A Review on Catalytic Hydrolysis of Ammonia Borane for Hydrogen Production

Abstract

Ammonia borane (NH₃BH₃, AB) is recognized as a promising hydrogen carrier due to its high hydrogen storage density (146 gL⁻¹, mass fraction 19.6%), safety, non-toxicity, and high chemical stability. The hydrolysis of AB has also become a research hotspot in recent years and offers a viable route for hydrogen production. However, the practical application of AB hydrolysis encounters substantial challenges, including undefined catalytic mechanisms, suboptimal catalytic performance, and intricate issues in AB regeneration. Thus, elucidating catalytic mechanisms, developing highly efficient catalysts, and exploring effective regeneration methods for NH₃BH₃ are critical and urgent. This paper delves into the catalytic hydrolysis process of AB, detailing the mechanisms involved, and simplifies the steps that affect AB hydrolysis activity into the adsorption, activation, dissociation of reactants, and the formation and desorption of H₂. It discusses the structural characteristics of metal catalysts used in recent studies, assessing their performance through metrics such as turnover frequency (TOF), activation energy (E_a), and reusability. On this basis, this paper conducts a relatively comprehensive analysis and summary of the strategies for optimizing the performance of AB hydrolysis catalysts, including three aspects, focusing on optimizing the number and dispersion of active centers, enhancing reactant adsorption and activation, and facilitating hydrogen desorption. In addition, it also addresses strategies for controlled hydrogen release during AB hydrolysis and methods for regenerating AB from spent solutions. Finally, corresponding conclusions and prospects are proposed, to provide a certain reference for the subsequent development of safe and efficient catalysts and research on the catalytic mechanism of AB hydrolysis.

1. Introduction

Fossil fuels, the main energy source now widely used, have effectively met the world’s growing energy needs. However, fossil fuel reserves are encountering escalating depletion problems. Simultaneously, the excessive utilization of fossil fuels has led to numerous environmental safety concerns [1,2]. Currently, various new energy sources have been developed as alternatives to fossil fuels, including solar energy, hydrogen energy, wind energy, nuclear energy, tidal energy, etc. Of particular note is hydrogen energy, which is non-toxic and odorless, whose only by-product after combustion is water. It is environmentally friendly and comes from a wide range of sources. In addition, it has features such as high energy concentration, storage capability, and ease of transportation. Therefore, hydrogen energy is an ideal alternative to traditional fossil fuels [3]. However, achieving high-quality high-density hydrogen energy storage and its rapid release under mild conditions remains a critical issue in promoting its application. Therefore, developing efficient, safe, and stable hydrogen storage materials is one of the foremost challenges in current research on hydrogen energy applications [4,5].

At present, there are four relatively mature hydrogen storage methods, namely high-pressure gaseous hydrogen storage, low-temperature liquefied hydrogen storage, solid material hydrogen storage, and amino and liquid organic hydrogen supports [6]. The first two methods fall under physical hydrogen storage and have relatively mature technology, but both have certain safety issues [7]. The latter two methods are classified as chemical hydrogen storage, utilizing hydrogen-containing chemicals to generate hydrogen gas in situ. Common chemical hydrogen storage materials mainly include metal hydrides (MgH₂, etc.), complex aluminum hydrides (LiAlH₄, etc.), metal hydrides (NaBH₄), hydrazine (N₂H₄·H₂O) and borane, etc. Compared to the traditional model of first preparing and then storing, hydrogen refueling stations can store these chemical hydrogen storage materials on a large scale instead of pure H₂, thereby improving the safety of hydrogen storage and transportation. In addition, chemical hydrogen storage materials also have the advantages of high volumetric hydrogen storage density, high energy efficiency, and superior safety [8,9].

As a type of chemical hydrogen storage material, ammonia borane (NH₃BH₃ or AB) has attracted much attention due to its high hydrogen capacity (19.6%), non-toxicity, safe storage, and high chemical stability [10]. As shown in Figure 1, the crystal structure of AB includes B-N bonds, B-H bonds, and N-H bonds with bond lengths of 1.45 Å, 1.080 Å, and 1.110 Å, respectively. Among them, the H atom connected to the N atom is positive ($H^{\delta +}$), while the H atom connected to the B atom is negative ($H^{\delta -}$); the existence of a dihydrogen bond ($H^{\delta +}$…$H^{\delta -}$) makes the AB crystal have good stability. Although Shore et al. reported the synthesis of NH₃BH₃ as early as 1955 [11], its application did not gain widespread recognition until the mid-2000s. Numerous studies have demonstrated that NH₃BH₃ can be utilized as an anode fuel [12], a potential liquid organic hydrogen carrier [13], and a solid hydrogen carrier [14]. However, the primary application of NH₃BH₃ is as a hydrogen storage and release medium. Its superior hydrogen storage density, chemical stability, low molecular weight, and high solution stability have made NH₃BH₃ a subject of extensive academic research. However, despite its frequent classification as a liquid organic hydrogen carrier, NH₃BH₃ lacks carbon atoms, meaning it does not conform to the definition of an organic compound. Given this distinction, its classification in the context of hydrogen storage materials warrants reconsideration and clarification.

The hydrogen evolution of NH₃BH₃ primarily occurs through three methods: pyrolysis, methanolysis, and hydrolysis [15]. The thermal desorption of hydrogen from NH₃BH₃ occurs in three stages, at temperatures of approximately 90–130 °C, 130–350 °C, and above 350 °C, respectively [16]. During the pyrolysis process, numerous by-products (NH₂BH₂, NHBH, NB) form, complicating the subsequent separation process [17]. NH₃BH₃ can also be dissolved in methanol and release hydrogen in the presence of a catalyst, but the hydrogen yield and production rate are not ideal. Furthermore, the methanolysis process involves methanol, which is an organic substance that is toxic and relatively expensive [18]. The reaction equations of pyrolysis and methanolysis are shown in Equations (1)–(4), respectively. In contrast, the hydrolysis of ammonia borane has been extensively used, in which water is used both as a reaction medium and a source of pure hydrogen fuel. It also offers the advantages of mild reaction conditions and the absence of CO production (which can easily poison the catalyst) [19,20]. Ideally, 1 mol of AB reacts with 2 mol of water to produce 3 mol of hydrogen, as shown in Equation (5). However, practical hydrolysis yields gas impurities and boron-containing by-products, such as borate ions $B{(OH)}_4^{-}$ [21]. Identifying these hydrolysates and developing regeneration processes are crucial for understanding the mechanisms of AB, as discussed in Section 4 of this paper. It is worth noting that the heat released during the hydrolysis process can significantly influence the thermodynamic and kinetic properties of the system. Studies have demonstrated that when the system temperature rises to 85–135 °C, hydrolysis and thermolysis can become coupled, leading to an enhanced hydrogen release efficiency [22,23,24,25,26]. This phenomenon is commonly referred to as hydrothermolysis. For a systematic comparison of hydrothermolysis mechanisms versus conventional hydrolysis, see Supplementary Materials. However, hydrothermolysis faces challenges related to by-product formation, system stability, and recyclability, limiting its practical application in sustainable hydrogen storage.

$$N{H}_{3}B{H}_3\to {\displaystyle \frac{1}{x}}{(N{H}_{2}B{H}_2)}_x+H_2\uparrow \hspace{1em}\hspace{1em}(90–130) °C$$

(1)

$${\displaystyle \frac{1}{x}}{(N{H}_{2}B{H}_2)}_x\to {\displaystyle \frac{1}{x}}{(NHBH)}_x+H_2\uparrow \hspace{1em}\hspace{1em}(130–350) °C$$

(2)

$${\displaystyle \frac{1}{x}}{(NHBH)}_x\to {\displaystyle \frac{1}{x}}{(NB)}_x+H_2\uparrow \hspace{1em}\hspace{1em}(>350) °C$$

(3)

$$N{H}_{3}B{H}_3+4MeOH\stackrel{catalyst}{\to }N{H}_4^{+}+B{(OMe)}_4^{-}+3H_2\uparrow$$

(4)

$$N{H}_{3}B{H}_3+2H_{2}O\stackrel{catalyst}{\to }N{H}_4^{+}+B{O}_2^{-}+3H_2\uparrow$$

(5)

The self-hydrolysis of AB at room temperature exhibits a low reaction rate due to its strong resistance to hydrolysis. Thus, selecting an appropriate catalyst to enhance generation efficiency is crucial for AB hydrolysis. In 2006, Xu et al. reported for the first time that transition metals can catalyze the hydrolysis of AB [27]. Subsequently, numerous experts and researchers have begun exploring the use of transition metals as catalysts for the hydrolysis of AB. Catalysts are generally classified into noble and non-noble metals based on their chemical properties and catalytic performance. Due to their unique electron density, pure noble metal catalysts demonstrate high TOF and low activation energy [28]. Common noble metals include Pt [29], Rh [30], Ru [31], and Pd [32], with their highest catalytic activity shown in Figure 2a. However, the scarcity and high cost of noble metals have gradually shifted attention towards non-noble metal catalysts, primarily Fe [33], Co [34], Ni [35], and Cu [36]. While non-noble metals have also demonstrated some potential in catalyzing AB hydrolysis, their catalytic performance is often suboptimal, with issues such as poor dispersion, low conductivity, and easy aggregation. Developing a low-cost catalyst with excellent catalytic performance remains a significant challenge in this field.

Given the limitations of pure metal-based catalysts, researchers have explored various design strategies, including ultrafine metal nanoparticles [37,38,39], using supports, morphological engineering [40,41], and ligand stabilization [42], to enhance the number and dispersion of active sites. The introduction of support plays a crucial role in improving catalytic performance, not only by dispersing metal particles but also by regulating the surface properties of the catalyst through metal–support interactions. Precisely controlling the surface properties—adjusting the electronic and geometric structure—is essential for creating highly efficient catalysts [43]. Furthermore, fine-tuning the electronic structure of catalysts through metal component control [44,45], defect engineering [46,47], interface engineering [48], and modulation of metal–support interactions [49] can refine electronic density and valence band structure. This improved electronic structure promotes adsorption and activation of reactant molecules, accelerates the rate-determining step (RDS), and reduces reaction energy barriers, significantly enhancing hydrogen evolution activity. Especially in recent years, researchers have achieved atomic-level regulation of metal catalysts [50], leading to the development of superior single-atom catalysts and greatly advancing AB metal-catalyzed hydrolysis systems. As shown in Figure 2b, these advancements have led to the achievement of TOF values for non-noble metal catalysts, particularly Co and Ni, that are comparable to or exceed those of certain noble metals, such as Pd, thereby underscoring significant progress in catalyst optimization [51,52,53,54].

Previous reviews on AB hydrolysis have primarily concentrated on the catalysts, examining differences in support structures, metal types, and synthesis methods. There is limited discussion on the intrinsic mechanisms affecting the catalytic activity and specific hydrolysis pathways. To our knowledge, no comprehensive review currently addresses hydrolysis reaction mechanisms to summarize optimization strategies for nanocatalysts utilized in AB hydrolysis. This paper begins by examining the hydrolysis process of AB and investigates the mechanisms of AB hydrolysis in detail. The main steps are summarized as follows: adsorption of reactants NH₃BH₃ or H₂O on active sites, activation of chemical bonds within reactants, surface diffusion of active intermediates, interaction between active intermediates, and the formation and desorption of hydrogen. Nevertheless, our understanding of the specific mechanism of the catalyst in AB hydrolysis remains limited, particularly concerning the formation of reaction intermediates. Therefore, this paper does not extensively cover the interaction and conversion of these intermediates. Instead, we focus on analyzing and comparing the structural characteristics and catalytic performance of metal catalysts used in AB hydrolysis over the past five years, evaluating their catalytic efficiency from a kinetic perspective using parameters such as turnover frequency (TOF) and activation energy (E_a). Based on this, we summarize the optimization and regulation strategies for AB hydrolysis catalysts, focusing on four main aspects: performance optimization based on the number and dispersion of active centers, enhancing the adsorption and activation of reactants, promoting hydrogen desorption, and regulating catalytic activity. Additionally, we summarize the hydrolysis by-products and the methods for regenerating AB from these by-products. Finally, corresponding conclusions and perspectives are presented to provide some new insights for the development of stable and efficient catalysts and to further the understanding of catalytic reaction mechanisms.

2. Exploration of the Catalytic Mechanism, Testing and Evaluation Methods

2.1. The Catalytic Mechanism of AB Hydrolysis

While the number of studies on the mechanism of hydrolysis of AB still lags behind experimental studies, there has been a gradual increase in the relevant literature in this area. Figure 3 illustrates the advances in the study of the catalytic mechanism. In 2006, Xu et al. [55] proposed that AB molecules initially adsorb onto the catalyst surface, forming an active metal–hydrogen (M-H) intermediate. The nucleophilic attack of water molecules on the B-N bond is the rate-determining step in the hydrolysis process. In 2014, Duan et al. [56] proposed that the rate-determining step in the dehydrogenation reaction of AB involves the concerted cleavage of both B-H and N-H bonds. In 2017, He et al. [57] identified the cleavage of the O-H bond in H₂O as the rate-determining step (RDS) of AB hydrolysis through kinetic isotope effect analysis. In the following studies, several common hydrolysis mechanisms for hydrolysis were proposed, including nucleophilic substitution [52], oxidative addition and reductive elimination [58], and proton activation [59]. In 2020, Li et al. [60] suggested that water molecules attack the transition state M-H, which is the rate-determining step in the reaction. Therefore, increasing the formation rate of M-H can significantly boost the hydrogen production rate. The formation and evolution of M–H intermediates (e.g., in Fe, Co, Ni systems) are critical to understanding the true active centers in AB hydrolysis, particularly for iron-subgroup catalysts undergoing redox and hydride-mediated transformations. Recently, Xu et al. [61] also verified this proposition through experimental studies. Furthermore, some researchers have proposed bimolecular activation mechanisms. For instance, Gui et al. [62] introduced the bimolecular activation mechanisms within Ni_1.2Fe_0.8@CN-G nanoalloys. Additionally, Zhang et al. [63] proposed a dual-active-site effect mechanism at the interface. In 2024, Guan et al. [64] proposed the construction of RuPt-Ti multi-site catalysts. Currently, there is no clear and unified understanding of the rate-determining step in the hydrolysis reaction. This paper aims to provide a comprehensive summary and analysis of these mechanisms, detailed in the following text.

(a)

Nucleophilic substitution mechanism

Nucleophilic substitution reactions (SN₂) gradually involve a positively charged carbon atom reacting with a nucleophile carrying a negative charge or partial negative charge, in which the carbon atom is replaced by the nucleophile [65]. This is one of the common mechanisms used to explain AB hydrolysis. In 2017, Huo et al. [52] prepared Ni_0.7Co_1.3P/GO to study its catalytic effects on AB hydrolysis. DFT calculations revealed that the nucleophilic substitution (SN₂) step is the rate-determining step in hydrogen evolution via AB hydrolysis. Qu et al. [66] believe that the synergistic electronic effect between NiCoP NPS and OPC-300 provides more convincing evidence for the nucleophilic substitution reaction (Figure 4a). In 2020, Wu et al. [67] identified several S-N reaction processes at the Co-Co₃O₄/CDS interface during synergistic catalysis (Figure 4b).

The general pathway for the nucleophilic substitution process is as follows: First, AB and H₂O molecules must overcome a certain energy barrier to adsorb onto the catalyst surface. During this process, the special interaction between the metal active sites on the catalyst and the reactant molecules weakens the chemical bonds within the reactant molecules. The OH* group in activated water attacks the B-N bond in NH₃BH₃, forming the intermediate BH₃OH* (OH* + BH₃NH₃* → BH₃OH* + NH₃*). Subsequently, the B-H bond breaks, releasing H* (BH₃OH* → BH₂OH* + H*). The H* released from the broken B-H bond combines with H* from water to form an H₂ molecule, which then desorbs from the catalyst surface. During this process, the SN₂ reaction repeats continuously, successively producing the intermediates BH₂(OH)₂* and BH(OH)₃*, ultimately generating 3 mol of H₂.

(b)

Oxidative addition and reductive elimination

Oxidative addition is a reaction where two atoms are added to a central atom simultaneously, increasing the oxidation state. Reduction elimination is the inverse process of oxidation addition, which involves the breaking of the group on the central atom and the release of a small molecule, usually accompanied by a decrease in the oxidation state. In 2017, Fu et al. [58] pointed out that the oxidative addition of the O-H bond in water is a rate-determining step in the hydrolysis process according to DFT calculation, and its high KIE value of 4.95 also confirms this claim. [H₃NBH₂H]⋯H-OH hydrogen bonding resulting from the hydridic property of the B-H bond is formed, laying the foundation for subsequent oxidative addition and reductive elimination. Peng et al. [68] further studied the reduction and elimination of H₂ (Figure 5a). After oxidative addition, the H* atoms from the B-H bond and the H* atom from the H₂O molecule are adsorbed on the Rh, respectively. The process of binding and releasing the H₂ molecule from the two active H atoms is regarded as reductive elimination. Wang et al. [69] used RuNi alloy for AB hydrolysis (Figure 5b). Due to the coordination of OH⁻ with metal NPs, the conductive effect of OH⁻ significantly enhances the electron density of alloy NPs, facilitating the oxidative addition of H₂O. The hydrolysis pathways based on oxidative addition and reduction elimination are as follows:

$$N{H}_{3}B{H}_3+H_{2}O\to N{H}_{3}B{H}_{2}OH+H_2\uparrow$$

(6)

$$N{H}_{3}B{H}_{2}OH+H_{2}O\to N{H}_{3}BH{(OH)}_2+H_2\uparrow$$

(7)

$$N{H}_{3}BH{(OH)}_2+H_{2}O\to N{H}_{3}BH{(OH)}_3+H_2\uparrow$$

(8)

$$N{H}_{3}BH{(OH)}_3+H_{2}O\to N{H}_4^{+}+B{(OH)}_4^{-}$$

(9)

(c)

Bimolecular activation

Previous studies have demonstrated that single-metal catalysts generally activate AB molecules effectively but face significant activation energy barriers for H₂O. Recently, there has been growing interest in the bimolecular activation of both AB and water molecules. In 2020, Chen pioneered the development and control of Pt-WO dual active sites for hydrogen production from AB [70]. In 2021, Yao et al. [32] introduced a metal–support bimolecular activation mechanism using Pd/alk-Ti₃C₂. In this system, Pd NPs, with high surface activation energy, dissociate the B-H bond of NH₃BH₃, while surface-modified alk-Ti₃C₂ activates the O-H bond of H₂O, facilitating the formation of hydrogen at the metal–oxide interface (Figure 6a). The specific reaction pathway is shown in Equations (10)–(18). In 2022, Zhang et al. [63] synthesized a crystalline B-Co-P dual-active-site system supported on hexagonal boron nitride (h-BN) to explore its catalytic effects on AB hydrolysis (Figure 6b). Structural characterizations (e.g., XRD and high-resolution TEM) confirmed the well-defined atomic-bridge configuration of Co-B and Co-P sites, providing a reliable basis for DFT modeling. Their DFT calculations, based on the experimentally validated crystalline structure, revealed that H₂O preferentially adsorbs on Co-B sites, while NH₃BH₃ binds more strongly to Co-P sites. This spatially aligns with the distinct electronic environments of Co-B (electron-deficient) and Co-P (electron-rich) sites, as evidenced by XPS and Bader charge analysis. The atomic-bridge structure synergistically weakens the B–H and O–H bonds via dual-site polarization effects, thereby lowering the overall activation energy. Zhao et al. [71] proposed a bimolecular activation mechanism within PdCu nanoalloys, where Pd and Cu can activate ammonia borane and water molecules, respectively. The specific reaction pathway is shown in Equations (19)–(29).

$$N{H}_{3}B{H}_3+Pd\to N{H}_{3}B{H}_2^{\ast }+H^{\ast }$$

(10)

$$H_{2}O+T{i}_3{C}_2\to H{O}^{\ast }-T{i}_3{C}_2+H^{\ast }$$

(11)

$$H^{\ast }+H^{\ast }\to H_2\uparrow$$

(12)

$$N{H}_{3}B{H}_2^{\ast }+Pd\to N{H}_{3}B{H}^{\ast }-Pd+H^{\ast }$$

(13)

$$H_{2}O+T{i}_3{C}_2\to H{O}^{\ast }-T{i}_3{C}_2+H^{\ast }$$

(14)

$$H^{\ast }+H^{\ast }\to H_2\uparrow$$

(15)

$$N{H}_{3}B{H}^{\ast }+Pd\to N{H}_3{B}^{\ast }-Pd+H^{\ast }$$

(16)

$$H_{2}O+T{i}_3{C}_2\to H{O}^{\ast }-T{i}_3{C}_2+H^{\ast }$$

(17)

$$H^{\ast }+H^{\ast }\to H_2\uparrow$$

(18)

$$N{H}_{3}B{H}_3+2P{d}^{\ast }=N{H}_{3}B{H}_2-P{d}^{\ast }+H-P{d}^{\ast }$$

(19)

$$H_{2}O+2C{u}^{\ast }=HO-C{u}^{\ast }+H-C{u}^{\ast }$$

(20)

$$H-P{d}^{\ast }+H-C{u}^{\ast }=H_2+P{d}^{\ast }+C{u}^{\ast }$$

(21)

$$N{H}_{3}B{H}_2-P{d}^{\ast }+HO-C{u}^{\ast }=N{H}_{3}B{H}_{2}OH-P{d}^{\ast }+C{u}^{\ast }$$

(22)

$$N{H}_{3}B{H}_{2}OH-P{d}^{\ast }+P{d}^{\ast }=N{H}_{3}BHOH-P{d}^{\ast }+H-P{d}^{\ast }$$

(23)

$$H-P{d}^{\ast }+H-C{u}^{\ast }=H_2+P{d}^{\ast }+C{u}^{\ast }$$

(24)

$$N{H}_{3}BHOH-P{d}^{\ast }+HO-C{u}^{\ast }=N{H}_{3}BH{(OH)}_2-P{d}^{\ast }+C{u}^{\ast }$$

(25)

$$N{H}_{3}BH{(OH)}_2-P{d}^{\ast }=N{H}_{3}BHO-P{d}^{\ast }+H_{2}O$$

(26)

$$H-P{d}^{\ast }+H-C{u}^{\ast }=H_2+P{d}^{\ast }+C{u}^{\ast }$$

(27)

$$N{H}_{3}BHO-P{d}^{\ast }+HO-C{u}^{\ast }=N{H}_{3}BOOH-P{d}^{\ast }+C{u}^{\ast }$$

(28)

$$N{H}_{3}BOOH-P{d}^{\ast }=N{H}_4^{+}+B{O}_2^{-}+P{d}^{\ast }$$

(29)

In addition, some scholars have used density functional theory (DFT) to simulate the reaction mechanisms of the catalysts. Huo et al. [72] reported that a Fe₃₆Co₄₄ alloy cluster was used to hydrolyze AB and constructed five reaction pathways with different atomic coordination numbers. They found that the reaction has the smallest rate-determining step (RDS) barrier in the case of N-H bond preferential cleavage. Mao et al. [73] designed four reaction paths for hydrogen evolution from NH₃BH₃, catalyzed by CoP, N@CoP, and S@CoP. They discovered that the density of electronic states near the Fermi level of the N@CoP catalyst increased. In contrast, the density of electronic states near the Fermi level of the S-doped catalyst decreased. They proposed that N doping is the key factor in improving the hydrogen evolution activity from NH₃BH₃ catalyzed by CoP.

2.2. Testing and Evaluation Methods of Catalyst Performance

2.2.1. Performance Test of Dehydrogenation in Laboratory

In AB hydrolysis experiments, two reactor types are commonly used: magnetic-stirrer-driven reactors (Figure 7a), where stirring enhances AB solution diffusion, and magnetic catalyst reactors (Figure 7b), where catalyst magnetism facilitates mixing. The standardized protocol involves ultrasonic dispersion of the catalyst, reaction under controlled temperature and stirring in a water bath, and hydrogen collection via water displacement. Performance evaluation follows standardized parameters, with turnover frequency (TOF) as a key catalytic activity metric. Additional evaluation criteria are discussed further.

2.2.2. Standardized Description of Catalytic Performance Evaluation Methods

The catalytic performance evaluation parameters commonly employed include hydrogen generation rate (HGR), turnover frequency (TOF), and apparent activation energy (E_a). Additionally, the stability of the catalyst is also evaluated, with parameters such as TOF typically measured at room temperature.

(a)

Hydrogen Generation Rate (HGR)

HGR is a key indicator of catalytic activity, quantifying the hydrogen released per unit time and mass. It is frequently used to assess the performance of borohydride-based reactions and electrochemical hydrogen evolution. HGR is determined by fitting the slope of the hydrogen evolution volume vs. time curve in its linear range, as expressed by Equation (30).

$$\mathrm{H}\mathrm{G}\mathrm{R}={\displaystyle \frac{{\mathrm{V}}_{{\mathrm{H}}_2}}{{\mathrm{m}}_{\mathrm{catalyst}}•\mathrm{t}}}$$

(30)

In the above formula, ${\mathrm{V}}_{{\mathrm{H}}_2}$ represents the volume of H₂ generated, ${\mathrm{m}}_{\mathrm{catalyst}}$ is the total mass of the catalyst used, and t is the time required to complete the hydrolysis reaction. Therefore, the unit for the hydrogen generation rate (HGR) can be expressed as either $\mathrm{mLg}·{\min }^{-1}·{\mathrm{g}}^{-1}$ or $\mathrm{L}·{\min }^{-1}·{\mathrm{g}}^{-1}$.

(b)

Turnover Frequency (TOF)

Turnover frequency (TOF), introduced by Boudart in 1966, quantifies catalytic activity as the number of molecular reactions per active site per second. Given the challenge of precisely determining active sites, TOF is often approximated using the molar quantity of the catalyst:

$$\mathrm{T}\mathrm{O}\mathrm{F}={\displaystyle \frac{{\mathrm{n}}_{{\mathrm{H}}_2}}{{\mathrm{n}}_{\mathrm{cat}}•\mathrm{t}}}$$

(31)

For metal-based catalysts, TOF is typically estimated using metal molar content, while in multifunctional systems, ligands and oxides may also contribute to active sites. The unit of TOF is technically expressed as ${\mathrm{mol}}_{{\mathrm{H}}_2}•{\mathrm{mol}}_{\mathrm{cat}}•{\mathrm{h}}^{-1}$, ${\mathrm{mol}}_{{\mathrm{H}}_2}•{\mathrm{mol}}_{\mathrm{cat}}•S^{-1}$, or ${\mathrm{mol}}_{{\mathrm{H}}_2}•{\mathrm{mol}}_{\mathrm{cat}}•{\min }^{-1}$, However, it is sometimes also expressed in units of min⁻¹, s⁻¹, or h⁻¹. In this paper, we will use min⁻¹ as the simplified unit for TOF.

(c)

Apparent Activation Energy (E_a)

Activation energy represents the minimum energy required for hydrolysis and serves as a key catalytic parameter. It is determined via the Arrhenius Equation (32):

$$\mathrm{L}\mathrm{n}k=\mathrm{L}\mathrm{n}\mathrm{A}-{\displaystyle \frac{{\mathrm{E}}_{\mathrm{a}}}{\mathrm{R}}}•{\displaystyle \frac{1}{\mathrm{T}}}$$

(32)

where the slope of lnk versus 1/T gives E_a.

(d)

Stability

A catalyst’s stability, a key indicator of its performance, is evaluated through recyclability and reusability tests. Recyclability is assessed via the total turnover number (TON), which quantifies the catalyst’s maximum utilization. Reusability testing involves recovering and reintroducing the catalyst into fresh AB solutions, measuring retained activity across cycles. A stable catalyst exhibits minimal variation in HGR or TOF over multiple cycles.

3. Strategies for the Optimization and Control of Catalytic Performance

Given the limitations of pure noble metals and non-noble metals, researchers have employed various strategies to enhance their ability to catalyze AB hydrolysis at room temperature. This section begins with a simplified overview of the AB hydrolysis steps, first outlining the strategies for optimizing the performance of AB hydrolysis catalysts, with a particular emphasis on methods for increasing the number and dispersion of active centers. Building on this, it further summarizes approaches to enhance the adsorption and activation of reactants and explores the impact of hydrogen desorption on the AB hydrolysis process. Additionally, this section discusses the controllability of hydrogen release during AB hydrolysis by adjusting the catalyst’s usage conditions. Compared to traditional catalysts, controllable catalysts may maintain higher activity over long-term use, thereby reducing the frequency of catalyst replacement and maintenance. Finally, Table 1 presents a summary of relevant catalytic parameters, including TOF(HGR), activation energy (E_a), particle size (nm), durability, and other relevant information on noble and non-noble metal catalysts.

3.1. Strategies for Performance Optimization Based on Number and Dispersion of Active Centers

From previous studies, it is known that not all catalyst sites exhibit catalytic activity; those that do are termed active sites or active centers [74]. These active sites can be specific metal atoms, ions, defects, or surface functional groups. The reactants or intermediates can be adsorbed onto these active sites and interact with the catalyst.

The number and dispersion of active centers are crucial for catalytic activity. A higher number of active centers provide more reaction sites, and good dispersion ensures that these active centers are uniformly distributed on the catalyst surface, enhancing the interaction between reactants and active centers. Both factors contribute to improved catalytic efficiency. In this section, we discuss optimization strategies based on the number and dispersion of active centers, focusing on four aspects: size effects, morphology, modification of support, and stabilizer ligands, as summarized in Figure 8. Furthermore, strategies for enhancing the adsorption and activation of reactants are elaborated in Section 3.2 of this paper.

3.1.1. Optimization of Active Center by Size Effect

The catalytic activity of AB hydrolysis on supported metal NPs is size-dependent; specifically, the specific surface area of the catalyst increases as the metal particle size decreases. Researchers estimate metal dispersion using the average size of metal nanoparticles, where dispersion refers to the uniform distribution of metal particles on the supports. Smaller metal NPs achieve more uniform dispersion. By precisely controlling the size and dispersion of metal NPs, the specific surface area of the catalyst can be effectively increased, enhancing the exposure of surface active sites and improving the catalytic performance of the metal NPs [75,85].

The size-dependent effect on catalytic activity is most significant among single-metal catalysts. In 2017, Guo et al. [76] explored the size effect on the hydrolysis of AB catalyzed by unsupported Ni NPs. They found that the TOF of Ni NPs with sizes ranging from 4.9 to 27.4 nm exhibited a trend of first increasing and then decreasing, commonly referred to as a volcanic trend. Chen and colleagues utilized Ru NPs with different crystal phases to catalyze the hydrolysis of AB, finding that the size effect was the primary factor determining the activity trend of hexagonal close-packed (hcp) Ru [86]. Ju et al. [87] encapsulated Pt NPs (1–4 nm) within dendrimers and proposed that changes in Pt size within the sub-nanometer range have a tunable impact on catalytic activity (Figure 9), providing clear experimental insights into its intrinsic size-dependent catalytic behavior.

The impact of catalyst particle size on AB hydrolysis is primarily due to significant changes in the coordination environment of the catalyst with decreasing size, coupled with an increase in specific surface area that enhances the exposure of active sites. However, as the size decreases, the activity normalized to the metal surface area diminishes. This reduction is mainly due to the altered surface composition of the catalyst, leading to the agglomeration of metal NPs. Consequently, researchers frequently support catalysts on materials with a large specific surface area to improve the dispersion of active sites. For instance, Aksoy et al. [77] reported carbon nanotube-graphene hybrid materials supported with Pt nanoparticles (Pt/CNT-G). Lu et al. [88] anchored ultrafine MoOx-doped Ni NPs with an average particle size of 3 nm onto g-C₃N₄@glucose-derived nitrogen-doped carbon nanosheets. Punzi et al. [89] employed a covalent triazine framework (CTFPh) to support ultrafine Ni NPs with a diameter of 2.2 nm. These catalysts all exhibit high catalytic activity and low activation energy in the hydrolysis reaction of AB at room temperature. We propose that the exceptional catalytic activity can be attributed to the effective dispersion of small-sized metal NPs on supports, which increases the exposure of surface active sites.

Furthermore, reducing metal NPs’ size to single atoms significantly enhances catalytic performance. Single-atom catalysts (SACs) exhibit remarkably high catalytic activity due to the full exposure of active centers and their distinct electronic and geometric properties [90]. Yang et al. [78] achieved the controlled synthesis of Pt/graphene sub-nanometer catalysts (including single atoms and dimer clusters), demonstrating a volcano-type trend in TOF and activation energy relative to the number of ALD cycles of Pt. This finding further corroborates the volcano-type dependence of intrinsic activity on the size of Pt sub-nanometer catalysts.

In summary, as the size of the metal NPs decreases to the nanometer scale, the coordination environment of the surface atoms changes significantly. Small-size metal NPs exhibit more uncoordinated atoms (e.g., at corners and edges), leading to increased exposure of surface atoms, which generally possess higher catalytic activity compared to host atoms. Furthermore, reducing the size of metal nanocatalysts to sub-nanoscale (such as single atom forms) can achieve ultra-high atomic utilization and superior catalytic performance. However, smaller catalyst size does not always correlate with better performance. The TOF and nanoparticle size of some catalysts exhibit a volcanic trend, likely due to a tradeoff between the increased catalytic activity from decreased nanoparticle size and the decreased activity from particle agglomeration. Therefore, supporting small-sized metal NPs on high-surface-area materials enhances dispersion and catalytic performance.

3.1.2. Optimization of Active Center by Morphology

In addition to controlling the size of metal NPs, tuning the surface morphology of metal active components or support materials is crucial for regulating catalytic activity. Engineering various morphologies, such as nanowires, nanospheres, and nanoplates, creates distinct structural characteristics like planes and edges, resulting in diverse surface active sites. Moreover, adjusting the surface morphology of supports can achieve a higher specific surface area, enhancing the dispersion and exposure of the metal active sites [79]. To objectively elucidate the impact of morphology on catalysts for hydrogen production via the hydrolysis of AB, this section examines both unsupported and supported catalysts, with particular emphasis on unsupported transition metal oxides and uniquely structured supported catalysts.

(a)

Unsupported transition metal oxides

Transition metal oxide hybrids have attracted much attention in recent years, with numerous studies showing that their catalytic properties strongly depend on their morphology. Feng et al. [91] synthesized three types of Ni_0.5Cu_0.5Co₂O₄ nanocomposites with distinct morphologies for AB hydrolysis: nanoplates, nanoparticles, and sea urchin-like microspheres composed of nanowires. Among these, the Ni_0.5Cu_0.5Co₂O₄ nanoplates exhibited the best catalytic performance. This enhancement is attributed to their ease of reduction and rapid formation of hydrolyzed products compared to nanoparticles and microspheres. In contrast, microspheres composed of hollow micro/nanostructures provide a high specific surface area and rich coordination environment, enhancing mass transfer and resulting in excellent catalytic performance [92]. Feng et al. [93] first developed a high-performance plate-like Cu₂O-CoO nanocomposite catalyst for the hydrolysis of NH₃BH₃. The synergistic interaction between Cu₂O and CoO led to an impressive turnover frequency (TOF) of 34.1 min⁻¹. (Figure 10a). Similarly, Feng et al. [53] used sea urchin-like hollow CuMoO₄-CoMoO₄ mixed microspheres containing nanorods. This intricate structure exhibited much higher TOF values than a single CuMoO₄ or CoMoO₄ microsphere, as illustrated in Figure 10b.

Numerous studies have demonstrated that metal oxides with nanowire morphologies often expose more surface active sites, enhancing catalytic activity in AB hydrolysis experiments. For instance, Zhang et al. [94] observed that hollow worm-like nanowires in Rh/h-NCNW create flexible channels that facilitate the exposure and mass transport of surface active sites, thus improving the hydrolysis of AB (Figure 10c). Lu et al. synthesized Cu_0.6Ni_0.4Co₂O₄ nanowires [95] and Mo-doped Cu_0.5Ni_0.5Co₂O₄ nanowires [96] (Figure 10d,e), both exhibiting excellent catalytic activity in AB hydrolysis, with TOF values of 119.5 min⁻¹ and 195.25 min⁻¹, respectively—among the highest for non-noble metal catalysts. Furthermore, Lu et al. suggested that specific surface area is not the primary factor influencing Cu_0.6Ni_0.4Co₂O₄’s catalytic performance, contrary to some previous studies. They hypothesized that certain nanowires might transform into nanorods or even nanoparticles after multiple reactions, and that the loss of nanowire morphology could account for diminished catalytic activity. Thus, morphological changes may critically determine the catalytic activity of transition metal oxides. However, catalysts with the same morphology but different compositions exhibit significant differences in catalytic performance. Adjusting the active components can further fine-tune catalytic performance. Nevertheless, the effect of different components in metal oxides on their activity is complex and necessitates further investigation.

(b)

Catalysts with special support morphology

Specialized support structures can significantly increase the specific surface area, facilitating the dispersion of small-sized metal NPs and creating more catalytically active centers. For instance, Chen et al. [97] report a CoNiP nanobox supported on graphene oxide (GO) for AB hydrolysis. The box-like configuration greatly increased the specific surface area and the density of active sites (Figure 10f). Similarly, Wang et al. [98,99] employed nickel foam to support Co-Fe-B (Figure 10g) and Co-Mo-B NPs (Figure 10h), both exhibiting twisted ribbon-like structures. These samples demonstrated exceptional catalytic activity due to their small size and distinctive morphological structures. Additionally, innovative composite nanostructures, such as those with core@void@shell configurations, are attracting increasing attention. For example, Li et al. [100] fabricated CuO/Co₃O₄@C microspheres using a simple hard template etching approach (Figure 10i). The unique rattle structure and synergistic effect between CuO and Co₃O₄ markedly enhanced the catalytic performance in AB hydrolysis. Yuan et al. [101] doped urchin-like titanium dioxide nanoribbon assemblies with B impurities and then deposited uniformly distributed ultrafine Ru NPs. This hierarchical structure notably enhanced the catalytic hydrogen evolution performance of AB, achieving a high activity of 1729 min⁻¹ without an alkaline promoter.

In summary, the morphology of catalysts significantly affects their specific surface area, the dispersion of metal particles, and the quantity and distribution of active sites, all of which determine the kinetic performance and efficiency of the catalyst. By precisely controlling the support morphology, such as utilizing nanobox and core@void@shell structures, catalytic effectiveness can be greatly optimized. Additionally, variations in metal oxide components can influence catalytic activity. It is crucial to recognize that in metal oxides, the active catalysts are the reduced metals or alloys, not the pure metal oxides themselves. Nonetheless, the synergistic effects between metal oxides and their mechanisms of action remain unclear and require further exploration.

3.1.3. Optimization of Active Center by Modification of Support

The morphology of the support plays a critical role in determining the number and dispersion of active sites. Equally important are the surface properties of the support, which play a crucial role in determining catalytic performance. In recent years, the modification of supports has become a research hotspot, with carbon and oxide supports being the most common. Heteroatom doping, involving non-metallic elements such as N, O, and P, allows scholars to introduce various functional groups and defects on the support surface. These functional groups and defects do not directly catalyze reactions but serve as active sites and influence the properties of active sites by regulating the electronic structure of the support. This section has reviewed the impact of different functional groups and defects introduced by heteroatom doping on active sites.

Heteroatom doping of carbon (such as N, O, and P) has been proven to be an effective strategy for controlling the size of metal NPs by anchoring their growth sites [102,103]. In particular, nitrogen (N) doping introduces various nitrogen-containing functional groups (e.g., pyridine nitrogen, indole nitrogen) that markedly improve the catalytic performance of metal sites [104,105]. For example, Zhong et al. [106] used hierarchical porous nitrogen-doped carbon (NPC) as a support to anchor small-sized RuCo NPs. The nitrogen functional groups improve the affinity of metal precursors via electrostatic interactions, enhancing the anchoring and stability of metal NPs. Liang et al. [107] synthesized Ru/Ti₃C_2−xN_x using nitrogen-doped MXene (Ti₃C_2−xN_x) as the support, which exhibited the highest TOF value of 1334 min⁻¹ in the hydrolysis of AB. Liu et al. [80] observed that CoSNC axially coupled with out-of-plane CoRu nanoalloy, and the increased exposure of pyridine nitrogen in the support further optimized its electronic structure, facilitating the activation of more reaction sites (Figure 11a). Similarly, Sun et al. [108] supported uniformly dispersed Ru nanoclusters (NCs) on oxidized carbon black, while Slot et al. [109] used surface-oxidized Ti₃C₂T_x to support Pt NPs. Both systems exhibited high catalytic activities for AB hydrolysis at room temperature, with TOF values of 602 min⁻¹ and 272 min⁻¹, respectively. This enhancement is primarily ascribed to the oxygenated groups that promoted the formation and uniform dispersion of small-sized Ru and Pt NPs, providing rich surface accessibility for catalytic active sites.

Furthermore, researchers have optimized the structure of catalysts by introducing multiple functional groups and defects on the support surface. For example, Guo et al. [81] and Jiang et al. [110] prepared ultrafine Ru NCs supported on O/N-doped carbon (ONC) with hollow and hierarchical porous structures. Both exhibited excellent catalytic performance in AB hydrolysis, surpassing most developed Ru catalysts. Notably, the Ru/ONC with hollow structures achieved a TOF of 1837 min⁻¹ in an alkaline solution (Figure 11b). This improvement is attributed to the abundant O/N doping, which provides numerous surface sites for anchoring Ru ions and regulating the nucleation and growth of NCs, thus forming a uniform distribution of Ru NCs with rich surface active sites. Zheng’s team [111] demonstrated that the N/O co-doped WSC exhibits a unique micropore-dominated porous structure, which not only promotes the uniform and compact dispersion of the RuNi alloy but also enhances the hydrophilicity of the resulting catalyst, thereby contributing to improved catalytic performance. Wu et al. [82] found that the co-doping of nitrogen and phosphorus introduced numerous defects in the CoP–CoO/NCDs nano-heterostructure, further regulating the reactivity of the carbon layer (Figure 11c). Similarly, Song et al. [112] utilized B/N co-doped carbon sulfide to support Ru NPs (Ru/BNC). The co-doping enhanced the defect sites and porosity of the carbon substrate, facilitating the formation of uniformly dispersed, ultrafine Ru NPs (Figure 11d).

Doping defects can also optimize the electronic properties of active sites, creating effective adsorption and activation sites. For example, Long’s team [113] developed N-doped urchin-like TiO₂ (N-U-TiO₂)-supported Rh nanoparticles, where oxygen vacancies induced by N-doping facilitated electron transfer to Rh, generating electron-rich Rh species that enhanced water dissociation. Wang et al. [114] proposed that oxygen vacancies (VOs) form unique electronic vacancies that act as electron donors (Figure 11e), fostering the formation of an electron-rich surface on Cu_0.76Co_2.24O₄. The instantaneous Cu-H generated by CuO reduction is also conducive to the generation of hydrogen. However, the exact role of VOs in AB hydrolysis, especially within this complex system of metal oxide catalysts, remains uncertain. Further studies are needed to determine whether VOs act as catalytic active sites or merely modulate the electronic structure of the catalyst. A more detailed discussion on this topic is provided in Section 3.2. Nevertheless, the doped heteroatoms may also reduce the defect degree of the catalyst. Zhang et al. [115] found that doping Co NPs into Rh/Ni@Ni-N-C resulted in the activation energy of Rh_0.75Co_0.25/Ni@Ni-N-C being slightly higher than that of Rh/Ni@Ni-N-C. From the XPS image (Figure 11f), the authors speculated that the doped heteroatoms occupied the position between Rh and Ni@Ni-N-C defects, which weakened the interaction between Rh and defect sites.

Figure 11. (a) Fine XPS spectra of N 1 s of Co_0.5Ru_0.5/Co_sNC, Co_0.5/Co_sNC, Ru_0.5/Co_sNC, and Co_sNC [80]. (b) Activity comparison of Ru/ONC with other studies [81]. (c) N and P co-doping and interface effect for improving AB hydrolysis [82]. (d) Raman spectra of Ru/BNC, BNC, Ru/C, and C [112]. (e) XPS spectra of Cu 2p in catalysts and VO defect strategy impact [114]. (f) Raman spectra peak diagram of Ni@Ni-N-C, Rh/Ni@Ni-N-C, and Rh_0.75Co_0.25/Ni@Ni-N-C [115].

Figure 11. (a) Fine XPS spectra of N 1 s of Co_0.5Ru_0.5/Co_sNC, Co_0.5/Co_sNC, Ru_0.5/Co_sNC, and Co_sNC [80]. (b) Activity comparison of Ru/ONC with other studies [81]. (c) N and P co-doping and interface effect for improving AB hydrolysis [82]. (d) Raman spectra of Ru/BNC, BNC, Ru/C, and C [112]. (e) XPS spectra of Cu 2p in catalysts and VO defect strategy impact [114]. (f) Raman spectra peak diagram of Ni@Ni-N-C, Rh/Ni@Ni-N-C, and Rh_0.75Co_0.25/Ni@Ni-N-C [115].

In summary, the anchoring effect of functional groups and defects on the support surface facilitates the formation of ultrafine, ligand-free metal NPs with high dispersion, thus creating more active centers. These functional groups and defects also serve as coordination sites, stabilizing the small metal NPs, promoting the mass transfer of reactants, and enhancing the accessibility of active sites. Moreover, certain multi-functional functional groups can strengthen the metal–support interaction, altering the adsorption state of reactants to stabilize metal ions and reduce the leaching of metal NPs. This aspect is described in more detail in Section 3.2.

3.1.4. Optimization of Active Center by Stabilizer

Ultrafine metal NPs are prone to instability and aggregation, which reduces the number of active sites. To address this, researchers often use stabilizers to maintain their stability. Various stabilizers, including synthetic polymers, surfactants, and natural macromolecular compounds, have been extensively explored to regulate NP growth and enhance dispersion.

Synthetic stabilizers such as polyvinyl alcohol (PVA), polyvinylpyrrolidone (PVP), cetyltrimethylammonium bromide (CTAB), and sodium citrate primarily stabilize metal NPs through physical adsorption or weak chemical interactions. For instance, sodium citrate has been shown to effectively reduce particle size, inhibit aggregation, and enhance stability. Chen et al. [83] demonstrated that sodium citrate can be used as a stabilizer to effectively reduce particle size, inhibit aggregation, and improve the stability of metal NPs. Similarly, PVP plays a critical role in morphology control. Yang et al. [116] synthesized Pt, Pd, and PtPd alloy NPs with different shapes and sizes (within 10 nm) by tuning the PVP content. The interaction of PVP with different crystallographic planes influenced the exposure of specific facets, thereby affecting catalytic performance. Surfactants have also been employed to create structured porous architectures that improve catalyst stability and activity. Umegaki et al. [117] demonstrated that CTAB, through micellar self-assembly, facilitates the formation of ordered nanoporous structures, significantly increasing the specific surface area and pore volume of catalysts. Meanwhile, sodium dodecyl sulfate (SDS), an anionic surfactant, has been found to regulate particle growth by inhibiting excessive nucleation, reducing primary particle size, and optimizing the deposition of active metals such as Ni and Ru, thereby enhancing catalytic efficiency for ammonia borane hydrolysis [118].

In recent years, natural macromolecular stabilizers have emerged as promising alternatives due to their biocompatibility and multifunctionality. Ozay et al. [119] utilized xanthan gum, a natural polymer, to effectively stabilize Ru₀ NPs, demonstrating its potential as a green stabilizing agent. Fu et al. [84] further expanded this approach by employing natural polyphenol-based stabilizers. Their study revealed that bayberry tannins (BTs), when immobilized on collagen fibers (CFs), not only served as stabilizers but also acted as carriers for Ru NPs. The phenolic hydroxyl groups of BT formed coordination bonds with Ru ions, significantly reducing NP size and enhancing dispersion.

However, some studies have shown that stabilizers can strongly bind to the metal surface and block active sites [120]. Although these stabilizers can be removed from metal NPs, the process will inevitably damage the catalyst’s surface structure. Therefore, researchers are developing green catalysts that either do not require stabilizers or use easily removable ones to avoid structural damage. For example, Wang et al. [121] synthesized ultrafine, highly dispersed Ru NCs using inexpensive urea as a nitrogen source and three-dimensional porous carbon for confinement. This method eliminates the need for surfactants, thereby avoiding potential activity loss associated with their removal. Additionally, some studies have demonstrated that ultrafine, ligand-free metal NPs can be synthesized using dispersants. For instance, Mao et al. [122] synthesized ligand-free Pd/IPC NPs using sodium hydroxide and a 3D IPCN structure as supports. The achieved Pd/IPCNs exhibited extremely high catalytic activity for hydrogen production via the hydrolysis of AB, with a TOF value of 122.8 min⁻¹, surpassing many reported Pd-based catalysts. This high activity is due to sodium hydroxide effectively controlling the growth of small Pd NPs on the IPCN surface.

The above studies demonstrate that adding stabilizers during catalyst synthesis can reduce the irregular aggregation, distribution, and regrowth of NPs, significantly mitigating activity loss. However, since stabilizers can cover active sites, developing catalysts that do not require stabilizers or use easily removable ones is crucial for achieving high-performance catalysts.

3.2. Strategies for Performance Optimization Based on Enhancing Adsorption and Activation of Reactants

Based on the above descriptions, researchers have implemented various measures to optimize the number and dispersion of active centers. However, enhancing the catalytic activity of AB hydrolysis also hinges on the effective adsorption and activation of reactant molecules at these active sites. The interaction between the metal active sites and the reactant molecules can activate the reactant molecules, thus weakening the chemical bonds in the reactant molecules and reducing the activation energy required for the reaction [123,124]. Specifically, promoting the adsorption and activation of water molecules on the catalyst surface is crucial, and this process relies on the precise tuning of the catalyst’s electronic structure. Scholars have found that the synergistic effect between different metals and the metal–support interaction can finely tune the electronic structure of the catalyst surface. This optimization significantly promotes the adsorption and activation of reactant molecules and accelerates the hydrolysis of AB. In this section, the metal alloys mainly include bimetallic alloys and polymetallic alloys. Moreover, by carefully selecting support materials and adjusting surface modifications, the metal–support interaction can be optimized at the molecular level, thereby enhancing overall catalytic performance.

3.2.1. Optimization of Adsorption and Activation of Reactants by Bimetallic Alloys

The poor catalytic activity of single-component non-noble metal catalysts in the hydrolysis of AB can primarily be attributed to their weak adsorption capacity for reactants, leading to the rapid dissociation of reactant molecules from the surface [125]. Additionally, most metals effectively activate AB molecules but exhibit low adsorption energy and high activation barriers for H₂O molecules. This limitation critically impedes their catalytic efficacy in AB hydrolysis. To address these limitations, researchers have increasingly turned to doping strategies to enhance the performance of non-noble metal catalysts. A notable advancement was made in 2009 by Xu et al., who synthesized FeNi alloy with Pt-like catalytic activity for the first time, demonstrating a significant enhancement in catalytic performance in AB hydrolysis [126]. Following this, various bimetallic alloys have been designed and synthesized. Compared to single-metal catalysts, bimetallic alloys offer improved catalytic performance by leveraging the synergistic effect between the two metals. This synergy enhances the surface electronic and geometric structures of the alloy, leading to better adsorption and activation of reactant molecules.

Due to the substantial difference in electronegativity, alloys can exhibit the geometric or ensemble effect of one metal diluted by the other and the “ligand effect” of electronic interactions [127]. Initially, the “ligand effect” referred to the second metal altering the catalytic or adsorptive properties of a bimetallic catalyst by modifying the electronic state of the parent metal [128]. Conversely, the “electronic effect” tended to refer to electron transfer or a charge re-arrangement during the catalytic process, resulting in electronic coupling effects [129]. Currently, the “ligand effect” is frequently equated with the electronic effect, both denoting the modulation of electronic structures to improve catalytic performance. For example, in CoRu alloys, Zhao et al. [130] found that the Ru-H bond forms metal–AB active intermediates due to the higher electronegativity of Ru, facilitating B-H bond activation and achieving a high TOF of up to 814.7 min⁻¹. Moreover, research indicates that a greater electronegativity difference between two metal elements favors electron enrichment. For instance, Kang et al. [131] encapsulated Au-M (M = Co, Ni, Cu) nanoalloys within the zeolitic framework ZIF-8, constructing their adsorption structure (Figure 12a). AB is expected to transfer a hydride ligand to either a Ni or Au site, where it forms an electrostatic bond with an acidic H atom of water, promoting easier O-H bond cleavage compared to free water. Wang et al. [132] synthesized a series of Co-based, non-noble metal alloy NPs, including Fe, Ni, Cu, and Zn, for the hydrolysis of AB. The d-band center and hydrogen generation rates of these Co-based alloys exhibit a volcano-type trend (Figure 12b), indicating that a catalyst with an optimal εd position can effectively adsorb and desorb H₂O molecules. Furukawa et al. [133] investigated a series of Co-based bimetallic alloys supported on SiO₂, finding that catalyst activity varied significantly with the type of second metal (Co-Ge > Co-Ga >> Co >> Co-Sn). They hypothesized that the difference in electronegativity between the two metals causes electronic transfer upon alloying, which is crucial to the superior performance observed. This variation in electronegativity can finely tune the electronic structure.

In addition to the electronic effect, geometric and ensemble effects can also regulate the electronic structure of catalysts, thereby influencing the adsorption and activation states of reactant molecules on the catalyst surface, leading to enhanced catalytic performance [134,135]. The optimal spatial distribution of active centers on the surface is crucial for facilitating specific catalytic reactions in metal catalysts. This spatial distribution is referred to as the geometric effect [136]. When the adsorption or activation on the alloy surface requires multiple sites, this geometric effect extends into an “ensemble effect”. For instance, Li et al. [137] doped trace amounts of precious metal Ru into Co to study their synergistic effect in AB hydrolysis, achieving a TOF of 754 min⁻¹ with the resulting Ru_0.075Co_0.925/NPC catalyst. As illustrated in Figure 12c, the enhanced catalytic activity is hypothesized to result from the random stacking of a rigid geometric structure and the ensemble effect of active clusters with more unfilled d-electron layers in the bimetallic RuCo alloy. Similarly, Abutaleb et al. [138] synthesized bimetallic Ni-Mn-decorated graphite nanofiber catalysts using electrospinning, achieving an initial TOF twice that of a manganese-free catalyst. The synergistic effects of strain and ligand interactions between the bimetals result in high catalytic activity. Unfortunately, it is challenging to accurately distinguish the dominant role of geometric and electronic effects in alloy catalysts due to the interactions between active atoms and their neighbors. Recent studies have suggested decoupling these effects through orthogonal experiments. For example, in 2019, Wang et al. [139], using atomic layer deposition, were the first to regulate electronic and geometric effects within the same catalytic system of Pd-catalyzed aerobic oxidation of benzyl alcohol, finding that geometric effect dominates the catalytic activity of larger particles, while electronic effect dominates smaller particles. Recently, Yang et al. [140] innovatively examined the impact of various Pt-based catalyst compositions on the dehydrogenation performance of AB hydrolysis. The Pt₃Cu alloy, with its optimal geometric structure and d-band center, demonstrated excellent performance, where the geometric effect outweighed the electronic effect (Figure 12d). As the proportion of Cu increased, the electronic effect began to dominate. Further in-depth research is required to elucidate the individual and combined impacts of electronic and geometric effects on the hydrolysis mechanism of AB. Such studies will facilitate a deeper understanding of the relationship between these effects and alloy composition and structure, thereby advancing the design of optimized catalysts.

Figure 12. (a) Calculated (DFT) adsorption structures of Au@ZIF-8 (left), Ni@ZIF-8 (middle), and Au_0.5Ni_0.5@ZIF-8 (right) [131]. (b) Relationships between TOF versus the εd position for Co_xCu_1−x [132]. (c) Ensemble effect in RuCo alloy-catalyzed AB hydrolysis [137]. (d) DFT-calculated relationship between d-band center and HER [140]. (e) Structure and electron density of Ru_0.6Co_0.4/P25 [141].

Figure 12. (a) Calculated (DFT) adsorption structures of Au@ZIF-8 (left), Ni@ZIF-8 (middle), and Au_0.5Ni_0.5@ZIF-8 (right) [131]. (b) Relationships between TOF versus the εd position for Co_xCu_1−x [132]. (c) Ensemble effect in RuCo alloy-catalyzed AB hydrolysis [137]. (d) DFT-calculated relationship between d-band center and HER [140]. (e) Structure and electron density of Ru_0.6Co_0.4/P25 [141].

Moreover, the optimized electronic structure within alloys can simultaneously activate both NH₃BH₃ and H₂O. For instance, Wen et al. [141] doped Co into Ru, causing electron transfer from Co to Ru. This electron transfer compressed the Ru lattice and lowered its d-band center, enhancing the activation of both NH₃BH₃ and H₂O (Figure 12e). The CoCu-NC-5 catalyst, prepared by Song et al., exhibited a TOF value in AB hydrolysis seven times higher than that of a single Co catalyst [142]. The active component structure at the CuCoO/Co interface in CoCu-NC-5 promoted the activation and dissociation of water molecules, leading to this enhanced performance. Xu et al. [143] discovered that the synergy of P-bridged Fe-X-Co coupled sites and the optimized electronic structure due to Fe doping can significantly improve the chemisorption and dissociation of NH₃BH₃ and H₂O molecules, thereby reducing the reaction barriers. This finding underscores the effectiveness of P-bridged Fe-X-Co coupled sites in hollow carbon spheres for efficient hydrogen generation. Other examples include Ru₂Fe₁/N-C [144] and Pt_0.1%Co_3%/TiO₂ [145], which demonstrate similar enhancements in catalytic performance, highlighting the role of dual-molecule activation in promoting efficiency.

In summary, the precise control of the composition and structure of bimetallic alloys enables the modulation of surface adsorption and activation states of catalysts through geometric or electronic changes induced by interactions between different metal atoms. This fine-tuning effectively stimulates the positive synergistic effect between metals, reducing activation energy. This synergistic effect may arise from the electronic effect (ligand effect), geometric effect (ensemble effect), strain effect, and bimolecular activation effect. These effects influence the electronic state of the metal center, including the position of the d-band center, subsequently affecting the adsorption and activation states of reactants by altering the electronic structure, coordination environment, or structural strain of the metal atoms. Generally, an elevated d-band center of metal atoms after alloying enhances the adsorption of AB or water molecules, thereby increasing catalytic activity. Nevertheless, excessive adsorption can hinder reactant desorption, decreasing catalytic activity, which explains the volcanic trend observed in some catalytic activities relative to the d-band position.

3.2.2. Optimization of Adsorption and Activation of Reactants by Polymetallic Alloys

Polymetallic catalysts have garnered significant attention due to their tunable electronic structures and enhanced catalytic properties. By precisely adjusting the composition of different metal elements, researchers have developed a range of polymetallic alloys that exhibit unique synergistic effects, optimizing both adsorption and activation of reactants. These alloys can be broadly categorized into noble-metal-containing and non-noble metal systems, each offering distinct catalytic advantages.

Many noble-metal-containing alloys have been found to exhibit superior catalytic activity in AB hydrolysis, attributed to their optimized electronic structure and enhanced charge transfer properties. For example, Fu et al. [146] constructed a monophase Pt₇₆Au₁₂Co₁₂ trimetallic nanoalloy with a TOF value of up to 450 min⁻¹ and an E_a as low as 18.47 kJ/mol, achieving one of the highest activities ever reported for a carrier-free catalyst. XPS and ICP measurements (Figure 13a) revealed that modified electronic interactions and enhanced charge transfer capabilities are key to the excellent catalytic activity of the Pt₇₆Au₁₂Co₁₂ nanoalloy. Similarly, Xu et al. [147] demonstrated that Pt₇₆Au₁₂Co₁₂/RGO catalysts achieved a TOF of 854.0 min⁻¹, surpassing most Pt-based monometallic catalysts. X-ray absorption spectroscopy (XAS) spectra (Figure 13b) attribute this high activity to significant internal synergistic effects among the constituent components, which contribute to an optimized electronic structure. In this system, platinum serves as the primary catalytic center, cobalt acts as an electron donor to form Co-Pt bonds, and copper modulates the chemical state of platinum by adjusting the electron density around cobalt. Liu et al. [148] attributed the exceptional catalytic activity of Pt/CoCu-NC to the synergistic effect of multiple active sites: Pt sites facilitate NH₃BH₃ activation, while Co/Cu sites enhance H₂O activation (Figure 13c). The Co–Cu heterostructure further accelerates the rate-determining step, significantly boosting hydrogen evolution. Beyond Pt-based systems, Zhao et al. [149] demonstrated that NiPdMo NPs achieve significant catalytic activity, characterized by a low activation energy of 52.3 kJ/mol and a high TOF of 252.7 min⁻¹. The observed enhancement in catalytic performance is attributed to the synergistic effects among Ni, Pd, and Mo, which optimize the electronic interactions essential for improved reactivity. Jin et al. [150] developed a multi-component alloy, np-RuNiFeCo, with exceptional catalytic performance for the hydrolysis of AB. Density functional theory (DFT) studies revealed that the synergistic interaction between Ni, Fe, and Co promotes electron transfer to Ru, enhancing the adsorption of NH₃BH₃ and H₂O molecules and reducing the activation energy of H₂O. Li et al. [151] developed a Rh_0.8Ru_0.2Ni_0.25@MMT-S catalyst, where the alloy structure significantly enhanced the adsorption and activation of H₂O molecules, facilitating the rate-determining step and reducing the activation energy in AB hydrolysis.

Given sustainability and economic considerations, researchers have increasingly explored non-noble metal polymetallic alloys, leveraging their intrinsic synergistic interactions to enhance catalytic activity. For instance, Liang et al. [152] proposed that Cu_0.8Ni_0.1Co_0.1@MIL-101 alloy exhibits superior catalytic activity due to the synergistic effect among its three metal components, far surpassing monometallic and bimetallic alloys. This synergistic effect is mainly attributed to the system’s electronic effect, where Cu acts as an electron donor. The activity of this ternary system exhibits a “V” shape as the molar ratio of copper increases (Figure 13d). As D-band electrons flow from Cu and Co to Ni, the resultant increase in holes within Ni reduces the electron density, facilitating the formation of the active metal–H intermediate with Cu and thereby enhancing catalytic activity. Similarly, Li et al. [60] believe that the substantial increase in the catalytic activity of Cu₆Fe_0.8Co_3.2@MIL-101 in AB hydrolysis can also be attributed to the synergistic effects among the Cu-Fe-Co trimetallic components. These effects mainly include electronic effects and geometric effects. Given the similar electronegativities of Cu (1.9), Fe (1.83), and Co (1.88), the authors suggest that geometric effects are the predominant factors influencing catalytic activity. This enhanced synergistic effect may be related to the composition and proportion of metals in the alloy. Nevertheless, a clear explanation of the specific roles of each component and the synergistic interactions among the components in polymetallic catalysts is still lacking. Further research is needed to elucidate these mechanisms in detail.

In summary, polymetallic alloys with tunable electronic structures often exhibit superior catalytic activity compared to monometallic and bimetallic alloys. The internal synergistic effects, primarily due to electronic and geometric interactions between metal nanostructures, are crucial for their enhanced performance. The composition and proportion of metals significantly influence these effects. Nevertheless, the underlying mechanisms that govern these interactions and the precise contributions of each metal within the alloy still necessitate comprehensive investigation. Moreover, while certain metal promoters are known to augment hydrolysis activity, others may conversely impede the hydrolysis of AB.

3.2.3. Optimization of Adsorption and Activation of Reactants by Metal–Support Interaction

As previously discussed, charge transfer in alloy catalysts is crucial for regulating reactant adsorption and desorption. This charge transfer also commonly occurs at the metal–support interface. Under reducing conditions, strong interactions between the support and metal NPs (MSIs) can form at this interface. MSIs significantly affect the electronic properties and surface geometric structure, potentially leading to synergistic effects that influence the catalyst’s activity and stability [153]. However, not all supports generate beneficial MSIs, as this depends on the chemical properties, surface structure of the support, and the characteristics of the metal NPs. Generally, reducible oxides can form abundant surface oxygen vacancies and electronic metal–support interactions, thereby promoting the hydrolysis of AB by activating the O-H bond in water. Moreover, other supports, such as carbon materials and zeolites, have also been shown to exhibit MSIs, synergistically enhancing AB hydrolysis. This section will discuss these supports in detail.

(a)

Oxide Supports

Because the interactions between oxide supports and metals can be simultaneously affected by size, strain, and shape effects, researchers typically limit their studies to a few empirically selected reducible metal oxides when investigating the electronic properties of metals. Common oxide supports include WO₃, Co₃O₄, Fe₃O₄, CoFe₂O₄, SiO₂, Al₂O₃, and TiO₂.

Co₃O₄, with its weak reducibility, often forms strong interactions with metal NPs. For example, Zhang et al. [154] synthesized Pt/Co₃O₄ nanocrystals (NCs) that showed remarkable catalytic activity, with a TOF of 721 min⁻¹. This high activity was attributed to the synergistic effects between MSI and different active sites, effectively promoting water molecule activation (Figure 14a). Li et al. [155] proposed that the synergy between Co, Co₃O₄, and TiO₂ reduces the activation energy for water, facilitating its dissociation and activation, thus increasing the hydrogen evolution rate significantly (Figure 14a). In addition, oxygen vacancies (VOs) on Co₃O₄ further enhance H₂O adsorption and dissociation [156]. These vacancies adjust the electronic structure of the catalyst rather than acting directly as active sites [157]. Increasing oxygen vacancies modulates the local charge distribution and electronic energy levels near defect sites, leading to electron transfer from the metal oxide to the metal [158,159,160]. Tian et al. [161] discovered that Co defect sites enable adjacent O atoms to share more electrons with other Co atoms, aiding H₂O adsorption on Co₃O₄, resulting in optimal hydrogen evolution rates (Figure 14b). Shen et al. [162] enhanced water dissociation on the Ru catalyst by establishing an oxygen vacancy (VO)–Ti ensemble engineering on the Ru catalyst. This strategy promoted electron enrichment on Ru and electron transfer between Ru and the support, facilitating water molecule activation and dissociation (Figure 14c). They also introduced an artificial atom (RuPd-TiO₂-VO) leveraging VOs and d-orbital coupling effects for efficient water dissociation [163]. The d-orbital coupling between Ru and Pd mimics Rh’s outer electron structure, while electron transfer from VOs to RuPd creates electron-rich active sites, further enhancing water dissociation. Similar enhancements were seen in Co₃O₄ nanocrystals with oxygen vacancies embedded in graphitic carbon nitride (Co-CN-O-100) [164] and nitrogen-doped carbon-coated CuCoMo nanoparticles (CoCu₁Mo₃-NC-O-15) [165].

In addition to Co₃O₄, oxide supports that can generate metal–support interactions include Fe₃O₄ [166], CoFe₂O₄ [167], SiO₂ [168], and MoO₃ [169]. Typically, metal NPs catalysts supported on non-reducible oxides such as SiO₂ and Al₂O₃ do not exhibit strong metal–support interactions. Nonetheless, Zhu et al. [170] performed an XPS analysis on core–shell composites (Pd@Ag/SiO₂) supported on porous silica. They identified electronic synergistic effects between the metal and support, which facilitated H adsorption and formed Pd-H species, thereby enhancing catalytic activity. While Fe₃O₄ and CoFe₂O₄ are typically considered inert, they can form strong interactions with metal NPs [171]. The resulting magnetic nanocatalysts often exhibit high catalytic activity and excellent reusability [172]. For instance, Li et al. [173] developed oxygen-vacancy-rich Rh/CoFe₂O₄-SB-H₂ structures. The strong electronic metal–support interaction and the synergistic effects of the oxygen vacancies significantly enhanced the catalytic performance of Rh/CoFe₂O₄-SB-H₂ for hydrogen production, achieving a TOF of 1894 min⁻¹ and retaining 75% of its initial activity after 10 cycles (Figure 14d).

(b)

Carbon Supports

The inherent chemical inertness of carbon-based materials, such as porous carbon [174], nitrogen-doped carbon [175], and MXene [176], uniquely positions them as highly effective supports for isolating variables like particle size effects and strain effects. For example, Ye et al. [177] attributed the superior catalytic activity of Ru@Co-NC to the electronic metal–support interaction (EMSI) between ultrafine Ru NPs and Co-NC, which facilitates the oxidative cleavage of the O-H bond in H₂O molecules—a mechanism illustrated in Figure 14e. Similar metal–support interaction are observed with nitrogen-doped carbon, where its exceptional conductivity and large surface area significantly enhance electron transfer [178]. Wang et al. [88] proposed that the strong electronic interaction between Ni-MoOx and (P)NCS induces electron-rich Ni species, which serve as the key factor contributing to the superior catalytic performance of Ni-MoOx/(P)NCS.

Furthermore, researchers have demonstrated that the structural properties of MXene and metals can be modulated through metal–support interaction. Slot et al. [179] reported that the catalytic performance of Rh/MXene is due to the favorable interactions between Rh(0) NPs and MXene. Karataş et al. [180] found that the metal–support interaction between Ru NPs and MXene led to a high TOF value of 582 min⁻¹. Additionally, recent studies have explored strategies for modulating MSI through surface functionalization and doping. Qin et al. [181] developed an amine-functionalized Ti₃C₂ (CuCo/PDA-Ti₃C₂) to immobilize CuCo NPs, inducing strong metal–support interaction (SMSI). This regulates local charge distribution, optimizes d-orbital centers, and enhances adsorption/desorption, accelerating O-H bond cleavage in water. Hou et al. [182] demonstrated that tuning -F and -OH groups on Ti₃C₂T_x modulates the electronic structure of metal NPs, facilitating electron transfer between Ti₃C₂T_x and Ni. This process generates electron-rich Ni and electron-deficient Ti₃C₂T_x, which synergistically activate AB and H₂O while enhancing reactant adsorption and lowering the reaction barrier (Figure 14f). Bian et al. [183] developed V-doped Ru/Ti_2.5V_0.5C₂ catalysts, where V doping promotes electron transfer from Ti to Ru, yielding electron-rich Ru species that enhance AB activation and weaken the B-H bond, while V-doped Ti₃C₂ improves H₂O activation. However, precisely tailoring metal electronic structures via carbon surface chemistry remains challenging. The influence of heteroatomic functional groups and dopants on catalytic performance is complex, and decoupling electronic effects from structural and chemical contributions remains a key challenge in catalysis research.

(c)

Other Supports

Several studies have demonstrated that supports with distinctive microstructures can synergistically interact with metals as well as their alloys, thereby enhancing the activation and dissociation of water molecules and significantly improving the catalytic performance for AB hydrolysis. For instance, zeolites with tunable acidity have shown such synergistic effects. Wang et al. [184] reported Rh_0.8Ru_0.2/SP-ZSM-5-100, with its tunable acidic zeolite structure, achieved an ultra-high TOF of 1006 min⁻¹ due to the synergistic interaction between bimetallic Rh-Ru clusters and the Brønsted acid sites of the zeolite. This performance was comparable to that of zeolite-supported single-atom Rh catalysts. Similarly, Wei et al. [185] utilized a core–shell structure of carbon-coated zeolite to support Ru NPs. DFT calculations indicated that the metal–support interaction between Ru and the zeolite enhanced the activation of AB and H₂O, effectively reducing the reaction energy barrier. Moreover, metal–organic frameworks (MOFs) with ordered porous structures have also exhibited synergistic metal–support interactions. Xu et al. [186] suggested that metal NPs/MOF catalysts with dual active centers substantially enhance catalytic performance, attributed to electron aggregation at the interface (Figure 14g). Furthermore, Liu et al. [187] proposed that the strong electronic interaction between COF and ultrafine PtCo₂ NCs can accelerate the transfer of photogenerated electrons from COF to PtCo₂ NCs, promoting the adsorption and activation of NH₃BH₃ and H₂O molecules, which accelerates hydrogen release.

Figure 14. (a) AB hydrolysis mechanism by CoCoO_x/TiO₂@N-C (COTC) [155]. (b) Comparative performance of Co-based and transition metal oxide catalysts in AB dehydrogenation [161]. (c) AB hydrolysis mechanism by TiO₂-VO-Ru [162]. (d) Remaining catalytic activity of initial Rh/CoFe₂O₄-SB-H₂ in each cycle [173]. (e) AB hydrolysis mechanism by Ru@Co-NC [177]. (f) AB hydrolysis mechanism by Ni/Ti₃C₂T_x−4 [182]. (g) Differential charge density of the interfaces between Cu NPs and Co-MOFs [186].

Figure 14. (a) AB hydrolysis mechanism by CoCoO_x/TiO₂@N-C (COTC) [155]. (b) Comparative performance of Co-based and transition metal oxide catalysts in AB dehydrogenation [161]. (c) AB hydrolysis mechanism by TiO₂-VO-Ru [162]. (d) Remaining catalytic activity of initial Rh/CoFe₂O₄-SB-H₂ in each cycle [173]. (e) AB hydrolysis mechanism by Ru@Co-NC [177]. (f) AB hydrolysis mechanism by Ni/Ti₃C₂T_x−4 [182]. (g) Differential charge density of the interfaces between Cu NPs and Co-MOFs [186].

On the other hand, designing complex multi-component metal-based catalysts by integrating various active components into unique structures is one of the most common strategies for enhancing the catalytic kinetics of AB hydrolysis catalysts. These hybrid systems, due to their structural and compositional advantages, can effectively modulate the electronic structure of the active centers, optimizing their interaction with AB molecules. Xu et al. [188] attributed the high activity of CuCo supported on zirconia/nitrogen-doped porous carbon/reduced graphene oxide (ZrO₂/NC/RGO) to the strong synergy between ZrO₂ and N species, which optimizes the d-band center and Fermi level of CuNi sites, enhancing ABH kinetics. Chen et al. [189] developed a NiCoP/CoP heterostructure catalyst with 3D porous nano-architectures. The interactions between NiCoP and CoP, combined with the unique structure, significantly enhanced catalytic performance, achieving a hydrogen generation rate of 6345 mL⋅min⁻¹⋅g⁻¹ and an activation energy of 25.89 kJ⋅mol⁻¹, while maintaining excellent stability. Likewise, catalysts like CuCo(O) supported on a carbon–nitrogen framework (CuCo(O)@CN) [190] and Co₂P/(0.59-Cu₃P) constructed on CoZn-ZIF-derived porous N-doped carbon [191] also exhibit synergistic compositional and structural advantages that activate AB and H₂O molecules and reduce their activation energy barriers.

In summary, the metal–support interaction (MSI) in heterogeneous metal catalysts plays a pivotal role in modulating electronic structures. This interaction can occur on reducible oxide supports and carbon supports, often leading to bonding between the metal and support, interfacial charge transfer, and alterations in metal structure. These changes can affect the adsorption and activation of reactant molecules, lowering the energy barriers, and promoting hydrolysis reactions. Additionally, tuning oxygen vacancies on oxide supports and functional groups on some carbon supports can enhance electron transfer, inducing metal–support interaction. This, in turn, promotes the adsorption and activation of water molecules, leading to their dissociation. Besides reducible oxide supports and certain carbon supports that readily undergo metal–support interactions, some multi-component hybrid catalysts and supports can also effectively activate H₂O molecules through synergistic metal–support interactions, thereby influencing the activity and stability of the catalysts.

3.3. Strategies for Performance Optimization Based on Hydrogen Desorption

The final step of the AB hydrolysis reaction is the generation and desorption of hydrogen gas. Two activated hydrogen atoms (H*) combine to form H₂, which then desorbs from the surface. Efficient hydrogen desorption helps to clear the hydrogen coverage layer from the catalyst surface, reducing agglomeration caused by surface hydrogen bonding interactions between catalyst particles, thus exposing more active sites. This section will discuss the optimization of hydrogen desorption performance in the hydrolysis of AB.

From the perspective of reaction dynamics, the adsorption of reactants provides a platform for the formation of intermediates, while the desorption of products frees active sites, allowing the continuous adsorption of new reactants. As previously discussed, the positive synergistic effects between alloy catalysts enhance the adsorption and activation of reactants. Additionally, this synergy can facilitate the adsorption and desorption of H* intermediates. Wen et al. [141] suggested that the optimal adsorption strength of H* intermediates is responsible for the high catalytic activity of the RuCo alloy. As shown in Figure 15, catalytic activity is inversely correlated with H adsorption energy (EH). Ru displays strong hydrogen adsorption, impeding desorption and subsequent reactions. Conversely, Co’s weaker H adsorption lowers EH with increasing Co content, enhancing catalytic activity. Similarly, Yang et al. [192] encapsulated monodisperse PdRu NPs in a gel network, creating Pd_xRu_yNP@Alyne-PVA gel, which achieved a TOF of 578.2 min⁻¹ in an alkaline solution. This performance is due to the strong synergistic effects between Pd and Ru NPs; Pd effectively captures hydrogen, while Ru facilitates hydrogen dissociation, promoting H₂ desorption from the catalyst surface. Recently, Guan et al. [64] developed a RuPt-Ti multi-site catalyst that optimizes H* intermediate adsorption through the multi-site synergistic effect of Ti, accelerating the hydrolysis reaction.

Studies have indicated that desorption is an endothermic process; the lower the heat absorbed, the more facile the desorption, which hinges on the catalyst’s moderate hydrogen affinity [193]. Excessive affinity results in high desorption energy barriers, whereas insufficient affinity leads to low catalytic activity. Additionally, H₂ adsorbed on the catalyst surface in molecular form is more easily desorbed. For example, Yan et al. [194] demonstrated that, compared to Pt₁ single atoms, the higher energy 5d orbital vacancies in Pt₂ dimers provide moderate adsorption energy, facilitating molecular hydrogen adsorption and desorption during AB hydrolysis, thus enhancing catalytic activity. From another perspective, the adsorption/desorption of hydrogen on the catalyst is highly dependent on the appropriate binding energy between the metal and H atoms [195]. Optimal hydrogen production is achieved when the standard free energy of hydrogen adsorption approaches zero. For instance, the hydrogen adsorption free energy of nickel phosphide and cobalt phosphide is currently known to be the closest to zero among catalysts, except for Pt, often exhibiting favorable hydrogen desorption performance. Examples include the heterostructure catalyst CoP/Co₂P [196], where charge transfer within the heterostructure promotes the adsorption of reactants and hydrogen desorption in AB hydrolysis. Zhang et al. [197] partially replaced oxygen with phosphorus in CoO_0.5P_0.5@CS-n, where the high surface charge distribution near P facilitates hydrogen desorption in the CoO_0.5P_0.5@CS-n catalyst.

In summary, effectively controlling hydrogen desorption not only preserves the dispersion of catalyst particles but also increases the availability of active sites on the catalyst surface, thus directly enhancing catalytic activity and efficiency. Hydrogen desorption is inherently an endothermic process, and its efficiency primarily depends on the catalyst’s moderate affinity for hydrogen. Hydrogen that is adsorbed in molecular form desorbs more readily. Furthermore, the processes of reactant adsorption and product desorption are mutually reinforcing. Enhancing the desorption of reactants can be achieved by optimizing the catalyst’s structure and composition, especially by harnessing the positive synergistic effects of metal elements within the alloy, which is essential for maintaining active sites and facilitating continuous catalytic reactions.

3.4. Strategies for Controlling Catalytic Performance

To efficiently utilize the produced hydrogen and mitigate the safety risks and high costs associated with hydrogen storage and transportation, designing “switches” for on-demand hydrogen production is crucial [198]. Here, this section summarizes conventional methods for achieving on-demand hydrogen generation via AB hydrolysis, which primarily includes four modes: temperature-regulated switches, pH-regulated switches, catalyst-transfer switches, and other innovative catalyst switches.

Firstly, since the activation energy (E_a) is positively correlated with the reaction temperature, temperature-regulated switches for controlled hydrogen evolution can generally be achieved by varying the reaction solution’s temperature. For instance, Wegner et al. [199] reported a metal-free bis(borane) Lewis acid catalyst that enables controlled dehydrogenation of AB by varying the temperature between 60 °C and room temperature. Additionally, recent advancements have highlighted the effectiveness of temperature-responsive polymer materials in such regulatory mechanisms. Specifically, Huang et al. [200] developed a luminescent, thermo-responsive catalyst based on SiO₂@Pt@PABI-Tb@PNIPAM, whose “on-off” state can be visually observed by its green luminescence. At room temperature, the swollen PNIPAM allows AB to approach the catalyst’s active sites, presenting an “on” state; at higher temperatures, the shrunken PNIPAM could inhibit the access of AB to active Pt NPs, thereby switching to an “off” state. Subsequently, the research group further optimized the thermo-responsive and visualization characteristics by creating a smart thermo-controlled catalyst, PNIPAM-coated MS@Pt@EuW10 [201]. As depicted in Figure 16a, this design leverages the temperature-responsive properties of PNIPAM, with the catalyst being in an “on” state at room temperature and in an “off” state at higher temperatures, successfully overcoming the application limitations caused by irreversible pore size changes in mesoporous hybrid catalysts.

A pH-regulated switch can effectively control hydrogen production by introducing H⁺ or OH⁻ into the reaction medium. Early studies indicated that increasing the concentration of H⁺ could accelerate the hydrolysis of AB. Nevertheless, in 2017, Astruc et al. [202] suggested that H⁺ might inhibit H₂ evolution, whereas OH⁻ could play a positive role. Subsequent investigations have focused on the role of OH⁻ in AB hydrolysis, with recent studies confirming that the addition of an appropriate amount of OH⁻ can effectively enhance AB hydrolysis [203]. This is because OH⁻ can act as a metal promoter, increasing the electron density of the NPs through coordination with their surface, thus facilitating the oxidative addition of the O-H bond in water molecules. However, an excessive amount of hydroxide can lead to competitive adsorption with water molecules near the catalyst’s active sites, ultimately inhibiting the hydrolysis process. Moreover, the addition of an alkaline can corrode the reaction vessel and pose significant challenges for subsequent waste liquid treatment. Compared to the direct addition of bases, researchers might consider utilizing alkaline supports or modifying catalysts with alkaline functional moieties.

Catalyst conversion switches play a pivotal role in on-demand hydrogen evolution from AB hydrolysis. Magnetic catalysts are especially effective, enhancing hydrogen production upon contact with reactants and easily removable from the reaction via magnetic or mechanical methods to halt hydrogen evolution [204,205,206]. Additionally, monolithic catalysts like foam nickel (NF) offer significant benefits in achieving controlled catalyst release. For instance, Asim et al. [207] used Co-doped nickel phosphide supported on foam nickel (NiCoP/NF), achieving the start or stop of the reaction simply by immersing or removing the NiCoP/NF from the reactor (Figure 16b). Luo et al. [208] documented the synthesis of the Ru-Ni-NF catalyst, emphasizing its facile separability and enhanced recyclability relative to conventional powdered carbon-supported catalysts. Similarly, other monolithic structures such as foam copper [209], three-dimensional nanoarrays [210], and carbon cloth [211], enable efficient and controlled hydrogen evolution.

In addition to the previously discussed “switches”, alternative approaches have also proven effective for achieving controllable hydrogen evolution from AB hydrolysis. For instance, photocatalytic regulation has emerged as a promising method, Zhang et al. [212] demonstrated that porous, broad-spectrum light-responsive graphitic carbon nitride nanosheets (CNNs) supporting Co and Ni NPs can serve as a photocatalytic switch in AB hydrolysis. Another effective strategy involves adding inhibitors to the reaction solution. Fortunately, researchers have found that Zn²⁺ ions exhibit stronger adsorption energy than AB on certain metal surfaces, allowing the introduced Zn^2‫ ions to occupy active sites, thereby inhibiting AB adsorption and suppressing its hydrolysis reaction. Moreover, the adsorbed Zn^2‫ ions can readily coordinate with EDTA²⁻, freeing the occupied active sites for reactant adsorption and thus reactivating the hydrolytic dehydrogenation reaction. Building upon this concept, Chen et al. [213] designed a multifunctional catalytic platform based on Rh@GQDs, which for the first time achieved switch control, multistep control, and stepless speed control of H₂ release, as depicted in Figure 16c. Similarly, Mi et al. [214] innovatively proposed a method to control hydrogen release in the AB@Co/HNTA solid-state system by regulating the amount of water added as depicted in Figure 16d. After complete hydrolysis of AB, the Co/HNTA can be reused to encapsulate AB, forming a large-volume solid composite for subsequent hydrolysis reactions. To date, the Zn^2‫/EDTA²⁻ system has been successfully employed for “on-off” control of AB hydrolysis across various catalytic systems, including Pt/ZIF-67 [215], Ru/MoAl_1−xB [216].

Figure 16. (a) Schematic illustration of catalysts at different temperatures [200]. (b) On-demand hydrogen generation using 4-NiCoP/NF as an on/off switch [207]. (c) Mechanism of controlled H₂ evolution by Rh@GQDs [213]. (d) AB@Co/HNTA applied in vehicle applications for controlled hydrogen release [214].

Figure 16. (a) Schematic illustration of catalysts at different temperatures [200]. (b) On-demand hydrogen generation using 4-NiCoP/NF as an on/off switch [207]. (c) Mechanism of controlled H₂ evolution by Rh@GQDs [213]. (d) AB@Co/HNTA applied in vehicle applications for controlled hydrogen release [214].

In summary, the controllable hydrogen production from AB hydrolysis is primarily achieved by altering the properties of the solution or by separating the metal catalyst from the solution. The former typically involves controlling the reaction rate by changing the solution properties, such as adjusting the temperature, pH, or light conditions. The latter involves the meticulous design of the catalytic system, particularly the metal catalysts. This controllable hydrogen release allows the hydrogen production process to be turned on or off based on market demand, thus enhancing safety and reducing the overall operational costs.

Table 1. Performance indicators for noble and non-noble metal catalysts.

Catalyst	TOF (min−1)/ (mL·min−1·gCat−1)	Temperature (K)	nmetal/nAB	Ea (kJ/mol)	Particle Size (nm)	Preparation Method	Durability	Reference
Rh/C-300A-350H	3308	298	—	35	1.65	Low-temperature oxidative thermal redispersion strategy	40.4%/5	[30]
Ru/VO-Co3O4	2114	298	—	58.8	2.8	—	92.5%/6	[31]
Pd/alk-Ti3C2	230.6	298	—	21.2	4.9	Direct reduction method	40%/5	[32]
Rh/C	1246	298	—	40.9	—	In situ reduction	61.2%/8	[37]
Ru/HPCM	440	303	—	43	1.41	Iron citrate pyrolysis	50%/8	[38]
Co/FeCeO2-0.6	92.8	298	—	—	10	—	—	[47]
Cu0.6Co0.4O@CN	57.5	298	—	38.4	7	—	—	[48]
Ru–MgO/HBC	784	298	—	50.09	1.7	Wet impregnation method	89.9%/10	[49]
1.5Co1.5Ni/α-MoC	321.1	298	—	—	—	Impregnation method	—	[51]
Ni0.7Co1.3P/GO	153.9	298	0.026	43.2	5	Two-step strategy	95.2%/7	[52]
CuMoO4−CoMoO4	104.7	298	—	38.4	2–3 μm	Template-free approach	—	[53]
Ni/FeNiOx-25	72.3	303	—	39.18	1–2	—	100%/6	[54]
Pt/CNTs-O-HT	567	303	0.0047	—	1.3	—	—	[56]
Co/CTF-1	42.3	298	0.05	42.7	7.3	—	—	[57]
Ni2Pt@ZIF-8	600	293	—	23.3	2	Co-reduction method	—	[58]
Rh1/VO2	—	—	—	38.7	—	—	—	[59]
CuFeCo@MIL-101	23.2	298	0.073	37.1	2.6	Impregnation–reduction method	60%/7	[60]
RuMoP@MOF-199	753.6	298	—	46.9	2.1	Liquid impregnation method	79%/5	[61]
Ni1.2Fe0.8@CN-G	23.25	298	—	38.24	4	Pyrolysis	—	[62]
Co3B-CoP/h-BN	37	303	—	51.8	—	—	—	[63]
RuPt-Ti/Ti3C2	1293	293	—	28.6	1.89	Impregnation–reduction method	—	[64]
NiCoP/OPC-300	68.03	298	0.042	38.9	1.2	One-step chemical reduction method	85%/5	[66]
Co-Co3O4/CDS	6816 mLH2·min−1·gCo−1	298	—	—	—	Hydrothermal process	—	[67]
Rh/OPNC	433	298	—	26.4	2.88	Air-mediated pyrolysis method	62%/5	[68]
Ru1Ni1.9/NCS	824	298	—	26.5	2.3	Impregnation–reduction method	67%/5	[69]
Pt/CNT-5W	710	303	—	27.8	1.4	Two-step method	45%/5	[70]
Pd0.1Cu0.9/T-PC	279	298	—	57	2.9	In situ reduction	44.4%/10	[71]
Rh/PCNs	513.2	298	—	46.5	2.3	Simple pyrolysis	66%/6	[77]
Ni−MoOx/(P)NCS	85.7	298	—	29.6	3	Phosphate-mediated method	54%/5	[88]
Ni/CTFPh	14.3	303	—	34.8	2.2	Metal vapor synthesis	—	[89]
Co/CoNx-CNT-33-800T	7833 mLH2 gCo−1min−1	313	0.0654	46.17	—	—	75%/40	[90]
5Pt/G1600-O3-60	618.9	298	—	—	—	Atomic layer deposition	88%/6	[78]
MoO3-doped MnCo2O4 (0.10)	26.4	298	—	34.24	—	In situ synthesis	—	[79]
Ni0.5Cu0.5Co2O4 nanoplatelets	80.2	298	—	28.4	200	—	—	[91]
Ni0.5Cu0.5Co2O4 microspheres	65.1	298	—	29.5	40	—	—	[91]
Ni0.5Cu0.5Co2O4 nanoparticles	45.5	298	—	43.2	50	—	—	[91]
Pt-Ni	302.3	—	—	42.1	—	—	—	[92]
Plate-like Cu2O–CoO	34.1	298	—	—	Thickness of 40 nm	—	—	[93]
Rh/h-NCNWs	1234	298	—	36.94	—	—	60%/5	[94]
Cu0.6Ni0.4Co2O4	119.5	298	—	33.91	—	—	70%/8	[95]
Cu0.5Ni0.5Co2O4 (Mo = 0.10)	195.25	298	—	—	—	—	—	[96]
CoNiP/GO	134.6	298	—	44.12	—	—	84.6%/5	[97]
Co–Fe–B@g-C3N4/NF	14,005 mLH2·min−1·gCat−1	298	—	45	—	Chemical deposition method	—	[98]
Co-Mo-B/NF	6027.1 mLH2·min−1·gCat−1	298	—	43.6	65	Electroless plating method	—	[99]
CuO/Co3O4@C-4	18.8	298	—	18.5	—	Hydrothermal method	—	[100]
Ru/B-U-TiO2	1287	298	—	37.96	—	—	65%/8	[101]
Co0.7Cu0.3@NHPC-800	—	303	—	26.2	—	—	—	[103]
Co@N-C-700	5.6	298		31	9	One-step thermolysis	97.2%/10	[104]
Ru/NPC	813	298	—	24.95	—	In situ reduction	67.3%/5	[105]
Rh/NPC	473.5	298	0.003	40.2	6.03	Pyrolysis method	54.0%/8	[106]
Ru/Ti3C2−xNx	1334	298	—	—	1.54	Microwave heating polyol method	—	[107]
Co0.5Ru0.5/CosNC	1068	298	—	18.96	—	—	100%/10	[80]
Ru/OCB	602	298	—	34.3	2.0	Microwave-assisted solid-state strategy	52%/5	[108]
Pt/MXene-O3	265	303	—	69	0.6	—	—	[109]
Ru/ONC	556	298	—	34.3	1.69	Gas-phase oxidation strategy	—	[81]
Ru@PC-5–700	405.9	303	—	—	1.3	—	58.3%/7	[110]
Ru0.50Ni0.50@WSC	251	303	—	45.3	—	Facile adsorption–NaBH4 reduction method	—	[111]
CoP–CoO/NCDs	89.56	298	—	41	58.82	—	—	[82]
Ru/BNC	1854	298	—	26.31	1.56	—	—	[112]
Rh/N-U-TiO2	721	298	—	20.05	3.28	—	62%/5	[113]
Cu/Cu0.76Co2.24O4-VO	28.46	298	—	24.36	—	—	—	[114]
Rh0.75Co0.25/Ni@ Ni-N-C	223.8	303	—	28.63	3.69	Maceration reduction method	—	[115]
Rh/C-SC	336	298	—	37.1	4.1	—	50%/5	[83]
PtPd3	4034 mL·min−1·gCat−1	298	—	14.56	10	—	70.8%/5	[116]
CF-BT-Ru	322	308	—	32.41	2.6	—	—	[84]
Ru/3DNPC-500	584	298	—	31	1.32	High-temperature pyrolysis	50%/7	[121]
Pd/IPCN	122.8	298	—	29.1	2.17	—	—	[122]
P2-Cu-Co3O4@CNF	35.6	303	—	29.86	—	Nanoconfinement method and a facile ion-doping approach	—	[123]
Cu0.5Ni0.5/h-BN	6.33	303	—	23.02	—	Adsorption–chemical reduction	—	[125]
CPFC-MS@NiAl-LDH@RhxNi1−x	13	298	—	40.3	—	—	—	[127]
Pt-Co/GQDs	520	303	—	45.3	13	—	—	[129]
CoRu0.5/CQDs	814.7	298	—	39.29	4.25	One-step hydrothermal	—	[130]
AuNi@ZIF-8	40	298	—	37.4	—	—	—	[131]
SCo0.43Cu0.57	5.68	298		31.06	18.23	Acid etching method	71.8%/5	[132]
Ni–Zn/SiO2	4.3	298	0.025	—	8.5	—	—	[133]
Ru0.8Ni0.2/g-C3N4-rGO	905	303	0.0016	27.2	1.4	Adsorption–chemical reduction	55%/6	[135]
Ru0.25Pd0.75@g-pC	214.49	303	—	47.3	1.5	One-pot calcination method	62%/18	[136]
Ru0.075Co0.925/NPC	754	298	—	30.5	26.68	Maceration reduction method	56%/5	[137]
NiMn-decorated CNFs	58.2	303	—	38.9	60	—	—	[138]
Ru0.6Co0.4/P25	443.7	298	—	43.9	25	—	—	[141]
CoCu-NC-5	8.12	298	—	34.25	7.95	—	—	[142]
Fe-CoP@C	183.5	298	—	30.6	—	—	—	[143]
Ru2Fe1/N–C	424	298	—	33.7	3	Impregnation–co-reduction	26%/5	[144]
Pt0.1%Co3%/TiO2	1530	298	0.0008	63.8	1.3	Step-by-step reduction method	100%/5	[145]
Pt76Au12Co12	450	298	—	18.47	—	Sequential Digestive reduction	56%/5	[146]
Cu0.4Co0.6 Pt0.0075O/RGO	854	—	—	39.8	2.89	—	80%/8	[147]
Pt1.5/CoCu0.4-NC	1636.82	298	—	41.78	9.31	Liquid-phase reduction method	—	[148]
Ni0.3Pd0.7Mo0.2 NPs	252.7	298	—	52.3	5.93	In situ reduction method	15%/5	[149]
np-RuNiFeCo	148.2	298	—	25.3	4	—	—	[150]
Rh0.8Ru0.2Ni0.25@MMT-S	2961	298	—	29.7	2	Impregnation method	—	[151]
Cu0.8Ni0.1Co0.1 @MIL-101	72.1	298	0.0027	29.1	2.8	Solvent evaporation method	73%/8	[152]
Pt/Co3O4 NCs	721	298	—	31.3	1.2	—	86.6%/10	[154]
Co-CoOx/TiO2@N-C	5905 mL·min−1·gCo−1	298	—	38.5	—	Sol–gel method	85%/5	[155]
Pd0/Co3O4	3048	298	—	62	—	Impregnation/reduction method	100%/10	[156]
Vo–Co–Sn5:2	17.6	298	—	45.95	5–15	Co-precipitation–calcination method	82.6%/14	[157]
1.5-PdTVO	240	298	—	34.6	2	—	—	[158]
A20-Pd5	17.4	—	—	41	20	Solid-state approach	—	[159]
Ru/Co2.28Cu0.72O4/C7.5	2020	298	—	26.3	1.4	—	—	[160]
Vco-Co3O4	934	298	—	32.65	45.9	—	—	[161]
1.5-RTVO-4	1370	298	—	46.3	1.9	—	—	[162]
RuPd-TiO2-VO	2750	298	—	32	2	—	90%/10	[163]
Co-CN-O-100	11,410 mL·min−1·gCat−1	313	—	39.41	—	—	80%/5	[164]
CoCu1Mo3-NC-O-15	24.44	298	—	28.44	—	—	—	[165]
Ru0/SiO2-Fe3O4	127	298	0.0079	54	3.75	Maceration reduction method	100%5	[166]
Ni0/CoFe2O4	38.3	298	—	62.7	—	Two-step impregnation–reduction method	38%/10	[167]
Pt/MoO3−x-500	2268.6	298	—	13.97	1.8	—	—	[169]
Pd0.75@Ag0.25/SiO2	109.99	303	—	42.26	10	Seed-mediated stepwise reduction	83.2%/5	[170]
Rh0/CoFe2O4	720	298	0.00048	66	2.18	Two-step impregnation–reduction method	100%/5	[172]
Rh/CoFe2O4-SB-H2	1894	298	—	59.3	—	Impregnation–reduction method	75%/10	[173]
Ru/PC	744	298	—	39.11	1.5	Salt template-assisted in situ construction	—	[174]
Co–Mo2C/NC	18,876 mLH2·min−1·gCo−1	298	—	49.8	4.71	One-step method	77.6%3	[175]
Pd1Rh4/Ti3C2	338	298	—	33.8	4	Microwave-assisted reduction method	—	[176]
Ru@Co-NC	568	298	—	24.2	2.25	Sodium chloride template method	59.8%/5	[177]
Cu0.9Ni0.1/Ti3C2Tx	2429 h−1	323	—	41.61	1.84	Wet-chemical co-reduction method	95.5%/5	[178]
Ni-MoOx/(P)NC	85.7	298	—	29.6	3	Phosphate-mediated method	54%/5	[88]
Ru/TASC-NaOH	582	303	—	43	1.6	—	—	[179]
Rh/MXene	288.4	298	—	54.2	2.55	Wet impregnation method	99%/7	[180]
CuCo/PDA-Ti3C2	71.8	293	—	45.89	1.8	Surficial alkaline functional strategy	—	[181]
Ni/Ti3C2Tx−4	161	298	0.074	59.3	3.07	—	—	[182]
Ru/Ti2.5V0.5C2	1072	298	—	43	—	—	—	[183]
Rh0.8Ru0.2/ SP-ZSM-5-100	1006	298	0.001	56.5	0.7	Incipient wetness impregnation method	—	[184]
Ru/S-1@C(RSC-2)	892	298	—	36.8	3	In situ reduction method.	—	[185]
Cu0.5@Co0.5-MOF/5	129.8	298	—	26.5	5.5	Hydrothermal method	95%/5	[186]
PtCo2@COF	486	293	—	34.5	2	—	—	[187]
Cu0.8Ni0.2@ZrO2/NC/RGO	40.9	303	—	33.24	2.5	—	—	[188]
NiCoP/CoP	30.3	298	—	25.89	—	Three-step hydrothermal–oxidation–phosphorization	—	[189]
CuCo(O)@CN	12.4	298	—	33.8	30	Thermal reduction	64.7%/5	[190]
Co2P/(0.59-Cu3P)-NC	729.6	298	—	63.5	—	—	—	[191]
Pd1Ru2NPs@Alkyne-PVA gel	247.93	298	—	33.02	2	In situ reduction	—	[192]
Ru1Ni4/APTS-rGO	1559	298	—	37.2	2	One-step in situ co-reduction	49.9%/5	[193]
Pt2Ox	2800	—	—	—	0.96	Bottom-up approach	100%/5	[194]
CoP/Co2Ps	64.1	298	—	38.8	—	Salt-induced phase transformation	100%/10	[196]
CoO0.5P0.5@CS	37	298	—	41	7.87	—	—	[197]
MS@Pt@EuW10@PNIPAM	51.1	298	—	74.89	—	—	—	[201]
NiNPs/ZIF-8	85.7	298	—	28	2.7	Deposition–precipitation (DP) method	—	[202]
Ru1Co9/TiO2	1408	298	—	33.25	—	Co-precipitation and reduction	—	[203]
Pd/CGP-GO-Fe3O4	36.5	303	—	27.4	—	Co-precipitation method	—	[204]
Rh0-Co3O4	1800	298	0.00024	61.7	—	Two-step impregnation–reduction method	100%/5	[205]
CuNi/Co3O4	30.5	298	—	41.8	100 in diameter	Impregnation–reduction method	50%/5	[206]
NiCoP/NF	—	298	—	58.95	—	Low-temperature induced phosphating method	—	[207]
Ru-Ni-NF	539.6	298	—	36.4	—	Spontaneous redox reaction	—	[208]
CFNP@CF foam	12.5	303	—	23.4	20	Electrodeposition	99.5%/3	[209]
CoP NA/Ti	6500 mLH2·min−1·g−1	—	—	41	—	—	100%/20	[210]
Co–Mo–B/CC	3916.1	298	—	25.2	—	Electroless plating	55.9%/5	[211]
Rh@GQDs	469	303	0.002	54.85	2.3	—	—	[213]
Co/HNTA	—	298	—	10.8	3.07	—	—	[214]
Pt/ZIF-67	687	303	—	45.43	2.15	—	—	[215]
Ru/MoAl1−xB	494	303	—	39.2	2.4	—	—	[216]

Table 1. Performance indicators for noble and non-noble metal catalysts.

Catalyst	TOF (min−1)/ (mL·min−1·gCat−1)	Temperature (K)	nmetal/nAB	Ea (kJ/mol)	Particle Size (nm)	Preparation Method	Durability	Reference
Rh/C-300A-350H	3308	298	—	35	1.65	Low-temperature oxidative thermal redispersion strategy	40.4%/5	[30]
Ru/VO-Co3O4	2114	298	—	58.8	2.8	—	92.5%/6	[31]
Pd/alk-Ti3C2	230.6	298	—	21.2	4.9	Direct reduction method	40%/5	[32]
Rh/C	1246	298	—	40.9	—	In situ reduction	61.2%/8	[37]
Ru/HPCM	440	303	—	43	1.41	Iron citrate pyrolysis	50%/8	[38]
Co/FeCeO2-0.6	92.8	298	—	—	10	—	—	[47]
Cu0.6Co0.4O@CN	57.5	298	—	38.4	7	—	—	[48]
Ru–MgO/HBC	784	298	—	50.09	1.7	Wet impregnation method	89.9%/10	[49]
1.5Co1.5Ni/α-MoC	321.1	298	—	—	—	Impregnation method	—	[51]
Ni0.7Co1.3P/GO	153.9	298	0.026	43.2	5	Two-step strategy	95.2%/7	[52]
CuMoO4−CoMoO4	104.7	298	—	38.4	2–3 μm	Template-free approach	—	[53]
Ni/FeNiOx-25	72.3	303	—	39.18	1–2	—	100%/6	[54]
Pt/CNTs-O-HT	567	303	0.0047	—	1.3	—	—	[56]
Co/CTF-1	42.3	298	0.05	42.7	7.3	—	—	[57]
Ni2Pt@ZIF-8	600	293	—	23.3	2	Co-reduction method	—	[58]
Rh1/VO2	—	—	—	38.7	—	—	—	[59]
CuFeCo@MIL-101	23.2	298	0.073	37.1	2.6	Impregnation–reduction method	60%/7	[60]
RuMoP@MOF-199	753.6	298	—	46.9	2.1	Liquid impregnation method	79%/5	[61]
Ni1.2Fe0.8@CN-G	23.25	298	—	38.24	4	Pyrolysis	—	[62]
Co3B-CoP/h-BN	37	303	—	51.8	—	—	—	[63]
RuPt-Ti/Ti3C2	1293	293	—	28.6	1.89	Impregnation–reduction method	—	[64]
NiCoP/OPC-300	68.03	298	0.042	38.9	1.2	One-step chemical reduction method	85%/5	[66]
Co-Co3O4/CDS	6816 mLH2·min−1·gCo−1	298	—	—	—	Hydrothermal process	—	[67]
Rh/OPNC	433	298	—	26.4	2.88	Air-mediated pyrolysis method	62%/5	[68]
Ru1Ni1.9/NCS	824	298	—	26.5	2.3	Impregnation–reduction method	67%/5	[69]
Pt/CNT-5W	710	303	—	27.8	1.4	Two-step method	45%/5	[70]
Pd0.1Cu0.9/T-PC	279	298	—	57	2.9	In situ reduction	44.4%/10	[71]
Rh/PCNs	513.2	298	—	46.5	2.3	Simple pyrolysis	66%/6	[77]
Ni−MoOx/(P)NCS	85.7	298	—	29.6	3	Phosphate-mediated method	54%/5	[88]
Ni/CTFPh	14.3	303	—	34.8	2.2	Metal vapor synthesis	—	[89]
Co/CoNx-CNT-33-800T	7833 mLH2 gCo−1min−1	313	0.0654	46.17	—	—	75%/40	[90]
5Pt/G1600-O3-60	618.9	298	—	—	—	Atomic layer deposition	88%/6	[78]
MoO3-doped MnCo2O4 (0.10)	26.4	298	—	34.24	—	In situ synthesis	—	[79]
Ni0.5Cu0.5Co2O4 nanoplatelets	80.2	298	—	28.4	200	—	—	[91]
Ni0.5Cu0.5Co2O4 microspheres	65.1	298	—	29.5	40	—	—	[91]
Ni0.5Cu0.5Co2O4 nanoparticles	45.5	298	—	43.2	50	—	—	[91]
Pt-Ni	302.3	—	—	42.1	—	—	—	[92]
Plate-like Cu2O–CoO	34.1	298	—	—	Thickness of 40 nm	—	—	[93]
Rh/h-NCNWs	1234	298	—	36.94	—	—	60%/5	[94]
Cu0.6Ni0.4Co2O4	119.5	298	—	33.91	—	—	70%/8	[95]
Cu0.5Ni0.5Co2O4 (Mo = 0.10)	195.25	298	—	—	—	—	—	[96]
CoNiP/GO	134.6	298	—	44.12	—	—	84.6%/5	[97]
Co–Fe–B@g-C3N4/NF	14,005 mLH2·min−1·gCat−1	298	—	45	—	Chemical deposition method	—	[98]
Co-Mo-B/NF	6027.1 mLH2·min−1·gCat−1	298	—	43.6	65	Electroless plating method	—	[99]
CuO/Co3O4@C-4	18.8	298	—	18.5	—	Hydrothermal method	—	[100]
Ru/B-U-TiO2	1287	298	—	37.96	—	—	65%/8	[101]
Co0.7Cu0.3@NHPC-800	—	303	—	26.2	—	—	—	[103]
Co@N-C-700	5.6	298		31	9	One-step thermolysis	97.2%/10	[104]
Ru/NPC	813	298	—	24.95	—	In situ reduction	67.3%/5	[105]
Rh/NPC	473.5	298	0.003	40.2	6.03	Pyrolysis method	54.0%/8	[106]
Ru/Ti3C2−xNx	1334	298	—	—	1.54	Microwave heating polyol method	—	[107]
Co0.5Ru0.5/CosNC	1068	298	—	18.96	—	—	100%/10	[80]
Ru/OCB	602	298	—	34.3	2.0	Microwave-assisted solid-state strategy	52%/5	[108]
Pt/MXene-O3	265	303	—	69	0.6	—	—	[109]
Ru/ONC	556	298	—	34.3	1.69	Gas-phase oxidation strategy	—	[81]
Ru@PC-5–700	405.9	303	—	—	1.3	—	58.3%/7	[110]
Ru0.50Ni0.50@WSC	251	303	—	45.3	—	Facile adsorption–NaBH4 reduction method	—	[111]
CoP–CoO/NCDs	89.56	298	—	41	58.82	—	—	[82]
Ru/BNC	1854	298	—	26.31	1.56	—	—	[112]
Rh/N-U-TiO2	721	298	—	20.05	3.28	—	62%/5	[113]
Cu/Cu0.76Co2.24O4-VO	28.46	298	—	24.36	—	—	—	[114]
Rh0.75Co0.25/Ni@ Ni-N-C	223.8	303	—	28.63	3.69	Maceration reduction method	—	[115]
Rh/C-SC	336	298	—	37.1	4.1	—	50%/5	[83]
PtPd3	4034 mL·min−1·gCat−1	298	—	14.56	10	—	70.8%/5	[116]
CF-BT-Ru	322	308	—	32.41	2.6	—	—	[84]
Ru/3DNPC-500	584	298	—	31	1.32	High-temperature pyrolysis	50%/7	[121]
Pd/IPCN	122.8	298	—	29.1	2.17	—	—	[122]
P2-Cu-Co3O4@CNF	35.6	303	—	29.86	—	Nanoconfinement method and a facile ion-doping approach	—	[123]
Cu0.5Ni0.5/h-BN	6.33	303	—	23.02	—	Adsorption–chemical reduction	—	[125]
CPFC-MS@NiAl-LDH@RhxNi1−x	13	298	—	40.3	—	—	—	[127]
Pt-Co/GQDs	520	303	—	45.3	13	—	—	[129]
CoRu0.5/CQDs	814.7	298	—	39.29	4.25	One-step hydrothermal	—	[130]
AuNi@ZIF-8	40	298	—	37.4	—	—	—	[131]
SCo0.43Cu0.57	5.68	298		31.06	18.23	Acid etching method	71.8%/5	[132]
Ni–Zn/SiO2	4.3	298	0.025	—	8.5	—	—	[133]
Ru0.8Ni0.2/g-C3N4-rGO	905	303	0.0016	27.2	1.4	Adsorption–chemical reduction	55%/6	[135]
Ru0.25Pd0.75@g-pC	214.49	303	—	47.3	1.5	One-pot calcination method	62%/18	[136]
Ru0.075Co0.925/NPC	754	298	—	30.5	26.68	Maceration reduction method	56%/5	[137]
NiMn-decorated CNFs	58.2	303	—	38.9	60	—	—	[138]
Ru0.6Co0.4/P25	443.7	298	—	43.9	25	—	—	[141]
CoCu-NC-5	8.12	298	—	34.25	7.95	—	—	[142]
Fe-CoP@C	183.5	298	—	30.6	—	—	—	[143]
Ru2Fe1/N–C	424	298	—	33.7	3	Impregnation–co-reduction	26%/5	[144]
Pt0.1%Co3%/TiO2	1530	298	0.0008	63.8	1.3	Step-by-step reduction method	100%/5	[145]
Pt76Au12Co12	450	298	—	18.47	—	Sequential Digestive reduction	56%/5	[146]
Cu0.4Co0.6 Pt0.0075O/RGO	854	—	—	39.8	2.89	—	80%/8	[147]
Pt1.5/CoCu0.4-NC	1636.82	298	—	41.78	9.31	Liquid-phase reduction method	—	[148]
Ni0.3Pd0.7Mo0.2 NPs	252.7	298	—	52.3	5.93	In situ reduction method	15%/5	[149]
np-RuNiFeCo	148.2	298	—	25.3	4	—	—	[150]
Rh0.8Ru0.2Ni0.25@MMT-S	2961	298	—	29.7	2	Impregnation method	—	[151]
Cu0.8Ni0.1Co0.1 @MIL-101	72.1	298	0.0027	29.1	2.8	Solvent evaporation method	73%/8	[152]
Pt/Co3O4 NCs	721	298	—	31.3	1.2	—	86.6%/10	[154]
Co-CoOx/TiO2@N-C	5905 mL·min−1·gCo−1	298	—	38.5	—	Sol–gel method	85%/5	[155]
Pd0/Co3O4	3048	298	—	62	—	Impregnation/reduction method	100%/10	[156]
Vo–Co–Sn5:2	17.6	298	—	45.95	5–15	Co-precipitation–calcination method	82.6%/14	[157]
1.5-PdTVO	240	298	—	34.6	2	—	—	[158]
A20-Pd5	17.4	—	—	41	20	Solid-state approach	—	[159]
Ru/Co2.28Cu0.72O4/C7.5	2020	298	—	26.3	1.4	—	—	[160]
Vco-Co3O4	934	298	—	32.65	45.9	—	—	[161]
1.5-RTVO-4	1370	298	—	46.3	1.9	—	—	[162]
RuPd-TiO2-VO	2750	298	—	32	2	—	90%/10	[163]
Co-CN-O-100	11,410 mL·min−1·gCat−1	313	—	39.41	—	—	80%/5	[164]
CoCu1Mo3-NC-O-15	24.44	298	—	28.44	—	—	—	[165]
Ru0/SiO2-Fe3O4	127	298	0.0079	54	3.75	Maceration reduction method	100%5	[166]
Ni0/CoFe2O4	38.3	298	—	62.7	—	Two-step impregnation–reduction method	38%/10	[167]
Pt/MoO3−x-500	2268.6	298	—	13.97	1.8	—	—	[169]
Pd0.75@Ag0.25/SiO2	109.99	303	—	42.26	10	Seed-mediated stepwise reduction	83.2%/5	[170]
Rh0/CoFe2O4	720	298	0.00048	66	2.18	Two-step impregnation–reduction method	100%/5	[172]
Rh/CoFe2O4-SB-H2	1894	298	—	59.3	—	Impregnation–reduction method	75%/10	[173]
Ru/PC	744	298	—	39.11	1.5	Salt template-assisted in situ construction	—	[174]
Co–Mo2C/NC	18,876 mLH2·min−1·gCo−1	298	—	49.8	4.71	One-step method	77.6%3	[175]
Pd1Rh4/Ti3C2	338	298	—	33.8	4	Microwave-assisted reduction method	—	[176]
Ru@Co-NC	568	298	—	24.2	2.25	Sodium chloride template method	59.8%/5	[177]
Cu0.9Ni0.1/Ti3C2Tx	2429 h−1	323	—	41.61	1.84	Wet-chemical co-reduction method	95.5%/5	[178]
Ni-MoOx/(P)NC	85.7	298	—	29.6	3	Phosphate-mediated method	54%/5	[88]
Ru/TASC-NaOH	582	303	—	43	1.6	—	—	[179]
Rh/MXene	288.4	298	—	54.2	2.55	Wet impregnation method	99%/7	[180]
CuCo/PDA-Ti3C2	71.8	293	—	45.89	1.8	Surficial alkaline functional strategy	—	[181]
Ni/Ti3C2Tx−4	161	298	0.074	59.3	3.07	—	—	[182]
Ru/Ti2.5V0.5C2	1072	298	—	43	—	—	—	[183]
Rh0.8Ru0.2/ SP-ZSM-5-100	1006	298	0.001	56.5	0.7	Incipient wetness impregnation method	—	[184]
Ru/S-1@C(RSC-2)	892	298	—	36.8	3	In situ reduction method.	—	[185]
Cu0.5@Co0.5-MOF/5	129.8	298	—	26.5	5.5	Hydrothermal method	95%/5	[186]
PtCo2@COF	486	293	—	34.5	2	—	—	[187]
Cu0.8Ni0.2@ZrO2/NC/RGO	40.9	303	—	33.24	2.5	—	—	[188]
NiCoP/CoP	30.3	298	—	25.89	—	Three-step hydrothermal–oxidation–phosphorization	—	[189]
CuCo(O)@CN	12.4	298	—	33.8	30	Thermal reduction	64.7%/5	[190]
Co2P/(0.59-Cu3P)-NC	729.6	298	—	63.5	—	—	—	[191]
Pd1Ru2NPs@Alkyne-PVA gel	247.93	298	—	33.02	2	In situ reduction	—	[192]
Ru1Ni4/APTS-rGO	1559	298	—	37.2	2	One-step in situ co-reduction	49.9%/5	[193]
Pt2Ox	2800	—	—	—	0.96	Bottom-up approach	100%/5	[194]
CoP/Co2Ps	64.1	298	—	38.8	—	Salt-induced phase transformation	100%/10	[196]
CoO0.5P0.5@CS	37	298	—	41	7.87	—	—	[197]
MS@Pt@EuW10@PNIPAM	51.1	298	—	74.89	—	—	—	[201]
NiNPs/ZIF-8	85.7	298	—	28	2.7	Deposition–precipitation (DP) method	—	[202]
Ru1Co9/TiO2	1408	298	—	33.25	—	Co-precipitation and reduction	—	[203]
Pd/CGP-GO-Fe3O4	36.5	303	—	27.4	—	Co-precipitation method	—	[204]
Rh0-Co3O4	1800	298	0.00024	61.7	—	Two-step impregnation–reduction method	100%/5	[205]
CuNi/Co3O4	30.5	298	—	41.8	100 in diameter	Impregnation–reduction method	50%/5	[206]
NiCoP/NF	—	298	—	58.95	—	Low-temperature induced phosphating method	—	[207]
Ru-Ni-NF	539.6	298	—	36.4	—	Spontaneous redox reaction	—	[208]
CFNP@CF foam	12.5	303	—	23.4	20	Electrodeposition	99.5%/3	[209]
CoP NA/Ti	6500 mLH2·min−1·g−1	—	—	41	—	—	100%/20	[210]
Co–Mo–B/CC	3916.1	298	—	25.2	—	Electroless plating	55.9%/5	[211]
Rh@GQDs	469	303	0.002	54.85	2.3	—	—	[213]
Co/HNTA	—	298	—	10.8	3.07	—	—	[214]
Pt/ZIF-67	687	303	—	45.43	2.15	—	—	[215]
Ru/MoAl1−xB	494	303	—	39.2	2.4	—	—	[216]

4. By-Product Analysis and Regeneration Strategies in AB Hydrolysis

Ideally, 1 mol of AB reacts with 2 mol of water to produce 3 mol of H₂. However, in practical reactions, the products contain significant impurities, primarily categorized into gaseous impurities and boron-containing by-products. Identifying the hydrolysis products is crucial for the processing and recycling of spent fuel, as well as for understanding the hydrolysis mechanism of AB. This section provides a detailed discussion on the identification of AB hydrolysis products, methods for its regeneration, and a complete life cycle recycling system, laying a foundation for further understanding of the hydrolysis mechanism.

4.1. Identification of By-Products of Hydrolysis Reaction

Hydrogen produced from the hydrolysis of AB can be used in hydrogen fuel cells and the construction of hydrogenation stations, where ensuring high-purity hydrogen is critical for efficient operation. The impurities generated from AB hydrolysis fall into two main categories: gaseous impurities mixed with hydrogen and solid by-products [217].

Ammonia (NH₃) is the primary gaseous impurity in the hydrolysis products of AB. Although NH₃ is also a potential hydrogen carrier, converting it to pure hydrogen typically requires high temperatures and extended conversion times. Additionally, the partial exchange of ions formed by NH₃ with protons reduces the proton conductivity in the anode catalyst layer [218], consequently shortening the lifespan of fuel cells. Therefore, minimizing NH₃ release is essential for efficient hydrogen generation from AB. Recently, researchers have explored some strategies to suppress or mitigate NH₃ formation, aiming to enhance the viability of AB-based hydrogen production systems. A key approach involves integrating functional materials with strong NH₃ adsorption capabilities to selectively capture ammonia by-products during hydrogen generation. For instance, Majumder et al. [219] developed CB7-functionalized Co:Ni alloy nanocomposites, which exhibit exceptional NH₃ adsorption capability, ensuring the continuous generation of high-purity, ammonia-free H₂ during the hydrolysis of ammonia borane. Banerjee et al. [220] designed a multifunctional catalyst, Cu-MOF-74, which enables simultaneous hydrogen generation from ammonia borane (AB) and selective capture of ammonia by-products, thereby enhancing the efficiency and purity of hydrogen production. In addition to physical adsorption, chemical suppression of NH₃ formation has also been explored. Nakagawa et al. [221] employed citric acid as a catalyst to not only facilitate the efficient hydrolysis of ammonia borane but also effectively suppress the delayed release of ammonia, thereby optimizing the reaction kinetics and hydrogen purity. These studies collectively highlight the potential of material engineering and reaction modulation strategies in minimizing NH₃ contamination, providing valuable insights into the development of advanced catalytic systems for AB-derived hydrogen production.

Beyond gaseous impurities, complex solid by-products that are challenging to remove from the reaction system can deactivate the catalyst by accumulating on its surface. Initially, scholars proposed that ammonium borates, such as NH₄BO₂, are the primary products post-hydrolysis [222]. Liu et al. [223] suggested that boric acid, rather than ammonium metaborate (NH₄BO₂), is the main hydrolysis product. Other studies have identified ammonium tetrahydroxyborate (NH₄B(OH)₄) as a primary product [224]. Demirci et al. [225] proposed that two types of borate, boric acid and tetrahydroxyborate anions could form, with the solution quickly evolving to create and maintain equilibrium with polyborates, as shown in Equation (33). In excess aqueous solutions, the hydrolysis by-product is ammonium tetrahydroxyborate [226]. Ramachandran et al. [227] found that at higher concentrations of the aqueous solution (10.8 M), polyborates (${(N{H}_4)}_2{B}_4{O}_5{\left(OH\right)}_4\times 1.4H_{2}O$) form, eventually resulting in a mixture of borates and polyborates. Recently, Valero-Pedraza et al. [228] observed the precipitation of borate crystals in the products of spontaneous AB hydrolysis in a concentrated aqueous alkaline solution (10 M). They identified a novel by-product—diammonium tetraborate dihydrate (NH₄₎₂B₄O₅(OH)₄·2H₂O), as shown in Equation (34). This discovery expands the list of potential borate by-products derived from AB hydrolysis.

$$2B{\left(OH\right)}_3+B{\left(OH\right)}_4^{-}\leftrightarrow B_3{O}_3{\left(\mathrm{OH}\right)}_4^{-}+3H_{2}O$$

(33)

$$4N{H}_{3}B{H}_3+11H_{2}O\leftrightarrow 12H_2+2N{H}_3+{(N{H}_4)}_2{B}_4{O}_5{(OH)}_4\cdot 2H_{2}O$$

(34)

4.2. The Regeneration Strategies of Ammonia Borane

As insights into the nature of solid borate by-products deepen, regenerating AB from spent fuels is essential for developing a sustainable hydrogen cycle in AB dehydrogenation processes. The efficiency of AB recovery is closely linked to the composition of by-products in the spent fuel, which varies with the dehydrogenation method employed. For instance, Ramachandran et al. [227] demonstrated the first successful hydrogen cycle from AB using a one-pot synthesis from ammonium tetramethoxyborate, as illustrated in Figure 17a and described by reaction Equation (35). Additionally, further studies have explored regenerating AB from solid pyrolysis by-product BNHx using methods like hydrazine and liquid ammonia regeneration, as well as the catalytic hydrodechlorination of BCl₃ [229,230]. For a systematic review of thermolysis byproduct regeneration mechanisms and recent advances, see Supplementary Materials. This section primarily concentrates on regeneration strategies that utilize the hydrolysis products of AB.

$$N{H}_{4}B{\left(OMe\right)}_4+N{H}_{4}Cl+LiAl{H}_4\underset{\begin{array}{l}0 °C to RT\\ 3-8 h\end{array}}{\overset{THF}{\to }}N{H}_{3}B{H}_3+Al{\left(OMe\right)}_3+MeOH+H_2+LiCl+N{H}_3$$

(35)

Borate is the primary product of AB hydrolysis, yet studies on the direct or indirect regeneration of AB from borates remain limited due to the diverse types of borates produced during hydrolysis. The B-O bond in borates is exceptionally stable, with a formation enthalpy of approximately 700 kJ/mol. Conversion of a B-O bond to a B-H bond typically requires a potent reducing agent. However, the U.S. Department of Energy has discontinued research on regeneration using traditional reducing agents because their preparation requires energy-intensive electrochemical methods on metals.

Currently, the methods for AB regenerating from hydrolysis products are primarily based on its common synthesis approach, which involves the displacement reaction of alkali metal borohydrides with ammonium salts in an organic solution. In 2012, Ramachandran et al. [231] used LiAlH₄ as a reducing agent. In tetrahydrofuran (THF) solution, AB could be regenerated by the reaction of (B(OMe)₃) with ammonium salt, as shown in Equation (36). They also found that the type of ammonium salt affects the yield of AB regeneration, with NH₄Cl yielding the best results. Liu et al. [223] successfully regenerated AB using NaBH₄, derived from methanol esterification of boric acid, mixed with ammonium sulfate in THF (Figure 17b). However, due to the competitive hydrolysis of ammonium borohydride, the yield of AB often falls below quantitative levels. To further optimize this regeneration method, the research group replaced traditional THF with recyclable 2-methyltetrahydrofuran and used water as a promoter for the metathesis of NaBH₄ and (NH₄)₂SO₄, synthesizing AB with extremely high purity (>99%) under ambient temperature and pressure [232]. Additional studies indicate that adding NH₃ in THF can facilitate the regeneration of pure AB; Ramachandran et al. [233] found that the NH₃ can nucleophilically attack ammonium borohydride while simultaneously dehydrogenating to form AB, thus promoting the synthesis of pure ammonia borane from sodium borohydride and ammonium sulfate at high yields under room temperature. Recently, Majumder [219] innovatively proposed a three-step process using a solution of NH₄BO₂ resulting from hydrolysis to regenerate ammonia borane (AB), as shown in Figure 17c. The regenerated AB has been demonstrated to release hydrogen at levels comparable to commercial AB. Additionally, regenerating AB from intermediates produced during the hydrolysis reaction is a viable approach. For instance, Davis et al. [234] showed that borazine formed during AB dehydrogenative coupling can restore B-H bonds in the presence of Lewis bases. When the Lewis base is NH₃, it directly forms AB and another by-product, “B(NH₂)₃”. Nagyházi et al. [235] introduced an efficient and rapid homogeneous AB hydrolysis and dehydrogenation regeneration system (Figure 17d).

$$B{\left(OMe\right)}_3+N{H}_{4}Cl\underset{0 °C 3h}{\overset{THF}{\to }}{H}_{3}B-N{H}_3+Al{\left(OMe\right)}_3+H_2+LiCl$$

(36)

In summary, AB regeneration depends on the dehydrogenation method and its by-products. Effective AB synthesis typically involves the displacement reaction between ammonium salts derived from hydrolysis by-products and alkali metal borohydrides (e.g., NaBH₄) in organic solvents. Additionally, AB can be regenerated from hydrolysis intermediates or through novel processes. However, current methods for regenerating AB encounter significant challenges. The stability of B-O bonds necessitates considerable energy input and hydrogenation, often requiring NaBH₄ as a reducing agent, which escalates costs. Furthermore, the competitive hydrolysis of NaBH₄ decreases regeneration efficiency. Additionally, the non-recyclability of reducing agents constrains their industrial feasibility.

Figure 17. (a) AB hydrogen cycle [227]. (b) Proposed total life cycle of ammonia borane for hydrogen generation [223]. (c) AB hydrolysis and the synthetic procedure for the regeneration of ammonia borane ammonium borate [219]. (d) Dehydrogenation and regeneration of AB in aqueous, hydrolytic conditions [235].

Figure 17. (a) AB hydrogen cycle [227]. (b) Proposed total life cycle of ammonia borane for hydrogen generation [223]. (c) AB hydrolysis and the synthetic procedure for the regeneration of ammonia borane ammonium borate [219]. (d) Dehydrogenation and regeneration of AB in aqueous, hydrolytic conditions [235].

5. Conclusions and Outlook

5.1. Conclusions

As the utilization of ammonia borane (AB) expands, supported metal-based catalysts are garnering significant interest for their potential to catalyze AB hydrolysis. However, challenges such as ambiguous catalytic mechanisms and suboptimal catalytic performance necessitate ongoing research and development in this domain. To present the complex knowledge and research system in front of readers, this paper reviewed the recent five-year advances in the AB hydrolysis field, including the elucidation of hydrolysis mechanisms, optimization and control strategies for catalytic performance, and methods for regenerating ammonia borane.

This study systematically explores the evolution of the AB hydrolysis mechanism, delineating the hydrolysis process into fundamental stages such as adsorption, activation, dissociation, and the subsequent formation and desorption of hydrogen. It employs comprehensive evaluation metrics such as hydrogen generation rate (HGR), turnover frequency (TOF), and activation energy (E_a) to conduct in-depth analyses of the structural and functional efficiencies of various catalysts. These insights are pivotal for optimizing catalyst performance to enhance AB hydrolysis effectively.

Additionally, this paper outlines strategies for optimizing and controlling catalytic performance. It begins with catalyst optimization, focusing on enhancing the number and dispersion of active centers. This involves optimizing catalyst size, morphology, support surface modification, and the impact of stabilizer ligands. It also details methods for enhancing the adsorption and activation of reactants, leveraging the synergistic effects between different metallic components in alloy catalysts and metal–support interaction to significantly promote the adsorption and activation of AB and/or H₂O on the catalyst surface, thus lowering reaction barriers and accelerating the hydrolysis of AB.

Moreover, common methods for achieving “on-demand” hydrogen generation switches from AB hydrolysis are reviewed, including temperature-regulated switches, pH-regulated switches, catalyst-transfer switches, and other innovative catalyst switches.

Finally, the paper concludes by discussing the by-products of AB hydrolysis, exploring potential regeneration methods, and outlining future research directions to advance the field. This comprehensive approach not only clarifies the underlying mechanisms of AB hydrolysis but also paves the way for developing more efficient and sustainable hydrogen energy solutions.

5.2. Outlook

Recent advancements in optimizing AB hydrolysis catalysts have been made, but significant gaps remain in understanding the hydrolysis mechanisms, reaction products, AB regeneration, and practical applications.

Most research focuses on experimental methods for enhancing hydrogen evolution rates, with an insufficient exploration of dehydrogenation performance and rate-determining steps under various conditions. Although density functional theory (DFT) calculations have suggested several mechanisms for AB hydrolysis, distinguishing these mechanisms is difficult, especially when multiple pathways may coexist. Additionally, the lack of precise techniques for identifying reaction intermediates and transition states limits the understanding of metal nanoparticle nucleation and growth mechanisms, as well as the relationship between active sites and electronic structures in multi-component catalysts. Future work should prioritize operando characterization techniques (e.g., in situ XAS, Raman) to resolve the transient hydride-mediated transformations of iron-group catalysts (Fe, Co, Ni) and their impact on active center evolution. This issue is particularly challenging for complex multi-metal catalysts, where the roles of each component are hard to discern. Moreover, the durability of AB hydrolysis catalysts needs significant improvement, as most studies assess performance over fewer than ten cycles. Enhancing catalyst durability is essential for the practical application of AB hydrolysis.

Secondly, researchers have predominantly focused on catalysts, often neglecting the importance of reactants and by-products. The impact of hydrolysis products, including hydrogen purity and borate management, must be considered. For instance, excess water in hydrolysis results in ammonium tetrahydroxyborate as a by-product. The exothermic nature of AB hydrolysis causes water to evaporate and carry borates with the hydrogen, leading to impurity. Additionally, water condensation can cause borates to precipitate in pipelines, causing blockages.

Lastly, regenerating AB from hydrolysis products and achieving practical application remain significant challenges. The regeneration process from borates has not been thoroughly studied, partly due to the formation of multiple borates during hydrolysis. For example, regenerating AB from NH₄BO₂ requires high energy and faces competition hydrolysis from the hydrolysis of reducing agents. Moreover, current research rarely addresses practical application scenarios for hydrogen production from AB hydrolysis. Existing regeneration methods fall short of the U.S. Department of Energy’s (DOE) 60% well-to-tank (WTT) efficiency target. There is also a lack of feasibility analysis for the engineering applications of AB preparation and hydrolysis product regeneration.

Our research group aims to tackle these challenges by developing and operating hydrolysis reactors and refining the comprehensive lifecycle evaluation system for hydrogen production from AB. This effort will offer deeper insights into the practical issues of using AB as an anode fuel and hydrogen carrier, thus guiding future research and development in this field.