Langmuir 2004, 20, 1945-1952
1945
Synthesis and Isolation of the Molecule-like Cluster Au38(PhCH2CH2S)24 Robert L. Donkers, Dongil Lee, and Royce W. Murray* Kenan Laboratories of Chemistry, University of North Carolina, Chapel Hill, North Carolina 27599-3290 Received September 12, 2003. In Final Form: December 8, 2003 The synthesis and characterization of phenylethanethiolate-coated monolayer protected clusters (MPCs) with Au140 (average) and Au38 cores are described. The latter 1.1 nm core diameter nanoparticles, whose Au38(PhC2S)24 composition was analytically established, are stable and isolable in relatively high purity and ∼200 mg quantities per reaction batch. Chemical shift effects in 1H and 13C NMR spectra reveal the existence of either closely related Au38 structures or differences between ligands on a given structure, or both. The Au38(PhC2S)24 MPCs undergo place exchange reactions with other thiolate ligands and, owing to their stability and ease of production, provide a gateway to other Au38 MPCs with partially or completely different monolayer compositions.
Introduction Reducing the dimensions of bulk chemical materials of known properties to unsupported, solution-soluble objects of the same chemicalssbut having nanoscale dimensions1s is receiving enormous attention because of the promise of such substances as components for nanoscaled devices and chemical analysis and because of interest in understanding fundamental size-property relationships of matter.2-8 Nanoparticles composed of semiconductor materials and whose properties are size dependent are called “quantum dots” and are being researched across a span of scientific disciplines. Nanoparticles that are based on very small metal nanoparticles9-34 that have sizedependent properties are also known, but the available * To whom email.unc.edu.
correspondence
may
be
addressed.
variety of these materials is both limited and mostly based on gold nanoparticles. As with semiconductors, advances in property investigations of metal quantum dots are strongly tied to advances in synthesis, as well as in theoretical models that describe their size-dependent properties. Also as with semiconductors, the stability of metal nanoparticles (against aggregation and other chemical changes), and the associated ease of their experimental investigation, depends on a successfully protective capping shell of ligands. Phosphines have been employed for Pd and Au nanoparticles, leading to nanoparticles with Pd59 and Pd145,9,10 Au11,31 and Au558,28 cores, among others. Thiolate ligands have been employed for stabilization of monolayerprotected cluster (MPC) Au cores that have been reported to contain 9,34 11,35 28,11,12 38,23,36 55,30 75,14,19 116,19 and
rwm@
(1) In Small Wonders, Endless Frontiers: A Review of the national Nanotechnology Initiative; Committee for the Review of the National Nanotechnology Initiative: National Academy Press: Washington, DC, 2002. (2) Alivisatos, A. P. Science 1996, 271, 933. (3) Brus L. E. J. Phys. Chem. 1986, 90, 2555. (4) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Science 1995, 270, 1355. (5) Wehrenberg, B. L.; Guyot-Sionnest, P. J. Am. Chem. Soc. 2003, 125, 7806. (6) Wang, C.; Shim, M.; Guyot-Sionnest, P. Science 2001, 291, 2390. (7) Molecular Electronics; Jortner, J.; Ratner, M., Eds.; Blackwell Science: Oxford, 1997. (8) Schmid, G. Chem. Rev. 1992, 72, 1709. (9) Tran, N. T.; Kawano, M.; Powell, D. R.; Dahl, L. F. J. Am. Chem. Soc. 1998, 120, 10986. (10) Tran, N. T.; Powell, D. R.; Dahl, L. F. Angew. Chem., Int. Ed. 2000, 39, 4121. (11) Schaaff, T. G.; Knight, G.; Shafigullin, M. N.; Borkman, R. F.; Whetten, R. L. J. Phys. Chem. B 1998, 102, 10643. (12) Link, S.; Beeby, A.; FitzGerald, S.; El-Sayed, M. A.; Schaaff, T. G.; Whetten, R. L. J. Phys. Chem. B 2002, 106, 3410. (13) Alvarez, M. M.; Khoury, J. T.; Schaaff, T. G.; Shafigullin, M. N.; Vezmar, I.; Whetten, R. L. J. Phys. Chem. B 1997, 101, 3706-3712. (14) Gutie´rrez, E.; Powell, R. D.; Furuya, F. R.; Hainfeld, J. F.; Schaaff, T. G.; Shafigullin, M. N.; Stephens, P. W.; Whetten, R. L. Eur. Phys. J. D 1999, 9, 647. (15) Hicks, J. F.; Templeton, A. C.; Chen, S.; Sheran, K. M.; Jasti, R.; Murray, R. W.; Debord, J.; Schaaff, T. G.; Whetten, R. L. Anal. Chem. 1999, 71, 3703. (16) Schaaff, T. G.; Shafigullin, M. N.; Khoury, J. T.; Vezmar, I.; Whetten, R. L.; Cullen, W. G.; First, P. N.; Gutie´rrez-Wing, C.; Ascensio, J.; Jose-Yacama´n, M. J. J. Phys. Chem. B 1997, 101, 7885-7891. (17) Schaaff, T. G.; Shafigullin, M. N.; Khoury, J. T.; Vezmar, I.; Whetten, R. L. J. Phys. Chem. B 2001, 105, 8785.
(18) Whetten, R. L.; Khoury, J. T.; Alvarez, M. M.; Murthy, S.; Vezmar, L.; Wang, Z. L.; Stephens, P. W.; Cleveland, C. L.; Luedtke, W. D.; Landman, U. Adv. Mater. 1996, 8, 428-422. (19) Alvarez, M. M.; Khoury, J. T.; Schaaff, T. G.; Shafigullin, M. N.; Vezmar, I.; Whetten, R. L. Chem. Phys. Lett. 1997, 266, 91. (20) Schaaff, T. G.; Whetten, R. L. J. Phys. Chem. B 1999, 103, 9394. (21) Aguila, A.; Murray, R. W. Langmuir 2000, 16, 5949. (22) Hostetler, M. J.; Wingate, J. E.; Zhong, C.-J.; Harris, J. E.; Vachet, R. W.; Clark, M. R.; Londono, J. D.; Green, S. J.; Stokes, J. J.; Wignall, G. D.; Glish, G. L.; Porter, M. D.; Evans, N. D.; Murray, R. W. Langmuir 1998, 14, 17-30. (23) Lee, D.; Donkers, R. L.; Desimone, J. M.; Murray, R. W. J. Am. Chem. Soc. 2003, 125, 1182-1183. (24) Miles, D. T.; Murray, R. W. Anal. Chem. 2003, 75, 1251. (25) Wuelfing, W. P.; Green, S. J.; Pietron, J. J.; Cliffel, D. E.; Murray, R. W. J. Am. Chem. Soc. 2000, 122, 11465. (26) Wuelfing, W. P.; Murray, R. W. J. Phys. Chem. B 2002, 106, 3139. (27) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, 33, 27. (28) Schmid, G. Inorg. Synth. 1990, 27, 214. (29) Toshima, N.; Shiraishi, Y.; Teranishi, T.; Miyake, M.; Tominaga, T.; Watanabe, H.; Brijoux, W.; Bo¨nnemann, H.; Schmid, G. Appl. Organomet. Chem. 2001, 15, 178. (30) Brown, L. O.; Hutchison, J. E. J. Am. Chem. Soc. 1997, 119, 12384. (31) Yang, Y.; Chen, S. Nano Lett. 2003, 3, 75. (32) Hainfeld, J. F.; Furuya, F. R. J. Histochem. Cytochem. 1992, 40, 177. (33) Hainfeld, J. F. J. Struct. Biol. 1999, 127, 93. (34) Hainfeld, J. F.; Powell, R. D. J. Histochem. Cytochem. 2000, 48, 471-480. (35) Woehrle, G. H.; Warner, M. G.; Hutchison, J. E. J. Phys. Chem. B 2002, 106, 9979. (36) Ha¨kkinen, H.; Barnett, R. N.; Landman, U. Phys. Rev. Lett. 1999, 82, 3264-3267.
10.1021/la035706w CCC: $27.50 © 2004 American Chemical Society Published on Web 01/20/2004
1946
Langmuir, Vol. 20, No. 5, 2004
140-14517,19,22,37 Au atoms, among others. These various atom numbers correspond to “magic numbers”, or stable closed shell core structures. The Brust synthesis38 has been a convenient method for preparing Au MPCs stabilized by thiolate ligands. In this reaction, gold(I) is reduced in the presence of an excess of the protecting thiol; competition between Au core growth and thiolate monolayer passivation determines the core diameter. Because the core sizes are influenced by a kinetic competition, a range of core sizes tends to be produced, with larger proportions of smaller cores, on average, resulting from larger thiol/Au reaction feed ratios22,39 and lowered reaction temperatures. Lowered temperatures presumably retard the rate of core growth more than that of thiolate passivation and lend dominance to the latter. Postreaction methods to reduce and/or analyze the dispersity of Au MPC core diameters have included fractional precipitation,17,40,41 size exclusion chromatography,14,42 reversed-phase high-performance chromatography (HPLC),39 electrophoresis,14,43 and chemical etching20 and annealing reactions.37 Generally, the postreaction methods are incompletely successful, and in most purified samples, impurity core diameters are still present. Despite the dispersity problem, some gold MPCs have been isolated (often in low yield) by others12,13,16,17,31,44 and us27,37,41,45,46 in sufficient purity for study of their electrochemical and photophysical properties as a function of core diameter. Core-size-dependent properties have been observed for Au cores with ∼140-145 atoms, which in our experiments have a core diameter of 1.6 nm and an average formula of Au140(C6)53. ‘The Au140 MPCs do not exhibit a significant HOMO-LUMO band gap, but instead have such a small double layer capacitance that a series of more or less evenly spaced one electron current peaks can be observed27,37,45,47 electrochemically. This electronic core-charging phenomenon is called quantized double layer (QDL) charging, and the spacing of the charging steps on the potential axis depends on the MPC core size45,47 and protecting monolayer thickness.15 Smaller MPCs become more overtly molecule-like, by the appearance of a significant HOMO-LUMO gap as evidenced both optically and electrochemically.45 Condensation of the continuum electronic structures of these smaller MPCs into more defined energy states is additionally evidenced by steplike features in UV-vis spectra.13,48 The above behavior is in contrast to that of larger core MPCs, which display a featureless decrease of high optical absorption in the UV to low values in the NIR region, interrupted for (37) Hicks, J. F.; Miles, D. T.; Murray, R. W. J. Am. Chem. Soc. 2002, 124, 13322. (38) Brust, M.; Walker, M.; Bethell, D.; Schriffrin, D. J.; Whyman, R. J. Chem. Soc., Chem. Commun. 1994, 801. (39) Donkers, R. L.; Lee, D.; Jimenez, V.; Georganopoulou, D.; Wang, G.; Harper, A. S.; Brennan, J.; DeSimone, J. M.; Murray, R. W. Manuscripts in preparation. (40) Chen, S.; Murray, R. W. J. Phys. Chem. B 1999, 103, 9996. (41) Chen, S.; Hostetler, M. J.; Evans, N. D.; Murray, R. W. Langmuir 1999, 15, 682. (42) Wilcoxon, J. P.; Martin, J. E.; Provencio Langmuir 2000, 16, 9912-9920. (43) Templeton, A. C.; Cliffel, D. E.; Murray, R. W. J. Am. Chem. Soc. 1999, 121, 7081-7089. (44) Li, D.; Li, J. Surf. Sci. 2003, 522, 105. (45) Chen, S.; Ingram, R. S.; Hostetler, M. J.; Pietron, J. J.; Murray, R. W.; Schaaff, T. G.; Khoury, J. T.; Alvarez, M. M.; Whetten, R. L. Science 1998, 280, 2098. (46) Ingram, R. S.; Hostetler, M. J.; Murray, R. W.; Schaff, T. G.; Khoury, J. T.; Whetten, R. L.; Bigioni, T. P.; Guthrie, D. K.; First, P. N. J. Am. Chem. Soc. 1997, 119, 9279. (47) Chen, S.; Murray, R. W.; Feldberg, S. W. J. Phys. Chem. B 1998, 102, 9898. (48) Quinten, M. Surf. Sci. 1986, 172, 557-577.
Donkers et al.
MPCs with g2 nm cores, only by a characteristic surface plasmon (SPR) absorbance at about 520 nm. Information on the properties of MPCs with core sizes e1.6 nm remains sparse, in substantial part because of the synthetic challenges of preparing stable, monodisperse metal quantum dots, including making them in sufficient quantity as to allow a wide range of property investigations. This paper describes the synthesis and isolation of monodisperse phenylethanethiolate-coated MPCs that are analytically ascertained to have Au38 cores (diameter 1.1 nm) and a shell of 24 phenylethanethiolate ligands (i.e., Au38(PhC2)24). The raw synthetic products contain a substantial quantity of larger core MPCs, mostly Au140 cores (average diameter 1.6 nm, but rather polydisperse), from which the Au38(PhC2)24 MPCs could be isolated, in ∼200 mg batches per reaction, based on solubility differences. Au38 nanoparticles have not been previously synthesized on a practical scale or isolated with high purity. The Au38(PhC2)24 synthesis adds a significant source of molecule-like nanoparticles to the alkanethiolatestabilized Au11 and Au55 MPCs reported by Yang and Chen,31 and Brown and Hutchison,30 and the glutathionecapped Au28 MPCs isolated by Schaaff, et al.12 The synthesis described here has facilitated a series of investigations of Au38 MPCs that are described in other papers: (a) the voltammetry under23 N2 and under39 liquid CO2 of a molecular melt comprised of a thiolated polyether derivative prepared with Au38(PhC2)24 as precursor, (b) the spectroscopy and voltammetry of Au38(PhC2)24 that explore its HOMO-LUMO gap properties, and (c) the synthesis, properties, and high-resolution HPLC of the related alkanethiolate-protected Au38 MPC.39 An important issue in molecule-like nanoparticle behavior is the extent to which properties are a function of the protecting ligand as well as of the actual number of core atoms. Significant results on that point are found in the reports by Hutchison et al.35 and Yang and Chen,31 showing that the stability and electronic band gap, respectively, are altered by replacing phosphine with thiolate ligands. The ready availability of the Au38(PhC2)24 MPC should allow inspection of the consequences of less drastic changes in nanoparticle ligation. Experimental Section Chemicals. 2-Phenylethane thiol (PhC2SH, 99%), decanethiol (98%), hexanethiol (98%), tetra-n-octylammonium bromide (Oct4NBr, 98%), and sodium borohydride (99%) from Aldrich, toluene, methylene chloride, and methanol (all reagent grade) and acetonitrile (Optima) from Fisher, tetrahydrofuran (>99.9% anhydrous) and 1 M lithium triethylborohydride (superhydride) in THF from Acros, and ethanol from Aarper Alcohol and Chemical Co. were all used as received. Hydrogen tetrachloroaurate trihydrate (from 99.999% pure gold) was prepared using a literature procedure49 and stored in a freezer at -20 °C. Low conductivity water was obtained with a Millipore Nanopure water purification system. Measurements. Thermogravimetric analysis (TGA) was performed with a Seiko SSC 5200 thermal analysis system on accurately weighed, carefully dried, g15 mg cluster samples, under N2 and in standard Al pans. In the temperature ramp (25 to 600 °C at 20 °C/min), the organic volatilization and ensuing mass loss occurred in a single step just above 200 °C (Figure S-1), leaving a shiny gold metal residue in the pan.22 Results of organic fraction of Au38(PhC2)24: expt, 32.8%; calcd, 33.4%. Au140(PhC2)53: expt, 24.2%; calcd, 20.8%. Transmission electron microscopy phase contrast images were obtained with a side-entry Phillips CM12 microscope operating at 120 keV, of MPC samples prepared by spreading a droplet of (49) In Handbook of Preparative Inorganic Chemistry; Brauer, G., Ed.; Academic Press: New York, 1965; p 1054.
Monolayer Protected Clusters 1 mg/mL MPC in toluene (drying in air for g24 h) on standard carbon-coated (20-30 nm) Formvar films on copper grids (600 mesh). Three typical regions of each sample were imaged at 510 or 660 K magnification. Core size histograms were read (and confirmed manually by ruler on an expanded printout) from digitized photographic images using Scion Image Beta Release 2 (www.scioncorp.com). Values from obvious twins or aggregates of particles were deleted manually. 1H and 13C NMR spectra (referenced to Si(CH ) ) were taken 3 4 in CD2Cl2 solutions with Bruker AC 200 MHz and Varian Inova 600 MHz spectrometers. Purification of synthetic batches of PhC2MPC was monitored from the NMR resonances from free phenylethane thiol and Oct4Br impurities. MPC concentrations for UV-vis spectra, taken with a Cary 50 UV-vis spectrometer in CCl4 solutions, were chosen13 to provide A ≈ 1.0 at 4 eV. Synthesis of PhC2-Coated MPCs. PhC2 MPCs were synthesized with the well-known Brust two-phase procedure22,38 and also with a single-phase version. Brust Procedure: After 3.10 g (7.2 mmol) of HAuCl4‚(H2O)3 in 100 mL of Nanopure water was added to a solution of 5.00 g (9.1 mmol) of Oct4NBr in 200 mL of toluene in a 500 mL Erlenmeyer flask and stirred for 30-40 min, the water phase has changed from yellow to colorless and the toluene layer from colorless to a dark red. The aqueous layer was removed and 3.43 mL (26 mmol) of phenylethanethiol was added, stirring until the solution became colorless, signaling reaction of the gold salt with thiol to form a soluble AuIthiol polymer. The flask was cooled to 0 °C in an ice water bath, and 3.80 g (100 mmol) of NaBH4 that had been dissolved in 60 mL of ice cold water was then added to the AuIthiol solution with vigorous stirring. A black color, indicative of MPC formation, was immediately produced in the solution mixture, and a gas is evolved. The reaction temperature was maintained at 0 °C in the ice bath and the biphasic mixture vigorously stirred for 24 h, after which the water layer was separated and the black organic phase layer was washed thrice with 50 mL of Nanopure water and rotary evaporated to produce a black oil. Two hundred milliliters of acetonitrile was added to the flask, and the mixture was stirred for 5 min and allowed to stand overnight. Part of the MPC product dissolved, leaving an acetonitrile-insoluble black powder that was collected on a medium porosity glass frit and washed generously with acetonitrile. Typically ∼1 g of this material, which is mainly Au140(PhC2S)53, is isolated. It is soluble in toluene, methylene chloride, and THF but not acetone, acetonitrile, or ethanol. The acetonitrile solution was rotary evaporated and the solid suspended in ethanol and collected on a fine porosity fritted glass filter, washing generously with ethanol, providing 0.20-0.25 g of a brown powder, Au38(PhC2S)24. 1H NMR of the acetonitrile-soluble brown powder generally showed the presence of residual Oct4N+ (see later). Removal of this impurity was achieved by dissolving 0.3 g of MPC in 2 mL of CH2Cl2 and precipitating by adding ethanol or 80/20 v/v ethanol water, collecting the precipitate on a fine glass frit. The sample of Au38(PhC2S)24 sent for elemental analysis had not been so purified, however, and since the 1H NMR showed about 1.5 equiv of Oct4N+, a working formula of Au38(PhC2)24(Oct4NBr)1.5 was assumed for comparison to the elemental analysis results (% weight). Anal. Calcd: S, 6.6 (mass); C, 24.8; Au, 64.6. Found: S, 7.0; C, 25.2; Au, 62.0. The S:Au mole ratio was as follows: expt, 0.69; calc, 0.63. For the material to which an average formula Au140(PhC2)53 is assigned: Anal. Calcd: S, 4.9 (mass); C, 14.6%; Au, 79.0%. Found: S, 5.6 (mass); C, 15.9; Au, 72.1. The S:Au mole ratio was as follows: expt, 0.47; calc, 0.38. A tabular summary of thermogravimetric and elemental analysis results is included in Supporting Information. In some preparations, the Au38(PhC2S)24 portion was contaminated with significant amounts of Au140(PhC2S)53. The smaller MPC was isolated by treating the mixture with acetonitrile, then filtering and evaporating the filtrate, and repeating this procedure until the Au38(PhC2S)24 product was verified to be pure by 1H NMR. Carrying out the Brust reaction at room rather than lowered temperature produces a large proportion of polydisperse MPCs in the Au309 size range41 that can be roughly separated from the now-lesser Au140 fraction by the former’s low solubility in acetone.
Langmuir, Vol. 20, No. 5, 2004 1947 Single Phase Procedure: A flame-dried Schlenk flask containing ∼15 g of activated 3 Å sieves was prepared. To avoid consumption of superhydride by water, all efforts were made to maintain the dryness of the Schlenk flask contents throughout the procedure. Fifty milliliters of anhydrous THF was added via syringe followed by 1.94 g (3.5 mmol) of Oct4NBr and 1.53 g (3.5 mmol) of HAuCl4‚3H2O. This mixture was stirred for 30 min, and then the red solution, with the excess water from the gold salt removed, was separated from the sieves via syringe and transferred to another flame-dried Schlenk flask. After adding 1.44 mL, 10.2 mmol of phenylethanethiol, the solution changed to a light yellow over 30 min. This solution was then cooled to -78 °C on a dry ice/acetone bath, and 44 mL of a 1 M lithium triethylborohydride solution in THF (44 mmol), also cooled to -78 °C, was added over 30 s with vigorous stirring. The black solution was stirred for 2 h, excess phenol was added to quench the remaining superhydride, and the flask was allowed to warm to room temperature. The flask contents were added to a separatory funnel and washed thrice with 30 mL of brine, dried over sodium sulfate, filtered, and rotary evaporated to a black oil. One hundred milliliters of acetonitrile was added to the product, and the mixture was allowed to stand overnight. A 0.48 g quantity of a black precipitate (acetonitrile-insoluble Au140(PhC2S)53) was collected on a medium porosity glass frit. The filtrate solution was rotary evaporated and allowed to stand under ethanol overnight, then a brown solid (Au38(PhC2S)24, 0.21 g) was collected on a medium porosity glass frit. In this procedure, although the HAuCl4‚3H2O salt dissolves in THF without the aid of Oct4NBr, omitting the latter from the synthesis results in a white precipitate upon the addition of thiol to the gold solution that turns to a light purple upon the addition of superhydride. The solution turned black upon warming to room temperature, was quenched with phenol, and resulted in insoluble clusters. Ligand Exchange Reactions. Thiolates of monomethylterminated thio-poly(ethylene glycol) (MePEGS-, PEG MW ) 365),23 decanethiol and hexanethiol were incorporated into the monolayer shell of Au38(PhC2S)24 by ligand-place exchange reaction.23 In a typical procedure, an excess of ligand (typically 100-150 µL) was added to 45 mg of MPC in 25 mL of THF and the solution was stirred for 4 days. The solvent was rotary evaporated, and the flask contents were rinsed several times with heptane for MePEGS- and with ethanol for the alkanethiolate ligands, until the R-thiol protons of free ligand were no longer observable in 1H NMR. The monolayer composition was determined (Figures S-3 through S-6) by decomposing the MPC with iodine50 and analyzing the liberated ligands as disulfides by 1H NMR.
Results and Discussion Synthesis and Characterization. General aspects of the synthesis (details are in the Experimental Section) will be discussed first; analytical evidence for the Au38(PhC2S)24 MPC composition is then presented. The synthesis of the very small core phenylethanethiolate-protected MPCs was executed with the two-phase Brust procedure22,38 and also by a single-phase procedure. Both involve dissolving chloroaurate in an organic phase (phase-transferring to toluene from water in the Brust procedure, or directly dissolving in tetrahydrofuran) in the presence of Oct4NBr, followed by reduction with phenylethanethiol to a AuI form and then further to Au0. The sizes of the resulting Au MPC cores are determined in part by the kinetic competition between core growth (following the initial nucleation) and passivation by the protecting thiolate monolayer and in part by the thermodynamic structural preference to form closed-shell, “magic number” cores.18 The competition is dependent on the reaction temperature and the relative reactant concentrations.22 The concentrations used here, ∼3:1 thiol (50) Templeton, A. C.; Hostetler, M. J.; Kraft, C. T.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 1906.
1948
Langmuir, Vol. 20, No. 5, 2004
to gold and 10:1 reducing agent to gold mole ratios, are the same as those developed for synthesis of MPCs containing significant quantities of 1.6 nm core diameter alkanethiolate-coated MPCs.22,37 Increasing the thiol:gold ratio (up to g10:1) did not change the product distribution significantly. The more significant synthetic change was the lowering of reaction temperature (to 0 and -78 °C) on the premise that core growth kinetics would be retarded more than passivation kinetics, thus producing larger portions of smaller MPCs. There were hints of the possible success of this tactic in an earlier paper.22 Indeed, the 0.20-0.25 g yield of the brown, powdery Au38(PhC2S)24 MPC obtained here is ∼20% based on moles of the chloroaurate reactant, with the other product being e1 g of a black, polydisperse material containing a significant fraction of an MPC, average composition Au140(PhC2S)53. However, while lowered temperature produced a significant yield of the very small Au38(PhC2S)24 MPC, its solubility properties were the most vital factor in isolating it from the larger portion of Au140(PhC2S)53. The Au38(PhC2S)24 proved to be soluble in acetonitrile, acetone, benzene, and halogenated solvents, while the larger black powder MPC product(s) are soluble in the same solvents, except for acetonitrile. The previously isolated but larger41 MPC, Au309(PhC2S)91 (average formula), was also insoluble in acetonitrile (and acetone). The acetonitrile solubility of the Au38(PhC2S)24 product was a key factor, allowing its dissolution from the larger reaction product mass and isolation in quite monodisperse form. Notably, neither product fraction was soluble in hexane or heptane, in sharp contrast to the ready solubility of alkanethiolate-protected MPCs. Our experience with MPCs has been that the smaller core MPCs tend to be more polar and more soluble in polar solvents,15,17 presumably owing to a polar Au-S interface, and transmonolayer dispersion forces and/or increasing solvent penetration into the MPC monolayer (at small core radii). The small core Au38(PhC2S)24 MPC is very likely also more polar than its larger cousins. However the nature of the protecting monolayer has an even stronger influence on solubility properties, as we often see by solubility changes resulting from ligand place-exchange reactions (at constant core diameter). We believe that a favorable solvation of the phenylethyl grouping by acetonitrile must augment the core size polarity effect in the case of Au38(PhC2S)24. Indeed, we have since isolated39 what appears to be a hexanethiolate-protected Au38 MPC that is not acetonitrile soluble. These results show that the solvation characteristics of the thiolate monolayer ligand must be considered in searches for new MPC core sizes. Another possible factor in the efficient production of Au38(PhC2S)24 MPCs is the bulkiness of the aryl ring lying only two carbons away from the thiolate binding site, making it a more efficient passivator of further core growth than are alkanethiolate ligands. The closeness of the aryl ring also allows for favorable orbital overlap of its π electrons with those of the sulfur atom, resulting in enhanced nucleophilicity (i.e., Au binding). The overlap is apparent in an energy minimized structure determined in an MM1 calculation51 showing a preferred coalignment of the sulfur and aromatic π-electron orbitals. This is speculative, however, since no experimental evidence is available to support this electronic interaction. Our approach22 to assessing average molecular formulas for MPCs has combined analytical measurements of the organic mass fraction of the MPC (determined with TGA), elemental analysis, and imaging of the average MPC core (51) Determined using CS ChemBats3D V3.5.
Donkers et al.
Figure 1. TEM images and corresponding histograms of the core diameters of the acetonitrile-insoluble fraction, average Au140(PhC2S)53 (top) and acetonitrile-soluble fraction, Au38(PhC2S)24, lower.
diameter with transmission electron microscopy (TEM). Representative TEM images and corresponding core size histograms are shown in Figure 1. Each histogram summarizes manual and computer-aided analyses of the core diameters of more than 400 different nanoparticles. The size difference between the acetonitrile soluble and insoluble products is qualitatively evident, as is the greater uniformity of the smaller material. The average 1.12 nm diameter of the Au38(PhC2S)24 MPC core is close to the TEM instrument’s resolution limit. HPLC of this material shows23 a main peak comprising 94% of the total sample; its voltammetry39 is clean and generally without extraneous peaks. From these other observations, we believe that the Au38 core size is more uniform than the TEM histogram would suggest. As for the acetonitrile-insoluble product, its average 1.67 core size (Figure 1, top) is appropriate for the average formulation Au140(PhC2S)53, but the material is obviously rather polydisperse, similar to previously isolated raw samples of alkanethiolate MPCs.22 The analytical measurements are given in the Experimental Section and in Table S1. For the acetonitrilesoluble, isolated brown powder, the 0.69 S:Au mole ratio determined by elemental analysis is close to the theoretical value (0.63) for Au38(PhC2S)24. This formula is also consistent with the TGA analysis, in which the thiolate ligands are volatilized as disulfides (along with any adsorbed halide or Oct4NBr impurity), leaving a nonvolatile Au metal residue. For Au38(PhC2S)24 the theoretical and experimental total organic mass percentages are 33.4 and 32.8%, respectively. The latter analysis is also
Monolayer Protected Clusters
consistent with the assumption (backed by elemental analysis data, see ref 17) of a Au38(CH3S)24 formulation by Landman36 in a theoretical study. A similar analysis of the black, acetonitrile-insoluble MPC fraction gives an experimental S:Au ratio of 0.47 as compared to the theoretical 0.38, and an experimental organic mass fraction of 24.2 (theoretical 20.8%). The average experimental core diameter determined by TEM 1.6(7) is consistent with a 140 atom core (1.6 nm). The differences between experiment and theory in the TGA and elemental analysis results are biased in a direction (higher Au content) consistent with the abundance of larger core MPCs evident in the Figure 1 histogram. The polydispersity of this average Au140(PhC2S)53 material is further evident by the absence of quantized double layer charging peaks in its voltammetry (not shown); these features are known to become evident only in much more monodisperse samples.37 The TEM resolution is inadequate to judge the shape of the Au38(PhC2S)24 MPC core. We assume a face-centered cubic structure with a truncated octahedron morphology, in accordance with calculations36 for a methanethiolateprotected Au38 nanoparticle. The core contains a central six-atom octahedron and 24 corner surface atoms and has an average diameter of 1.1 nm, identical to the TEM histogram average. For the average Au140(PhC2S)53, we assume the same truncated octahedron geometry and structure as has been assigned to the related Au140(hexanethiol)53 MPC.18,22,36,52,53 The good yield and efficient isolation of Au38(PhC2S)24 are highly favorable considering the previous difficulties of producing large quantities of monodisperse MPCs with core diameters e1.6 nm. This new material represents a new route to the practical production of 1.1 nm MPC cores with other monolayer shells. 1H and 13C NMR. Previous 1H NMR studies22,54,55 of alkanethiolate monolayers on MPCs have described significant peak broadening compared to the free ligands. Three main reasons have been given: (a) the tight packing of protons close to the Au core causes rapid spin-spin relaxation from dipolar interactions; (b) there are different chemical shifts for ligands attached at different thiolate binding sites on the Au core surface (vertexes, edges, terraces), and the chemical shift differences also vary with core size and structure; and (c) the size-dependent rotation diffusion of the clusters leads to size-dependent spinspin relaxation broadening. The latter has been shown to scale with MPC core radii.22 1H NMR spectra of the acetonitrile-insoluble and soluble MPC fractions are shown in spectra a and c of Figure 2. The spectrum of the former fraction (average composition ca. Au140(PhC2S)53, Figure 2a) shows the characteristic peak broadening seen before22 for alkanethiolate MPCs. The free ligand spectrum is shown in Figure 2b. The Au38(PhC2S)24 MPC (Figure 2c) is a much smaller entity than the Au140 nanoparticle and exhibits resonances nearly as sharp as those of the free ligand (Figure 2b). The ligand’s proton chemical shifts are, however, substantially altered on the MPC. The ethylene peaks of the free ligand (B1 and B2) are changed to a broad featureless peak (C1) and (52) Garzo´n, I. L.; Rovira, C.; Michaelian, K.; Beltra´n, M. R.; Ordejo´n, P.; Junquera, J.; Sa´nchez-Portal, D.; Artacho, E.; Soler, J. M. Phys. Rev. Lett. 2000, 85, 5250-5251. (53) Garzo´n, I. L.; Michaelian, K.; Beltra´n, M. R.; Posada-Amarillas, A.; Ordejo´n, P.; Artacho, E.; Sa´nchez-Portal, D.; Soler, J. M. Phys. Rev. Lett. 1998, 81, 1600-1603. (54) Badia, A.; Lennox, R. B.; Reven, L. Acc. Chem. Res. 2000, 33, 475. (55) Badia, A.; Gao, W.; Singh, S.; Demers, L.; Guccia, L.; Reven, L. Langmuir 1996, 12, 1262.
Langmuir, Vol. 20, No. 5, 2004 1949
two slightly broadened triplets (C2 and C3), and the aryl resonances change to a complex pattern centered at about δ ) 7.2 ppm. The ratio of the aryl proton integral to the sum of peaks C1, C2, and C3 is 1.33, close to the expected 5/4 ) 1.25 ratio. Interestingly, the ratios for the C1, C2, and C3 peaks are 2:3:1. Figure 3 shows the 1H decoupled 13C NMR spectra for the aromatic carbons of the Au38(PhC2S)24 MPCs. Again, and unlike previous 13C NMR spectra of alkanethiolate MPCs,55,56 the resonances are well resolved and relatively unbroadened. Comparing the spectrum of as-prepared Au38(PhC2S)24 MPCs in Figure 3b to that of the free thiol monomer, 2-phenylethanethiol (Figure 3a), we see that the four distinct aromatic carbon peaks of the thiol are somewhat shifted and are now doublets. These results show that there are two distinct monolayer ligand environments on the Au38(PhC2S)24 MPCs. Similar doublet peaks are found for the ethylene carbons (not shown). The proton and 13C NMR spectra of the Au38(PhC2S)24 MPCs are consistent with one another by showing more complex spectra than the free thiol ligand. It seems clear from these results that substantial differences exist in the chemical environment of ligands bound to the MPC cores. The source of these differences is not yet evident. The two most obvious possibilities to explain the two sets of peaks in the 13C NMR spectrum (Figure 3b), are that (a) there are two distinctly different kinds of binding sites (and consequent chemical shifts) on the Au38 core surface and/or that (b) there are two (at least) nearly identically sized core structures (or core isomers) present with differing chemical environments for ligands on their surfaces. These possibilities relate of course to whether the Au38(PhC2S)24 MPCs are truly “monodisperse” nanoparticles. Our previously reported23 reversed-phase HPLC analysis of Au38(PhC2S)24 MPCs showed a single elution peak, suggestive of a high degree of uniformity of the nanoparticles. An analogous reversed-phase HPLC analysis of a related Au38(hexanethiolate)24 MPC preparation39 also gives a single elution peak, but a higher resolution double-column HPLC experiment reveals several peaks for closely related MPCs with essentially identical ∼Au38 cores. The possibility of a mixture of closely related Au38 (b) cannot be ruled out at this time. Returning to Figure 2c, additional resonances appear at a chemical shift (3.05 ppm) between the C2 and C3 protons, corresponding to R-CH2 protons of the (Oct4)N+ phase transfer agent employed in the synthesis. The area of these minor peaks amounts to ∼1.5 equiv of the (Oct4)N+ cation. (Other (Oct4)N+ proton peaks at higher field are shown in the raw spectrum, Figure S-2a.) Recently the Schiffrin laboratory reported57 the removal of this surfactant from larger core dodecanethiolate-coated MPCs using Soxhlet extraction and attributed its presence as a counterion of halide adsorbed on the Au core and/or attached by favorable alkyl chain interactions with the dodecanethiolate ligands. Halide adsorption on Au surfaces is known28,29,58-63 to occur, and we favor this (56) Terrill, R. H.; Postlethwaite, T. A.; C., C.; Poon, C.; Terzis, A.; Chen, A.; Hutchison, J. E.; Clark, M. R.; Wignall, G. D.; Londono, J. D.; Superfine, R.; Falvo, M.; Johnson Jr., C. S.; Samulski, E. T.; Murray, R. W. J. Am. Chem. Soc. 1995, 117, 12537. (57) Waters, C. A.; Mills, A. J.; Johnson, K. A.; Schiffrin, D. J. Chem. Commun. 2003, 540. (58) Watanabe, S.; Sonobe, M.; Arai, M.; Tazume, Y.; Matsuo, T.; Nakamura, T.; Yoshida, K. Chem. Commun. 2002, 2866. (59) Ikezawa, Y.; Terashima, H. Electrochim. Acta 2002, 47, 178. (60) Kerner, Z.; Pajkossy, T. Electrochim. Acta 2002, 47, 2055. (61) Sa´nchez-Corte´s, S.; Garcı´a-Ramos, J. V. Surf. Sci. 2001, 473, 133. (62) Melendres, C. A.; Hahn, F. J. Electroanal. Chem. 1999, 463, 258.
1950
Langmuir, Vol. 20, No. 5, 2004
Donkers et al.
Figure 2. 1H NMR spectra in CD2Cl2 solutions of (a) Au140(PhC2S)53 MPCs, (b) 2-phenylethanthiol, and (c) Au38(PhC2S)24 MPCs.
explanation since the hydrophobic interactions should be relatively weaker for the short aryl ligands used. The (Oct4)NBr impurity could be removed (as seen by 1 H NMR) by simple chromatography of the reaction mixture on a silica gel column, which also serves to clean out any larger size MPCs from the Au38(PhC2S)24 sample. This treatment strikingly alters both the 1H and 13C NMR spectra from those in Figures 2c and 3b, without changing the relative 1H aryl:methylene peak area ratio, or the UVvis and FTIR spectra, or the HOMO-LUMO gap seen in the cyclic voltammetry.39 From the fact that the electrochemical rest potentials were more positive in solutions of Au38(PhC2S)24 after silica gel treatment, and from other evidence, we suspect that the chemical shift changes represent effects of the Au core becoming less electron rich as a result of an oxidation process. This phenomenon is under study and will be described more fully later.39 UV-Visible Spectra. UV-vis spectroscopy of metal nanoparticles has received considerable attention.13,16,48,64-68 (63) Bellier, J. P. J. Electroanal. Chem. 1982, 140, 391. (64) Barnett, R. N.; Cleveland, C. L.; Ha¨kkinen, H.; Luedtke, W. D.; Yannouleas, C.; Landman, U. Eur. Phys. J. D 1999, 9, 95-104. (65) Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C. J. Phys. Chem. B 2003, 107, 668. (66) Mie, G. Ann. Phys. 1908, 25, 377. (67) Kreibig, U.; Genzel, L. Surf. Sci. 1985, 156, 678.
Noteworthy is the account by Whetten and co-workers13,16 of the evolution of spectra of Au MPCs with cores ranging from 3.2 to 1.7 nm. Generally, the strong absorption in optical spectra of larger Au MPCs decays smoothly from UV through visible wavelengths, interrupted only by a characteristic broad surface plasmon resonance (SPR) band centered at ∼2.5 eV. This latter band decreases in intensity for smaller Au MPC cores, becoming indistinctto-absent for core diameters below 2 nm. The low energy absorption onset of the spectra is attributed to an interband gap 5d f 6sp transition. For the smallest MPCs, the smooth UV-vis absorption changes to a steplike set of transitions, as the electronic band structure begins to condense into a more discrete energy level structure. Indeed, all of these features can be seen in spectra of various PhC2S- coated Au MPCs, as shown in Figure 4b. The top curve, the spectrum of a 2.1 nm core diameter Au309(PhC2S)91 nanoparticle (made as described previously41) exhibits the typical SPR band of a larger core MPC, whereas this band is absent in spectra of the smaller core diameter Au140(PhC2S)53 and Au38(PhC2S)24 MPCs. The latter (lower curve, Figure 4b) displays the steplike fine structure characteristic of a noncontinuum electronic (68) Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles; Wiley: New York, 1983.
Monolayer Protected Clusters
Figure 3.
1
H-decoupled
Langmuir, Vol. 20, No. 5, 2004 1951
13
C NMR spectra in CD2Cl2 solutions of (a) 2-phenylethanthiol and (b) Au38(PhC2S)24 MPCs.
Figure 4. UV-vis spectra of PhC2MPC in CCl4 solution: (a) derivative absorption spectrum of Au38(PhC2S)24 MPC; (b) UVvis spectra of (top to bottom) (i) Au309(PhC2S)91 MPCs, (ii) Au140(PhC2S)53 MPCs, and (iii) Au Au38(PhC2S)24 MPCs. The lower spectra were normalized by choosing the concentrations of the samples to have A ∼ 1 at 4 eV.
structure, similar to other very small MPCs.13,16,31,32,35 The fine structure is illustrated more clearly in the derivative spectrum of Figure 4a. The overall absorbance of Au38(PhC2S)24 decreases with increasing wavelength, reaching a small absorption maximum at 1.61 eV and an absorption edge at ∼1.3 eV that we interpret as the HOMO-LUMO gap energy. Electrochemical measurements on the Au38(PhC2S)24 MPCs are consistent with the spectroscopy and will be discussed in detail in another publication.39
Ligand Place Exchange Reactions. Incubation of MPCs in solutions with free thiols results in ligand place exchange and MPCs with mixed-ligand monolayers.69,70 The place exchange reaction is important since it can produce multifunctional nanoparticles, allows the introduction of ligands with functional groups not compatible with the reductive MPC synthetic conditions, and permits systematic control over the proportion of the second ligand in the MPC monolayer. MPCs with mixed monolayers of alkanethiolate and mercaptoalkanoic acid ligands have been self-assembled into network polymer films on electrodes.71 We have studied72,73 the dynamics of the ligand place exchange, using Au140(hexanethiolate)53 MPCs; the reactions show a rich complexity of behavior including wide variations in surface site reactivitysdue apparently to chemical differences between vertex, edge, and terrace core sites. The variation of surface chemistry is consonant with the variations in NMR chemical shifts described in Figures 2 and 3 above. The place exchange dynamics of positively charged MPCs also reveal the presence of multiple mechanistic pathways for exchange. We have postulated that the vertex and edge sites on the Au core, being defect sites in the classical sense, are the most reactive in place exchange, while terrace sites are relatively unreactive. On that account, one would expect that ligand place exchange for Au38(PhC2S)24, whose core surface is nearly all vertex/edge36 should be both more (69) Ingram, R. S.; Hostetler, M. J.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 9175. (70) Hostetler, M. J.; Green, S. J.; Stokes, J. J.; Murray, R. W. J. Am. Chem. Soc. 1996, 118, 4212. (71) Zamborini, F. P.; Leopold, M. C.; Hicks, J. F.; Kulesza, P. J.; Malik, M. A.; Murray, R. W. J. Am. Chem. Soc. 2002, 124, 8958-8964. (72) Song, Y.; Murray, R. W. J. Am. Chem. Soc. 2002, 124, 70967102. (73) Song, Y.; Huang, T.; Murray, R. W. J. Am. Chem. Soc. 2003, 125, 11694-11701.
1952
Langmuir, Vol. 20, No. 5, 2004
facile and prone to be driven to completion than is the case for larger core MPCs. Indeed, both partial and complete exchange of the PhC2S- ligand with a thiolated poly(ethylene) glycol is possible, resulting in mixed monolayer Au38(PhC2S)5(MPEG365)19 and Au38(PhC2S)13(MPEG135)11 MPCs and, using two consecutive exchange reactions, completely exchanged Au38(MPEG365)24. These materials are nanoparticle-based molecular melts, and it is possible to carry out voltammetric investigations of inter-MPC electron transfers in them.23 Ligand exchange reactions of Au38(PhC2S)24 MPC resulted in greater than 85% incorporation of dodecanethiolate and hexanethiolate ligands when the corresponding thiols were used at a 20fold molar excess relative to the PhC2S- ligands present, with quantitative exchange being achieved with a second ligand exchange reaction. Degradation of the Au38 core structure does not accompany place exchange, based on the optical absorption spectrum and voltammetry.23 This remarkable core maintenance indicates this MPC can be used as a “seed” structure to make other small-core MPCs, analogous to
Donkers et al.
other reported work where alkanethiols have been used31,35 to replace triphenylphosphine ligands on Au11 and Au55 nanoparticles. This fact, and the efficient production of large quantities of relatively monodisperse MPCs, is leading us to further studies of Au38 nanoparticles that will be described in coming publications. Acknowledgment. This research is supported in part by grants from the National Science Foundation and the Office of Naval Research. R.L.D. acknowledges fellowship support from the Canadian Natural Sciences and Engineering Research Council. The authors thank Yang Song and Marc Terhorst for NMR assistance, W. Ambrose for help with TEM images, and L. Zannoni for TGA analyses. Supporting Information Available: Thermogravimetric analysis traces, results of elemental and TGA analysis, and NMR spectra. This material is available free of charge via the Internet at http://pubs.acs.org. LA035706W
