Ceria–Zirconia-Supported Ruthenium Catalysts for Hydrogen Production by Ammonia Decomposition

Abstract

Commercial cerium–zirconium oxide supports (Ce_0.5Zr_0.5O₂, Ce_0.75Zr_0.25O₂, and Ce_0.4Zr_0.5Y_0.05La_0.05O₂) were used to prepare Ru/CeZrO_x catalysts. According to the XRD and IR spectroscopy data, the supports consist of ceria-based substitutional solid solutions. The specific surface areas of supports and catalysts are similar and range from 71–89 m²/g. As shown by TEM and XRD methods, the size of support particles equals 6–11 nm. According to the TEM data, the size of ruthenium particles does not exceed 1.3 nm. The catalyst activity in the ammonia decomposition process was studied. The Ru/Ce_0.75Zr_0.25O₂ catalyst at temperature 500 °C and GHSV 120,000 h⁻¹ demonstrated the highest hydrogen productivity of 53.3 mmol H₂/(g_cat·min) and compares well with the best results reported in the literature. The kinetics of ammonia decomposition reaction were calculated using the Temkin–Pyzhov exponential expression. The developed mathematical model well described the experimental data. The studied catalysts demonstrated high activity for the ammonia decomposition reaction.

1. Introduction

Recently, hydrogen has attracted particular attention as an environmentally conscious energy carrier which is zero-carbon and provides high energy conversion efficiency when used in fuel cells [1]. However, low volumetric energy density, low boiling point, and the hard-to-handle storage and transportation of hydrogen are still limiting factors for its wide application as a fuel [2]. In 2020, a demonstration project was implemented for hydrogen transportation from Brunei to Japan, using a liquid organic hydrogen compound (LOHC), namely, toluene-methylcyclohexane (TOL-MCH) [3,4]. Currently, the gravimetric storage capacity of the most promising LOHCs does not exceed 7 wt.% hydrogen; in particular, TOL-MCH contains only 6.1 wt.% hydrogen, which is much lower compared to ammonia [5]. Another downside of the LOHC technology is the need to transport the dehydrogenated organic compound back to the hydrogenation plant. The ammonia-based technology is free of this problem; the released nitrogen can either be used for local needs, or simply emitted to the atmosphere. Exactly ammonia seems to be the most attractive hydrogen carrier owing to its environmental safety, carbon-neutrality, and high hydrogen content (17.8 wt.%, or 121 kgH₂/m³ at 10 bar); this material is easy to handle and can be readily liquefied, transported, and converted to hydrogen [6].

Hydrogen production from the decomposition of ammonia has been the subject of significant research in recent years. The ammonia decomposition reaction is endothermic and a high operating temperature is required to perform this process at high conversions. Thus, in recent years, significant progress has been made in the research of ammonia decomposition catalysts, and various catalyst systems based on Fe [7,8,9], Co [9,10], and Ni [8,9,11,12,13] have been developed. It is known that ruthenium-based systems are the best catalysts for low-temperature decomposition of ammonia. A large number of studies have been devoted to the research and development of high-performance ruthenium catalysts for the process of ammonia decomposition to produce CO_x-free hydrogen [4,5,6,14,15,16,17,18,19,20]. Ruthenium catalysts supported by various materials—MgO [9,19,21,22], carbon materials [16,17,18,23,24], Al₂O₃ [6,18,25,26], ZrO₂ [14,27,28,29] zeolites [30], SiO₂ [31,32,33], etc.—have been studied in detail. Particular attention is focused on the oxides of rare earth metals La₂O₃ [27,34,35,36], Pr₆O₁₁ [27,37,38,39] and, especially, CeO₂ [18,25,26,35,36], which have specific redox and structural characteristics that enhance the catalyst performance in the reaction of ammonia decomposition. For example, Hu et al. found that CeO₂ was the best support for the Ru catalyst compared to CNTs and Al₂O₃, particularly at low temperatures [18]. This result is attributed to the strong metal support interaction and the electronic modification of Ru by ceria.

In order to improve the catalyst activity, thermal stability, and dispersion of the active component, ceria is often doped with oxides of other rare earth metals, for example, ZrO₂ [29,40,41,42]. As reported by several research groups, the introduction of trivalent rare earth ions (La³⁺, Nd³⁺, Pr³⁺, Y³⁺ and Sm³⁺) into the lattice of CeO₂–ZrO₂ solid solutions could further improve their reducibility and thermal stability [43], and markedly improve the interaction between the active component and support.

Most often, the synthesis of ruthenium catalysts is performed by impregnating the support with the active component precursor, followed by thermal treatment [18,24,26,27,31,36]. In recent years, attempts have been made to optimize the synthesis procedure in order to obtain the most efficient catalysts containing highly dispersed ruthenium particles and possessing advanced activity in the process of ammonia decomposition.

The present work is devoted to the preparation, physicochemical characterization, and studies of catalytic properties of the ceria–zirconia-supported ruthenium catalysts in the process of hydrogen production by ammonia decomposition. A kinetic model of the process of catalytic decomposition of ammonia was reliably developed describing the experimental results.

2. Experimental

2.1. Synthesis of Ruthenium Catalysts

Catalyst samples containing ~2–3 wt.% ruthenium were prepared. As supports, we chose Ce_0.5Zr_0.5O₂, Ce_0.75Zr_0.25O₂, and Ce_0.4Zr_0.5Y_0.05La_0.05O₂, to be used as constituent parts of catalysts for automotive exhaust gases afterburning. The active component precursor was the [Ru(NH₃)_nCl_m]Cl_p (n = 5–6; m = 0–1; p = 1–2) complex prepared according to the procedure described in [44]. The synthesized ruthenium–ammonia complex was loaded on the supports by impregnation followed by drying in air at 120 °C for 3 h and reduction in H₂ flow (60 mL/min) at 450 °C for 4 h.

2.2. Physicochemical Methods for the Supports and Catalysts Characterization

X-ray diffraction analysis (XRD) of the samples was performed using a D8 Advance diffractometer (Bruker, Germany, CuKα radiation, λ = 0.15418 nm). The diffraction patterns were obtained by scanning in the range 2θ = 10–80 °C, step size 0.05 ° (2θ), and scan speed 2 s/step. Based on the XRD data, the positions of intensity maxima, interplanar spacings d_hkl, and full width at half maximum (FWHM) were calculated; the qualitative phase composition of the samples, the unit cell parameters, and dispersion (the coherent scattering regions (CSR) size) were determined.

The low temperature nitrogen adsorption (LTNA) method was used to determine the specific surface area of the supports and catalysts using an ASAP 2400 instrument.

X-ray photoelectron spectroscopy (XPS) measurements were performed using an ES300 spectrometer (KRATOS Analytical, USA) with MgKα (hν = 1253.6 eV). The power of the X-ray source in all cases was 78 W. The instrument was calibrated by the position of the maximum of the Au4f7/2 and Cu2p3/2 lines at 84.0 and 932.7 eV for bulk metallic gold and copper, respectively. The spectra were calibrated by the position of the Ce 3D U‴ peak. In the case of cerium oxide, the position of the U‴ peak was assumed at 916.7 eV [45]. The relative element composition ratio of the sample was determined from perspective peak areas using empirical atomic sensitivity factors (ASF) after recording narrow regions of the corresponding atomic layers [46].

Transmission electron microscopy (TEM) studies were performed using a JEM-2100 transmission electron microscope (JEOL, Japan) at grating resolution 0.14 nm and accelerating voltage 200 kV.

CO Chemisorption

The dispersion of ruthenium in the prepared catalysts was determined by the method of pulsed chemisorption of CO with a ChemBET Pulsar TPR/TPD instrument (Quantachrome Instruments, USA). CO chemisorption was performed after the preliminary reduction of samples in a 10% H₂/Ar mixture at 400 °C for 10 min followed by purging with helium for cooling the sample to room temperature. The pulses of 10 vol.% CO/He mixture at regular time intervals was fed into a flow of inert carrier gas (helium). After oxidation in O₂ at room temperature, the ruthenium dispersion was redetermined. The values of dispersion and particle sizes were calculated by assuming linear CO chemisorption on ruthenium (the stoichiometric ratio, CO:Ru = 1:1) [47].

The IR spectra of the supports and catalysts were recorded using an IRPresige-21 instrument (Shimadzu, Japan), wavenumber range 350–7900 cm^–1, spectral resolution 4 cm^–1, and scans accumulation number 50. To record IR spectra, a small amount of the sample was mixed with KBr powder and compressed into a tablet.

Determination of the catalyst activity in the ammonia decomposition process was performed using a fixed-bed flow reactor-based setup (Figure 1).

The reaction of ammonia decomposition was performed in a 12 cm³ glass reactor, at atmospheric pressure, in the temperature range 400–500 °C, and at ammonia flow rates of 20, 60, and 100 mL/min. For the studies, catalyst weights were used: 0.0579 g Ru/Ce_0.5Zr_0.5O₂ 05, 0.0511 g Ru/Ce_0.75Zr_0.25O₂, and 0.0525 g Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂. A thermocouple in the reactor was placed in the middle of the catalyst bed. The reaction products were analyzed using a TSVET-500M chromatograph (Russia) with a thermal conductivity detector, the carrier gas—hydrogen; the temperature of the column packed with the Haeyesep C sorbent (which enabled NH₃ and N₂ separation) was 70 °C. Continuous sampling was provided by a six-way computer-controlled valve. The kinetics were studied in the temperature range 350–510 °C in increments of 20 °C. The catalyst weight was 20.3 mg, so that the degree of decomposition of ammonia did not exceed 40%, even at 510 °C. The best catalyst (Ru/Ce_0.75Zr_0.25O₂) was used to conduct durability tests (Figure S5), with its thermal stability checked for 2 days. Moreover, two such experiments were made—the first with ammonia from a mixture of NH₄Cl and NaOH, and the second with pure ammonia. For the first experiment, 0.1077 g was used, and for the second, 0.1058 g. However, for the control in the first experiment, pure ammonia was supplied in the morning and evening.

3. Results and Discussion

According to the LTNA data, the specific surface area of the supports amounted to 71–81 m²/g (

Table 1. The main characteristics of mixed oxides Ce_1-xZr_xO₂.

Support	SBET, m2/g	DXRD, nm	UCP, nm	DTEM, nm
Ce0.75Zr0.25O2	87	10.5	0.5397	8.2
Ce0.5Zr0.5O2	81	7.5	0.5282	7.4
Ce0.4Zr0.5Y0.05La0.05O2	71	6.8	0.5261	6.4

). After ruthenium loading, the specific surface area changed insignificantly (

Table 2. The main characteristics of ruthenium catalysts.

Catalyst	Ru Content *, wt.%	SBET, m2/g	UCPsup, nm	Particle Size, nm
				Support		Ruthenium
				XRD	TEM	TEM	CO Chemisorption
Ru/Ce0.75Zr0.25O2	2.61	89	0.5395	10.8	8.2	-	2.6
Ru/Ce0.5Zr0.5O2	3.32	80	0.5285	7.6	7.4	1.3	2.0
Ru/Ce0.4Zr0.5Y0.05La0.05O2	2.26	84	0.5256	6.8	6.4	1.2	1.6

* According to X-ray fluorescence analysis.

).

According to the XRD data, the support material is a mixed oxide with a fluorite-type structure (Figure S1). Table 1 presents the values of the unit cell parameter (UCP) and CSR size for mixed oxides Ce_1−xZr_xO₂. In each case, the UCP value is lower than that of unmodified CeO₂ (a = 5.411Å) [48], which proves the inclusion of zirconium cations (of a smaller radius) into the CeO₂ structure with the formation of a substitutional solid solution. It is known that the introduction of Y³⁺ and La³⁺ ions into the lattice of the CeO₂–ZrO₂ solid solution has a positive effect on the redox properties and oxygen capacity owing to the formation of oxygen vacancies in the initial oxide [43,49]. The XRD patterns of the supports (Figure S1) contain no individual peaks of pure oxides (Ce, Zr, Y, La) because the oxides of rare earth elements completely dissolve in the cerium oxide lattice with the formation of a solid solution.

Table 2 presents the UCP values of the mixed oxide Ce_1−xZr_xO₂, and the CSR sizes for ruthenium catalysts. The XRD diffraction patterns of all catalysts (Figure S2) have no peaks corresponding to ruthenium, which means that ruthenium particles present in a highly dispersed state. Since the measurement accuracy of the UCP and CSR values varies, it seems reasonable to conclude that the catalyst preparation procedure had no effect on the support characteristics.

According to the TEM data, ruthenium particles in the Ru/Ce_0.5Zr_0.5O₂ and Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂ catalysts are uniformly distributed over the support surface (Figure 2a–c). However, the TEM image of the Ru/Ce_0.75Zr_0.25O₂ catalyst showed no ruthenium particles on the support surface. Most likely, ruthenium particles in the latter sample occur in a highly dispersed state, and therefore possess insufficient contrast for visualization in a TEM image. Furthermore, ruthenium can occur in a form of oxidized Ru4+ species which are also invisible in a TEM image due to low contrast ability. Figure 2 also presents the size distribution of the support particles for all catalysts and that of ruthenium particles for Ru/Ce_0.5Zr_0.5O₂ and Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂. It can be seen that the average particle size of the support was 6.4 nm for Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂, 7.5 nm for Ru/Ce_0.5Zr_0.5O₂, and 8.2 nm for Ru/Ce_0.75Zr_0.25O₂ (Table 2). The average particle size of ruthenium was 1.3 nm in Ru/Ce_0.5Zr_0.5O₂ and 1.2 nm in Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂ (Table 2). It should be noted that ruthenium particles of similar size were also obtained in other supports. For example, as reported in [50], a catalyst with ruthenium particles of size 1–2 nm was obtained by impregnating an La₂O₂CO₃–Al₂O₃ support. In our earlier works, we obtained 3–5 nm ruthenium particles on carbon supports by the impregnation method [44]; with decreasing ruthenium content, the particle size decreased to 2–3 nm [51].

In the present work, the dispersion of ruthenium particles was also determined by the method of CO pulse chemisorptions and compared with the respective TEM data (Table 2).

It can be seen that the high dispersion of ruthenium is confirmed by both methods. In contrast to TEM, the CO chemisorption method allowed the determining ruthenium particle size in Ru/Ce_0.75Zr_0.25O₂, which equaled 2.6 nm. This observation is associated with the reduction of ruthenium from the oxidized state under experimental conditions.

The supports and catalysts were studied by the IR spectroscopic method. The IR spectra of the supports demonstrate intensive absorption bands in the range of 373–380 cm⁻¹, which is characteristic for vibrations of the F_1u symmetry in cubic CeO₂ [50,51] (Figure 3, spectra 1–3). In addition, broad low-intensive absorption bands were observed in the range of 560–620 and 720–740 cm^–1, which are attributed to vibrations of the F_1u symmetry in cubic ZrO₂ [52]. Thus, in the studied supports, both CeO₂ and ZrO₂ maintain their cubic structure, in good agreement with the XRD data (Figure S2, Table 1).

After depositing the Ru active component, the vibrational frequency of CeO₂ shifts to the high-frequency region by 5–22 cm⁻¹ (Figure 3, spectra 4–6). A similar increase in the vibrational frequency was observed earlier with the introduction of 7 mol.% Ho³⁺ ions in CeO₂ [52].

It should be noted that it is hardly feasible to reveal the structure of RuO_x compounds by IR spectroscopy, because some of the compounds (for example, RuO₂) have no absorption bands in the 250–1000 cm⁻¹ range. Therefore, auxiliary physicochemical methods should be applied to elucidate the structure of ruthenium compounds.

The XPS method was used to determine the elemental composition of the catalyst surface, and the element chemical states. According to the XPS data, the surface of all catalysts contained the following elements: Ru, Ce, Zr, C, and O (Figure S3).

The XPS spectra of all samples were decomposed in the C 1s and Ru 3D spectral regions and the contribution from the C 1s signal was subtracted. The obtained spectra were a combination of the main Ru 3d_5/2 doublet peak at 281.6 ± 0.1 eV and the satellite peak at E_b(Ru 3d_5/2) ~ 283 eV (Figure 4). The maximum of the Ru 3p_3/2 peak in all cases was close to 463.4 eV. Based on the literature data, the observed spectral characteristics (E_b(Ru 3d_5/2) ~ 281.6 eV and E_b(Ru 3p_3/2) ~ 463.4 эB) were predominantly assigned to the Ru⁴⁺ state [53,54]. The satellite peak at E_b(Ru 3d_5/2) ~282.5–283 eV was ascribed to the excitation of plasmons from RuO₂-type structures [55], or to the effect of ruthenium cations final state in the rutile structure [56]. Furthermore, the presence of the satellite peak and its intensity can be determined by the crystallization degree of the RuO₂ particles, and by the hydration degree of their surface [53,57].

The presence of ruthenium on the catalyst surface, predominantly in an oxidized state, as proved by the XPS data, is most likely explained by the formation of an oxide film resulting from the contact of pre-reduced active component particles with atmospheric air.

For all the catalysts, the Ce 3D spectral regions were decomposed into peaks corresponding to the Ce³⁺ and Ce⁴⁺ states. It was found that in all samples, the content of the Ce³⁺ species on the support surface was almost the same and equaled 21 ± 3%; cerium in the Ce⁴⁺ state dominated, in good agreement with the literature data [58,59].

The Zr 3D spectra of all samples were correctly described by a single doublet component with the Zr 3d_5/2 binding energy at ~ 181.9 eV, corresponding to a Zr⁴⁺ state similar to that in the ZrO₂ composition (Figure 5) [46,60].

In the Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂ catalyst, the elements La and Y were not revealed by the XPS method (Figure S4), but both elements were detected in the catalyst by chemical analysis (Table S2). This observation is most likely explained by their localization in the bulk and low content on the support surface.

The catalysts were tested in the process of ammonia decomposition. The results of catalytic tests conducted at various ammonia flow rates and temperatures (400, 440, and 500 °C) are presented in Table S2. Figure 6 illustrates the results of catalyst testing at an ammonia flow rate of 100 mL/min. It can be seen that with increasing temperature, the conversion of ammonia over all catalysts increases significantly. At 500 °C, the ammonia conversion over catalysts Ru/Ce_0.5Zr_0.5O₂ and Ru/Ce_0.75Zr_0.25O₂ is practically the same, but the latter catalyst demonstrates much higher hydrogen productivity (53.3 mmol H₂/min·g_cat). In addition, compared to other catalysts, this sample showed a sharper increase in activity with increasing temperature, i.e., it has the maximum activation energy. Note also that the specific activity of this catalyst exceeded the activity of the known catalyst synthesized in [61].

The kinetics of ammonia decomposition over the Ru/Ce_0.5Zr_0.5O₂ catalyst were studied as a function of the hydrogen content in the reaction mixture formed by mixing the reagent flows to provide a total flow rate of the mixture of 100 mL/min. The hydrogen concentration was varied in the range of 5–90 vol.% by changing its flow rate from 5 mL/min to 90 mL/min at a reaction temperature of 490 °C. At a reaction temperature of 410 °C, the hydrogen flow rate was varied from 10 mL/min to 90 mL/min. At a constant ammonia content of 10 vol.%, this means that the H₂/NH₃ ratio was 0.5 ÷ 9.0. To maintain a constant flow rate of the reaction mixture, helium was introduced into the system (up to 100 mL/min). The results of catalytic tests, linearized using the Van ’t Hoff differential method, are shown in Figure 7a.

As Figure 7a shows, when the hydrogen content in the reaction mixture increases from 5 to 55 vol.% (corresponding lnC(H₂) interval is 1.5 to 4), the reaction rate decreases linearly (at 490 °C). The calculation shows that in this H₂ concentration interval, the apparent reaction order with respect to hydrogen is −0.27. When the hydrogen content in the reaction mixture varies from 65 to 90 vol.% (corresponding lnC(H₂) interval is 4–4.5), the reaction rate decreases much more sharply, and the apparent reaction order with respect to hydrogen approaches −2.59. Obviously, Ru/Ce_0.5Zr_0.5O₂, similarly to other ruthenium catalysts, is susceptible to the hydrogen inhibition effect described in [62]. In the literature, the effect of hydrogen inhibition is explained by hydrogen displacement from the catalyst surface of nitrogen-containing transitional forms (N, NH, NH₂), which have a lower adsorption heat than that of hydrogen. The concentration of these nitrogen species is an important factor for both decomposition and synthesis of ammonia; thus, the effect of hydrogen inhibition is significant for both forward and reverse reactions. In the range of 50–100 vol.% hydrogen concentrations (Figure 7a, 410 °C, purple dots), the apparent reaction order with respect to hydrogen is −0.53.

The effect of ammonia content in the reaction mixture on the catalytic properties of the Ru/Ce_0.5Zr_0.5O₂ was studied in the range of 10–100 vol.%. inlet NH₃ concentrations. The results of catalytic tests, linearized using the Van ’t Hoff differential method, are shown in Figure 7b. Indeed, in the 10–50 vol.% NH₃ concentration range (corresponding lnC(NH₃) interval is 2–4 (490 °C)), the reaction rate increases proportionally to increasing ammonia content. As the ammonia concentration increases within 50–100 vol.% (490 °C) (corresponding lnC(NH₃) range is 4 to 4.5), the apparent reaction order with respect to ammonia decreases to 0.25. The observed point, in which the character of the dependence alters (the inflection point), corresponds to the reaction mixture composition of 50 vol.% NH₃, 50 vol.% He.

At the reaction temperature of 410 °C, no change in the apparent reaction order is indicated; we can instead discuss the common order of the reaction in the studied NH₃ concentration interval, equal to 0.13. The reason for this catalyst behavior at 410 °C is most likely related to the rearrangement of the catalyst active centers owing to the influence of the support. This reminds that cerium–zirconium mixed oxides are phases characterized by high oxygen lability, which can transit to an oxygen-lean state in a reducing medium, such as that containing ammonia and hydrogen. The oxygen deficit most likely promotes the formation of a partial positive charge in the support phase that affects the character of its electronic interaction with ruthenium particles. Probably, depending on the temperature, the surface oxygen deficit varies, thus altering the electronic state of the active component in contact with the reaction medium.

In general, it can be assumed that the observed dependence results from the competitive adsorption of ammonia and hydrogen on the ruthenium surface. The concentration ratio of the released H₂ and undecomposed NH₃ in the gas mixture exceeds 0.8 at the inlet reaction mixture composition of 50 vol.% NH₃ and 50 vol.% He (inflection point at 490 °C). At high concentrations of ammonia in a mixture with helium, its conversion does not exceed 30%, and the NH₃-to-hydrogen concentration ratio falls below 0.5. As the NH₃ concentration decreases, its conversion increases and the hydrogen concentration also increases. As a result, the competition between hydrogen and ammonia for the active centers of the catalyst becomes more profound.

The observed dependencies are characteristic for the competitive adsorption of reagents on the catalyst surface. A similar situation was observed in our recent study [63] on acetylene hydrogenation to ethylene over the Pd-based catalyst: the competitive adsorption of acetylene and hydrogen was proved by the inflection point on the curve of the catalyst activity depending on the acetylene concentration in the reaction mixture.

Temperature is an important factor that significantly affects both the catalyst activity and the reaction product distribution. Figure 8 presents the results of studying the influence of temperature increasing from 350 to 510 °C with a step of 20 °C on the catalytic performance of the properties of Ru/Ce_0.5Zr_0.5O₂.

Obviously, after linearization, all points fit on a straight line. The apparent activation energy E_a calculated by the Arrhenius equation equals ~90 kJ/mol.

The activation energy value complies with that reported in the literature for ruthenium catalysts on oxide supports [62,64,65,66].

Simulation of Experimental Data on Ammonia Decomposition over the Ru/Ce_0.5Zr_0.5O₂ Catalyst Using Kinetic Data Reported in the Literature

Strictly speaking, comparative analysis of the activity of various catalysts in terms of specific productivity in a flow reactor is correct only at a low conversion of the initial reagent that ensures an insignificant concentration gradient along the reactor length. Otherwise, if the apparent reaction order is positive with respect to the inlet reagent (ammonia in our case) and negative with respect to the product (hydrogen), the apparent productivity for any catalyst will noticeably decrease with increasing reagent (ammonia) conversion. To overcome this problem, it seems reasonable to compare different catalysts at similar conversions (see Table 2), but this approach is hardly feasible experimentally, especially when comparison with the literature data is needed.

An alternative approach to compare the catalysts’ activity is based on mathematical modeling of the reactor in which the tests were performed, using global kinetics. Note that for the studied catalyst, the kinetics themselves may still be unknown. However, we can try to model the experimental data on the ammonia conversion for a particular catalyst using the kinetic data obtained by other authors for other catalysts (with a similar active component and support) and, thus, compare this catalyst in terms of volumetric activity with other catalysts with already known kinetic data.

This approach was applied to analyze the activity of one of the catalysts studied in this work, namely, Ru/Ce_0.5Zr_0.5O₂. Unfortunately, there are very few published kinetic data appropriate for simulating a fixed-bed flow reactor for catalysts of this composition. We were able to find only two relatively close examples, which are briefly considered below.

The authors of [64] used a flow differential fixed-bed flow reactor and a 3%LiOH/1%Ru/α-Al₂O₃ catalyst in the temperature range of 350–475 °C and obtained a Temkin–Pyzhev-like power-law expression with reversibility term:

$$r_{N{H}_3}=\frac{k_0 \exp \left(\frac{-E_A}{R T}\right) {\left(P_{N{H}_3}\right)}^{ 0.24}}{{\left(P_{H_2}\right)}^{ 0.54}} \left(1-{\left[\frac{{\left(P_{H_2}\right)}^{ 1.5} {\left(P_{N_2}\right)}^{ 0.5}}{K_{eq} P_{N{H}_3}}\right]}^2\right)$$

(1)

with $E_A$ = 82.6 kJ/mol and $k_0$ = 36.18 mol(NH₃)*kPa^0.3/g_cat/s.

In [24], using a fixed-bed flow reactor for x%Ni-y%Ru/CeO₂ bimetallic catalysts at x = 1.5–10 wt.% and y = 0.3–2 wt.%, the authors developed the Langmuir–Hinshelwood– Hougen–Watson (LHHW) kinetic model and obtained the best agreement between the calculated and experimental results under the assumption that the rate-limiting stage is the dehydrogenation of the first hydrogen atom (the remaining stages were considered as quasi-stationary):

$$N{H}_3\left(ads\right)+[*]\rightleftarrows N{H}_2\left(ads\right)+H\left(ads\right)$$

(2)

The respective kinetic equation is as follows:

$$r_{N{H}_3}=\frac{k_0 \exp \left(\frac{-E_A}{R T}\right) K_{N{H}_3}{C}_{N{H}_3}}{{\left(1+K_{N{H}_3}{C}_{N{H}_3}+\sqrt{\frac{C_{H_2}}{K_{H_2}}} \right)}^2}$$

(3)

For the 5%Ni-1%Ru/CeO₂ catalyst, the authors of [24] assumed the following kinetic parameters: $E_A$ = 124 kJ/mol, $k_0$ = 3.9 × 106 mol(NH₃)/g_cat/s, $K_{N{H}_3}$ = 1.1 m³/mol, and $K_{H_2}$ = 4.7 × 10⁻² mol/m³.

Calculations were performed according to an isothermal model of a plug–flow reactor using kinetic expressions and parameters (1) and (3) under the operating conditions of experiments with the Ru/Ce_0.5Zr_0.5O₂ catalyst (layer bulk density ~1000 kg/m³): normal pressure, inlet composition—NH₃ 100%, ammonia flow rate—100 mL(gas)/min, and GHSV—296,000 h⁻¹. For the calculations, we used the in-home-made program code. The calculations were performed in the temperature range of 350–500 °C with a step of 10 °C. A correction multiplier for the pre-exponential factor was selected in order to provide the coincidence of the calculated ammonia conversion at 450 °C with the experimental value of 10.5% for both kinetic Equations (1) and (3).

Figure 9 compares the calculated and experimental results for ammonia conversion depending on the temperature. Curve 1 was obtained by 10-fold increasing the pre-exponential factor for kinetic Equation (1). Curve 2 was obtained by 3-fold increasing the kinetic Equation (3). That is, at 450 °C, our Ru/Ce_0,5Zr_0,5O₂ catalyst is an order of magnitude more active than the 3%LiOH/1%Ru/α-Al₂O₃ catalyst (which, in turn, is close in activity to unpromoted Ru/α-Al₂O₃ catalysts [64]), and three times more active than the 5%Ni-1%Ru/CeO₂ catalyst. Taking into account that the ruthenium content in the latter catalyst is 2–3 times lower, it seems reasonable to conclude that our Ru/Ce_0.5Zr_0.5O₂ catalyst has almost as high activity per 1 g Ru as that of the 5%Ni-1%Ru/CeO₂ [24], and exceeds it 3-fold in terms of volumetric activity.

A comparison table of our catalysts with other catalysts, as well as correlations between the physical and catalytic properties of our catalysts are given in additional materials. Additional materials also contain references [67,68,69,70,71,72,73,74,75,76,77,78,79] and explanations to the comparison table and correlations.

4. Conclusions

• The IR spectroscopy and XRD data proved the formation of solid CeO₂–ZrO₂ solutions in the supported ruthenium catalysts. With increasing zirconium content, the crystallite size decreases and equals 8.2 nm for Ru/Ce_0.75Zr_0.25O₂, 7.5 nm for Ru/Ce_0.5Zr_0.5O₂, and 6.4 nm for Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂. The synthesized catalysts are characterized by high dispersion of the active component. According to the TEM data, the ruthenium particle size in Ru/Ce_0.5Zr_0.5O₂ and Ru/Ce_0.4Zr_0.5Y_0.05La_0.05O₂ was 1.3 and 1.2 nm, respectively.
• According to the XPS data, ruthenium on the catalyst surface occurs predominantly in the Ru⁴⁺ state, most probably due to its oxidation at contact with air. Cerium in all samples presents predominantly in the Ce⁴⁺ state.
• Ruthenium catalysts were tested in the process of ammonia decomposition. The best activity showed the catalyst in support with the highest content of cerium (Ru/Ce_0.75Zr_0.25O₂). It provided a hydrogen productivity of 53.3 mmol H₂/min•g_cat at 500 °C and retained high activity during 48 h testing in the ammonia decomposition process.
• The kinetics of NH₃ decomposition were calculated using the Temkin–Pyzhov exponential expression. The mathematic model describes well with the experimental data.

5. Notations

• W is the specific rate of reaction, mmol(H₂)/(g_cat min);
• $r_{N{H}_3}$ is the specific rate of reaction, mol(NH₃)/g_cat/s;
• $C_{N{H}_3}$ and $C_{H_2}$ are the ammonia and hydrogen concentrations, respectively, mol/m³;
• $P_{N{H}_3}$, $P_{H_2}$ and $P_{N_2}$ are the ammonia, hydrogen, and nitrogen partial pressure bars, respectively;
• $k_0$ is the pre-exponential factor;
• $K_{eq}$ is the temperature-dependent equilibrium constant;
• $E_A$ is the activation energy;
• $K_{N{H}_3}$ is the ammonia adsorption constant;
• $K_{H_2}$ is the hydrogen desorption constant.