Catalytic Activity of Thermolyzed [Co(NH3)6][Fe(CN)6] in CO Hydrogenation Reaction

Currently, the processes of obtaining synthetic liquid hydrocarbons and oxygenates are very relevant. Fischer-Tropsch synthesis (FTS) is the most important step in these processes. The products of thermal destruction in argon of the mixture [Co(NH3)6][Fe(CN)6] and Al(OH)3 were used as catalysts for CO hydrogenation. The resulting compositions were studied using powder X-ray diffraction, IR spectroscopy, elemental analysis, SEM micrographs. The specific surface area, pore and particle size distributions were determined. It was determined that the DCS-based catalysts were active in the high-temperature Fischer-Tropsch synthesis. The effect of aluminum in the catalyst composition on the distribution of reaction products was revealed.


Introduction
Fischer-Tropsch synthesis (FTS) is catalyzed by Group VIII metals (iron, cobalt, nickel). In the presence of cobalt-based catalysts, yields of products and formation of long-chain alkanes are strongly dependent on the nature of the catalyst support [1]. The yield of olefins' and oxygenates' high activity in the water shift reaction are significant on iron catalysts [2]. To obtain systems that have the advantages of both metals, catalysts with bimetallic ironcobalt containing particles are created. In [3] the effect of the nature of the substrate (oxide and carbon carriers) on the composition of the products obtained is shown. In [2] the formation of an alloy of metals was found, which enhances the ability to incorporate CO, which leads to an increase in the selectivity for alcohols. It was found [4], that the supported Fe-Co bimetallic catalysts showed higher activity in comparison with catalysts from pure metals. In [5] it was found, that joint impregnation leads to the formation of mixed Fe-Co oxides, the reduction of which leads to the formation of alloys. It was also found that the introduction of an increase in the cobalt content in the system accelerates the hydrogenation of carbon monoxide, and the formation of mixed particles of cobalt and iron promotes the formation of light olefins in comparison with the monometallic cobalt catalyst, and a decrease in activity during the Fischer-Tropsch synthesis. In [6], it is shown that the addition of PdTe and PdTe 1−x Bi x to the standard cobalt catalyst increases both yield and selectivity with respect to the liquid phase and is the shift in the conversion products toward the diesel and wax fractions.
When operating catalysts for the CO conversion, a necessary stage is the preliminary activation of the catalyst in a stream of hydrogen, CO, or synthesis gas at elevated temperatures. Usually, this process takes place in a special apparatus and then the reduced catalyst is transferred to the reactor. This is complicated by the fact that the catalysts are pyrophoric and this procedure should be carried out very carefully. The creation of catalytic systems that do not require an activation stage is highly relevant.
Previously, the authors of the article studied the methods of forming active catalytic systems for CO hydrogenation without a preliminary activation step. For example, in [7,8] the catalysts were formed by the method of organic matrixes with further IR-pyrolysis of a joint metal salt (iron/cobalt) and various nature polymer solution. It was found that the nature of the polymer significantly affects the selectivity and composition of the resulting products. In the presence of Fe-copolymer of styrene and divinylbenzene, the C 5+ hydrocarbon selectivity was reached 60-80%, while an increased content of normal paraffins is observed. And in the case of Co-polyvinylalcohol, the selectivity for methane exceeds 70% with a CO conversion of 100%.
The solid decomposition products of double complex salts (DCS) exhibit promising catalytic properties. The research group of Potemkin D.I. is engaged in the systematic research of catalysts for total and preferential CO oxidation. We will consider this reaction further. These authors frequently use DCS as precursors for catalysts, which include one 3d-metal and one platinum metal. Often, catalysts obtained by thermolysis of DCS in various gas atmospheres are more efficient than catalysts that consist only of platinum metals. In [9], nanopowders of unordered solid solution Pt 0. 5  In the work shown on [12], it was found that the active phase of the reduction product of the [Co(NH 3 ) 6 ][Fe(CN) 6 ]+Al(OH) 3 mixture consists of the CoFe intermetallic compound. It changes the course of the Fischer-Tropsch synthesis reaction and alkenes become the main product (>90%). The aim of this work is to study the thermolysis products of [Co(NH 3 ) 6 ][Fe(CN) 6 ] in argon as catalytic compositions for the FTS process. The effect of the addition of Al(OH) 3 to [Co(NH 3 ) 6 ][Fe(CN) 6 ] is shown by the example of changes on thermal behavior, X-ray phase composition of thermolysis products, and, consequently, catalytic activity.  6 ] and their mixtures were obtained for a more complete picture of the thermal decomposition of the mixed catalytic composition and elucidation of the possible mutual influence of the precursor components.

Results and Discussion
Thermal decomposition of Al(OH) 3 was characterized by three stages of weight loss ( Figure 1): 60-140 • C (4 wt.%-removal of sorption water occurs); 140-290 • C (48 wt.%) and 290-550 • C (10 wt.%). The last two weight losses were associated with the dehydration of Al(OH) 3 . X-ray amorphous products were obtained by calcining Al(OH) 3 at 350 and 550 • C. A slightly crystallized γ-Al 2 O 3 was obtained by calcining at 900 • C. The endothermic peak of the evaporation of sorbed water, a relatively large exothermic peak of decomposition and, presumably, a weakly pronounced endothermic peak before the main stage of decomposition can be seen on the DSC curve. Weight loss in the range 290-550 • C was not accompanied by a pronounced peak.
average temperature was chosen in the indicated range, 650 °C.
Secondly, as it was written above, in the work of [12], promising data were obtained on catalysis in the FTS reaction. The composition [Co(NH3)6][Fe(CN)6] + Al(OH)3 was used as a precursor for the catalyst, the calcination temperature in hydrogen was 380 °C. Since the same composition was also used in this study, a similar temperature was chosen, namely, 350 °C.  The characteristics of the investigated catalytic compositions were sho Average values were given, each parameter was determined at least 3 time Diffraction patterns of samples I and II are shown in Figure S1: Pow fraction patterns of [Co(NH3)6][Fe(CN)6] and catalytic compositions. No bands for NH3 and CN − were found in the IR spectra of samples I and II spectra of [Co(NH3)6][Fe(CN)6] and catalytic compositions). Consequently not present in unchanged/slightly altered forms in the composition of the ucts.
Photomicrographs of catalytic compositions were shown in Figure 2.  (Figure 1). The decomposition processes of DCS were usually multistage and are characterized by several parallel processes of decomposition of ligands [13]. According to DSC data, the first, fourth, and fifth stages of decomposition were characterized by endothermic and two exothermic peaks. The processes, that were occurring in the temperature range 230-550 • C, did not have significant heat effects.
The thermal behavior of a mixture of Al(OH) 3 and [Co(NH 3 ) 6 ][Fe(CN) 6 ] was characterized by the following four main stages, in temperature ranges similar to the individual components: 60-140 • C (2.5 wt.%, this is similar Al(OH) 3 ), 140-230 • C (41 wt.%), 230-350 • C (12 wt.%) and 350-550 • C (6 wt.%) ( Figure 1). Weight loss stopped after 550 • C. However, the heat effects of the processes occurring during the thermolysis of the mixture were very different from the individual components. One distinct exothermic peak could be found on the DSC curve, corresponding to the main process of weight loss in the range 140-230 • C. Note that, based on the DSC data, the position of the peaks corresponding to the steps of mass loss changes and the sign of the heat effect changes. The decomposition of the mixture ended earlier than the DSC itself. A significant change in the mechanism of decomposition of DCS in the presence of Al(OH) 3 occurs.
The choice of temperatures for obtaining catalytic compositions was due to the following factors. First, according to the literature it was determined [14,15] that the formation of the CoFe intermetallic compound during thermolysis in argon [Co(NH 3 ) 6 ][Fe(CN) 6 ] occurs at temperatures of 600-675 • C. Therefore, in this work, the average temperature was chosen in the indicated range, 650 • C.
Secondly, as it was written above, in the work of [12], promising data were obtained on catalysis in the FTS reaction. The composition [Co(NH 3 ) 6 ][Fe(CN) 6 ] + Al(OH) 3 was used as a precursor for the catalyst, the calcination temperature in hydrogen was 380 • C. Since the same composition was also used in this study, a similar temperature was chosen, namely, 350 • C.
The characteristics of the investigated catalytic compositions were shown in Table 1. Average values were given, each parameter was determined at least 3 times. Diffraction patterns of samples I and II are shown in Figure S1: Powder X-ray diffraction patterns of [Co(NH 3 ) 6 ][Fe(CN) 6 ] and catalytic compositions. No characteristic bands for NH 3 and CN − were found in the IR spectra of samples I and II ( Figure S2: IR spectra of [Co(NH 3 ) 6 ][Fe(CN) 6 ] and catalytic compositions). Consequently, ligands were not present in unchanged/slightly altered forms in the composition of thermolysis products.
Photomicrographs of catalytic compositions were shown in Figure 2.
According to microprobe analysis, it was determined that sample I is a hollow carbon tube. A ball of intermetallic compound Co 0.5 Fe 0.5 was located at the end of these tubes. Sample II had a powdery structure without a definite appearance. Sample II contains a homogeneous mixture of Co, Fe, and Al.
The distribution of particles and pores by size was given in Tables 2 and 3. Mixing DCS with Al(OH) 3 led to an increase in the size of the main part of particles in thermolysis products from 4 to 27 µm. This can be explained by the presence of a larger amount of aluminum particles in the sample. Based on Figure 2, it can be seen, that in the case of sample I, the main part of the specific surface area was formed by hollow carbon tubes, and in sample II, by aluminum particles.

Catalytic Tests
It could be seen ( Table 4) that the thermal destruction products DCS and DCS + Al(OH) 3catalysts are active in the CO hydrogenation reaction without a preliminary activation step. With an increase of the synthesis temperature, a raise in the conversion of CO is observed on both catalysts: up to 93% for I and up to 35% for II. A high formation of by-products, such as gaseous hydrocarbons, with a selectivity of up to 61%, is observed on the catalyst without aluminum. The specific activity (A, molCO·gMe −1 c −1 ) of the catalysts also rises with increasing temperature, while the values are higher for the II catalyst, especially in the region of 290-310 • C. The selectivity for oxygenates was low and did not exceed 10%. On a catalyst containing aluminum, a higher selectivity for the target C 5+ hydrocarbons was observed. According to microprobe analysis, it was determined that sample I is a hollow carbon tube. A ball of intermetallic compound Co0.5Fe0.5 was located at the end of these tubes. Sample II had a powdery structure without a definite appearance. Sample II contains a homogeneous mixture of Co, Fe, and Al.
The distribution of particles and pores by size was given in Tables 2 and 3. Mixing DCS with Al(OH)3 led to an increase in the size of the main part of particles in thermolysis products from 4 to 27 μm. This can be explained by the presence of a larger amount of aluminum particles in the sample. Based on Figure 2, it can be seen, that in the case of sample I, the main part of the specific surface area was formed by hollow carbon tubes, and in sample II, by aluminum particles.

Catalytic Tests
It could be seen ( Table 4) that the thermal destruction products DCS and DCS + Al(OH)3-catalysts are active in the CO hydrogenation reaction without a preliminary activation step. With an increase of the synthesis temperature, a raise in the conversion of  In the works of [16][17][18][19], they used different carbonaceous materials such as activated carbon, carbon nanofiber, carbon nanotube and carbon spheres. On these systems, the selectivity for C 5+ was 40-80% with a CO conversion of 5-60%. It should be noted that all catalytic systems were preliminarily reduced in a stream of hydrogen for 20 h at temperatures of 350-400 • C. Whereas the catalysts formed in this work exhibited high activity in CO hydrogenation without a preliminary activation step. This makes them extremely promising for further study. Figure 3 shows that the distribution of the products of the Fischer-Tropsch synthesis on catalysts was significantly different. Thus, on the I catalyst, the maximum hydrocarbon distribution was a hydrocarbon with a chain length of 6-8 carbon atoms and the addition of aluminum to the catalyst composition shifts the maximum to C 9 -C 10 . It was also worth noting that on the II sample, more intense formation of higher molecular weight synthesis products occurs: the content of C 19+ hydrocarbons reaches 15%, while on the I catalyst it does not exceed 5%. This is because the introduction of aluminum into the system leads to the formation of weak Lewis centers, which are responsible for the polymerization activity of the Fischer-Tropsch synthesis catalyst, which leads to the formation of hydrocarbons with a longer chain length in all fractions [20]. Molecules 2021, 26, x FOR PEER REVIEW 8 of 12 From the data in Table 4, it could be seen that the introduction of aluminum into the composition of the catalyst leads to a twofold decrease in the content of olefins in the products. Most likely, this was due to an increase in the fineness of cobalt due to the presence of aluminum in the catalyst composition and the contribution of cobalt to the bimetallic catalyst became more significant, which led to an increase in the yield of hydrocarbons of normal structure and a decrease in the yield of olefins [22].
Authors of [23] found that an increase in the porosity of the support leads to a higher degree of metal reduction during the Fischer-Tropsch synthesis, which contributes to a more significant formation of olefins, which is in good agreement with the data in Table  5. Alcohols, as can be seen from Figure 4, fit quite well on a straight line, which indicated that, most likely, oxymethylene radicals were formed on the surface of the catalyst, followed by polycondensation with the formation of alcohols [24]. This type of mechanism was typical for chain growth on cobalt Fischer-Tropsch synthesis catalysts [20]. On the other hand, the composition of the products is influenced by the pore size distribution. From the data in Tables 2 and 3, the porosity of sample I is higher than for sample II. Most likely, the smaller pores of sample II increase internal diffusion hindrances during the reaction, which leads to the formation of hydrocarbons with a longer chain length. The larger pores of the sample favor a lighter hydrocarbon product formation due to a decrease in internal diffusion hindrances [21].
From the data in Table 4, it could be seen that the introduction of aluminum into the composition of the catalyst leads to a twofold decrease in the content of olefins in the products. Most likely, this was due to an increase in the fineness of cobalt due to the presence of aluminum in the catalyst composition and the contribution of cobalt to the bimetallic catalyst became more significant, which led to an increase in the yield of hydrocarbons of normal structure and a decrease in the yield of olefins [22].
Authors of [23] found that an increase in the porosity of the support leads to a higher degree of metal reduction during the Fischer-Tropsch synthesis, which contributes to a more significant formation of olefins, which is in good agreement with the data in Table 5. Alcohols, as can be seen from Figure 4, fit quite well on a straight line, which indicated that, most likely, oxymethylene radicals were formed on the surface of the catalyst, followed by polycondensation with the formation of alcohols [24]. This type of mechanism was typical for chain growth on cobalt Fischer-Tropsch synthesis catalysts [20]. Molecules 2021, 26, x FOR PEER REVIEW 9 of 12

Obtaining Catalytic Compositions
The sample I was the product of thermal destruction of [Co(NH3)6][Fe(CN)6] in argon at 650 °C. Sample II was obtained by mixing [Co(NH3)6][Fe(CN)6] with freshly precipitated Al(OH)3. The resulting mixture was dried to constant weight and calcined in argon at 350 °C.

STA Measurements
Simultaneous thermal analysis was made using the equipment STA 409 PC Luxx (NETZSCH-Gerätebau GmbH, Selb, Germany, Selb, 2009). Experiments were provided in a dynamic atmosphere of argon. The heating rate was 10 °C•min −1 , the temperature range was 30-1000 °C. Later, 10 mg of the sample was placed in Al2O3-crucible.

Isothermal Experiments
Thermal decomposition of the catalytical mixtures under isothermal conditions was carried out in a tubular reactor in a stream of Ar. A sample of ~0.5 g was heated to a prescribed temperature in a quartz boat with a heating rate of 10 °C/min and was held at this temperature for 1 h. After the end of the soaking time, the reaction tube was removed from the furnace and cooled to room temperature in a stream of Ar. The tubular muffle furnace Nabertherm RT 50-250/11 (Nabertherm GmbH, Lilienthal, Germany, 2013) was used. Argon was pre-dried by passing through concentrated H2SO4, the flow rate was 12-15 L•h −1 .

Physicochemical Research Methods
IR spectra were recorded on a Nicolet 6700 FT-IR spectrophotometer (Thermo Fisher Scientific Inc., Hillsboro, OR, USA, 2010) in the 4000-400 cm −1 region (KBr tablets). X-ray phase analysis of products of thermal destruction of DCS was made in 2θ range 5°-90° on  6 ] was obtained by stoichiometric mixing of equivalent amounts of solutions of the starting cationic and anionic precursor complex. [Co(NH 3 ) 6 ]Cl 3 (Vekton, Russia), K 3 [Fe(CN) 6 ] (Neva Reaktiv, Russia) with less than 10 −5 mol% impurities were used. Powder X-ray diffraction (XRD), IR spectroscopy, elemental analysis results are the same as described in [14,15,[25][26][27]. The curves of thermal analysis in argon, which were obtained in this study and reported in [14,26], have a similar character.

Obtaining Catalytic Compositions
The sample I was the product of thermal destruction of [Co(NH 3 ) 6 ][Fe(CN) 6 ] in argon at 650 • C. Sample II was obtained by mixing [Co(NH 3 ) 6 ][Fe(CN) 6 ] with freshly precipitated Al(OH) 3 . The resulting mixture was dried to constant weight and calcined in argon at 350 • C.

STA Measurements
Simultaneous thermal analysis was made using the equipment STA 409 PC Luxx (NETZSCH-Gerätebau GmbH, Selb, Germany, Selb, 2009). Experiments were provided in a dynamic atmosphere of argon. The heating rate was 10 • C·min −1 , the temperature range was 30-1000 • C. Later, 10 mg of the sample was placed in Al 2 O 3 -crucible.

Isothermal Experiments
Thermal decomposition of the catalytical mixtures under isothermal conditions was carried out in a tubular reactor in a stream of Ar. A sample of~0.5 g was heated to a prescribed temperature in a quartz boat with a heating rate of 10 • C/min and was held at this temperature for 1 h. After the end of the soaking time, the reaction tube was removed from the furnace and cooled to room temperature in a stream of Ar. The tubular muffle furnace Nabertherm RT 50-250/11 (Nabertherm GmbH, Lilienthal, Germany, 2013) was used. Argon was pre-dried by passing through concentrated H 2 SO 4 , the flow rate was 12-15 L·h −1 .

Sample Preparation for Micrographs
The samples were placed on an aluminum table. The table was covered with nonsprayed conductive carbon tape. TESCAN AMBER (Brno, the Czech Republic, 2021),-a two-beam scanning electron microscope based on a field emission cathode of the Schottky type with an additional Orage ion column with a Ga ion source, was used. Energy dispersive detector EDS Ultime 100 manufactured by Oxford Instruments (High Wycombe, England, 2021) was applied.

Catalytic Tests
The synthesis of hydrocarbons from CO and H 2 was carried out in a flow catalytic unit with a fixed-bed reactor. The catalyst was diluted with an equal volume of quartz to reduce the resistance of the powder layer to gas flow and placed on the quartz layer to prevent the material from spilling out of the reaction zone.
The synthesis was carried out under continuous operation using synthesis gas CO:H 2 = CO:H 2 = 1:1 (mol.) at a pressure of 2 MPa and gas hourly space velocity of 300 h −1 . The temperature rise in the range of 210-330 • C was carried out stepwise (by 20 • C every 12 h). 12 h of the experiment are required to enter the mode and to accumulate the necessary amount of reaction products for analysis. At the end of each isothermal regime, gas and liquid samples were taken for analysis.

Analysis of Reagents and Reaction Products in FTS
The initial synthesis gas and gaseous products of synthesis were analyzed by gas chromatography on a Krystallux-4000 chromatograph (Russia, Yoshkar-Ola, 2017) with a TCD detector and the carrier gas is helium. Two chromatographic columns were used. A column filled with CaA molecular sieves (3 m × 3 mm) was used to separate CO and N 2 . Temperature condition: isothermal, 80 • C. To separate CO 2  Oxygen-containing products in the aqueous phase were analyzed by gas-liquid chromatography on a Kristallux-4000M chromatograph (Russia, Yoshkar-Ola, 2017). The detector was flame ionization. Gas flow rate: helium −20 mL min −1 , hydrogen −25 mL min −1 , air −250 mL min −1 . An HP-FFAP capillary column (Nitroterephthalic Modified Polyethylene Glycol) (50 m × 0.32 mm × 0.50 µm) was used for analysis. The sample volume is 0.3 µL. Temperature mode: programmable, 70 • C (8 min), 70 • C-110 • C, 10 • C min −1 , 110 • C-220 • C, 15 • C min −1 −220 • C (10 min). For quantitative calculation, the internal standard method was used (the standard is isobutyl alcohol).
The following indicators were used to evaluate the activity of the catalyst: CO conversion (percentage of the mass of reacted carbon monoxide to the mass of CO entering the reaction zone), product yield (number of grams of the product obtained by passing 1 m 3 of synthesis gas through the catalyst reduced to normal conditions), activity (A)-the number of moles of CO that reacted per 1 g of the sum of metals in 1 s, selectivity (S)-the percentage of carbon used to form product related to the total amount of carbon reacted.

Conclusions
Intermetallic compound Co 0.5 Fe 0.5 and X-ray amorphous carbon in the form of hollow tubes is obtained by thermal decomposition in argon [Co(NH 3 6 ] and Al(OH) 3 . It was shown that the synthesized DCS-based catalysts are active in the high-temperature Fischer-Tropsch synthesis. It was found that the introduction of aluminum into the composition of the catalyst leads to the formation of higher molecular weight hydrocarbons and a decrease in the content of olefins in the reaction products. This is most likely associated with the formation of a more dispersed cobalt phase on the one hand and the formation of weak Lewis centers, which are responsible for the polymerization activity of the catalyst on the other hand.
Supplementary Materials: The following are available online. Figure S1: Powder X-ray diffraction patterns of [Co(NH 3