Formation of Silver Nanoparticles and Nanocraters on Silicon Wafers

Aug 1, 2006 - Formation of Silver Nanoparticles and Nanocraters on Silicon Wafers. Junhui He†,‡ and Toyoki Kunitake*,†. Frontier Research System...
84 downloads 9 Views 494KB Size
Langmuir 2006, 22, 7881-7884

7881

Formation of Silver Nanoparticles and Nanocraters on Silicon Wafers Junhui He†,‡ and Toyoki Kunitake*,† Frontier Research System, The Institute of Physical and Chemical Research (RIKEN), Hirosawa 2-1, Wako, Saitama, 351-0198 Japan, and Functional Nanomaterials Laboratory, Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Zhongguancun Beiyitiao 2, Haidianqu, Beijing 100080, China ReceiVed April 16, 2006. In Final Form: June 30, 2006 Silver nanocraters and monodisperse nanoparticles were formed on silicon wafers by spin-coating of an aqueous AgNO3/PVA solution and calcination of the resulting Ag+/PVA composite film. The monodisperse Ag nanoparicles were formed from small Ag+/PVA aggregates and were uniformly and stably distributed on the substrate surface. They were located as close as 2.8 nm apart (edge to edge) without coalescence. This nanoparticle stability was apparently derived from their interaction with the oxidized wafer surface. On the other hand, Ag metallic nanocraters with and without nanodots at their centers were produced from large Ag+/PVA aggregates. The explosive decomposition of AgNO3 and PVA by calcination could explain their formation. When Ag+ ions were reduced to Ag nanoparticles prior to calcination, larger Ag nanoparticles were produced probably due to aggregation of closely situated nanoparticles. Those nanoparticles that were located far enough stayed intact. Perspectives are discussed in terms of potential applications.

Introduction

Experimental Section

Nanomaterials are attracting interests of scientists from various fields as electronic, magnetic, mechanical, catalytic, and composite materials, as they have unique physicochemical properties not attainable by conventional materials. Current strategies of fabricating nanosized objects include the well-known top-down approach (e.g., lithographic fabrication) and the bottom-up approach (e.g., self-assembly and nanoscale and atomic scale manipulation by scanning probe techniques).1 Among the latter, the self-assembly approach is becoming more and more attractive, due to the high efficiency of the fabrication process. For example, 1D arrangements,2,3 2D surface patterns,4,5 and 3D superlattices6 were assembled from nanospheres of different sizes. Additional physicochemical treatments can further transform discrete metallic nanoparticles into semicontinuous and continuous nanostructures.7,8 Very recently, we reported fabrication of porous and nonporous metallic nanostructures using cellulose fiber as a template.9 As metal nanoparticles were in situ synthesized and homogeneously distributed in cellulose fibers, networks of metallic nanostructures were readily obtained by subsequent removal of the organic components. It is important to examine whether porous and nonporous metallic films could be fabricated by applying a similar approach on 2D substrate surfaces. In this article, we report our results toward this goal. Unexpectedly, metallic nanoparticles and nanocraters were produced instead of porous and nonporous metallic films.

An aqueous mixture containing 100 mM of AgNO3 and 100 unit mM of poly(vinyl alcohol) (PVA; MW ) ∼78 000) was prepared by mixing aqueous AgNO3 and aqueous PVA. It was spin-coated at 3000 rpm for 2 min on silicon wafer (polished silicon wafers, Mitsubishi, Tokyo, Japan). The film was allowed to dry in air for several hours and then in a vacuum overnight. The surface morphology was observed directly without any metallic coating on a Hitachi S-5200 field emission scanning electron microscope (FESEM) (Hitachi, Tokyo, Japan). The as-prepared Ag+/PVA film on silicon wafer was either reduced in aqueous NaBH4 or subjected to calcination without reduction. In the former case, the film was immersed in 10 mM aqueous NaBH4 for 10 min, rinsed with pure water, and dried in a vacuum overnight. In the latter, film on silicon wafer was placed in a programmable KDF-S70 furnace (Denken, Kyoto, Japan). It was heated at a rate of 2.4 K/min from room temperature to 450 °C, kept at this temperature for 3 h, and finally allowed to cool. XPS measurements were carried out on an Escalab 250 (VG) using Al KR (1486.6 eV) radiation. The applied power was operated at 15 kV and 20 mA. The base pressure in the analysis chamber was less than 10-8 Pa. Smoothing, background removal, and peak fitting were carried out with a VG analysis software package, ECLIPS. All the peaks were corrected with C 1s (285 eV) as the reference. AFM measurements were carried out by noncontact mode on an Explorer scanning probe microscope (TMX2100, TopoMetrix, Santa Clara, CA).

* To whom correspondence should be addressed. E-mail: kunitake@ ruby.ocn.ne.jp. † RIKEN. ‡ Chinese Academy of Sciences. (1) Moriarty, P. Rep. Prog. Phys. 2001, 64, 297-381. (2) Wyrwa, D.; Beyer, N.; Schmid, G. Nano Lett. 2002, 2, 419-421. (3) Teranishi, T.; Sugawara, A.; Shimizu, T.; Miyake, M. J. Am. Chem. Soc. 2002, 124, 4210-4211. (4) Kiely, C. J.; Fink, J.; Brust, M.; Bethell, D.; Schiffrin, D. J. Nature 1998, 396, 444-446. (5) Kitaev, V.; Ozin, G. A. AdV. Mater. 2003, 15, 75-78. (6) Motte, L.; Billoudet, F.; Lacaze, E.; Pileni, M.-P. AdV. Mater. 1996, 8, 1018-1020. (7) Fullam, S.; Cottel, D.; Rensmo, H.; Fitzmaurice, D. AdV. Mater. 2000, 12, 1430-1432. (8) Onoue, S.; He, J.; Kunitake, T. Chem. Lett. 2006, 35, 214-215. (9) He, J.; Kunitake, T.; Watanabe, T. Chem. Commun. 2005, 795-796.

Results and Discussion A colorless precursor solution was prepared by mixing aqueous AgNO3 and aqueous poly(vinyl alcohol) (PVA). Surface plasmon absorption was not observed for the mixed solution (colorless), indicating that Ag+ ions were stable in PVA solution and Ag nanoparticles were not formed at least at room temperature. The precursor solution was spin-coated on a silicon wafer, and the resulting film was allowed to dry in air and then in a vacuum. The surface morphology was observed directly without any metallic coating by scanning electron microscopy (SEM). As shown in Figure 1, two types of nanoparticle are noticeable on the surface of the as-prepared film. One is nearly monodisperse nanoparticles (size: 2 nm) that cover most of the substrate surface.

10.1021/la0610349 CCC: $33.50 © 2006 American Chemical Society Published on Web 08/01/2006

7882 Langmuir, Vol. 22, No. 18, 2006

Figure 1. SEM image of the Ag+/PVA film (one layer) spin-coated on a silicon wafer. Inset: magnified image, where solid and dotted arrows point to large particles and small particles, respectively.

He and Kunitake

Figure 3. SEM image of the Ag+/PVA film (one layer) after reduction by NaBH4. Inset: magnified image, where solid and doted arrows point to large particles and small particles, respectively.

Figure 2. SEM image of the Ag+/PVA film (eight layers) spincoated on a silicon wafer. Inset: magnified image, where solid and dotted arrows point to large particles and small particles, respectively.

The other has a smaller population and larger sizes (mean diameter d ) 7.9 nm, standard deviation σ ) 1.9 nm). As the precursor solution was homogeneous and colorless and spin coating was conducted under nonreductive conditions, these nanoparticles (small and large) must be composed of aggregated AgNO3 and PVA. They fully cover the silicon wafer surface. Multilayer films were prepared by repeating the spin-coating process. The surface morphology of an eight-layer film is shown in Figure 2. The monodisperse nanoparticle keeps its spherical shape, and its size increased to 4 nm. The population and size of the large particle also increased, and the mean diameter (d) and standard deviation (σ) were estimated as 10.3 and 3.3 nm, respectively. Some large particles have nonspherical shapes that may be attributed to aggregation of smaller particles. To transform Ag+ ions into metallic silver, the as-prepared Ag+/PVA film on silicon wafer was immersed in aqueous NaBH4, rinsed with pure water, and dried in a vacuum overnight. Figure 3 shows the surface morphology of the as-treated specimen. It is clear that Ag nanoparticles were produced and distributed on the substrate surface. X-ray photoelectron spectroscopy (XPS) measurements were carried out to check whether the nanoparticles were metallic. The modified Auger parameter (R ) EK(MNN) (356.3 eV) + EB(3d5/2) (368.3 eV)) of silver in the specimen was estimated as 724.6 eV, indicating that the nanoparticles are metallic.10-12 Their mean diameter and standard deviation were estimated to be 3.7 and 1.1 nm, respectively. Clearly, Ag atoms

Figure 4. (a) SEM image of the Ag particle/PVA film (one layer; see Figure 3) after calcination at 450 °C, with inset showing a magnified image of the squared area. (b) Schematic illustration of the changes in surface morphology that occurred upon thermal treatment.

formed by reduction of Ag+ ions were mobile in PVA film and coalesced to form Ag metallic nanopaprticles. The specimen was calcined at 450 °C to remove the organic components. This thermal treatment resulted in a significant change in surface morphology. As shown in Figure 4a, large nanoparticles of ca. 30 nm in size were formed with a lesser population together with areas where large particles were not contained. A magnified image of the area (Figure 4a, squared area) between them, however, shows that many smaller nanoparticles stay intact. We propose a mechanism to explain the observed surface morphology, as schematically illustrated in Figure 4b. The distribution of Ag nanoparticles is not homogeneous on the substrate surface (Figure 3). Those particles that are close enough would collide and fuse with each other upon (10) He, J.; Ichinose, I.; Kunitake, T.; Nakao, A. Langmuir 2002, 18, 1000510010. (11) He, J.; Ichinose, I.; Fujikawa, S.; Kunitake, T.; Nakao, A. Chem. Commun. 2002, 1910-1911. (12) He, J.; Ichinose, I.; Kunitake, T.; Nakao, A.; Shiraishi, Y.; Toshima, N. J. Am. Chem. Soc. 2003, 125, 11034-11040.

Formation of Ag Nanoparticles and Nanocraters

Figure 5. (a) SEM image of the Ag+/PVA film (one layer) after calcination at 450 °C, where solid and dotted arrows point to nanocraters with and without a nanodot at their center, respectively. (b) Magnified image, where arrows point to monodisperse nanoparticles.

thermal treatment, forming a larger nanoparticle and leaving a blank area surrounding it. In contrast, when the particles are located far enough, they would stay intact. In our previous study, nonporous metallic nanostructures were obtained by in situ synthesis (adsorption of metal salts and reduction) of metal nanoparticles in cellulose fiber and subsequent removal of the organic components (calcination). In contrast, porous metallic nanostructures were formed by introduction of metal salts into cellulose fiber and subsequent calcination (without prior reduction to nanoparticles). Simultaneous thermal decomposition of the metal salt (e.g., AgNO3) and organic components led to porous metallic nanostructures in the latter case. Such morphological variation should be interesting as well in the case of organic film matrixes (we expected formation of porous films). Thus, Ag+/PVA thin films on silicon wafer were directly calcined at 450 °C for 3 h. Figure 5 shows the surface morphology of the as-calcined specimen. Very interestingly, nanocraters of ca. 20 nm in diameter are predominantly observed instead of a porous film. A close look shows that some of the nanocraters have a nanodot at their centers, though many of them are empty (Figure 5a). A magnified image (×250 000) shows the existence of monodisperse nanoparticles of 5 nm in addition to the nanocraters (Figure 5b, pointed by arrows). It is known that AgNO3 readily decomposes at temperatures above 440 °C into metallic silver, nitrogen, oxygen, and nitrogen oxides. Thus, the observed nanocraters, nanodots, and monodisperse nanoparticles must consist of metallic silver. In fact, the modified Auger parameter (R ) EK(MNN) (355.9 eV) + EB(3d5/2) (368.5 eV)) of silver was estimated to be 724.4 eV in XPS measurements of the specimen, confirming that they are indeed metallic.10-12 The formation of metallic nanocraters was also confirmed by atomic force microscopy (AFM), as shown in Figure 6. The height profile gives the diameter and the depth of the nanocrater as ca. 20 and ca. 4.5 nm, respectively. The diameter of the nanocrater is nearly five times as large as its height, indicating that it is a rather flat nanocrater. The observed height profile would be affected by its contour, since the AFM tip used is not small enough. Unfortunately, both the nanodot within the crater and the small monodisperse nanoparticles as revealed by SEM were not observed by AFM under the current conditions, probably because the lateral resolution was not sufficient. It is useful to speculate on the mechanism of nanocrater formation at this point (Figure 7). The organic components are removed at 450 °C,13,14 and AgNO3 is decomposed at this temperature into metallic silver, releasing nitrogen, oxygen, and nitrogen oxides. The Ag atoms produced were labile and would

Langmuir, Vol. 22, No. 18, 2006 7883

Figure 6. AFM image and height profile of metallic nanocraters.

Figure 7. Schematic illustration of the formation of Ag nanoparticles and nanocraters on the silicon surface.

aggregate to form Ag clusters. Silicon wafer substrates usually have an ultrathin SiO2 layer on their surface. The oxide surface strongly interacts with metal atoms and will decrease the mobility of metal atoms. The mobility must be reduced with aggregation and the increase in particle size. The metal nanoparticle will be ultimately stabilized by strong bonding interaction between the outmost orbital of its surface atoms and the contacting oxygen atoms of the SiO2 surface.12,15 It was also discussed previously that aggregation of a metal (Ag, Cu) adatom with a metal cluster becomes increasingly difficult with increasing cluster sizes due to enhanced repulsion.16 PVA serves as a binder for formation of Ag+/PVA thin films. Its hydroxyl groups are ligands to Ag+ ions and are essential for homogeneous dispersion of Ag+ ions. The Ag+/PVA particles on a silicon wafer contain NO3- ions and a PVA moiety that would decompose to release gaseous N2, O2, CO2, H2O, and nitrogen oxides upon thermal treatment. The large Ag+/PVA particles may expand upon calcination to collapse eventually to metal nanocraters due to gas release. In contrast, the small Ag+/ PVA particles (ca. 2 nm) produce smaller amounts of Ag atoms and gaseous molecules, only giving monodisperse Ag nanoparticles. The monodisperse feature of these small Ag nanoparticles is in agreement with that of the small Ag+/PVA nanoparticles before calcination. (13) He, J.; Ichinose, I.; Fujikawa, S.; Kunitake, T.; Nakao, A. Chem. Mater. 2002, 14, 3493-3500. (14) He, J.; Fujikawa, S.; Kunitake, T.; Nakao, A. Chem. Mater. 2003, 15, 3308-3313. (15) Franke, R.; Rothe, J.; Pollmann, J.; Hormes, J.; Bo¨nnemann, H.; Brijoux, W.; Hindenburg, Th. J. Am. Chem. Soc. 1996, 118, 12090-12097. (16) Fichthorn, K. A.; Merrick, M. L. Phys. ReV. B 2003, 68, 041404-1041404-4.

7884 Langmuir, Vol. 22, No. 18, 2006

He and Kunitake

Conclusions

promise in catalytic materials, sensors, high-density recording devices, and functional and biocompatible surfaces. In fact, it was found that the persistent current in aluminum nanorings could produce large magnetic fields at the tip of nanotube templates.17 Metallic nanorings may be also used as microarrays for inoculation with biological molecules.18 Ordered patterns of metal nanocraters and monodisperse nanoparticles on substrate surfaces may be additionally advantageous. Further efforts are now in progress in our laboratory.

In summary, metallic nanocraters and monodisperse nanoparticles were formed on silicon wafers by spin-coating and calcination of Ag+/PVA composite films. The monodisperse nanoparicles are homogeneously and stably distributed on the substrate surface and can be located as close as nearly 2.8 nm (edge to edge) without coalescence. This stability appears to come from their interaction with the oxidized wafer surface. On the other hand, metallic nanocraters were produced from large Ag+/PVA nanoparticles. In contrast, if Ag+ ions were reduced to form Ag nanoparticles prior to calcination, no gaseous molecules were released from inside the Ag nanoparticles and, thus, no Ag nanocraters were produced. Instead, larger nanoparticles were formed due to aggregation of neighboring nanoparticles that were close enough. Those nanoparticles that are located far enough, however, stayed intact. Metal nanoparticles, particularly nanocraters that are densely and homogeneously distributed on solid surfaces, are expected to have interesting physicochemical properties and may possess

Acknowledgment. J. He is grateful to the National Natural Science Foundation of China (Grant No. 20471065), “Hundred Talents Program” of CAS, and the President Fund of CAS. We are grateful to Ms. Rie Takaki for assistance in AFM measurements. LA0610349 (17) Bagci, V. M. K.; Gu¨lseren, O.; Yildirim, T.; Gedik, Z.; Ciraci, S. Phys. ReV. B 2002, 66, 045409-1-045409-5. (18) Xu, C. W. PCT Int. Appl., 2003; application WO 2002-US36979 20021115.