Hydrogen Evolution upon Ammonia Borane Solvolysis: Comparison between the Hydrolysis and Methanolysis Reactions

Abstract

Hydrogen (H₂) production is a key challenge for green carbon-free sustainable energy. Among the H₂ evolution methods from H-rich materials, ammonia borane (AB) solvolysis stands as a privileged source under ambient and sub-ambient conditions given its stability, non-toxicity, and solubility in protic solvents, provided suitable and optimized nanocatalysts are used. In this paper dedicated to Prof. Avelino Corma, we comparatively review AB hydrolysis and alcoholysis (mostly methanolysis) in terms of nanocatalyst performances and discuss the advantages and inconveniences of these two AB solvolysis methods including AB regeneration.

1. Introduction

Hydrogen (H₂) generation is one of the main focuses of current research, particularly because it is a green energy source toward the replacement of harmful fossil-based fuel [1,2]. H₂ energy has appeared to be essential, because it is carbon-free, without pollution, and it involves an excellent energy storage density [1]. Besides water (hydrogen evolution reaction, HER [3]), liquid-phase hydrogen generation systems include a variety of metal borohydrides, ammonia and its borane derivatives (ammonia borane, amine boranes, hydrazine borane, etc.), formic acid, and several hydrocarbons [1]. Among all these possibilities, since the seminal work by Chandra and Xu [4], ammonia borane (AB) has appeared to be by far the most studied H₂ source. During the first decade of the century, AB thermolysis was the subject of various studies [5,6], but the high temperatures required (only 15% H₂ is obtained at 200 °C [7]), long induction times, and ill-defined mixture of B_xN_y derivatives obtained made nanocatalyzed AB solvolysis a privileged route [2,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. Therefore, we are briefly reviewing here the two major AB solvolysis routes, i.e., hydrolysis and methanolysis, with the goal of comparing them, because, despite similarities, there are important differences in terms of environmental concern, rates, efficiencies, facilities, and side reactions. Publications generally focus on either of them without comparison with the alternative solvolysis method. In conclusion, researchers should be able to consider the parameters and state of the art toward selecting their choice. Finally, in the conclusion, prospects are formulated toward improving the hydrogen generation process upon AB solvolysis.

2. Ammonia Borane (AB)

Ammonia borane, currently abbreviated as AB, is a white solid, due to the intermolecular H bonding (NH—HB distance about 2 Å [24]). Its molecular weight M_w is 30.8 g·mol⁻¹ and its density at 25 °C is ρ = 0.78 g·mL⁻¹. It is stable in water at neutral pH, whereas NaBH₄, another H₂ source, slowly reacts with water. AB is stabilized, formally by a dative bond from N to B with a net electron transfer from N to B (Figure 1, left) of approximately 0.5 electrons and with the formation of a compound that is 65% ionic; the ionic structure is represented on the right of Figure 1 [25]. Thus, the N-H bonds are acidic, whereas the B-H bonds are hydridic.

AB is soluble in ether, 1,2-dimethoxyethane (DME), water, alcohols, and many ionic liquids; slightly soluble in benzene, dichloromethane, and chloroform; insoluble in alkanes.

AB was first synthesized by Shore and Parry and reported in 1955 (Equation (1)) [26].

$${\mathrm{LiBH}}_4+{\mathrm{NH}}_4\mathrm{Cl}\to {\mathrm{NH}}_3{\mathrm{BH}}_3+\mathrm{LiCl}+{\mathrm{H}}_2$$

(1)

Since then, many variations of this commonly utilized metathesis-type synthesis between metal borohydrides MBH₄ and ammonium salts NH₄X have appeared, summarized six decades later in Shore’s review [27].

The synthesis of AB was improved by the use of methyl borate, B(OMe)₃ [28]. Methyl borate had indeed long been available from boric acid, B(OH)₃, and methanol, a reaction reported in 1953 by Schlesinger et al. (Equation (2)) [29].

$${\mathrm{H}}_3{\mathrm{BO}}_3+4{\mathrm{CH}}_3\mathrm{OH}\to \mathrm{B}{\left(\mathrm{OMe}\right)}_3+\mathrm{MeOH}+3{\mathrm{H}}_2\mathrm{O}$$

(2)

In particular, Ramachandran’s group developed this especially practical synthesis from B(OMe)₃ using LiAlH₄ in THF added to a mixture of B(OMe)₃ and NH₄Cl in THF at 0 °C over 1 h, followed by stirring 3 h at 0 °C (90% isolated yield after extraction with ether) (Equation (3)) [30].

$$\mathrm{B}{\left(\mathrm{OMe}\right)}_3+{\mathrm{NH}}_4\mathrm{Cl}\underset{\mathrm{THF}, 0{ }^{°}\mathrm{C}, 3 \mathrm{h}}{\overset{{\mathrm{LiAlH}}_4}{\to }}{\mathrm{BH}}_3{\mathrm{NH}}_3+\mathrm{Al}{\left(\mathrm{OMe}\right)}_3+\mathrm{LiCl}+{\mathrm{H}}_2$$

(3)

In 2015, Ramachandran’s group even reported an astute extension of the metathesis reaction of Equation (1) to the syntheses of a variety of amine borane derivatives in high yields and purity by reactions of alkylammonium sulfate derivatives with NaBH₄ [29].

3. H₂ Generation upon AB Hydrolysis

Besides properties such as a mild reducing agent and other properties that are reviewed elsewhere [2,5,6], AB is an excellent source of H₂, which is due to its low molecular weight, high hydrogen content (19.6 wt%), high solubility in water (35 g of AB for 100 g of water), stability, and lack of toxicity.

3.1. Use of an Acidic Medium

Although AB is stable in neutral or slightly basic aqueous solution, H₂ is formed in acidic solutions owing to the hydricity [31] of the BH groups of AB. Chandra and Xu reported that 3 mol H₂ was released from AB in aqueous solution in the presence of solid acids such as Dorex, Amberlyst, and Nafion-H, in a few minutes under ambient conditions, according to Equation (4) [31]. CO₂ also reacts as an acid, upon dissolution into water, releasing H₂ together with the formation of boric acid [32].

$${\mathrm{NH}}_3{\mathrm{BH}}_3+2{\mathrm{H}}_2\mathrm{O}\stackrel{{\mathrm{H}}^{+}}{\to }{\mathrm{NH}}_4^{+},{\mathrm{BO}}_2^{-}+3{\mathrm{H}}_2$$

(4)

3.2. Use of a Transition-Metal Catalyst

In neutral aqueous medium, the presence of a transition-metal-based catalyst is necessary to achieve AB hydrolysis with the formation of 3 mol H₂ and hydrated ammonium borate (Equation (5)).

$${\mathrm{NH}}_3{\mathrm{BH}}_3+4{\mathrm{H}}_2\mathrm{O}\stackrel{\begin{array}{c}\text{metal-based}\\ \mathrm{catalyst}\end{array}}{\to }{\mathrm{NH}}_4^{+},\mathrm{B}{\left(\mathrm{OH}\right)}_4^{-}+3{\mathrm{H}}_2$$

(5)

A large variety of transition-metal-based catalysts including homogeneous and heterogeneous catalysts, transition-metal complexes, oxides, transition-metal nanoparticles (NPs), and heterobinuclear catalysts have been reported and reviewed by Xu’s group and others [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. Ramachandran’s group reported that upon catalysis with RuCl₃, NH₃ was also produced along with H₂ evolution during AB hydrolysis, and the amount of NH₃ produced increased upon increasing the AB concentration, which would be a problem for the realization of fuel cells. The setup for H₂ production from AB and measurement is shown in Figure 2 [33].

These authors also found that the borate produced in the reaction of Equation (5) was ammonium tetraborate {[(NH)₄]₂[B₄O₅(OH)₄]·1.41H₂O} with the nature of the borate anions depending on the pH, and polyborate anions being formed at higher pH [32]. In early studies, transition-metal complexes were shown to be efficient for AB hydrolysis with H₂ evolution. Mechanistic studies privileged both B-H and N-H cleavage in the rate-determining step (RDS) [33,34]. The intermediacy of organometallic or inorganic radicals during the catalyst process might play a significant role [35]. Other studies with molybdenum complexes showed that the mechanism depended on the oxidation state of the catalyst [36]. Finally, with Co-Ni and Ni NPs, O-H activation by the catalyst in the RDS was proposed [37,38].

Here, the focus is on nanocatalysts that have attracted most of the attention during the last two decades, for which the NP support has played an essential role [39,40,41,42,43,44,45,46,47,48,49]. In particular, porous materials including metal–organic frameworks (MOFs) have been a privileged type of NP support, because the NP catalyst can be formed and encapsulated inside MOFs for efficient nanocatalysis [50,51,52,53,54,55,56,57]. Therefore, hydrophilic late transition-metal salts are introduced in water, for instance, into the cavity of the zeolitic imidazolate framework (ZIF) [58], for which ZIF-8 [59,60,61] has been very successful in our laboratory using reduction with NaBH₄ [62] of the transition-metal cation to NP inside the ZIF-8 cavity [18,63,64,65,66,67] (Scheme 1).

Under these conditions, Ni was shown to be the most catalytically effective first-row transition metal with an Ni NP size of 2.7 nm inside ZIF-8. In the presence of 0.3 N that appears to favor the catalysis, the turnover frequency (TOF) at 25 °C was of 85.7 mol_H2·mol_cat.⁻¹·min⁻¹ [63]. The reaction is first-order in catalyst concentration and zero-order in AB concentration, an indication that the role of AB in the transition state is not significant. This is confirmed by a kinetic isotope effect (KIE) with D₂O (KIE = k_H2O/k_D2O) of 2.49, suggesting that water O-H bond cleavage is the rate-determining step (RDS). The favorable influence of OH⁻ on the reaction rate was taken into account by OH⁻ coordination onto the Ni NP surface, increasing the electron density on Ni, which is favorable for water O-H oxidative addition onto surface Ni atoms. Meanwhile, on–off control of the H₂ evolution was demonstrated upon successive addition of OH⁻ that activated H₂ evolution and H⁺ that stopped the reaction (Figure 3) [63].

3.3. Use of a Transition-Metal-Alloyed Catalyst

Heterobimetallic catalysts often result in positive synergistic effects [68,69], and Hou et al. synthesized graphene-oxide-supported bimetallic Ni-Co catalysts reaching a TOF of 154 mol_H2·mol_cat.⁻¹·min⁻¹ at ambient temperature [68]. A remarkable result was obtained in our group with Pt-Ni NPs in ZIF-8 (Ni₂Pt@ZIF-8) for which the TOF value reached 669.3 mol_H2·mol_cat.⁻¹·min⁻¹. The synthesis of NiPt NP@ZIF-8 is shown in Scheme 2 [64].

This high TOF value was 87 times higher than the TOF of Pt@ZIF-8 under similar conditions, as a result of a positive volcano-type type synergy between Ni and Pt. The suggested mechanism was related to that with Ni@ZIF-8, and the primary KIE with D₂O was high (4.67), showing that water O-H cleavage was the RDS [63]. A considerable volcano-type positive synergy in catalytic efficiency was also obtained with Co-Pt NPs embedded in “click” dendrimers. In these “click” dendrimers that contain crucial triazole ligands, the nanocatalysts were encapsulated inside the dendrimer by coordination onto these ligand groups bridging the dendrimer core to the dendrimer periphery (Figure 4). In the presence of 0.3 M NaOH, the TOF value was 476.2 mol_H2·mol_catal⁻¹·min⁻¹ (952.4 mol_H2·mol_Pt⁻¹·min⁻¹). In both the ZIF-8 (endoreceptor) and click dendrimers (exoreceptors), the N atoms of the support played an important role in the confinement of the catalyst together with the substrates, and similar mechanistic trends were observed with both supports (Scheme 3) [70].

Plasmon excitation by visible light is useful in nanocatalysis [71,72,73,74,75,76,77,78,79,80], including for H₂ evolution from AB [76,77,78]. When the bimetallic nanocatalyst included Au as one of the two metals, excitation of the Au plasmon by illumination with visible light boosted H₂ evolution, with the most efficient light effects being observed upon combining Au with M = Co, Ni, Ru, Rh, or Pt. The evolution of 3 mol H₂ was observed in down to 1.3 min at 25 °C, 3 times faster than in the dark. Co was found to be the best first-raw transition metal alloyed with Au [78]. This boosting effect was taken into account by the plasmon-induced “hot” electron transfer, from Au to the other late transition metal M in the alloy, enriching the electron density on M, which enhances activation of the substrates [78,79]. Indeed, in the reaction mechanism, it is the transition metal atom M at the surface that is responsible for water O-H bond cleavage by oxidative addition of this bond in the rate-determining step (RDS). The increase in electron density at this metal center, such as the one provided from the plasmonic metal by “hot” electron transfer, largely favors this oxidative addition (vide infra). Primary KIEs observed using D₂O under dark conditions and found to be even larger under visible-light irradiation confirmed that O-H bond cleavage was largely involved in the RDS. In the initial hydride transfer from AB to the nanocatalyst, it appears that water is not significantly involved (contrary to NaBH₄ as the H₂ source, because the negatively charged BH₄⁻ anion is more significantly hydrogen-bonded to water than in neutral AB). On the other hand, after such a hydride transfer (or oxidative addition of the B-H bond immediately followed by B to NP electron transfer), the nanocatalyst becomes negatively charged, which increases the ability of water activation by oxidative addition onto the active late transition metal M (but not Au alone, which is a very poor nanocatalyst of the reaction). The increase in electron density at the active metal site is still considerably enhanced by the plasmonic hot electron transfer. Thus, the anionic [NP-H]⁻ intermediate is likely to form a hydrogen bond with water, [NP-H]⁻–H–OH, removing some electron density from the water O–H bond, which further facilitates oxidative addition of this bond onto the activated transition-metal surface. This first reaction phase ends by reductive elimination of the two hydride ligands forming H₂ and the OH and NH₃BH₂ ligands providing NH₃BH₂OH that, in turn, undergoes H₂ formation according to a similar process (Scheme 1) [78].

4. Catalyzed H₂ Generation upon AB Methanolysis

It is striking that AB methanolysis (Equation (6)) has been so much less studied than AB hydrolysis.

$${\mathrm{NH}}_3{\mathrm{BH}}_3+4{\mathrm{CH}}_3\mathrm{OH}\stackrel{\mathrm{catalyst}}{\to }{\mathrm{NH}}_4\mathrm{B}{\left({\mathrm{OCH}}_3\right)}_4+3{\mathrm{H}}_2$$

(6)

The seminal AB methanolysis publication is that of Ramachandran and Gagare who reported in 2007 both optimized ammonia borane preparation and its methanolysis [33]. These authors utilized, for AB methanolysis, late transition-metal catalysts, both of the first-row metals and noble metals (CoCl₂, Raney Ni, Pd/C, RuCl₃ and RhCl₃), with RuCl₃ being found as the best catalyst. With this catalyst, they obtained a TOF value of 150 mol_H2·mol_catal⁻¹·min⁻¹, which already was a very good performance compared to all those obtained afterward (vide infra) and up to now. Jagirdar’s group reported Co, Ni, and Cu-based nanocatalysts (including oxides and bimetallic Ni-Cu alloys), avoiding the use of noble metals [81,82]. Among all the AB methanolysis reports up to now (most of the reported TOF values are below 100 mol_H2·mol_catal⁻¹·min⁻¹), a remarkable performance in terms of TOF value for this reaction was obtained by Sun’s group with 366.4 mol_H2·mol_catal⁻¹·min⁻¹ using a Ag-Pd nanocatalyst, Ag₃₀Pd₇₀, synthesized by the coreduction of AgNO₃ and PdCl₂ in the presence of oleylamine, supported on carbon [83]. Another very efficient catalyst reported by Lu’s group contained ultrafine Rh NPs (1.4–2.6 nm) confined in a covalent organic framework (COF) synthesized from piperazine and cyanuric chloride, and it produced a TOF of 505 mol_H2·mol_catal⁻¹·min⁻¹ at 298 K. Given the temperature effect decreasing the TOF value upon decreasing the temperature, such high TOF values ensure applicability for low-temperature devices [84].

The same group recently reported an even better catalyst, a porphyrin-derived N-doped porous carbon-confined Ru NP for AB methanolysis with a very high TOF value of 727 mol_H2·mol_catal⁻¹·min⁻¹ at 25 °C [85]. Luo et al. reported L-proline (PRO)-capped Rh NPs as a catalyst with an AB methanolysis TOF of 1035 mol_H2·mol_catal⁻¹·min⁻¹ and 0.3 M NaOH that is known to boost late transition-metal nanocatalysts of AB solvolysis (vide supra). The presence of a suitable O- and N-rich PRO template also contributes to the positive synergy. At 0 °C, this nanocatalyst was still effective in ethanol that is less toxic than methanol, with these trends being useful toward low-temperature portable applications [86]. The best AB methanolysis performance so far was obtained by Zahmakiran’s group using nanocatalyst MIL-101-decorated Pd NPs with TOF = 1080 mol_H2·mol_catal⁻¹·min⁻¹ at room temperature [87].

The literature on AB methanolysis has recently been reviewed by Xu’s group [8], Li and Wang’s group [88], and Özkar [89], and reports on nanocatalysts have appeared based on first-row late transition-metals [8,90], and also on noble metals [90,91,92].

H₂ evolution upon AB methanolysis presents several advantages over the method using AB hydrolysis toward applications [8,88,91]: (i) AB is highly soluble in methanol under ambient conditions (23 wt%) [33,92]; (ii) AB is very stable in methanol in the absence of a catalyst; (iii) H₂ evolution upon methanolysis of AB is pure without ammonia (contrary to AB hydrolysis), which is an impurity poisoning fuel cells; (iv) AB methanolysis produces H₂ at low temperature if the nanocatalyst is efficient enough (methanol freezes at (−97.8 °C), which could be useful toward portable devices at cold temperatures); (v) NH₄B(OCH₃)₄ that is formed upon AB methanolysis readily reacts with LiAlH₄ and NH₄Cl at room temperature to regenerate AB, whereas the regeneration of NH₄B(OH)₄ formed upon AB hydrolysis is more complicated [33,89].

AB methanolysis is, in general, as for AB hydrolysis, first-order in catalyst concentration and zero-order in AB concentration, and the mechanism has been suggested to involve O-H cleavage in the RDS [81,82,83,84,85,86,87,88,89,90,91,92,93]. For instance, along this line, Lu’s group, in their work noted above on porphyrin-derived N-doped porous carbon-confined Ru NPs, indicated that AB methanolysis of both NH₃BD₃ and ¹⁵NH₃BH₃ in CH₃OH showed secondary KIEs of 1.29 and 1.05, respectively, and methanolysis of NH₃BH₃ in CD₃OD involved a primary KIE of 5.78. These results clearly showed the key role of methanol in the RDS step of AB methanolysis [85].

AB methanolysis was also recently catalyzed by Au-Pd@ ZIF-8, generated by reducing equimolar amounts of HAuCl₄ and Na₂PdCl₄ by NaBH₄ in methanol in the presence of ZIF-8, with the reaction being 3.7 times faster under visible-light illumination than in the dark. KIE values measured in CD₃OD were 2.2 and 3.4, respectively, in agreement with O–H bond cleavage as the RDS. Synergy effects between Pd and Au were taken into account in DFT calculations by the d-orbital density of states (DOS) of different NPs, with the d-band center of Au (111) and Pd (111) being −3.2 and −2.4 eV, respectively, whereas, upon alloying, the d-band center of AuPd (111) was −1.5 eV, at a higher energy than those of both Au (111) and Pd (111). Thus, the d orbital of the metal in AuPd is closer to the energy level of the antibonding orbital of the adsorbate species than with the related Au and Pd nanocatalysts, signifying a higher AB adsorption energy on AuPd. Upon illumination, a hot plasmon Au electron is injected into the Pd-AB adsorbate, increasing the electron density on Pd toward the methanol O–H oxidative addition onto the Pd surface center of the alloy, as confirmed by the DFT calculations (Scheme 4) [93].

Alternatively, Yamashita’s group examined photocatalytic H₂ generation over the Au/TiO₂ nanocatalyst from AB decomposition. This reaction was shown to be initiated by the generation of plasmon-induced charge pairs. It is not known, however, how the RDS would change upon AB methanolysis compared to hydrolysis in this process [94].

5. Comparison between AB Hydrolysis and Methanolysis

AB solvolysis is a superb and the most studied efficient method of H₂ generation, because AB is safe, stable in water and methanol under ambient conditions, and can now be obtained in >99% purity. Both AB hydrolysis and methanolysis produce 3 mol H₂ per mol AB, but the H₂ produced by hydrolysis contains all the more NH₃ as the AB concentration is higher, which is damaging, because NH₃ is poisoning fuel cells, whereas the H₂ produced by AB methanolysis is pure from NH₃ impurities. This is a very important advantage of methanolysis over hydrolysis, although nanocatalyzed AB hydrolysis proceeds faster than methanolysis. Another important parameter is the temperature, because hydrolysis is limited to reactions above 0 °C, the freezing point of water, whereas AB methanolysis can be operated at low to very low temperatures that are not limited by the extremely low freezing point of methanol as long as the catalyst is reactive enough at these low temperatures. Another essential parameter is the regeneration of AB from the AB solvolysis side product, NH₄B(OH)₄ in the case of hydrolysis, and NH₄B(OCH₃)₄ in the case of AB methanolysis. This AB regeneration is much more complicated in the case of hydrolysis than in that of methanolysis, because NH₄B(OCH₃)₄ is easily converted to AB upon reaction with LiAlH₄ and NH₄Cl, as reported by Ramachandran [33], whereas NH₄B(OH)₄ must first be converted to NH₄B(OCH₃)₄ by reaction with methanol (Equation (7), Scheme 5). However, the recycling of LiAl(OMe)₄ has not yet been conducted (Scheme 5) [33,88]

$$\begin{array}{c}{\mathrm{NH}}_4\mathrm{B}{\left({\mathrm{OCH}}_3\right)}_4+{\mathrm{NH}}_4\mathrm{Cl}+{\mathrm{LiAlH}}_4\to {\mathrm{NH}}_3{\mathrm{BH}}_3+\mathrm{Al}{\left({\mathrm{OCH}}_3\right)}_3+{\mathrm{CH}}_3\mathrm{OH}+{\mathrm{H}}_2+\mathrm{LiCl}+{\mathrm{NH}}_3\\ 81\% \mathrm{yield}\hspace{1em}\hspace{1em} \\ >98\% \mathrm{purity} \end{array}$$

(7)

The energy-consuming step in AB regeneration is the reduction of the strong bonds B-O in B(OCH₃)₃ or NH₄B(OH)₄, produced by AB solvolysis, which must proceed using a hydride such as LiAlH₄ Equation (7) [92].

A comparison of the mechanisms of nanocatalyzed AB dehydrogenative hydrolysis and methanolysis suggests that methanolysis is more boosted (3.7 times) by visible-light illumination compared to the reaction in the dark than hydrolysis (1.4 time). The KIE increased from 2.2 in the dark to 3.4 under illumination for AB methanolysis, but only from 4 to 4.3 for AB hydrolysis. This shows that visible light has much more influence on AB methanolysis than on AB hydrolysis. Upon examining the hydrogen bonding between the anionic hydridic species [Pd–H]⁻, formed by hydride transfer from AB to the NP, and the acidic solvent O–H bond, this bond is stronger upon AB hydrolysis than upon AB methanolysis, as water is more acidic than methanol. Consequently, solvent O–H bond weakening is more significant with water than with methanol, and solvent O–H bond oxidative addition on the nanocatalyst surface is easier with water than with methanol. This explains that the additional activation by visible light is necessary upon AB methanolysis compared to AB hydrolysis [66,78,93].

6. Conclusions

Among the H-rich H₂ sources [1], AB stands as one of the best suppliers upon solvolysis, either hydrolysis or alcoholysis. This issue has been raised by the seminal report from Chandra and Xu in 2006 on AB hydrolysis [4]. Since then, considerable efforts have been devoted to improvements toward safe and efficient H₂ production upon AB solvolysis. In particular, interest in these H₂ generation methods has been boosted by the in-depth seminal work by the Ramachandran group, since 2007, who disclosed that the AB methanolysis reaction improved AB synthesis and proposed its regeneration from AB solvolysis reactions [28,30,33].

Given the AB stability in water and alcohols, catalysis, particularly nanocatalysis by late transition-metal NPs and alloys stabilized by optimized supports, plays the central role in H₂ production from AB solvolysis. As in most catalytic reactions, noble metals (Ru, Rh, Pd, and Pt), despite their toxicity, high price, and scarcity, are more efficient than non-noble metals, i.e., those of the first-row transition metals (Co, Ni, Cu), but synergies among these metals and with the support can increase catalytic efficiencies, particularly doping with other elements such as B or P. Studies involving single-atom catalysts are also awaited [95,96]. Nanoporous supports are especially efficient and useful [97], because they favor the formation of ultrasmall and thus very efficient nanocatalysts and protects them from aggregation.

In comparing AB hydrolysis and methanolysis, it quickly appeared that AB hydrolysis encountered a major problem in that the AB hydrolysis product was all the more contaminated by NH₃ as the AB concentration was higher, which was damaging due to the NH₃ poisoning of fuel cells [98,99]. On the other hand, the H₂ produced is pure upon methanolysis.

The other drawback of AB hydrolysis is the regeneration after hydrolysis that is longer and more complex than regeneration after AB methanolysis. Finally, the last drawback is that the temperature window is much more limited for hydrolysis, due to water freezing at 0 °C, compared to methanolysis.

Favorable points for hydrolysis in comparison with methanolysis are that water is cheaper and less toxic than methanol and that the hydrolysis reactions are faster than the methanolysis reactions. Yet, TOFs of the AB methanolysis reaction with sophisticated nanocatalysts have very recently approached or overtaken 1000 mol_H2·mol_cat.⁻¹·min⁻¹, allowing perspectives for the application to low-temperature devices [84,85,86]. Ethanol is less toxic than methanol, but nanocatalyzed reactions are slower in ethanol than in methanol.

In summary, the most critical issue for AB solvolysis is the high cost of AB and of its regeneration after solvolysis, even more so for hydrolysis than for methanolysis. More research is required for improvements in catalysts and supports, particularly with emphasis on non-noble metal nanocatalysts. Up-to-date investigation tools such as in situ XRD, electron microscopies, X-ray absorption, theoretical calculations, and Machine Learning [100] should help the improvements, scaling, and transfer to industry for device applications toward an “H₂ economy”.