Contribution of metal-hydride species on the catalysts to catalytic performances

Metal hydride particles

Abundant M-H species exist in both the surface and bulk phase of M-H particles, which have good stability and can participate in catalytic reactions, so they have been extensively studied in catalysis and hydrogen storage. Among them, rare-earth-based AB₅-type metal hydrides were known as good hydrogen storage materials, but they were one of the first metal hydrides to be considered for catalytic properties. As early as the latter half of the last century, Soga et al. studied the catalytic properties of intermetallic AB₅ such as CaNi₅, LaNi₅, and alloy AB₅ such as LaNi₄M in olefin hydrogenation reaction and found that they exhibited high catalytic activity^[17-19]. Whereafter, Johnson et al. studied the catalytic properties and mechanism of the intermetallic LaNi₅H_x catalyst using 1-undecene hydrogenation as a model, and quantitatively confirmed the catalytic activity difference of a metal hydride precursor in different states^[20]. The authors found that the improvements in the catalytic activity of metal hydrides can be attributed to three factors:

(1) There was highly concentrated H in the metal hydride bulk phase, which can provide H on the catalyst surface.

(2) The presence of the hydride phase caused the volume expansion of the catalyst particles, which helped the H to migrate from the bulk phase to the catalyst surface through microcracks and grain boundaries.

(3) When the M-H species in the bulk phase had high activity, the surface hydrogen had weak chemisorption, which can improve the hydrogenation reaction rate.

Recently, Yu et al. prepared homogeneous LaNi_5.5 particles with a size of about 100 nm in molten salt by an improved CaH₂ reduction method, showing a unique core-shell structure with LaNi₅ core and Ni-rich shell [Figure 1A]^[21]. LaNi_5.5 showed high activity and cyclic stability for reversible hydrogen absorption and desorption of N-ethylcarbazole (NEC), comparable to that of noble metal catalysts. The fast H transfer rate at the LaNi₅-H solid solution and Ni/LaNi₅ interfaces was attributed to the excellent catalytic performance [Figure 1B and C]. Zhong et al. found that LaNi₅H₅ exhibited high activity and stability in the hydrogenation of CO₂ to methane, and based on the microstructure and chemical state of the catalyst, the high performance was attributed to the shortening of the H diffusion distance and the in-situ formation of Ni nanoparticles, which prevented agglomeration^[22]. However, the raw material cost of AB₅-type metal hydrides is too high, so much research has tried to replace the rare earth elements with other metals^[23,24]. In general, AB₅-type metal hydrides are good catalysts with small particle size, relatively large surface area, abundant transition metal sites, and less sensitivity to oxygen and water.

Figure 1. (A) Structure illustration of a LaNi_5.5+x; (B) Hydrogenation process of NEC on LaNi₅H_x at 7 MPa H₂; (C) Dehydrogenation process of 12H-NEC on LaNi₅ at 0.1 MPa H₂. Reproduced from Ref^[21]. Copyright (2021), with permission from Elsevier; (D) The Co 2p XP spectra of ZrCoH_1.2 measured at 1.0 mbar H₂, at 25 and 240 °C, respectively; (E) Reaction path for ZrCoH_x catalyst in CO₂ reduction. Reproduced from Ref^[25]. Copyright (2016), with permission from John Wiley and Sons. NEC: N-ethylcarbazole.

Mg-based A₂B-type and Zr-based metal hydrides have also been reported in catalytic studies. Kato et al. explored the source of catalytic activity of ZrCoH_x and Mg₂NiH₄ in CO₂ hydrogenation^[25,26]. The catalytic performances of ZrCoH_x (x = 0, 0.1, 1.2, 2.9) with different hydrogen contents were evaluated in a fixed-bed reactor, and it was found that the methane produced gradually increased with the hydrogen content. No methane production was observed when pristine intermetallic ZrCo was used as a catalyst. Time-of-flight secondary ion mass spectroscopy (TOF-SIMS) and near ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) were used to analyze the surface changes of the catalyst during the reaction. It was found that ZrCoH_x was initially covered by surface oxides, and the Co on the surface was reduced with the temperature increase under the inert air flow [Figure 1D]. In the hydrogen atmosphere, the degree of Co reduction was higher and the surface hydrogen concentration was also greater. A possible reaction mechanism was therefore proposed: CO₂ adsorbed on zirconia to form formates, while the overflow of M-H species in the bulk phase led to reduction of cobalt oxide at the interface. Gaseous hydrogen molecules then struck the reduced Co atoms to dissociate into hydrogen atoms, and all the surface H spilled over to zirconia. Finally, the formates reacted with the H on the zirconia to form methane. Hydrogen desorption and absorption occurred throughout the process to achieve cyclic catalysis [Figure 1E]. However, the reaction path and mechanism lacked characterizations of in situ spectroscopy. Mg₂NiH₄ also had hydrogen desorption and hydrogen absorption steps in CO₂ hydrogenation, but it showed a completely different phenomenon: in the process of hydrogen desorption under CO₂ stream, Mg was selectively oxidized, which resulted in disproportionation of Mg₂NiH₄, thus forming a Mg-rich oxide layer on the surface, along with the formation of Ni particles and MgNi₂. The higher concentration of basic MgO layers was prone to absorbing CO₂; Ni particles served as active sites for H₂ adsorption and dissociation. The dissociative adsorption of CO₂ and subsequent hydrogenation led to further oxidation of the surface.

In addition to catalysis, the aforementioned rare-earth-based AB₅-type and Mg-based A₂B-type metal hydrides are widely used in hydrogen storage; Zr/Ti-based AB₂-type and FeTi-based AB-type metal hydrides are also common hydrogen storage materials^[27]. Metal hydrides utilize metals/alloys to absorb and release hydrogen gas at a certain temperature and pressure for hydrogen storage. The ability of metal hydrides to reversibly absorb and desorb H₂ makes them very attractive for hydrogen storage applications. Some metals combine with hydrogen to form hydrides which are volatile compounds with low boiling points and cannot be used as hydrogen storage materials, such as CuH^[28]; the chemical properties of the M-H species are one of the important influencing factors for hydrogen storage applications, hydrogen bonds that are too strong or too weak are not suitable for hydrogen storage. At present, the preparation technology and process of metal hydrides for hydrogen storage are relatively mature.

Alkali and alkaline-metal hydrides (AHs) such as LiH, NaH, KH, BaH₂ and CaH₂ that do not contain transition metals are also proven to be active for some catalytic reactions. Gao et al. reported a looped ammonia synthesis process, in which BaH₂ acted as mediators and nitrogen carriers^[29]. In this cycle, N₂ was reduced by BaH₂ to generate BaNH, which further decomposed to obtain BaH₂ and NH₃. The valence state of H in BaH₂ varied between -1, 0, and +1. Although BaH₂ demonstrated catalytic activity in the ammonia synthesis reaction, Ni-BaH₂ yielded by the combination of BaH₂ with Ni through ball milling exhibited significantly superior catalytic performance. The yield rate of NH₃ was much higher than that of Cs-Ru/MgO under mild conditions below 250 °C^[13]. Consequently, although transition metal-free hydrides have been demonstrated to be active and play crucial roles in some reactions, the activities are typically not remarkable before being combined with transition metals.

A series of catalysts, synthesized by combining LiH, NaH, BaH₂ and other AHs with transition metals such as Co and Mn, demonstrated significant catalytic activities in ammonia synthesis [Figure 2]^[30,31]. Therefore, researchers used CaH₂ and MgH₂ as functional carriers to support transition metal particles to obtain a series of composite catalysts with high performances, such as Ru/BaO-CaH₂^[32], defect-rich MgH₂/Cu_xO^[33], Ni-MgO/MgH₂^[34] and Al-doped MgH₂^[12]. Ru/BaO-CaH₂ was employed in ammonia synthesis, and the high electron-donating capacity of the resultant BaH₂ was deemed pivotal to the high activity of the catalyst at low temperatures [Figure 3A]. The high electron-donating ability of BaH₂ facilitated the adsorption and dissociation of N₂, thus promoting the synthesis of ammonia. Concurrently, the absorption of H₂ by metal hydrides diminished the adsorption of H₂ on the surface of Ru, thereby impeding the phenomenon of hydrogen poisoning [Figure 3B]. Defect-rich MgH₂/Cu_xO and carbon-confined MgH₂ nanoparticles can achieve high C₂-C₄ selectivity and high CO₂ conversion at low H₂/CO₂ ratios when applied to CO₂ reduction. The abundant oxygen vacancies and Mg/O-heteroatom defects on MgH₂/Cu_xO can facilitate the adsorption and activation of CO₂, while the lattice H^- in MgH₂ with low concentration can promote the hydrogenation of CO₂ and the formation of carbon chains to produce short-chain olefins. Ni-MgO/MgH₂ and Al-doped MgH₂ exhibited high selectivity for methane in the hydrogenation of CO₂. Theoretical calculations^[34] indicated that in the Ni-MgO/MgH₂ system, CO₂ preferentially adsorbed and activated on the surface of MgO, and hydrogenated by H⁺ on the surface of Ni-doped MgO to form CO^* species (O-terminal hydrogenation). Subsequently, it reacted with H^- on the surface of MgH₂ to produce CH₄ (C-terminal hydrogenation). The H^- species provided by the surface of MgH₂ ensured the continuous hydrogenation of CO₂, which can be supplemented by H₂ dissociation even when the surface H^- species were consumed [Figure 3C].

Figure 2. (A) Temperature dependence of the NH₃ synthesis rates of BaH₂-Co/CNTs and BaO-Co/CNTs. The insert of (A) shows the three steps of NH₃ synthesis catalyzed by BaH₂-Co/CNTs. Reproduced from Ref^[30]. Copyright (2017), with permission from American Chemical Society; (B) Comparison of NH₃ yield of different catalysts, reaction condition: 10 bar syn-gas (N₂:H₂ = 1:3), 573 K, 60,000 mL·g·cat^-1·h^-1 WHSV; (C) Scheme for designing highly active ETM catalysts stimulated by Mn₄N-AHs was proposed. ETMs (active center 1) activate the reactant to form adsorbent A’, and active center 2 pulls A’ away from center 1 to form more stable A”, thereby releasing center 1 and allowing A to continue to be activated. A” can be combined with another reactant B or the active substance B’ to produce the final product. Reproduced from Ref^[31]. Copyright (2018), with permission from American Chemical Society. CNTs: Carbon nanotubes; ETM: early 3d transition metal; AHs: alkaline-metal hydrides.

Figure 3. (A) H₂-TPD profiles of used Ru/CaH₂, Ru/BaH₂, and Ru/BaO-CaH₂, which were performed under Ar flow (1 °C·min^-1). The three catalysts were used in the NH₃ synthesis reaction at 340 °C for 50 h; (B) Schematic diagram of the Ru/BaO-CaH₂ catalyzed NH₃ reaction. Reproduced from Ref^[32]. Copyright (2018), with permission from American Chemical Society; (C) Reaction path with minimum energy for absorbed CO^* hydrogenation to CH₄. Reproduced from Ref^[34]. Copyright (2021), with permission from Elsevier. TPD: Temperature-programmed desorption.

Based on the aforementioned examples, heterogeneous metal hydride catalysts typically played the following several roles in reactions: (1) providing additional H sources and H transfer routes; (2) facilitating hydrogen dissociation; (3) establishing new reaction mechanisms; (4) regulating the electronic structure of the catalyst. The intrinsic properties of M-H in metal hydrides and reaction conditions can both influence their promoting effects in catalysis. In the case of interstitial hydrides, such as LaNi₅H₅, high H-mobility was typically the determining factor in terms of high activity. In contrast, MgH₂, CaH₂ and other ionic hydrides generally exhibited low H-mobility, rendering them ineffective for hydrogenation reactions such as olefin hydrogenation which proceeded under milder conditions. In the context of ammonia synthesis, the typical reaction temperature exceeded 250 °C. At this temperature, the tuning effect of metal hydrides on the electronic structure became dominant, thereby rendering metal hydrides with low H-mobility effective catalysts.

However, it is obvious that the majority of relevant works on mechanism research relied on theoretical calculation and lacked in-situ characterization data and direct experimental evidence. Because of the heterogeneous structure of particle catalyst, the activity data and mechanism obtained can only reflect the average of catalyst. Concurrently, heterogeneous metal hydride catalysts were mostly synthesized via traditional metallurgical methods, which usually required a long time activation treatment with H₂ under high pressure before catalytic reaction, and usually had a particle size in the micron range and a relatively low surface area.

Molecular metal hydrides

Molecular metal hydrides, which are important homogeneous catalysts, have a long study history started from the discovery of the first example H₂Fe(CO)₄^[35] in the 1930s. These materials are concentrated in d-zone transition metals and f-zone elements, with a small number of alkaline earth metal complexes also reported. Transition-metal polyhydrides complexes L_nMH_x have been used as homogeneous catalysts in over 40 organic reactions, such as hydrogenation of olefins and alkynes, formation of amines, hydrosilylation, and so on, which were summarized in detail by Babón et al. in terms of catalytic reactions^[36]. Compared with heterogeneous catalysts, molecular metal hydride catalysts not only have higher activity, selectivity and mild reaction conditions, but also have great advantages in studying catalytic mechanisms and the dynamic changes of M-H during reactions. Molecular metal hydrides have given rise to a series of conceptual studies on M-H, such as the mechanisms of the collective motions of the H surrounding the metal center^[37], their thermodynamic acidity^[38-40] or hydricity^[41], the insight of interconversion processes between different H ligands, and the impact of the presence of M–H bonds of different types on the catalytic properties of the complexes. These are difficult to study by experiments and calculations in heterogeneous catalyst systems.

The versatility of molecular metal hydrides can be attributed to the manner, in which H coordinates with metal centers and the diversity of chemical properties of M-H. In general, H ligands can coordinate with metals in the form of H⁺/H⁰/H^- or in the form of H₂. Molecular metal hydrides may have multiple H ligands with different coordination modes simultaneously.

A classical molecular metal hydride RuHCl(PPh₃)₃ which contains a hydride ligand was used by Sasson and Rempel to catalyze the rearrangement of unsaturated alcohols into saturated ketones^[42]. RuHCl(PPh₃)₃ containing M-H showed a threefold greater activity than RuCl₂(PPh₃)₃, demonstrating the critical role of M-H in catalysis. Yadav and Gupta synthesized and characterized two groups of Ru complexes supported with coumarin-amide-based ligands, one group consisting of three [Ru-Cl] without hydrogen ligands and the other consisting of three [Ru-H] containing one hydrogen ligand^[43]. The authors applied them to transfer hydrogenation and found that [Ru-H] exhibited a remarkable catalytic activity, achieving a yield of over 90% without base. In contrast, [Ru-Cl] required the presence of base for catalysis. As depicted in Figure 4A, [Ru-H] catalyzed via an inner-sphere mechanism. Firstly, the substrate replaced one of the PPh₃ ligands in the catalyst to form intermediate A. Secondly, the carbonyl substrate was inserted into the Ru–H bond, and PPh₃ was re-attached, resulting in the Ru-alkoxide species B. Thirdly, the alkoxide species was substituted by isopropanol, resulting in the formation of the targeted alcohol product and intermediate species C. Finally, the catalyst was regenerated through β-hydride elimination, with acetone producing concurrently, which was detected by gas chromatography. Pandey et al. employed (^iPrPP^RP)CoH(PMe₃) to facilitate formic acid dehydrogenation, achieving catalytic turnover numbers of up to 7,122 with a single load of formic acid and 10,338 with a continuous addition of formic acid^[44]. The detailed ³¹P{¹H} nuclear magnetic resonance (NMR) and infrared (IR) spectrum revealed the reaction mechanism and the catalyst deactivation pathway. The dissociation of -PMe₃ promoted the formation of reactive intermediates, and M-H also played a key role in catalysis, which not only participated in the formation of intermediates, but also facilitated the formation of intermediates. Additionally, it promoted H₂ release and CO₂ generation [Figure 5A]. In these examples of molecular metal hydrides, the H ligands were characterized by a negative charge. However, there were also some positive and zero-valence H ligands for molecular metal hydrides. For example, the H ligands in [(MeCp)WH(CO)₃] were negative, which applied by Hammarstrom to proton-coupled electron transfer (PCET) reaction^[45]; the H ligands in CpW(CO)₂(IMes)H, [CpW(CO)₂(IMes)H]^•+ and [CpW(CO)₂(IMes)(H)₂]⁺ [IMes = 1,3-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene] reported by Roberts et al. were electrically neutral^[46].

Figure 4. (A) Inner-sphere mechanism for aldehyde/ketone transfer hydrogenation catalyzed by [Ru–H] supported with coumarin-amide-based ligands. Reproduced from Ref^[43]. Copyright (2023), with permission from American Chemical Society; (B) Stoichiometric C=C coupling reaction between [{(^Dippnacnac)Mg(μ-H)}₂] and CO, and CO reduction with PhSiH₃ catalyzed by [{(^Dippnacnac)Mg(μ-H)}₂]; (C) Proposed reaction cycle for CO reduction with PhSiH₃ catalyzed by [{(^Dippnacnac)Mg(μ-H)}₂]. Reproduced from Ref^[55]. Copyright (2015), with permission from John Wiley and Sons.

Figure 5. (A) Reaction path for dehydrogenation of formic acid catalyzed by (^iPrPP^RP)CoH(PMe₃), and the variation of (^iPrPP^RP)CoH(PMe₃) during the whole process. Reproduced from Ref^[44]. Copyright (2024), with permission from American Chemical Society; (B) Hypothesized mechanism for hydrogenation of Quinolines catalyzed by [Ir(η⁴-COD)(Me₂Bzim)(PPh₃)]⁺. Reproduced from Ref^[54]. Copyright (2011), with permission from American Chemical Society.

Many molecular transition metal polyhydrides have also shown high activities in numerous catalytic reactions. Osmium trihydrides OsH₃(acac)(PⁱPr₃)₂ (Hacac = acetylacetone) and OsH₃{k²-N_py,N_imine(BMePI)}(PⁱPr₃)₃ [HBMePI = 1, 3-bis (6’-methylpyridine-2’-imine) isoindoline] reported by Esteruelas et al. achieved high yield of H₂ in cyclic amine dehydrogenation^[47,48]. Zeiher et al. employed Rhenium-heptahydride ReH₇(PCy₃)₂ in the dehydrogenation of 1-arylethanols^[49], and cobalt-trihydride CoH₃{κ³-P,N,P-[py(CMe₂PⁱPr₂)₂]} reported by Lapointe et al. can obtain 4-octyne with a yield of 99% in hydrogenation of trans-4-octene^[50]. The dihydride-ruthenium(II)-bis(dihydrogen) complex RuH₂-(η²-H₂)₂(PCy₃)₂ with two H ligands and two dihydrogen ligands, originally applied in dehydrogenative cyclization, corresponding cyclic products were obtained in moderated-good yields after 3-72 h at 25 °C^[51], and Borowski et al. found the catalyst to be interesting for hydrogenation of aromatic hydrocarbons as well^[52,53]. [Ir(η⁴-COD)(Me₂Bzim)(PPh₃)]⁺ (Me₂Bzim = N,N-dimethyl-benzimidazolylidene) reported by Dobereiner et al. promoted the hydrogenation of quinolines under mild conditions^[54]. Stoichiometric studies and intermediate separation were conducted to reveal an outer-sphere mechanism involving sequential proton offering by dihydrogen ligands and Hydride transferring to protonated substrate [Figure 5B]. The density functional theory (DFT) calculation also showed the rationality of this mechanism.

The molecular metal hydrides were transition metal-based, with only a few alkaline earth metal hydrides having been reported thus far and primarily those magnesium-based. That was due to the large ionic radius of the heavier alkaline metals, namely Ca, Sr, Ba and Ra, which resulted in high activities of hydrides and facilitated the Schlenk rearrangement to form insoluble inorganic salts [AeH₂]_∞ (Ae = alkaline earth metals). Anker et al. made many efforts in the synthesis and catalysis of Mg-based molecular metal hydrides. Among them, they synthesized [{(^Dippnacnac)Mg}₂(μ-H)}₂] and used it for catalytic reduction of CO in the presence of PhSiH₃ to produce PhSiH₂-CH₃ and (PhH₂Si)₂O [Figure 4B]. The reaction mechanism was shown in Figure 4C and the driving force of the catalytic reaction was derived from the formation of siloxane (PhH₂Si)₂O^[55]. Shi et al. successfully synthesized a series of semi-sandwich heavy alkaline hydrides (Cp^Ar)Ae[CH(SiMe₃)₂](S) [Cp^Ar = C₅Ar₅, Ar = 3, 5-ⁱPr₂-C₆H₃; S = tetrahydrofuran (THF) or 1,4-diazabicyclo[2.2.2]octane (DABCO)], and hydrogenated in n-hexane solvent under medium pressure to obtain the corresponding dinuclear hydrides [(CpAr)Ae(μ-H)(S)]₂ (Ae = Ca, S = THF; Ae = Sr, Ba, S = DABCO)^[56]. These hydrides can efficiently catalyze the hydrogenation of olefin under mild conditions (30 °C, 6 bar H₂, 5% Cat.), and the activities of the catalyst and the application ranges of the reaction substrate increased with the increase of the ionic radius of metals (Ca²⁺ < Sr²⁺ < Ba²⁺).

Molecular metal hydride catalysts have great advantages in the study of catalytic mechanism, but they are susceptible to inactivation and difficult to recycle, which limit their practical applications.

Hydride-doped metal nanoclusters

Metal nanoclusters have recently emerged as a subject of considerable interest within the field of catalysis. They exhibit the characteristics of both homogeneous and heterogeneous catalysts, acting as a bridge between nanoparticles and small molecule complexes. The precise atomic composition and crystal structure, clear surface bonding mode and monodispersion of nanoclusters afford them significant advantages in the study of structure-activity relationships and the establishment of catalytic models, as well as in the investigation of catalytic mechanisms^[57-59]. Despite the numerous reports on hydride-doped metal nanoclusters, the difficulty of synthesis and poor stability limit their applications in catalysis research^[60,61].

Sun et al. reported a novel atomically accurate mercaptan protected Cu-hydride nanocluster [Cu₂₅H₁₀(SPhCl₂)₁₈]^3- (Cu₂₅H₁₀), which can catalyze the hydrogenation of ketones toward corresponding alcohols under mild conditions^[62]. The chemical environments of ten Cu-H in the cluster were studied by ¹H/²H NMR and their location in the cluster was calculated by DFT. A specific H site (Hc) with low resistance and high reactivity was predicted by DFT calculation, and an energy-competing reaction pathway was proposed: during the reaction, the reactant ketone/aldehyde reacted with a μ₃-H (H_c) at a specific location in the cluster to form Cu₂₅H₉-OCH₃ intermediates, which then continued to react with an adjacent μ₆-H to form alcohols, and finally the two H vacancies on the cluster were supplemented by H₂ heterolytic cleavage to complete the catalytic reaction cycle [Figure 6A]. The catalytic path was verified by ¹H/²H NMR and electrospray ionization mass spectrometry (ESI-MS) detection of clusters before and after the reaction under H₂/D₂ gas [Figure 6B]. Cu₂₅H₁₀ was one of the first nanocatalysts whose catalytic mechanism was studied with atomic precision. Liu et al. reported [Cu₂₀H₉(Tf-dpf)₁₀]BF₄ (Cu₂₀H₉) and [Cu₂₀H₈(Tf-dpf)₁₀](BF₄)₂ (Cu₂₀H₈) recently^[63]. They were very close in composition, differing only in the presence of a single H. This subtle difference gave rise to significant variations in their geometric and electronic structures, which in turn resulted in markedly distinct catalytic properties: in the conjugate reduction of cinnamaldehyde, the yield of Cu₂₀H₈ was 96.7%, which was 25 times higher than that of Cu₂₀H₉ (3.7% yield). According to the ESI-MS test results of the clusters after the reaction, combined with theoretical calculations, it was inferred that the high activity of Cu₂₀H₈ was due to its more prone to dissociation of Tf-dpf ligand during the catalytic process, thus exposing the copper active site. In contrast, no ligand dissociation was observed for Cu₂₀H₉ during catalysis, leading to its lower activity. M-H affected the catalytic performance by regulating the structure of the catalyst; whether it participated in the reaction and its dynamic change during the catalytic process had not been studied.

Figure 6. (A) Energy change and structures of intermediates in of a supposed route involves two H from Cu₂₅H₁₀; (B) ²H NMR spectra of Cu₂₅D₁₀ after catalysis with H₂ and fresh Cu₂₅D₁₀. Reproduced from Ref^[62]. Copyright (2019), with permission from American Chemical Society; (C) Conversion for 4-NP reduction over Ag₂₅Cu₄-1 (1 for short), Ag₂₅Cu₄H₈-2_H (2_H for short), and Ag₂₅Cu₄-1 in the presence of H₂O₂; (D) Schematic diagram of the conversion from 1 and 2_H and their performance on the catalytic reduction of 4-NP to 4-AP. With successively addition of H₂O₂ and NaBH₄, 1 can gradually convert to 2_H with better catalytic performance. Reproduced from Ref^[65]. Copyright (2022), with permission from American Chemical Society. NMR: Nuclear magnetic resonance; 4-NP: 4-nitrophenol; 4-AP: 4-aminophenol.

The reaction conditions for the reduction of 4-nitrophenol (4-NP) by NaBH₄ to 4-aminophenol (4-AP) are mild; therefore, it is a commonly used “model catalytic reaction” to study the properties of hydride-doped metal nanoclusters. Ni et al. reported three structurally similar PdAg nanoclusters [PdHAg₁₉(dtp)₁₂] [dtp = S₂P(OⁱPr)₂] (PdHAg₁₉), [PdHAg₂₀(dtp)₁₂]⁺ (PdHAg₂₀) and [PdAg₂₁(dtp)₁₂]^+[64]. It was observed that PdHAg₁₉ and PdHAg₂₀ contained interstitial hydrogen while PdAg₂₁ did not. In the reduction of 4-NP to 4-AP, to achieve the complete transformation, PdHAg₁₉ needed 25 min, PdHAg₂₀ needed 30 min, and PdAg₂₁ needed 60 min. Although there was no other experimental evidence, the authors attributed this performance difference to interstitial hydrogen: its presence led to a significant distortion of the nanocluster core, and this structural change may affect the catalytic performance; and interstitial hydrogen may be directly involved in the catalytic process. Yuan et al. reported two bimetal clusters: Ag₂₅Cu₄Cl₆(dppb)₆(PhC≡C)₁₂(SO₃CF₃)₃ (Ag₂₅Cu₄-1) and Ag₂₅Cu₄Cl₆H₈(dppb)₆(PhC≡C)₁₂(SO₃CF₃)₃ (Ag₂₅Cu₄-2_H), with the only difference being that Ag₂₅Cu₄-2_H contained 8 M-H while Ag₂₅Cu₄-1 has no M-H^[65]. Ag₂₅Cu₄-2_H showed good stability in 4-NP reduction and can be recovered and reused many times, with TON reaching 3538, while Ag₂₅Cu₄-1 had poor activity [Figure 6C]. Two clusters with such similar structures can have such different catalytic properties. Deuteration experiments revealed that M-H did not participate in the reaction. Consequently, the authors concluded that the M-H species in Ag₂₅Cu₄-2_H contributed to the stability of the cluster structure and the +1 oxidation state of Ag/Cu by forming μ₃-H-Ag₃ and μ₄-H-Ag₄ bonds. The electron-deficient Ag and Cu sites were conducive to the hydrogenation of BH₄^- and thus exhibited better catalytic activity [Figure 6D]. The above two works compared the distinctions in catalytic performance between structurally similar clusters comprising and lacking M-H, indicating that M-H species have a significant influence on the catalytic performance of catalysts, regardless of whether they directly participate in the reaction.

Liu et al. reported two structurally similar copper hydride nanoclusters protected by amido-ligands (Tf-dpf):Cu₁₁H₃(Tf-dpf)₆(OAc)₂ (Cu₁₁) and [Cu₁₂H₃(Tf-dpf)₆(OAc)₂]·OAc (Cu₁₂)^[66]. According to the DFT calculations, the two nanoclusters had distinct hydride locations, with Cu₁₁ having three interfacial μ₅-H and Cu₁₂ having three interfacial μ₆-H. In the reduction of 4-NP to 4-AP, the activity of Cu₁₁ was much higher than that of Cu₁₂:Cu₁₁ can completely convert 4-NP to 4-AP within 10 min, while the yield of equivalent Cu₁₂ was only 5% when reacting for 30 min. Because the structures of the two clusters were similar, the authors focused on the different Cu-H species of the two clusters. According to deuteration experiments, μ₅-H in Cu₁₁ participated in the catalytic cycle, while μ₆-H in Cu₁₂ was inactive. This work showed that the different positions of M-H could exhibit different catalytic properties.

The catalytic properties of hydride-doped metal nanoclusters in hydrogen evolution reaction (HER) and electrocatalytic CO₂ reduction reaction (CO₂RR) have also been reported. Kulkarni et al. reported a gold nanocluster [Au₂₄(NHC)₁₄Cl₂H₃]³⁺ (Au₂₄H₃) stabilized by N-heterocyclic carbene (NHC), which was the first time that gold nanoclusters containing hydrides were prepared^[67]. Based on DFT calculations, three metal hydrides acted as bridging ligands to connect the two Au₁₂ cores, and their wavenumbers in the IR spectrum ranged from 1,114 to 1,316 cm^-1. Au₂₄H₃ exhibited a CO Faraday efficiency of over 90% in CO₂RR, and calculations suggested that this may be due to the metal hydride assisting in the activation of the CO₂ into an easily reducible configuration. [Au₂₂H₃(dppe)₃(PPh₃)₈]³⁺ (Au₂₂H₃) reported by Gao et al. also had three bridge-H connect two Au₁₁ cores^[68]. The electrocatalytic reduction of CO₂ to CO showed high selectivity and reactivity [92.7% CO Faradaic efficiency (FE_CO), 134 A/g_Au mass activity], and the catalytic performance was significantly better than that of Au₁₁. Figure 7A and B showed the catalytic pathway: bridge-H on Au₂₂H₃ could directly hydrogenate CO₂ to form adsorbed ^*COOH, with a much lower Gibbs free energy (ΔG) than Au₁₁. Then absorbed ^*COOH formed by electrochemical proton reduction absorbed CO^* and water. Finally, the CO desorbed, and the consumed bridge-H were replaced by proton reduction. However, Au-H in clusters did not all promote CO generation. The [Au₇(PPh₃)₇H₅](SbF₆)₂ (Au₇H₅) clusters reported by Tang et al. showed high hydrogen selectivity (98.2% FE H₂) in CO₂RR as catalyst^[69]. Based on Hirshfeld charge analysis, the five H on the surface of Au₇ were in negative valence states. The conclusion was also supported by XPS, showing that the valence states of Au atoms in Au₇H₅ were close to +1. Compare to Au₇H₅, [Au₈(PPh₃)₇]²⁺(Au₈) showed high CO selectivity. Combined with DFT calculations, it was speculated that CO₂RR electroreduction was inhibited by hydrides presenting as ligands in Au₇H₅ [Figure 7C and D].

Figure 7. (A) Structures obtained from DFT of intermediates for Au₂₂H₃ in electrocatalytic Reduction of CO₂ to CO; (B) Free energy change for Au₁₁ and Au₂₂H₃ in CO₂RR. Reproduced from Ref^[68]. Copyright (2022), with permission from American Chemical Society; Free energy change for Au₇H₅ (C) and Au₈ (D) in CO₂RR and HER. Reproduced from Ref^[69]. Copyright (2023), with permission from John Wiley and Sons. DFT: Density functional theory; HER: hydrogen evolution reaction.

Brocha Silalahi et al. reported two clusters: [PdHCu₁₁{S₂P(OⁱPr)₂}₆(C≡CPh)₄] (PdHCu₁₁) and [PdHCu₁₂{S₂P(OⁱPr)₂}₅{S₂PO(OⁱPr)}(C≡CPh)₄] (PdHCu₁₂)^[70]. Their metal hydrides were located in the tetrahedron of PdCu₃ and strongly bound to Pd. PdHCu₁₁ exhibited excellent HER activity with a record initial potential of -0.05 V (at 10 mA·cm^-2), and maintained stable after 1,000 cycles in 0.5 M H₂SO₄. While the catalytic performance of PdHCu₁₂ was significantly inferior. Deuteration experiments demonstrated that Pd-H did not participate in the reaction. Combined with theoretical calculations, the high activity of PdHCu₁₁ was attributed to its open structure, which made it easier for the reactants to access the central Pd atom as the active site, while the encapsulated H was observed to play a role in maintaining the stability of the cluster structure.

Although hydride-doped nanoclusters have unique advantages in the study of catalytic mechanisms, the current relevant work has not been deeply studied in terms of its dynamic changes and the roles of M-H species in catalysis, with little experimental evidences and mainly relying on theoretical calculations.