Letter pubs.acs.org/NanoLett
Adding Two Active Silver Atoms on Au25 Nanoparticle Chuanhao Yao,† Jishi Chen,†,‡ Man-Bo Li,† Liren Liu,§ Jinlong Yang,§ and Zhikun Wu*,† †
Key Laboratory of Materials Physics, Anhui Key Laboratory of Nanomaterials and Nanostructures, Institute of Solid State Physics, Chinese Academy of Sciences, Hefei 230031, China ‡ Department of Materials Science and Engineering, University of Science and Technology of China, Hefei, Anhui 230026, China § Hefei National Laboratory for Physics Sciences at the Microscale, University of Science and Technology of China, Hefei, Anhui 230026, China S Supporting Information *
ABSTRACT: Alloy nanoparticles with atomic monodispersity is of importance for some fundamental research (e.g., the investigation of active sites). However, the controlled preparation of alloy nanoparticles with atomic monodispersity has long been a major challenge. Herein, for the first time a unique method, antigalvanic reduction (AGR), is introduced to synthesize atomically monodisperse Au25Ag2(SC2H4Ph)18 in high yield (89%) within 2 min. Interestingly, the two silver atoms in Au25Ag2(SC2H4Ph)18 do not replace the gold atoms in the precursor particle Au25(SC2H4Ph)18 but collocate on Au25, which was supported by experimental and calculated results. Also, the two silver atoms are active to play roles in stabilizing the alloy nanoparticle, triggering the nanoparticle fluorescence and catalyzing the hydrolysis of 1,3-diphenylprop-2-ynyl acetate. KEYWORDS: Alloy, nanoparticles, atomic monodispersity, antigalvanic reduction, fluorescence, active sites
C
reactive (i.e., more noble) metal atoms, that is, the more reactive metal nanoparticles act as templates for the substitution with the less reactive metal atoms in order to prepare nanoalloys.43,44,46,47 Because of the instant and spontaneous occurrence of GR, it is difficult to precisely control the reaction process at the atomic level and, until now, no monodisperse TAN has been reported by this method. We recently revealed an unexpected antigalvanic reduction (AGR),45,48 which provides a facile and mild way to synthesize nanoalloys that are otherwise difficult to obtain. Herein we demonstrate that AGR45,48,49 can be used to synthesize atomically monodisperse TAN, Au25Ag2(SC2H4Ph)18 (denoted as Au25Ag2) in gram scale and high yield (89%) within 2 min. Surprisingly, only a single-component TAN is formed, that is, no other compositional Au−Ag TAN such as Au25Ag(SC2H4Ph)18 with an 8-electron closing shell50,51 is observed in the product through wide-ranged reaction conditions (note: Au25Ag2(SC2H4Ph)18 is of 9-shell-closing electron count based on the superatom concept 50,51 ). The gold atoms in Au 25 (SC 2 H 4 Ph) 18 (denoted as Au 25 , counterion: [N(C8H17)4]+) are not replaced by the silver atoms, which is supported by the inductively coupled plasma atomic emission spectrometry (ICP-AES) together with other characterizations
ontrolling the nanoparticles with atom precision has long been a major goal for nanoscientists and chemists due to the significance for achieving precise structure (composition)property correlations and understanding some essential issues of nanostructures (e.g., active sites in catalysis).1−7 Since the early work by Brust et al.,8 some important advances have been achieved on the thiolated monometallic nanoparticles (e.g., Au, Ag nanoparticles) with atomic precision through wet chemical synthesis in the past decade.9−27 Alloy nanoparticles exhibit unusual structures and properties28−37 and are long pursued in research community; however, the synthesis of thiolated alloy nanoparticles (TAN for short) with atomic monodispersity38−40 is still a major challenge, which hampers the subsequent research of TAN on the structure, mechanism, and so forth. To the best of our knowledge, atomically monodisperse Au−Ag TAN (silver as the doping metal) are rarely attained by now. The main reasons are the diversity and similarity of the products based on the current synthesis methods (mainly the synchro-synthesized method, that is, the nanoalloys were formed after the mixed metals or salts were synchronously manipulated by physical or/and chemical protocols).30,40,41 Galvanic reduction (GR)42−47 is another well-recognized way to engineer the compositions, structures, and properties of metal nanostructures, but it requires the preparation of more reactive (i.e., less noble) metal nanoparticles initially, followed by the GR step in which the metal atoms in the as-prepared nanoparticles are replaced by less © XXXX American Chemical Society
Received: November 21, 2014 Revised: January 1, 2015
A
DOI: 10.1021/nl504477t Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
nm), which is also confirmed by transmission electron microscopy (TEM) measurement, Supporting Information Figure S1b. The absorption spectrum of the product (Figure 1a) is distinctly different from that of the synchro-synthesized Au25−xAgx (Figure 1b) but somewhat resembles that of Au25 (Figure 1c) although some differences do exist (e.g., Au25 has two absorption bands centered at 445 and 400 nm in the shortwavelength visible region, while the new product has only one peak centered at 412 nm in the same region). The absorption differences between the product and Au25−xAgx (or Au25) indicate that the product is not identical with Au25−xAgx (or Au25), however the resemblances between the product and Au25 remind us that they may have some structural similarities. As for the structure of Au25, it can be viewed as a core−shell structure composed of an icosahedral Au13 core and a shell of 12 Au atoms that are face-capped on the Au13 core, that is, Au13(core)/Au12 (shell).13,15,52 As far as we know, for Au25 the lowest energy band at 685 nm corresponds to the HOMO− LUMO transition that is due entirely to the electronic and geometric structure of the Au13 core in Au25 skeleton, and the other two featured bands (associated with the exterior Au12 shell in Au25 skeleton) at 445 and 400 nm are assigned to mixed intraband (sp ← sp) and interband (sp ← d) transitions, an interband transition (sp ← d), respectively.1,15 Thereby, the slight shift from 685 nm (Au25) to 700 nm (product) indicates that the Au13 core in Au25 is not influenced greatly, while the disappearance of absorption bands at 445 and 400 nm and the emergence of a new band at 412 nm indicate that the exterior Au12 shell in Au25 skeleton is distinctly effected. Matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS), which is a very powerful and well-recognized tool to identify metal nanoclusters, is employed to analyze the product. Interestingly, it shows a distinct peak at ∼7610 Da (highest intensity in the mass spectrum, Figure 2). The unity spacing of the isotopes (Figure 2 inset) implies that the ionized clusters bear a +1 charge (acquired in the positive ion mode); the m/z value of the peak therefore represents the molecular ion mass. It is known that the formula weight of Au25(SC2H4Ph)18 is ∼7394 Da.20 The deviation between
(e.g., mass spectrometry (MS) analyses). The composition, UV/vis/NIR, and MS analyses also indicate that the structural framework of Au25 remains in Au25Ag2 and the two silver atoms are just deposited on Au25; computer calculation provides further support for this. Importantly, the as-prepared TAN (with two silver atoms enclosed) exhibit enhanced ambient stability, fluorescence, and catalytic activity compared with the precursor particle Au25 and the synchro-synthesized Au25−xAgx (SC2H4Ph)18 (x = 2, 3, 4, denoted as Au25−xAgx). Below we present more detailed information and discussions. A facile and mild approach is developed for obtaining Au25Ag2 (details are seen in the Supporting Information). Briefly, in a typical experiment (note: the entire experimental process is conducted under ambient conditions), Au 25 precursor particles (39.3 mg, 0.005 mmol) were dissolved in acetonitrile, then a freshly prepared AgNO3 (1.7 mg, 0.01 mmol) solution was added drop by drop within 2 min. Over the period of 2 min, the solution quickly turned from red to graygreen and a large amount of black solid precipitated out (see inset of Supporting Information Figure S1a), indicating the occurrence of the AGR reaction. The collected solid was washed by excess CH3CN, then recrystallized to yield tiny cubic crystals (Supporting Information Figure S1a). It is noteworthy that one-batch synthesis of Au25Ag2 particles on a scale of grams has been achieved by scaling up the reaction (Supporting Information Figure S2). As a comparison, the mixed Au25−xAgx (Supporting Information Figure S3) nanoparticles are also prepared by modifying a previous method.9,20 Compared with the synthesis of Au25−xAgx, the synthesis of Au25Ag2 from Au25 is rather facile, time-saving, high-yield, and can be reproduced in large scale. The product (dissolved in toluene) shows two distinct absorption bands centered at 700 and 412 nm, respectively (Figure 1a). The UV/vis/NIR spectrum of the crude product is
Figure 1. UV/vis/NIR absorption spectra of Au25Ag2 before and after purification (a), Au25−xAgx (b), and Au25 (c). (Insets: digital photos of the corresponding products dissolved in toluene).
indeed almost superimposable with that of the pure clusters, indicating an extraordinarily high purity of the as-synthesized product, which was evidenced by further characterization such as MS analyses and elemental analyses (see below). The absence of plasma resonance peak at ∼520 nm implies that the nanoparticles are smaller than ∼2 nm (often called nanoclusters, opposite to plasmonic nanocrystals larger than ∼2
Figure 2. MALDI-MS spectra of Au25Ag2(SC2H4Ph)18 (a) and Au25(SC2H4Ph)18 (b). (Inset a: comparison of the experimental and calculated isotope patterns; resolution, 15 000; calcd for Au25Ag2C144H162(M)+, 7608.7406 Da; found, 7608.7197 Da; error, 2.7 ppm). Note: the mass calibration was performed using Au25(SC2H4Ph)18 as an internal standard, the error measurement was referred to a previous method.53 B
DOI: 10.1021/nl504477t Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 3. MALDI-MS spectra of Au25Ag2 (SC2H4Ph)18 under various laser intensity (a); MALDI-MS spectra of Au25Ag2(SC2H4Ph)18 with or without further recrystallization for twice under identical laser irradiation (25%) (b). Inset: thin-layer chromatography of Au25 (1 and 1′), Au25Ag2 (2 and 2′) and the mixture of Au25 and Au25Ag2 (3 and 3′). Note: 2, without further recrystallization; 2′, with further recrystallization.
7394 Da and the 7610 Da peaks is 216 Da, which is equal to the mass of two Ag atoms. Upon the basis of the exact mass, the formula can be readily deduced, that is, Au25Ag2(SC2H4Ph)18, which is supported by the excellent agreement of the isotopic patterns between experiment and calculation (Figure 2a inset, the more accurate molecule ion peak is observed at 7608.7197 Da after calibration with Au25 as an internal standard). Figure 2a also shows some minor peaks in the range of m/z lower than 7000 Da. By comparing with that of Au25 (Figure 2b), it is rationally deduced that they are fragments from Au25. Of note, herein Au25 is also a fragment of Au25Ag2 due to three facts: (i) the peak intensity at 7394 Da is influenced by the laser intensity (see Figure 3a); (ii) the peak intensity does not vary notably under identical laser irradiation after further recrystallization of the product for twice (see Figure 3b); and (iii) thin-layer chromatography (TLC) results also exclude the coexistence of Au25 (see Figure 3b, inset). To further confirm the composition and purity, both thermogravimetric analysis (TGA) and X-ray photoelectron spectroscopy (XPS) were performed. The ratio of metals (Au and Ag) to organic component (ligands) can be determined by TGA, and the atomic number ratio of Au/Ag/S can be determined by XPS. Note that these methods rely on the nanoparticles’ purity. If the nanoparticles are pure, these approaches readily give rise to a precise formula.12 But for a size-mixed sample, such approaches are less useful in nanoparticle formula determination because these methods only provide a mean value of the total sample. First, TGA was performed. A weight loss of 32.6 wt % indicates that the nanoparticles contain 32.6 wt % organic component and 67.4 wt % metal (Au and Ag), see Figure 4a. This is in perfect agreement with the theoretical values of Au25Ag2(SC2H4Ph)18 (organic, 32.5 wt %; metal, 67.5 wt %). For comparison, the TGA of Au 25−x Ag x (SC 2 H 4 Ph) 18 (x = 2, 3, 4) and Au25(SC2H4Ph)18 were conducted as well (Figure 4b,c). Au25−xAgx are mixed nanoparticles in terms of the x values (x = 2, 3, 4), thus 38.7 wt % is a mean value of weight loss (the theoretical losses: for Au23Ag2, 38.2 wt %; for Au22Ag3, 38.7 wt %; for Au21Ag4, 39.1 wt %). Of note, both Au25 and Au25−xAgx are anionic nanoparticles (counterion: N(C8H17)4+). When one calculates the weight loss, the mass of N(C8H17)4+ should be
Figure 4. TGA of Au25Ag2(SC2H4Ph)18 (a), Au25−xAgx(SC2H4Ph)18 (b), and Au25(SC2H4Ph)18 (c).
involved. The weight loss of 36.5 wt % for Au25 is in good agreement with the theoretical value (37.3 wt %). For Au25Ag2, in case that the counterion-N(C8H17)4+ was enclosed, the expected weight loss would be 36.4 wt %, which has a large deviation from the experimental value (32.6 wt %), thus N(C8H17)4+ should not be present in the as-prepared nanoparticle, and indeed XPS analysis also excludes the existence of element N. The absence of N(C8H17)4+ counterion indicates that Au25Ag2 is not anionic as the case of Au25. The fact that both Cl− and Br− are not detected in the XPS spectrum and salt tests (see Figure S4 and the Experimental Procedure in Supporting Information) indicates that Au25Ag2 is not cationic (with Cl− or Br− as counterion), either. Therefore, it should be charge neutral. The XPS analysis shows that the Au/Ag/S atomic ratio is 14.67/1.17/10.68 (i.e., 25/1.99/18.1), which is consistent very well with the expected ratio of 25/2/18 for Au25Ag2(SC2H4Ph)18. Besides, elemental analysis provides another strong support for the purity and composition (calculated for Au25Ag2S18C144H162: C, 22.73; H, 2.15; S, 7.58. Found: C, 22.51; H, 2.11; S, 7.46). With respect to the neutral state of the silver atoms in the alloy nanoparticles, XPS C
DOI: 10.1021/nl504477t Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters shows that the binding energies of Ag 3d5/2 (368.2 eV) and Ag 3d3/2 (374.2 eV) are identical to those of standard neutral silver (368.2 eV for Ag 3d5/2, 374.2 eV for Ag 3d3/2),45 see Supporting Information Figure S5a. A slight negative shift of the binding energy of Ag 3d5/2 similar to the previous report35 was not found and may related to the increase of electropositivity of Au in Au25Ag2, see Supporting Information Figure S5b,c; a positive 0.2 eV shift of the binding energy of Au4f was detected from Au25 to Au25Ag2 (note: other possibilities may also exist). Moreover, salt test involving the addition of excess I− to the nanoparticles solution (see the Supporting Information for experimental details) did not result in any precipitate, excluding the presence of Ag+ enclosed in the nanoparticles. Taken together, these experimental results unambiguously demonstrate that the nanoparticles are pure with a molecule-like composition, Au25Ag2(SC2H4Ph)18. To be noted, the preparation reaction involves adding two silver atoms on Au25 but also changing the charge state from −1 to 0. The monodispersity of the as-prepared alloy nanoparticles without the coexistence of other compositional TAN (e.g., Au25Ag(SC2H4Ph)18 with a shell-closing electronic structure50,51) is indeed a surprise, considering the previous efforts to make pure silver-doped gold nanoparticles.30,40,41 “Sizefocusing”20,54 is a recognized strategy to attain monodisperse nanoparticles, however, herein it should not occur because the product was collected immediately without aging after the completion of the addition of silver ions. The starting Ag+/Au25 molar ratio in the reaction could be a nontrivial factor that influences the composition of the product. To investigate this, various Ag+/Au25 molar ratios were tested for the AGR reaction (see Supporting Information). Very surprisingly, no matter how to tune the starting Ag+/Au25 molar ratios (from 0.1 to 7), the alloy product is only one with unchanged composition, Au25Ag2(SC2H4Ph)18, which is evidenced by the UV/vis/NIR spectrum and critical MALDI-MS analyses (see Supporting Information Figure S6). Of note, the yield of the synthetic reaction is lower in the cases of less than 2 equiv. of silver. The exclusive composition is another surprise for it implies that the two incorporated silver atoms do not replace the gold atoms in the parent nanoparticle, which is supported by the UV/vis/NIR spectral resemblance to Au25 (Figure 1), together with the identification of the fragment ion peak at 7394 Da (Figure 2). ICP-AES analyses provide further support. The concentration of gold in the supernatant after AGR reaction is below the detection limit of ICP-AES, while in the case where two gold atoms were replaced by two silver atoms it would be ca. 490 ppm as expected in the preset conditions. As mentioned above, the UV/vis/NIR spectral resemblance to Au 25 (especially the spectral shape similarity in the NIR region that is assigned to the feature absorption of Au13 core15), together with the identification of fragment ion peak at 7394 Da, indicates that Au25Ag2 might retain the structural framework of Au25 and the two silver atoms could be deposited on Au25, which is further supported by the absence of fragment containing silver atom(s) in the mass spectrum (see Figure 2) and the similar XRD patterns of Au25Ag2 and Au25, see Figure 5. Besides, the loss of two Ag atoms from independent sites seems very unlikely without an intermediate M-Ag (i.e., Au25Ag(SC2H4Ph)18) being in evidence, and the present data cannot rule out a structure in which two silvers are collocated on the Au25. The emerging multiple peaks (in the range from 2.5 to 4.5 ppm) assigned to methylene protons of phenylethanethiolate in 1H NMR spectrum of Au25Ag2 provides
Figure 5. XRD patterns of Au25Ag2 and Au25..
another support for the structure, see Supporting Information Figure S7 (as a comparison, the 1H NMR spectra of Au25−/0 are also shown). Theoretical calculation verified this possibility: the simulated UV/vis/NIR spectrum of the proposed structure (after optimizations, see Figure 6b) is very close to the
Figure 6. Structure of Au25 (a) and Au25Ag2 (b) and the comparison of the experimental and calculated spectrum of Au25Ag2 (c). (Color labels: yellow, S; green, Au; red, Ag. All H and C atoms are omitted for clarity) .
experimental one (Figure 6c). Of course, the exact structure of Au25Ag2 remains to be attained by single-crystal X-ray diffraction measurement in future efforts. Structure (composition)-property correlation for nanoparticles is a puzzling but fundamental issue that has stimulated enormous interests to nanoscientists. For relatively large nanoparticles (>2 nm), due to the difficulties in both controlling the nanoparticles with atomic precision and accessing the single crystal structure it is very difficult to achieve precise structure (composition)-property correlation, while for the ultrasmall nanoparticles (
