J. Phys. Chem. B 2005, 109, 19657-19663
19657
Preparation and Characterization of Cr(CO)4dpp (Chromium Tetracarbonyl 2,3-Bis(2′-pyridyl)pyrazine) Adsorbed on Silver Nanoparticles Hua Tan,† Lingkai Wong,† Mei Ying Lai,§ G. S. M. Kiruba,† Weng Kee Leong,† Ming Wah Wong,† and Wai Yip Fan*,† Department of Chemistry, National UniVersity of Singapore, 3 Science DriVe 3, Singapore 117543, and Institute of Materials Research and Engineering (IMRE), 3, Research Link, Singapore 117602 ReceiVed: July 4, 2005; In Final Form: August 19, 2005
A transition metal carbonyl species, Cr(CO)4dpp, has been successfully attached to bare silver nanoparticles prepared by laser ablation of a metal foil in ethanol. Transmission electron microscopy (TEM) images have shown that at least a portion of the silver nanoparticles have been capped by the chromium species, and ligand shells corresponding to Cr(CO)4dpp multilayer adsorption onto the silver nanoparticles of 30-50 nm diameter have been observed. The detection of the strongest Raman-active νCO band of Cr(CO)4dpp at 2004 cm-1 revealed that the species has been adsorbed without decomposition. The time-of-flight secondary ion mass spectrometry (TOF-SIMS) signals recorded of the chromium-capped silver nanoparticles were also consistent with the nondecomposition adsorption process. Density functional calculations have been used to reproduce the Raman spectrum using Ag7+ as a model surface. A large binding energy of about 122 kJ/mol has also been computed between silver and nitrogen atoms thus lending support to Cr(CO)4dpp being chemisorbed onto the silver surface.
Introduction There has been tremendous interest in the study of nanoparticles because of their wide-ranging applications in molecular electronics, catalysis, and chemical sensing systems.1-3 Silver and gold nanoparticles stabilized by neutral and ionic organic surfactants have been synthesized by chemical reduction methods and in recent years by pulsed laser ablation of metal foils.4-7 Transmission electron microscopy (TEM), together with UV-visible absorption spectroscopy, has been invaluable in providing direct evidence of the size and shape of these metallic nanoparticles. In addition, Ag and Au surfaces have been used as attachment sites for nitrogen- or sulfur-based adsorbates, such as pyridines or thiols, in which the nature of the adsorbatesurface interaction has been probed using surface-enhanced Raman scattering (SERS) spectroscopy.8-12 There have been very few studies of transition metal carbonyls as adsorbates on nanoparticles. Both metallic nanoparticles and transition metal carbonyls are well-known as efficient catalysts in many chemical processes. Since nanoparticles can also serve as surface templates, such systems may also serve as molecular models for studying the nature of the adsorbate-surface interactions, a potentially powerful alternative to the more conventional atmospheric and ultrahigh vacuum surface studies employing thin films.13-15 To our knowledge, there has been only one study of transition metal carbonyl species attached to silver colloids, which was investigated using SERS spectroscopy. Vlckova et al. obtained the Raman spectra of Re(CO)3BrL species (L ) dpp (2,3-bis(2′-pyridyl)pyrazine), bpm (2,2′bipyrimidine), and bpy (2,2′-bipyridine)) attached to silver colloids dispersed in water and dimethyl sulfoxide (DMSO).16 * Corresponding author. E-mail:
[email protected]. Fax: 6567791691. † National University of Singapore. § Institute of Materials Research and Engineering.
The adsorption of the Re-dpp and Re-bpm complexes was shown to occur without decomposition, but the absence of an extra coordination site for the Re-bpy species precluded its interaction with silver. The formation of an Ag+-L-Re(CO)3BrL surface complex was postulated in order to account for their UV-visible and SERS spectroscopic data. The ways in which the choice of solvent (aqueous or nonaqueous), addition of a coadsorbate or surfactant, and size of the metal colloid affect the production of optimal SERS signals have also been examined extensively.17-19 Recently, we have embarked on a project to synthesize and characterize metal carbonyl catalysts attached to nanoparticles to be used as a potential semi-heterogeneous catalyst. One of the main problems of organometallic homogeneous catalysis is the difficulty in separating the catalyst from the organic products. However, although heterogeneous catalysts are used extensively in industry, many complex organic transformations could be performed only by organometallic species under homogeneous media. We aim to bridge the gap by making use of the ease of separation of a nanoparticulate organometallic catalyst from the organic products by centrifugation. In this work, we have synthesized and characterized Cr(CO)4dpp adsorbed on Ag nanoparticles dispersed in ethanol. We chose the dpp ligand because it has been known to act as a linker between the Re(CO)3 moiety and the surface of the silver nanoparticles.16 The Cr(CO)4 moiety is also interesting, as it has been shown to photocatalyze hydrosilylation of dienes using phenylsilanes and photoinduce the hydrogenation of dienes, although this paper will initially focus only on the structural characterization of the nanoparticles.20,21 The nanoparticles have been produced by pulsed YAG laser ablation of a silver foil in ethanol solution in the absence of surfactants. The choice of ethanol over water as a solvent is to enable dissolution of the chromium carbonyl species and hence facilitate its direct adsorption onto the nanoparticle. The techniques used for characterization in this
10.1021/jp0536347 CCC: $30.25 © 2005 American Chemical Society Published on Web 09/29/2005
19658 J. Phys. Chem. B, Vol. 109, No. 42, 2005
Tan et al.
work are TEM imaging, SERS, UV-visible absorption spectroscopy, and time-of-flight secondary ion mass spectrometry (TOF-SIMS). To gain a better insight into the adsorption mechanism and SERS spectrum of Cr(CO)4dpp on silver nanoparticles, density functional theory (DFT) calculations of a model of the Ag surface (Ag7+) complex of Cr(CO)4dpp were also performed. Experimental Section Cr(CO)6 and dpp were purchased from Aldrich and used without further purification. Toluene, hexane, dichloromethane, and absolute ethanol were dried before use. Cr(CO)4dpp was synthesized according to the procedure by Ruminski and Wallace.22 Laser ablation was carried out with a Continuum Surelite III-10 Nd:YAG pulsed laser operating at 10 Hz. UVvisible absorption spectra were recorded with a Shimadzu 2550 UV-visible absorption spectrophotometer, in quartz cells of 1 cm path length. SERS measurements were carried out on a Renishaw inVia confocal Raman microscope, equipped with a 50× objective lens and a 514.5 nm laser of 1 mW. TEM images were taken using a JEOL-3010 (300 keV) instrument. The TOFSIMS spectra were acquired using an ION TOF SIMS IV spectrometer, with 25 keV Ga+ bombardment of a 500 × 500 µm sample area in the mass range of 0 < m/z < 1000, with an acquisition time of 400-1000 s. 1. Sample Preparation. A Nd:YAG pulsed laser (1064 nm, 10 ns, 200 mJ/pulse) was focused onto a piece of silver foil, and placed in ethanol for 5 min while the solvent was continuously stirred. The experiment was performed in air. A UV-visible absorption spectrum was taken when the solution turned light yellow, which shows the formation of silver nanoparticles. A 532 nm laser beam (10 ns, 50 mJ/pulse) from the same laser was then focused into the solution itself to induce melting and boiling of the nanoparticles, as these effects have been known to give a narrow range of nanoparticulate size.6 Depending on the number of nanoparticles produced, about 0.08-0.3 mL of a 1 × 10-4 M Cr(CO)4dpp solution in ethanol was then added to 9.9 mL of the dispersed Ag nanoparticles solution until a slight precipitation was observed. After 30 min of stirring, the dark purple solution was centrifuged (Jouan A14, 8000 rpm) for 20 min in order to separate the nanoparticles from the excess Cr(CO)4dpp. After centrifugation, the solution appeared light purple with precipitated black solids, which was later shown to be a combination of pure silver nanoparticles and silver nanoparticles adsorbed with Cr(CO)4dpp (the solid mixture is henceforth known as A). The solid A was dried and kept in the dark, or it could be dispersed in ethanol to be used for characterization studies. Both the solid and solution were used immediately or within 1 day after preparation to minimize decomposition of the chromium species. Ag-dpp nanoparticles were prepared in the same way for comparison studies. 2. Characterization. For the SERS spectroscopic measurements, a drop of solution A was placed on a microscope glass slide and air-dried for a few minutes to allow complete evaporation of the ethanol. The slide was then placed under a confocal Raman microscope (Renishaw inVia, 50× objective lens, 514.5 nm laser of 1 mW), and an optical image was captured of the nanoparticles distributed on the glass slide. Since the microscope has a spatial resolution of only 1 µm, the nanoparticles appeared as dark micron-size spots in the image even though their actual sizes were expected to be smaller. The Raman excitation laser was then focused on one of these spots to obtain its SERS spectrum. The Stokes spectrum was recorded by scanning the spectrometer from 500 to 3200 cm-1 relative
Figure 1. UV-visible absorption spectra with increasing concentration of Cr(CO)4dpp added into the ethanol solution containing Ag colloid (λmax ) 400 nm). The initial concentration of the Ag colloid is estimated to be 10-4 M (on the basis of the number of Ag atoms).
to the 514.5 nm laser line. Raman spectra were also taken for Ag-dpp, pure Ag nanoparticles, and pure Cr(CO)4dpp deposited on glass slides for comparison studies. For the sample preparation for TEM imaging and TOF-SIMS, a drop of solution A was deposited onto a carbon-coated copper grid (diameter ) 3 mm) and a glass slide, respectively. The grid was dried in a desiccator under reduced pressure for a few hours before TEM images were taken. 3. Computational Details. The structures of Cr(CO)4dpp, Ag7+, and Cr(CO)4dpp‚‚‚Ag7+ were fully optimized using Becke’s three-parameter exchange functional23 and PerdewWang correlation functional24 together with the LANL2DZ basis set.25,26 Raman spectra and zero-point energy corrections were computed at the DFT-optimized geometries. The interaction (binding) energy of the complex was calculated as the difference between the energy of the complex and the total energy of the two monomers, i.e., Cr(CO)4dpp and Ag7+. Charge density analysis was performed using the natural bond orbital (NBO)27 approach based on the B3P86/LANL2DZ wave function. All calculations were performed using the Gaussian 98 series of programs.28 Results and Discussion The UV-visible absorption spectrum of the light yellowcolored silver colloidal solution shown in Figure 1 is similar to that of those previously prepared by laser ablation or chemical reduction in water or organic solvents.6 The surface plasmon band at λmax ) 400 nm signifies the presence of dispersed silver nanoparticles with a mean size of about 15 ( 10 nm after annealing with the laser pulse. By comparing the intensity of the plasmon band to the intensity of the band from nanoparticles produced using chemical reduction methods, we estimated the concentration of the Ag nanoparticles to be 10-4 M (based on the number of silver atoms). Compared to chemical reduction methods, the laser ablation method is simple, convenient, and able to produce very stable nanoparticles (lasting at least two months) in the absence of surfactants. An advantage of nanoparticles without surfactants is the enhanced direct adsorptivity of Cr(CO)4dpp onto silver.29 Unlike silver-thiol interactions, the silver-nitrogen interaction is weaker, and thus even weakly bound surfactants such as sodium dodecyl sulfate (SDS) could compete for the adsorption sites on silver.9 However, bare particles tend to be less stable, and can easily aggregate and
Preparation and Characterization of Cr(CO)4dpp TABLE 1: SERS and Calculated Spectra of dpp and Cr(CO)4dpp Adsorbed on Ag Nanoparticles dpp-Agn lit (cm-1)a
expt (cm-1)
1628 1602 1535 1482 1400 1370 1302 1293 1231
1630 1587 1533 1495 1415 1329 1295 1268 1216 1191
1059
1069
1007
1006
J. Phys. Chem. B, Vol. 109, No. 42, 2005 19659 TABLE 2: TOF-SIMS Data for Cr(CO)4dpp Adsorbed on Ag Nanoparticlesa
Cr(CO)4dpp-Agn
Cr(CO)4dpp-Ag7+
species
mass
relative intensity
expt (cm-1)
calcb (cm-1)
2004 1642 1580 1531 1464 1422 1393 1318 1248 1217 1189 1105 1066 1039 1008
2004 (679) 1644 (284) 1585 (567) 1516 (469) 1455 (137) 1442 (149) 1353 (99) 1302 (430) 1277 (773) 1214 (21) 1179 (6) 1092 (163) 1055 (141)c 1023 (157) 1015 (75)
Cr dpp Cr-dpp Cr(CO)dpp Ag-dpp Ag-dppCr(CO)
52 234 286 314 343 423
1 0.6 0.3 0.05 0.55 0.01
a
Literature values, taken from ref 16. b B3P86/LANL2DZ level, Raman intensities (A4 amu-1) are given parentheses. c There is another Raman frequency of 1053 cm-1 (163) in proximity.
eventually precipitate out of solution in the presence of a large excess of adsorbates. To investigate the interaction between Cr(CO)4dpp and Ag nanoparticles, we titrated the colloidal solution of Ag nanoparticles in ethanol with 10-4 M Cr(CO)4dpp in ethanol. As Figure 1 shows, the surface plasmon peak position (λmax ) 400 nm) of Ag nanoparticles in ethanol is largely unaffected by the addition of Cr(CO)4dpp. Although the intensity of the band appears to increase with addition of the Cr(CO)4dpp, this may be attributed to the additional absorption across the wavelength region of Cr(CO)4dpp. This is corroborated by the increase in intensity of the MLCT transition of Cr(CO)4dpp (∼560 nm) after each titration. However, A precipitated out upon crossing a threshold concentration of Cr(CO)4dpp, which is to be expected because the nanoparticles were not stabilized with surfactants. At this point, the surface plasmon band of silver disappeared, and only electronic bands due to the free Cr(CO)4dpp could be seen in the spectrum. In the previous reports on the pyridine and Re(CO)3Br(bpm) systems, both of which also bind to silver via the nitrogen atom,8,16 the plasmon bands were unaffected until an excess of adsorbates had been added. In some of these cases, depending on the solvent used, an additional band around 600-700 nm that was attributed to aggregation of the silver colloids was also observed. Such an aggregation was not observed here, even in the presence of excess Cr(CO)4dpp. In contrast, adsorption of thiols on silver is known to broaden the plasmon band as well as red-shift its band center to as far as 600 nm.11 Another reason for the seemingly unchanged silver plasmon band in the presence of Cr(CO)4dpp observed here is that only a small percentage of the silver nanoparticles is capped by the chromium species. The main peaks of the spectrum are therefore those of pure silver nanoparticles and free Cr(CO)4dpp in solution; the plasmon band of silver adsorbed with Cr(CO)4dpp may not be strong enough to be observable. To verify this, we have examined the SERS spectrum and TEM image of A. The SERS spectra of Ag-dpp and A measured at 514.5 nm are shown in Figure 2a,b. The vibrational frequencies of the observed SERS-active modes are listed in Table 2. These modes can be categorized into different groups, as suggested by Vlckova et al. for the adsorption of Re(CO)3Br(dpp) on silver.16 First, a majority of the spectral bands from 1100 to 1700 cm-1 have their counterparts in the SERS spectrum of the dpp ligand of the Ag-dpp system. The assignment is further supported by the calculated Raman spectrum (Table 1). The spectrum of Ag-
a
Recorded using a Ga ion (25 keV) source.
dpp is not identical to that of A, although some coincidences and small frequency shifts of the corresponding bands have been measured between the two systems. Because of the spectral complexity, the vibrations of the dpp segment have not been assigned. Second, the characteristic ν(CO) band of Cr(CO)4dpp is observed at 2004 cm-1, which corresponds to the in-phase symmetric CO stretching vibration (breathing motion). Although there are three other CO vibrational modes (see later section), only the strongest Raman-active band could be observed here. Unfortunately, we were unable to analyze the low-frequency vibrations of the sample because of a fault with the Raman notch filter of the spectrometer. The accessibility of these vibrations could have allowed the Ag-N vibrational frequency to be determined. We were unsuccessful in many attempts to record the SERS spectrum of Cr(CO)4dpp deposited on the surface of a glass slide. The absence of any Raman signals serves to show that silver is indeed required for the detection of SERS signals. Similarly, the SERS spectrum of only silver nanoparticles on the glass surface did not yield any Raman bands. The simultaneous observation of the metal carbonyl stretching modes with the vibrational modes of dpp, which were not identical to those of the adsorbed dpp ligands in the SERS spectra of Cr(CO)4dpp, shows that this species has been adsorbed on the Ag surface without decomposition. It is not possible to show experimentally whether Cr(CO)4dpp has been chemisorbed or physisorbed on the Ag surface, although density functional calculations (see later) indicated that a strong interaction between the nanoparticles and the pair of nitrogen atoms on dpp is expected. From the SERS measurement, it is not possible to establish the extent of adsorbate coverage on the silver nanoparticles. As evidenced in the UV-visible spectrum, a significant portion of solid A is believed to consist of the uncoated silver nanoparticles. The TEM images of A (Figure 3) show a size distribution of spherical silver nanoparticles, with an average diameter of 20 nm. The electron diffraction data of the nanoparticles correspond very well to the lattice spacings of silver itself.6 The size distribution is fairly narrow considering that the absence of surfactants in the system precludes size control of the nanoparticles. This probably shows the effect of the 532 nm laser treatment on the solution of nanoparticles in which size control has been demonstrated.6 The size distribution is also consistent with the appearance of the surface plasmon band centered at 400 nm because the λmax of the band could be used as an indication of spherical nanoparticle size. Interestingly, a closer inspection of some of the TEM images revealed that it is the larger nanoparticles that provided evidence for the adsorption of Cr(CO)4dpp on silver. Figure 3b shows a silver nanoparticle of 50 nm diameter (large dark spot); a narrow ring or shell (in lighter contrast) of about 3.8 nm width surrounding this nanoparticle is observed. A survey of the silver nanoparticles in the TEM image also reveals the presence of such shells of 2-4 nm thickness around silver nanoparticles of sizes ranging from 30 to 50 nm. Closer inspection shows very thin shells
19660 J. Phys. Chem. B, Vol. 109, No. 42, 2005
Tan et al.
Figure 4. Segments of the TOF-SIMS spectra from m/z ) 230 to 350 recorded for (a) pure Ag nanoparticles deposited on a glass slide and (b) Cr(CO)4dpp adsorbed on silver nanoparticles deposited on a glass slide. Figure 2. SERS spectrum of (a) dpp adsorbed on Ag nanoparticles and (b) Cr(CO)4dpp attached to Ag colloid. The peak at 2004 cm-1 in (b) corresponds to the in-phase symmetric CO stretching motion of the Cr(CO)4dpp moiety.
Figure 3. (Top) Transmission electron microscope image of AgCr(CO)4dpp nanoparticles showing narrow size distribution. (Bottom) TEM image of a single Ag nanoparticle (dark particle of diameter 40 nm) with adsorbed multilayers of Cr(CO)4dpp with a thickness of 3.8 nm (5 layers).
around some of the smaller nanoparticles to be discerned. In contrast, no such shell structures were observable in the TEM
images of pure silver nanoparticles. Since surfactants were not used in the preparation of the nanoparticles, the presence of the shells can be attributed to Cr(CO)4dpp multilayer adsorption on the silver surface. The thickness of the shell in Figure 3b corresponds to about 5 layers of Cr(CO)4dpp stacked onto the silver surface, using an estimated value of 0.8 nm for the length of the chromium carbonyl species measured from one of the nitrogen atom pair interacting with the silver to the oxygen atom of the carbonyl bond furthest from the surface. Of course this does not mean that all silver cores were covered with 5 layers of the chromium shell; rather, this calculation can give only an order-of-magnitude estimate. Hence, the Cr(CO)4dpp could be found in two environments, either as the species directly adsorbed on silver or as a weakly physisorbed species found in the multilayer shell. The latter species would also be present in a larger quantity. Such ligand shells have also been observed in water-soluble phosphine-coated gold nanoparticles. Schmid et al. found that a shell of about 7 nm thickness, corresponding to 12 phosphine layers, has been attached to a 44 nm diameter gold colloid.30 As in our case, the TEM image of the phosphine shell was faint because of the absence of long-range order within the adsorbed layers. Unlike silver, the chromium adsorbants in our experiments are not deposited in polycrystalline form, and hence their electron diffraction patterns could not be obtained, as is often the case for any organic surfactant-capped nanoparticles. TOF-SIMS (positive and negative) spectra of A and pure silver nanoparticles deposited on a glass slide were also recorded (Figure 4). A broad scan of the mass spectrum of A from m/z ) 0 to 1000 reveals some features not found in the TOFSIMS recorded for pure silver nanoparticles on glass. The signalto-noise ratios of the extra features were generally low because only patches of nanoparticles were distributed on the glass surface. The optical image observed with the Raman microscope suggests that many of these patches, as well as part of the bare glass surface, will still be exposed to the gallium ion beam, which has an estimated diameter of a few micrometers. It was therefore essential to obtain a reference spectrum of deposited, uncoated Ag nanoparticles because the high sensitivity of the TOF-SIMS technique allows even a small amount of impurity present on the bare glass slide to be detected. The absence of
Preparation and Characterization of Cr(CO)4dpp surfactants from the sample also greatly simplified the interpretation of the TOF-SIMS spectrum. Although both positive and negative spectra of the sample yielded approximately the same information, better signal-to-noise ratios were obtained for the former spectrum, and hence only the features of the positive spectrum will be discussed. Table 2 lists the main features, together with the intensity of the peaks relative to the Cr ion peak at m/e ) 52, in the positive mass spectrum. The Cr+ peak is the most intense of the signals because of the chromium-containing species, although by far the largest peaks in the spectrum belong to pure silver species, Agn, which are easily identified by their isotopic pattern. The next strongest peaks correspond to molecular fragment ions of A from species such as Ag-dpp, Ag-dppCr, and AgdppCr(CO). Unfortunately, the Cr(CO)4dpp molecular ion peak, with or without silver attached to it, could not observed. However, the observed signals from fragment ions are consistent with the adsorbed molecule being extensively dissociated by the energetic Ga+ beam. Delcorte et al. have observed extensive fragmentation of ruthenium clusters cast on a silicon substrate;31 the dominant peaks of their TOF-SIMS spectrum arose from clusters that lost the hydrocarbon residues and a number of CO ligands. In our case, ion bombardment has also led to extensive CO dissociation from Cr(CO)4dpp because the largest observable fragment is Ag-dppCr(CO), followed by a loss of CO to yield Ag-dppCr. It is also noteworthy that fragments that do not contain dpp, such as Ag-Cr or Ag-Cr(CO)n, are absent in the mass spectrum. The presence of such signals may indicate the decomposition of Cr(CO)4dpp leading to direct Cr(CO)n binding on silver. However, this is unlikely because molecular species of the type Cr(CO)nAgm are unknown. In contrast, mass spectral peaks due to species such as Ag-dpp and Ag-dppCr could be observed quite readily, presumably originating from ion beam dissociation of the parent molecule. The overall pattern in the mass spectrum lends further support to the nondissociative nature of Cr(CO)4dpp adsorption on silver. The nature of the adsorption, either chemisorption or physisorption, of the chromium species is more difficult to establish experimentally. If the chromium species had been weakly physisorbed on the surface, it is more likely to have fragments containing only dpp or only Ag since the Ag-N interaction would be very weak and readily detached upon ion bombardment. Hence, the detection of fragments having both Ag and dpp species such as Ag-dpp, Ag-dppCr, and Ag-dppCr(CO) in reasonable yields lends more support to a chemisorption of Cr(CO)4dpp on silver. To further probe the mechanism of chemisorption and modes of interaction, we performed DFT calculations on a cluster model of the Ag surface. Roy and Furtak have suggested that a small ion cluster, Agn+, may provide a suitable representation of the silver surface.32 In a recent theoretical study, we have established that a cluster ion of seven silver atoms (Ag7+) provides a good theoretical model for the silver surface.33 The Ag7+ cluster adopts a pentagonal bipyramidal geometry with D5h symmetry. This geometry is preferred over the two-dimensional planar geometry (D6h symmetry). Hence, we have employed this Ag7+ cluster ion with D5h symmetry to represent the silver nanoparticles in this study. The optimized geometry of the uncomplexed Cr(CO)4dpp is shown in Figure 5. The pyridyl nitrogen (ring 3) and pyrazyl nitrogen (ring 2) are in a trans arrangement (τ(NCCN) ) -140.0°). This conformational preference is readily understood in terms of the strong repulsion between the nitrogen lone pairs. These two donor nitrogen atoms are potential sites of interaction
J. Phys. Chem. B, Vol. 109, No. 42, 2005 19661
Figure 5. Optimized (B3P86/LANL2DZ) geometries of Cr(CO)4dpp and Cr(CO)4dpp‚‚‚Ag7+, bond lengths in Å and angles in degrees.
with the active sites on the silver nanoparticles. Two plausible modes of interaction between Cr(CO)4dpp and Ag7+ are envisaged: monodentate and bidentate. On the basis of calculations of a smaller model complex, Cr(CO)4dpp‚‚‚Ag+, we have found that the bidentate complex is significantly more stable than the two possible monodentate complexes. More importantly, we notice that adsorption of Cr(CO)4dpp in a monodentate fashion will lead to strong steric repulsion between certain functional groups of Cr(CO)4dpp with the Ag atoms adjacent to the binding site. Thus, we can safely rule out the possibility of a monodentate Ag‚‚‚N interaction between Cr(CO)4dpp and the Ag surface. It is also worth noting that there is no splitting of the SERS bands observed, suggesting that only one mode of Ag‚‚‚N interaction was observed experimentally. The optimized geometry of the Cr(CO)4dpp‚‚‚Ag7+ bidentate complex is depicted in Figure 5. The two Ag‚‚‚N interaction distances are 2.366 and 2.462 Å. For comparison, the Ag‚‚‚N bond length in the pyridine‚‚‚Ag7+ complex is 2.200 Å at the same level of theory.28 The two donor nitrogen atoms of dpp in the coordinated Cr(CO)4dpp are almost coplanar (τ(NCCN) ) -35.8°) to allow simultaneous interaction with a silver atom of Ag7+ in a bidentate fashion. Constrained optimization of Cr(CO)4dpp indicates that the dpp moiety is sufficiently flexible; only 20 kJ mol-1 is required to adopt the conformation of dpp required in the Ag complex. Interestingly, the interaction between the chromium carbonyl moiety and dpp is stronger in
19662 J. Phys. Chem. B, Vol. 109, No. 42, 2005
Tan et al.
TABLE 3: Calculateda Carbonyl Stretching Frequencies (cm-1) and Raman Intensities (A4 amu-1) of Cr(CO)4dpp and Cr(CO)4dpp‚‚‚Ag7+ Cr(CO)4dpp‚‚‚Ag7+
Cr(CO)4dpp frequency
Raman intensity
relative intensity
frequency
Raman intensity
relative intensity
1879 1907 1913 1996
76 98 30 446
0.17 0.22 0.07 1.00
1905 1929 1934 2004
130 64 172 679
0.19 0.09 0.25 1.00
a
At the B3P86/LANL2DZ level.
the silver complex. This is readily reflected in the shorter Cr-N bond lengths (2.021 and 2.061 Å, cf. 2.051 and 2.066 Å, respectively, in the free Cr(CO)4dpp). The C-O bonds are correspondingly slightly weaker. The Cr(CO)4dpp‚‚‚Ag7+ complex is predicted to have a large binding energy of -122.5 kJ mol-1 (B3P86/LANL2DZ + ZPE). On the basis of the NBO charge density analysis, there is a significant amount of electron transfer (0.10 e) from Cr(CO)4dpp to Ag7+ in the bidentate complex. The computed Raman spectral lines are summarized in Tables 1 and 3. There are several important features that warrant discussion. First, we note that the computed Raman frequencies are in excellent agreement with the observed values in the SERS spectrum, with a mean absolute deviation of 12 cm-1 (Table 1). Second, substantial Raman intensity enhancement is observed for all Raman-active vibrations of Cr(CO)4dpp upon complexation with Ag7+. For instance, there is a 5-fold increase in the Raman intensity of the symmetric CdC stretching vibration of the pyrazine ring at 1585 cm-1. This theoretical finding is consistent with the characteristic SERS effect of a “chemisorbed” molecule on a metal surface.34 Third, the four carbonyl vibrations have different Raman intensities (Table 3). The CO vibration with the highest frequency (2004 cm-1) is predicted to have a substantially larger Raman intensity. As the CO absorption bands in the SERS spectrum are weak (Figure 4b), it is not surprising that only the strongest band was observed. Fourth, all four CO vibrational frequencies undergo blue shift (8-28 cm-1) on going from the free Cr(CO)4dpp molecule to the surface complex (Table 3). For the CO vibration with the strongest Raman intensity (i.e., 2004 cm-1), the frequency shift is just 8 cm-1. These small changes are perhaps not unexpected as the CO moieties are far from the site of interaction with the Ag cluster. Finally, we predict the Ag‚‚‚N stretching vibration to be 175 cm-1. For pyridine absorbed on a silver surface, a vibrational frequency of 239 cm-1 has been assigned to the Ag-N stretching vibration.35 The excellent agreement between the calculated and experimental Raman spectra offers strong support that our Ag cluster model employed here is a good theoretical model for the investigation of chemisorption of small molecules on Ag nanoparticles. The calculated Raman enhancement and charge transfer in the Cr(CO)4dpp‚‚‚Ag7+ complex are thus consistent with the chemisorption model of adsorption. Summary A transition metal carbonyl species, Cr(CO)4dpp, has been successfully attached to silver nanoparticles. Bare silver nanoparticles were prepared by laser ablation of a metal foil followed by dropwise addition of Cr(CO)4dpp to affect the adsorption process. Observation of the strongest Raman-active νCO band of Cr(CO)4dpp at 2004 cm-1 revealed that the species adsorbed without decomposition on silver. TEM images showed that at least a portion of the silver nanoparticles have been capped by
the chromium species. Ligand shells corresponding to a multilayer of Cr(CO)4dpp species surrounding silver nanoparticles of 30-50 nm diameter have been observed. The mass spectral signals of the fragments observed in the TOF-SIMS spectrum were also consistent with the nondecomposition nature of the adsorbed Cr(CO)4dpp on silver. A computational investigation using DFT calculations at the B3P86/LANL2DZ level was employed to study the adsorption mechanism and Raman spectrum of Cr(CO)4dpp adsorbed on silver nanoparticles. On the basis of the Cr(CO)4dpp‚‚‚Ag7+ model calculation, Cr(CO)4dpp is chemisorbed via bidentate coordination of pyridyl and pyrazyl nitrogen atoms of dpp to an active site of Ag nanoparticles. The calculated Raman spectrum of Cr(CO)4dpp‚‚‚Ag7+ also agrees well with the SERS spectrum. We are also exploring the synthesis and applications of such bimetallic nanoparticles in homogeneous catalytic systems. For example, ruthenium carbonyl or phosphine species, which are known to catalyze a variety of organic reactions, could be adsorbed on gold or silver nanoparticles, which themselves are also efficient alkyne activators, so that their efficencies against more conventional homogeneous catalysts could be compared and contrasted. Work is already underway to characterize these organometallic species using a variety of ligands as linkers to the nanoparticles. Acknowledgment. The project was supported by the Agency of Science, Technology and Research (ASTAR) under Grant 143-000-198-305. H.T. thanks the National University of Singapore for a research scholarship. References and Notes (1) Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281, 2013. (2) Daniel, M.; Astruc, D. Chem. ReV. 2004, 104, 293. (3) Abid, J. P.; Girault, H. H.; Brevet, P. F. Chem. Commun. 2001, 9, 829. (4) Pileni, M. P. Langmuir 1997, 13, 3266. (5) Petit, C.; Lixon, P.; Pileni, M. P. J. Phys. Chem. 1993, 97, 12974. (6) Mafune, F.; Kohno, J.; Takeda, Y.; Kondow, T.; Sawabe, H. J. Phys. Chem. B 2000, 104, 8333. (7) Henglein, A. J. Phys. Chem. 1993, 97, 5457. (8) Joo, T. H.; Kim, K.; Kim, M. S. Chem. Phys. Lett. 1984, 112, 65. (9) Maeda, Y.; Kitano, H. J. Phys. Chem. 1995, 99, 487. (10) Temperini, M. L. A.; Sala, D.; Lacconi, G. I.; Gioda, A. S.; Macagno, V. A.; Arvia, A. J. Langmuir 1988, 4, 1032. (11) Szafranski, C. A.; Tanner, W.; Laibinis, P. E.; Garrell, R. L. Langmuir 1998, 14, 3580. (12) Xue, G.; Ma, M.; Zhang, J.; Lu, Y.; Carron, K. T. J. Colloid Interface Sci. 1992, 150, 1. (13) Miles, D. T.; Murray, R. W. Anal. Chem. 2001, 73, 921. (14) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, 33, 27. (15) Badia, A.; Cuccia, L.; Demers, L.; Morin, F.; Lennox, R. B. J. Am. Chem. Soc. 1997, 119, 2682. (16) Vlckova, B.; Matejka, P.; Vanoutersterp, J. W. M.; Snoeck, T. L.; Stufkens, D. J. Inorg. Chem. 1994, 33, 2132. (17) Hu, J.; Zhao, B.; Xu, W.; Fan, Y.; Li, B.; Ozaki, Y. Langmuir 2002, 18, 6839. (18) Lepp, A.; Siiman, O. J. Phys. Chem. 1985, 89, 3494. (19) Vlckova, B.; Barnett, S. M.; Kanigan, T.; Butler, I. S. Langmuir 1993, 9, 3234. (20) Abdelqader, W.; Chmielewski, D.; Grevels, F. W.; Oskar, S.; Peynircioglu, N. B. Organometallics 1996, 15, 604. (21) Platbrood, G.; Wilputte-Steinert, L. J. Mol. Catal. 1975/76, 1, 265. (22) Ruminski, R. R.; Wallace, I. Polyhedron 1987, 6, 1673. (23) Becke, D. A. J. Chem. Phys. 1993, 98, 5648. (24) Perdew, J. P. Phys. ReV. B 1986, 33, 8822. (25) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299. (26) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry; Schaefer, H. F., III, Ed.; Plenum Press: New York, 1976; Vol. 3. (27) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899. (28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Preparation and Characterization of Cr(CO)4dpp Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98; Gaussian, Inc.: Pittsburgh, PA, 1998.
J. Phys. Chem. B, Vol. 109, No. 42, 2005 19663 (29) Garrell, R. L.; Schultz, R. H. J. Colloid Interface Sci. 1985, 105, 483. (30) Schmid, G.; Lehnert, A.; Kreibig, U.; Adamczyk, Z.; Belouschek, P. Z. Naturforsch. 1990, 45b, 989. (31) Delcorte, A.; Hermans, S.; Devillers, M.; Lourette, N.; Aubriet, F.; Muller, J. F.; Bertrand, P. Appl. Surf. Sci. 2004, 231, 131. (32) Roy, D.; Furtak, T. E. Chem. Phys. Lett. 1986, 124, 299. (33) Kiruba, G. S. M.; Zhang, J.; Wong, M. W. To be published. (34) Cheng, R. K.; Furtak, T. E. Surface Enhanced Raman Scattering; Plenum Press: New York, 1982. (35) Creighton, J. A. Surf. Sci. 1983, 124, 209.
