Research on Catalysts for Online Ammonia Hydrogen Production in Marine Engines: Performance Evaluation and Reaction Kinetic Modeling

Abstract

One viable technical approach for achieving hydrogen-blended combustion in marine ammonia-fueled engines is to utilize online ammonia decomposition to produce hydrogen, which is then introduced into the engine for combustion. This work carried out ammonia decomposition experiments using various catalysts, examining the effects of temperature and space velocity on Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts. Based on the experimental data obtained, the kinetic parameters of ammonia decomposition were fitted using four different models: mass action law, first-order reaction, Langmuir, and Temkin–Pyzhev kinetics across two catalysts, with the subsequent mechanistic analysis of catalytic reaction processes within the reactor. The results revealed that the NH₃ conversion rate of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst was superior to that of the Ni/Ce_0.36Zr_0.64O₂ catalyst, with temperature activity windows of 250–450 °C and 400–600 °C, respectively. Within the range of 2000–32,000 mL·g⁻¹·h⁻¹), an increase in space velocity led to a decrease in NH₃ conversion rate by approximately half. All four models were able to predict NH₃ conversion rates for the different catalysts with reasonable accuracy. The activation energies for Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts were found to be 37.7 kJ·mol⁻¹ and 66 kJ·mol⁻¹, respectively. Targeting hydrogen requirements of 10–40% vol for ammonia engines, the corresponding catalytic temperatures for Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ were above 267 °C and 500 °C, respectively.

1. Introduction

Currently, the greenhouse effect has emerged as one of the major climate issues. For the shipping industry, the mainstream international approaches to carbon reduction in ships include overall vessel energy efficiency improvement, the optimization of ship power systems, the adoption of alternative marine fuels, and carbon capture and storage technologies [1,2]. Among these, the adoption of alternative marine fuels is the ultimate direction for achieving long-term carbon reduction goals in shipping. Among zero-carbon fuels, ammonia (NH₃) has a high energy density (10,800 kJ·kg⁻¹), mature production technology, and is easily liquefied (298 K, 1 MPa) with established transportation technology. Its complete combustion products are water and nitrogen, while the NOx in incomplete combustion products can be removed through Selective Catalytic Reduction [3,4]. The use of ammonia as a propulsion fuel on ships holds great promise [3,4].

Ammonia, when used as an internal combustion engine fuel, exhibits disadvantages such as slow combustion speed, low combustion temperature, high ignition temperature, and high ignition energy [5,6,7]. Igniting ammonia with other highly reactive fuels can significantly enhance its combustion process. Among zero-carbon fuels, hydrogen (H₂) can significantly enhance the combustion process in ammonia-fueled engines. When the hydrogen blending ratio is between 10 and 40 vol%, it notably improves thermal efficiency and engine performance [8]. However, the low volumetric energy density of hydrogen necessitates its storage at an extremely low temperature or very high pressure. As a result, the transportation and storage costs of hydrogen fuel are significantly higher than those of ammonia and other mainstream fuels. Ammonia can act as a hydrogen carrier, producing hydrogen through a decomposition reaction, as shown in Equation (1). Ammonia has advantages such as high hydrogen storage capacity (17.6 wt%), low gas production cost, and no CO₂ emission upon decomposition, making it considered an ideal hydrogen storage carrier [9,10,11,12]. Therefore, utilizing ammonia as a hydrogen storage carrier and generating hydrogen through its decomposition to blend with ammonia-fueled engines represents a valuable technical approach to enhance ammonia fuel combustion.

$$2\mathrm{N}{\mathrm{H}}_3={\mathrm{N}}_2+3{\mathrm{H}}_2,\mathsf{\Delta }\mathrm{H}(298 \mathrm{K})=46.22\mathrm{kJ}\text{·}{\mathrm{mol}}^{-1}$$

(1)

Ammonia decomposition is an endothermic reaction that begins to occur significantly above 800 °C without the presence of a catalyst, reaching a decomposition rate of over 99% above 1000 °C. A high reaction temperature leads to increased energy consumption, significantly reducing the efficiency of hydrogen production. Currently, the main methods for hydrogen production from ammonia include thermal catalysis [13], electrolysis [14], photocatalysis [15], plasma methods [16], electron beam methods [17], and radiation methods [18]. Among these, thermal catalysis is a mature technology with low energy consumption and relatively low equipment cost, making it the mainstream technology for ammonia decomposition to produce hydrogen [13]. The primary optimization direction is to use efficient catalysts to lower the temperature window for the ammonia decomposition reaction. To achieve the online catalytic decomposition of hydrogen in ammonia engines, utilizing exhaust energy to promote ammonia decomposition for hydrogen production can significantly enhance energy efficiency. However, the exhaust temperature of ammonia marine engines is around 400 °C, and, considering the heat loss in the reactor’s heat exchange process, the actual achievable temperature in the ammonia reactor would be lower [19]. Therefore, it is necessary to determine a suitable NH₃ decomposition catalyst based on actual exhaust temperature conditions and hydrogen blending needs, providing a reference for the selection of a hydrogen blending catalyst in ammonia-fueled engines.

Research on ammonia decomposition catalysts currently focuses primarily on Ru catalysts and non-precious-metal catalysts. For the experimental research, Yin et al. [20] tested the NH₃ conversion rate of a 4.8% Ru/CNTs catalyst under conditions of 450 °C and 30,000 mL·g⁻¹·h⁻¹, achieving a conversion rate of 43.4%. Lucentini et al. [21] investigated the effect of temperature on NH₃ conversion rate with a 5% Ru/CeO₂ catalyst. The results showed that, at 450 °C and a space velocity of 6000 mL·g⁻¹·h⁻¹, a maximum conversion rate of 98.0% could be achieved, and a synergistic effect was observed between the CeO₂ support and the active component Ru. Pinzón et al. [22] investigated the ammonia decomposition activity of Ru/SiC catalysts using a commercial β-SiC as a support and found that the Ru/SiC catalyst with a Ru mass fraction of 2.5% exhibited the best catalytic performance. With 5% NH-Ar as the feedstock, under conditions of 400 °C and 60,000 mL·g⁻¹·h⁻¹, the ammonia conversion rate reach 99%. Kocer et al. [23] investigated the ammonia decomposition activity of Ru/GA catalysts using graphene aerogel (GA) as a support and found that the Ru/GA catalyst with a Ru mass fraction of 13.6% achieved an ammonia conversion rate of 97.6% at 450 °C and a space velocity of 30,000 mL·g⁻¹·h⁻¹. A Ru-based catalyst is the most active ammonia decomposition catalyst to date. In recent years, achieving a synergistic effect between the support and Ru using basic metal oxides like CeO₂ has been one of the key measures to enhance their ammonia decomposition activity [21]. Non-noble-metal catalysts have a significant price advantage compared to Ru catalysts. Among non-precious metals, Ni-based catalysts were considered more effective for ammonia decomposition. Yan et al. [24] tested the NH₃ decomposition efficiency of a 43.8%Ni/Al₂O₃-CeO₂ catalyst. The results showed that, under 600 °C and 90,000 mL·g⁻¹·h⁻¹, a conversion efficiency of 85% was attained. Sato et al. [25] tested a 15%Ni/Mg-Al catalyst and investigated the impact of temperature on its NH₃ decomposition efficiency. The results showed that, at 500 °C and 3000 mL·g⁻¹·h⁻¹, the maximum decomposition efficiency was 59%. Hu et al. [26] studied the maximum NH₃ conversion efficiency of a 5%Ni/ZSM-5 catalyst under conditions of 500 °C and 30,000 mL·g⁻¹·h⁻¹, achieving a conversion efficiency of 41%. Yu et al. [27] investigated a 10% Ni/La₂O₃ catalyst and, under conditions of 550 °C and 30,000 mL·g⁻¹·h⁻¹, the maximum NH₃ conversion efficiency was 59%. Gu et al. [28] measured the NH₃ conversion efficiency of a 25% Ni/Al₂O₃ catalyst under conditions of 450 °C and 24,000 mL·g⁻¹·h⁻¹, and found a conversion efficiency of 29.2%. Although the aforementioned studies have shown the potential of different types of catalysts in ammonia decomposition reactions, it is difficult to assess their applicability in marine engine online ammonia-to-hydrogen systems utilizing exhaust energy due to differences in experimental conditions such as space velocity.

To gain deeper insights into the reasons for differences in decomposition efficiency among different catalysts, kinetic parameters and reaction processes can be investigated. In investigating the kinetic models of ammonia decomposition reactions, the performance characteristics of catalysts can be analyzed through reaction kinetic parameters such as pre-exponential factors and activation energy. Chellappa et al. [29] used a first-order reaction rate model for the Ni-Pt/Al₂O₃ catalyst and experimentally fitted NH₃ decomposition rate data under different temperature and space velocity conditions, obtaining activation energy and pre-exponential factor values of 49 kJ·mol⁻¹ and $1.309\times {10}^{12}$ s⁻¹, respectively. Hesam Maleki et al. [30], based on the Co_0.5Ce_0.1Al_0.4O(sa) catalyst, employed the Temkin–Pyzhev model to fit NH₃ catalytic conversion rate data under different temperatures and space velocities. The pre-exponential factor and activation energy obtained were 1.22 × 10⁷ mol²·L⁻²·s⁻¹ and 128 kJ·mol⁻¹, respectively, with an R² of 0.995 for the model. For the simulation study of reaction processes, Rei-Yu Chein et al. [31] calculated the material conversion process within the reactor using a first-order reaction rate model. Deshmukh et al. [32] analyzed the concentration changes in adsorbed substances and the NH₃ catalytic decomposition reaction pathways under different conditions based on a first-order reaction kinetics model with a Ru/Al₂O₃ catalyst. In summary, there are various models for calculating the reaction rate of ammonia decomposition for hydrogen production. However, further verification and analysis are necessary for the reaction rate calculation models and their performance in the reaction process under different catalysts at marine engine exhaust temperatures.

This work proposes a reaction kinetics approach that combines experimental and numerical simulations to study the performance of ammonia decomposition catalysts and the reaction process of ammonia decomposition. This method aims to provide appropriate catalyst selection and mechanistic support for the study of ammonia decomposition reforming hydrogen production systems in ammonia-fueled engines. The effects of reaction temperature and space velocity on the NH₃ conversion rate of the prepared catalysts determine the catalysts that can meet the hydrogen blending ratio (10–40 vol%) under engine exhaust temperature and flow rate conditions. Then, based on the experimental results, kinetic parameters of both catalysts were fitted using four kinetic rate expressions, and the performance differences in the experimental data were comparatively analyzed based on parameters such as the pre-exponential factors and activation energies of different catalysts. A reaction kinetics model was established based on parameters such as the pre-exponential factors and activation energies of catalysts aimed at accurately predicting the reaction time characteristics of component mole fractions and NH₃ conversion rates under marine engine online ammonia-to-hydrogen conditions. Then, by analyzing the kinetic processes of NH₃ consumption and H₂ generation at different temperatures, the distribution of reaction rates and component concentrations with respect to reaction time was revealed, further clarifying the kinetic characteristics of the reaction process.

2. Results and Discussion

2.1. Comparison of Conversion Rates Under Different Temperatures and Space Velocities

Figure 1 shows the experimental NH₃ conversion rates for Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts under different temperatures at a space velocity of 8000 mL·g⁻¹·h⁻¹. Overall, the NH₃ conversion rates of both catalysts increase with the temperature rise because ammonia decomposition is an endothermic reaction, and increasing the temperature favors the forward reaction. Regarding the temperature characteristics of different catalysts, it is evident that the conversion rate for the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst is significantly higher than that for the Ni/Ce_0.36Zr_0.64O₂ catalyst. The activation temperatures for the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts are approximately 250 °C and 450 °C, respectively. As the temperature increases, both reach their maximum NH₃ conversion rates at approximately 500 °C and 650 °C, respectively, at which point NH₃ is considered to have completely decomposed into N₂ and H₂. The temperature activity windows for the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts are 250–450 °C and 450–600 °C, respectively. Therefore, the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst exhibits significantly better catalytic performance for ammonia decomposition than the Ni/Ce_0.36Zr_0.64O₂ catalyst.

In this work, the selection of temperature and space velocity conditions was based on the ammonia flow rate and exhaust gas temperature during the operation of the ammonia engine. Specifically, the exhaust gas temperature ranged from 380 °C to 580 °C, with ammonia flow rates between 68 and 114 L·min⁻¹. At the maximum flow rate of 114 L·min⁻¹, the corresponding space velocity was 8000 mL∙g⁻¹·h⁻¹. Under these flow conditions, the waste heat from exhaust gases could preheat the ammonia to 260–470 °C. As shown in Figure 2, within this temperature range (260–470 °C), the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst achieved hydrogen blending ratios of 10–74%, while the Ni/Ce_0.36Zr_0.64O₂ catalyst only reached 0–7%. These results demonstrate that, under actual exhaust temperature conditions, the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst can meet the required hydrogen blending ratio specifications, whereas the Ni/Ce_0.36Zr_0.64O₂ catalyst fails to satisfy these requirements.

Figure 3 shows the NH₃ conversion rates for Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts at different weight hourly space velocities. It is observed that, as the space velocity increases, the NH₃ conversion rates for both catalysts gradually decrease. Additionally, the decrease in NH₃ conversion rates is more pronounced at higher space velocities. This is likely because higher space velocities mean that more gas flows through the catalyst bed, reducing the time gas molecules have to come into contact and react with the catalyst. The shortened residence time reduces the reaction time of the reactants, causing ammonia to not fully decompose before passing through the catalyst bed, thus leading to a decrease in NH₃ conversion rates. In this work, the Ru catalyst demonstrated superior performance with 91% ammonia decomposition efficiency at 470 °C and 16,000 mL·g⁻¹·h⁻¹, outperforming the 73% conversion rate reported by Huang et al. [33] for their Ru/La₂O₃ catalyst under comparable conditions.

2.2. Reaction Kinetic Parameters

To determine the reaction kinetic parameters of different catalysts, the kinetic parameters were fitted using various reaction rate models based on the experiment data presented in Section 2.1. Figure 4 illustrates the experimental and simulated NH₃ conversion rates for different catalysts at varying temperatures. In general, various models can accurately predict the NH₃ conversion rates of different catalysts at different temperatures. To further quantify the prediction accuracy of different reaction rate models for the NH₃ conversion, we calculated the R² between the results of different models and experiments, as shown in

Table 1. R² (Coefficient of Determination) of four different kinetic models over Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts.

	Ru/Ce0.33Zr0.58La0.03Nd0.03Pr0.03O2.09	Ni/Ce0.36Zr0.64O2
Model l	0.959	0.992
Model 2	0.968	0.993
Model 3	0.943	0.983
Model 4	0.939	0.991

. It can be observed that the experimental and simulated NH₃ conversion rates exhibit a strong linear correlation with R² values greater than 0.9, indicating that various reaction rate models can accurately predict the NH₃ conversion rates for different catalysts. Specifically, for the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst, the R² of the simulated results for various reaction rate models are in the following order: model 2 > model 1 > model 3 > model 4. For the Ni/Ce_0.36Zr_0.64O₂ catalysts, the R² for the various reaction rate models are in the following order: model 2 > model 1 > model 4 > model 3. Thus, it can be concluded that model 2 provides the best predictive performance of the NH₃ conversion rates for both catalysts. Furthermore, while the Temkin–Pyzhev model and nitrogen desorption model can theoretically describe surface adsorption/desorption processes, their complexity and specific applicability make them prone to overfitting or deviations from actual experimental conditions, often resulting in unsatisfactory fitting performance. In contrast, the first-order reaction model offers simpler mechanistic interpretation and demonstrates broader applicability for simulating the straightforward reaction pathway of ammonia decomposition.

In summary, the kinetic parameters obtained from model 2 better reflect the reaction kinetics characteristics of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts. Based on the kinetic parameter fitting results from model 2, the pre-exponential factors and activation energies of both catalysts are shown in

Table 2. Pre-exponential factors and activation energies for catalysts under model 2.

	Ru/Ce0.33Zr0.58La0.03Nd0.03Pr0.03O2.09	Ni/Ce0.36Zr0.64O2
Ea (kJ·mol−1)	37.7	66.0
A (s−1)	128.0	1621.2

. Comparing the kinetic parameters of both catalysts, it is evident that, although the pre-exponential factor of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst (128 s⁻¹) is an order of magnitude lower than that of the Ni/Ce_0.36Zr_0.64O₂ catalyst (1621.2 s⁻¹), its activation energy is as low as 37.7 kJ·mol⁻¹, about half that of the Ni/Ce_0.36Zr_0.64O₂ catalyst (66 kJ·mol⁻¹). This value closely matches the 34.7 kJ·mol⁻¹ reported by Huang et al. [33] for their Ru/La₂O₃ catalyst. This phenomenon indicates that, in heterogeneous catalytic reaction systems, a reduction in activation energy plays a decisive role in enhancing catalyst activity, whereas the pre-exponential factor, as a kinetic parameter related to collision frequency, typically has a secondary influence.

2.3. Analysis of the Catalytic Reaction Process

Figure 5 displays the time distribution of NH₃ and H₂ volumetric fractions in the reactor of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst at different temperatures, calculated via model 2. On the one hand, it can be observed that, over time, the NH₃ volumetric fraction gradually decreases, while the H₂ volumetric fraction gradually increases over time at different temperatures. This is because, as the NH₃ decomposition reaction progresses in the reactor, NH₃ is continually decomposed into N₂ and H₂. Simultaneously, as the reaction proceeds, the NH₃ concentration gradually decreases, leading to a decrease in reaction rate, which, in turn, reduces the NH₃ consumption rate and H₂ production rate. On the other hand, at the same reaction time, as the temperature increases, the NH₃ volumetric fraction decreases and the H₂ volumetric fraction gradually increases. According to Equation (3), this is because, as the temperature increases, the NH₃ decomposition reaction rate constant k increases, resulting in a higher reaction rate. Furthermore, it can be observed that the final H₂ volumetric fractions at different temperatures are approximately 17%, 40%, and 72%, respectively. At 465 °C, NH₃ is nearly completely decomposed, achieving the maximum H₂ production rate. According to the literature requirements for improving ammonia-fueled engines, a H₂ volumetric fraction of 10–40 vol% is needed [8], which can be achieved using the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst above 267 °C.

Figure 6 illustrates the variations in NH₃ consumption rate over reaction time for the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst at different temperatures. As the decomposition progresses, the NH₃ consumption rate steadily decreases. Furthermore, at 465 °C, the consumption rate varies more significantly over time, as the reactant concentration decreases more markedly at this temperature, as shown in Figure 5a.

Figure 7 displays the volume fraction distribution of NH₃ and H₂ in the reactor of the Ni/Ce_0.36Zr_0.64O₂ catalyst at different temperatures over reaction time, calculated using model 2. On the one hand, the change trends of NH₃ and H₂ volume fraction over reaction time and temperature are similar to those of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst, but the overall magnitude of variation is lower. This is because the activation energy of the Ni/Ce_0.36Zr_0.64O₂ catalyst is higher than that of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst, resulting in a lower reaction rate constant k under the same conditions.

Figure 8 displays the reaction rate distribution over reaction time for the Ni/Ce_0.36Zr_0.64O₂ catalyst at different temperatures. It can be observed that the trend of reaction rate changes is generally similar to Figure 6. However, due to the higher activation energy of the Ni/Ce_0.36Zr_0.64O₂ catalyst, its overall reaction rate is lower than that of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst. Additionally, the final volume fraction of H₂ produced at different temperatures is approximately 0.6%, 5%, 20%, and 57%, respectively. According to the literature, the required volumetric fraction of H₂ for improving ammonia-fueled engine combustion is 10–40vol% [8], indicating that the Ni/Ce_0.36Zr_0.64O₂ catalyst requires a temperature around 450 °C, which is much higher than the temperature requirements of the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst.

3. Experiment

3.1. Preparation of Catalysts

In this study, two ammonia decomposition catalysts for hydrogen production were designed and prepared, namely the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts.

For the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst, the active component Ru loading was 0.7 wt%, with the support material being a composite metal oxide. First, the composite metal oxide support was prepared using a co-precipitation method. Ce(NO₃)₃·6H₂O, ZrOCl₂·8H₂O, La(NO₃)₃·9H₂O, Nd(NO₃)₃·9H₂O, and Pr(NO₃)₃·9H₂O (5 mmol) were dissolved in anhydrous ethanol at a molar ratio of 33:58:3:3:3, and NH₃·H₂O was added dropwise at 50 °C until the solution’s pH reached ≥10, forming a precipitate. After washing the precipitate with deionized water, the sample was dried in a drying oven at 120 °C for 10 h, and then calcined at 500 °C for 5 h with a heating rate of 10 °C·min⁻¹. The resulting support was denoted as Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09. Next, Ru(NO₃)₃·6H₂O was used as the precursor for the Ru, and the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst was prepared using the impregnation method. A suitable amount of Ru(NO₃)₃·6H₂O was dissolved in deionized water and stirred vigorously for 1 h, and then transferred to deionized water containing the support oxide under vigorous stirring. The solution was stirred for 1 h, and then the solvent was evaporated at 70 °C. The solid residue was dried in a drying oven at 120 °C for 10 h and calcined at 500 °C for 5 h.

For the Ni/Ce_0.36Zr_0.64O₂ catalyst, the active component Ni loading was 5 wt%, with the support material being a composite metal oxide. First, the composite metal oxide support was prepared using a co-precipitation method. Ce(NO₃)₃·6H₂O and ZrOCl₂·8H₂O (5 mmol) were dissolved in anhydrous ethanol at a molar ratio of 36:64, and NH₃·H₂O was added dropwise at 50 °C until the solution’s pH reached ≥10, forming a precipitate. After washing the precipitate with deionized water, the sample was dried in a drying oven at 120 °C for 10 h, and then calcined at 500 °C for 5 h with a heating rate of 10 °C·min⁻¹. The resulting support was denoted as Ce_0.36Zr_0.64O₂. Next, Ni(NO₃)₂·6H₂O was used as the precursor for the Ni, and the Ni/Ce_0.36Zr_0.64O₂ catalyst was prepared using the impregnation method. A suitable amount of Ni(NO₃)₂·6H₂O was dissolved in deionized water and stirred vigorously for 1 h, and then transferred to deionized water containing the support oxide under vigorous stirring. The solution was stirred for 1 h, and then the solvent was evaporated at 70 °C. The solid residue was dried in a drying oven at 120 °C for 10 h and calcined at 500 °C for 5 h.

Prior to use, the catalyst powder was uniaxially pressed at 50 MPa for 60 s to form pellets. The compressed material was then ground and sieved to obtain 20–40 mesh (0.425–0.85 mm) particles. Finally, the catalyst was reduced under flowing ultrapure H₂ (99.999%) at 500 °C for 2 h with a controlled heating rate of 10 °C·min⁻¹.

3.2. Catalytic Experiments

In the catalyst performance tests, the catalyst activity was measured on the test bench shown in Figure 9. NH₃ (99.999%) flowed into an isothermal flow reactor (inner diameter = 8 mm) for the reaction, and the temperature was controlled by a furnace in the tubular reactor. The volumetric fractions of NH₃, N₂, and H₂ at the reactor outlet were measured using a gas chromatograph equipped with a TCD (Thermal Conductivity Detector). The catalytic performance was evaluated through two sets of experiments: temperature-dependent activity tests were conducted at a fixed space velocity of 8000 mL·g⁻¹·h⁻¹ under 1.01 bar, where the Ni/Ce_0.36Zr_0.64O₂ catalyst was tested from 201 to 650 °C and the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst from 244 to 513 °C, with a heating rate of 10 °C·min⁻¹ and 60 min isothermal holds at each temperature; space-velocity-dependent tests were performed across 2000–32,000 mL·g⁻¹·h⁻¹ at constant temperatures of 650 °C for Ni/Ce_0.36Zr_0.64O₂ and 470 °C for Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalysts, with 30 min stabilization at each space velocity. The ammonia conversion rate can be calculated from the ammonia fraction in the products [33,34], as shown in Equation (2):

$$\mathsf{\eta }={\displaystyle \frac{1-\mathsf{\alpha }}{1+\mathsf{\alpha }}}\times 100\%$$

(2)

where $\mathsf{\eta }$ represents the ammonia conversion(%) and $\mathsf{\alpha }$ is the ammonia volume fraction(%).

4. Kinetic Model

As shown in Equation (1), the ammonia decomposition reaction leads to significant volume expansion. In the catalytic performance test reactor, the small reactor diameter ensures relatively uniform radial spatial distribution. Based on reactor parameters including gas flow rate and reactor volume, we used COMSOL Multiphysics 6.1 to convert the axial spatial distribution of the continuous reactor into the temporal distribution of a zero-dimensional batch reactor model with constant volume, as shown in Equation (3). The reaction rate is defined in Equation (4). The simulation conditions are shown in

Table 3. Simulation conditions.

Simulation Conditions
Residence time, $\mathsf{\tau }$(s)	0.327
Catalyst volume, Vcat (cm3)	0.727
Catalyst density, ${\mathsf{\rho }}_{\mathrm{c}\mathrm{a}\mathrm{t}}$ (g∙cm−3)	1.375
Pressure, p (bar)	1.01
Inlet gas composition, α (%)	${\mathsf{\alpha }}_{{\mathrm{N}\mathrm{H}}_3}$ = 99.98%, ${\mathsf{\alpha }}_{{\mathrm{N}}_2}$ = 0.01%, ${\mathsf{\alpha }}_{{\mathrm{H}}_2}$ = 0.01%
Temperature, T (°C)	240–650

. In this way, this study can obtain the distribution of reaction rate and species concentration with residence time through numerical simulation, which can reflect the spatial distribution characteristics of reaction rate and species concentration along the flow direction in the actual reaction process:

$$\mathsf{\tau }={\displaystyle \frac{{\mathrm{V}}_{\mathrm{c}\mathrm{a}\mathrm{t}}}{\mathrm{Q}}}={\displaystyle \frac{{\mathrm{w}}_{\mathrm{c}\mathrm{a}\mathrm{t}}}{{\mathsf{\rho }}_{\mathrm{c}\mathrm{a}\mathrm{t}}\mathrm{Q}}}$$

(3)

where $\mathrm{Q}$ is the ammonia flow rate(mL∙s⁻¹) and ${\mathrm{V}}_{\mathrm{cat}}$, ${\mathrm{w}}_{\mathrm{cat}}, \mathrm{a}\mathrm{n}\mathrm{d} {\mathsf{\rho }}_{\mathrm{cat}}$ represent the catalyst volume(cm³), catalyst weight(g), and catalyst density(g∙cm⁻³), respectively:

$${\mathrm{r}}_{{\mathrm{N}\mathrm{H}}_3}={\displaystyle \frac{\mathrm{d}{\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_3}}{\mathrm{d}\mathrm{t}}}$$

(4)

where ${\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_3}$ is the NH₃ concentration (mol∙m⁻³) and ${\mathrm{r}}_{{\mathrm{N}\mathrm{H}}_3}$ is the reaction rate (mol∙m⁻³∙s⁻¹).

In the numerical simulation process, this study first fitted the kinetic parameters of the ammonia decomposition for the hydrogen production reaction process for different catalysts based on the experimental data obtained, and performed a comparative analysis of the accuracy differences among various ammonia decomposition reaction rate models. Building on this, numerical simulations were conducted, combined with reaction process analysis, to investigate the evolution of reaction rates and species concentrations during the ammonia decomposition for the hydrogen production process.

For the calculation of the NH₃ catalytic decomposition reaction rate, current research commonly employs a first-order reaction model and the Temkin–Pyzhev model [29,30]. This work incorporates the mass action law model and the Langmuir model for comparison. The ammonia decomposition reaction for the four reaction rate kinetic models is conducted under low-pressure conditions, with reverse reactions neglected. In this study, the law of conservation of mass model, first-order reaction model, Langmuir model, and Temkin–Pyzhev model are designated as model 1–4, respectively. To obtain accurate catalyst kinetic parameters, this study employs these four models for the parameter fitting of catalysts, identifies the model that can accurately predict the reaction results based on the fitting parameters, and further determines the reasonable kinetic parameters.

4.1. Law of Conservation of Mass Model

The law of mass action applies to simple homogeneous reactions. When considering multi-step reactions, the reaction order is determined directly based on the stoichiometric numbers of the reactants involved. In the ammonia decomposition reaction, there are distinct rate-controlling steps [19,35,36,37,38]. The reaction rate of these steps can be described using the law of mass action. By using equilibrium or steady-state approximations, the concentration of intermediate products can be expressed as a function of reactants or products, thereby deriving the expression for the overall reaction rate [33,34]. In the model based on the law of conservation of mass, the reaction rate of ammonia decomposition is proportional to the square of the concentration of the reactant, and the form of its reaction rate expression is consistent with that of a second-order reaction:

$${\mathrm{r}}_{\mathrm{N}{\mathrm{H}}_3}=\mathrm{k}{\mathrm{C}}_{\mathrm{N}{\mathrm{H}}_3}^2=\mathrm{A}\exp ({\displaystyle \frac{-{\mathrm{E}}_{\mathrm{a}}}{\mathrm{R}\mathrm{T}}}){\mathrm{C}}_{\mathrm{N}{\mathrm{H}}_3}^2$$

(5)

where ${\mathrm{r}}_{\mathrm{N}{\mathrm{H}}_3}$ represents the ammonia decomposition reaction rate (mol∙m⁻³∙s⁻¹), $\mathrm{k}$ is the reaction rate constant (mol⁻¹∙m³∙s⁻¹), ${\mathrm{C}}_{\mathrm{N}{\mathrm{H}}_3}$ is the ammonia concentration (mol·m⁻³), $\mathrm{A}$, ${\mathrm{E}}_{\mathrm{a}}$, $\mathrm{R},$ and T are the pre-exponential factor (mol⁻¹∙m³∙s⁻¹), activation energy (kJ·mol⁻¹), molar gas constant, and the reaction temperature (K).

4.2. First-Order Reaction Model

When the rate of ammonia decomposition is solely dependent on the concentration of ammonia, without significant limitations from adsorption, surface reactions, and other factors leading to complex kinetics, the first-order reaction model for ammonia decomposition can be adopted. The first-order reaction model is more applicable under high-temperature and high-space-velocity conditions, as these conditions may diminish the effects of complex surface phenomena such as adsorption and desorption. The expression for the catalytic reaction rate of NH₃ is shown in Equation (6).

$${\mathrm{r}}_{\mathrm{N}{\mathrm{H}}_3}=\mathrm{k}{\mathrm{C}}_{\mathrm{N}{\mathrm{H}}_3}=\mathrm{A}\exp ({\displaystyle \frac{-{\mathrm{E}}_{\mathrm{a}}}{\mathrm{R}\mathrm{T}}}){\mathrm{C}}_{\mathrm{N}{\mathrm{H}}_3}$$

(6)

4.3. Langmuir Model

The Langmuir adsorption isotherm equation is used to describe the adsorption behavior of ammonia on the catalyst's surface [39,40]. When the desorption of N is the rate-determining step in the ammonia catalytic decomposition reaction [19,35,36,37,38], and the influence of hydrogen concentration from the reverse reaction is disregarded, the ammonia decomposition reaction rate is influenced by the surface coverage of N atoms on the catalyst and can be determined using Equation (7):

$${\mathrm{r}}_{\mathrm{N}{\mathrm{H}}_3}=\mathrm{k}{\mathsf{\theta }}_{\mathrm{N}}=\mathrm{A}\exp ({\displaystyle \frac{-{\mathrm{E}}_{\mathrm{a}}}{\mathrm{R}\mathrm{T}}}){\displaystyle \frac{\mathrm{K}{\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_3}}{1+\mathrm{K}{\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_3}}}$$

(7)

where ${\mathsf{\theta }}_{\mathrm{N}}$ represents the fractional coverage of N* adsorbates on the catalyst and $\mathrm{K}$ is the adsorption equilibrium constant of N on the surface of the catalyst.

4.4. Temkin–Pyzhev Model

The Temkin–Pyzhev model is suitable for describing scenarios with low temperatures or significant surface limitations where adsorption is the rate-controlling step. In this case, the ammonia decomposition reaction rate depends on the concentrations of NH₃ and H₂ [19], with the reverse reaction being neglected. The reaction rate expression is provided in Equation (8):

$${\mathrm{r}}_{\mathrm{N}{\mathrm{H}}_3}=\mathrm{k}({\displaystyle \frac{{\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_3}^2}{{\mathrm{C}}_{{\mathrm{H}}_2}^3}}{)}^{\mathsf{\beta }}=\mathrm{A}\exp ({\displaystyle \frac{-{\mathrm{E}}_{\mathrm{a}}}{\mathrm{R}\mathrm{T}}})({\displaystyle \frac{{\mathrm{C}}_{{\mathrm{N}\mathrm{H}}_3}^2}{{\mathrm{C}}_{{\mathrm{H}}_2}^3}}{)}^{\mathsf{\beta }}$$

(8)

where $\mathsf{\beta }$ represents a fit parameter, $\mathsf{\beta }$ = 0.25, and ${\mathrm{C}}_{{\mathrm{H}}_2}$ is the concentration of hydrogen (mol·m⁻³).

5. Conclusions

This work studied the effects of temperature and space velocity on the NH₃ conversion rate of Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts through a catalyst performance evaluation experiment. Based on the experimental data obtained, the catalytic kinetic parameters of the two catalysts were fitted using four models: the law of mass action, first-order reaction, Langmuir, and the Temkin–Pyzhev model, and the reaction processes on the surface of the catalyst were analyzed. nder identical space velocity conditions, the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst demonstrates significantly higher NH₃ conversion rates than the Ni/Ce_0.36Zr_0.64O₂ catalyst across all tested temperatures. The active temperature windows were determined to be 250–450 °C for the Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst and 450–600 °C for the Ni/Ce_0.36Zr_0.64O₂ catalyst. Both catalysts showed similar responses to varying space velocities (2000–32,000 mL·g⁻¹·h⁻¹), with NH₃ conversion efficiency decreasing by approximately 50% across this range. The Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 catalyst emerged as the superior low-temperature option, achieving the required 10–40 vol% hydrogen blending ratio in marine engine exhaust conditions (260–470 °C), while the Ni catalyst failed to meet this specification.All four tested NH₃ decomposition rate models effectively predicted the temperature-dependent conversion characteristics (R² > 0.9) for all model–experiment comparisons). Among these, the first-order reaction model showed the highest accuracy for both Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts, offering broad applicability for simulating the straightforward ammonia decomposition pathway. Kinetic analysis revealed activation energies of 37.7 kJ·mol⁻¹ (Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09) and 66 kJ·mol⁻¹ (Ni/Ce_0.36Zr_0.64O₂), providing fundamental parameters for reactor design optimization in ammonia-fueled engines. The reaction process of Ru/Ce_0.33Zr_0.58La_0.03Nd_0.03Pr_0.03O_2.09 and Ni/Ce_0.36Zr_0.64O₂ catalysts are analyzed, clarifying the distribution of reaction rates and component concentrations over reaction time, and reflecting the kinetic behavior of reactants flowing axially in a practical tubular reactor. The establishment of the space–time equivalent relationship obtained from the zero-dimensional kinetic model developed in this work provides an important theoretical basis for translating laboratory-scale kinetic data into industrial reactor design.