Small Gold Clusters Formed in Solution Give Reaction Turnover Numbers of 107 at Room Temperature

See allHide authors and affiliations

Science  14 Dec 2012:
Vol. 338, Issue 6113, pp. 1452-1455
DOI: 10.1126/science.1227813


Very small gold clusters (3 to 10 atoms) formed from conventional gold salts and complexes can catalyze various organic reactions at room temperature, even when present at concentrations of parts per billion. Absorption and emission ultraviolet-visible spectroscopy and matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry revealed that, for example, the ester-assisted hydration of alkynes began only when clusters of three to five gold atoms were formed. The turnover numbers and turnover frequencies associated with these catalyzed reactions can be as high as 107 and 105 per hour, respectively.

The current interest in gold catalysis stems from the discovery 25 years ago that nanosized gold can catalyze the hydrochlorination of acetylene and the oxidation of carbon monoxide at room temperature (1, 2). Gold nanoparticles (AuNPs) can activate O2, H2, sp3, sp2, and sp. C-X bonds (X = C, H, halogen, boronic acid, etc.), among others, giving access to new reaction pathways (35). Gold salts and complexes have catalytic activity in solution, particularly for unsaturated C-C bonds; new routes for organic synthesis have taken advantage of this activity (6, 7). In particular, readily available Au(I) and Au(III) chloride salts or complexes have been used as Lewis catalysts for many of these homogeneous reactions. However, their ubiquity has puzzled chemists because these catalysts perform similarly irrespective of the gold oxidation state. Some authors have speculated that Au(III) is reduced to Au(I) under reaction conditions and that the latter is the true catalyst; others have attributed this similar catalytic performance to the disproportionation of Au(I) to Au(III) and Au(0) (810). Here, we show that different Au(I) and Au(III) salts or gold complexes form 3- to 10-atom gold clusters in solution at room temperature and can act as extremely active catalysts, with turnover number (TON) values up to ~107 and turnover frequency (TOF) values up to ~105 hour−1. We present results for two types of representative Au-catalyzed reactions in solution: the ester-assisted intermolecular hydration of alkynes and the bromination of arenes (1113). Efficient formation of these very small gold clusters in diluted suspensions achieved a catalytic activity nearly five orders of magnitude higher than those previously reported.

Several di- and triatomic gold(I) cationic complexes have been successfully isolated (14) and used as catalysts for particular reactions (1518), as have multiatomic gold centers and well-defined AuNPs (1921). However, the sub-nanosized regime remains little explored in catalysis, despite results suggesting that multiatomic gold entities with sizes intermediate between well-defined gold complexes and AuNPs are potential active species (22, 23).

When catalytic species form during a reaction, a reaction-induction period can generally be observed, provided that the rate of formation of the catalyst is slower than the rate of the reaction. No reaction-induction periods are generally found for Au-catalyzed homogeneous reactions, even though there are several examples in which the active species is not the original source of gold (4, 7). The absence of an induction period can be explained by assuming that the active species is formed very quickly under the nominal amounts of Au used (1 to 5 mol%) and cannot be detected. However, if the amount of precursor is reduced to a level where the formation of the active gold species is measurable, then the catalytically active species and the rate of the reaction can be decoupled. To check this hypothesis, we carried out the ester-assisted hydration of alkynes in the presence of different amounts of Au(I) and Au(III) salts and, after optimization, performed kinetic experiments at 100-ppm concentrations of gold. The results in Fig. 1B show an induction period followed by a similar reaction rate for AuCl and HAuCl4, indicating that the real gold active species must be formed before the catalysis starts and could be the same regardless of the nature of the gold salts used. Note that low yields of product 4 were observed when no gold was introduced in the reaction and that a similar kinetic profile was found when Au(OH)3 was used as the gold source at 100 ppm; hence, this induction period was not due to autocatalytic or exothermic processes.

Fig. 1

Ester-assisted hydration of in situ–formed alkyne 3. (A) Structures of 1, 2, 3, and 4. (B) Plot-time conversion for AuCl (squares) and HAuCl4 (diamonds) at 100 ppm, after correction with the blank experiment. (C) Turnover number (TON) and turnover frequency (TOF) for different amounts of AuCl, calculated as moles of 4 formed per mole of AuCl at final conversion (TON) and as the initial reaction rate after the induction time per mole of AuCl (TOF). The final yield of 4 is >90% in all cases, except for 0.01 ppm, where it is ~50%. (D) Absorption measurements (a.u., arbitrary units) for the hydration of 3 containing the Au active species along the induction time (A) and when the reaction proceeds (B) with the corresponding fluorescence (inset, irradiated at 349 nm).

Because the ester-assisted hydration of alkynes has been reported with gold(III) salts and gold phosphine complexes at 1 to 5 mol% (10, 11), we also tested complexes AuPPh3NTf2 5 and AuPPh3Cl 6 as catalysts (Ph, phenyl; Tf, triflyl) but at low concentration (0.1 mol%). The evolution of the complexes was monitored by 1H and 31P nuclear magnetic resonance (NMR) spectroscopy (fig. S1) (24). Decomposition of these complexes should give the diphosphine adduct Au(PPh3)2X 7 and AuX, and indeed, the quantitative formation of complex 7 was observed after adding alkyne 2. Because the formation of ketone 4 occurred while 7 was present in the reaction mixture, it might be initially concluded that 7 was the active gold catalyst. However, when 7 was independently prepared and used as catalyst, the reaction did not take place. Thus, if 7 is not the catalytic species, the AuX salt is the only species involved in the catalysis, which connects with the results described above for gold(I) chloride salts (25).

The reaction was carried out with different concentrations of AuCl; the results are given in Fig. 1C. Gold formed an active catalyst at concentrations of 10 ppb, yielding a TON value of ~107. To our knowledge, this value is substantially greater than any value reported for a nonenzymatic catalyst at room temperature.

We investigated the possible generation of AuNPs during the induction period, and their evolution during the reaction, with the use of matrix-assisted laser desorption/ionization–time-of-flight (MALDI-TOF) mass spectrometry and ultraviolet-visible (UV-vis) spectroscopy. We took samples periodically while checking the progress of the reaction by gas chromatography (GC). Mass fragments and fragmentation patterns indicated that very small Au clusters formed in a few seconds and that these clusters did not aggregate to more than 13 atoms (26) at any moment [<2500 daltons; for MALDI-TOF spectra, mass analyses, and formula assignment, see fig. S2 (24)]. Initially, clusters were formed mainly by 5 to 9 Au atoms, which rearranged over time. The hydration of 3 started only when species with 3 to 5 gold atoms (Au3-Au5 clusters) predominated. Absorption UV-vis spectroscopy (Fig. 1D) confirmed these observations: The original absorption band for AuCl rapidly decomposed into three new bands at ~350, 450, and 570 nm, which correspond to the jellium model predictions for 3 to 5, 8 to 10, and 13 gold atoms, respectively, and finally collapsed into a main band at ~349 nm (4 to 5 Au atoms) when the reaction started and proceeded (27). Fluorescence measurements confirmed the energetic absorption of the Au cluster at the band gap level (see below) (23, 27). After 6 hours (~80% conversion), the reaction practically stopped, and in this situation, few Au3-Au5 clusters were observed. These results would indicate that for the ester-assisted hydration of alkynes, Au3-Au5 clusters can catalyze the reaction with very high turnover frequencies.

The results in fig. S3 show that the presence of H2O was not necessary to form the Au3-Au5 clusters because the reaction started just after their addition. In contrast, when solvent 2 was absent, the Au nanoclusters were not observed even after 2 hours. However, if solvent 2 was added after 2 hours, the reaction started after an induction period. In other words, the Au3-Au5 clusters were only formed in the presence of alkyne 2, and when they were formed, the induction period disappeared; hence, these clusters were very probably the catalytic active species. It is not surprising that very small gold clusters possess Lewis acidic character; recent studies have shown that the so-called molecular clusters (fewer than 20 atoms) present an available low unoccupied molecular orbital (LUMO) for nucleophilic interactions (28, 29). The role of HCl was also investigated with reactions performed with different acids and chloride salts. The reaction also proceeded in HBr, but no reaction occurred when alkyl ammonium or sodium chloride salts were used instead, which indicates that a low pH was key for the catalysis and that chlorides were apparently not necessary for the reaction to proceed.

The generality of the reaction was studied; the results are presented in Fig. 2. Alkyl and aryl propargyl esters with different functionalities were transformed to the corresponding ketones in good yields and high selectivity with AuCl concentrations of 200 ppm to 10 ppb.

Fig. 2

(A) Reaction scheme. (B) Isolated yields for the ester-assisted hydration of alkynes at low AuCl loadings at different reaction times (24). Asterisk indicates GC yield.

These results indicate that very small Au clusters of 3 to 5 atoms were formed from different Au sources at room temperature in acidified propargyl alcohol, and that these clusters could catalyze the ester-assisted intermolecular hydration of alkynes with catalytic efficiencies up to five orders of magnitude greater than those previously reported. In addition, the work described here rationalizes the similar catalytic efficiency observed for Au(I) and Au(III) compounds. To determine whether small gold clusters, rather than cationic gold, were the active species for other homogeneous reactions, we tested the AuCl3-catalyzed bromination of arenes (11). This reaction has been reported with catalytic amounts of AuCl3 ranging from 1 to 0.01 mol%, and the authors ascribed this high catalytic efficiency to the Lewis acidity of Au(III). However, if it is not Au(III) but the small clusters formed that are responsible for the catalytic activity, the less acidic AuCl should also work efficiently. The results in Fig. 3B show that when using AuCl at ppm concentrations under the reported reaction conditions—namely, dichloroethane (DCE) as a solvent at room temperature (11)—a catalytic efficiency similar to that with AuCl3 after an induction period was observed, and UV-vis measurements (fig. S4) indicate that small clusters of 7 to 9 atoms were responsible for the catalysis. Thus, it seems that different-sized clusters catalyze the two homogeneous reactions studied here.

Fig. 3

(A) Studied reaction schemes. (B to E) Plot-time conversion for the bromination of arene 16 (B) and for the same reaction in the presence of the reactants for the ester-assisted hydration of alkyne under the following reaction conditions (bromination, diamonds; hydration, squares): (C) All the reactants are added from the beginning. (D) The reactants 2, 16, and 17 are added from the beginning; at 1.5 hours of reaction time, 1 and 1 equivalent of water are added to start the hydration reaction. (E) The reactants 1, 2, and water are added from the beginning; at 1.5 hours of reaction time, 16 and 17 are added.

The formation of these Au7-Au9 clusters was slower in DCE than in alkyne 2, so the bromination of arenes in alkyne 2 should proceed independently of the ester-assisted hydration of alkynes, given that the catalytic gold clusters formed at different rates. To test this hypothesis, we performed the one-pot hydration of 3 in the presence of the reactants for the bromination 16 and 17, and the results in Fig. 3C show that the bromination of 16 rapidly started according to the early formation of Au8-Au9 clusters in acidified solvent 2 (fig. S2). However, as the concentration of Au3-Au5 clusters increased, the number of Au5-Au9 clusters decreased and the formation of 16 was inhibited, so the hydration of 3 was the only reaction taking place. This effect was also observed if the reactants for the hydration, 1 and water, were added after 1.5 hours of reaction time (Fig. 3D). Accordingly, no bromination occurred if the reactants 16 and 17 were added after 1.5 hours reaction time when no Au8-Au9 clusters were present (Fig. 3E). These results demonstrate that different small Au clusters catalyze different reactions and can be differentiated by kinetic experiments.

To further confirm these findings, we independently synthesized Au5 and Au8 clusters stabilized on the dendrimer Poly(amineamide-ethanol) (PAMAM-OH) (30). The results in Fig. 4 show that Au5-PAMAM at ppm concentrations did indeed catalyze the ester-assisted hydration of alkynes, whereas Au8-PAMAM preferentially catalyzed the bromination reaction, with no induction time in both cases. However, the catalytic efficiency of the gold clusters when supported on the dendrimer was at least two orders of magnitude lower than formed in situ in the propargyl alcohol 2.

Fig. 4

(A) Studied reaction schemes. (B and C) Plot-time conversion for the ester-assisted hydration (B) and for the bromination of p-dimethoxybenzene (C) with Au5-PAMAM (diamonds) and Au8-PAMAM (squares).

A systematic study of the cluster formation in different solvents and acids revealed that very small gold clusters formed under a variety of reaction conditions, with a size distribution controlled by the nature of both the solvent and the acid, as determined by UV-vis spectroscopy (fig. S5). These results imply that very small gold clusters can be formed under different experimental conditions and can be used in other catalytic processes.

Regardless of the gold salt [AuCl, AuCl3, Au(OH)3] or gold complex (AuPPh3Cl, AuPPh3NTf2) used, we observed rapid decomposition during the reaction to give gold clusters preferentially formed by 5 to 13 atoms. These clusters formed new Au3-Au5 clusters. Only when a critical concentration of the last clusters formed did the ester-assisted hydration of alkynes occur. However, when Au3-Au5 clusters were added at time zero, the ester-assisted hydration of alkynes immediately started with no induction period. With a second reaction (i.e., bromination of p-dimethoxybenzene), no induction period was observed with gold salts in acidified propargyl alcohol 2, indicating that the larger Au5-Au9 clusters initially formed are active for this reaction. The bromination of p-dimethoxybenzene readily stopped when the Au5-Au9 clusters were transformed into Au3-Au5 clusters. When both reactions were performed simultaneously, the bromination of p-dimethoxybenzene readily started, but no hydration reaction was observed. After ~2 hours, the bromination reaction stopped and the hydration rapidly proceeded, which reflected the transformation to a different type of clusters in solution. An independent synthesis of the different clusters and their use as catalysts confirmed the nature of the catalytic species. Our results may lead to more efficient Au catalytic systems and to a greater understanding of the chemistry of sub-nanosized particles in general.

Supplementary Materials

Materials and Methods

Figs. S1 to S5

References and Notes

  1. See supplementary materials on Science Online.
  2. As pointed out by the reviewers, other potential decomposition pathways such as oxidative dimerization of the alkyne or reduction of gold can also occur (31).
  3. Acknowledgments: Supported by an Instituto de Tecnología Química postgraduate scholarship (J.O.-M.); an FPU contract through the Ministerio de Ciencia e Innovacion (MCIINN) (J.R.C.-A.); Consejo Superior de Investigaciones Científicas (A.L.-P.); Consolider-Ingenio 2010 (proyecto MULTICAT); the MCIINN PLE2009 project; and Subprograma de Apoyo a Centros y Universidades de Excelencia Severo Ochoa (proyecto SEV 2012 0267).
View Abstract

Navigate This Article