Identification of Surface Structures on 3C-SiC Nanocrystals with

Nov 6, 2009 - This provides the opportunity to design suitable surface structures by terminating the surface dangling bonds with different atoms there...
4 downloads 10 Views 2MB Size
Download PDF
NANO LETTERS

Identification of Surface Structures on 3C-SiC Nanocrystals with Hydrogen and Hydroxyl Bonding by Photoluminescence

2009 Vol. 9, No. 12 4053-4060

X. L. Wu,*,†,‡ S. J. Xiong,† J. Zhu,§ J. Wang,† J. C. Shen,† and Paul K. Chu*,‡ National Laboratory of Solid State Microstructures and Department of Physics, Nanjing UniVersity, Nanjing 210093, People’s Republic of China, Department of Physics and Materials Science, City UniVersity of Hong Kong, Tat Chee AVenue, Kowloon, Hong Kong, China, and College of Physics Science and Technology, Yangzhou UniVersity, Yangzhou 225002, People’s Republic of China Received July 11, 2009; Revised Manuscript Received October 19, 2009

ABSTRACT SiC nanocrystals (NCs) exhibit unique surface chemistry and possess special properties. This provides the opportunity to design suitable surface structures by terminating the surface dangling bonds with different atoms thereby boding well for practical applications. In this article, we report the photoluminescence properties of 3C-SiC NCs in water suspensions with different pH values. Besides a blue band stemming from the quantum confinement effect, the 3C-SiC NCs show an additional photoluminescence band at 510 nm when the excitation wavelengths are longer than 350 nm. Its intensity relative to the blue band increases with the excitation wavelength. The 510 nm band appears only in acidic suspensions but not in alkaline ones. Fourier transform infrared, X-ray photoelectron spectroscopy, and X-ray absorption near-edge structure analyses clearly reveal that the 3C-SiC NCs in the water suspension have Si-H and Si-OH bonds on their surface, implying that water molecules only react with a Si-terminated surface. First-principle calculations suggest that the additional 510 nm band arises from structures induced by H+ and OH- dissociated from water and attached to Si dimers on the modified (001) Si-terminated portion of the NCs. The size requirement is consistent with the observation that the 510 nm band can only be observed when the excitation wavelengths are relatively large, that is, excitation of bigger NCs.

The interaction between water and material surfaces is an important research area in multiple disciplines including biophysics, semiconductor technology, and corrosive processes. Particularly, characterization and better understanding of the interface between water and solid biomaterials are crucial to device design in biomedical engineering and medicine. A typical example is surface patterning on the molecular level and then using these patterns to control adsorption of proteins while maintaining their activity.1-4 The search for biocompatible, patterned substrates for the development of biosensors is an active field of research.5,6 Silicon carbide (SiC) is a promising material in biophysics and biomedical applications due to its low weight, high strength, extreme hardness, wear and corrosion resistance, and inertness. Hence, silicon carbide (SiC) nanostructures have attracted considerable interest because of their novel * To whom correspondence should be addressed. (X.L.W.) E-mail: [email protected]. Fax: 86-25-83595535. Tel: 86-25-83686303. (P.K.C) E-mail: [email protected] . † Nanjing University. ‡ City University of Hong Kong. § Yangzhou University. 10.1021/nl902226u CCC: $  2009 American Chemical Society Published on Web 11/06/2009

morphologies,7-12 quantum-confined optical emission,13,14 and applications in biophysics such as fluorescent biological labels.15 In these applications, the surface physical and chemical properties are very important because they directly determine the growth of the nanocrystals (NCs), stability in a solvent, and luminescence properties. Since SiC NCs consists of carbon or silicon outermost layers, they exhibit interesting and complicated surface chemistry in different surroundings. This provides the opportunity to design suitable surface bonding structures by terminating the surface dangling bonds with different atoms.16 On the basis of some theoretical studies on the surface characteristics of polytypic SiC exposed to water, acetic acid, and methanol,1,17,18 dependence of the band gap on the crystallite size and surface bonding structure is revealed.10,19,20 However, experimental studies on this kind of surface modification of SiC NCs are difficult because of the demand for a very clean surface and perfect detection conditions such as very high vacuum.16,21 Even so, the surface modifications of SiC bulk and particles are still a subject of current experimental investigations due to various potential applications. Recently, Rosso et al.22,23

reported for the first time the formation of the covalently bound alkyl monolayers on SiC surfaces by reacting alkenes with HF-etched surfaces or by ultraviolet irradiation in the presence of alkenes to produce well-defined, chemically tunable, and highly stable coatings. Their results show promising applications of functionalized SiC. Iijima et al.24 modified the surface of SiC NCs with azo radical initiators in a reaction with unsaturated hydrocarbons. This rendered the hydrophobic SiC surface hydrophilic and the dispersion stability of SiC NCs in the aqueous solution was improved significantly. After functionalization with perfluoroalkysilane, Niu et al.25 showed that the aligned SiC nanowire crossed nets had excellent superhydrophobic properties. Zinovev et al.26 found that SiC chlorination strongly depended on the hexagonal surface orientation. Because of the thermodynamically more favorable reaction of chlorine with silicon rather than carbon, the C-terminated side (0001j) clearly underwent noticeable changes, resulting in coverage by a black carbon film, whereas the Si-side (0001) surprisingly remained unreacted visually. This result enhanced our understanding on the otherwise largely unpredictable behavior that often takes place during processing of SiC industrial ceramics via chlorination. Photoluminescence (PL) is sensitive to the surface composition and bonding structure of SiC NCs. The complicated surface on SiC NCs provides numerous sites for many potential species to attach, thereby making it an ideal host for sensing applications such as luminescent probes in medical imaging.27 Cicero et al.1 have theoretically predicted that the interactions between water molecules and bulk 3CSiC (001) Si-terminated surface lead to dissociation of water molecules accompanied by surface reconstruction. Hence, it is possible to use PL to monitor such reconstructed surface and dissociated H2O molecules. In this work, we investigate the PL characteristics of aqueous suspensions containing 3CSiC NCs less than 6.5 nm in size with different pH values. An additional PL band at 510 nm is observed when the excitation wavelength is longer than 350 nm and it is shown to arise from the optical transitions associated with these surface complexes. Since it is very difficult to conduct ab initio calculation directly on NCs due to their sizes, we focus on the effects of the surface bonding structure in water on the spectrum and theoretically calculate the electronic structure of the layered 3C-SiC with surface reconstruction and dissociation of adsorbed water molecules based on the density functional theory (DFT) which agrees well with experimental results. Fabrication of the water suspensions of 3C-SiC NCs has been described previously.28 Because of the existence of very weak surplus acid, the pH value of the initial water suspension was about 5.5. To adjust the pH value, several drops of diluted HCl or NaOH solutions were added to the suspension and stirred for several min. The details regarding the transmission electron microscope (TEM) observation and PL measurements are similar to those reported previously.13,28 Fourier transform infrared (FTIR) spectra were acquired from the suspension put on a clean silicon wafer using a Nicolet 170SX. X-ray photoelectron spectroscopy (XPS) analysis was 4054

Figure 1. (a) A representative TEM image of the 3C-SiC NC distribution. The inset shows the high-resolution TEM image of a NC. (b) Size distribution obtained from many TEM images. (c,d) Normalized PL spectra of the water and ethanol suspensions with excitation in the wavelength range of 310-410 nm. (e) Dependence of the blue PL peak positions on excitation wavelengths for the water and ethanol suspensions.

carried out on a Kratos AXIS Photoelectron Spectrometer (Japan). The narrow-scan spectra were obtained under ultrahigh vacuum conditions and using monochromatic Al KR X-ray radiation. The Si L3,2-, and O K-edge X-ray absorption near-edge structure (XANES) measurements were conducted on the X-ray Magnetic Circular Dichroism Endstation at The Hefei Synchrotron Radiation Facility in The University of Science and Technology of China. Prior to the XPS and XANES measurements, the samples were prepared by dropping the water suspension of 3C-SiC NCs on a clean silicon wafer coated with a thin silver layer by electron beam evaporation and dried in vacuum for a long time until all the water was vaporized leaving a thin layer of about 200 nm on the substrate. To eliminate possible contributions from surface contamination to the spectral results, Ar ion bombardment was performed to remove about 1 nm of materials from the surface prior to XPS. The reference SiO2 film was also prepared by electron beam evaporation directly on a clean silver thin film. Figure 1a exhibits a representative TEM image of the 3CSiC NCs. The well-dispersed NCs are nearly spherical and have diameters of 1.5-6.5 nm. The inset in Figure 1a depicts the high-resolution TEM image of a NC revealing that the NC has lattice fringes corresponding to the (111) plane of 3C-SiC. Here, we would like to stress that for a spherical NC, the 〈111〉 growth orientation does not imply that all outer crystallographic planes of the NCs belong to the {111} Nano Lett., Vol. 9, No. 12, 2009

Figure 2. PL intensities of the blue and green bands versus excitation wavelength.

family. Owing to the large surface-to-volume atom ratio, rearrangement of a large number of surface atoms enables such spherical NC surface to eventually exhibit various different structures as a result of surface energy minimization. This situation is similar on the NC surface with other growth orientations. A Gaussian fit of the size distribution showing the most probable diameter of 3.6 nm is displayed in Figure 1b. Figure 1c depicts the normalized PL spectra of the aqueous suspension acquired at different excitation wavelengths. When the excitation wavelength is shorter than 350 nm, the PL spectra only show a blue band at a fixed peak location. When the excitation wavelength is increased to 350 nm, the blue band exhibits an obvious redshift and an additional green band starts to emerge at 510 nm as a shoulder. As the excitation wavelength is increased, the intensity of the green band gradually increases while the peak position remains unchanged. This blue band has been reported to be due to the band-to-band recombination of photoexcited carriers in the quantum confined 3C-SiC NCs13,14 corresponding to the band gap of the 3C-SiC NCs suspended in water. Figure 2 plots the PL intensities of the blue and green bands versus excitation wavelength. The intensity changes in the green band are similar to those observed from the blue band. Its intensity reaches a maximum when excited by a wavelength of ∼370 nm. This behavior has been previously observed in porous silicon related nanostructured materials.29-32 A typical example is that when the surface of porous silicon forms a SidO double bond, a new electronic state with an energy ∼2.0 eV appears in the band gap.31,32 This pins the PL energy level to that of the surface state. In the luminescence process, the photoexcited carriers are produced in the cores of the quantum-confined porous silicon NCs, whereas the radiative recombination takes place in the surface/defect states. Thus, the photoexcitation efficiency of carriers depends on the NC size, similar to that observed from the blue band in the current PL spectra. According to the behavior of the green band, it is believed to arise from the surface and water molecule states, but the photogenerated carriers occur in the 3C-SiC NC cores. Scanning tunneling microscopy reveals several Si-terminated surfaces on 3CSiC NCs having different structures.21,33 Theoretical studies have predicted that water molecules will dissociate into -OH Nano Lett., Vol. 9, No. 12, 2009

Figure 3. PL spectra of the water suspensions with different pH values, taken under the excitation wavelength of 360 nm.

and -H groups bonding onto the surface of 3C-SiC with Si-terminations.1,19,34 With regard to the C-terminated surface, a similar situation takes place but it depends strongly on the temperature.35,36 In our experiments, a large portion of the new surfaces with Si- and C-terminations is formed after ultrasonic treatment of the etched 3C-SiC grains. At this low temperature, only Si-terminated surfaces are expected to react with H2O molecules to form some OH- and H+ bonding structures.1 To identify the role of the solvent (water), an ethanol suspension of 3C-SiC NCs is produced by the same way but no similar green band or pinned blue PL can be observed, as shown in Figure 1d. The dependence of the blue PL peak position on the excitation wavelength is presented in Figure 1e. It is evident that in the ethanol suspension, the blue PL peak position redshifts linearly with increasing excitation wavelength. The results indicate that the surface/water-molecule states caused by the OH- and H+ bonding processes play an essential role in the emergence of the green band,37 because dissociation of ethanol cannot produce OH- groups. Naturally, only the C-H or Si-H structures should not be the origin of the green band. To identify the role of OH- and H+, we examine the PL spectra of the water suspensions with different pH values and the corresponding results are presented in Figure 3. The green band appears mainly in the acidic suspensions and its intensity relative to the blue band decreases with decreasing pH value. It almost vanishes in the strong acidic suspension with a pH value of 1.0. For the blue band, it is pinned to almost the same position in these acidic suspensions. In contrast, in all the alkaline suspensions, the green band cannot be observed and the blue band shows an obvious redshift with increasing pH value. Appearance of the green band may raise that question why a similar green band was not observed in water suspensions of 3C-SiC NCs described in our previous papers.13,14 The previous suspension is obtained by subjecting a porous 3C-SiC film fabricated by electrochemical etching of a polycrystalline 3C-SiC target to ultrasonic vibration in water. Such a suspension commonly has a pH value of about 7.0. Thus, the green band cannot be observed. In contrast, the current water suspension is obtained by chemical etching of 3C-SiC powders in a mixture of HNO3 and HF. Hence, it often has a pH value smaller than 6 due to incomplete cleaning, thereby leading to the appearance of the green band. The above results imply that 4055

Figure 5. C 1s, Si 2p, and O 1s core level XPS spectra acquired from the water suspension film deposited on a silver substrate. Figure 4. FTIR spectra of the as-fabricated water and ethanol suspensions dropped on the silicon wafers.

a specific concentration range of the OH- and H+ bonding structures is needed to produce the green emission. On the other hand, the introduction of excess OH- can easily damage the surface luminescent centers. The results in Figures 1 and 3 indicate that the green luminescent centers are complexes of the structurally changed sections of the Si-terminated (001) surfaces and absorbed water molecules dissociated into H+ and OH- groups, as they are energetically favorable.1 To obtain more evidence on the existence of -OH and -H groups on the surface of the 3C-SiC NCs and simultaneously investigate the surface structures, we examine the FTIR spectra of the thin films produced from the NC solutions in H2O and ethanol. Figure 4 shows that the two spectra have obvious differences except the strong SiC optical phonon at 813 cm-1 and weak Si-CHdCH2 vibration at 1410 cm-1.38 First, the C-Hx (x ) 1-3) bonds at 2850, 2925, and 2960 cm-1 have high intensities according to the spectrum acquired from the ethanol suspension film.39 In comparison, they are very weak in the spectrum acquired from the water suspension film. Second, a weak broadband at 3000-4000 cm-1 related to Si-O-H and OH vibrations and a sharp 740 cm-1 band from C-Si-H vibration emerge only in the spectrum from the water suspension film.38,40 Third, a narrow and very weak band at 1020 cm-1 emerges from the spectrum taken from the ethanol suspension film whereas it becomes a strong broadband at 1076 cm-1 in the spectrum from the water suspension film. The narrow band can be attributed to C-OH vibration.38 The broadband does not arise from only the longitudinal optical mode of the Si-O-Si stretching vibration since no increased absorption is observed on the side of the 480 cm-1 absorption peak.41 In fact, some shoulders are observed on the high energy side. These shoulders should stem from the Si-OH vibration.38 Finally, a strong broad vibration band at 619 cm-1 appears only in the spectrum taken from the water suspension film. This band is related to the vibration of Si-H bonds.39 The above FTIR results clearly reveal that the Si-terminated 3CSiC NC surfaces are hydrophilic and can dissociate water molecules into -OH and -H groups to bond mainly on the Si-terminated NCs, whereas the C-terminated surface does not interact with water molecules at such a low temperature. Figure 5 shows the C 1s, Si 2p, and O 1s core level XPS 4056

spectra of the water suspension film deposited on a silver substrate. These spectra do not reveal the presence of bonding with water molecules due to careful evaporation and removal of the surface layer from the sample prior to the analysis. In the C 1s spectrum, the strongest peak at 283.2 corresponds to the SiC component,22,27 whereas the two small shoulders at 284.6 and 286.8 eV can be attributed to the chemical bonding of CHn and O-CH3,23 respectively. In the Si 2p spectrum, the strongest peak corresponds to the SiC component, whereas the peak at 100.7 eV can be ascribed to the Si1+ state (a Si atom bonding to one oxygen atom).42,43 Since the 531.7 eV peak in the O 1s spectrum is due to surface hydroxyls,44,45 we suggest that the 100.7 eV peak originates from the chemical bonding of Si-OH. Here, it should be mentioned that the 530.4 eV peak in the O 1s spectrum is related to the Ag film substrate because it matches the chemical bonding of silver oxide (O1s-Ag states, Ag2O).46 These XPS results indicate that the Si-terminated NC surfaces are hydrophilic and are connected to -OH, whereas the C-terminated surfaces are hydrophobic. After integration and correction for the carbon and oxygen contents in a reference sample (a clean silver surface), the Si/C/O atomic ratios are 1:1.01:0.14. Using a shell approximation,47,48 we can calculate the number of the atoms in a NC, NSiC ) 4πd3/3a3, where d and a are the NC diameter and lattice constant (0.436 nm for 3C-SiC), respectively. The NC with a diameter of 3.6 nm has approximately 2356 atoms and the number of silicon and carbon is almost the same at 1178. The Si/O (or C/O) ratio determined from the XPS can be used to calculate approximately the area occupied on the NC surface by each OH- capping ligand. For a ratio of 1:0.14, the 3.6 nm NC with 1178 core Si (or C) atoms have 164 surrounding O atoms. Dividing the NC surface area by 164 indicates that each OH- ligand occupies an average area of 0.252 nm2. This value is about twice that expected for a close packed monolayer of H2O ligands surrounding the NC, indicating that both OH- and H+ ligands coat the NCs with about 50% surface coverage. This is in good agreement with our PL and FTIR spectral analyses. X-ray absorption fine structure is an effective technique to study the near-neighbor local structure in complicated materials.49-52 In particular, it is known that oxygen near K-edge spectrum is sensitive to the change in the coordination Nano Lett., Vol. 9, No. 12, 2009

XANES analysis further confirms our assignment on the NC surface structure.

Figure 6. The O K-edge (a) and Si L3,2-edge (b) XANES spectra obtained from the water suspension film deposited on a silver substrate.

number in silicon oxide.52 Therefore, the Si L3,2-, and O K-edge XANES spectra of the water suspension film are obtained and depicted in Figure 6a,b in which the corresponding XANES spectra from the amorphous SiO2 film deposited by electron beam on the silver thin film are also shown for comparison. The O K-edge spectrum from the SiC sample shows the following main features compared to that observed from SiO2 film. First of all, a new peak emerges at 531.5 eV. This peak has a shoulder at 530 eV. The two peaks have been assigned to an O atom bonding to Si1+ at the NC surface.50,53 This is consistent with our XPS result because the surface of the current 3C-SiC NCs mainly consists of Si1+ states. This indicates that the Si atom at the NC surface has only a nearest oxygen coordination. Second, the intensity of feature a at 538 eV decreases and that of feature c at 544 eV increases. Feature a has been ascribed to resonance stemming from multiple scattering inside two adjacent tetrahedrons with constructive interference in the spectrum of R-SiO2 and feature c to O 2p-Si 3sp hybridization.54,55 The intensity enhancement observed from feature c implies the formation of a large ordered structure on the NC surfaces. This is also in agreement with OH- bonding on the Si-terminated surfaces. In the Si L3,2-edge spectrum, the main features are similar to those observed from the a-SiO2 thin film. A key difference is that a shoulder peak appears at 105.5 eV. Feature A with a double-peak structure appears as a spin-orbit doublet associated with the transitions of Si 2p3/2 and 2p1/2 core states to the antibonding Si 3s derived states.55 When the bridging oxygen in the Si-O-Si bond is replaced by the -OH group on the surface, the Si 2p binding energy decreases. This leads to a downshift of the energy (peak position) of feature A.56 Thus, the appearance of the shoulder peak is a result of -OH groups bonded to the Si-terminated 3C-SiC NC surfaces. The Nano Lett., Vol. 9, No. 12, 2009

To identify the effect of the surface bonding structures in water at different pH values on the PL characteristics, we carry out an ab initio computational study on the Siterminated surface of the 3C-SiC with absorbed H2O molecules or only -OH groups. The calculation is performed using the DFT within the generalized gradient approximation of Perdew, Burke, and Ernzerholf57 under package CASTEP58 in which a plane-wave norm-conserving pseudopotential method59 is used. We use a kinetic energy cutoff of 470 eV to represent the single-particle wave functions. In the water suspensions, since we are interested in the interactions between H2O molecules and surfaces rather than in the size effects of NCs, we adopt slabs to investigate the adsorption of water molecules on the Si- or C-terminated surfaces of the NCs. Since there are various surfaces for the current NC growth orientation and it has been indicated that the surfaces with the Si- and C-terminated (001) growth planes have significantly different hydrophilic properties,1 we focus on them in the DFT studies. The SiC (001) Si- or C-terminated surfaces are represented by slabs carved along the (001) direction with six atomic layers. The top layer is Si- or C-terminated and covered with 1.0 nm thick vacuum slab and the bottom layer is matched to a perfect 10-layer slab of the SiC crystal so that a surface configuration is built up for which both the DFT calculation and the geometric optimization can be carried out within the CASTEP package. There are four planar cells in one on-plane supercell of this structure so that we can study the modification of the Si- or C-terminated surface with a water molecule (or an -OH group) initially absorbed onto the top four Si or C atoms. The geometry of the surface configuration is optimized using BFGS minimizer in the CASTEP package with default convergence tolerances of 2 × 10-5 eV for energy, 0.05 eV/Å for maximum force, and 0.002 Å for maximum displacement.60 Figure 7 shows the results of geometry optimization for (a) bare Si-terminated (001) surface, (b) H2O molecule on Si-terminated (001) surface, (c) -OH group on Siterminated (001) surface, and (d) H2O molecule on Cterminated (001) surface (only several topmost layers of the surfaces are shown). When a water molecule approaches the Si-terminated surface to about 0.1 nm, it is dissociated into OH- and H+ and attached to a Si dimer, similar to that obtained by molecular dynamics.1 On the C-terminated surface, however, the geometry optimization procedure leads to further separation of the H2O molecule from the surface, and in the resulting configuration shown in Figure 7d neither dissociation of H+ and OH- nor bonding to the surface can be seen. This clearly shows the hydrophobic nature of the C-terminated surfaces again consistent with the moleculardynamics calculation. To investigate the situation in the alkaline environment, we also perform the same calculation but with a -OH group instead of water molecule to simulate the situation that H+ ions are mostly neutralized, as shown in Figure 7c. Compared to the situation of a bare Siterminated surface shown in Figure 7a, both the dissociated (-OH, -H) and -OH groups can repel the nearby Si atoms 4057

Figure 7. Optimized surface structures for (a) bare Si-terminated (001) surface, (b) H2O molecule on Si-terminated (001) surface, (c) -OH group on Si-terminated (001) surface, and (d) H2O molecule on C-terminated (001) surface. Yellow biggest balls and red smaller ones in the bulk represent Si and C atoms, while blue bigger balls and green smallest ones are for O and H atoms, respectively. Bond angles and interatomic distances are shown in (a-d).

on the surface thereby changing the related bond lengths and bond angles. However, this type of surface modification involving dissociated (-OH, -H) groups is weaker than that in the case with the -OH ones. With regard to the C-terminated surface, although H2O molecules are not dissociated, they can still repel nearby C atoms on the surface even more strongly. As the green band is associated with dissociated (-OH, -H) groups bonding on the Si-terminated surface, our results suggest that too drastic surface modification does not favor the emergence of the green band. In the geometry optimization, we include one H2O molecule or one -OH group per surface supercell. Since every surface supercell includes 4 Si or 4 C atoms, the intermolecule distance of absorbed groups is about 0.61 nm. On the basis of the optimized geometry, we calculate the band structure of electrons. It is known that the DFT calculation underestimates the energy gaps of SiC and their nanostructures.1,19 To compensate for this, we set the scissors operator in the CASTEP package to be 1.95 eV, which shifts the calculated conduction band up with respect to the valence band in order to get the optical gap to be consistent with experiments.13 The obtained densities of states (DOS) are shown in Figure 8. In the case of H2O attachment to the Si-terminated surface, there appears a peak (marked with A) just below the bottom of the conduction band which is almost absent in the case of H2O attachment to the C-terminated surface and -OH group attachment to the Si-terminated surface. Thus, the transition from this peak to the valence band can contribute a PL line which is slightly lower in energy than the transition from the conduction band to valence band. However, this line should disappear in an alkaline environment in which the valence band edge is shifted upward. To demonstrate this more clearly, we calculate the squared optical matrix element averaged over all polarization directions to mimic the random orientations of NCs in the water suspension. Figure 9 indicates that in the case of dissociated water molecule attachment, a green line at about 2.49 eV (∼510 4058

Figure 8. Calculated densities of states for SiC modified Si- and C-terminated (001) surfaces with H2O molecules and -OH groups.

Figure 9. Calculated squared optical matrix elements for SiC modified Si-terminated (001) surfaces with dissociated H2O molecules and -OH groups.

nm) occurs below the blue line at 2.75 eV of the band-toband transition. In contrast, in the alkaline environment the former disappears but the latter exhibits an evident redshift due to the valence band edge shift. All these features agree well with our experimental results implying that the green line is related to the complex consisting of the rearranged surface together with -OH and -H groups. The lack of attached H+ in the alkaline environment not only eliminates the green line but also reduces the optical gap due to different surface bonding structures, consistent with the trend shown in Figure 3. In the water suspension, there are a large number of NCs with various sizes and different shapes. Consequently, the energy levels of the NCs also spread over into certain ranges. This stochastic situation can be described by a function F(E,V), which is the probability of finding a level in energy E (related to the valence band edge) in the NCs of size (volume) V. Thus, when the suspension is irradiated with energy E, the probability of exciting electrons from the valence band in NCs of size V is proportional to F(E,V). The excited electrons in these NCs will relax to a lower level ε and then emit photons of energy ε by transition back to the Nano Lett., Vol. 9, No. 12, 2009

valence band with a probability proportional to F(ε,V). By averaging over all possible sizes and possible energies, we can obtain a formula for the average emitted photon energy (band-to-band PL line) of the suspension under irradiation j ) AεAF(ε,V)F(E,V)dVdε, where of photons E: E AF(ε,V)F(E,V) in the integrand is the probability of the PL process (including the exciting and the emitting probabilities), and A is the normalization factor. After integration over V j is only a function of photon energy E. In a and ε, E neighborhood of E0, one has a linear expansion of this j (E) ) E j (E0) + R(E0)(E - E0), where the slope function: E R(E0) ) AεAF(ε,V){[∂F(E,V)]/(∂E)}|E)E0dVdε. If R(E0) is positive and constant in the whole studied energy range, we j with E, as reflected by the curve have a linear increase of E obtained from the ethanol suspension and the low-energy part of the curve acquired from the water suspension in Figure 1e. Saturation of the curves in the water suspension in the high-energy part suggests specific reconstruction of NCs with small sizes and it strongly depends on the aqueous environments of the NCs. This leads to more scattering level distribution F(E,V) in this region and a vanishing slope R(E0) after averaging. As the green PL band arises from the modified (001) Siterminal surface and dissociated H2O molecules, a large enough section of the NC surface is necessary. Hence, only the NCs with larger sizes that can accommodate such complexes may contribute to the green band. This explains why the green band can be observed by using longerwavelength excitation which can excite more NCs with larger size. In conclusion, we have examined the PL spectra of the water suspensions of the 3C-SiC NCs with sizes less than 6.5 nm. It is found that in addition to the blue band resulting from the quantum confinement effect, an additional PL band appears at 510 nm when the excitation wavelengths are more than 350 nm. The 510 nm band occurs mainly in the acidic suspensions and vanishes in the alkaline ones. The pH value dependence observed for the PL spectra and first-principle calculations suggest that the 510 nm band is closely associated with the complex of modified (001) Si-terminated surface of 3C-SiC NC and dissociated H2O molecules having the form of OH- on one end of the Si-Si dimer and H+ on the other end. This work is of importance to applications of nanoscale SiC materials. Acknowledgment. The authors sincerely thank Dr. S. D. Wang in the Nanomaterials and Soft Matter Laboratory, Suzhou University and Dr. W. S. Yan in the Synchrotron Radiation Laboratory, University of Science and Technology of China for performing the XPS and XANES measurements, respectively. This work was jointly supported by Grants (60876058,10874071,60676056,BK2008020,andBK2006715) from the National and Jiangsu Natural Science Foundations. Partial support was also from National Basic Research Programs of China under Grants 2007CB936301 and 2006CB921803 as well as Hong Kong Research Grants Council (RGC) General Research Grant (GRF) CityU 112307. Nano Lett., Vol. 9, No. 12, 2009

References (1) Cicero, G.; Catellani, A.; Galli, G. Phys. ReV. Lett. 2004, 93, 016102. (2) Wojtyk, J. T. C.; Tomietto, M.; Boukherroub, R.; Wayner, D. D. M. J. Am. Chem. Soc. 2001, 123, 1535–1536. (3) Gillmor, S. D.; Thiel, A. J.; Strother, T. C.; Smith, L. M.; Lagally, M. G. Langmuir 2000, 16, 7223–72228. (4) Buriak, J. M. Chem. ReV. 2002, 102, 1271–1308. (5) Tiefenauer, L.; Ros, R. Colloids Surf., B 2002, 23, 95–114. (6) Blanchard, A. P.; Kaiser, R. J.; Hood, L. E. Biosens. Bioelectron. 1996, 11, 687–690. (7) Dai, H. J.; Wang, E. W.; Lu, Y. Z.; Fan, S. S.; Lieber, C. M. Nature 1995, 375, 769–772. (8) Zhang, D. Q.; Alkhateeb, A.; Han, H. M.; Mahmood, H.; Mcllroy, D. N.; Norton, M. G. Nano Lett. 2003, 3, 983–987. (9) Xi, G. C.; Peng, Y. Y.; Pan, S. M.; Li, T. W.; Yu, W. C.; Qian, Y. T. J. Phys. Chem. B 2004, 108, 20102–20104. (10) Wang, C. H.; Chang, Y. H.; Yen, M. Y.; Peng, C. W.; Lee, C. Y.; Chiu, H. T. AdV. Mater. 2005, 17, 419–422. (11) Zhou, J.; Liu, J.; Yang, R. S.; Lao, C. S.; Gao, P. X.; Tummala, R.; Xu, N. S.; Wang, Z. L. Small 2006, 2, 1344–1347. (12) Bechelany, M.; Brioude, A.; Stadelmann, P.; Ferro, G.; Cornu, D.; Miele, P. AdV. Funct. Mater. 2007, 17, 3251–3257. (13) Wu, X. L.; Fan, J. Y.; Qiu, T.; Siu, G. G.; Chu, P. K. Phys. ReV. Lett. 2005, 94, 026102. (14) Fan, J. Y.; Wu, X. L.; Li, H. X.; Liu, H. W.; Siu, G. G.; Chu, P. K. Appl. Phys. Lett. 2006, 88, 041909. (15) Fan, J. Y.; Li, H. X.; Jiang, J.; So, L. K. Y.; Lan, Y. W.; Chu, P. K. Small 2008, 4, 1058–1062. (16) Derycke, V.; Soukiassian, P. G. ; Amy, F.; Chabal, Y. J.; D’angelo, M. D.; Enriquez, H. B.; Silly, M. G. Nat. Mater. 2003, 2, 253–258. (17) Chang, H.; Wu, J.; Gu, B. L.; Liu, F.; Duan, W. H. Phys. ReV. Lett. 2005, 95, 196803. (18) Cicero, G.; Galli, G.; Catellani, A. J. Phys. Chem. B 2004, 108, 16518– 16524. Kanai, Y.; Cicero, G.; Selloni, A.; Car, R.; Galli, G. J. Chem. Phys. B 2005, 109, 13656–13662. (19) Reboredo, F. A.; Pizzagalli, L.; Galli, G. Nano Lett. 2004, 4, 801– 804. (20) Peng, X. H.; Nayak, S. K.; Alizadeh, A.; Varanasi, K. K.; Bhate, N.; Rowland, L. B. J. Appl. Phys. 2007, 102, 024304. (21) Soukiassian, P.; Semond, F.; Douillard, L.; Mayne, A.; Dujardin, G.; Pizzagalli, L.; Joachim, C. Phys. ReV. Lett. 1997, 78, 907–910. (22) Rosso, M.; Giesbers, M.; Arafat, A.; Schroen, K.; Zuilhof, H. Langmuir 2009, 25, 2172–2180. (23) Rosso, M.; Arafat, A.; Schroen, K.; Giesbers, M.; Roper, C. S.; Maboudian, R.; Zuilhof, H. Langmuir 2008, 24, 4007–4012. (24) Iijima, M.; Kamiya, H. J. Phys. Chem. C 2008, 112, 11786–11790. (25) Niu, J. J.; Wang, J. N.; Xu, Q. F. Langmuir 2008, 24, 6918–6923. (26) Zinovev, A. V.; Moore, J. F.; Hryn, J.; Pellin, M. J. Surf. Sci. 2006, 600, 2242–2251. (27) Lee, K. H.; Lee, S. K.; Jeon, K. S. Appl. Surf. Sci. 2009, 255, 4414– 4420. (28) Zhu, J.; Liu, Z.; Wu, X. L.; Xu, L. L.; Zhang, W. C.; Chu, P. K. Nanotechnology 2007, 18, 365603. (29) Koch, F.; Petrova-Koch, V.; Muschik, T. J. Lumin. 1993, 57, 271– 281. (30) Kanemitsu, Y.; Uto, H.; Masumoto, Y. Phys. ReV. B 1993, 48, 2827– 2830. (31) Wolkin, M. V.; Jorne, J.; Fauchet, P. M.; Allan, G.; Delerue, C. Phys. ReV. Lett. 1999, 82, 197–200. (32) Wu, X. L.; Xiong, S. J.; Fan, D. L; Gu, Y.; Bao, X. M.; Siu, G. G.; Stokes, M. J. Phys. ReV. B 2000, 62, R7759–R7762. (33) (a) Semond, F.; Soukiassian, P.; Mayne, A.; Dujardin, G.; Douillard, L.; Jaussaud, C. Phys, ReV. Lett. 1996, 77, 2013–2016. (b) Soukiassian, P.; Semond, F.; Mayne, A.; Dujardin, G. Phys, ReV. Lett. 1997, 79, 2498–2501. (34) Okamato, Y. Phys. ReV. B 1999, 60, 10632–10635. (35) (a) Derycke, V.; Soukiassian, P.; Mayne, A.; Dujardin, G.; Gautier, J. Phys. ReV. Lett. 1998, 81, 5868–5871. (b) Derycke, V.; Soukiassian, P.; Mayne, A.; Dujardin, G. Surf. Sci. Lett. 2000, 446, L101–L107. (36) Catellani, A.; Galli, G.; Rigolli, P. L. Phys. ReV. B 2000, 62, R4794– R4797. (37) Shim, H. W.; Kim, K. C.; Seo, Y. H.; Nahm, K. S.; Suh, E. K.; Lee, H. J. Appl. Phys. Lett. 1997, 70, 1757–1759. (38) Arkles, B.; Anderson, R.; Larson, G. L. Infrared Analysis of Organosilicon Compounds: Spectra-Structure Corrections; Petrarch Systems Silanes & Silicones: New York 1987; pp 100-103. 4059

(39) Wu, X. L.; Siu, G. G.; Gu, Y.; Tang, N.; Gao, T.; Bao, X. M. Appl. Phys. Lett. 1999, 74, 827–829. (40) Saleh, R.; Munisa, L.; Beyer, W. Thin Solid Films 2003, 426, 117– 123. (41) Song, H. Z.; Bao, X. M.; Li, N. S.; Wu, X. L. Appl. Phys. Lett. 1998, 72, 356–358. (42) Ohishi, K.; Hattori, T. Jpn. J. Appl. Phys. 1994, 33, L675–L678. (43) Hollinger, G.; Himpsel, F. J. Appl. Phys. Lett. 1984, 44, 93–95. (44) Verne, E.; Bretcanu, O.; Balagna, C.; Biabchi, C. L.; Cannas, M.; Gatti, S.; Vitale-Brovarone, C. J. Mater. Sci: Mater. Med. 2009, 20, 75–87. (45) Jin, R. B.; Wu, Z. B.; Liu, Y.; Jiang, B. Q.; Wang, H. Q. J. Hazard. Mater. 2009, 161, 42–48. (46) Kuo, Y. L.; Chen, H. W.; Ku, Y. Thin Solid Films 2007, 515, 3461– 3468. (47) Holmes, J. D.; Ziegler, K. J.; Doty, R. C.; Pell, L. E.; Johnston, K. P.; Korgel, B. A. J. Am. Chem. Soc. 2001, 123, 3743–3748. (48) Rosso-Vasic, M.; Spruijt, E.; van Lagen, B.; De Cola, L.; Zuilhof, H. Small 2008, 4, 1835–1841. (49) For example, see a review report: Sham, T. K. Int. J. Nanotechnol 2008, 5, 1194–1246. (50) Yamashita, Y.; Yamamoto, S.; Mukai, K.; Yoshinobu, J.; Harada, Y.; Tokushima, T.; Takeuchi, T.; Takata, Y.; Shin, S.; Akagi, K.; Tsuneyuki, S. Phys. ReV. B 2006, 73, 045336.

4060

(51) Sham, T. K.; Gordon, R.; Heald, S. M. Phys. ReV. B 2005, 72, 035113. (52) Lin, J. F.; Fukui, H.; Prendergast, D.; Okuchi, T.; Cai, Y. Q.; Hiraoka, N.; Yoo, C. S.; Trave, A; Eng, P.; Hu, M. Y.; Chow, P. Phys. ReV. B 2007, 75, 012201. (53) Wallis, D. J.; Gaskell, P. H.; Brydson, R. J. Microsc. 1995, 180, 307– 312. (54) Wu, Z. Y.; Jollet, F.; Seifert, F. J. Phys.: Condens. Mater. 1998, 10, 8083–8092. (55) Tsai, H. M.; Ray, S. C.; Pao, C. W.; Chiou, J. W.; Huang, C. L.; Du, C. H.; Pong, W. F. J. Appl. Phys. 2008, 103, 013704. (56) Xiong, W. H.; Peng, J.; Hu, Y. F. Phys. Chem. Minerals [Online early access]. DOI 10.1007/s00269-009-0301-7. (57) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77, 3865–3868. (58) Clark, S. J.; Segall, M. D.; Pickard, C. J.; Hasnip, P. J.; Probert, M. J.; Refson, K.; Payne, M. C. J. Crystallogr. 2005, 220, 567–570. (59) Hamann, D. R.; Schluter, M.; Chiang, C. Phys. ReV. Lett. 1979, 43, 1494–1497. (60) Pfrommer, B. G.; Cote, M.; Louie, S. G.; Cohen, M. L. J. Comput. Phys. 1997, 131, 233–240.

NL902226U

Nano Lett., Vol. 9, No. 12, 2009