Chiral Functionalization of Optically Inactive Monolayer-Protected

Feb 6, 2008 - Nicole Cathcart , Pretesh Mistry , Christy Makra , Brendan Pietrobon , Neil Coombs , Masoud Jelokhani-Niaraki and Vladimir Kitaev. Langm...
6 downloads 0 Views 545KB Size
Download PDF
Langmuir 2008, 24, 2759-2766

2759

Chiral Functionalization of Optically Inactive Monolayer-Protected Silver Nanoclusters by Chiral Ligand-Exchange Reactions Naoki Nishida, Hiroshi Yao,* and Keisaku Kimura Graduate School of Material Science, UniVersity of Hyogo, 3-2-1 Koto, Kamigori-cho, Ako-gun, Hyogo 678-1297, Japan ReceiVed October 27, 2007. In Final Form: December 14, 2007 We report the ligand-exchange reaction between the optically inactive racemic penicillamine monolayer on a silver nanocluster surface and enantiopure D- or L-penicillamine dissolved in solution. Emergence of the identical band positions in the gel electrophoresis separation assures the presence of size-invariant silver nanoclusters (1.05 and 1.30 nm in core diameter) during the ligand-exchange reaction and allows us to further examine the optical/chiroptical properties of these nanoclusters. Consequently, chiral functionalization of the achiral silver nanoclusters has been demonstrated, yielding large Cotton effects in metal-based electronic transitions with an almost mirror-image relationship between the enantiomeric compounds. The ligand-exchange experiments as well as the normal syntheses of the silver nanoclusters revealed that their absorption profiles and anisotropy factors were strongly dependent on the enantiomeric purity (or enantiomeric excess) of surface chiral penicillamine, so that (several-fold) larger chiroptical responses of the silver nanoclusters as compared to those of the analogous gold clusters with a comparable size could be induced by the metal core deformation or rearrangement along with a universally influential vicinal contribution from the chiral ligand field.

Introduction Metal nanoclusters with the diameter reduced to the order of the metal’s Fermi wavelength are a new class of materials made up of tens to hundreds of atoms, and they can be seen in intermediates between single atoms/molecules and bulk materials. Specifically, synthesis and functionalization of the size-controlled, ligand-stabilized (or monolayer-protected) metal nanoclusters have attracted strong attention in nanochemistry1 because such metal nanoclusters offer a significant possibility to fabricate building blocks for intriguing applications in catalysis, photonics, biosensing, and molecular electronics.2 These building blocks are expected to have unique size-dependent physical and chemical properties (quantum size effects) that have been documented mostly in studies of gold nanoclusters; that is, gold clusters with core diameters smaller than 1-2 nm are small enough to possess discrete electronic states not found in larger particles.3 In addition, some theoretical studies show that the lowest energy isomers of such small gold clusters (e.g., Au28(SCH3)16) are able to have a disordered low-symmetry geometry and a chiroptical effect.4 Passivation of small metal nanoclusters with chiral molecules also results in unique electronic and chiroptical responses that * Corresponding author. Tel.: +81-791-58-0160; fax: +81-791-58-0161; e-mail: [email protected]. (1) (a) Fendler, J. H. Chem. Mater. 2001, 13, 3196. (b) Daniel, M.-C.; Astruc, D. Chem. ReV. 2004, 104, 293. (2) (a) 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 (Washington, DC, U.S.) 1998, 280, 2098. (b) Schwerdtfeger, P. Angew. Chem., Int. Ed. 2003, 42, 1892. (c) Whetten, R. L.; Shafigullin, M. N.; Khoury, J. T.; Schaaff, T. G.; Vezmar, I.; Alvarez, M. M.; Wilkinson, A. Acc. Chem. Res. 1999, 32, 397. (d) Huang, T.; Murray, R. W. J. Phys. Chem. B 2001, 105, 12498. (3) (a) Kubo, R. J. Phys. Soc. Jpn. 1962, 17, 975. (b) Sanchez, A.; Abbet, S.; Heiz, U.; Schneider, W.-D.; Ha¨kkinen, H.; Barnett, R. N.; Landman, U. J. Phys. Chem. A 1999, 103, 9573. (c) Menard, L. D.; Gao, S.-P.; Xu, H.; Twesten, R. D.; Harper, A. S.; Song, Y.; Wang, G.; Douglas, A. D.; Yang, J. C.; Frenkel, A. I.; Nuzzo, R. G.; Murray, R. W. J. Phys. Chem. B 2006, 110, 12874. (4) (a) Garzo´n, I. L.; Reyes-Nava, J. A.; Rodrı´guez-Herna´ndez, J. I.; Sigal, I.; Beltra´n, M. R.; Michaelian, K. Phys. ReV. B: Condens. Matter Mater. Phys. 2002, 66, 73403. (b) Garzo´n, I. L.; Beltra´n, M. R.; Gonza´lez, G.; Gutı´errezGonza´lez, I.; Michaelian, K.; Reyes-Nava, J. A.; Rodrı´guez-Herna´ndez, J. I. Eur. Phys. J. D 2003, 24, 105. (c) Roma´n-Vela´zquez, C. E.; Noguez, C.; Garzo´n, I. L. J. Phys. Chem. B 2003, 107, 12035.

are unlike those of the component parts.5-7 For example, we reported large chiroptical responses of size-selected gold nanoclusters (0.6-1.8 nm) protected by a pair of enantiomers of chiral penicillamine.7 These optically active nanoclusters exhibited a clear mirror-image relationship in their circular dichroism (CD) signals, with the anisotropy factors decreasing with the cluster size increase. We also succeeded in preparing chiral penicillamine-protected silver nanoclusters with core sizes ranging 1-3 nm.8 In comparison with the gold nanoclusters bearing the same ligands, a steep enhancement of optical activity was observed in the silver nanoclusters when the diameter became less than ∼1.5 nm. More interestingly, the silver nanoclusters covered with penicillamine racemate exhibited different absorption profiles from those covered with enantiopure (D- or L-) penicillamine in the same core size.8 Although the induction of chirality on the clusters and its mechanism of clarification are definitely important, the origin of observed optical activity in metal-based electronic transitions is still fragmentary, probably due to few examples of well-defined optically active clusters and the lack of geometrical information. Meanwhile, it has been elucidated that the reactivities of monolayer-protected metal nanoclusters respond to both the metal core and the nature of the monolayer ligand shell.9 An important example is the ligand-exchange reaction. Surface ligands on a nanoparticle surface can be exchanged by stronger ligands, and this ligand-exchange reaction provides an important means for the chemical functionalization of nanoparticles and greatly extends (5) (a) Schaaff, T. G.; Knight, G.; Shafigullin, M. N.; Borkman, R. F.; Whetten, R. L. J. Phys. Chem. B 1998, 102, 10643. (b) Schaaff, T. G.; Whetten, R. L. J. Phys. Chem. B 2000, 104, 2630. (6) (a) Gautier, C.; Bu¨rgi, T. J. Am. Chem. Soc. 2006, 128, 11079. (b) Tamura, M.; Fujihara, H. J. Am. Chem. Soc. 2003, 125, 15742. (c) Yanagimoto, Y.; Negishi, Y.; Fujihara, H.; Tsukuda, T. J. Phys. Chem. B 2006, 110, 11611. (d) Shemer, G.; Krichevski, O.; Markovich, G.; Molotsky, T.; Lubitz, I.; Kotlyar, A. B. J. Am. Chem. Soc. 2006, 128, 11006. (7) Yao, H.; Miki, K.; Nishida, N.; Sasaki, A.; Kimura, K. J. Am. Chem. Soc. 2005, 127, 15536. (8) Nishida, N.; Yao, H.; Ueda, T.; Sasaki, A.; Kimura, K. Chem. Mater. 2007, 19, 2831. (9) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, 33, 1906.

10.1021/la703351p CCC: $40.75 © 2008 American Chemical Society Published on Web 02/06/2008

2760 Langmuir, Vol. 24, No. 6, 2008

Nishida et al.

Figure 1. (a) Chemical structure of D- or L-penicillamine (D-Pen or L-Pen). An asymmetric carbon is denoted by an asterisk in each formula. (b) Photographs of the PAGE separation. Bands 1 and 2 are common in all samples and thus labeled in the order of mobility of the nanoclusters: (i) Ag-D-Pen, (ii) Ag-L-Pen, (iii) Ag-rac-Pen, (iv) Ag-D50-Pen, (v) Ag-L50-Pen, (vi) D-Pen ligand-exchanged sample, and (vii) L-Pen ligandexchanged sample. The average diameters of separated 1 and 2 already have been estimated to be 1.05 and 1.30 nm, respectively. In iv-vii, the right-side images are “shadowgraphs” on a fluorescent screen taken under UV-light (365 nm) irradiation to resolve the fractions more clearly.

the versatility of these systems.10 In the studies on thiol ligandexchange reactions, the following schemes have been revealed: (i) The reaction has a 1:1 stoichiometry, liberating one thiol from the original nanoparticle monolayer for every new thiolate incorporated into it, (ii) it is an associative reaction, that is, the incoming ligand penetrates the nanoparticle monolayer and in some manner replaces a bonded thiolate ligand, and (iii) it probably occurs at a higher rate on vertexes and edges of the core surface as opposed to core terrace sites. If such ligand-exchange reactions can be utilized for introducing chirality or optical activity onto the achiral monolayer-protected metal nanoclusters, they will offer a useful synthetic route for fabricating chirally functionalized nanoclusters. We report here the results of chiral ligand-exchange reactions between the racemate monolayer on silver nanoclusters with a core diameter smaller than 1.5 nm and enantiopure D- or L-penicillamine ligands. Functionalization of the large optical activity in the natively achiral silver nanoclusters was successfully achieved. Furthermore, through the normal syntheses of monolayer-protected silver nanoclusters using penicillamine as a surface capping agent as well as the ligand-exchange experiments, we find that both chiroptical and optical responses in the metalbased electronic transitions strongly depend on the enantiomeric excess (ee) of the D- or L-isomers on the nanocluster surface. On the basis of such ee-dependent properties, we discuss the implications of the chiral core structure on the silver nanocluster. Experimental Procedures Materials. Silver nitrate (AgNO3, 99.5%), sodium borohydride (NaBH4, > 90%), methanol (GR grade), and ethanol (GR grade) were received from Wako Pure Chemicals and used as received. Electrophoresis grade acrylamide, N,N′-methylenebisacrylaminde, tris(hydroxymethylaminomethane) (Tris), glycine, ammonium peroxodisulfate, and N,N,N′,N′-tetramethylethylenediamine were received from Nacalai Tesque and used without further purification. Enantiopure D- or L-penicillamine (abbreviated as D-Pen (99%) or L-Pen (99%), respectively) and racemic DL-penicillamine (abbreviated (10) (a) Hostetler, M. J.; Green, S. J.; Stokes, J. J.; Murray, R. W. J. Am. Chem. Soc. 1996, 118, 4212. (b) Song, Y.; Murray, R. W. J. Am. Chem. Soc. 2002, 124, 7096. (c) Song, Y.; Huang, T.; Murray, R. W. J. Am. Chem. Soc. 2003, 125, 11694. (d) Zhang, J.; Lakowicz, J. R. J. Phys. Chem. B 2005, 109, 8701.

as rac-Pen (97%)) were received from Aldrich and used as received. The chemical structures of D-Pen and L-Pen are shown in Figure 1a. Pure water was obtained by an Advantec GS-200 automatic waterdistillation supplier. Synthesis. Penicillamine-protected silver nanoclusters were prepared using a similar method to that reported previously.8 Briefly, 0.5 mmol of AgNO3 dissolved in water (0.121 M) and 0.5 mmol of penicillamine were at first mixed in methanol (100 mL), followed by the addition of a freshly prepared 0.2 M aqueous NaBH4 solution (25 mL) under vigorous stirring. After further stirring for 1.5 h, the addition of ethanol (200 mL) into the solution gave a brown crude precipitate. The precipitate was then thoroughly washed with ethanol. Finally, a nanocluster powder was obtained by a vacuum-drying procedure. Note that the samples prepared by this method are called normally synthesized silver nanoclusters. In the syntheses, we used five kinds of penicillamine with different ee values: (i) rac-Pen (0% ee), (ii) pure D-Pen (100% ee of D-isomer), (iii) pure L-Pen (100% ee of L-isomer), (iv) mixture of D-/L-Pen at a molar ratio of 3:1 (50% ee of D-Pen), and (v) mixture of D-/L-Pen at a molar ratio of 1:3 (50% ee of L-Pen). Each nanocluster sample prepared was called Agrac-Pen, Ag-D-Pen, Ag-L-Pen, Ag-D50-Pen, or Ag-L50-Pen, respectively. On the basis of our previous IR spectroscopic study, we revealed that penicillamine molecules chemisorb on silver surfaces as thiolates by forming an Ag-S bond (as revealed by the disappearance of the S-H stretch mode) and that the amino acid moiety was present in the carboxylate form (as revealed by the appearance of two characteristic peaks of COO- vibration).7,8 When conducting chiral ligand-exchange reactions, we used an optically inactive silver nanocluster sample as the starting material that was prepared in a similar procedure described previously but with a molar ratio of rac-Pen/Ag of 2:1. The reaction was conducted by stirring an aqueous solution (1.0 mL) of the as-synthesized racPen-protected silver nanocluster powder (10.0 mg) and enantiopure D- or L-Pen (5.0 mg) overnight at room temperature under an inert (N2-saturated) atmosphere. The uncapped optically active penicillamine residue could be excluded by polyacrylamide gel electrophoresis (PAGE) since the molecular penicillamine has a different (and large) mobility from the relevant silver nanoclusters. The nanocluster size fractionation was carried out using PAGE in a similar procedure reported previously.8 Briefly, the stacking and separating gels were at first buffered at pH 6.8 and 8.7 with the Tris-HCl solution, respectively. The sample solutions were then loaded onto a stacking gel top and eluted with the running buffer consisting of a glycine (192 mM) and Tris (25 mM) solution for 5

Optically InactiVe Au Nanoclusters h at a constant voltage mode (150 V). To extract the silver nanoclusters in aqueous solution, a part of the gel containing each fraction was carefully cut out and homogenized, followed by the addition of distilled water. The gel lumps were removed by centrifugation. Instrumentation. Absorption spectra were recorded using a Hitachi U-4100 spectrophotometer. CD spectra were recorded with a JASCO J-820 spectropolarimeter. Rectangular 1 cm cuvettes made of quartz were used for the measurements. The mean core sizes of fractioned silver nanoclusters were determined by a solution-phase small-angle X-ray scattering (SAXS) technique.8 The SAXS analyses were based on the assumption that spherical clusters are distributed with the simple Γ-distribution function.11

Results and Discussion PAGE Separation. The penicillamine-protected silver nanoclusters can be separated according to their size and charge by PAGE because of their negatively charged nature. Photographs of a typical PAGE separation under normal illumination for the as-synthesized Ag-rac-Pen, Ag-D-Pen, Ag-L-Pen, Ag-D50-Pen, and Ag-L50-Pen together with the ligand-exchanged samples are shown in Figure 1b. Note that right-side images in iv-vii in Figure 1b are “shadowgraphs” of their left-side images on a fluorescent screen taken under UV-light (365 nm and 4 W) irradiation to resolve the fractions more clearly.12 Many bands are observable in the gel for each sample, but the appearance of discrete bands suggests the presence of magic number compounds. We hereafter focus on two selected fractions denoted as 1 and 2 (their relative mobilities expressed as Rf are about 0.90 and 0.76, respectively) because (i) these bands were practically detectable (discernible) in all nanocluster samples at the almost identical positions with each other and (ii) our previous study demonstrated that the rac-Pen-protected silver nanocluster compounds in these bands exhibited slightly different absorption properties from those of the enantiopure D-/L-Pen-protected clusters.8 The identical band location (1 and 2) means that the silver nanoclusters to be fractioned are composed of similar size and chemical components.5-8 When distinguishing these compounds with regard to their stereochemistry, a suffix (rac, D, L, D50, or L50) is added at the end the compound number; for example, 1D50, 2D50, ... for the Ag-D50-Pen nanocluster sample. Moreover, a suffix e is added when the sample was obtained via the ligand-exchanged reaction; for example, 1eD, 2eD, ... for the D-Pen ligand-exchanged nanocluster sample. On the basis of our previous SAXS analyses as well as the excellent linear relationship between the logarithm of the average size and the Rf value, the core diameter of each fractioned cluster compound 1 or 2 was determined to be 1.05 or 1.30 nm, respectively.8,13 In any case, the appearance of the identical bands in every nanocluster sample (11) Sasaki, A. Riguku J. 2005, 22, 31. (12) Silver nanoclusters to be fractioned strongly absorb UV light at 365 nm due to the metal-based electronic transitions, so the irradiated light cannot pass through the lanes (bands) of silver nanoclusters in the plate gel. On the other hand, the residual light can pass through the gel and excite a fluorescent screen. Then, the lanes containing silver nanoclusters make “shadows” on the screen. (13) (a) Density of bulk silver is reported to be 10.5 g/cm3 (58.6 atoms/nm3) (Stiger, R. M.; Gorer, S.; Craft, B.; Penner, R. M. Langmuir 1999, 15, 790). Taking the silver core as a sphere with the bulk density covered with an outermost layer of close-packed silver atoms (13.92 atoms/nm2; this value can be estimated from the radius of a silver atom 0.144 nm), the cores of 1 and 2 contain ∼36 and ∼68 Ag atoms, of which ∼25 and ∼45 atoms lie on the nanocluster surface, respectively. Note that, for the estimation of the number of surface atoms, the skin is taken as the center of the outermost silver atoms (Terrill, R. H.; Postlethwaite, T. A.; Chen, C.-h.; Poon, C.-D.; Terzis, A.; Chen, A.; Hutchison, J. E.; Clark, M. R.; Wignall, G.; Londono, J. D.; Superfine, R.; Falvo, M.; Johnson, C. S., Jr.; Samulski, E. T.; Murray, R. W. J. Am. Chem. Soc., 1995, 117, 12537). (b) If we approximately assume a surface coverage of the thiol on the silver nanocluster to be a typically reported value of 8.5 × 10-10 mol/cm2 (5.1 molecules/nm2) (Mohtat, N.; Byloos, M.; Soucy, M.; Morin, S.; Morin, M. J. Electroanal. Chem. 2000, 484, 120; this packing density was calculated for the Ag(111) surface), 1 and 2 would have 9-10 and 16-17 thiol molecules on their surface, respectively.

Langmuir, Vol. 24, No. 6, 2008 2761

helps our further investigation of chiral ligand-exchange reactions and the relevant optical activity functionalization. Optical and Chiroptical Properties of Normally Synthesized Silver Nanocluster Compounds. UV-vis absorption spectra, their first derivative spectra, and patterns of the anisotropy factor (or g-factor), which is defined as the ratio of the molar dichroic absorption to the extinction coefficient, of 1D/1L/1rac and 2D/ 2L/2rac are shown in Figure 2a-c and 2d-f, respectively. Similarly, the optical/chiroptical properties of 1D50/1L50 and 2D50/ 2L50 are shown in Figure 3a-c and 3d-f, respectively. Note that since the 1D50 exhibited a broader fraction band than those of the others (see images(i-v in Figure 1b), we carefully cut and collected this band as similar as possible to the others to minimize the influence of this poor quality. The separated compounds 1D/1L/1D50/1L50 and 2D/2L/2D50/2L50 exhibited large Cotton effects, whereas 1rac and 2rac were optically inactive. Absorption Properties. In enantiopure D- and L-Pen-protected nanocluster compounds (1D/1L and 2D/2L), the absorption spectral shapes are almost identical within the same compound numbers, supporting that each numbered compound has a similar core size, geometry, and chemical composition.5,14,15 The fact is corroborated by the identical patterns of the first derivative spectra (Figure 2b,e). In addition, their spectra were rather structured, indicating that the electronic structures of 1 and 2 are in a quantum confinement regime and better viewed as nonmetallic clusters rather than metallic particles covered by thiols.16 In silver nanocluster compounds covered with penicillamine in a lower ee value (1rac/1D50/1L50 or 2rac/2D50/2L50), on the other hand, the spectral shapes involving peak or shoulder positions were slightly different from those of 1D/1L or 2D/2L, respectively. This interesting feature also could be confirmed by the patterns of the first derivative spectra of these compounds. For example, the first derivative spectra of 1D50/1L50 exhibit neither a hump at around 400 nm nor a trough at around 375 nm, while those of 1D/1L have a hump at ∼400 nm and that of 1rac a trough at ∼375 nm. See the elliptically marked regions in Figures 2b and 3b or Figures 2e and 3e, in which slight but distinct differences are shown in the optical responses among the respective nanocluster compounds.17 It should be noted that partial oxidation of the silver nanocluster surfaces may blur the spectral features. However, even if the partial oxidation is likely on the silver nanocluster surface, all cluster samples (1D/1L/1rac/1D50/1L50 and 2D/2L/2rac/2D50/2L50) are expected to be in an equally oxidized state under the same preparation conditions, so that the observed spectral differences are not due to this phenomenon. Hence, we can conclude that the ee value of the chiral surface ligand influences the optical properties of the silver nanoclusters in the same core size. This means that the core geometry of the penicillamine-protected silver nanoclusters is a function of the enantiomeric purity (or ee value) of the surface ligand. This behavior would be induced by statistical (or random) ligation of the thiols onto a silver cluster surface as has been proven for chiral cysteine adsorption on a gold surface.18 Interestingly, it appears that the absorption properties of the silver clusters do not depend on the handedness of the surface ligand. Chiroptical Properties. The ligand enantiomeric purity also had an influence on the Cotton effects of the present silver (14) Negishi, Y.; Nobusada, K.; Tsukuda, T. J. Am. Chem. Soc. 2005, 127, 5261. (15) (a) Sosa, I. O.; Noguez, C.; Barrera, R. G. J. Phys. Chem. B 2003, 107, 6269. (b) Noguez, C. Opt. Mater. 2005, 27, 1204. (16) Nobusada, K. J. Phys. Chem. B 2004, 108, 11904. (17) To further demonstrate the spectral differences, absorption spectra of the nanocluster compounds (1D/1L/1rac/1D50/1L50) were normalized at around 255 nm and compared with each other. See the Supporting Information for more detail. (18) Ku¨hnle, A.; Linderoth, T. R.; Hammer, B.; Besenbacher, F. Nature (London, U.K.) 2002, 415, 891.

2762 Langmuir, Vol. 24, No. 6, 2008

Nishida et al.

Figure 2. Panels a-c show absorption spectra, first derivative spectra, and anisotropy factors of 1D/1L/1rac, respectively. Panels d-f also show absorption spectra, first derivative spectra, and anisotropy factors of 2D/2L/2rac, respectively. For clarity, spectra of absorption of 1D/1L and 2D/2L, and first derivatives of 1L/1rac and 2L/2rac, are offset by adding (subtracting) a constant. In panels c and f, the red, green, and black curves represent the cases of D-Pen-, L-Pen-, and rac-Pen-protected silver nanoclusters, respectively.

nanoclusters (Figures 2c,f and 3c,f). Note that optically active penicillamine contributes to the CD signal only in an UV region shorter than ∼270 nm in wavelength and has a maximum anisotropy factor of ∼2.0 × 10-3 at ∼228 nm.7,8 The pairs of anisotropy factors of 1D/1L, 1D50/1L50, 2D/2L, and 2D50/2L50 show a nearly mirror-image relationship in the range of metal-based electronic transitions, exhibiting that enantiomeric ligands can produce the respective stereochemically controlled silver nanoclusters. This finding indicates that the separated silver nanoclusters have well-defined stereostructures even under a relatively low ee condition of the surface ligand. For the 1rac and 2rac compounds, no CD signals could be reasonably detected. The Cotton effect would be induced through the different (quantized) transitions and their interactions in the silver clusters.4,8 For further elucidation of the core size and ligand ee effects on the optical activity of the silver nanoclusters, comparison of the maximum anisotropy factors (termed as gmax-factors) is effective. Slight pattern differences among the same numbered compounds with a different ee value (1D/1L and 1D50/1L50 or

2D/2L and 2D50/2L50) come from those in their different absorption and CD spectra. The size dependence could be observed in that the anisotropy factor decreased as the nanocluster core size increased. More importantly, the gmax-factor is approximately proportional to the ligand ee value: In compound 1, for example, the gmax-factor was 1.5 × 10-3 for enantiopure D-/L-Pen-protected silver nanoclusters (1D and 1L), whereas it decreased to 7-8 × 10-4, scaled by a factor of ∼0.5, for the clusters covered with 50% ee penicillamine (1D50/1L50). In 2, a similar trend was observed in the gmax-factors (5-6 × 10-4 for 2D/2L and 2-2.5 × 10-4 for 2D50/2L50). This result indicates that enantiomeric purity of the surface ligand quantitatively controls the chiroptical responses in chiral penicillamine-protected silver nanoclusters. Optical and Chiroptical Properties of Ligand-Exchanged Silver Nanocluster Compounds. As shown in the PAGE separation images of the ligand-exchanged samples (Figure 1), a noticeable modification was observed in lanes, but the positions of 1 and 2 were invariant as the magic-numbered silver

Optically InactiVe Au Nanoclusters

Langmuir, Vol. 24, No. 6, 2008 2763

Figure 3. Panels a-c show absorption spectra, first derivative spectra, and anisotropy factors of 1D50/1L50, respectively. Panels d-f also show absorption spectra, first derivative spectra, and anisotropy factors of 2D50/2L50, respectively. Absorption spectra of 1D50 and 2D50, and first derivatives of 1L50 and 2L50, are offset by adding (subtracting) a constant. In panels c and f, the red and green curves represent the cases of D-Pen- and L-Pen-enantioenriched silver nanoclusters, respectively.

nanoclusters.19 UV-vis absorption spectra, first derivative spectra, and patterns of the anisotropy factors of 1eD/1eL and 2eD/2eL are displayed in Figure 4a-f. Absorption Properties. For 1eD/1eL and 2eD/2eL, their spectral shapes were found to be different from those of 1D/1L/1rac and 2D/2L/2rac but almost identical to those of 1D50/1L50 and 2D50/2L50, respectively. This can be confirmed by the patterns of the first derivative spectra shown in Figure 4b,e. On the basis of the result that the absorption property of the present silver nanoclusters is a function of the ee value of the surface chiral ligand, the high similarity of absorption spectra (and the corresponding first (19) (a) Actually, the fractioned bands 1eD/1eL and 2eD/2eL were somewhat diffusive, indicating that the qualities of these compounds were relatively poor as compared to those of the corresponding normally prepared samples. Hence, we carefully cut the lanes of 1eD/1eL and 2eD/2eL as sharp and similar as possible to the other cases. (b) In addition, we conducted field-emission scanning transmission electron microscopy (FE-STEM) for the D-Pen-exchanged compounds 1eD and 2eD. Unfortunately, clear images could not be obtained probably due to the very small size and high reactivity of the clusters under the electron-beam irradiation, but we confirmed that there was no size growth in these compounds.

derivative spectra) suggests a partial (probably nearly 50% ee) ligand-exchange reaction in 1eD/1eL and 2eD/2eL. The partial chiral ligand-exchange reaction is supported by the magnitudes of the gmax-factors. Chiroptical Properties: Chiral Functionalization. A key feature is the appearance of intense Cotton effects from the metal clusters modified through the chiral ligand-exchange reaction (Figure 4c,f). An attempt to bestow the optical activity in citratecapped silver nanoparticles has been found in the literature;20 however, well-defined, size-regulated examples of chiral functionalization using a chiral ligand-exchange reaction were given first. The pattern of the anisotropy factor of 1eD/1eL or 2eD/2eL is similar to that of 1D50/1L50 or 2D50/2L50, respectively. Its intensity is smaller than that of 1D/1L or 2D/2L but almost identical to that of 1D50/1L50 or 2D50/2L50, respectively. These results also suggest that the surface ligands would be partially (probably nearly 50% (20) Li, T.; Park, H. G.; Lee, H.-S.; Choi, S.-H. Nanotechnology 2004, 15, 660.

2764 Langmuir, Vol. 24, No. 6, 2008

Nishida et al.

Figure 4. Panels a-c show absorption spectra, first derivative spectra, and anisotropy factors of 1eD/1eL, respectively. Panels d-f also show absorption spectra, first derivative spectra, and anisotropy factors of 2eD/2eL, respectively. Absorption spectra of 1eD and 2eD, and first derivatives of 1eL and 2eL, are offset by adding (subtracting) a constant. In panels c and f, the red and green curves represent the D-Pen- and L-Pen-enantioenriched silver nanoclusters, respectively.

ee) displaced by penicillamine with specific handedness through the exchange reaction to yield the enantioenriched chiral clusters. Note that, at this stage, we have no information on the ratio between D- and L-isomers in the surface monolayer. Therefore, we next assess the reaction yield of the chiral ligand-exchange or ee value. Estimation of Enantiomeric Purity of Surface Penicillamine Monolayer. The absorption spectral profile and intensity of the g-factor of the silver nanoclusters depends on the ee value of surface penicillamine, so the ligands of 1eD/1eL or 2eD/2eL are expected to be modified by a partial displacement of the enantiopure species as discussed previously. To evaluate this hypothesis, we developed a simple assay using acidic decomposition of the small silver nanoclusters:21 In this method, the addition of a concentrated acid (e.g., HCl) induced the detachment of ligands and subsequent decomposition (agglomeration) of silver nanoclusters,8 resulting in the appearance of CD signals (we focus on the signals) mostly originating from the optically active penicillamine moiety along with the disappearance of those

originating from the metal-based transition of silver. At first, by using normally prepared cluster compounds, a calibration curve showing plots of the CD intensity at 225-230 nm (chiral penicillamine origin) for the decomposed sample versus the ee value of the surface ligand was obtained as follows: Only a small portion (10 µL) of concentrated hydrochloric acid (6 M) was added to the solution (3 mL) of the normally synthesized 1D/1L or 1D50/1L50 (2D/2L or 2D50/2L50) of known ee to decompose (or agglomerate) the clusters, followed by centrifugation to yield a supernatant containing detached penicillamine moieties. Before acidification, absorption spectra of the samples are measured, with their absorbance as close as possible to each other at the (21) (a) Lee, H.; Kim, M. S.; Suh, S. W. J. Raman Spectrosc. 1991, 22, 91. (b) In ref 21a, it was reported that cysteine-capped silver nanoparticles were decomposed to form gray precipitates under highly acidic conditions (pH < 3). (c) Although this detail of the decomposition mechanism is not clear at present, it is probably due to a high chemical reactivity of small silver nanoclusters in the reaction with hydrochloric acid, which might produce silver chloride (Li, L.; Zhu, Y.-J. J. Colloid Interface Sci. 2006, 303, 415.)

Optically InactiVe Au Nanoclusters

Langmuir, Vol. 24, No. 6, 2008 2765

Figure 5. Panels a and b show absorption spectra of 1D/1D50/1eD and 1L/1L50/1eL before acidification (acid-untreated samples). Panels c and d show CD spectra of the respective compounds after acidification. (e) Relationship between the initial ee value of the surface ligand and the CD signal intensity (averaged at 225-230 nm) normalized to the absorbance (at ∼300 nm) of the acid-untreated nanocluster 1. The plots and linear regression lines could be obtained for the 1D/1D50 (circles) and 1L/1L50 (diamonds). The expected positions of 1eD and 1eL are also shown by the arrow.

metal-based transition region. Then, the almost colorless supernatant was subjected to CD measurements. The ordinary absorption of the nanoclusters and CD spectra of the supernatants of 1D/1D50/1eD are shown in Figure 5a,c, and those of 1L/1L50/1eL are in Figure 5b,d, respectively. Note that a similar procedure, the acidification-centrifugation-CD measurement, was used for the ligand-exchanged compound 1eD/1eL (or 2eD/2eL). From the measurements of 1D/1D50 and 1L/1L50, we can know the relationship (calibration curve, shown in Figure 5e) between the initial ee value of the surface ligand and the detected CD intensity. In Figure 5e, the CD intensity was normalized at the absorbance of ∼300 nm for each acid-untreated silver nanocluster component.22 As a consequence, a linear regression line can be obtained. On the basis of this calibration line, we obtained ee values of ∼47 or ∼55% for 1eD or 1eL, respectively. Similarly, the ee values of 2eD and 2eL were ∼48 (22) When the absorption of the supernatant solution was used for the calibration, we failed to obtain the relationship probably because the solution still contained the running buffer components and thus interfered with the appropriate estimation.

and ∼59%, respectively,23 exhibiting a similar enantiomeric purity with each other and with those obtained for 1eD/1eL. If we assume a linear relationship between the gmax-factor and the ee value, the maximum anisotropy factors of 1e and 2e can be estimated to be ∼8.0 × 10-4 and ∼3.2 × 10-4, respectively, which agree well with the experimentally obtained values (8-10 × 10-4 for 1e and 3.0 × 10-4 for 2e). Therefore, we can conclude that the racemic ligands on normally synthesized 1 and 2 are partially exchanged by chiral D-/L-Pen through ligand exchanges to yield the optically active nanoclusters having a surface ee of ∼5060%. Possibility of Silver Core Deformation in Optically Active Nanoclusters. Since the enantioenriched penicillamine-protected silver nanocluster compounds behave like a sterically regulated molecular complex such as a transition metal complex bearing (23) Calibration curve (relationship between ee and CD/Abs values) was carefully examined for 2 (2D/2L/2D50/2L50). Although a linear relationship could be obtained, the slopes of the regression line were different between 2D/2D50 and 2L/2L50 series. See the Supporting Information for details. However, by using the respective calibration curve for the D- or L-Pen-exchanged samples, we estimated the reasonable enantiomeric excess of 2eD or 2eL to be ∼48 or ∼59% ee, respectively.

2766 Langmuir, Vol. 24, No. 6, 2008

chiral ligands, a dissymmetric vicinal field effect from asymmetric carbon in the ligands is expected to universally contribute to their origin of chiroptical activity.8,24,25 Furthermore, in a tiny cluster region smaller than ∼1.5 nm in diameter (e.g., 1 and 2), the possibility of additional contributions from a chiral silver core and/or chiral adsorption pattern has been proposed on the basis of the enhancement of the anisotropy factor in comparison with the analogous gold nanoclusters having the same ligand.8 Here, we further discuss their origin from the present results. The ee-dependent absorption properties of penicillamineprotected silver nanoclusters suggest that the nanoclusters covered with chiral penicillamine with a low ee value (1rac, 2rac, 1D50/1L50, and 2D50/2L50) might have a different core geometry from enantiopure (or 100% ee) D-Pen- and L-Pen-protected nanoclusters (1D/1L and 2D/2L) and that the relevant core deformation or rearrangement is probably caused by a sterically favorable adsorption pattern under the statistical/kinetic ligation.26 In addition, the anisotropy factors of the silver nanoclusters (1.05 and 1.30 nm in diameter) were on the order of 1 × 10-3 and thus are considerably large. Hence, unlike the case of chiral penicillamine-protected gold nanoclusters,7 the previous considerations point more toward a deformed chiral core situation in chiral penicillamine-protected silver nanoclusters, and thus, it is tempting to speculate that the silver core can be much more readily deformable than the gold core in the size range of less than ∼1.5 nm. The origin of the difference in the metal core deformability between gold and silver is unclear at present, but some possibilities can be stated: (i) A high chemical reactivity of silver toward oxygen, which would increase the probability of surface oxide formation, might help to break symmetry when chiral ligands are attached onto the nanocluster surface. (ii) The difference in the electronic structure between gold and silver might influence their deformability. The electronic structure of bulk noble metals is characterized by a half-filled band formed from the atomic s-orbitals, which are overlapped by the d-band existing below the Fermi energy.27 In gold, the d-wavefunctions are more extended than those in silver, whereas the s-wavefunctions are more contracted due to relativistic effects that are stronger in heavier atoms.27 Since it is expected that electronic coupling between the s-orbitals in the metal core and the ligand molecular orbitals is different between gold and silver, it might also help symmetry breaking, yielding a larger optical activity. (iii) The balance between the metal cohesive energy and the metal-sulfur bond energy might contribute to the core deformation. The cohesive energy of noble metals as well as the bond energy between sulfur and metal atoms is reported: The cohesive energy of Au or Ag is about 385 or 284 kJ mol-1, respectively,28 and the bond energy of Au-S or Ag-S is about 200 or 217 kJ mol-1, (24) Goldsmith, M.-R.; George, C. B.; Zuber, G.; Naaman, R.; Waldeck, D. H.; Wipf, P.; Beratan, D. N. Phys. Chem. Chem. Phys. 2006, 8, 63. (25) Yasui, T.; Hidaka, J.; Shimura, Y. Bull. Chem. Soc. Jpn. 1966, 39, 2417. (26) Negishi, Y.; Takasugi, Y.; Sato, S.; Yao, H.; Kimura, K.; Tsukuda, T. J. Phys. Chem. B 2006, 110, 12218. (27) Ha¨kkinen, H.; Moseler, M.; Kostko, O.; Morgner, N.; Hoffmann, M. A.; Issendorff, B. Phys. ReV. Lett. 2004, 93, 93401.

Nishida et al.

respectively.29 Despite the similar values of Au-S and Ag-S bond energies, the metal cohesive energy of gold is much larger than that of silver, and thus, the silver cluster core might be more readily deformable than gold through strong thiolate ligation. It is evident that chirality in the surface monolayer can transfer optical activity to the relevant metal nanoclusters, so that such tiny silver nanoclusters covered with chiral adsorbates will be one of the most promising and interesting candidates for new optically active nanomaterials.

Conclusion In summary, chiral functionalization of optically inactive monolayer-protected silver nanoclusters was successful via a ligand-exchange reaction between racemic penicillamine on a silver nanocluster surface and enantiopure D- or L-penicillamine. The ligand-exchanged nanoclusters displayed a different PAGE separation pattern from that of the normally synthesized samples, but the emergence of the identical band positions in the separating gel assured the presence of size-invariant nanoclusters (1.05 and 1.30 nm in core diameter) and allowed us to investigate the optical/chiroptical properties of these nanoclusters. The ligandexchange reaction resulted in the appearance of large Cotton effects in metal-based electronic transitions with an almost mirrorimage relationship between the respective enantiomeric compounds. To approximately estimate the enantiomeric excess (ee) of the D-/L-isomer on the ligand-exchanged cluster surfaces, we developed an assay on the basis of the fact that the present silver nanoclusters were decomposed to release the surface ligand molecules under a strongly acidified condition. Consequently, the ee value of the surface enantiopure penicillamine was ∼5060%. During the course of the experiments, we found that ordinary absorption profiles and anisotropy factors of such small silver nanoclusters were strongly dependent on the ee value of the surface penicillamine, so that the larger optical activity as compared to that of the analogous gold nanoclusters with a comparable size might be due to the silver core deformation or rearrangement along with the universally influential vicinal contribution from the chiral ligand field. Acknowledgment. The present work was financially supported by Grant-in-Aids for Scientific Research B: 19310076 (H.Y.) and S: 16101003 (K.K.) from JSPS, Japan. Supporting Information Available: Normalized absorption spectra of the nanocluster compounds (1D/1L/1rac/1D50/1L50), and relationship (calibration curve) between ee value of the surface ligand and detected CD intensity normalized to the absorbance at ∼300 nm for acid-untreated samples 2. This material is available free of charge via the Internet at http://pubs.acs.org. LA703351P (28) Kambe, K. Phys. ReV. 1955, 99, 419. (29) (a) Ulman, A. An Introduction to Ultrathin Organic Films: From Langmuir Blodgett to Self-Assembly; Academic Press: New York, 1991. (b) Pradeep, T.; Sandhyarani, N. Pure Appl. Chem. 2002, 74, 1593. (c) Monchev, B.; Petkov, P.; Petkova, T.; Popov, C. J. Optoelectron. AdV. Mater. 2005, 7, 1293.