Introduction

Atomically precise Au nanoclusters (NCs) serving as effective catalysts have been widely explored for a variety of organic reactions1, such as the oxidation2,3,4,5, hydrogenation6,7,8, C-C coupling reactions9 and even C-H activation10. Among them, as homogeneous or heterogeneous catalysts, Au25 NCs have attracted particular attention due to their precise number of atoms and stable structure, which is important to understand the catalytic mechanism at the atomic and molecular level4,11,12. To our knowledge, there are three stable charge states (−1, 0, +1) for Au25 NCs13. It is notable that the charge states (−1, 0) have been further confirmed by the corresponding single-crystal structures14,15. Besides, the three different states are reciprocally transformed by oxidation and reduction16. Our previous results also indicated the single electron of Au25 was capable to be transferred to organic molecule17. Notably, Au25z NCs employed as ET mediators, which provide electrons or deliver electrons, may offer novel insights into the mechanism of Au-catalyzed chemical reactions and the correlation between structure and catalytic property. The electron transfer (ET) reactions of the Au25 NCs have been discussed by nuclear magnetic resonance18, electrochemistry and thermochemical calculations19. However, to date, there are no reports of Au25 NCs functioning as ET catalysts in solution-phase organic reactions. Herein, we first demonstrate the ET catalysis of Au25 in intramolecular cascade reaction of 2-nitrobenzonitrile, from which 2-amniobenzamide (a precursor of significant pharmaceuticals20,21) is produced with high yield.

Results

The synthesis protocol of Au25(SR)18TOA+ (abbreviation as Au25, SR = SCH2CH2Ph) was referred to the previous literature22. The ultraviolet-visible (UV-vis) spectrum of the obtained NCs is shown in Supplementary Fig. S1. The composition of the as-prepared NCs is further verified by matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF-MS, see Supplementary Fig. S2). In this work, the intramolecular cascade reaction of 2-nitrobenzonitrile (abbreviated as 1a) for 2-amniobenzamide (2a) was investigated using Au25 as ET catalysts (See Fig. 1).

Figure 1
figure 1

Au25z catalyzed intramolecular cascade reaction of 2-nitrobenzonitrile.

Cyclic voltammograms of 1a and Au25 in 0.1 M Bu4NClO4/CH2Cl2 solution are shown in Fig. 2 and Supplementary Table S1. It is obvious that half-wave potential of 1a is positive than that of Au250/−1, suggesting 1a is capable to grab the electron from Au25. In the ET mediated system, Au25 is susceptible to the substrate (1a) and easily oxidized into Au250. Meanwhile, N free radicals can be formed after 1a getting electrons from Au25 clusters, proceeding with an intramolecular cascade reaction. Additionally, while treated with NaBH4, Au250 could be reduced back to Au25, indicating Au25 functioned as an electron bridge and excess NaBH4 provided continual electron source. The electron transfer reactions are presented as follows (Fig. 3), which are indeed established by our investigation.

Figure 2
figure 2

The cyclic voltammograms of Au25TOA+ and 2-nitrobenzonitrile in 0.1 M Bu4NClO4/CH2Cl2 at 298 K (degassed by N2, scan speed: 0.05 V*s−1).

(a), (b) The reduction peaks of Au25 (0/−1) and substrate, respectively; (c), (d) the oxidation peaks of Au25 (0/−1) and substrate, respectively.

Figure 3
figure 3

The electron transfer process.

To achieve the essential understanding, UV-vis and electron spin-resonance (ESR) spectra were initially employed to investigate the electron transfer process. The distinguished UV-vis absorption characters of Au25z (z = −1, 0) can indeed serve as spectroscopic fingerprints and be capable for ready determination of the negative and neutral states15. When increased amounts of 1a were added to Au25 in tetrahydrofuran, the UV-vis spectra showed obvious changes, in contrast to pure Au25. As shown in Fig. 4a, the shoulder peak at 800 nm, the fine spectral peak of Au25, successively declined and eventually disappeared. Simultaneously, a new shoulder dominated at ~620 nm emerged and gradually enhanced, indicating the progressive formation of Au250 from Au25. On the other hand, after excess addition of sodium borohydride, the 800 nm absorption shoulder arised again (Fig. 4b), suggesting the recovery of Au25. The above spectroscopic evidence confirmed that Au25z was an electron mediator and it kept delivering electrons from NaBH4 to substrate. After a circle, a molecule of 1a gained an electron from Au25 to form a radical anion and further converted into 2-amniobenzamide. More significantly, we achieved ESR data to further confirm the occurrence of the initial ET process. Pure Au25 shows no signal while the mixture of Au25 and 1a displays a triplet peaks, suggesting the formation of N radical (Fig. 4c). In addition, more complicated radical was formed with the addition of NaBH4 (Supplementary Fig. S3), indicating the N radical proceeded with further reaction.

Figure 4
figure 4

UV-vis and ESR monitoring spectra of Au25 mixed with 1a and/or NaBH4 (single crystals of Au25 dissolved in tetrahydrofuran).

(a): Only 1a was added into Au25 (The spectra gradually changed on addition of 1a as the arrows showed); (b): blue line is UV-vis spectrum of pure Au25, red line is UV-vis spectrum of mixed Au25 and 1a, black line displays the UV-vis spectrum of mixed Au25 and 1a after NaBH4 added; (c): ESR signals of pure Au25 (black line) and the reaction mixture of Au25 and 1a (red line) (0.8 mg 1a dissolved in 0.55 mL THF added to 0.55 mL THF of 1 mg Au25 clusters).

To get a deep insight into the catalytic reaction, a range of control experiments were carried out, the conversion was determined by GC. The catalytic reduction was operated in THF at 20°C for 2 h. Au25 could achieve high conversion (87%) and selectivity (96%) (Entry 10, table 1), which was considered as highly efficient at ambient condition. For comparison, a range of functionalities like nitro, amino, nitrile groups were introduced to assess the electronic effects. Comparing with 1a, solo-functionalized nitrobenzene was reduced to azoxybenzene in relatively low yield in the presence of Au25 (Entry 1, table 1). It is worth mentioning that no reductive or additive product of cyanobenzene was detected even through the prolonged reaction time, neither did the substituted cyanobenzene with an electron-donating group or an electron-withdrawing group (Entry 3, 4, 5, table 1), indicating the reaction of 1a was not originated from electron effects. Moreover, para-substituted 4-nitrobenzonitrile was selectively reduced (Entry 2, table 1) and partially reduced products of 1a gave no conversion (Entry 6, 7, table 1). Hence, only when nitro group and nitrile group were placed at ortho positions could both of them be reduced in high yields. When the H2O in THF was replaced by D2O, two more Daltons (138 Da vs 136 Da, see Supplementary Fig. S4, S5) of the molecule mass for the products were detected, implying water involved in the transformation. In contrast, few products were achieved without any catalysts (Entry 8, table 1). When Au NPs were employed as catalysts in the system, a few products could be achieved either. The occurrence may be ascribed to the charge transfer from the surface of Au NPs. However, this kind of effect is too weak and Au NPs can't deliver electrons efficiently from NaBH4 to substrates. As a result, low yield is doomed (Entry 9, table 1, TEM image of Au NPs was presented in Supplementary Fig. S6). Limited to their uncertain and complicated structure, it is difficult to get deep in understanding the mechanism of Au NPs catalyzed reactions. Notably, Au25 is a great candidate to make up the deficiency due to its well-defined structure and properties, for which is easy to shift between negative and neutral states.

Table 1 Au25(SR)18 catalyzed reduction of different substratesa

Given nitrile group and nitro group coexist in 1a and they were both reduced, three possible paths were postulated to further probe the reaction route (see Supplementary Fig. S7): (1) 2-aminobenzonitrile (4a; Supplementary Fig. S7) acted as intermediate; (2) a five-member ring intermediate was formed; (3) nitrile group was reduced at the beginning (5a; Supplementary Fig. S7). The transforming process was supposed to follow one of the three paths. Given the whole reaction was conducted in strong reducing environment (excessive NaBH4 present), oxidative addition to form carbonyl bond from 4a to 2a, or from 1a to 5a was clearly impossible, which meant both path 1 and 3 were implausible. Hence, path 2 was considered to be the plausible route. The two-step reaction proceeded as an intramolecular cascade reaction. As nitro group was sensitive to the catalyst, after the nitrile group at ortho position was reduced, they were supposed to form a five-member heteroring (3a; Supplementary Fig. S7). Similar theories were previously reported by Baker's yeast catalysis23 and transition metal catalysis24,25.

As showed in Fig. 5, the possible reaction mechanism was speculated to disclose the process. The single electron was first transferred from Au25 to 1a with the formation of N radical. Importantly, this electron transfer step and the formation of N radical were supported by ESR spectra. Then, the radical could be readily transformed into the intermediate of 2-(hydroxyamino) benzonitrile. Before the final product (2a) was eventually formed, a five-member heterocyclic intermediate emerged to grab another electron from Au25. Meanwhile, one molecule of water was generated after a circle which could be involved in the intramolecular cascade reaction. Notably, solo-functionalized and para-substituted substrates were not capable to attain the active intermediate five-member ring. When reacted with NaBH4, Au250 could be reduced to Au25, suggesting Au25 served as the ET mediator and NaBH4 provided consequent electron source. UV-vis spectra also confirmed the circle of catalytic species. After the reaction reached a chemical equilibrium, the mixture was centrifugated to test the supernate for UV-vis spectrum and MALDI-TOF-MS. Typical absorption bands centered at 670 nm and 450 nm were observed (See Supplementary Fig. S8). Besides, molecular ion peak of 7391 (See Supplementary Fig. S9) further verified that the Au25 remained after catalyzing the intramolecular cascade reaction, which was consistent with our hypothesis.

Figure 5
figure 5

Proposed mechanism of Au25z catalyzed intramolecular cascade reaction.

Discussion

In summary, we have demonstrated an intramolecular cascade reaction catalyzed by Au25z which served as electron mediators. It was worth noting that this reaction achieved high conversion (87%) and selectivity (96%) with Au25z NCs as ET catalysts at ambient conditions. Furthermore, the catalytic mechanism was systematically investigated. Au25z shifted between Au25 and Au250 in order to ensure the intermolecular electron flow. Excess NaBH4 provided hydrogen source and kept Au25 at negative state. More importantly, the catalytic species Au25z was proved to be still existed after the reaction finished, which meant it could be recovered and recycled. It is believed that our research may open up wide opportunities for ET catalysis of Au25 in extensive organic reactions.

Methods

General experimental detail

All the chemicals and reagents are commercially available. TLC plates (Merck Silica Gel 60 F254) were used for analytical TLC and Merck Kieselgel 200–300 was used for preparative column chromatography. 1H NMR spectra were acquired on a Brucker AM 400 operating at 400 MHz. 13C NMR spectra were acquired on a Brucker AM 400 operating at 100 MHz. Conversion and selectivity was measured with Shimadzu GC 2010 plus. UV tests were conducted with a Hewlett-Packard (HP) 8453 diode array spectrophotometer. TEM images were obtained by JEM 2100. Room temperature electron spin resonance (ESR) was conducted on JEOL JES-FA200 EPR spectrometer (300 K, 9048 MHz, X-band). Cyclic voltammetric measurements (CV) were performed on an electrochemical workstation (CHI 660D) at a scan rate of 0.05 V·s-1 under ambient conditions.

Preparation of Au25(SCH2CH2Ph)18TOA+

Briefly, HAuCl4·3H2O dissolved in deionized water along with TOAB dissolved in toluene, were combined in a 25 mL tri-neck round bottom flask. After stirring for ~15 min, the aqueous phase was excluded by syringe. The toluene solution was protected with N2 and cooled down in an ice bath. Hereafter, PhCH2CH2SH was added with a very low speed. 1 h later, the stirring speed was adjusted to fast stirring; an aqueous solution of NaBH4 was quickly added all at once. The synthesis proceeded overnight. The final product was purified by precipitation with acetonitrile three times.

Typical procedure of reduction reaction

0.2 g 2-nitrobenzonitrile (1.35 mmol), 0.102 g NaBH4 (2.7 mmol) and 2 mg catalyst was added into 10 mL of THF in round-bottom flask. The mixture was stirred for 2 h at room temperature before being evaporated to dryness under vacuum. The crude product was dissolved in 8 mL methylene dichloride and washed by 3 mL deionized water. Finally the pure product was obtained by column chromatography.