Influences of Co-Content on the Physico-Chemical and Catalytic Properties of Perovskite GdCo_xFe_1−xO₃ in CO Hydrogenation

Abstract

The effect of the substitution of cobalt into the GdFeO₃ perovskite structure on the selective hydrogenation of CO was investigated. A series of GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1) samples were synthesized by sol-gel technology and characterized by XRD, BET specific area, DSC, TG, EDX and XPS. The experimental data made it possible to reveal a correlation between the state of iron and cobalt atoms, the fractions of surface and lattice oxygen, and catalytic characteristics. It has been found that varying the composition of GdCo_xFe_1−xO₃ complex oxides leads to a change in the oxygen-metal binding energy in Gd-O-Me, the ratio of metals in various oxidation states, and the amount of surface and lattice oxygen, which affects the adsorption and catalytic characteristics of complex oxides.

1. Introduction

Light olefins are among the basic petrochemical products. Currently, the global production of these products is estimated at 185 million tons [1], and the global production of propylene already exceeds 100 million tons per year [2]. The continuous growth in demand for light olefins is largely determined by the increase in the consumption of polyethylene and polypropylene and the expansion of their technological applications. To solve this problem, it is necessary to create a variety of effective catalytic technologies to produce light olefins from available raw materials.

One of the promising methods of ethylene production is the process of hydrogenation of CO. This process implies the use of noble metal-based catalysts such as Ru, Pt, Pd and Rh because of their inherent resistance to carbon deposition on the surface. However, the use of these catalysts is not economically viable. The industry also prefers nickel and cobalt catalysts because these metals are inexpensive and readily available. However, nickel- and cobalt-based catalysts are more prone to coke deposition under these synthesis conditions, resulting in reduced activity and deactivation [3,4,5].

Thus, the development of low-budget, effective and sintering-resistant catalysts for obtaining ethylene is an important problem. The issue of studying and developing new catalytic systems, improving already used processes and creating new methods aimed at obtaining olefins under conditions of hydrogenation of carbon monoxide remains relevant nowadays. Oxides with a perovskite structure (ABO₃) are considered an alternative to traditional catalysts for many industrial processes due to their high thermal stability, the ability to control physicochemical properties and low cost [6,7,8]. One of the advantages of perovskite-like oxides is the possibility of creating a wide range of compositions by complete/partial, as well as isovalent/non-isovalent substitution of A and B cations while maintaining the structure. This allows for controlling the stability of the structure [9], electronic [10], redox and surface properties of complex oxides [11], creating new functional materials [12,13,14,15,16,17,18,19,20].

Perovskites have good flexibility and diversity in their chemical composition and can accommodate solid defects, such as vacancies at both the cation and anion sites [21], [22]. The B-site transition metals on the surface are believed to be the active centers owing to the exposed 3d electron orbitals. Considering that the B-site transition metals embedded in the perovskite lattice are atomically dispersed, we can expect to develop highly active and anti-coking catalysts.

Perovskites have been widely studied in oxidation reactions [23,24,25]. However, the study of Fe- and Co-containing perovskites in the hydrogenation of carbon monoxide is insufficiently described in the literature [26,27,28,29,30,31]. From the analysis of the existing works, it can be concluded that there is a close relationship between the nature and chemical state of the elements responsible for the activity and the type of surface C-containing intermediates and product distributions. Accordingly, A(Fe/Co)O₃ perovskite-type oxides with reducible B³⁺ ions under typical reaction conditions are excellent model catalysts for gaining new and useful insight into the surface process taking place during CO activation and carbon chain growth.

Thereby the purpose of this work was to study the catalytic activity of systems based on perovskite-type iron- and cobalt-containing complex oxides in the process of carbon monoxide hydrogenation and to establish the relationship between the physicochemical properties of complex oxides and their catalytic characteristics. It should be noted that catalysts of similar composition worked well for dry reforming of methane [32] and diesel soot oxidation [33].

2. Results and Discussion

2.1. Characterization

Complex oxides GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1) were studied by X-ray phase analysis, X-ray photoelectron spectroscopy and SEM+EDX. Detailed XFA and XPS results are presented in [26]. All samples are single-phase and have an orthorhombic structure with the spatial group Pbnm and Pnma. All diffraction peaks can be attributed to the orthorhombic structure of GdFeO₃ (PDF-ICDD 01-072-9908), GdCoO₃ (PDF-ICDD 00-025-1057) and their solid solutions.

Analysis of the dependence of the structural parameters a, b and c on the composition of GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1) (Figure 1) shows that the values of parameters a and c decrease slightly with a greater iron substitution for cobalt in ferrite, while the lattice parameter b remains practically unchanged for all complex oxides except for the ferrite samples with cobalt fraction x = 0.8 and completely substituted GdCoO₃. This indicates that the lattice geometry distorts with increasing cobalt content.

The crystallite size of the obtained samples estimated using the Scherrer formula is in the range from ~47 to 65 nm (

Table 1. Results of studying GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1) surface composition by X-ray spectral, microanalysis, XRD crystallite size and S_BET.

Compound	XRD Crystallite Size, nm	SBET (m2/g)	Content (at %)				Fe/Co
			Gd	Fe	Co	O
GdFeO3	53.4	9.9	18.53	19.15	-	62.32	-
GdCo0.2Fe0.8O3	47.3	3.3	17.58	16.35	4.67	61.40	3.5
GdCo0.5Fe0.5O3	50.5	2.7	17.43	8.77	9.49	64.31	0.92
GdCo0.8Fe0.2O3	65.2	3.1	16.84	3.96	17.21	61.98	0.23
GdCoO3	59.9	2.3	16.02	-	17.80	66.18	-

) and does not show a dependence on the composition [33].

The specific surface area of the obtained compounds was determined by the BET method (Table 1); for GdFeO₃, the highest value was observed, which is consistent with the results of scanning electron microscopy. For gadolinium cobaltite and solid solutions GdCo_xFe_1−xO₃ (x = 0.2; 0.5; 0.8), the low values of the specific surface area are due to the relatively high calcination temperature (800 °C). At the same time, the values of the surface area of the obtained oxides were found to be higher than for similar compounds prepared by the ceramic method (less than 1.0 m²/g) [34]. Thus, the values of the specific surface area do not show a dependence on the composition, and the conditions of preparation have a great influence.

The surface morphology of gadolinium ferrite and cobaltite, as well as their solid solutions obtained after calcining at 800 °C for 1 h, was studied using scanning electron microscopy. Figure 2 shows photomicrographs of the surface of oxides with the composition GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1); all studied compounds demonstrate a similar surface morphology, which indicates that the complete or partial substitution of iron for cobalt in the perovskite structure does not affect surface morphology.

The micrographs of the surface in Figure 2 show a uniform morphology without particle agglomeration. The surface is characterized by a “porous” structure, possibly associated with combustion and gas evolution during oxide production. It should be noted that, in contrast to gadolinium cobaltite and partially substituted ferrites, gadolinium ferrite (GdFeO₃) exhibits a more porous morphology, which is due to a lower preparation temperature and is consistent with the results of determining the specific surface area (Table 1).

To study the composition of the surface, X-ray spectral analysis was used and elemental mapping of the surface was carried out. The results show that the amount of Gd and Fe/Co on the surface of the obtained compounds is close to the theoretical value (Table 1).

XPS was used to study the surface composition, the oxidation state and the distribution of atoms in the obtained compounds. Traditional photoelectron spectroscopy is a surface-sensitive method in which the depth of probing in the energy range 200 eV–1500 eV does not exceed 2 nm. Figure 3 presents the overview XPS spectra of oxides GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1) in the binding energy range of 0–1400 eV and represents the intense peaks of Gd3d, Fe2p, Co2p, O1s and C1s (signals of other elements were not found). The chemical states of the atoms in the compounds under study were refined by decomposing the obtained spectra described in [33].

The partial substitution of iron for cobalt in the samples GdCo_xFe_1−xO₃ (x = 0.2; 0.5; 0.8) resulted in the heterovalent state Fe²⁺, Fe³⁺(711.0; 724.0 eV) and Co²⁺/Co³ (779.97; 794.0 eV) (

Table 2. Results of studying the surface composition and state of atoms by XPS [33].

Compound	Fe	Co	Ratio
			Co2+/ ∑ Con+	Fe2+/ Fe2+ + Fe3	Os/Ol
GdFeO3	+3	-	-	0.64	1.52
GdCo0.2Fe0.8O3	+2; +3	+2; +3	0.51	0.43	0.59
GdCo0.5Fe0.5O3	+2; +3	+2; +3	0.50	0.58	0.81
GdCo0.8Fe0.2O3	+2; +3	+2; +3	0.13	0.58	0.82
GdCoO3	-	+2; +3	0.47	-	0.83

). The heterovalent state causes the appearance of oxygen vacancies in the structure of Co-containing samples and changes in the bond lengths Me-O-Me [35].

The O1s photoelectron spectra of compounds GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1) shown in Figure 4 are asymmetric and wide, which confirms the presence of various types of oxygen in these compounds [36,37]. Analysis and decomposition of the spectra showed that the obtained compounds contained various types of oxygen (at binding energies of ~533.00 eV, 530.90 and 528.57–529.67 eV), which are referred to as physically absorbed oxygen (H₂O), chemically adsorbed oxygen (O_s:O₂⁻, O₂²⁻ and O⁻) and lattice oxygen (O_l), respectively. Table 2 presents a quantitative assessment of various types of oxygen in the studied oxides. The O_s/O_l ratio in GdCo_xFe_1−xO₃ at x = 0.2; 0.5; 0.8; 1 increases with an increase in the amount of cobalt, and these four samples have almost the same values of the surface area. At the same time, for the GdFeO₃ (x = 0), a higher value of the surface area and a higher value of the O_s/O_l ratio are observed. The study [38] shows that chemisorbed oxygen (O_s) plays an important role in low-temperature oxidative reactions, while lattice oxygen (O_l) affects the oxidizing ability of compounds in high-temperature reactions. Thus, the O_s/O_l ratio can be used to assess the oxidative characteristics of the compounds under study.

As was shown in [33], when using the H₂-TPR and O₂-TPD methods, the mobility of lattice oxygen increases with increasing cobalt content in GdCo_xFe_1−xO₃ samples. This tendency confirmed that there are a large number of cationic vacancies in the Co-rich samples due to the easier Co³⁺→Co²⁺ transition.

2.2. CO Hydrogenation

The catalytic activity in the reaction of carbon monoxide hydrogenation was evaluated in the temperature range 523–708 K at a CO:H₂ ratio = 1:2 (CO conversions are shown in Figure 5). The substitution of small amounts of cobalt (x < 0.5) led to the suppression of CO conversions compared to the initial gadolinium ferrite. On the sample GdCo_0.2Fe_0.8O₃ (x = 0.8), degrees of conversion of carbon monoxide did not exceed 45%. Substitution of iron for cobalt (x = 0.2; 0.5) leads to an improvement in oxygen mobility and a decrease in the amount of surface oxygen O_s. This is reflected in the decrease in CO conversion as CO is adsorbed through the surface oxygen, but the introduction of more cobalt into ferrite gadolinium (x > 0.5) promoted the growth of CO conversion and it reached 90–95%.

For GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8) samples, the conversions barely depended on temperature and reached almost maximum values at 523 K. This is due to the participation in the reaction of carbon particles formed because of the dissociative adsorption of CO at lower temperatures [21]. A slight increase in the CO conversion with increasing temperature may be due to the involvement of not only surface oxygen but lattice oxygen as well. The amount of CO in the gas phase correlated with its conversions and was determined by the type of catalyst. At the same time, on all catalysts, the process was accompanied by the formation of carbon dioxide, the amounts of which also depended on the composition of perovskite (Figure 6).

The amount of formed carbon dioxide in the region up to 623 K was approximately the same, but further increasing the temperature led to an increase in the rate of CO₂ formation. It should be noted that on the sample GdCoO_3, the process of CO₂ formation was much more intense than on the other samples.

It is known that CO adsorption mainly occurs at the A-centers of perovskite with the formation of carbonate complexes Gd₂O₂CO₃ [39]. The decomposition of these complexes produces carbon dioxide (Equation (1)).

Gd₂O₂CO₃(s) → CO₂(g) + Gd₂O₃(s)

(1)

In the studied perovskites, the parameters a and c of the crystal lattice change; hence, the oxygen-metal bond energy in Gd-O-Me will also change, which in turn leads to a change in the gadolinium-oxygen bond and is reflected in the value of CO adsorption. The formation of CO₂ is also possible in the interaction of the adsorbed CO_ads molecule with perovskite (O_S) oxygen (Equation (2)):

CO_ads + O_S → CO₂

(2)

As stated above, the substitution of iron for cobalt improves oxygen mobility and redox properties and changes the ratio of surface and bulk oxygen in the structure of complex oxides: the share of O_s decreases, while the share of O_l increases. It can be assumed that with increasing temperatures, both surface and lattice oxygen begin to participate in the reaction (2) (the amount of lattice oxygen depends on the composition of the oxide, i.e., with cobalt enrichment the oxidation capacity of the samples increases).

The products of hydrogenation of carbon monoxide on GdCo_xFe_1−xO₃ catalysts (x = 0; 0.2; 0.5; 0.8; 1) were C₁–C₅ hydrocarbons, and the main ones were methane, ethylene, ethane and propylene. The rate of formation of hydrocarbons began at 573 K and increased with increasing temperatures.

A comparison of the rates of formation of the main products (methane, ethylene, propylene and CO₂) showed that the substitution of Fe in gadolinium ferrite for Co leads to an increase in the specific catalytic activity (Figure 7), but on the GdCo_0.2Fe_0.8O_3, a decrease in the rate of product formation was observed. It is worth noting that as the proportion of cobalt in the initial ferrite increases, the formation of propylene stops. The formation rates of methane and ethylene increased in the series:

GdCo_0.2Fe_0.8O₃ ˂ GdFeO₃ ˂ GdCo_0.5Fe_0.5O₃ ˂ GdCoO₃ ˂ GdCo_0.8Fe_0.2O₃

(3)

The appearance of a synergistic effect on GdCo_0.8Fe_0.2O₃ is due to the fact that this sample has different structural characteristics from the others.

Varying the composition of catalysts led to a change in the quantitative ratio of products. Thus, when performing the reaction on GdFeO3, the content of methane, ethylene and propylene at T = 648 K was 76%, 14% and 6%, respectively. The complete substitution of iron for cobalt increased the amount of methane to 85%. However, the GdCo_0.2Fe_0.8O₃ sample showed a decrease in methane to 38%, ethylene to 9% and propylene to 5%. The sample with the cobalt substitution fraction x = 0.8 had an ethylene content of 13%.

The dependence of the hydrocarbon chain growth factor α_i on the number of carbon atoms (i) in the chain of intermediate products is shown in Figure 8.

Regardless of the catalyst composition, the highest values were for ${\alpha }_1$, i.e., the formation of hydrocarbons with two carbon atoms was most likely in the studied systems, and on the GdCo_0.2Fe_0.8O₃ catalyst, this probability reached a maximum of 50%.

From the literature data [35] it is known that iron hydrogen is adsorbed mainly in atomic form and molecular forms are practically absent, but cobalt hydrogen can be adsorbed in both atomic and molecular forms. Moreover, on the iron centers, hydrogen is more mobile relative to the surface, unlike cobalt, because in the series Fe, Co and Ni, bond energy Me-O grows [25,35]. Thus, increasing the proportion of cobalt in the B-position leads to the growth of the molecular forms of hydrogen and their ordering. The probability of the formation of CH₂ particles increases and results in the formation of hydrocarbons with an even number of carbon atoms.

The calculation of selectivity for target products (ethylene, propylene and butylene) showed that the substitution of iron for cobalt in the perovskite led to a decrease in selectivity for the target reaction products. Thus, at 573 K, the ${\mathrm{C}}_2^{=}\text{-}{\mathrm{C}}_4^{=}$ selectivity on GdFeO₃ was 29% and 20% on GdCoO₃ (Figure 9). The replacement of iron with a small amount of cobalt led to an increase (of approximately 32%) in the selectivity of light olefins. The most selective with respect to ethylene turned out to be the sample with a fraction of cobalt x = 0.2, for which the Co²⁺/∑ Coⁿ⁺ ratio is maximal and Fe²⁺/∑ Feⁿ⁺ is minimal.

To determine kinetic characteristics, such as the effective activation energies of product formation and the logarithms of the pre-exponential multiplier (an indirect characteristic of the number of the active centers of the catalyst surface), the interpolation of experimental data in linear coordinates of the Arrhenius equation was performed. On the Arrhenius dependences of the rates of formation of ethylene for samples with a content of cobalt x > 0.5 two areas of linearity with a transition temperature of about T* ≈ 598–609.7 K were observed. The presence of two linear sections with different values of the activation energies E_a indicates either a change in the state of the active Meⁿ⁺ centers depending on the temperature (change in the oxidation state of Co³⁺ ↔ Co²⁺) or the course of different processes of a complex reaction in each temperature interval (T < T* is chemisorption of reactants, and T > T* is the conversion of chemisorbed complexes into products). The calculated values of the effective activation energies for the formation of reaction products and the logarithms of the prefactor for all the catalysts studied are shown in

Table 3. Resulting atomic state, CO conversion, selectivity for olefins, experimental activation energies and pre-exponential multipliers for GdCo_xFe_1−xO₃ catalysts.

Compound	GdFeO3	GdCo0.2Fe0.8O3	GdCo0.5Fe0.5O3	GdCo0.8Fe0.2O3	GdCoO3
Co2+/∑ Con+	-	0.52	0.50	0.13	0.47
Fe2+/Fe2+ + Fe3+	0.64	0.43	0.58	0.58	-
Os/Ol	1.52	0.59	0.81	0.82	0.83
α(CO), %	80	44	63	93	89
S(CnH2n), %	27.1	38.4	26.4	2.4	0.55
Ea(CH4), kJ/mol	70	77	99	177/252	159
ln K0(CH4)	5.7	5.6	11.3	29.6/46.6	25.1
R2	0.96	0.97	0.99	0.99/0.99	0.99
Ea(C2H4), kJ/mol	112	88	115/186	112/244	99/226
ln K0(C2H4)	18.5	6.7	12.9/28.2	15.0/43.1	11.5/37.1
R2	0.97	0.97	0.96/0.99	0.96/0.99	0.98/0.99

.

As the data in Table 3 show, the introduction of a small amount of cobalt led to a decrease in the activation energies of ethylene formation, while the value of the pre-exponential multiplier also decreased. The activation energies of methane formation and the logarithms of the prefactor were comparable to those of the original ferrite. For the GdCo_0.8Fe_0.2O₃ sample, the effective activation energies were the highest. However, in spite of the high activation energy, high values of ln K₀ are also observed. For the same sample, the minimum Co²⁺ content can be seen. For samples with a fraction of cobalt x ≥ 0.5, there are comparable values of O_s/O_l and Fe²⁺/(Fe²⁺ + Fe³⁺), but only the samples with x = 0.8 the minimum Co²⁺/∑ Coⁿ⁺ ratio and the highest rates of product formation were observed. At the same time, the most selective was the sample with the cobalt fraction x = 0.2, for which the Co²⁺/∑ Coⁿ⁺ ratio is maximal and Fe²⁺/∑ Feⁿ⁺ is minimal.

It can be assumed that Co³⁺ and Fe³⁺ in the Gd-O-Me bond are active centers, and the presence of Co²⁺ somewhat complicates the process and shifts it towards the formation of olefins.

The determination of the deposited carbon on the surface of the catalytic systems because of the reaction was carried out by thermogravimetric analysis (TG) and differential scanning calorimetry (DSC). Figure 10 shows the mass loss curves for GdCo_xFe_1−xO₃ samples (x = 0; 0.5; 1) as an example. In the temperature range of 553–593 K, a slight increase in the mass of the GdCo_0.5Fe_0.5O₃ sample was observed, which is most likely due to the process of oxygen chemisorption from the air environment on the surface carbon and the subsequent decrease in mass with exo-effects in the 573–753 K range is due to carbon desorption in the form of CO₂.

TG analysis showed a loss of mass of the samples up to temperatures of ~893–913 K (620–640 °C). The amount of deposited carbon increases with increasing cobalt content. The largest amount was observed in GdCoO₃. However, for all investigated perovskite systems, the formation of the so-called “resin” as well as the “encapsulation” of the catalyst particles was not observed.

3. Materials and Methods

3.1. Catalyst Preparation

The synthesis of GdCo_xFe_1−xO₃ complex oxides (x = 0; 0.2; 0.5; 0.8; 1) was performed using the citrate-nitrate sol-gel technique [40,41]. Stoichiometric Gd(NO₃)₃·6H₂O (99.9%, Vecton, St. Petersburg, Russia), Fe(NO₃)₃·9H₂O (98%, Vecton, St. Petersburg, Russia) и Co(NO₃)₃·9H₂O (98.5%, Vecton, St. Petersburg, Russia) nitrates were dissolved in distilled water, and then the flask was placed on a magnetic stirrer with a thermostat. Citric acid was added to the resulting salt solution with constant stirring. After the complete dissolution of citric acid, to set the pH of the solution at 6, an ammonia solution was added, and then the solution was evaporated at a temperature of ~120 °C until the resulting gel ignited and a black powder was formed. Then the temperature was raised to 450 °C and the resulting powder was calcined for 2 h. The samples containing cobalt were additionally calcined at 800 °C for 1 h.

3.2. Characterization

Powder X-ray diffraction (XRD) was used to determine the phase composition and structural data of the obtained compounds. The analysis was carried out using a Rigaku MiniFlex II diffractometer with CuK_α radiation. Diffraction patterns of all investigated compounds were obtained under the same conditions sufficient to detect all characteristic reflections: the range of angles is 2θ = 10–60°, and the scanning speed is 5°/min. The obtained diffraction patterns were analyzed by comparing the present reflections with the data of the ICDD-PDF2 database. The data obtained allowed us to conclude the phase composition of the studied compounds.

X-ray photoelectron spectroscopy (XPS) was used to analyze the surface composition and determine the state of atoms. The measurements were carried out on a Thermo Fisher Scientific Escalab 250Xi spectrometer using radiation AlK_α = 1486.6 eV, spectral resolution-0.5 eV. The characteristic C1s line (C-C bond), which is observed at a bond energy of 284.8 eV, was used as a control line. Experiments on temperature-programmed hydrogen reduction (H₂-TPR) were carried out on a Micrometrics AutoChem II 2920 analyzer. Samples with a mass of <0.05 g were placed in a U-shaped quartz reactor and heated with a constant flow of a gas mixture (10% H₂/90% Ar, feed rate = 50 mL/min) up to 1000 °C (heating rate = 10 °C/min). The amount of absorbed gas was measured using a thermal conductivity detector (TCD). In order to exclude the effect of water vapor on the operation of the TCD, a trap was placed in front of the detector and in a Dewar with ethyl alcohol (T = −139 °C). Sample temperature, valve position and gas flow rate data were monitored and controlled using a computer equipped with Micromeritics AutoChem II software.

Thermogravimetric analysis using a Netzsch STA 449 F5 Jupiter was performed to estimate the amount of deposited carbon after catalytic tests. The studies were carried out on samples with a suspension of 20–30 mg in a stream of air (rate = 50 mL/min⁻¹) at a heating rate of 10 K/min⁻¹ in the temperature range 303–1173 K. The specific surface area was determined by the method of low-temperature nitrogen adsorption at T = 77 K on Nova 4200e (Quantachrome, Boynton Beach, FL, USA) and QuadrasorbSI devices. The samples were preliminarily degassed at 300 °C for 5 h. The obtained adsorption-desorption isotherms were used to estimate the specific surface area of the samples by applying the BET method.

3.3. Catalytic Activity Tests

Catalyst performance of the samples in the carbon monoxide hydrogenation reaction was evaluated in a quartz tubular microreactor containing 0.1 g of the sample diluted in 0.5 g of quartz in order to avoid the formation of hot spots in the catalytic bed alongside increasing the volume. Activity tests were carried out in the temperature range 523–708 K under atmospheric pressure with the feed composition CO:H₂ = 1:2 and total flow rate of 1.5 L⋅h⁻¹, which corresponds to GHSV of 8700 h⁻¹. Gaseous effluent was analyzed online using a gas chromatograph instrument (Crystal 5000.2, a column of stainless steel filled with Porapack Q, argon as a carrier gas) fitted with thermal conductivity and flame ionization detectors.

The catalytic characteristics were calculated using the following equations:

$$X_i, \%=\frac{n_{int}-n_{out}}{n_{int}}\times 100$$

(4)

$$S_i=\frac{R_i}{\sum R_i}·100\%,$$

(5)

$$R_i=\frac{n_i{}_{out}\omega }{Vm}$$

(6)

where n_int и n_out is the initial amount of substance and mole yield of products, ∑n_i is the mole yield of hydrocarbon reaction products (mole), R_i is the rate of formation of the i-th reaction product per 1 g catalyst (mol/h⋅g), ∑R_i is the sum of yields of the i-th reaction products per 1 g catalyst (mol/h⋅g), ω is volumetric rate of the reaction mixture (L/h), V is the volume of the chromatograph loop (0.153 × 10⁻³ L) and m is the mass of the catalyst, g.

The carbon chain growth factor α_i was calculated using Equation (7) [42]:

$${\alpha }_i=\frac{{\sum }_{k>i}\frac{Y_k}{k}}{{\sum }_{k\ge i}\frac{Y_k}{k}}$$

(7)

where ${\alpha }_i$ is carbon chain growth factor for the intermediate with the number of carbon atoms I, Y_k is the yield of the component with the number of carbon atoms k.

The coefficient α₁ is the probability of the addition of CO to the intermediate containing one carbon atom to form an intermediate with two carbon atoms; α₂ is the probability of the next step of the addition of CO to the intermediate with two carbon atoms, etc.

4. Conclusions

Complex perovskite-type oxides GdCo_xFe_1−xO₃ (x = 0; 0.2; 0.5; 0.8; 1.0) prepared by the sol-gel method were studied as catalysts for the hydrogenation of carbon monoxide. The effect of iron substitution by cobalt on the physicochemical and catalytic properties in the process was investigated. It was found that the metals in the samples were in heterovalent states Fe²⁺/Fe³⁺ and Co²⁺/Co³⁺, which is compensated by oxygen vacancies. It was found that the partial substitution of iron for cobalt in GdCo_xFe_1−xO₃ (0.5 < x < 1) and a decrease in the Co²⁺ share in the sample resulted in an increase in CO conversions; nevertheless, x = 0.2 with the maximum Co²⁺/∑ Coⁿ⁺ ratio and minimum Fe²⁺/∑ Feⁿ⁺ ratio was the most selective for ethylene. It has been suggested that Co³⁺ and Fe³⁺ are active centers, while the presence of Co²⁺ somewhat complicates the process and shifts it towards the formation of olefins. It was also found that varying the composition of GdCo_xFe_1−xO₃ complex oxides leads to changes in the oxygen-metal bond energy in Gd-O-Me, the ratio of metals in different oxidation states, and the amount of surface and lattice oxygen, which is reflected in the adsorption and catalytic characteristics of complex oxides and change of the quantitative ratio of olefins in the process of CO hydrogenation.