, hydrogen temperature-programmed reduction, temperature-programmed desorption of CO, TG, and SEM, revealed that Mg efficiently optimized physicochemical properties of the Fe-based catalyst by generating a MgFe2O4 spinel structure. Further investigation demonstrated that the MgFe2O4 spinel structure could increase the syngas adsorption area, facilitating the reduction and carburization of the Fe phase, while Mg decreased CO2 selectivity (31.26 to21%) by restraining the water-gas shift reaction and improved the utilization efficiency of carbon. At the same time, alkali metal Na changed the surface electronic environment of the FTS catalyst to enhance CO adsorption as an electronic promoter, which suppressed methane formation by restraining over hydrogenation. Therefore, the synergism that existed between Mg and Na during the reaction escalated the CO conversion and aromatics selectivity to 96.19 and 51.38%, respectively.

Introduction

Aromatics <span class="Chemical">are a key feedstock in the chemical industry for the synthesis of a wide range of products, including synthetic materials, medicine, and so forth. Aromatics are mainly derived from the catalytic reforming of naphtha in ad traditional petroleum refining process.[1] A complementary path of aromatic production is required with the rapid economic development and the increasing demand for aromatics. Syngas (CO and H2) extracted from coal, biomass, and shale gas has been developed as a promising substitute for petroleum in recent years.[2,3]

According to the amount of catalyst bed, syngas to aromatics (STA) reactions have been categorized into two-step and one-step processes, respectively. Both of the processes produce aromatics from coal-derived syngas, which could realize full utilization of abundant coal resources. The two-step process following the light olefins route or methanol route is a mature process nowadays where the olefins and methanol are produced as intermediates in the first reactor and then converted to aromatics on zeolite in the second reactor. The one-step process requires only one reactor, whereas the tandem reactions in the two-step process go on simultaneously. Consequently, the one-step process could improve the conversion and target product selectivity by consuming the intermediate products (light olefins and methanol) from the first reaction, which simplifies the reaction procedure. The STA one-step process has a great potential application because it is low cost and energy saving.[4−6]

A variety of catalysts have been designed and developed for producing <span class="Chemical">aromatic hydrocarbons from syngas. Chang et al.[7] first combined Fe oxides and ZSM-5 zeolite as the STA catalyst and obtained 25.2% aromatics selectivity. Recently, the study from Ma et al.[8] integrated Na–Zn–Fe5C2 and hierarchical HZSM-5 with uniform mesopores, attaining a 51% aromatics selectivity with over 85% CO conversion. Further, they found that it was necessary to control an appropriate Bro̷nsted acid strength and density of zeolite to reach a better aromatics yield. Liu et al.[9] developed a bifunctional catalyst composed of K–FeMnO and MoNi–ZSM-5 zeolite for direct aromatics synthesis from syngas and found the addition of potassium not only promoted the CO adsorption performance but also increased the stability of the catalyst.

Although considerable works have been conducted on converting syngas to aromatics by t<span class="Chemical">he one-step process in recent years, many problems still remain in this field, such as low aromatics selectivity, high byproduct selectivity (CO2 and CH4), and shorter catalyst lifetime, leading to the relevant studies still being in the laboratory exploration stage. Therefore, improving the selectivity to the target product (benzene, toluene, and xylene, BTX), decreasing byproduct selectivity, and prolonging the catalyst lifetime are the most attractive and difficult challenges for scientific researchers. The metals Mg and Na are two important elements in STA catalysts. The former could enhance the reduction and carburization of Fe as a good promoter, and it has not been used in catalysts for the one-step synthesis of aromatics from syngas; the latter has a good electron donating characteristic, which could strengthen adsorption and dissociation of CO and subsequent C–C coupling process.[10−12] In this study, we designed the co-modified catalysts with Mg and Na for the use in direct conversion of syngas to aromatics for the first time. The cooperation of Mg and Na on an Fe-based catalyst facilitated the Fischer–Tropsch synthesis (FTS) to produce more lower olefins (intermediate product), which underwent aromatization (isomerization, oligomerization, cyclization, and H transfer) on HZSM-5 zeolite; therefore, the selectivity of aromatics (BTX) was enhanced.[8,13]

The present work studied t<span class="Chemical">he catalytic effect of Mg and Na on the STA reaction, especially focusing on the catalyst textural properties, activity, stability, and selectivity of aromatics. The experimental results were optimized by adjusting the contents of Mg and Na in the catalyst. A distinct improvement of CO conversion (92.13–96.19%) and aromatics selectivity (29.63–51.38%) was obtained after using the modified FeNaMg catalysts; meanwhile, the byproduct formation (CO2 and CH4) was restrained. The successful design of the FeNaMg catalyst provides a new route for metal-modified Fe-based catalysts, which makes the industrialization of the one-step synthesis from syngas to aromatics possible.

Experimental Section

Catalyst Preparation

The Fisc<span class="Chemical">her–Tropsch synthesis catalyst, FeNaMg, was synthesized by the sol-precipitation (SP) method. A certain amount of Fe(NO3)3 and Mg(NO3)2 with different molar ratios was added into 200 mL of water to make a solution and continuously stirred at 60 °C. NaOH (1.5 mol/L) was dropped slowly into the prepared solution as a precipitant, and the end pH value was maintained at 10. The obtained suspension was aged for 2 h followed by filtration and washing with deionized water for three times to maintain the concentration as 0.89 (Na/Fe molar ratio) for the Na ion. The precipitates were dried in the oven for 10 h at 120 °C and then calcined at 450 °C for 4 h. The resultant catalysts were denoted as FeNaMgx (number x represents the molar content of Mg in the Fe-based catalyst, x = 0, 10, 15, 20, 25, and 30). The optimal Fe/Mg molar ratio for the FeNaMg catalyst in this work is observed as 80:20 denoted as FeMg20. The catalysts without Na were also prepared for the comparison by the same method.

The re<span class="Chemical">ference catalyst FeNaMg20 prepared by the incipient wetness impregnation method in which an amount of Mg ion was impregnated on the surface of the FeNa catalyst was denoted as FeNaMg20 (IWI).[8] FeNaMg20 synthesized by the coprecipitation method was denoted as FeNaMg20 (CP).[11]

The commercial H<span class="Chemical">ZSM-5 zeolites (SiO2/Al2O3 = 100) from CNOOC Tianjin Chemical Research and Design Institute Co., Ltd. was pressed, crushed, and sieved into the particles with a size range of 20–40 mesh for use. Our study on metal-modified HZSM-5 showed that Ni could enhance the stability and improve the selectivity of aromatics, thus Ni-HZSM-5 was chosen to be the aromatization catalyst. Ni/HZSM-5 zeolites were prepared by incipient wetness impregnation using a Ni(NO3)2 precursor and then calcined at 550 °C for 4 h.

After the prep<span class="Chemical">aration, each catalyst was tested by inductively coupled plasma (ICP) to confirm its compositions; the data were similar to their feed ratios.

Catalyst Characterization

The Brunner–Emmet–Teller (<span class="Gene">BET) surface area, pore volume, and pore diameters of catalysts were measured by N2 physical adsorption–desorption at −196 °C using the micromeritics ASAP 2020 system.

X-ray diffraction (XRD) patterns of the prep<span class="Chemical">ared catalyst were obtained from an Ultima IV X-ray powder diffractometer using Cu Kα radiation under 30 kV and a 15 mA operating condition to characterize the phases in the structure of catalysts and determine the crystallinity of the catalysts.

X-ray photoelectron spectroscopy (XPS) was performed on a Thermal XPS ESCALAB 250Xi spectrometer to investigate t<span class="Chemical">he chemical state of the prepared catalysts. The binding energy (BE) used a C 1s peak at 284.8 eV.

Hydrogen temperature-programmed reduction (<span class="Chemical">H2-TPR) was studied using a Micromerities Autochem II 2920 chemisorption machine. Approximately 0.1 g catalyst sample was loaded in a U-tube and pretreated at 773 K under a He flow for 1 h. Then, the sample was reduced in a flow of 10% H2/90% Ar (50 mL/min). The alteration of reducing the gas concentration was tested through a thermal conductivity detector.

Temperature-programmed desorption of CO (CO-TPD) was performed on an Auto C<span class="Chemical">hem II equipment (Micromeritics, USA). The catalyst (∼0.10 g) loaded in a U-shaped quartz tube was reduced at 400 °C under a H2/Ar (10% H2 and 90% Ar, 50 mL min–1) flow. Pure CO was injected when the temperature was cooled down to 30 °C, then the catalyst was then purged with the carrier gas (20 mL min–1) to remove the gas, and the TPD curve was obtained via heating the catalyst from 30 to 800 °C at a heating rate of 10 °C min–1. The amount of the desorbed adsorbate was monitored by TCD.

Scanning electron microscopy (SEM) was performed to image the surfaces of catalysts using a Hitachi S-4700 (II) electron microscope (Tokyo, Japan) with an accelerating voltage of 15 kV.

Reactor System and Operating Procedures

The catalytic reaction was implemented in a <span class="Chemical">stainless tubular fixed-bed reactor (MRT-6113 microtube catalyst evaluation device) with an inner diameter of 12 mm and an effective bed length of approximately 12 cm. The catalysts composed of accurate 2.0 g of FTS catalysts and 2.0 g of HZSM-5 were loaded in the reactor after a full mix. Prior to the STA reaction, the catalysts were activated in a flow of H2/N2 (50/50 vol) at 400 °C for 4 h wherein the temperature was increased at a rate of 3 °C/min from room temperature to 270 °C and 2 °C/min from 270 to 400 °C. A catalytic test was performed in a reactor containing H2/CO/N2 with a volume ratio of 8/4/1 under 370 °C and 4 MPa. The ratio of H2 to CO was kept consistent with the molar ratio of molecules forming light olefins. N2 was the internal standard gas.

For noncondensable gases, H2, CO, <span class="Chemical">CH4, and CO2 were separated with a 13× molecular sieve packed column and were analyzed by a thermal conductivity detector and flame ionization detector (Agilent 7890B gas chromatographic analyzer). The condensed liquid phase included oil and water. The oil phase collected was analyzed on an off-line gas chromatograph (PE Clarus 580). The CO conversion and product selectivity were calculated as

where XCO is t<span class="Chemical">he CO conversion, Si is the selectivity of component i, I is the carbon number in component i, and n is the molar of the component.

The balances of C, H, and O in experiments <span class="Chemical">are all within the range of 95–105%.

Results and Discussion

Characterizations of Catalyst Samples

Since Mg and Na may have a huge stimulation on t<span class="Chemical">he Fe-based catalyst, the following characterizations were proceeded to explore the detailed promotional effects of these two additives.

N2 adsorption–desorption was performed in order to under<span class="Chemical">stand the effects of Mg on textural properties of catalysts, and the results are shown in Figure and Table S1. The isotherms were exhibited differently after Mg addition. The shape of the hysteresis loop changed from type H1 (Fe, FeNa) to type H2 (FeMg20, FeNaMg20),[14,15] and the BET surface area got a remarkable improvement with Mg decoration. It was suspected that the MgFe2O4 spinel phase was generated during the catalyst preparation process, which altered the stacking form of the catalyst crystal and increased the surface area, according to many previous papers.[16−19] The result could also be evidenced from SEM images (Figure S1) that the catalyst grains became smaller from probably 40 to 30 nm after adding Mg. The BET surface area increased with the proportion of Mg from 10 (FeNaMg10) to 20% (FeNaMg20) and then displayed a downward tendency, which may be due to excessive Mg addition, partially covering the Fe surface.[11]

Figure 1

N2 adsorption/desorption isotherms of (a) Fe-based catalysts and (b) FeNaMg catalysts with different Mg contents.

N2 adsorption/desorption isot<span class="Chemical">herms of (a) Fe-based catalysts and (b) FeNaMg catalysts with different Mg contents.

The <span class="Chemical">N2 adsorption–desorption behavior of the FeNaMg/Ni-HZSM-5 composite catalyst was shown in Figure S2. The N2 adsorption–desorption isotherm of the mixed catalyst had no distinct change compared with Ni-HZSM-5 zeolite, indicating that the FTS catalyst did not destroy the structure of zeolite.[20]

XRD and XPS analyses were conducted to investigate the phase and c<span class="Chemical">hemical state of catalysts before and after the reaction. The XRD pattern of the fresh catalyst was shown in Figure a. The characteristic diffraction peaks at 2θ values of 24.3, 33.3, 35.8, 40.9, 49.6, 54.2, 57.6, 62.6, and 64.1° were assigned to the standard characteristic peaks of α-Fe2O3 (JCPDS card no. 33-0664).[21] Na peaks were not detected in the FeNa catalyst, which meant that the small amount of Na had been evenly embedded in the catalyst’s crystal.[8] Mg-doped α-Fe2O3 with different Mg/Fe ratios presented similar XRD patterns with the FeNa catalyst, indicating that the small amount of Mg did not impact the crystal structure of α-Fe2O3. However, the intensity of α-Fe2O3 peaks declined with the further increase of Mg; meanwhile, a novel diffraction peak of MgFe2O4 was observed at 2θ 29.3°, revealing the promotional effect of Mg to Fe dispersion and MgFe2O4 spinel generation in a higher Mg content, which verified the conjecture in BET characterization.[11] No obvious characteristic peaks of MgO could be identified in X-ray diffraction patterns due to the high dispersion of elements in the synthesized catalysts.[11] These speculations could be further proved by XPS characterization. The whole XPS survey spectra in Figure S3 suggested that the FeNaMg20 catalyst surface was composed of Fe, C, O, Na, and Mg, which proved that Na and Mg had been successfully introduced into the catalyst. As demonstrated in Figure b, the Fe 2p XPS spectrum of the fresh FeNaMg20 catalyst presented two strong peaks at about 710.2 and 724.7 eV, which can be split into Fe2O3, Fe3O4, and MgFe2O4 peaks, respectively.[22,23] Meanwhile, the peak intensity of Fe2O3 was stronger than Fe3O4, indicating that Fe2O3 was the major component in the fresh catalyst.

Figure 2

(a) XRD patterns of the fresh catalysts. (b) Fe 2p and (c) Mg 1s XPS spectrum of the fresh FeNaMg20 catalyst.

(a) XRD patterns of the fresh catalysts. (b) <span class="Chemical">Fe 2p and (c) Mg 1s XPS spectrum of the fresh FeNaMg20 catalyst.

Figure shows the phase and c<span class="Chemical">hemical state of catalysts after the reaction. As demonstrated by XRD results of spent catalysts in Figure b, catalysts were constituted by magnetite (Fe3O4, JCPDS 19–0629) and Hägg carbide (χ-Fe5C2, JCPDS51-0997). Some of Fe2O3 was converted to the χ-Fe5C2 phase during the reaction shown at 42–45°, which has been proposed as the active phase for CO hydrogenation.[24−26] The χ-Fe5C2 peak intensity of FeNa and FeMg increased compared with the Fe catalyst, indicating that Na and Mg promoted the formation of iron carbide.[27] It can be also observed that the χ-Fe5C2 peak intensity became stronger with the further increase of Mg; on the contrary, the peak intensity of Fe3O4 decreased gradually. The XRD patterns of the spent catalyst showed the synergistic effect of Mg and Na on the process of carburization. The appearance of the Fe5C2 characteristic peak in Fe 2p XPS spectra (Figure b) proved the existence of Fe5C2 in used catalysts, confirming the results in Figure a. The peaks of both Fe 2p1/2 and Fe 2p3/2 could be divided into the following four peaks: Fe2O3, Fe3O4, Fe5C2, and MgFe2O4 where the main phase was Fe3O4 with the largest peak area. It can be also found that compared with the fresh catalyst in Figure b, the peak position of Fe 2p in the used catalyst moved a little to the lower binding energy. The possible explanation was that the electron of Mg moved to the Fe surface during the reaction, as a result, the peak position of Mg 1s shifted to the higher binding energy, as shown in Figures c and 3c.[28] Therefore, electron transfer between Mg and Fe may contribute to the carburization of Fe.

Figure 3

(a) XRD patterns of the used catalysts. (b) Fe 2p and (c) Mg 1s XPS spectrum of the used FeNaMg20 catalyst.

(a) XRD patterns of the used catalysts. (b) <span class="Chemical">Fe 2p and (c) Mg 1s XPS spectrum of the used FeNaMg20 catalyst.

The O 1s XPS profiles of <span class="Chemical">FeNa and FeNaMg20 catalysts are exhibited in Figure to show the effect of Mg on oxygen vacancy. The whole area in two graphs consisted of three peaks located at 532.6, 531.2, and 529.3 eV, which corresponded to the characteristic peaks of hydroxyl oxygen, oxygen vacancy, and lattice oxygen respectively.[29] Evidently, the intensity of the oxygen vacancy peak after decorating with Mg was much higher than that in the FeNa catalyst. The alteration of the lattice oxygen peak was opposite to oxygen vacancy that its peak intensity decreased after Mg introduction. Mg addition can promote the dissociation of lattice oxygens from the catalyst crystal during the reduction process, and oxygen vacancies formed subsequently at the position of dissociated lattice oxygen. The differences between two images showed that Mg promoted the formation of oxygen vacancies, which could enhance the adsorption and dissociation of CO2 according to the study by Sun et al.[30] who found that an O atom was extracted from the adsorbed CO2 to fill in oxygen vacancies and generated a CO, which could take part in the reaction continuously. It may explain why the CO2 selectivity declined after addition of Mg. Meanwhile, Na being the second promoter had no significant effect on the formation of oxygen vacancy, as shown in Figure S4.

Figure 4

XPS spectra of (a) FeNa and (b) FeNaMg20 catalysts before the reaction.

XPS spectra of (a) FeNa and (b) <span class="Chemical">FeNaMg20 catalysts before the reaction.

H2-TPR was performed to furt<span class="Chemical">her confirm the reduction effect of Mg on Fe, and the reduction profiles are presented in Figure . The Fe curve showed that there were three reduction processes that occurred when the temperature was increased; the peak at 380, 550, and 670 °C corresponded to the reduction from Fe2O3 to Fe3O4, Fe3O4 to FeO, and FeO to Fe, respectively.[13] It can be noticed that the addition of Na had no effect on the reduction process because there was nearly no deviation of FeNa peaks compared with the Fe curve.[31] When the FeNa catalyst was further modified with Mg, all three reduction peaks obviously shifted to a lower temperature. The profile clearly showed that Mg could facilitate the reduction process of Fe, which was beneficial to the subsequent carburization process.

Figure 5

H2-TPR profiles of the Fe, FeNa, and FeNaMg20 catalysts.

H2-TPR profiles of t<span class="Chemical">he Fe, FeNa, and FeNaMg20 catalysts.

CO-TPD was carried out to explore t<span class="Chemical">he effect of Na and Mg on CO adsorption (Figure ). Each curve contained roughly two peaks, which were located at a low temperature zone of 30–130 °C and high temperature zone of 430–800 °C , which represent the desorption of CO adsorbed in the form of a molecular state and the desorption of dissociated and adsorbed CO, respectively.[32] Meanwhile, the two desorption peaks of the original Fe catalyst were located at ∼70 and ∼570 °C. It was observed that the peak intensity in the low temperature region was reduced by doping Na, and the position of the high temperature desorption peak shifted to a higher temperature (∼720 °C). Since the reaction temperature is 370 °C, the effect of Na and Mg on CO adsorption in the high temperature zone is mainly discussed here. The distinct right shift of the desorption peak in the high temperature zone showed that Na promoted the dissociation adsorption of CO.[24,33] The promotion of Na was due to the formation of a highly polarized dipole layer on the surface of alkali metal ions, enhancing the electronic feedback to CO and weakening the C–O bond.[34] When the FeNa catalyst was further decorated with 20% Mg, the desorption peak shifted to a lower temperature, indicating that Mg did not promote the adsorption of CO, while the peak area increased, possibly due to the increase of the BET surface area of the Mg-decorated catalyst (Table S1). The co-modification of Na and Mg on the Fe-based catalyst promoted the adsorption of CO since Na made catalyst easier to adsorb CO, while Mg increased CO adsorption sites. When the amount of Mg was increased to 30% (i.e., the catalyst FeNaMg30), the intensity of the two peaks decreased, indicating that excessive Mg inhibited the dissociative adsorption of CO.

Figure 6

CO-TPD profiles of different FTS catalysts.

CO-TPD profiles of different <span class="Chemical">FTS catalysts.

From the <span class="Chemical">H2-TPR and CO-TPD profiles, the synergy effect of Mg and Na can be speculated in the aromatics synthesis reaction. Since Mg enhanced the H2 adsorption of the catalyst while Na was beneficial to the adsorption of CO, these two promoters achieved the adsorption equilibrium of H2 and CO, thereby promoting the formation of olefinic intermediates together.

TG profiles of the catalysts were shown in Figure a. It could be observed that t<span class="Chemical">he catalyst lost weight quickly due to the removal of absorbed water and organics during 50–220 °C.[35] Due to the oxidation of Fe5C2 and Fe3O4 to Fe2O3, the catalyst weight got an increase between 260 and 320 °C. It can be found that FeNaMg20/Ni–HZSM-5 had a larger weight increment, implying a larger portion of Fe5C2 was generated under Mg promotion, which was consistent with the result in Figure b. At the same time, it could be seen that the weight-increase peak moved to a higher temperature after Mg decoration, indicating that the carbide status was getting more stable and difficult to be oxidized. The weight presented a downward tendency after 350 °C, which was mainly caused by the combustion of carbon deposits. The amount of carbon deposit decreased with the introduction of Mg, which was beneficial to improve the stability. As shown in Figure b, the CO conversion and aromatics selectivity of the FeNa/Ni–HZSM-5 catalyst began to decrease after 60 h, indicating that it started to deactivate. Meanwhile, the activity of the FeNaMg20/Ni–HZSM-5 catalyst displayed an obvious decrease after 72 h, which exhibited a better stability than FeNa/Ni–HZSM-5.

Figure 7

(a) Thermogravimetry of the catalyst. (b) Stability of the catalyst.

(a) Thermogravimetry of t<span class="Chemical">he catalyst. (b) Stability of the catalyst.

Catalytic Behavior of the Synthesized Samples

The addition of <span class="Chemical">Mg and Na optimizes the structure and physicochemical properties of catalysts and also improves the catalytic performance, which are evidenced from the STA reactions. The STA reaction was evaluated at 370 °C and 4.0 MPa to test the catalytic performance. Different contents of Na and Mg were investigated to assess their promoting roles. The experiments of each catalyst formulation were repeated under the same reaction condition for at least two times, and the results adopted the average value. Figure illustrates the CO conversion and product selectivity in the STA reaction for 24 h. The CO conversion by the catalysts with the addition of Na was all significantly higher than the conversion without Na in catalysts (92.13%), suggesting that Na improves the CO conversion. This is probably due to the electron donation by Na, which promotes CO dissociation and C–C coupling (Figure ). The CO conversion also increased with the addition of Mg, although it has been reported that Mg did not promote the dissociation of CO directly, according to the study from Qiao et al.[36] This phenomenon maybe attributed to the more oxygen vacancies generated by Mg, which weakened the C–O bond and promoted the dissociation of CO, as shown in Figure .[29] Another possible reason could be related to the increased BET surface area of the Mg-modified Fe-based catalyst, providing more active sites to adsorb and activate CO. CO conversion increased with the Mg content and peaked at 20% (i.e., FeNaMg20). However, a further increase in the Mg content had an adverse effect on CO conversion because the excessive Mg promoter blocked the pore channels and covered active sites.

Figure 8

(a) CO conversion and product selectivity of catalyst samples. (b) BTX selectivity. Z5 refers to HZSM-5.

(a) CO conversion and product selectivity of catalyst samples. (b) BTX selectivity. Z5 re<span class="Chemical">fers to HZSM-5.

The productivity selectivity is anot<span class="Chemical">her index of importance for the catalytic performance. Na decoration suppressed the CH4 selectivity by inhibiting hydrogenation. CO2 selectivity showed a declining tendency, which is similar to the cited literature, and the product stream from FeNaMg20/Ni–HZSM-5 contained the least amount of carbon dioxide.[32] On one hand, oxygen vacancies played a role in decreasing CO2 selectivity via adsorbing and dissociating CO2.[30] On the other hand, Mg decoration effectively inhibited the water–gas shift reaction.[37]

All aromatics consisted an <span class="Chemical">oil phase, and the total selectivity got an obvious improvement from 29.63in the Fe catalyst to 51.38% in the catalyst with 20% Mg, which may be attributed to the promotion of Mg to Fe reduction and χ-Fe5C2 generation, concluding that Mg addition indirectly improved aromatics selectivity by promoting the formation of intermediate products. At the same time, the catalysts decorated by Mg presented high selectivity to BTX as shown in Figure b. It should be mentioned that the best results did not appear at the 30% content of Mg, which had the strongest density peak of Fe5C2 (Figure a). The possible reason is that excessive Mg blocked the catalyst channels and coverd the adsorption sites of CO on Fe; the BET surface area and pore volume of FeNaMg30 decreased in comparsion with FeNaMg20, as shown in Table S1.[38]

Effect of the Catalyst Preparation Method and Integration Manner

To find an efficient catalyst system, the catalyst prep<span class="Chemical">aration method and integration manner were tested. The co-precipitation (CP) method and incipient wetness impregnation (IWI) method were widely used to prepare FTS catalysts in many papers,[17,27] thus the investigation of these two methods is aimed to compare them with sol-precipitation and find out the most appropriate method for the FeNaMg catalyst system; the content of Mg in these three methods is consistent. The CO conversion and product selectivity of catalysts prepared with different methods were illustrated in Figure . It is clear that the SP method presented a better catalytic effect than CP and IWI methods with a higher CO conversion and lower byproduct (CO2 and CH4) selectivity. It could be seen that preparation methods affected the CO2 selectivity obviously, and Mg was only attached on Fe in the catalyst prepared by IWI, thus the amount of oxygen vacancies was less than that from SP. Moreover, aromatics selectivity, especially high-value aromatics (BTX) in the product stream, was much higher than that of the other two catalysts, as shown in Figure b. The probable reason is that the SP method formed the smaller Fe grains with a larger surface area and pore volume, as shown in Table S2. Therefore, the SP method is more suitable to the prepare Fe-based catalyst in the FeMg catalyst system than the other two methods.

Figure 9

(a) CO conversion and product selectivity of catalysts prepared with SP, IWI, and CP methods. (b) BTX selectivity of catalysts synthesized by different methods. Reaction conditions: 370 °C, 4.0 MPa, GHSV = 1800 h–1, 24 h, FeNaMg20/Ni–HZSM-5 mass ratio = 1.

(a) CO conversion and product selectivity of catalysts prepared with SP, IWI, and CP methods. (b) <span class="Chemical">BTX selectivity of catalysts synthesized by different methods. Reaction conditions: 370 °C, 4.0 MPa, GHSV = 1800 h–1, 24 h, FeNaMg20/Ni–HZSM-5 mass ratio = 1.

The prep<span class="Chemical">ared FeNaMg20 catalyst needs to be compounded with the Ni–HZSM-5 catalyst. Hence, the integration manner and proximity of the two components will also affect the catalytic performance. In this study, we compared the catalyst activities with different integration manners (dual bed, granule stacking, mortar mixing, and ball milling), and results are shown in Figure . In a consecutive dual-bed reactor, catalysts show a lower CO conversion (92.74%) and aromatics selectivity (32.02%) due to the lack of a thermodynamic driving force with a higher selectivity of byproducts (CO2 and CH4), resulting in low availability of carbon. The granule-stacking catalyst (20–40 mesh) presented a higher CO conversion rate (96.19%), higher aromatics selectivity (51.38%), and lower CO2 amount compared to the dual-bed catalysts, suggesting that the shorter distance was beneficial to the transmission of lower olefins (C2–C4) to the pores of zeolite, which promoted the occurrence of the aromatization reaction. However, a further decrease in the distance between the Fe-based catalyst and zeolite was detrimental to the reaction as shown in samples mixed by mortar mixing (80–100 mesh) and ball milling (<200 mesh). A reasonable explanation is that the Na on the Fe-based catalyst might have migrated to zeolite and affected its surface acidity, which restrained the subsequent aromatization process.[39,40] Interestingly, this phenomenon is distinct to the traditional thinking that the closer the catalyst particles are, the better the reaction performance. The experimental results showed that the suitable integration manner and distance between two components played an important role in the STA reaction.[41,42]

Figure 10

Performance of catalysts from different integration manners. Reaction conditions: 370 °C, 4.0 MPa, GHSV = 1800 h–1, 24 h.

Conclusions

In conclusion, direct synthesis of <span class="Chemical">aromatics from syngas has been demonstrated over a bifunctional catalyst composed of FeNaMg and Ni–HZSM-5. The modified catalyst with Mg and Na presented a better catalytic performance by improving the structure and surface electron characteristic of the Fe-based catalyst. The CO conversion and aromatics selectivity were significantly improved compared with the parent Fe catalyst, reaching 96.19 and 51.37%, respectively. The addition of Na could promote the adsorption of CO and C–C coupling as the electron-donating promoter. Under an appropriate Mg content, the BET surface area increased obviously with the formation of MgFe2O4, providing an enhanced adsorption area for syngas, while the reduction behavior of Fe oxides was also strengthened, which was beneficial to the carburization process by the electron transfer between Fe and Mg. The synergism between Na and Mg realized the adsorption equilibrium of CO and H2, facilitating the Fe5C2 formation, which promoted the synthesis of olefinic intermediates and restrained the generation of byproducts (CH4 and CO2). At the same time, the suitable catalyst preparation manner and proximity between two components also played a significant role in the STA reaction. These findings are essential for the future development of other metal-modified STA catalyst systems.