CO₂ Hydrogenation to Methanol on CuO-ZnO/SiO₂ and CuO-ZnO/CeO₂-SiO₂ Catalysts Synthesized with β-Cyclodextrin Template

Abstract

A series of mixed copper (II)—zinc oxide catalysts supported on unmodified and ceria-modified silica supports were synthesized using β-cyclodextrin as a template. The novelty of this work lies in the use of cyclosextrins for the template synthesis of catalyst supports. The obtained samples were analyzed by XRD, SEM-EDX, low-temperature nitrogen physisorption, XPS, and EPR. The magnetic properties of the catalysts were also measured. The thermal decomposition of precursors was analyzed by TGA combined with mass-spectrometric analysis of the evolved gases. The effects of the support pore size, the nature of the active phase and its loading, as well as the sequence of component deposition on the catalyst performance in the CO₂ conversion to methanol were studied. The catalysts with cerium added at the gelation stage demonstrated the best performance. The selectivity of these samples reaches values of more than 90% over a fairly ide temperature range, with the productivity reaching 480 g/kg cat·h at 300 °C.

1. Introduction

It is well understood that the increase in the CO₂ concentration in the atmosphere leads to climate change and negatively affects the human health. There are three main ways to diminish the CO₂ concentration in the atmosphere [1]. The first is to directly reduce the emission of this harmful pollutant. However, despite many efforts and international agreements on CO₂ emission reduction, its concentration has continued to increase constantly. Therefore, this method does not seem to be technologically effective. The second is to store huge amounts of CO₂. However, this method has not developed far beyond pilot projects. In the third way, CO₂ is used as a raw material for the production of alkanes, some other hydrocarbons, carboxylic acids, alcohols, etc. This approach simultaneously solves two problems: utilization of carbon dioxide and industrial production of valuable substances.

The CO₂ molecule is relatively inert, so breaking its chemical bonds is not easy [2]. This process is not usually performed due to the extreme conditions required. Therefore, conversion of CO₂ into a particular product is a complex problem. As such, catalysts are usually used in this process [3,4,5].

Methanol is one of the products of the catalytic conversion of CO₂ [6,7,8]. It is an important industrial intermediate that can be used to produce formaldehyde, acetic acid, hydrocarbons, olefins, gasoline, etc. In addition, it is a promising energy carrier and alternative fuel. Methanol is also used as a solvent, as an antifreeze, and as a precursor of other commodity chemicals [9].

CO₂ conversion to methanol can be described as a two-stage process [7,10]. In the first stage, CO₂ is reduced to CO, and in the second stage CO is hydrogenated to the final product (CH₃OH). However, this representation is not accurate enough, since the process is more complicated than indicated above, and plenty of by-products are formed. Some authors [11,12,13,14] believe that the reaction can proceed via two routes with different intermediates. CO₂ can be hydrogenated to methanol through the formate or hydrocarboxyl route. On the one hand, the formation barrier of HCOO• was found to be lower than that of •COOH. On the other hand, the hydrocarboxyl route is more kinetically favorable than the formate route. Even though formate species were found to be the most abundant on the catalyst surface during CO₂ conversion, the formate route is suggested to play a secondary role. However, the mechanism of CO₂ conversion to methanol is still a controversial issue [13,14].

Nowadays, a variety of catalysts are used in the process of CO₂ conversion into several significant chemical products, including alcohols, hydrocarbons, formaldehyde, and others. Among them, noble metals catalysts (Au, Pd), transition-based (Mo, Cu, Co), and bi- and multimetallic systems are most common [15,16,17,18,19]. It is important to consider the fact that CO₂ adsorbs on the medium strength basic sites to form covalent, hydrogen, and bidentate carbonates. Several organic-based systems are also used [16,18]. Carbon materials have some unique advantages, including accessible active sites, a very tailorable structure, high electrical density, etc. Several other factors also have an impact: the properties of solvent, the conditions used, additional insertion of promoters, nature of the support [15].

Copper or copper oxide promoted with aluminum or zinc oxides is the most used industrial catalyst for CO₂ conversion to methanol. Chromium, cobalt, nickel, and platinum oxides or their alloys can be also applied [20,21], but they are more expensive. Therefore, CuO and Cu catalysts are widely described in the literature [22,23,24,25,26,27]. Copper cations in the catalyst adsorb, bind, and stabilize the intermediate products via direct bonding with a carbon atom (hydrocarboxyl route) or through the oxygen-bridge (formate route) [7,25]. Various promoters (in particular, zinc oxide [28]) increase binding and adsorption of carbon di- and monoxides, accelerating the process.

Among other metal oxides, ceria-based catalysts are promising for different chemical processes [29,30,31,32,33]. Ceria demonstrates unique properties. It has a high oxygen capacity, the ability to form solid solutions with a wide range of metal oxides that can act as additional active sites, etc. Ceria-based catalysts are now widely used in environmentally important processes and in energy conversion systems [30]. Various supports can be used to improve the catalyst performance by increasing, for example, the specific surface area of the catalyst [34,35,36].

Silica gels have a high porosity and, as a result, an active inner surface, which makes them attractive for different applications, including catalysts, adsorbents, and catalyst supports. The efficiency of utilization strongly depends on the porosity of the silica gel, which, in turn, depends on the synthesis method [37,38,39]. The sol-gel method is commonly used to synthesize silica gel [40]. In this method, two processes proceed simultaneously during aging: condensation and polymerization. The rates of these processes depend on different parameters, such as the temperature, the amount of water in the system, the concentrations of substances used in the silica gel synthesis, etc.

Other techniques, such as templated synthesis, are also used to tune the pore structure of the catalyst [41,42,43]. For example, cyclodextrins can be used as templates. Cyclodextrins are cyclic oligosaccharides consisting of the macrocyclic ring of glucose subunits joined via α-1,4 glycosidic bonds, whereas α, β-, and γ-cyclodextrins contain 6, 7, and 8 glucose subunits, respectively. During aging, cyclodextrins form columnar structures over which silica gel chains are grown [41,42,43,44,45]. Large pores are formed in the silica gel after burning out the template (cyclodextrin). The disadvantage of cyclodextrins is their high solubility in water, which sharply reduces their associating ability. Therefore, β-cyclodextrin is used most often because of its lowest water solubility and higher ability to form columns. However, it must be dissolved in water. For this purpose, additives of different substances, in particular urea, are used. They can form complexes with cyclodextrins, thus increasing their solubility [42].

This work focuses on the synthesis of CuO-ZnO/CeO₂-SiO₂ catalysts using a β-cyclodextrin template. The catalysts are characterized by different physicochemical methods and are tested in the catalytic conversion of CO₂ to methanol. The influence of the synthesis parameters on the catalyst structure and performance is traced.

2. Results

2.1. Thermal Analysis

The unannealed catalyst precursors (all cerium-free and one cerium-containing) were studied by TGA-DTG analysis (Figure 1). Decomposition proceeds in two main stages. In the first stage, water starts to desorb at around 100 °C. Nitrates are decomposed in the second stage at 200–400 °C. The TGA curve of the cerium-containing catalyst precursor shows a more pronounced mass loss in the narrow temperature range of 200–250 °C because of the cerium nitrate decomposition. As an example, a thermogram of the precursor of the sample 5CZ/CeSi-5, combined with the results of the mass-spectrometric analysis of the evolved gases, is presented (Figure S1).

2.2. X-ray Diffraction Analysis

Figure 2 displays the XRD patterns of the cerium-free catalysts. A broad halo at 20–25° resulted from amorphous SiO₂, while all the observed reflections can be attributed to copper oxide. No reflections of zinc-containing phases were detected. This is likely due to the low concentration of zinc oxide when compared to the mass of the entire system. With increasing copper loading in the catalyst (20CZ/Si-5 versus 5CZ/Si-5), the CuO reflections became more intense. Additionally, the use of a larger water content during the synthesis (5CZ/Si-20 versus 5CZ/Si-5) also leads to more intense reflections. However, the CuO crystallite size, determined by the Debye–Scherrer equation, remained the same (about 15 nm).

The XRD patterns of the cerium-containing catalysts are displayed in Figure 3. The catalysts in which cerium was added at the gelation stage (5CZ/CeSi-5 and 5CZ/CeSi-5(m)) demonstrate similar patterns. Only broad ceria reflections are observed in these patterns. The low crystallinity of ceria perhaps resulted from the incorporation of ceria species into the amorphous matrix of silica gel. Ceria and copper oxide reflections were detected in the XRD patterns of the catalysts impregnated with the cerium salt (5CZ/Ce/Si-5 and 5CZCe/Si-5), with the former sample demonstrating better crystallinity. The ceria crystallite size in 5CZ/Ce/Si-5, determined by the Debye–Scherrer equation, was approximately 8 nm; it was about 5 nm in 5CZCe/Si-5. Thus, the introduction of cerium by incipient wetness impregnation improves the crystallinity of both ceria and copper oxide. However, the crystallinity of copper oxide in the cerium-containing catalysts is lower than in the cerium-free ones.

2.3. Textural Properties

The catalysts were analyzed by low-temperature nitrogen physisorption. The sorption isotherms are shown in Figures S2 and S3. The calculated textural parameters are summarized in

Table 1. Textural properties of the samples.

Sample	S, m2/g	Dmax, nm	Vtot, cm3/g	Vmeso, cm3/g	Vmicro, cm3/g
Si-5	542	6.4	0.89	0.89	-
5CZ/Si-5	363	10.2	0.90	0.90	-
5CZCe/Si-5	376	6.4	0.66	0.66	-
5CZ/Ce/Si-5	325	6.4	0.59	0.59	-
Si-20	354	12.2	1.06	1.06	-
CeSi-5	247	3.9, 9.5	0.46	0.46	-
CeSi-5 (m)	554 (273 *)	1.2, 1.6, 2.0	0.29	0.17	0.12
5CZ/CeSi-5 (m)	380 (80 *)	0.7, 2.1	0.22	0.17	0.05

* Micropore specific surface area—“t-plot”; DFT.

and Figure 4a,b. The specific surface area of the catalysts strongly depends on their composition and synthesis technique. The support with ceria incorporated at the gelation stage (CeSi-5 synthesized with β-CD and urea) has the lowest specific surface area. The other ceria-containing support synthesized with m-β-CD (CeSi-5 (m)) showed the largest specific surface area due to the presence of micropores. The support synthesized from a mixture with a lower water content (Si-5) showed smaller pores than its counterpart synthesized from a mixture with a larger water content (Si-20).

The pore size of the catalysts obtained using these supports changes slightly, while the specific surface area and total pore volume decrease because of the inclusion of copper, zinc, and cerium oxides. Moreover, it is remarkable that, in the case of the sample with micropores, they are filled, while the volume of mesopores remains the same.

2.4. Energy-Dispersive X-ray Spectroscopy

The catalyst composition was studied by EDX. The measured atomic ratios for Si, Cu, Zn, and Ce are shown in Figure 5, while the theoretical values are presented in parentheses. The SEM image for the sample 5CZ/CeSi-5 is presented in Figure S4.

The first diagram in Figure 5a shows the silicon to copper ratios. In the samples without ceria, copper is distributed more non-uniformly, with a higher copper concentration at the surface, as the experimental Si:Cu ratios is almost two times lower than the theoretical ones. This fact points to the positive effect of cerium(IV) oxide on the uniformity of copper(II) oxide distribution.

The second diagram in Figure 5b shows the silicon to zinc ratios. As expected, these ratios are higher than the Si:Cu ratios because of the molar ratios of copper and zinc used in the synthesis. The difference between the experimental and theoretical Si:Zn ratios is similar to that for the Si:Cu ratios. The ceria-containing catalysts show an almost uniform distribution of zinc, but the catalysts without ceria demonstrate less uniform zinc distribution. Furthermore, the experimentally determined Si:Zn ratios in the cerium-free samples are also almost two times lower than the theoretical ones. Therefore, the surface of these samples is enriched with zinc in the same way as it is enriched with copper.

The last diagram in Figure 5c shows the silicon to cerium ratios. The difference between theoretical and experimental values is almost negligible except for the sample with sequentially supported nitrates. Therefore, ceria is evenly distributed in all the samples. The amount of cerium in the surface is only slightly larger in the two samples with cerium deposited by the incipient wetness impregnation (5CZCe/Si-5 and 5CZ/Ce/Si-5) than in the samples with cerium added at the gelation stage.

It can be concluded that the addition of ceria increases the uniformity of copper(II) and zinc oxide distributions. Moreover, ceria is distributed more uniformly than copper oxide. Zinc distribution is more difficult to evaluate because of its lower content and the higher uncertainty in determining its concentrations.

2.5. X-ray Photoelectron Spectroscopy

Samples 5CZ/Ce/Si-5, 5CZCe/Si-5 and 5CZ/CeSi-5 were analyzed by XPS. The XPS spectra are shown in Figures S5–S7. To evaluate the Ce⁴⁺/Ce³⁺ ratio, the Ce3d spectra of the catalysts were fitted with the synthetic Ce³⁺ and Ce⁴⁺ components as described in method B in the relevant reference [46]. At the same time, the Cu2p spectra were fitted with a narrow component at about 932.2 eV attributed to Cu⁺ species and a broader component of Cu²⁺ species at higher binding energy and with a set of shake up satellites [46]. Both Cu2p and Ce3d spectra were recorded within a short acquisition time at the start and at the end of the XPS experiments. The fitting of these spectra demonstrated a strong reduction of Cu²⁺ and Ce⁴⁺ species during XPS analysis (

Table 2. Percentages of different Cu and Ce species in the catalysts calculated from their high-resolution XPS spectra.

Sample	ν (Ce), %		ν (Cu), %
	Ce4+	Ce3+	Cu2+	Cu+
5CZ/Ce/Si-5
fast acquisition (start *)	97	3	77	23
long-term acquisition	90	10	53	47
fast acquisition (end **)	88	12	37	63
5CZCe/Si-5
fast acquisition (start)	97	3	87	13
long-term acquisition	93	7	63	37
fast acquisition (end)	91	9	58	42
5CZ/CeSi-5
fast acquisition (start)	92	8	97	3
long-term acquisition	71	29	62	38
fast acquisition (end)	65	35	54	46

* Cu2p and Ce3d spectra were recorded immediately after X-ray source and neutralizer were switched on. ** Cu2p and Ce3d spectra were recorded at the end of XPS experiments.

). Therefore, the reliable data on the oxidation states of copper and cerium can only be evaluated from the spectra acquired within a short acquisition time immediately after the X-ray source and neutralizer were switched on. This evaluation shows (see Table 2) that, within the fitting uncertainty (at least ±5%), cerium at the surface of the samples is totally in the +4 oxidation state. Copper also mainly exists in the highest oxidation state (+2), but a contribution from Cu+ species is also observed, and it decreases in the order: 5CZ/Ce/Si-5 > 5CZCe/Si-5 > 5CZ/CeSi-5.

2.6. Magnetic Properties

The magnetic properties of the catalysts were analyzed, and the results are summarized in

Table 3. Magnetic properties of the catalysts.

Sample	Magnetic Susceptibility, emu/g	Magnetic Susceptibility Taking into Account Pascal Constants, emu/g	Magnetic Susceptibility, emu/mol Cu	Cu+2 Content, %
5CZ/Ce/Si-5	4.26 × 10−7	4.34 × 10−7	6.90 × 10−4	54
5CZ/CeSi-5	6.76 × 10−7	6.76 × 10−7	1.07 × 10−3	84

.

The spin only effective magnetic moment for the ion Ce⁺³ is 2.54 µB; for the ion Cu⁺² it is 1.79 µB. Ions Ce⁺⁴ and Cu⁺ are diamagnetic. The molar magnetic susceptibility of ions (${\mathsf{\chi }}_{\mathrm{i}}$) can be calculated from the equation ${\mathsf{\chi }}_{\mathrm{i}}=\frac{{\mathsf{\mu }}_{\mathrm{i}}^2}{8\mathrm{T}}$. The molar magnetic susceptibilities of Cu⁺² and Ce³⁺ ions are 1.27 × 10⁻³ and 2.73 × 10⁻³, respectively. The magnetic susceptibility of the samples can be calculated as $\mathsf{\chi }={\mathsf{\chi }}_{\mathrm{C}\mathrm{e}+3}+{2.6\mathsf{\chi }}_{\mathrm{C}\mathrm{u}+2}$. The empirical composition of 5CZCeSi-5 is Cu₁Ce_2.6Si_17.6O_41.5−x. Thus, the molar magnetic susceptibility for the samples containing only paramagnetic ions will be 8.56 × 10⁻³, which is much higher than the measured values (Table 4). Thus, the magnetic susceptibility of the samples is determined by the ratio of Cu⁺² to Cu⁺ species. It is obvious that copper and cerium are not in an individual oxidation state but represent a mixture. For sample 5CZ/CeSi-5, the susceptibility is higher than for 5CZ/Ce/Si-5, which can be explained by the high content of Ce⁺³ in the sample. This trend is also confirmed by the XPS data.

2.7. Electron Paramagnetic Resonance Spectroscopy

The room temperature EPR spectra of the samples are presented in Figure 6. The spectra do not contain the Ce³⁺ signal reported for similar systems [47], which indicates that cerium is completely oxidized to Ce⁴⁺. The literature data [48] allows attributing the signal in the spectra to a thin film or nanoparticles of CuO rather to a bulk oxide.

The integral intensities of the signals differ significantly for the catalysts, which may be due to the formation of an antiferromagnetic copper phase, the proportion of which increases in the order 5CZ/CeSi-5—5CZ/Ce/Si-5—5CZCe/Si-5. The latter sample shows an extremely low signal, which additionally complicates the interpretation.

The spectra represent a superposition of the signals of paramagnetic Cu²⁺ ions in different environments. It is obvious that there is a signal with HFS from copper nuclei (I = 3/2), g^‖ ≈ 2.32, g^┴ ≈ 2.06, A^‖ ≈ 500 MHz (signal A). Similar signals are observed for lead-silicate glasses [49] and CuO-ZrO₂ [50]. This signal refers to copper ions in the axially symmetrical environment. It should be noted that interatomic distances change in the disordered medium; therefore, the signal was described by a Lorentzian line shape with a variable intrinsic width and broadening due to the normal distribution of the spin Hamiltonian parameters (g and A), and a negative correlation was considered. A large line width indicates the inhomogeneity of the local environment in amorphous phases.

For a more adequate description of the line shape, three more signals were added. The choice of initial parameters was based on the work [51], i.e., fixed g_iso and varied the g-tensor parameters. The ideal description of the real line shape has not been achieved; in particular, it is not possible to accurately describe the part of the spectrum in low fields.

Upon refinement, one of the signals (B) degenerates into an isotropic one with g ≈ 2.14, which was assigned to bulk defects of the CuO phase [52]. There is a wide signal with unresolved HFS from copper nuclei (I = 3/2), g^‖ ≈ 2.31, g^┴ ≈ 2.05, A^‖ ≈ 80 MHz (signal C). The nature of signal D with HFS from copper nuclei (I = 3/2), g^‖ ≈ 2.47, g^┴ ≈ 2.05, A^‖ ≈ 420 MHz is not completely clear. Narrower A and D signals may refer to Cu²⁺ ions magnetically isolated in the amorphous phase. Broad signals are more likely to be related to bulk phases containing copper.

The EPR spectra have a complex shape and represent a superposition of signals from several types of paramagnetic centers. The proposed description is not completely exact; however, there are no experimental grounds for introducing new types of paramagnetic centers into the model, and even the interpretation of the available signals is too complicated.

2.8. Catalytic Activity

The catalytic activity of the samples is presented in Figure 7 as temperature dependences of productivity. The samples with cerium added at the stage of gel formation (catalysts 5CZ/CeSi-5 and 5CZ/CeSi-5(m)) showed the highest catalytic activities, probably because of the uniform distribution of elements (Cu, Zn and Ce) in these catalysts, as it was found by EDX. The activity of 5CZ/CeSi-5(m) slightly decreased after 250 °C.

By increasing the copper loading in the catalyst (20CZ/Si-5 versus 5CZ/Si-5), the rate of hydrogenation increases, which indicates the positive role of copper oxide. The increase in the water amount during the synthesis (5CZ/Si-20 versus 5CZ/Si-5) also increases the catalyst activity, probably due to the changes in the textural properties (the pore size increases).

The lowest catalytic activity was observed for the catalysts with cerium added by impregnation (5CZ/Ce/Si-5 and 5CZCe/Si-5). This probably resulted from a higher degree of crystallization of oxides (according to the XRD data) or because of the high content of Cu⁺ species (according to the XPS data).

The temperature dependences of the selectivity to methanol over the catalysts are presented in Figure 8. The highest selectivity was observed for the catalysts with cerium added at the stage of gel formation (5CZ/CeSi-5 and 5CZ/CeSi-5 (m)). Also, relatively high values were detected for the catalysts containing 20 wt.% of copper(II) oxide (20CZ/Si-5). The lowest selectivity, the same as the lowest catalytic activity, was observed for the catalysts synthesized via impregnation of the support with the cerium salt.

It is worth noting the difference in the temperature dependencies of the selectivity for the cerium-containing and cerium-free catalysts. The selectivity of the latter ones decreased with increasing temperature above about 210 °C, while for the ceria-containing catalysts, it only slightly changed.

The catalyst activities were compared with data available in the literature for similar catalytic systems. It has already been stated that a huge number of different catalysts with various additives were tested in attempts to improve the efficiency of the catalytic hydrogenation of carbon dioxide. Copper and its oxides are among the most used catalysts in this reaction. Copper-nickel alloys were tested [53], and the Cu_0.5Ni_0.5 and Cu_0.75Ni_0.25 showed the best performance; the selectivity reached 95%, while the productivity was up to about 600 g/(kg cat·h). Complex Cu/ZnO/ZrO₂ catalysts were also investigated and showed a sufficiently high catalytic activity due to high ability for CO₂ adsorption [54]. In this case, the selectivity to methanol reached 80%, while the productivity was 500 g/(kg cat·h). The catalytic activity of a ternary CuO-ZnO-Al₂O₃ catalyst promoted with SiO₂, TiO₂, or SiO₂-TiO₂ was studied [55]. It appears that the CO₂ conversion and methanol yield increased with the addition of promoters, and the selectivity to methanol reached up to 75–80%. However, the productivity of methanol in the reported work was 350 g/(kg cat·h). Cobalt-based catalysts are also widely used in this process. Thus, the selectivity of CoO nanoparticles at high temperatures can reach 73% at a productivity of 380 g/(kg cat·h) [56]. Various CoO/SiO₂ catalysts demonstrated the selectivity of about 50–70% and the productivity of 250–400 g/(kg cat·h) depending on the CoO loading used [57]. The results of our work are mainly comparable with the best literature data (Table 4). This work also confirms the beneficial effect of the porous support and method of the active phase deposition.

Table 4. Performances of catalysts of CO₂ conversion to methanol.

Reference	Catalyst	Selectivity, %	Productivity, g/(kg cat·h)
[53]	CuxNi1−x	75	400
[54]	CuO/ZnO/ZrO2	80	500
[55]	CuO/ZnO/Al2O3	85	350
[56]	CoO	75	380
[57]	CoO/SiO2	75	400
This work	5CZ/CeSi-5	95	480

Table 4. Performances of catalysts of CO₂ conversion to methanol.

Reference	Catalyst	Selectivity, %	Productivity, g/(kg cat·h)
[53]	CuxNi1−x	75	400
[54]	CuO/ZnO/ZrO2	80	500
[55]	CuO/ZnO/Al2O3	85	350
[56]	CoO	75	380
[57]	CoO/SiO2	75	400
This work	5CZ/CeSi-5	95	480

3. Materials and Methods

Tetraethyl orthosilicate (TEOS) (Component-reaktiv, Moscow, Russia) distilled at 167 °C and 748 Torr, cerium(III) nitrate hydrate (98%, CHEMCRAFT Ltd., Kaliningrad, Russia), copper(II) nitrate trihydrate (99%, Acros Organics, Geel, Belgium), zinc nitrate hexahydrate (99%, Alfa Aesar, Lancaster, UK), β-cyclodextrin (β-CD) (99%, Sigma-Aldrich, Moscow, Russia), methyl-β-cyclodextrin (m-β-CD) (99%, Sigma-Aldrich, Moscow, Russia), urea (98%, Acros Organics, Geel, Belgium), concentrated nitric acid (99%, Component-reaktiv, Moscow, Russia), and distilled water were used for the synthesis. The equipment that was used for different methods of sample analysis and measurement conditions are described in the Supplementary Information.

3.1. Synthesis of the Supports

Two mixtures were prepared simultaneously. The first mixture consisted of TEOS and a 0.1 M solution of nitric acid (3.7 mL of TEOS and 1.3 mL of the acid solution per 1 g of produced SiO₂). This mixture was constantly stirred with a magnetic mixer for two hours until the phase boundary became almost indistinguishable. The second reaction mixture was a solution consisting of water, urea, and β-CD. Urea was used to increase the solubility of β-CD. In one of the syntheses, β-CD was replaced by m-β-CD (the mass of the latter was taken from the calculation of the same number of moles of these substances; due to the high solubility of m-β-CD, urea was not used). The mass fraction of the template was:

w = (m(CD) + m(u))/(m(CD) + m(u) + m(SiO₂)) = 0.6,

where m(CD) is the mass of β-CD and m(u) is the mass of urea; their ratio (m(CD):m(u)) was 1:3.

In all cases, the water to SiO₂ ratio of 5 was used (except one sample with the water to SiO₂ ratio of 20). In several cases, CeO₂ was used to modify the support by the addition of cerium(III) nitrate to the second mixture. The weight of the nitrate was chosen so that the CeO₂ loading in the catalyst was 30 wt.%.

In each synthesis, the second mixture was added to the first one drop by drop during stirring, and then stirred for 1 h. Next, the mixture was sealed with a film and left for 24 h. After that, the film was punctured to let the hydrolysis products and water evaporate, and the mixture was left for two weeks at room temperature to age the gel and increase its viscosity. Subsequently, the gel was crushed into small pieces, dried at 120 °C for 24 h, and annealed at 600 °C for 2 h.

Support names were constructed using Ce for ceria, Si for silica gel, and the number indicating the water to SiO₂ ratio. For example, the name Si-20 was used for the SiO₂ support synthesized with water to SiO₂ ratio of 20.

3.2. Synthesis of the Catalysts

The solutions of copper and zinc nitrates were prepared in order to provide the CuO loadings in the synthesized catalyst of 5 and 20 wt.% and the molar ratio of Cu to Zn of 2:1. These solutions were supported by incipient wetness impregnation onto the synthesized supports. Copper and zinc nitrates were added simultaneously to all catalysts. In two syntheses (see Table 1), cerium nitrate was also deposited by incipient wetness impregnation up to CeO₂ loading of 30 wt.%. Moreover, in one case (sample 5CZ/Ce/Si-5 in Table 1), the support was initially impregnated with a cerium nitrate solution, and then, after drying and annealing at 600 °C for 2 h, with a mixture of solutions of copper and zinc nitrates. In the second synthesis (sample 5CZCe/Si-5 in Table 1), the support was impregnated simultaneously with three nitrates. Then the samples were annealed at 450 °C for 4 h to decompose nitrates. The compositions and designations of the catalysts are presented in

Table 5. Compositions and designations of the samples.

Catalyst	Support	CD Type	m(H2O)/m(SiO2)	wt.% CuO
5CZ/Si-5	Si-5	β-CD	5:1	5
5CZ/Si-20	Si-20	β-CD	20:1	5
20CZ/Si-5	Si-5	β-CD	5:1	20
5CZ/CeSi-5	CeSi-5	β-CD	5:1	5
5CZ/CeSi-5(m)	CeSi-5 (m)	m-β-CD	5:1	5
5CZCe/Si-5 *	Si-5	β-CD	5:1	5
5CZ/Ce/Si-5 **	Si-5	β-CD	5:1	5

* The support was simultaneously impregnated with a solution of copper, zinc, and cerium nitrates. ** The support was initially impregnated with a solution of cerium nitrate and then with the mixture of solutions of copper and zinc nitrates.

. Names of the obtained catalysts are constructed using five main parameters. The first one is the number used to denote the weight percent of copper oxide (5 wt.% or 20 wt.%). The second one is CZ designating copper and zinc oxides. The third part is letters used for naming the support. The fourth one is a number used for the mass ratio of water during synthesis of the support (five times or twenty times). The last part is the letter ‘m’, used to denote m-β-CD. The forward-slash ‘/’ was used to explain the order in which the phases were supported. For example, the designation 5CZ/Si-20 is used to name the catalyst obtained from the support SiO₂ with the amount of water 20 times larger than that of SiO₂, with the solutions of copper and zinc nitrates being supported to obtain the catalyst with 5 wt.% of CuO.

3.3. Catalytic Tests

Catalyst samples were tested in the CO₂ hydrogenation reaction at the pressure of 50 atmospheres and in the temperature range of 150 °C to 300 °C in a flow-through apparatus equipped with a 6 mm diameter stainless steel reactor with a fixed catalyst bed. In the experiment, the catalyst fraction of 0.25–0.5 mm was used. The catalyst loading was 150 mg. The catalyst was mixed with quartz of the same fraction as the catalyst, resulting in a volume of 1.4 mL. The ratio of H₂:CO₂ flows was 3:1. The total volumetric flow rate of the initial gases (H₂ + CO₂) was 80 mL/min. The reaction products were analyzed by using the online sampling mode with a Crystal 5000 chromatograph (Chromatec, Yoshkar-Ola, Russia). Before the reaction, a series of samples were reduced in a reactor in a hydrogen flow of 50 mL/min at a temperature of 260 °C for 4 h in all experiments.

4. Conclusions

It was shown that the synthesis conditions directly affect the physicochemical and catalytic properties of the CuO-ZnO/CeO₂-SiO₂ catalysts. The amounts of gases released during the decomposition of nitrates, urea and cyclodextrin affect the pore size that, in turn, also influences the properties of obtained catalysts. The incorporation of ceria into the support structure changes the interaction between copper and zinc oxide particles and the support, which affects the uniformity of the distribution of elements over the surface.

The catalysts with cerium added at the stage of gelation showed the highest efficiency in the catalytic conversion of CO₂ to methanol. These catalysts appear to have uniform distribution of the elements over the surface according to EDX data. The cerium-containing samples synthesized by the impregnation method also show the uniform distribution of the elements, but their catalytic activity is lower probably because of the presence of Cu⁺¹ species. Among other synthesis parameters, the amount of water used in the synthesis of the support seems to be noted. It was shown that with increasing the amount of water, the pore size in the support and the degree of crystallinity of the CuO phase increase. A higher water content during catalyst synthesis led to a higher productivity, but at the same time to a noticeably lower selectivity to methanol.