Effect of polar surface on the growth of Au

Abstract

The growth of Au on faceted MgO(111) and MgO(100) films was investigated by Auger electron spectroscopy, low energy electron diffraction, scanning electron microscopy, atomic force microscopy, and Raman spectroscopy. On the polar MgO(111) surface, a continuous Au film is easily formed because of the large surface energy compared with MgO(100). Heating the Au films grown on MgO(111) results in the formation of nanoparticles of Au. A relatively high Raman activity of the Au nanoparticles is observed. The size of Au nanoparticles might be adjusted by controlling the density of facets on (111) through changing the thickness of the polar oxide films. Our results provide a new insight into the preparation of nanoparticles of metals by choosing a polar surface as a patterned substrate, which may also be applied to other metal/oxide systems.

---

1. Introduction

The synthesis and characterization of metallic particles with well-dispersed and controllable size on the nanoscale have attracted considerable attention because of their potential application in information storage, sensing, catalysis, photonics, plasmonics, biomedical imaging and so on.^1–7 A major challenge remains in controlling the shape and size of metal nanoparticles, which can effectively tailor the special physical and chemical properties for desired applications. Among various nanoscale materials, gold and silver nanoparticles have been extensively studied due to their importance in many areas such as in catalysis, strong surface plasmon resonance, and surface enhanced Raman scattering (SERS).^8–12

To synthesize gold nanoparticles, various chemical and physical techniques have been employed.^2,10–13 In chemical methods, the size and properties of as-grown Au nanoparticles are strongly associated with many factors, such as the concentrations of reagents and ionic species, surface capping, reaction temperature and the substrate morphology. Physically, related experimental studies on controlling size and shape of nanoparticles are less since the formation of the particles generally results from the native characters of deposited metal and substrate materials.^1,11 The character of metal oxide substrate can play an important role in adjusting the nucleation and growth of gold nanoparticles, and these particles can be activated via charge transfer from the substrate, or storing/releasing oxygen in the process of catalysis.^13,14

The metal oxides with the polar surfaces, e.g. MgO(111), NiO(111), FeO(111), Fe₃O₄(111) and ZnO(0001), are interesting because of net charge within each plane and a dipole moment in the repeat unit perpendicular to the (111) face.^15,16 This dipole moment leads to diverging in the electrostatic energy. As indicated by theoretical calculations, the surface energy is infinite for bulk-terminated polar oxide surfaces and very large even for ultrathin films, leading in a “polar instability”.¹⁷ However, the polar surfaces can be stabilized by surface reconstructions, surface metallization, surface adsorption of charged species as well as faceting into neutral planes.¹³ Those processes can lead to a different environment for local surface atoms compared to the bulk or non-polar terminations, and some peculiar electronic states from such surfaces may appear in the gap of the oxides.¹⁸ Thereby, these complicated cases hinder the understanding of the stability mechanisms of polar surfaces.

Among metal oxides, MgO is widely chosen as the substrate due to its simple crystal structure (rock-salt), easy fabrication and good thermal stability.¹¹ The polar MgO(111) is of hexagonal symmetry with Mg²⁺ or O²⁻ ions on the surface, namely, planes of metal cations alternate with equidistant planes of oxygen anions along the [111] direction. Thus, the polar substrate of MgO(111) can be used to grow metals with specific conformations, and they may drive specific growth modes for the metal particles.^12,19

In this paper, we report Au deposition on MgO(100) and MgO(111) films, and their grown behavior on the different surfaces is compared. The surfaces were studied in situ by using surface analytical technologies including Auger electron spectroscopy (AES) and low energy electron diffraction (LEED) in an ultrahigh vacuum system. The morphology of Au on MgO films were imaged ex situ by scanning electron microscopy (SEM) and atomic force microscopy (AFM). The results indicated that the Au forms smaller particles on (100) face but layer-like on (111) face at initial coverages. At higher coverages, the Au epitaxially grows on the films as Au(100)/MgO(100) and Au(111)/MgO(111), and heating those surfaces leads to formation of nanoparticles of Au. The surface activity of Au particles grown on the films was checked by SERS using 4-methylbenzenethiol (4-MBT) as the probe molecules.

2. Experimental

The experiments were carried out in an ELS-22 (Leybold-Heraeus GmbH) ultrahigh vacuum chamber with a base pressure of 1 × 10⁻¹⁰ mbar.^16,19,20 The chamber is equipped with an Auger electron optics with single-pass cylindrical mirror analyzer for AES (PHI 3017, Φ ULVAC-PHI, Inc.) and an optics for LEED. In AES measurements, primary electron beam energy of 3 keV was used.

The single crystals of Mo(110) and (100) were used as the substrates to grow MgO(111) and MgO(100) films, respectively. The sample temperature was monitored via a C-type (W–5% Re/W–26% Re) thermocouple by spot-welding to the edge of the substrates. Before MgO films growth, the substrates were annealed at ∼1200 K in ∼10⁻⁷ mbar oxygen using an electron beam heater, followed by a subsequent flash to 1500 K without oxygen until no impurities were detected by AES and a sharp Mo(110)-(1 × 1) or Mo(100)-(1 × 1) LEED pattern was observed. The metal Mg and Au sources were made of a pure magnesium ribbon (purity > 99.9%) and gold wire (purity > 99.95%) wrapped tightly around a tungsten wire, respectively. The metal sources were thoroughly degassed before deposition. The deposition rates of Mg and Au were about 0.2 and 0.1 monolayer (ML) per min, respectively, calibrated via a quartz crystal oscillator. However, the deposition of Au on magnesia films is not necessarily layer-by-layer growth, therefore, the monolayer equivalent (MLE) is used to scale the Au thickness/coverage, which means that 1 MLE is not equal to one complete single Au atomic layer covered on the MgO surfaces.

The MgO(111) and (100) thin films with ∼20 ML thickness were epitaxially grown on the Mo(110) and (100) substrates by evaporating Mg in ∼10⁻⁶ mbar O₂ at 500 K, followed by annealing at 650 K in ∼10⁻⁷ mbar O₂, respectively. Then, the gold was grown on as-prepared MgO films at room temperature (RT). For comparison, all the experiments of Au–MgO(111) and Au–MgO(100) systems were performed under the same conditions.

The samples were characterized in situ by AES and LEED, and some of them were analyzed ex situ using a field emission SEM (XL30 S-FEG, FEI Company), the AFM (Alone AFM, MFP-3D™-Stand), and the Raman spectroscopy (JY-T64000). In Raman measurements, the SERS was checked by recording Raman spectra of 4-MBT molecules, which were absorbed on the sample surfaces by immersing the sample into 0.01 M 4-MBT ethanol solutions for 2 hours and then washing the surfaces with ethanol. The laser line of 633 nm wavelength was used as the excitation source, and the instrument resolution is better than 1 cm⁻¹. The data in all experiments were collected at RT.

3. Results and discussion

We have previously reported the preparation of MgO(111) and (100) films on Mo(110) and (100) substrates, respectively.^16,19–21 The films were characterized in situ by using various surface analytical technologies including X-ray photoelectron spectroscopy, high resolution electron energy loss spectroscopy, ultraviolet photoelectron spectroscopy, AES, and LEED.

The MgO(111) is of polarity differing from neutral MgO (100) face, which is structurally constructed by alternative stacking of anion and cation layers. Thus, the MgO(111) face has a net dipole moment normal to the surface. The (111) surface is unstable and easily results in a divergence in the surface energy to form {100} facets as observed by SEM.²² Also, the surface of MgO(111) films consists of faceted structure with (100), (010) and (001) face as we described previously.²¹ However, the density and size of the facets depend on the thickness of films because the electric dipole moment layers along the [111] directions are thickness dependent, i.e. below a critical thickness, stable ultrathin polar films without facets can be obtained.¹⁶

Fig. 1(a) and (b) show the SEM images with different coverage of Au on MgO(100) surface. For an initial deposition of 3 MLE Au on MgO(100) surface, it exhibits a surface consisting of compact islands with the about same size 5–8 nm, as seen in Fig. 1(a). At 20 MLE of Au coverage, a worm-lake structure appears (Fig. 1(b)), and a corresponding (1 × 1) LEED pattern with a surface structure of square symmetry is obtained (inset of Fig. 1(b)), showing the epitaxial correlation of Au(100) on MgO(100).

Fig. 1 SEM images (300 × 300 nm²). (a) 3 MLE and (b) 20 MLE Au on MgO(100) films; (c) 3 MLE and (d) 20 MLE Au on MgO(111) films. The insets in (b) and (d) give the corresponding LEED patterns with E_p = 88 and 72 eV, respectively.

Because of the weak interaction between metal and insulating substrate surface, e.g. SiO₂, Al₂O₃ and MgO(100), the growth processes of metal films pass through a sequence of morphological evolution with the increase of metal coverage.^23–26 At the initial stage of growth, metal atoms form isolated and compact islands as seen on (100) face. When more metal atoms are deposited, these islands become bigger and further coalesce into larger but compact islands. Once the quantity of deposited metal reaches critical value, these large islands start to form elongated islands and further develop into percolating structure with gaps/holes. Finally, these gaps/holes are filled and the entire metal films are formed. The whole evolution processes have been clearly exhibited by SEM.^24,27 A similar behavior has been found for Au deposition on TiO₂(110) too.²⁸

Au was also deposited on faceted MgO(111) films with the same experimental parameters, and measured by SEM (Fig. 1(c) and (d)). At an initial stage of growth (3 MLE of Au), a layer-like surface was observed (Fig. 1(c)). With increasing Au coverage to 20 MLE, the gaps/holes appear on the surface as shown in Fig. 1(d). The inset in Fig. 1(d) gives the corresponding LEED pattern of as-grown 20 MLE Au on the faceted-MgO(111) surface. The clear (1 × 1) LEED pattern with hexagonal symmetry indicates an epitaxial growth of Au(111) films.

It is notable that we observed the completely different growth mode for Au on faceted MgO(111) and (100) surfaces at initial Au coverage. The growth mode of metal on oxide is vital not only to surface structure but also to surface energy. Thermodynamically, the growth mode of Au on MgO relates to their surface energies of the oxide substrate (γ_MgO), the growing metal (γ_Au), as well as the interfacial energy between them (γ_i).¹ According to the equilibrium condition of growth, the total energy of the system is written as S = γ_Au + γ_i − γ_MgO.²⁹ For S < 0 a layer-by-layer growth mode is followed, while for S > 0 a three-dimensional (3D) structure appears. In general, metal has higher surface energy than oxide so that most of metals form 3D islands on oxide surfaces. For the (100) face of MgO, there is little interaction with Au due to its low surface energy (γ_MgO(100) = 0.9 J m⁻²), which induces the 3D nucleation.³⁰ In our experiments, it is most likely that the (111) face is O terminated since we prepared the MgO films in O₂. For the polar (111) surface of MgO, its large surface energy (γ_MgO(111) = 5.3 J m⁻² for O-terminated surface) benefits the formation of two-dimensional (2D) films.³¹ Because the faceted (111) face has higher surface energy than that of MgO(100), the Au growth on the faceted MgO(111) is flatter, which is similar to that in Ag/MgO system.²¹ Our results clearly illustrate that the surface energy plays an important role in growth mode of Au on MgO surfaces.

The 20 MLE Au films on MgO(100) and (111) surfaces were step-by-step annealed from RT to 1300 K. The temperature ramp is about 5–10 K per second, and it stayed for 20 minutes at each given temperature. Fig. 2 shows the intensity ratio of NOO Auger line of Au (∼74 eV) to KLL Auger line of O (∼510 eV) as a function of annealing temperatures. Considering that the mean free path of the Auger electrons of Au and O is about several nm, the detected Auger signals are from surface. Therefore, the Auger line intensity ratio (I_Au/I_O) can reflect the relative surface area occupied by Au atoms on MgO before and after annealing. For both of Au/MgO(100) and Au/MgO(111), the Auger signal of Mg or O is almost not detected due to thick Au layer covered on the surfaces before heating. However, in case of Au/MgO(100), the value of I_Au/I_O rapidly decreases as increasing the temperature, and kept stable from 750 to 950 K. At the temperature higher than 950 K, the ratio decreases again. For Au/MgO(111) system, the change of the I_Au/I_O as a function of temperature can be obviously divided into three stages as seen in Fig. 2. The intensity ratio decreases rapidly from RT to 550 K, and follows a slow decrease from 550 to 750 K (marked region I in Fig. 2). From 750 to 1100 K, the value of the ratio almost keeps constant (region II). When the annealing temperature is higher than 1100 K, the value of the ratio decreases again until it reaches zero at 1250 K (region III). Accordingly, the decrease of the I_Au/I_O is an indication of either desorption, agglomeration or diffusion of Au into the substrate. At the temperatures <700 K, desorption of Au is excluded because of an intrinsic thermal stability of Au. A diffusion of Au into MgO can also be excluded since the atom radius of Au is larger than these of Mg and O. Therefore, it is most likely that the Au is agglomerated upon heating.

Fig. 2 Intensity ratio of Au_NOO to O_KLL Auger lines as a function of annealing temperature for 20 MLE Au films on MgO(100) and (111) surfaces, respectively. For Au on MgO(111), the evolution passes through three steps: agglomeration at 400–700 K, thermal stable at 700–1100 K and desorption above 1100 K.

Fig. 3(a) shows the SEM image obtained after annealing the 20 MLE Au on MgO(100) at 700 K. An agglomeration of Au is observed after annealing, which leads to a formation of the Au nanoparticles with different size (20–80 nm). The area of MgO surface covered by gold is obviously decreased after annealing, which is consistent with the I_Au/I_O change in AES. The evolution of temperature-induced gold particles is further confirmed by using AFM as shown in Fig. 3(b). The height of Au particles is about 20–30 nm. Fig. 4(a) shows the SEM image after annealing the 20 MLE Au films on MgO(111) surface. Compared with the Au/MgO(100) system, the density of Au particles is much less, and the particle size is larger (100–200 nm) and the bare area of MgO(111) surface is obviously more. This result agrees well with the statistic AES data shown in Fig. 2. It is noted that the AFM image from Fig. 4(b) shows the average height of Au particles on MgO(111) surface is 80–100 nm, which is greater than these on MgO(100) surface. From SEM image, the Au particles are of hexagonal symmetry because the inherent crystallinity of gold or silver in nanometer scale easily results in the growth along [111] direction.^32,33 The results confirm that Au films are agglomerated to bigger particles on the faceted MgO(111) surface after annealing. On the (111) face, the value of I_Au/I_O from AES at 500–1150 K is lower compared with (100), which also indicates a formation of the bigger Au particles as seen from SEM and AFM. It is generally accessible that the density of particles decrease and average particle size becomes larger with the increasing temperature, i.e. the agglomeration and coalescence of Au atoms.³⁴

Fig. 3 (a) SEM images (1.5 × 1.5 μm²) of Au nanoparticles on MgO(100) surface after annealing 20 MLE Au films at 700 K; (b) corresponding AFM image (1 × 1 μm²).

Fig. 4 (a) SEM images (1.5 × 1.5 μm²) of Au nanoparticles on MgO(111) surface after annealing 20 MLE Au films at 700 K; (b) corresponding AFM image (1 × 1 μm²).

The Au particle size and shape might be varied with other growth parameters such as deposition rate of Au, substrate temperature as well as vacuum conditions. For instance, as reported, the slowly deposited films produce higher SERS intensities than rapidly deposited films.³⁵ In present study, we have a low, fixed deposition rate and the deposition of Au was carried out only at RT under ultra-high vacuum system. Therefore, the change of the Au morphology we found is dominated by different structures of the substrate.

Fig. 5 shows the SERS spectra for 4-MBT absorbed on Au/MgO(100) and Au/MgO(111), respectively. For a Raman spectrum of 4-MBT, there are two stronger bands around 1078 and 1590 cm⁻¹, corresponding to the C–S bonding and C–C aromatic ring bonding, respectively.^36,37 While 4-MBT is absorbed on 3 MLE Au covered MgO films, the main Raman signals of 1082 and 1595 cm⁻¹ on (100) face (curve 1 in Fig. 5(a)) are higher compared with (111) surface (curve 1 in Fig. 5(b)). It is known that the intensity of Raman signal depends on the surface roughness, particle size and shape.^11,12 The higher Raman signal suggests a rougher surface, which benefits the absorption and vibration of 4-MBT. When the thickness of Au reaches 20 MLE, there is an obvious increase of Raman signal on MgO(100) surface (curve 2 in Fig. 5(a)) due to a rougher surface (the SEM image is given as inset). In this case, the Raman peaks of 1182 cm⁻¹ (SO₂), 1331 cm⁻¹ (CH), 1376 cm⁻¹ (CH₃) and 1486 cm⁻¹ (OCH₃) are observed. In contrast, for 20 MLE Au on MgO(111) surface a weaker Raman signal testifies the formation of continuous Au films on MgO(111) surface (curve 2 in Fig. 5(b)), which is almost the same as that of 3 MLE Au. In principle, there are two main mechanisms responsible for SERS: electromagnetic (EM) mechanism and chemical charge transfer (CT) mechanism.^11,37,38 The former relates to the excitation of surface plasmon on metal nanoparticles, in which the localized electromagnetic field is obviously enhanced through the surface plasmon resonance. The latter involves the charge transfer between nano-particles and the polarized molecules. For both mechanisms, to form nano-structures is necessary in order to obtain the enhanced Raman signals. The weaker Raman peaks from curve 2 in Fig. 5(b) reveal a smoother Au layer on MgO(111) surface before annealing. Those results are in good agreement with the SEM and AFM results.

Fig. 5 SER spectra of 4-MBT adsorbed on Au/MgO(100) (a), and Au/MgO(111) (b). Curves 1–3 in (a) and (b) represent 3, 20 MLE Au, and Au particles obtained by annealing 20 MLE Au films at 700 K, respectively. The corresponding SEM images are given as inset.

Fig. 5(a) and (b) also show SERS spectra after annealing the 20 MLE Au on MgO(100) and MgO(111) at ∼700 K, respectively. On the Au/MgO(100) surface, the intensity of main Raman signal from 4-MBT is reduced because of decrease of Au particles (curve 3 in Fig. 5(a)). However, on Au/MgO(111) surface, the SERS signal is remarkably enhanced compared with that before annealing (curve 3 in Fig.5(b)). The density of Au particles on MgO(111) is much less than that on MgO(100) as shown in Fig. 3 and 4. The stronger Raman signals on MgO(111) indicate an obvious Raman activity of Au nanoparticle. In addition, the SERS spectra obtained from these different gold nano-structures are always dominated by the vibrational bands of C–S and C–C bonding (at 1078 and 1590 cm⁻¹), implying that the enhancement occurs principally via an EM mechanism.³⁹ The weak enhancement of the other peaks (in-plane C–H bending, stretch C–C bending, stretch C–C + in-plane C–H bending of the benzene ring) excludes the CT mechanism of the metal to the adsorbed 4-MBT molecules. However, although the change of Raman signal further approves the results obtained by SEM and AFM, a designed study is necessary to identify the Raman activity with Au nanoparticles in details.

One interesting result is that the Au particles on MgO(111) have a wide window of thermal stability between 750 and 1100 K, as shown in Fig. 2. Compared with Au on MgO(100), the formation of the Au nanoparticles on MgO(111) is strongly associated with the polar character. As we discussed above, the MgO(111) films is of instability with a larger surface energy, which is easily diverged into {100} facets to decrease its surface energy.^16,21,22 The faceted MgO(111) surface containing {100} facets looks like a “trigonal pyramid” shape as described in ref. 21. While Au is initially deposited on this surface, the Au nucleates in the valley, i.e. the bottom of “trigonal pyramid” rather than on {100} facets because the islands can in fact relax strain more effectively in the valley.⁴⁰ Further growth of Au results in the formation of a continuous Au(111) films. During higher temperature annealing, the increasing thermal energy enhances the lattice vibration and atoms transfer on MgO(111) surface. This accelerates the migration of Au atoms and prompts the formation of lager Au particles.

It seems most likely that the facets play an important role in the observed behavior. Since formation of the facets reduces the surface polarity, the Au should adhere less strongly to them. Thus, the activation energy for desorption of Au from the {100} facets is lower than on (111) terraces causing Au atoms to migrate away from the facets and toward terrace sites giving rise to nanoparticles. A similar result is also reported recently that Au transformed into a fairly uniform distribution of ellipsoidal nanoparticles after annealing on SiN substrate.⁴¹ The absence of small Au particles on MgO(111) surface illustrates that the lattice vibration of MgO(111) is too strong to make small particles stable because the metal bonding as well as their formation energy are weak in smaller particles, making them coalesce into bigger ones. During formation of larger Au particles, the interfacial interaction between Au and oxide surface should be weakened. Those Au particles are stable because of the role of strong metal bonding between gold atoms until they start to desorb at >1100 K. Moreover, we can see that the Auger intensity of I_Au/I_O starts to decrease at temperature >950 K for Au particles on MgO(100) due to the smaller particle size (as shown in Fig. 2), which also clearly indicates that the thermal stability of Au particles is size dependent.

As we discussed above that the density of facets on (111) face is films thickness-dependent, the thinner the films, the weaker the polarity. Also, the density and the size of the facets are related to the surface polarity, i.e. increasing the films thickness results in strong polarity and bigger facets.¹⁶ Thus, the size of Au particles grown on those faceted surfaces will be varied accordingly. The polar surface can be used for the metal particles growth. The polar surface with different structure, for instance the faceted surface, can affect the metal particles in size.

4. Conclusions

We have investigated the effect of MgO films with different surface structure on the size of Au particles. The results indicate that, compared with neutral MgO(100), the MgO(111) surface is of advantage to form continuous Au films owing to the larger surface energy. The polar instability-induced facets structure can be used as patterned-substrate to direct the preferential nucleation sites for nano-material preparation. These Au nanoparticles are of a high Raman activity. It is expected that the nanoparticles of Au with different size can be prepared by adjusting the thickness of polar films. This study develops a new method to prepare nano-metal materials by using the instability character of polar surfaces, which can also be used in other metal-oxide systems.

Acknowledgements

We acknowledge with pleasure the support of this work by the National Science Foundation of China (NSFC) (21273276) and National Basic Research Program of China (2012CB921700).

References

1. C. T. Campbell, Surf. Sci. Rep., 1997, 27, 1.
2. R. Sardar, A. M. Funston, P. Mulvaney and R. W. Murray, Langmuir, 2009, 25, 13840.
3. B. Yoon and U. Landman, Phys. Rev. Lett., 2008, 100, 056102.
4. W. Yan, S. H. Mahurin, Z. Pan, S. H. Overbury and S. Dai, J. Am. Chem. Soc., 2005, 127, 10480.
5. F. Pezzoli, M. Stoffel, T. Merdzhanova, A. Rastelli and O. G. Schmidt, Nanoscale Res. Lett., 2009, 4, 1073.
6. B. Lim, H. Kobayashi, T. Yu, J. Wang, M. J. Kim, Z. Y. Li, M. Rycenga and Y. Xia, J. Am. Chem. Soc., 2010, 132, 2506.
7. Y. Han, D. S. He and A. Y. Li, J. Nanopart. Res., 2013, 15, 1941.
8. Y. Pan, Y. Gao, G. Wang, D. Kong, L. Zhang, J. Hou, S. Hu, H. Pan and J. Zhu, J. Phys. Chem. C, 2011, 115, 10744.
9. M. M. Schubert, S. Hackenberg, A. C. van Veen, M. Muhler, V. Plzak and R. J. Behm, J. Catal., 2001, 197, 113.
10. C. Xu, W. S. Oh, G. Liu, D. Y. Kim and D. W. Goodman, J. Vac. Sci. Technol. A, 1997, 15, 1261.
11. N. Zhao, Y. Wei, N. Sun, Q. Chen, J. Bai, L. Zhou, Y. Qin, M. Li and L. Qi, Langmuir, 2008, 24, 991.
12. P. N. Njoki, I.-Im S. Lim, D. Mott, H.-Y. Park, B. Khan, S. Mishra, R. Sujakumar, J. Luo and C.-J. Zhong, J. Phys. Chem. C, 2007, 111, 14664.
13. L. Giordano, G. Pacchioni, J. Goniakowski, N. Nilius, E. D. L. Rienks and H.-J. Freund, Phys. Rev. Lett., 2008, 101, 026102.
14. G. Prevot, O. Meerson, L. Piccolo and C. R. Henry, J. Phys.: Condens. Matter, 2002, 14, 4251.
15. C. Noguera, J. Phys.: Condens. Matter, 2000, 12, R367.
16. M. Xue and Q. Guo, J. Chem. Phys., 2007, 127, 054705.
17. J. Goniakowski, F. Finocchi and C. Noguera, Rep. Prog. Phys., 2008, 71, 016501.
18. C. Noguera and J. Goniakowski, J. Phys.: Condens. Matter, 2008, 20, 264003.
19. M. Xue, Q. Guo, K. Wu and J. Guo, J. Cryst. Growth, 2009, 311, 3918.
20. K. Zheng, Y. Yu, Q. Guo, S. Liu, E. G. Wang, F. Xu and P. J. Moller, J. Phys.: Condens. Matter, 2005, 17, 3073.
21. M. Xue, S. Liu, J. Guo and Q. Guo, RSC Adv., 2013, 3, 18916.
22. V. E. Henrich, Surf. Sci., 1976, 57, 385.
23. X. Yu, P. M. Duxbury, G. Jeffers and M. A. Dubson, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 44, 13163.
24. K. Luo, X. Lai, C.-W. Yi, K. A. Davis, K. K. Gath and D. W. Goodman, J. Phys. Chem. B, 2005, 109, 4064.
25. F. Cosandey, L. Zhang and T. E. Madey, Surf. Sci., 2001, 474, 1.
26. Y. Han, R. Ferrando, Z. Y. Li and J. Phys, Chem. Lett., 2014, 5, 131.
27. G. Jeffers, M. A. Dubson and P. M. Duxbury, J. Appl. Phys., 1994, 75, 5016.
28. L. Zhang, F. Cosandey, R. Persaud and T. E. Madey, Surf. Sci., 1999, 439, 73.
29. W. Weiss and W. Ranke, Prog. Surf. Sci., 2002, 70, 1.
30. J. Goniakowski and C. Noguera, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 60, 16120.
31. A. Gibson, R. Haydock, J. P. LaFemina and J. Vac, Sci. Technol, 1992, 10, 2361.
32. F. Cosandey, L. Zhang and T. E. Madey, C-MRS Int. Symp. Proc., 1997, 440, 383.
33. L. Armelao, D. Barreca, G. Bottaro, A. Gasparotto, S. Gross, C. Maragno and E. Tondello, Coord. Chem. Rev., 2006, 250, 1294.
34. M. Zinke-Allmang, L. C. Feldman and M. H. Grabow, Surf. Sci. Rep., 1992, 16, 377.
35. P. Dawson, K. Alexander, J. Thompson, J. Haas III and T. Ferrell, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 44, 6372.
36. M. A. Khan, T. P. Hogan and B. Shanker, J. Nano Sys. Technol., 2009, 1, 1.
37. J. Stropp, G. Trachta, G. Brehm and S. Schneider, J. Raman Spectrosc., 2003, 34, 26.
38. X.-L. Tang, P. Jiang, G.-L. Ge, M. Tsuji, S.-S. Xie and Y.-J. Guo, Langmuir, 2008, 24, 1763.
39. M. Bechelany, P. Brodard, J. Elias, A. Brioude, J. Michler and L. Philippe, Langmuir, 2010, 26, 14364.
40. H. Hu, H. J. Gao and F. Liu, Phys. Rev. Lett., 2008, 101, 216102.
41. P. Pavaskar, I.-K. Hsu, J. Theiss, W. H. Hung and S. B. Cronin, J. Appl. Phys., 2013, 113, 034302.

---