Facile, Large Scale Synthesis of Dodecanethiol Stabilized Au₃₈ Clusters

Abstract

It has long been a major challenge to achieve synthetic control over size and monodispersity of gold thiolate nanoclusters. Among the reported Au_n thiolate clusters, Au₃₈ has been shown to be particularly stable, but was only obtained as a minor product in previous syntheses. In this work, we report a bulk solution synthetic method that permits large scale, facile synthesis of truly monodisperse Au₃₈ nanoclusters. This new method explores a two-phase ligand exchange process utilizing glutathione-capped Au_n clusters as the starting material. The ligand exchange process with neat dodecanethiols causes gold core etching and secondary growth of clusters, and eventually leads to monodisperse Au₃₈ clusters in high purity, which eliminates nontrivial postsynthetic separation steps. This method can be readily scaled up to synthesize Au₃₈(SC₁₂H₂₅)₂₄ in large quantities, and thus makes the approach and Au₃₈ nanoclusters of broad utility.

Metal nanoclusters that are composed of ~10 to hundreds of atoms bridge the gap of small organometallic compounds and nanocrystals.¹ Among precious metals, gold nanoclusters capped by thiolate ligands have attracts wide research interests due to their extraordinary chemical stability and interesting electronic and optical properties fundamentally different than their larger counterparts—gold nanocrystals. Unlike gold nanocrystals whose electronic properties resemble the bulk state, gold clusters with core diameter less than 2 nm exhibit quantized electronic states due to quantum confinement effect of electrons. These ultrasmall nanoparticles hold promise for a number of applications such as catalysis^2–4, biological labeling^5,6 and as building block for nanoscale devices.⁷

With respect to the chemical synthesis of Au_n thiolate clusters (n: the number of atoms), Whetten and coworkers previously synthesized glutathione (G-SH) stabilized gold clusters, Au_n(SG)_m, and separated the mixture by polyacrylamide gel electrophoresis (PAGE).⁸ The most abundant cluster species separated was determined to be Au₂₈(SG)₁₆;^8,9 note that it has recently been corrected as Au₂₅(SG)₁₈.¹⁰ Tsukuda and coworkers performed careful isolation and electrospray ionization mass spectrometry (ESI-MS) analysis of isolated Au:SG clusters, and identified a range of Au_n cluster species from Au₁₀ to Au₃₉.¹⁰ Thus far, Au:SR (where, R=the tail of thiolate) clusters with core masses of 5 kDa (25 atoms), 8 kDa (38 atoms), 14 kDa (~75 atoms), 22 kDa (~101 atoms) and 29 kDa (~140 atoms) have been reported.^11–13 Thiol etching experiments demonstrate that Au:SR clusters with core masses of 5 kDa (Au₂₅), 8 kDa (Au₃₈) and 29 kDa are extraordinarily stable.^12–16 For Au₂₅(SR)₁₈ clusters, improved syntheses have been reported in recent years.^17–19 And, the single crystal structure of [Au₂₅(SCH₂CH₂Ph)₁₈]⁻[TOA]⁺ has been independently determined by Murray²⁰ and Jin^21a groups, including the crystal structure [Au₂₅(SCH₂CH₂Ph)₁₈]⁰ clusters.^21b Both the anionic and neutral of charge neutral Au₂₅(SCH₂CH₂Ph)₁₈ clusters consist of an icosahedral Au₁₃ core and an exterior shell composed of six orthogonal Au₂(SCH₂CH₂Ph)₃ motifs.²¹ This structure was also theoretically described by Hakkinen et al.²² Among the extraordinarily stable 5 kDa (Au₂₅), 8 kDa (Au₃₈) and 29 kDa species, the 29 kDa Au:SR clusters (~1.6 nm core size, where R is C4, C6, C12, C18, or benzyl) were found to be predominantly formed in synthesis.^17,23–25 Recently, Tsukuda et al. improved the synthesis and formulated the composition of the 29 kDa species as Au₁₄₄(SR)₅₉ on the basis of detailed ESI-MS analysis.²⁶ The 8 kDa species was, however, only be obtained in very small quantities as a minor product following column chromatography separation^11,12,26 or a thiol etching method.^13,14,16 Nevertheless, the exact composition of the 8 kDa species has been determined to be Au₃₈(SR)₂₄ with ESI-MS.^11–13

Despite the significant amount of synthetic work that has been done on size and monodispersity control of gold thiolate clusters, most reported methods suffer from the production of a mixture of different sized clusters (except the case of Au₂₅(SR)₁₈ clusters), and the clusters of a desired size are often in low yield and small quantities. Thus, to develop bulk solution synthetic methods that offer precise size control of gold thiolate clusters at the atomic level is of critical importance if the full potential of this new type of nanomaterial is to be realized.

In this work, we report a bulk solution synthetic method that permits facile, large scale synthesis of truly monodisperse Au₃₈(SC₁₂H₂₅)₂₄ nanoclusters. This new method explores a two-phase ligand exchange process in which glutathione-capped Au_n clusters (a mixture) are utilized as the starting material. Neat dodecanethiol causes gold core etching of the starting material and secondary growth of clusters, ultimately resulting in monodisperse Au₃₈ clusters in high purity.

In a typical experiment, the two-phase synthesis of Au₃₈ thiolate clusters involves two major steps, i) synthesis of Au_n(SG)_m clusters and ii) two-phase ligand exchange induced growth of Au₃₈ clusters from the Au_n(SG)_m mixture (see Supporting Information for experimental details). Briefly, the Au:SG clusters were synthesized according to a well established literature method.^{8–10, 28} The aqueous Au:SG clusters were then subject to ligand exchange using neat 1-dodcanethiol (C₁₂-SH) in a two phase system, Figure 1. According to the previous work by Tsukuda et al.,¹⁰ the as-prepared Au:SG clusters contain a mixture of Au_n clusters ranging from Au₁₀ to Au₃₉. It has also been found that in the presence of excess G-SH and at 55 °C, Au_n clusters with n < 25 were oxidized to Au(I)-SG complexes while Au_n clusters with n ≥ 25 were converted to Au₂₅(SG)₁₈ clusters.¹⁵ In our work, we use C₁₂-SH (insoluble in water, i.e., a two phase system) to effect core etching of Au:SG clusters during the course of the two-phase ligand exchange process, and finally attain highly pure Au₃₈(SC₁₂H₂₅)₂₄ clusters. It is worth noting that Shichibu et al. previously explored a ligand-exhange approach to the synthesis of Au₂₅(SG)₁₈ clusters from phosphine-capped Au₁₁ clusters.²⁷

Figure 1.

Phase transfer of gold-glutathione clusters from water to dodecanethiol and further conversion to truly monodisperse Au₃₈(SC₁₂H₂₅)₂₄ clusters.

A key to our new synthetic approach is to effect an efficient phase transfer of Au:SG clusters from the aqueous solution to the C₁₂-SH phase (neat C₁₂-SH). We found that the Au:SG clusters can not be directly transferred into the C₁₂-SH phase due to a surface tension issue, hence, acetone was added to the water/C₁₂-SH two phase system to reduce the surface tension²⁹ and prompt phase transfer of Au:SG clusters from water to the C₁₂-SH phase. The ratio of water to acetone was found to influence the phase transfer efficiency. The optimal water to C₁₂-SH ratio was determined to be 1.0–1.5. More acetone results in precipitation of Au:SG clusters (black precipitate forms at the water/C₁₂-SH interface, while less acetone slows down the phase transfer rate. Figure 1 shows the phase transfer of Au:SG clusters from water to C₁₂-SH. The water phase is initially at the bottom of the flask and the C₁₂-SH phase forms the upper layer of the mixture. Prior to the phase transfer reactions, gold clusters are solely protected by G-SH, which makes them only soluble in water. After the solution was vigorously stirred for ~3 hours at 80 °C, gold clusters were almost completely transferred to the C₁₂-SH phase (Figure 1, right). During this process, gold cluster core etching by C₁₂-SH and secondary growth of clusters are expected to occur, which ultimately results in monodisperse Au₃₈(SC₁₂H₂₅)₂₄ clusters (vide infra).

The Au clusters in the C₁₂-SH phase were isolated after the two phase reaction process was completed. The excess C₁₂-SH was removed by washing the solution with methanol for several times. To isolate Au₃₈ clusters, toluene was used to extract the Au₃₈ clusters (see supporting Figure S1 for the absorption spectrum of the extracted clusters). The residue is poorly soluble in typical organic solvents (e.g. toluene or CH₂Cl₂, etc.), which is possibly Au(I)-SC₁₂ polymers since Au(I)-SC₁₂ polymer tends to form in thiol etching processes¹⁴ and has a poor solubility in organic solvents. To confirm that, we characterized the residue by thermal gravimetric analysis (TGA) and found that the weight loss is 50.3% (Figure S2), consistent with the calculated organic content (50.5%) of Au(I)-SC₁₂ polymer.

To further increase the purity of Au₃₈ clusters, a mixed toluene/acetone (1:5 v:v) solvent was used to extract Au₃₈ clusters from the crude product. The UV-vis spectrum of as-isolated Au₃₈ clusters shows a step-wise, multiple absorption band spectrum (Figure 2A), which is typical of gold thiolate clusters. The spectrum is almost superimposable with previously reported Au₃₈(SR)₂₄ optical spectrum.^11–13 The optical band edge energy is ~ 0.9 eV, and there is a prominent peak centered at 625 nm (2.0 eV). Other distinct peaks are at 1060 nm (1.17 eV), 755 nm (1.64 eV), and less intense ones at 2.2, 2.4 and 2.6 eV (Figure 2A and inset).

Figure 2.

(A) The UV-vis spectrum of Au₃₈ clusters. The inset shows the absorbance (abs (E)) vs photo energy (eV). The wavelength dependent absorbance is converted to energy dependent absorbance according to the equation: Abs(E) ∝ Abs(w) × w². (B) Typical chromatogram of Au₃₈ sample detected by DAD at 630 nm. Inset: UV-vis spectra are obtained by DAD at 14.0 min (black line), 14.8 min (red line), and 15.3 min (blue line).

Size exclusion chromatography was used to further confirm the purity of the as-prepared Au₃₈ clusters. A diode array detector (DAD) in situ monitors the UV-Vis spectra of eluted Au clusters. Figure 2B shows one single symmetric peak; the inset shows the spectra (190–950nm) corresponding to the eluted components at 14.0 min (left side of the peak), 14.8 min (the peak position) and 15.3 min (right side of the peak), respectively. The three spectra are indeed superimposable, which confirms high purity of the as-prepared Au₃₈ clusters.

The as-prepared Au₃₈ clusters were also analyzed by laser desorption ionization (LDI) mass spectrometry. In previous work, the formula of Au₃₈ clusters has been determined to be Au₃₈(SR)₂₄.^13,30 The LDI mass spectrum (negative mode) of the Au₃₈ clusters is shown in Figure 3A. A broad peak centered at ~8.2 kDa was observed. The 8.2 kDa peak mass corresponds to [Au₃₈S₂₄]⁻. The broadness of the mass peak is due to fragmentation of the [Au₃₈S₂₄] ⁻ ions as well as a series of recombination processes of the fragments in gas phase. Unlike the previous work in which the Au₃₈ clusters were found to always coexist with the Au₁₄₄ clusters.¹³ Herein, our new synthetic approach does not involve or generate Au₁₄₄ clusters, which greatly improves the purity of Au₃₈ clusters.

Figure 3.

(A) LDI mass spectrum and (B) ESI mass spectrum of Au₃₈(SR)₂₄ clusters, (C) NMR of Au₃₈(SR)₂₄ prepared from ligand exchange (black line) and Au₃₈(SC₁₂)₂₄ prepared by literature procedure (red line).¹³ The asterisks indicate the peaks from residual solvents. (D) TGA of Au₃₈(SR)₂₄ clusters.

Since our synthesis of Au₃₈ clusters involve two types of thiol ligands (G-SH and C₁₂-SH), it is important to find out whether the final Au₃₈ clusters possess a mixed ligand shell or sole dodecanethiolates. Given that the Au₃₈ clusters are only soluble in organic phase (toluene, CH₂Cl₂, etc.), the ligand shell of as-prepared Au₃₈ clusters may be only dodecanethiolates. To determine the exact ligand composition, we have performed electrospray mass spectrometry (ESI-MS), nuclear magnetic resonance (NMR), and thermogravimetric analysis (TGA).

Figure 3B shows the ESI-MS result of the Au₃₈ clusters. The peak at m/z 12,451 corresponds to the adducted species Au₃₈(SR)₂₄Cs⁺. After subtracting Cs⁺, the m/z 12,318 indeed matches the formula Au₃₈(SC₁₂H₂₅)₂₄. This is further confirmed by observing the intact cluster ion signal at m/z 12,318 (Figure 3B). These results show that the ligand exchange process between C₁₂-SH and G-SH on the Au cluster surface is thorough, which can be attributed to three reasons. First of all, a large excess of C₁₂SH was used (the ratio of C₁₂SH/GSH ~ 45). Secondly, the high temperature (80 °C) accelerates the ligand exchange process and tends to drive the process to completeness. Thirdly, the highly hydrophilic GSH ligand would be difficult to enter the C₁₂SH phase where Au₃₈ clusters are formed. The pure ligand composition of the Au₃₈ clusters was also confirmed by NMR analysis. We prepared and isolated Au₃₈(SC₁₂H₂₅)₂₄ clusters following a literature procedure,¹³ albeit the yield is very low. The ¹H NMR spectra of these two types of Au₃₈ clusters are indeed almost identical (Figure 3C), which confirms that no GSH ligands should exist in the final Au₃₈ clusters prepared by our two-phase exchange approach. In both NMR spectra, significantly broadened peaks (compared to free ligands) were observed, similar to previous report.²³ A set of broad peaks in the 2.8–4.1 ppm range are tentatively assigned to the H1 position (Figure 3C inset). The peaks in the 1.6–2.1 and 1.1–1.6 ppm range are assigned to H2 and H3-11, respectively. Finally, the triplet at ~0.9 ppm is assigned to the C12 position. TGA analysis also indicates that the Au₃₈ cluster composition should be Au₃₈(SC₁₂H₂₅)₂₄, Figure 3D; the weight loss of Au₃₈(SC₁₂H₂₅)₂₄ clusters prepared by the ligand exchange approach is 40.6%, which is very close to the theoretical value (39.2%).

It is worthy of comparing our two-phase exchange approach with previous work;^12–14 the latter has demonstrated that Au₃₈ clusters were formed during the etching process of large Au clusters (14 kDa)¹⁴ or a mixture of 14 kDa, 22 kDa and 29 kDa Au clusters.^12,13 The 22 kDa and 29 kDa Au clusters were found to coexist with Au₃₈ clusters even after reaction with neat C₁₂-SH for 24 hrs. To obtained pure Au₃₈ clusters, Tsukuda et al. used recycling size exclusion chromatography^12,26 or a complicated solvent extraction method¹³ to isolate 8 kDa, 22 kDa, 29 kDa Au clusters. This time-consuming procedure requires significant effort to obtain pure Au₃₈ in bulk quantities. In our work, we use an Au_n(SG)_m mixture (n=10–39) as the starting material to prepare Au₃₈ by ligand exchange with neat C₁₂-SH. Analogous to the reaction between excess GSH and small Au:SG clusters,¹⁵ those small, relatively unstable Au:SG clusters would decompose to form Au(I)-SC₁₂ complexes and only the thermodynamically stable Au₃₈ clusters survive the rather harsh thiol etching reaction at 80 °C. It is interesting that we did not observe Au₂₅ cluster formation in the exchange reaction with neat thiols at high temperature (80 °C), albeit Au₂₅ clusters are very stable under mild environment due to their highly symmetric and stable structure.²¹ A similar observation (i.e. not observing Au₂₅ cluster formation) has been noticed by Quinn et al.¹⁷ In our procedure, the Au₃₈ clusters are the sole cluster species after ligand exchange with neat C₁₂-SH. The extraordinary stability of Au₃₈(SR)₂₄ clusters at elevated temperatures (e.g. 80 °C) should account for the favorable growth of highly pure Au₃₈ clusters from the Au_n(SG)_m mixture in the two phase exchange process.

The above results exclusively demonstrate that the Au₃₈ clusters prepared by the ligand exchange process are Au₃₈(SC₁₂H₂₅)₂₄ in high purity, which is of critical importance for practical applications such as catalysis and biological labeling. This method allows easy scale-up for synthesizing Au₃₈(SC₁₂H₂₅)₂₄ in large quantities, which makes this synthetic approach and Au₃₈ clusters of broad utility. The crystal structure of Au₃₈(SR)₂₄ cluster is still unknown, but several groups have done theoretical calculations^30–34 and predict that Au₃₈(SR)₂₄ has 22 to 26 Au atoms as core and mixed Au(SR)₂ and Au₂(SR)₃ motifs as protecting groups.^13,33,34 This prediction is based upon some empirical structural rules learned from the single crystal structures of Au₁₀₂(p-MBA)₄₄ (p-MBA= p-mercaptobenzoic acid)³⁵ and [Au₂₅(SCH₂CH₂Ph)₁₈]⁻[TOA]^{+ 20, 21} reported recently.

In summary, we have developed a facile method for synthesizing Au₃₈(SC₁₂H₂₅)₂₄ clusters in high purity and yield via ligand exchange reactions of Au:SG clusters with neat C₁₂SH in a two phase reaction system. A key to this approach is the use of acetone to effect phase transfer of Au:SG clusters from the water phase to the neat C₁₂SH phase. Surprisingly, the ligand shell of Au₃₈ clusters was found to be sole dodecanethiolates, rather than a mixed ligand shell of dodecanethiolates and glutathionates. The exact formula of the as-prepared Au₃₈ clusters is determined to be Au₃₈(SC₁₂H₂₅)₂₄ based upon ESI-MS, NMR, and TGA analyses. The crystal structure of Au₃₈(SR)₂₄ cluster is yet to be unraveled.