Skip to main content
NCBI home page
Nature Communications logoLink to Nature Communications
. 2025 Aug 23;16:7862. doi: 10.1038/s41467-025-63124-8

Atomically precise gold nanocluster boosting selective hydrogenation of nitroarene by H2 in water

Xiao-Xiao Lai 1,#, Jianyu Wei 2,#, Ting-Ting Liu 1, Jing Li 1, Jiao-Jiao Li 3, Xian-Kai Wan 1,✉, Quan-Ming Wang 3,✉
PMCID: PMC12375014  PMID: 40849407

Abstract

The chemoselective hydrogenation of molecules containing multiple reducible groups using H2 presents inherent challenges. Here we show a homoleptic nanocluster [Au40(ArC≡C)22](Et4N)2 (Au40 for short, ArC≡C is 3,5-bis(trifluoromethyl)-phenylacetylide) is synthesized in high yield and its structure is elucidated using single-crystal X-ray diffraction. DFT calculations reveals that Au40 features a superatomic 20-electron configuration of (1S)2(1P)6(1D)10(1F)2. Au40/TiO2 exhibits 100% selectivity and activity for the hydrogenation of 4-nitroacetophenone in water, free of base and under mild conditions (80 °C, H2 10 bar). The turnover numbers reach a value of 335,569, and the turnover frequencies 5829 h−1 is an order of magnitude higher than those observed in the well-established Au/TiO2 system. The improved catalytic performance of Au40 is attributed to the synergy of its enhanced durability, unique molecular structure. This work demonstrates that tailoring the surface coordination structure is an effective way to modulate the catalytic performance of cluster catalysts.

Subject terms: Heterogeneous catalysis, Molecular self-assembly, Ligands

The chemoselective hydrogenation of molecules containing multiple reducible groups using molecular hydrogen presents inherent challenges. Here, the authors synthesize a homoleptic gold nanocluster that demonstrates exceptional catalytic selectivity, activity, and yield for the hydrogenation of 4-nitroacetophenone in water under mild conditions.

Introduction

The role of hydrogenation processors within the chemical industry is of considerable significance. It is estimated that approximately 25% of chemical transformations involve at least one hydrogenation step1–3. Consequently, it is not surprising that hydrogenation reactions represent one of the most extensively studied topics within the field of catalysis. The selective hydrogenation of chemicals with multiple reducible groups, including -C≡C, -C=O, -NO2, and -C≡N, has captured numerous interests due to its inherent advantages to facilitate the production of high-value functional chemicals1,4,5. Among these, aromatic amines containing carbonyl groups are crucial intermediates in diverse fine chemical industries, including pharmaceuticals, specialty chemicals, agrochemicals, polymers, and pigments5–9. The selective hydrogenation of nitroaromatics represents an effective way to prepare aminoaromatics8,10,11. The most commonly used reducing agents include H2 and indirect hydrogen sources such as N2H4, NaBH4, NH3·BH3, and RnSiH4-n1,12–14. The indirect hydrogen sources require not only pre-preparation and activation but also the use of dry and inert atmospheres, which makes them costly and difficult to control. Such requirements pose challenges for scalability and are associated with issues related to reaction control, product separation, and catalyst degradation15–17. Therefore, it is highly desirable to use clean and cheap H2 for selective hydrogenation. Gold nanoparticles (Au NPs) have been extensively investigated as heterogeneous catalysts for the chemoselective hydrogenation of carbonyl-substituted nitroaromatic compounds to aromatic amines using H25,18–20. Unfortunately, they not only suffer from poor activity but also frequently require environmentally unfriendly solvents, additives, and harsh reaction conditions (pH2 > 3 MPa, T > 120 °C), which hinders practical applications.

Corma et al.21,22. have demonstrated that low-coordinated Au atoms are essential for the dissociation of H2 in Au NPs-based catalysts for hydrogenation. Therefore, the precise construction of low-coordinated Au centers on the surface of Au-based catalyst is the key to improve the catalytic performance of hydrogenation with H2. Gold nanoclusters protected by organic ligands, which have abundant low-coordinated Au centers and atomically tunable steric hindrance of the Au-ligand shell (usually called staple motifs), have emerged as a promising class of catalyst with improved catalytic activity and selectivity toward selective hydrogenation with H223–27, for instance, in the chemoselective hydrogenation of α,β-unsaturated aldehydes to the corresponding α,β-unsaturated alcohols28, the semihydrogenation of alkynes to alkenes29,30, and the chemoselective hydrogenation of aromatic nitro compounds31–33. The atomically tunable structure and composition of the nanoclusters facilitate the synthesis of cluster-based catalysts, which combine the characteristics of homogeneous and heterogeneous catalysts and exhibit excellent activity and durability34–39. Therefore, the construction of atomically precise cluster catalysts contributes not only to the practical application of selective hydrogenation, but also to the study of structure-property relationships and the clarification of the mechanism at the atomic level.

Herein, we report the synthesis and total structure determination of a superatomic [Au40(ArC≡C)22](Et4N)2 (Au40) nanocluster. Density functional theory (DFT) calculations revealed that the Au40 features a superatomic Au26 core with a free electron number of 20 in an unusual electronic configuration of (1S)2(1P)6(1D)10(1F)2. The molecular and electronic structure of Au40 is significantly different from that of the thiolate protected Au40 reported by Wu et al.40 and Jin et al.41. Remarkably, the Au40/TiO2 catalyst exhibited 100% conversion and selectivity towards the chemoselective hydrogenation of carbonyl-substituted nitroaromatic compounds to aromatic amines in water, free of base, and under mild reaction conditions (80 °C, H2 10 bar). The turnover numbers (TON) are 335569 and the turnover frequencies (TOF) 5829 h−1 is an order of magnitude higher than that observed for TOF in the well-established Au/TiO2 system towards selective hydrogenation. Au40/TiO2 catalyst exhibits general applicability to various nitroaromatics. DFT calculations revealed that the unsaturated Au(I) sites derived from ArC≡C-Au-C≡CAr staple motif are the active center for H2 activation and dissociation. The alkynyl ligands on the surface of Au40 remain attached throughout the entirety of the reaction. Therefore, the excellent catalytic performance of Au40/TiO2 can be attributed to the synergistic effect of the unique molecular structure, the steric hindrance of ligands, the excellent stability of Au40, and the carrier effect.

Results

Synthesis and characterization

A direct reduction strategy was employed to synthesize Au40. The ArC≡CAu precursor was directly reduced by NaBH4 in a mixed solution of dichloromethane/methanol (v/v = 3/1) containing MeONa, followed by preparative thin-layer chromatography (PTLC) separation (Supplementary Fig. 1). As reported that the species that reduces gold(I) is BH3OH, which originates from the hydrolysis of BH4−42. The hydrolysis was catalyzed by H+, so in basic (MeONa) solutions, the reduction process becomes more slowly. This can inhibit over-reduction and increase the yield of Au40 nanoclusters. The separation yield is 56% (based on the molecular weight with the Na+ form and the moles of the used Au), which is higher than that of the majority of gold nanoclusters that have been previously reported. Finally, high-quality crystals of Au40 were obtained for SCXRD through the diffusion of n-hexane into a dichloromethane solution (detailed synthesis procedures are described in the method). The composition of Au40 was first characterized by electrospray ionization time-of-flight mass spectrometry (ESI-TOF-MS) in negative mode (Fig. 1a). The peak at m/z 6547.48 corresponds to the molecular ion [Au40(ArC≡C)22]2- of Au40. The observed isotopic pattern of the dianion cluster was in perfect agreement with the simulation results (Fig. 1a, inset). The presence of the countercation [NEt4]+ in Au40 was confirmed by ESI-TOF-MS in positive mode (Fig. 1b).

Fig. 1. Mass spectra of Au40.

Fig. 1

Open in a new tab

a, b The ESI mass spectra of Au40 in negative and positive mode respectively. Inset: Measured (black trace) and simulated (red trace) isotopic patterns of [Au40(ArC≡C)22]2- and [(CH3CH2)4N]+.

The 1H NMR analysis of Et4NBF4 and Au40 was performed using CD2Cl2 (Supplementary Fig. 2a and b). The peaks observed at δ = 1.33, 3.24 and 6.85-8.40 ppm are attributed to the CH3 of Et4N+, the CH2 of Et4N+, and the H on the aromatic ring of ArC≡C in Au40, respectively. The integration ratio of the phenyl H in the ArC≡C ligands, methylene H, and methyl H in Et4N+ was calculated to be7.96: 2.00: 3.12 (Supplementary Fig. 2b), which is consistent with the theoretical value of 8.25: 2.00: 3.00 in Au40. This suggests that cluster Au40, with the molecular formula [Au40(ArC≡C)22](Et4N)2, remained intact in CD2Cl2 solution. The multiple chemical shifts of 1H and 19F NMR spectra demonstrated that the ArC≡C ligands in Au40 nanoclusters feature different coordination environments (Supplementary Fig. 2b and c). Thermogravimetric analysis (TGA) results demonstrated that Au40 remained intact when the temperature was below 200 °C. When the temperature was increased to 400 °C, the sample lost 22.2% of its weight, corresponding to the loss of 12 ArC≡C ligands from the Au40 nanoclusters, and when the temperature reached 750 °C, the ligands were fully decomposed (Supplementary Fig. 3). This indicates that Au40 exhibited excellent thermal stability. The bonding energies of Au 4f7/2 and Au 4f5/2 in Au40 were determined to be 84.4 and 88.1 eV, respectively, by X-ray photoelectron spectroscopy (XPS) (Supplementary Fig. 4). The absence of bands at 84.9 and 83.8 eV for AuI and Au0, respectively, indicates that the Au atoms in Au40 are partially in the Au0 state. As illustrated in Supplementary Fig. 5, the C≡C stretching peak at around 1960 cm-1 in the Raman spectrum of Au40 clearly showed a blue shift, where the C≡C stretching peak for ArC≡CH is around 2120 cm-1 upon binding to Au atoms in Au4043.

Molecular structure of Au40

The molecular structure of Au40 was analyzed by SCXRD (Supplementary Table 7), and it was found that the [Au40(ArC≡C)22]2- in Au40 had quasi C2 symmetry and consisted of 40 gold atoms and 22 alkynyl ligands (Fig. 2). Au40 comprises a C2-symmetric Au26 kernel (Supplementary Fig. 6) and surface “staple” motifs, including four monomeric Au(ArC≡C)2, two V-shaped dimeric Au2(ArC≡C)3, and two h-shaped trimeric Au3(ArC≡C)4 “staple” motifs (Supplementary Fig. 7). The Au26 core in Au40 consisted of two Au14 units, which are composed of Au944 and Au745, through the sharing of two Au atoms (Fig. 3a), by sharing two Au atoms (Fig. 3b, highlighted in green). The four monomeric Au(ArC≡C)2 “staple” motifs were distributed at the upper and lower ends of the Au26 core (Fig. 3c, highlighted in pink). The two V-shaped dimeric Au2(ArC≡C)3 “staple” motifs were distributed on the left and right sides of the Au26 core (Fig. 3d, highlighted in violet). The two h-shaped trimeric Au3(ArC≡C)4 “staple” motifs were distributed on the front and back sides of the Au26 core (Fig. 3e, highlighted in blue). Therefore, the molecular structure of Au40 was characterized by a core-shell configuration (Supplementary Fig. 8a).

Fig. 2. The total structure of [Au40(ArC≡C)22]2- anion in Au40.

Fig. 2

Open in a new tab

The H atoms have been omitted for clarity.

Fig. 3. Anatomy of the molecular structure of Au40.

Fig. 3

Open in a new tab

a Assembling of Au9 and Au7 into Au14. b Assembling of two Au14 into Au26 core in Au40. c The Au26 kernel surrounded by four Au(ArC≡C)2 monomeric “staple” motifs (pink). d The Au26 kernel surrounded by two Au2(ArC≡C)3 dimeric “staple” motifs (violet). e The Au26 kernel surrounded by two Au3(ArC≡C)4 trimeric “staple” motifs (blue). Color labels: orange/turquoise/green/pink/violet/blue = Au; broght-green = F; gray = C.

The molecular structure of Au40 differs from the bi-icosahedron-based thiolate Au40(SR)24, which was predicted by Häkkinen et al.46 and Jiang et al.47 through theoretical calculations. As shown in Supplementary Fig. 8b, the molecular structure of Au40 is distinct from that of Au40(o-MBT)24 (o-MBT = 2-methylbenzenethiolate), as reported by Jin et al.41. The latter comprises a snowflake-like Au25 kernel (highlighted in orange) and nine surface-protecting staples, including six Au(o-MBT)2 monomeric staples and three Au3(o-MBT)4 trimeric staples (highlighted in pink and blue, respectively). Remarkably, Au40(S-Adm)22 (S-Adm = adamantanethiolate), as reported by Wu et al.40, possess the same molecular formula with Au40 but exhibit completely disparate molecular structure. As shown in Supplementary Fig. 8c, Au40(S-Adm)22 comprised an Au29 kernel (highlighted in orange) and an exterior shell consisting of six Au(S-Adm)2 monomeric staples, one Au2(S-Adm)3 dimeric staple, and one Au3(o-MBT)4 trimeric staple (highlighted in pink, violet, and blue, respectively). The Au40 and Au40(S-Adm)22 isomers provide a possible platform for investigations of ligand-regulated structures and properties of cluster-based materials at the atomic level. Moreover, the molecular structure of Au40 is distinct from that of alkynyl and phosphine co-protected [Au40(PhC≡C)20(dppm)4]4+ (PhC≡C=phenylacetylene, dppm = bis(diphenylphosphino) methane), as reported by Wang et al.48. The latter comprises a face-centered cubic (fcc) Au34 kernel and six surface-protecting staples (Supplementary Fig. 8d, highlighted in orange and pink, respectively), including two Au(PhC≡C)2 monomeric staples and four PhC≡C–Au–P^P staple motifs. The structure of Au40 differs from that of all previously reported Au nanoclusters with 40 Au atoms, suggesting that Au40 probably possesses unique physicochemical properties and has potential applications. The Au-Au distances in the Au26 core of Au40 are in the range of 2.6128(10)− 3.4395(12) Å, with an average of 2.84 Å. While, the Au-Au bond length between the Au atoms in the staple motifs, including monomeric Au(ArC≡C)2 and dimeric Au2(ArC≡C)3 “staple” motifs, and the Au26 core ranged from 2.8085(12) to 3.5745(10) Å, giving an average of 3.22 Å. This is 13.4% longer than the average Au-Au bond length in the Au26 core. The disparate Au-Au bond length distributions indicate that it is reasonable to divide the molecular structure of Au40 into an Au26 kernel and peripheral organometallic staple motifs shell.

The monomeric Au(ArC≡C)2 and V-shaped dimeric Au2(ArC≡C)3 staple motifs in Au40 were similar to those observed in previously reported alkyny-protected [Au23(PhC≡C)9(Ph3P)6](SbF6)249 and Au44(PhC≡C)2850. In contrast, the h-shaped trimeric Au3(ArC≡C)4 staple motif (Supplementary Fig. 9a) observed in Au40 differed from the π-like trimeric Au3(tBuC≡C)4 staple motif (Supplementary Fig. 9b) previously reported for Au22(tBuC≡C)1851. The Au-Au distance under the alkynyl group in the h-shaped trimeric Au3(ArC≡C)4 staple motif ranges from 3.1710(11) to 3.5741(10) Å, with an average of 3.41 Å. In the π-like trimeric Au3(tBuC≡C)4 staple motif, the Au-Au distance ranged from 3.4404(12) to 3.5136(11) Å, with an average of 3.48 Å. The Au-Cα-Au angles in the h-shaped trimeric Au3(ArC≡C)4 staple exhibit a large range of 98.0(6)−115.2(7) °, and that in π-like trimeric Au3(tBuC≡C)4 staple exhibit a narrow range of 110.2(12)−112.6(11) °, with an average of 107.1 and 111.4°, respectively. This suggests that the Au-Cα-Au angles in the h-shaped trimeric Au3(ArC≡C)4 staple are more flexible and variable than those in the π-like trimeric Au3(tBuC≡C)4 staple. Moreover, the Au-Au distances under the S-Adm and Au-S-Au angles in the trimeric Au3(S-Adm)4 staple (Supplementary Fig. 9c) of Au40(S-Adm)2240 were observed to range from 3.1270(2) to 3.6293(21) Å and 83.4(3) to 104.3(4) °, with average of 3.30 Å and 90.6°, respectively. The Au-Au distance and Au-S-Au angles in the trimeric Au3(S-Adm)4 staple are observed to be smaller than those in Au40(3.30 Å vs 3.41 Å, and 90.6° vs 107.1°). This indicates that the h-like trimeric Au3(ArC≡C)4 staple has a larger volume, which is more conducive to exposing the Au active center in the staple motif. Molecular packing diagrams of Au40 in the crystal are shown in Supplementary Fig. 10. Each unit cell contained four [Au40(ArC≡C)22]2- molecules. As illustrated in Supplementary Fig. 11, the results of the Hirschfeld surface analysis indicate that the volume and surface area of the [Au40(ArC≡C)22]2- anion are 6706.09 Å3 and 2443.44 Å2, respectively.

Electronic structure and optical properties

The electronic structure of [Au40(ArC≡C)22]2- was investigated using DFT calculations. The ideal symmetry of Au40 is C2; however, DFT calculations revealed that it was slightly distorted to C1 following relaxation. As illustrated in Supplementary Table 1, the DFT-optimized structure (Supplementary Fig. 12) exhibited comparable Au-Au distances with those observed by SCXRD. The computed highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) gap of 0.97 eV, calculated at the GGA level of theory, is in accordance with the experimental result of 0.76 eV, exhibiting a slight blue shift. The calculated average natural atomic orbital (NAO) charges of the Au atoms within the Au26 inner core (Aucore) and the periphery Aum(ArC≡C)n (m = 1/2/3, n = 2/3/4) staple motifs (Austaple) are of +0.17 and +0.52, respectively (Supplementary Table 2). This corresponds to natural electronic configurations of 5d9.81 6s0.96 6p0.04 (Aucore) and 5d9.58 6s0.89 6p0.01 (Austaple). These results are consistent with the proposed cluster structure, which describes a mixed-valent [Au26]6+ kernel surrounded by 14 Au(I) centers and 22 [ArC≡C]- ligands. The number of free electrons in Au40 is given by the equation N* = NAu - Nalkynyl – Q, where N* is the number of electrons, NAu is the number of Au atoms, Nalkynyl is the number of alkynyl ligands, and Q is the charge (−2) of Au40. Therefore, N* = 40 − 22 + 2 = 20 corresponds to a closed-shell superatomic electron count. According to the superatom model52, Au40 is expected to exhibit a closed-shell jellium configuration of 1S21P61D102S21F0. Nevertheless, a detailed analysis of its Kohn-Sham frontier molecular orbitals (KS-MOs) diagrams (Fig. 4a, b) indicate that the Au40 features an unconventional electronic configuration, characterized by a partially occupied 1F2 shell and a vacant 2S0 level. Such an unconventional configuration originates from the Jahn–Teller distortion away from spherical symmetry53,54, resulting in significant level splitting of the superatomic shell for Au40. In this case, the rod-shaped geometry of the Au kernel gives rise to a stabilized (occupied) 1Fz3 level (HOMO) and destabilized (vacant) 2S level. In this case, Au40 can be described as a Jahn–Teller distorted superatom with a jellium configuration of 1S21P61D101F2.

Fig. 4. DFT calculation results for [Au40(ArC≡C)22]2- under BP86/TZP/ZORA level of theory.

Fig. 4

Open in a new tab

a DFT-calculated KS-MOs diagram of [Au40(ArC≡C)22]2-, the alkynyl ligand-based MOs are not shown, and the occupied low-lying bonding MOs are associated with the σAu-C bonds. b Its corresponded plots of superatomic orbitals. c Experimental absorption spectrum of Au40 in dichloromethane (inset: the photo of solution). d The comparation of experimental and TD-DFT + TB simulated UV-vis-NIR spectra of [Au40(ArC≡C)22]2-, the blue sticks represented the computed oscillator strength for each excited state.

The ultraviolet-visible-near infrared (UV-vis-NIR) spectrum of Au40 in dichloromethane exhibits four distinct characteristic absorption peaks at 550, 642, 786, and 1175 nm (Fig. 4c). The molar extinction coefficient (ε) of Au40 at 550 nm is 33075.7 M−1cm−1 (Supplementary Fig. 13). The excitation properties of Au40 were investigated by TD-DFT + TB55 approach, wherein the excitation energies and oscillator strengths of the lowest 1200 excited states were calculated. The simulated UV-vis-NIR spectrum shows little difference from the experimental spectrum (Fig. 4d). The absorption spectrum at −66 °C was similar to that at 25 °C, except that the absorption intensity changed, indicating that the structure of Au40 in solution did not change (Supplementary Fig. 14), and the differences between the observed and simulated absorption spectra caused by the unideal condition for experiment rather than the structural change in solution. As shown in Supplementary Fig. 15, the first three lowest energy bands (α, β, and γ) are the metal-to-metal charge transition (MMCT) and exhibit superatomic 1 F (HOMO) → 1 F (LUMO), 1 F (HOMO) → 1Gz4 (LUMO + 2), and 1D (HOMO-2, HOMO-3) → 1 F (LUMO, LUMO + 1) natures, respectively. The excitation band δ corresponded to the metal-to-ligand charge transition (MLCT) from the superatomic 1D (HOMO-1) to the π*(ligand)-type LUMO + 8 (Supplementary Fig. 15). The highest energy band ε is composed of several mixed MMCT and MLCT involving 5 d(Au) MOs, superatomic 1D and 1 F MOs, and π*(ligand) MOs (Supplementary Fig. 15). The stability of Au40 in CH2Cl2 was monitored by UV/vis spectroscopy (Supplementary Fig. 16). No decomposition occurred after seven weeks of storage under either light-avoiding or non-light-avoiding conditions, indicating that Au40 features good solution stability.

Catalytic performance of Au40

The excellent stability and unique geometrical and electronic structures of Au40 make it a valuable candidate for a wide range of potential catalytic applications. Moreover, the application of Au40 to heterogeneous catalysis facilitates the homogenization of heterogeneous catalysts, which not only facilitates the recycling of catalysts but also enables the investigation of catalytic mechanisms at the atomic level. Chemoselective hydrogenation of 4-nitroacetophenone with H2 in water was performed as a probe reaction to investigate the potential application of Au40 in catalyzing the selective hydrogenation of substituted nitroaromatics (Fig. 5). TiO2 (anatase) was employed as a support, and an impregnation method was utilized to prepare the Au40/TiO2 catalyst. The Au40 nanocluster was adsorbed onto the TiO2 via electrostatic interactions. Typically, TiO2 powder was immersed in a CH2Cl2 solution of Au40, which was then subjected to centrifugation, drying, and annealing at 130 °C for 1 h in a vacuum tube under vacuum. The detailed process is described in the Supplementary Information. To evaluate the stability of Au40 during the preparation of the Au40/TiO2 catalyst, it was necessary to bring Au40 to the adsorption equilibrium on the TiO2 surface. This was achieved by the addition of excess Au40, which allowed the stability of Au40 to be determined by testing the UV/Vis spectrum of the supernatant. As demonstrated in Supplementary Fig. 17, the UV/Vis spectrum of the supernatant was identical to that of the Au40 nanocluster, indicating that the Au40 nanocluster remained intact after immobilization. To further investigate the thermal stability of Au40 in the Au40/TiO2 catalyst during the thermal treatment process, we characterized a CH2Cl2 solution of unsupported Au40 that was subjected to a thermal treatment similar to that applied to Au40/TiO2 (130 °C for 1 h under vacuum conditions) by UV/vis spectroscopy. As illustrated in Fig. 6a, the Au40 nanocluster exhibited identical profiles before and after thermal treatment, indicating that Au40 remained intact after thermal treatment. This result suggests that Au40 in the Au40/TiO2 catalyst remains intact throughout the immobilization and thermal treatment process.

Fig. 5. Proposed reaction equation for the hydrogenation of 4-nitroacetophenone.

Fig. 5

Open in a new tab

Heterogeneous catalyst is Au40/TiO2. H2O and H2 act as solvent and hydrogen source respectively.

Fig. 6. Evaluation of catalyst stability and activity.

Fig. 6

Open in a new tab

a UV/vis spectra of Au40 in CH2Cl2 to confirm the stability of Au40 after thermal treatment at 130 °C for 1 h under vacuum conditions. b chemoselective hydrogenation of 4-nitroacetophenone activity for Au40/TiO2. c plot of ln(Ct/C0) versus time for Au40/TiO2 (red) and TiO2 (black). d recycling tests of chemoselective hydrogenation of 4-nitroacetophenone over Au40/TiO2. Reaction conditions: 4-nitroacetophenone (8.26 mg, 0.05 mmol), 1.0 mL H2O, 80 °C, H2 10 bar, 40 mg catalyst.

The Au40/TiO2 catalyst was employed in the chemoselective hydrogenation of 4-nitroacetophenone; the detailed process is described in the Supplementary Method section. Subsequently, the resulting products were analyzed by 1H NMR spectroscopy (Fig. 7). It should be noted that the most commonly used reducing agents for the reduction of nitroaromatics are NaBH4 or aminoborane. In contrast, our approach involves the use of H2 and water instead of environmentally unfriendly organic solvents (e.g., toluene) and a base-free system. Neat TiO2 exhibited essentially no catalytic activity, whereas the Au40/TiO2 catalyst displayed excellent catalytic activity (100% conversion and selectivity) toward the chemoselective hydrogenation of 4-nitroacetophenone to 4-aminoacetophenone under mild conditions of 80 °C and H2 10 bar (Fig. 6b and Supplementary Fig. 18). However, the same reaction catalyzed by Au nanoparticles usually requires a minimum temperature higher than 100 °C or a pressure higher than 20 bar18. The catalytic performance of Au40/TiO2 with different solvents, including water, ethanol, methanol, and acetonitrile, was examined. As shown in Supplementary Table 3, when acetonitrile was utilized as the solvent, the conversion was only 10.5%. In contrast, when water, ethanol, and methanol were used as solvents, both the conversion and selectivity were found to be 100%. However, when ethanol and methanol are used as solvents, the Au40 nanocluster tends to detach from the TiO2 carrier, leading to faster deactivation of the Au40/TiO2 catalyst. It was determined that water would be the most suitable solvent for this experiment, given its insolubility with the Au40 nanocluster, environmentally friendly, as well as its capacity to facilitate the formation of a solid-liquid-gas three-phase interface on the hydrophobic surface of Au40 nanocluster thus promoting the H2 adsorption and activation. To investigate the carrier effect, we selected several other oxides as carriers, including Al2O3, hydroxyapatite (HAP), and CeO2. The Au40/Al2O3, Au40/HAP, and Au40/CeO2 catalysts were prepared using the same protocol as for Au40/TiO2, and the catalytic performances of different oxides and catalysts for the chemoselective hydrogenation of 4-nitroacetophenone were studied under the same conditions. Supplementary Table 4 demonstrated that all the oxides, Au40/Al2O3 and Au40/HAP exhibited no catalytic activity. However, for the selective hydrogenation of 4-nitroacetophenone, Au40/TiO2 exhibited 100% selectivity and conversion, whereas Au40/CeO2 achieved selectivity and conversion of 3.2% and 86.8%, respectively. By the way, Au40 nanocluster without TiO2 did not exhibit catalytic activity for the chemoselective hydrogenation of 4-nitroacetophenone. These indicate that the TiO2 plays a crucial role in the catalytic performance of Au40-based heterogeneous catalyst for the chemoselective hydrogenation of 4-nitroacetophenone. As demonstrated in Supplementary Table 5, when the reaction temperature was reduced from 80 to 65 °C and the H2 pressure was increased from 10 to 25 bar, the conversion and selectivity for 4-aminoacetophenone both reached 100%. However, when the temperature was further reduced to 60 °C and the H2 pressure increased to 30 bar, the conversion and selectivity for 4-aminoacetophenone were 49.3% and 100%, respectively. These indicate that to ensure complete conversion, the reaction temperature must not be lower than 65 °C.

Fig. 7. 1H NMR (400 MHz) spectral analysis of catalytic products at different reaction times.

Fig. 7

Open in a new tab

Reaction conditions: 4-nitroacetophenone (8.26 mg, 0.05 mmol), 1.0 mL H2O, 80 °C, H2 (10 bar), 40 mg Au40/TiO2 catalyst.

The plot of ln(Ct/C0) (where C0 represents the initial concentration of 4-nitroacetophenone and Ct is the concentration at reaction time t) versus time demonstrated that neat TiO2 is inactive, and that the hydrogenation of 4-nitroacetophenone with H2 on Au40/TiO2 follows first-order kinetics (Fig. 6c). The Haber model allows for two generally accepted reaction pathways for the catalytic reduction of aromatic nitro compounds: direct and condensation routes (Fig. 8a)56. As shown in Fig. 7 and Supplementary Fig. 19, only two products, 4-aminoacetophenone and hydroxylamine, were detected during the reaction. This indicates that Au40/TiO2 facilitated the hydrogenation of the nitro group in nitroaromatics through the direct route (Fig. 8a). Furthermore, the removal of the catalyst by hot filtration after 40 min of reaction resulted in the cessation of the catalytic conversion, indicating that the process was heterogeneous (Supplementary Fig. 20). The TOF of Au40/TiO2 was calculated to be 5829 h-1 based on the substrate conversion, which reached 23.7%. The TON of Au40/TiO2 was calculated to be 335569, which is much higher than the reported gold nanocluster-based catalysts for this reaction. The detailed process for TOF and TON calculation was described in the Supplementary Method section. As shown in Fig. 6d, after five reuse cycles, Au40/TiO2 demonstrated essentially undiminished catalytic activity and selectivity, indicating that Au40/TiO2 displays excellent catalytic cyclability and stability toward the chemoselective hydrogenation of 4-nitroacetophenone to 4-aminoacetophenone with H2 in water. To further demonstrate the stability of Au40 in the catalytic process, the catalyst and the solution after the reaction were characterized by transmission electron microscopy (TEM), XPS and gas chromatography-mass spectrometry (GC-MS), respectively. As shown in Supplementary Fig. 21, the particle size of Au40 did not change significantly both the catalyst preparation and subsequent catalytic reaction. A marginal increase was recorded for the ratio of Au(0)/Au(I), from 1.37 to 1.38, indicating negligible decomposition of Au40 and keeping it intact during the hydrogenation reaction (Supplementary Fig. 22). Furthermore, the GC-MS spectra of the reaction solution did not detect the ArC≡C ligand (Supplementary Fig. 23, 24), indicating that the ArC≡C ligand on the surface of Au40 was not released during the catalytic reaction. These results underscore the robustness of Au40/TiO2 catalyst.

Fig. 8. Chemoselective hydrogenation of nitroaromatics.

Fig. 8

Open in a new tab

a Proposed reaction pathways for hydrogenation of nitroaromatics to different products. b The mechanism of chemoselective hydrogenation of 4-nitroacetophenone to 4-aminoacetophenone.

A series of nanocluster-based catalysts was synthesized (the detail synthesis processes were described in the Supplementary Method section) and catalytically tested to gain a deeper understanding of the genesis of the observed activity and the pivotal issues inherent to the catalytic system. The catalytic results were presented in Table 1. The absence of a product when only TiO2 was present suggested that the catalytic site was derived from Au40 (Table 1, entries 1 and 2). The conversion of the reaction was examined for ArC≡CAu (AuArC≡C)49 supported on TiO2, resulting in a conversion of 23.8% (Table 1, entry 3). This indicates that intact Au40 contributed to the remarkable catalytic performance, rather than the decomposition-produced ArC≡CAu. [Au11(Ph3P)8Cl2]Cl (Au11)57 protected by Ph3P, [Au25(ArC≡C)18]Na (Au25RA)58, and Au25(PET)18 (Au25RS)59 (PET = phenylethanethiol), which have similar structures and are protected by alkynyl and thiols, respectively, were synthesized and examined (Supplementary Fig. 25a–c). The conversions of the Au40/TiO2, Au11/TiO2, Au25RA/TiO2, and Au25RS/TiO2 catalysts were 100%, 2.9%, 35.5%, and 0.0%, respectively (Table 1, entries 2, 4, 5, and 6). The alkynyl ligand is identical in both Au40 and Au25RA. However, the surface of Au40 contains the Au (ArC≡C)2, Au2(ArC≡C)3, and Au3(ArC≡C)4 staple motifs, whereas the surface of Au25RA displays solely the Au2(ArC≡C)3 staple motif. These results indicate that the Aux(ArC≡C)y (where x and y are the numbers of Au and alkynyl ligands, respectively) staple motifs on the surface of gold nanoclusters play a role in the observed activity, while thiol and phosphine ligands are inactive to the hydrogenation reaction. To investigate the effect of the metal core on catalytic performance, homoleptic Au36(PhC≡C)24 (Au36)50 and Au44(TBPA)28 (Au44)50 (TBPA = 4-tert-butylphenylacetylene), which have similar surface structures but differ in their metal core structures related to Au40, were synthesized and tested (Supplementary Fig. 25d, e). The conversion of 4-nitroacetophenone was 18.0% and 13.3%, respectively (Table 1, entries 7 and 8). The results suggest that the metal kernel is another important factor in determining the performance of nanocluster-based catalysts. Furthermore, the Au nanoparticles (AuNPs)60 (Supplementary Fig. 25f) synthesized in this study were inactive in the chemoselective hydrogenation of 4-nitroacetophenone under the same conditions (Table 1, entry 9). Therefore, we attribute the excellent catalytic performance of Au40 to the synergistic effect of its unique metal core structure and Aux(ArC≡C)y staple motifs on its surface.

Table 1.

Hydrogenation of 4-aminoacetophenone with different catalystsa

Entry Catalyst Conv. (%)* Select. B/E (%)*
1 TiO2 n.r.b --
2 Au40/TiO2 100 100.0/0.0
3 AuArC≡C/TiO2 23.8 88.9/11.1
4 Au11/TiO2 2.9 66.7/33.3
5 Au25RA/TiO2 35.5 63.6/36.4
6 Au25RS/TiO2 n.r. --
7 Au36/TiO2 18.0 97.0/3.0
8 Au44/TiO2 13.3 71.0/29.0
9 AuNPs/TiO2 n.r. --
Open in a new tab

aReaction conditions: 40 mg of Au catalyst [0.4 weight % (wt %)], 4-nitroacetophenone (8.26 mg, 0.05 mmol), 1.0 mL H2O, 80°C, H2 (10 bar), 1.5 h. *The conversion and selectivity for 4-aminoacetophenone (B) and hydroxylamine (E) were determined by 1H NMR. bn.r. = no reaction.

The process of catalytic hydrogenation of substituted nitroarenes by gold clusters remains uncertain. Previous studies by Corma et al.21,22 have demonstrated that low-coordination Au atoms are essential for the dissociation of H2 in Au-based catalysts. With respect to the remarkable performance of base-free hydrogenation catalyzed by Au40, the possible mechanism related to the H2 activation process was investigated by H2 temperature-programmed desorption (H2-TPD) and DFT calculations. We performed H2-TPD of TiO2 and Au40/TiO2. As shown in Supplementary Fig. 26, TiO2 exhibited a distinct H2 desorption peak at 160 °C, while Au40/TiO2 showed two distinct H2 desorption peaks at 110 and 160 °C. The peak at 160 °C corresponds to the H2 desorption peak of TiO2, and the peak at 110 °C corresponds to the H2 desorption peak of the H2-adduct of the Au40 nanocluster, indicating the formation of the H2-adduct of nanocluster during the hydrogenation process. DFT calculations revealed that the optimal H2 absorption site is Au(I) on the ArC≡C-Au-C≡CAr monomeric staple, with an H2 absorption energy (∆Eabs) of −1.85 kcal/mol, indicating moderate H2 absorption on the cluster surface (Fig. 9a). The possible H2 activation process was further investigated, and two types of stable intermediates were obtained. The first intermediate (IM1) describes that the H2 molecule transferred one H atom to the Au(I) site and the other H atom to the neighboring C site of the alkynyl ligand, followed by H − H bond breaking and an elongation of the Au(I) − C bond from 1.982 to 2.044 Å, leading to an endothermy of 11.76 kcal/mol (Fig. 9a). The second intermediate (IM2) exhibits a similar H − H bond breaking process but is accompanied by complete breakage of the Au(I) − C bond, resulting in the formation of a terminal coordinated H−Au(I)-C≡CR motif and a hydrogenated alkyne ligand (Fig. 9a). This process leads to endothermy of 18.45 kcal/mol, which is higher than that of IM1. Therefore, the activation of H2 on Au40 was more likely to occur via the IM1 pathway. The above analysis indicates that the surface alkynyl ligands could be crucial for H2 activation, potentially acting as a base that aids in cleaving the H − H bond across the cluster surface. In addition, to investigate the mechanism of H2 dissociation, we introduced the radical scavenger 2,6-di-tert-butyl-4-methylphenol (BHT) during the hydrogenation reaction. If H2 undergoes homolytic dissociation, active hydrogen is consumed by BHT, thereby significantly affecting the conversion. The addition of BHT did not cause significant changes in the conversion and selectivity (Supplementary Table 6), indicating that H2 dissociation is heterolytic rather than homolytic61,62. Furthermore, only hydroxylamine was detected as an intermediate during the reaction and DFT calculations indicated that H2 dissociation is an energy-increasing process. We proposed that the rate-determining step of the chemoselective hydrogenation of 4-nitroacetophenone includes H2 dissociation and activation of the N-O bond in hydroxylamine63,64.

Fig. 9. DFT predicted energies and geometric structure variations of the cluster catalytic system under BP86/TZP/ZORA level of theory.

Fig. 9

Open in a new tab

a The different hypothetical hydrogen activation process. bThe substrate molecules are perpendicular to the catalyst surface. c The substrate molecules are parallel to the catalyst surface. Color labels: yellow/orange = Au; bright green = F; gray = C; blue = N; red = O; pink= activited hydrogen atoms; incanus = H.

Previous studies by Koel et al.65 have shown that when acetone is absorbed on the Au(111) surface, the C = O double bond is oriented nearly parallel to the surface plane. In chemoselective hydrogenation of nitroaromatic by the Au NPs-based catalyst (Au/TiO2), the orientation of the substrate molecules relative to the catalyst surface varies depending on the type of the substrate employed. Corma et al.21,22 have demonstrated that when the nitro group interacts with Au sites, the substrate molecules adopt a perpendicular orientation relative to the catalyst surface. These findings indicate that in the chemoselective hydrogenation of 4-nitroacetophenone using the Au/TiO2 catalyst, the substrate molecules are parallel to the catalyst surface when the C = O group interacts with the Au sites. Conversely, when the nitro group interacts with Au sites, the substrate molecules adopt a perpendicular orientation relative to the catalyst surface. Following the activation and cleavage of H2 on the surface of Au40, 4-nitroacetophenone will approach the cluster surface to undergo the hydrogenation process. As shown in Fig. 9b and c, the 4-nitroacetophenone is more favorable to approach the activated hydrogen atoms in a perpendicular orientation rather than a parallel one. This preference is attributed to the steric hindrance of surface ArC≡C ligands in Au40, which results in a highly selective reduction of NO2 over the C = O group in 4-nitroacetophenone by Au4066. As demonstrated in Fig. 8b, it has been shown that Au40 catalysis the chemoselective hydrogenation of 4-nitroacetophenone to 4-aminoacetophenone, including the formation of nitroso and hydroxylamine. Since Au40/TiO2 has been demonstrated to possess excellent catalytic performance in the chemoselective hydrogenation of 4-nitroacetophenone to 4-aminoacetophenone, its general applicability to various nitroaromatics (Table 2, entries 1-7) was evaluated, and its catalytic activity and selectivity were analyzed by 1H NMR spectroscopy (Supplementary Fig. 27–32). The Au40/TiO2 catalyst exhibited satisfactory conversion (100%) and selectivity (100%) for a diverse range of nitroaromatics (Table 2, entries 1-7), indicating the general applicability of Au40/TiO2 for the chemoselective hydrogenation of nitroaromatics. Furthermore, the use of 4-methylacetophenone instead of 4-nitroacetophenone resulted in the absence of the hydrogenation product (Supplementary Fig. 33), indicating that Au40/TiO2 is inactive toward the hydrogenation of the C=O group (Table 2, entry 8). In addition, to determine the yield of 4-aminophenylacetone in the chemoselective hydrogenation of 4-nitroacetophenone, we introduced 1,3,5-trimethoxybenzene as the internal standard during the catalytic reaction. As demonstrated in Supplementary Fig. 34, the yield of 4-aminophenylacetone was 97%, based on the integral values of each proton of 4-aminophenylacetone and 1,3,5-trimethoxybenzene in 1H NMR. The enhanced catalytic performance of Au40 can be attributed to the synergistic effect of the surface ArC≡C-Au-C≡CAr staple motifs, the unique metal core, the steric hindrance effect of the alkynyl ligands, and the carrier effect.

Table 2.

The chemoselective hydrogenation of a variety of substrates with nitro or alkenyl groups was achieved using Au40/TiO2 as the catalysta

Entry Substrate Product Conv. (%) Select. (%)c
1 graphic file with name 41467_2025_63124_Taba_HTML.gif graphic file with name 41467_2025_63124_Tabb_HTML.gif 100 100
2 graphic file with name 41467_2025_63124_Tabc_HTML.gif graphic file with name 41467_2025_63124_Tabd_HTML.gif 100 100
3 graphic file with name 41467_2025_63124_Tabe_HTML.gif graphic file with name 41467_2025_63124_Tabf_HTML.gif 100 100
4 graphic file with name 41467_2025_63124_Tabg_HTML.gif graphic file with name 41467_2025_63124_Tabh_HTML.gif 100 100
5 graphic file with name 41467_2025_63124_Tabi_HTML.gif graphic file with name 41467_2025_63124_Tabj_HTML.gif 100 100
6 graphic file with name 41467_2025_63124_Tabk_HTML.gif graphic file with name 41467_2025_63124_Tabl_HTML.gif 100 100
7 graphic file with name 41467_2025_63124_Tabm_HTML.gif graphic file with name 41467_2025_63124_Tabn_HTML.gif 100 100
8 graphic file with name 41467_2025_63124_Tabo_HTML.gif graphic file with name 41467_2025_63124_Tabp_HTML.gif n.r.[b] --
Open in a new tab

aReaction conditions: 40 mg of Au40/TiO2 catalyst [0.4 weight % (wt %)], 0.05 mmol of substrates, 1.0 mL of H2O, 80 °C, H2 (10 bar), 2 hours. bn.r. = no reaction.cChemoselective hydrogenation of nitro group in substrates. *The conversion and selectivity for were determined by 1H NMR.

Discussion

In summary, an homoleptic [Au40(ArC≡C)22](Et4N)2 nanocluster was synthesized in 56% yield via direct reduction of gold-alkynyl precursors. The molecular structure of Au40 differs significantly from that of previously reported Au nanocluster with 40 Au atoms. The superatomic Au26 core endows Au40 a 20-electron superatom with the configurations of (1S)2(1P)6(1D)10(1F)2. The Au40/TiO2 exhibits 100% conversion and selectivity, and 97% yield toward the chemoselective hydrogenation of carbonyl-substituted nitroaromatics to aromatic amines in water, free of base, under mild conditions (80 °C) using H2 (10 bar). The TON and TOF for the Au40/TiO2 catalyst are up to 335569 and 5829 h-1, respectively. Au40/TiO2 demonstrates general applicability to various nitroaromatics. The excellent catalytic performance of Au40/TiO2 can be attributed to the synergistic effect of the unsaturated Au(I) sites derived from the ArC≡C-Au-C≡CAr staple motif, the improved durability, the steric hindrance effects caused by ArC≡C ligands in Au40, and carrier effect. This study presents an example of a surface structure designed to modulate the catalytic performance. It offers valuable insights for the rational design and synthesis of catalysts with excellent catalytic performance for the selective hydrogenation. Compared to Au NPs-based catalysts, Au40/TiO2 not only demonstrates enhanced activity and selectivity but also operates under mild reaction conditions (80 °C, 10 bar H2), without additive and with environmentally friendly solvents (water vs toluene), positioning the Au nanocluster as a potential alternative catalyst for practical applications.

Methods

Synthesis of [Au40(ArC≡C)22](Et4N)2

ArC≡CAu (43.4 mg, 0.1 mmol) was dispersed in 5.0 mL of a mixed solvent of dichloromethane/methanol (v/v = 3/1). After stirring for 5 min, 100 μL of methanol solution of MeONa (0.5 M) was added, causing the yellow suspension to turn brown. Subsequently, within 1 min, 1.0 mL of a freshly prepared ethanol solution of NaBH4 (0.63 mg/mL) was added dropwise to the suspension under stirring. The solution changed from brown to brown-green. The reaction continued in the dark for 24 h at 15 °C. The mixture was then rotary evaporated to dryness, resulting in a dark solid. The solid was washed with n-hexane (3.0 mL × 2), and extracted with 3.0 mL of dichloromethane. After centrifuging at 10280 × g for 3 min, the supernatant was collected and concentrated for separation by preparative thin-layer chromatography (PTLC) below 20 °C. The solution of nanoclusters was pipetted onto the PTLC plate and separated in a developing tank using a developing solvent of dichloromethane/methanol (v/v = 7/1) (Supplementary Fig. 1). The second band, corresponding to the title nanocluster, was cut off and extracted with methanol. Centrifugation and rotary evaporation of the supernatant gives a black solid of the title cluster of 18.3 mg (56 % yield based on the molecular weight with the Na+ form and the mols of Au). In order to obtain high-quality single crystals, the obtained product separated by PTLC was dissolved in 2.0 mL of dichloromethane, and an excess of tetraethylammonium tetrafluoroborate was added, and stirred for 30 min in the dark. The mixture was centrifuged at 10280 × g for 2 min and filtered, and the supernatant was subject to diffusion of n-hexane to obtain blocklike black crystals of [Au40(ArC≡C)22](Et4N)2 after 2 weeks at 4 °C.

Preparation of Au40/TiO2 and evaluation of catalytic performance

Typically, 1.0 mg of Au40 was dissolved in 4.0 mL of dichloromethane, and 250 mg of TiO2 was added. After stirring for 2 h at room temperature in air. The Au40/TiO2 catalyst was collected by centrifugation at 10280 × g for 2 min and dried in vacuum. The catalyst was then annealed at 130 °C for 1 h in a vacuum tube to get Au40/TiO2 heterogeneous catalyst. For chemoselective hydrogenation of 4-nitroacetophenone, the reaction conditions were as follows: 40 mg of Au40 (0.4 wt%)/TiO2 catalyst, 4-nitroacetophenone (8.26 mg, 0.05 mmol), 1.0 mL of H2O, 80 °C, H2 (10 bar), and 1.5 h. The conversion and selectivity were determined by 1H NMR. The detailed procedures were described in the Supplementary Method section.

Supplementary information

Supplementary Information (4MB, pdf)
Transparent Peer Review file (688.1KB, pdf)

Source data

Source data (39.2MB, zip)

Acknowledgements

This work was supported by the Natural Science Foundation of China (22201188 (X.-K.W.), 92361301 (Q.-M.W.)), the National Key R&D Program of China 2022YFA1503900 (Q.-M.W.), the Startup Funds of Sichuan University (YJ202196 (X.-K.W.)), and the Fundamental Research Funds for the Central Universities. We thank Meng Yang, Feng Yang, and Dong-Yan Deng from the Comprehensive Training Platform of the Specialized Laboratory in the College of Chemistry at Sichuan University for SCXRD, TEM, and NMR testing, respectively. We thank Prof. Fan Zhang and Prof. Zhen-Feng, Cai for the help of H2-TPD and Raman testing, respectively. We also thank Peng-Chi, Deng from the Analytical & Testing Center at Sichuan University for NMR testing at low temperature.

Author contributions

X.-K.W. Conceived, supervised, and guided the project. X.-X.L. carried out the synthesis, crystallization of the clusters, characterization, catalysis test, and analyzed the data. T.-T. L., J.L., and J.-J.L. conducted the characterization. J. W conducted the density functional theory calculations and wrote the manuscript. X.-K.W. and Q.-M.W. discussed and analyzed the data, wrote and revised the manuscript. All authors discussed the results and contributed to the manuscript.

Peer review

Peer review information

Nature Communications thanks Rajendra S. Dhayal, Katsuhiro Isozaki, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The data that support the findings of this study are available from the corresponding author upon request. The X-ray crystallographic data for the structures reported in this article have been deposited at the Cambridge Crystallographic Data Centre (CCDC) under deposition numbers CCDC 2394293.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Xiao-Xiao Lai, Jianyu Wei.

Contributor Information

Xian-Kai Wan, Email: xkwan@scu.edu.cn.

Quan-Ming Wang, Email: qmwang@tsinghua.edu.cn.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-63124-8.

References

  • 1.Zhang, L., Zhou, M., Wang, A. & Zhang, T. Selective hydrogenation over supported metal catalysts: from nanoparticles to single atoms. Chem. Rev.120, 683–733 (2020). [DOI] [PubMed] [Google Scholar]
  • 2.Lan, X. & Wang, T. Highly selective catalysts for the hydrogenation of unsaturated aldehydes: a review. ACS Catal.10, 2764–2790 (2020). [Google Scholar]
  • 3.Vilé, G., Albani, D., Almora-Barrios, N., López, N. & Pérez-Ramírez, J. Advances in the design of nanostructured catalysts for selective hydrogenation. ChemCatChem8, 21–33 (2016). [Google Scholar]
  • 4.Zhao, M. et al. Metal–organic frameworks as selectivity regulators for hydrogenation reactions. Nature539, 76–80 (2016). [DOI] [PubMed] [Google Scholar]
  • 5.Corma, A. & Serna, P. Chemoselective hydrogenation of nitro compounds with supported gold catalysts. Science313, 332–334 (2006). [DOI] [PubMed] [Google Scholar]
  • 6.Grirrane, A., Corma, A. & García, H. Gold-catalyzed synthesis of aromatic azo compounds from anilines and nitroaromatics. Science322, 1661–1664 (2008). [DOI] [PubMed] [Google Scholar]
  • 7.Wang, L. et al. Single-site catalyst promoters accelerate metal-catalyzed nitroarene hydrogenation. Nat. Commun.9, 1362 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Chen, G. et al. Interfacial electronic effects control the reaction selectivity of platinum catalysts. Nat. Mater.15, 564–569 (2016). [DOI] [PubMed] [Google Scholar]
  • 9.Furukawa, S., Takahashi, K. & Komatsu, T. Well-structured bimetallic surface capable of molecular recognition for chemoselective nitroarene hydrogenation. Chem. Sci.7, 4476–4484 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Serna, P. & Corma, A. Transforming nano metal nonselective particulates into chemoselective catalysts for hydrogenation of substituted nitrobenzenes. ACS Catal.5, 7114–7121 (2015). [Google Scholar]
  • 11.Wu, Q. et al. Full selectivity control over the catalytic hydrogenation of nitroaromatics into six products. Angew. Chem. Int. Ed.63, e202408731 (2024). [DOI] [PubMed] [Google Scholar]
  • 12.Wang, N. et al. Impregnating subnanometer metallic nanocatalysts into self-pillared zeolite nanosheets. J. Am. Chem. Soc.143, 6905–6914 (2021). [DOI] [PubMed] [Google Scholar]
  • 13.Guan, Z.-J. et al. Ligand engineering toward the trade-off between stability and activity in cluster catalysis. Angew. Chem. Int. Ed.61, e202116965 (2022). [DOI] [PubMed] [Google Scholar]
  • 14.You, Q. et al. Bottom-up construction of metal–organic framework loricae on metal nanoclusters with consecutive single nonmetal atom tuning for tailored catalysis. J. Am. Chem. Soc.146, 9026–9035 (2024). [DOI] [PubMed] [Google Scholar]
  • 15.Zhang, S. et al. Enhanced catalysis of Au/TiO2 for transfer hydrogenation of unsaturated nitro compounds by surface engineering. Appl. Surf. Sci.654, 159503 (2024). [Google Scholar]
  • 16.Yao, C. et al. Mechanism and kinetics guided design of catalysts for functionalized nitroarenes hydrogenation. ChemCatChem16, e202400027 (2024). [Google Scholar]
  • 17.Orlandi, M., Brenna, D., Harms, R., Jost, S. & Benaglia, M. Recent developments in the reduction of aromatic and aliphatic nitro compounds to amines. Org. Process Res. Dev.22, 430–445 (2018). [Google Scholar]
  • 18.Shimizu, K. -i et al. Chemoselective hydrogenation of nitroaromatics by supported gold catalysts: mechanistic reasons of size- and support-dependent activity and selectivity. J. Phys. Chem. C.113, 17803–17810 (2009). [Google Scholar]
  • 19.Corma, A., Serna, P., Concepción, P. & Calvino, J. J. Transforming nonselective into chemoselective metal catalysts for the hydrogenation of substituted nitroaromatics. J. Am. Chem. Soc.130, 8748–8753 (2008). [DOI] [PubMed] [Google Scholar]
  • 20.Wang, L. et al. Activity and selectivity in nitroarene hydrogenation over Au nanoparticles on the edge/corner of anatase. ACS Catal.6, 4110–4116 (2016). [Google Scholar]
  • 21.Boronat, M. et al. A molecular mechanism for the chemoselective hydrogenation of substituted nitroaromatics with nanoparticles of gold on TiO2 Catalysts: A cooperative effect between gold and the support. J. Am. Chem. Soc.129, 16230–16237 (2007). [DOI] [PubMed] [Google Scholar]
  • 22.Corma, A., Boronat, M., González, S. & Illas, F. On the activation of molecular hydrogen by gold: a theoretical approximation to the nature of potential active sites. Chem. Commun.32, 3371–3373 (2007). [DOI] [PubMed] [Google Scholar]
  • 23.Cai, X., Wang, H., Tian, Y., Ding, W. & Zhu, Y. Catalytic Application of atomically precise metal clusters in selective hydrogenation processes. ACS Catal.14, 11918–11930 (2024). [Google Scholar]
  • 24.Jing, W. et al. Surface and interface coordination chemistry learned from model heterogeneous metal nanocatalysts: from atomically dispersed catalysts to atomically precise clusters. Chem. Rev.123, 5948–6002 (2023). [DOI] [PubMed] [Google Scholar]
  • 25.Du, Y., Sheng, H., Astruc, D. & Zhu, M. Atomically precise noble metal nanoclusters as efficient catalysts: a bridge between structure and properties. Chem. Rev.120, 526–622 (2020). [DOI] [PubMed] [Google Scholar]
  • 26.Jin, R., Li, G., Sharma, S., Li, Y. & Du, X. Toward active-site tailoring in heterogeneous catalysis by atomically precise metal nanoclusters with crystallographic structures. Chem. Rev.121, 567–648 (2021). [DOI] [PubMed] [Google Scholar]
  • 27.Guan, Z.-J., Li, J.-J., Hu, F. & Wang, Q.-M. Structural engineering toward gold nanocluster catalysis. Angew. Chem. Int. Ed.61, e202209725 (2022). [DOI] [PubMed] [Google Scholar]
  • 28.Zhu, Y., Qian, H., Drake, B. A. & Jin, R. Atomically Precise Au25(SR)18 nanoparticles as catalysts for the selective hydrogenation of α,β-unsaturated ketones and aldehydes. Angew. Chem. Int. Ed.49, 1295–1298 (2010). [DOI] [PubMed] [Google Scholar]
  • 29.Li, G. & Jin, R. Gold nanocluster-catalyzed semihydrogenation: a unique activation pathway for terminal alkynes. J. Am. Chem. Soc.136, 11347–11354 (2014). [DOI] [PubMed] [Google Scholar]
  • 30.Wan, X.-K., Wang, J.-Q., Nan, Z.-A. & Wang, Q.-M. Ligand effects in catalysis by atomically precise gold nanoclusters. Sci. Adv.3, e1701823 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Li, G., Zeng, C. & Jin, R. Thermally robust Au99(SPh)42 nanoclusters for chemoselective hydrogenation of nitrobenzaldehyde derivatives in water. J. Am. Chem. Soc.136, 3673–3679 (2014). [DOI] [PubMed] [Google Scholar]
  • 32.Liu, C., Abroshan, H., Yan, C., Li, G. & Haruta, M. One-Pot Synthesis of Au11(PPh2Py)7Br3for the highly chemoselective hydrogenation of nitrobenzaldehyde. ACS Catal.6, 92–99 (2016). [Google Scholar]
  • 33.Tan, Y. et al. ZnAl-Hydrotalcite-Supported Au25 nanoclusters as precatalysts for chemoselective hydrogenation of 3-Nitrostyrene. Angew. Chem. Int. Ed.56, 2709–2713 (2017). [DOI] [PubMed] [Google Scholar]
  • 34.Liu, X., Cai, X. & Zhu, Y. Catalysis synergism by atomically precise bimetallic nanoclusters doped with heteroatoms. Acc. Chem. Res.56, 1528–1538 (2023). [DOI] [PubMed] [Google Scholar]
  • 35.Shen, H. et al. N-Heterocyclic Carbene-stabilized gold nanoclusters with organometallic motifs for promoting catalysis. J. Am. Chem. Soc.144, 10844–10853 (2022). [DOI] [PubMed] [Google Scholar]
  • 36.Alamer, B. et al. Planar core and macrocyclic shell stabilized atomically precise copper nanocluster catalyst for efficient hydroboration of C–C multiple bond. J. Am. Chem. Soc.146, 16295–16305 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Biswas, S. et al. Luminescent Hydride-Free [Cu7(SC5H9)7(PPh3)3] Nanocluster: Facilitating highly selective C–C bond formation. J. Am. Chem. Soc.146, 20937–20944 (2024). [DOI] [PubMed] [Google Scholar]
  • 38.Jia, T. et al. Atomically precise ultrasmall copper cluster for room-temperature highly regioselective dehydrogenative coupling. Nat. Commun.14, 6877 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Zhang, Y. et al. Preorganized Nitrogen Sites for Au11 Amidation: A generalizable strategy toward precision functionalization of metal nanoclusters. J. Am. Chem. Soc.145, 12164–12172 (2023). [DOI] [PubMed] [Google Scholar]
  • 40.Zhao, Y. et al. A dual purpose strategy to endow gold nanoclusters with both catalysis activity and water solubility. J. Am. Chem. Soc.142, 973–977 (2019). [DOI] [PubMed] [Google Scholar]
  • 41.Zeng, C. et al. Gold tetrahedra coil up: Kekule-like and double helical superstructures. Sci. Adv.1, e1500425 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Okinaka, Y. An electrochemical study of electroless gold-deposition reaction. J. Electro chem. Soc.120, 739 (1973). [Google Scholar]
  • 43.Maity, P., Takano, S., Yamazoe, S., Wakabayashi, T. & Tsukuda, T. Binding Motif of terminal alkynes on gold clusters. J. Am. Chem. Soc.135, 9450–9457 (2013). [DOI] [PubMed] [Google Scholar]
  • 44.Takano, S., Hirai, H., Muramatsu, S. & Tsukuda, T. Hydride-doped Gold Superatom (Au9H)2+: synthesis, structure and transformation. J. Am. Chem. Soc.140, 8380–8383 (2018). [DOI] [PubMed] [Google Scholar]
  • 45.Zeng, C., Liu, C., Chen, Y., Rosi, N. L. & Jin, R. Gold–Thiolate Ring as a Protecting Motif in the Au20(SR)16 Nanocluster and Implications. J. Am. Chem. Soc.136, 11922–11925 (2014). [DOI] [PubMed] [Google Scholar]
  • 46.Malola, S. et al. Au40(SR)24 Cluster as a Chiral Dimer of 8-Electron superatoms: structure and optical properties. J. Am. Chem. Soc.134, 19560–19563 (2012). [DOI] [PubMed] [Google Scholar]
  • 47.Jiang, D. -e The expanding universe of thiolated gold nanoclusters and beyond. Nanoscale5, 7149–7160 (2013). [DOI] [PubMed] [Google Scholar]
  • 48.Wang, T., Zhang, W. H., Yuan, S. F., Guan, Z. J. & Wang, Q. M. An alkynyl-protected Au40 nanocluster featuring PhC≡C-Au-P^P motifs. Chem. Commun.54, 10367–10370 (2018). [DOI] [PubMed] [Google Scholar]
  • 49.Wan, X.-K., Yuan, S.-F., Tang, Q., Jiang, D.-E. & Wang, Q.-M. Alkynyl-protected Au23 Nanocluster: A 12-Electron System. Angew. Chem. Int. Ed.54, 5977–5980 (2015). [DOI] [PubMed] [Google Scholar]
  • 50.Wan, X.-K., Guan, Z.-J. & Wang, Q.-M. Homoleptic Alkynyl-protected gold nanoclusters: Au44(PhC≡C)28 and Au36(PhC≡C)24. Angew. Chem. Int. Ed.56, 11494–11497 (2017). [DOI] [PubMed] [Google Scholar]
  • 51.Han, X.-S. et al. Total structure determination of alkynyl-protected gold nanocluster Au22(tBuC≡C)18 and the Thermochromic luminescence. Angew. Chem. Int. Ed.59, 2309–2312 (2020). [DOI] [PubMed] [Google Scholar]
  • 52.Walter, M. et al. A unified view of ligand-protected gold clusters as superatom complexes. Proc. Natl. Acad. Sci.105, 9157–9162 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Gam, F., Wei, J., Kahlal, S., Saillard, J.-Y. & Halet, J.-F. in 50th Anniversary of Electron Counting Paradigms for Polyhedral Molecules: Bonding in Clusters, Intermetallics and Intermetalloids Vol. 188 (ed D. Michael P Mingos) 69-102 (Springer International Publishing, 2021).
  • 54.Kang, S.-Y., Nan, Z.-A. & Wang, Q.-M. Superatomic orbital splitting in coinage metal nanoclusters. J. Phys. Chem. Lett.13, 291–295 (2022). [DOI] [PubMed] [Google Scholar]
  • 55.Rüger, R., van Lenthe, E., Heine, T. & Visscher, L. Tight-binding approximations to time-dependent density functional theory — A fast approach for the calculation of electronically excited states. J. Chem. Phys.144, 184103 (2016). [DOI] [PubMed] [Google Scholar]
  • 56.Corma, A., Concepción, P. & Serna, P. A Different reaction pathway for the reduction of aromatic nitro compounds on gold catalysts. Angew. Chem. Int. Ed.46, 7266–7269 (2007). [DOI] [PubMed] [Google Scholar]
  • 57.McKenzie, L. C., Zaikova, T. O. & Hutchison, J. E. Structurally similar triphenylphosphine-stabilized undecagolds, Au11(PPh3)7Cl3 and [Au11(PPh3)8Cl2]Cl, exhibit distinct ligand exchange pathways with glutathione. J. Am. Chem. Soc.136, 13426–13435 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Li, J.-J., Guan, Z.-J., Lei, Z., Hu, F. & Wang, Q.-M. Same magic number but different arrangement: alkynyl-protected Au25 with D3 symmetry. Angew. Chem. Int. Ed.58, 1083–1087 (2019). [DOI] [PubMed] [Google Scholar]
  • 59.Zhu, M., Aikens, C. M., Hollander, F. J., Schatz, G. C. & Jin, R. Correlating the crystal structure of a thiol-protected Au25 cluster and optical properties. J. Am. Chem. Soc.130, 5883–5885 (2008). [DOI] [PubMed] [Google Scholar]
  • 60.Yu, S., Wilson, A. J., Heo, J. & Jain, P. K. Plasmonic control of multi-electron transfer and C–C coupling in visible-light-driven CO2 reduction on Au nanoparticles. Nano Lett.18, 2189–2194 (2018). [DOI] [PubMed] [Google Scholar]
  • 61.Serna, P., Concepción, P. & Corma, A. Design of highly active and chemoselective bimetallic gold–platinum hydrogenation catalysts through kinetic and isotopic studies. J. Catal.265, 19–25 (2009). [Google Scholar]
  • 62.Cano, I. et al. Air-stable gold nanoparticles ligated by secondary phosphine oxides as catalyst for the chemoselective hydrogenation of substituted aldehydes: a remarkable ligand effect. J. Am. Chem. Soc.137, 7718–7727 (2015). [DOI] [PubMed] [Google Scholar]
  • 63.Shao, Z.-J., Zhang, L., Liu, H., Cao, X.-M. & Hu, P. Enhanced Interfacial H2 activation for nitrostyrene catalytic hydrogenation over Rutile Titania-Supported Gold by Coadsorption: A First-Principles Microkinetic Simulation Study. ACS Catal.9, 11288–11301 (2019). [Google Scholar]
  • 64.Wang, H., Shi, F., Pu, M. & Lei, M. Theoretical study on nitrobenzene hydrogenation by N-doped carbon-supported late transition metal single-atom catalysts. ACS Catal.12, 11518–11529 (2022). [Google Scholar]
  • 65.Syomin, D. & Koel, B. E. IRAS studies of the orientation of acetone molecules in monolayer and multilayer films on Au(111) surfaces. Surf. Sci.498, 53–60 (2002). [Google Scholar]
  • 66.Wan, X.-K. et al. Tailoring atomically precise gold nanoclusters for boosting selective hydrogenation of Nitrostyrene with H2. ACS Nano19, 11371–11380 (2025). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information (4MB, pdf)
Transparent Peer Review file (688.1KB, pdf)
Source data (39.2MB, zip)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon request. The X-ray crystallographic data for the structures reported in this article have been deposited at the Cambridge Crystallographic Data Centre (CCDC) under deposition numbers CCDC 2394293.

Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES