Temperature Dependence of Formation of Nanorods and Dots of

Apr 30, 2004 - Akihito Imanishi , Takeshi Hayashi , Kenta Amemiya , Toshiaki Ohta and Yoshihiro ... Akihito Imanishi , Satoshi Yamane and Yoshihiro Na...
0 downloads 0 Views 648KB Size
Download PDF
4604

Langmuir 2004, 20, 4604-4608

Temperature Dependence of Formation of Nanorods and Dots of Iodine Compounds on an H-Terminated Si(111) Surface in a Concentrated HI Solution Akihito Imanishi,† Takeshi Hayashi,† and Yoshihiro Nakato*,†,‡ Division of Chemistry, Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka 560-8531, Japan, and Core Research for Evolutional Science and Technology (CREST), JST, Japan Received September 18, 2003. In Final Form: March 16, 2004 Immersion of atomically flat, H-terminated Si(111) surfaces in 7.6 M HI for 0.5-4 h caused spontaneous formation of nanosized clusters at the Si surface. X-ray photoelectron spectroscopy analysis suggested that the clusters were composed of silicon iodides (such as SiHxI4-x), produced most probably by Si etching with HI. Atomic force microscopy inspection revealed that the immersion at a low temperature below about 30 °C led to the formation of long rod-shaped clusters, oriented in the 〈1 h1 h 2〉 direction or equivalents, whereas the immersion at a high temperature above 30 °C led to the formation of circular dot clusters, their size and shape changing abruptly at about 70 °C. It is shown experimentally that the formation of dot clusters at a high immersion temperature is explained on the basis of thermodynamics, whereas that of oriented rod clusters at a low temperature is explained by a kinetics-controlled mechanism.

Introduction Formation of regulated nanostructures on semiconductor surfaces constitutes the basis of future hugely integrated and intelligent devices.1,2 Conventional photolithography has been used for microfabrication for a long time, but this method has the severe limit in that the control size is limited to about 50 nm3. Recently, atomicscale fabrication by use of surface probe microscopes such as scanning tunneling microscopy (STM) and scanning near-field optical microscopy (SNOM)4,5 has been attracting much attention as a powerful method for nanostructuring, but this method has also a severe limit in that it is not adaptable to mass production, which is eventually necessary for practical application. We have another method for nanostructuring, namely, a chemical method to make use of self-assembling or selforganizing abilities of molecular systems. This method has a strong merit in that it is adapted to mass production. In addition, this method is based on chemical properties of substances and thus resultant nanostructures are expected to have high chemical stability and self-restoration ability, contrary to the cases of physical methods such as photolithography and surface probe methods. A number of studies have been reported on spontaneous formation of aligned nanostructures of metals or semiconductors on solid surfaces by deposition under ultrahigh vacuum * To whom correspondence may be addressed. Fax: +81-6-68506236. E-mail: [email protected]. † Osaka University. ‡ CREST, JST. (1) Nanomaterials: Synthesis, Properties and Applications; Edelstein, A. S., Cammarata, R. C., Eds.; Institute of Physics Publications: London, 1998. (2) Electrochemistry of Nanomaterials; Hodes, G., Ed.; John Wiley & Sons, Inc.: Weinheim, 2001. (3) Marumoto, K.; Yabe, H.; Aya, S.; Kise, K.; Ami, S.; Sasaki, K.; Watanabe, H.; Itoga, K.; Sumitani, H. J. Vac. Sci. Technol., B 2003, 21, 207. (4) Hosaka, S.; Hosoki, S.; Hasegawa, T.; Koyanagi, H.; Shintani, T.; Miyamoto, M. J. Vac. Sci. Technol., B 1995, 13, 2813. (5) Kolb, D. M.; Ullmann, R.; Ziegler J. C. Elecrochim. Acta. 1998, 43, 2751.

(UHV) conditions.6,7 On the other hand, little is known on spontaneous formation of ordered nanostructures at the solid/solution interface,8,9 except for the formation of nanorods or dots by making use of appropriate templates.10,11 Recently we discovered some unique examples of spontaneous formation of ordered nanostructures at the solid/solution interfaces: (1) the formation of oriented nanowires of nickel on H-terminated atomically flat Si(111) surfaces by electrodeposition in an acidic solution;12,13 (2) the formation of oriented nanorods on the H-terminated Si(111) surfaces during immersion in 7.6 M HI at room temperature (∼25 °C);13,14 (3) the formation of regularly oriented nanoholes and grooves at the singlecrystal n-TiO2(rutile) surface by photoelectrochemical etching in aqueous H2SO4.15 Interestingly, in the second example, only round nanodots were formed when the Si(111) was immersed in 7.6 M HI containing 0.05 M I2, contrary to the case of immersion in 7.6 M HI with no I2. The above examples of spontaneous formation of oriented nanostructures are of great interest in view of exploration of the self-organizing ability of molecular systems and its development to nanotechnology. In the present work, we have made detailed studies on the formation of oriented nanorods on the H-Si(111) surfaces when they were immersed in 7.6 M HI, with an emphasis placed on the elucidation of the growth mechanism. The (6) Mo, Y.-W.; Savage, D. E.; Swartzentruber, B. S.; Lagally, M. G. Phys. Rev. Lett. 1990, 65, 1020. (7) Mundschau, M.; Bauer, E.; Telieps, W. Surf. Sci. 1989, 213, 381. (8) Hara, K.; Ohdomari, I. Jpn. J. Appl. Phys. 1998, 37, L1333. (9) Morisawa, K.; Ishida, M.; Yae, S.; Nakato, Y. Electrochim. Acta 1999, 44, 3725. (10) Martin, C. R.; Mitchell, D. T. Electroanal. Chem. 1999, 21, 2. (11) Nikoobakht, B.; Wang, Z. L.; El-Sayed, M. A. J. Phys. Chem. B 2000, 104, 8635. (12) Imanishi, A.; K. Morisawa and Y. Nakato, Electrochem. SolidState Lett. 2001, 4, C69. (13) Nakato, Y.; Murakoshi, K.; Imanishi, A.; Morisawa, K. Electrochemistry 2000, 68, 556. (14) Imanishi, A.; Ishida, M.; Zhou, X.; Nakato, Y. Jpn. J. Appl. Phys. 2000, 39, 4355. (15) Tsujiko, A.; Kisumi, T.; Magari, Y.; Murakoshi K.; Nakato Y. J. Phys. Chem. B 2000, 104, 4873.

10.1021/la035749c CCC: $27.50 © 2004 American Chemical Society Published on Web 04/30/2004

Formation of Nanorods of Iodine Compounds

Langmuir, Vol. 20, No. 11, 2004 4605

investigation of the effect of the immersion temperature has revealed that the size and shape of clusters and hence the growth mechanism drastically change with the immersion temperature. Experimental Section Two-types of single-crystal n-Si(111) wafers were used, both having a resistivity of 10-15 Ω‚cm. One is normal n-Si wafers having more or less vicinal (111) surfaces with no specified tilting directions and angles, obtained from Shin-Etsu Handoutai Co. Ltd., Japan. The other is n-Si wafers having well-specified vicinal (111) surfaces, tilting in the 〈1 h1 h 2〉 direction at angles of 0.36 ( 0.1° and 4.0 ( 0.1°, obtained from Osaka Tokushu Gokin Co. Ltd., Japan. All experiments in the present work were done with the former normal n-Si wafers with no specified tilting directions and angles, unless otherwise noted. Nearly atomically flat and H-terminated Si(111) surfaces with a step and terrace structure were obtained by the conventional RCA cleaning,16 followed by etching with 5% HF for 5 min and with 40% NH4F for 15 min.17 The Si surfaces thus treated were then immersed in 7.6 M HI for certain periods of time in a range from 30 to 480 min. The Si immersion was carried out in the dark under a nitrogen atmosphere in order to prevent the oxidation of the Si surface by either photogenerated holes or oxidants in solution such as dissolved oxygen and I2 (or I3-). A double-walled glass vessel was used for the Si immersion, and the temperature of 7.6 M HI, put in the inner part of the vessel, was kept constant by circulating temperature-regulated water through the outside opening of the vessel. Chemical analysis of the Si(111) surface was carried out with an X-ray photoelectron spectrometer (XPS, Shimadzu ESCA1000) having an Al KR line as the X-ray source. The surface atomic (I/Si) ratio was calculated from the integrated intensities of XPS I-3d and Si-2p peaks, the contribution from inside Si atoms of Si crystal being corrected by a method of Himpsel et al.18 The Si surfaces were also investigated with an FTIR spectrometer (Bio-Rad FTS 575C). The multiple internal reflection (MIR) method was adopted to get a high spectral sensitivity.19,20 The Si wafers were cut into a prism shape, and the number of light reflections in the Si wafer (prism) was calculated from its geometrical shape to be about 100. The sample chamber was purged with nitrogen gas before and during measurements. A Si wafer with a chemically oxidized surface, prepared by immersing in boiling (30% H2O2 + 98% H2SO4) (1:1 in volume) for 10 min, was used as the spectral reference. Surface morphology was inspected with an atomic force microscope (AFM, Digital Instruments NanoScope IIIa). The AFM images were obtained by a tapping mode under an ex situ condition with the Si samples placed in air. The radius of an AFM tip was about 10 nm according to a catalog of a production company (Nanosensors).

Results and Discussion Figure 1 shows AFM images of HF- and NH4F-etched Si(111) surfaces (a) before and (b-d) after immersion in 7.6 M HI, where the temperature of the HI solution is changed from (b) 5 °C to (c) 63 °C to (d) 81 °C. The scalelike pattern in Figure 1a represents a step and terrace structure of the NH4F-etched Si(111) surface, indicating that a nearly atomically flat and H-terminated Si(111) surface is produced by the etching. This conclusion was confirmed by the fact that the FTIR spectra of the Si(111) surfaces of this type showed a strong sharp peak at 2083.6 cm-1, assigned to the stretching mode of Si-H bonds at (16) Kern, W.; Puotinen, D. A. RCA Rev. 1970, 31, 187. (17) Higashi, G. S.; Chabal, Y. J.; Trucks, G. W.; Raghavachari, K. Appl. Phys. Lett. 1990, 56, 656. (18) Himpsel, F. J.; McFeely, F. R.; Taleb-Ibrahimi, A.; Yarmoff, J. A.; Hollinger, G. Phys. Rev. B 1988, 38, 6084. (19) Zhou, X. W.; Ishida, M.; Imanishi A.; Nakato, Y. Electrochim. Acta 2000, 45, 4655. (20) Zhou, X. W.; Ishida, M.; Imanishi A.; Nakato, Y. J. Phys. Chem. B 2001, 105, 156.

Figure 1. AFM images for H-terminated Si(111) surfaces (a) before and (b-d) after immersion in 7.6 M HI. The immersion times and temperatures are (b) 480 min, 5 °C, (c) 240 min, 63 °C, and (d) 60 min, 81 °C. White lines forming a triangle, added to (b), indicate the 〈1 h1 h 2〉 direction and equivalents of the Si(111) surface, determined from the directions of trigonal etch pits whose edges are known to be in parallel to the 〈1 h1 h 2〉 direction or equivalents.32,33

a terrace, together with some weak peaks assigned to vibration modes of Si-H and SiH2 bonds at steps, as reported in previous papers.19,20 For the immersion in 7.6 M HI at 5 °C for 480 min (Figure 1b), a number of long rod clusters are formed everywhere on the Si(111) surface. The AFM inspection showed that the width, height, and length of the rod clusters were about 20 nm, 1-2 nm, and 100-200 nm, respectively, indicating that the rods are of a nanosize and rather flat. It is to be noted also that the rods show a high degree of alignment, that is, they are all aligned in the 〈1 h1 h 2〉 direction or equivalents of the Si(111) surface. A similar result was obtained for the Si immersion at 25 °C, as reported previously.14 For the immersion at a slightly higher temperature of 63 °C for 240 min (Figure 1c), on the other hand, not the rods but dot clusters were formed at terraces and steps of the Si(111) surface. For the immersion at a further high temperature of 81 °C for 60 min (Figure 1d), dot clusters were also formed, but their size and shape were quite different from those in Figure 1c. We can roughly say that the dot clusters in panel c are small and round, the diameter and height being about 20 nm and 10-15 nm, respectively, whereas the dot clusters in panel d are large and flat, the diameter and height lying about 100 and 1 nm, respectively. The results of Figure 1 indicate that the size and shape of the clusters depend strongly on the temperature of the HI solution, indicative of different mechanisms of the cluster growth at different temperatures. To investigate the growth mechanism in more detail, we inspected a change in the size and shape of the clusters with the immersion time. Figure 2 shows a time course of the AFM image of the Si surface for the immersion at 5 °C. After 60-min immersion (a), a large number of small dustlike

4606

Langmuir, Vol. 20, No. 11, 2004

Imanishi et al.

Figure 2. AFM images for H-terminated Si(111) surfaces after immersion in 7.6 M HI at 5 °C. The immersion times were (a) 60, (b) 120, and (c) 240 min. A triangle in (c) is drawn in the same way as in Figure 1b.

Figure 3. AFM images for H-terminated Si(111) surfaces after immersion in 7.6 M HI at 63 °C. The immersion times were (a) 60, (b) 120, and (c) 240 min.

Figure 4. AFM images for H-terminated Si(111) surfaces after immersion in 7.6 M HI at 81 °C. The immersion times were (a) 30, (b) 60, and (c) 120 min.

dot clusters were formed all over the Si surface. The dot clusters in (a) then changed into wormlike clusters after 120 min of immersion (b) and into rod clusters after 240 min of immersion (c). The rod clusters were aligned in the 〈1 h1 h 2〉 direction or equivalents, as mentioned earlier. It may be noted also that the change in the cluster size and shape is accompanied by a decrease in the cluster density. Figure 3 shows a time course of the Si surface for the immersion at 63 °C. For such a high immersion temperature, clearly separated large-sized dot clusters were already formed after 60 min of immersion (a), different from the case of Figure 2. With the increasing immersion time, the number of the clusters was kept nearly constant, but their diameter (d) and height (h) increased much, from d ) ∼10 nm and h ) 1-2 nm for the 60-min immersion (a) to d ) 10-20 nm and h ) 4-5 nm for the 120-min immersion (b) to d ) 50-60 nm and h ) 10-15 nm for the 240-min immersion (c), respectively. Figure 4 shows a time course for the immersion at 81 °C. In this case, not round but flat dot clusters were formed, as already mentioned. With the immersion time, the

diameter increased more and more, from ∼50 nm for 30min immersion (a) to ∼100 nm for 60-min immersion (b), to 1000 nm for 120-min immersion (c), whereas the height of the clusters was kept nearly constant (about 1 nm). The results of Figures 2-4 strongly suggest that small dustlike clusters are first formed, and they aggregate with each other with the immersion time, probably by surface diffusion of a cluster constituting molecular compound(s). Here arise some questions. The first one is what is the chemical composition of the clusters, or in other words, what is the surface-diffusing species for aggregation. The atomic force microscopy (AFM) images of Figures 1-4 show that the step and terrace patterns of the Si surface changed during the immersion in 7.6 M HI, indicating that Si etching with HI occurred at the surface. The conclusion was confirmed by detailed experiments using n-Si wafers having a well-defined vicinal (111) surface tilting in the 〈1 h1 h 2〉 direction at an angle of 0.36 ( 0.1°. Figure 5 shows AFM images for this type of Si(111) surface (a) before and (b) after immersion in 7.6 M HI at 5 °C for 240 min. A fairly well arranged step structure is

Formation of Nanorods of Iodine Compounds

Figure 5. AFM images for H-terminated Si(111) surfaces (a) before and (b) after immersion in 7.6 M HI at 5 °C for 240 min. Well-controlled vicinal Si(111) surfaces with a tilting angle of 0.36 ( 0.1° were used to see the etching process clearly.

Figure 6. FTIR spectra for H-terminated Si(111) surfaces with a large tilting angle of 4.0 ( 0.1° as a function of the time of immersion in 7.6 M HI at 5 °C. The immersion times were (a) 0, (b) 120, and (c) 240 min.

seen in panel a, whereas it is strongly disturbed in panel b, which clearly indicates that Si etching with HI occurred during the immersion. The Si etching with HI should produce silicon iodides, designated as SiHxI4-x (x ) 0, 1, 2, or 3) or their oligomers, which will be nearly insoluble in 7.6 M HI. It is thus most likely that such etching products form the nanoclusters on the Si surface. This expectation is supported by the following XPS observations: (1) The XPS spectra for the Si surface after the HI immersion gave clear I-3d5/2 and I-3d3/2 peaks together with a Si-2p peak, indicating the presence of iodine at the Si surface (or in the nanoclusters). (2) The I-3d5/2 peak was at 619.4-619.8 eV, very close to that for I2 molecules (619.9 eV), suggesting that iodine at the Si surface was in a nearly neutral (noncharged) state. This is reasonable because Si-I bonds in SiHxI4-x should have only a slight polar character owing to a small electronegativity difference between Si and I. (3) The Si2p peak contained no high-energy peaks assigned to highly oxidized Si (Si3+ or Si4+). This implies that no Si-O bond is present at the Si surface (or in the nanoclusters). We also investigated the surface process by FTIR measurements. Figure 6 shows an example of FTIR spectra for an H-terminated Si(111) surface as a function of the time of immersion in 7.6 M HI at 5 °C. A Si(111) wafer with a large tilting angle of 4.0 ( 0.1° and thus with a high density of surface steps was used in this case in order to increase the absorption intensities of Si-H vibrations at step sites. In Figure 6, bonds peaked at 2083.6 and 2071.4 cm-1 are reported to be assigned to SiH (terrace)

Langmuir, Vol. 20, No. 11, 2004 4607

and SiH (step), respectively.21,22 On the other hand, three bands peaked at 2093.8, 2101.5, and 2134.8 cm-1 are reported to originate from SiH2 (step) and adjoining SiH (terrace) that interacts with the SiH2 (step).21-25 The first 120 min of immersion caused a change from spectrum a to spectrum b, in which the intensity of the 2083.6-cm-1 band became weak, whereas the intensities of the 2071.4cm-1 band and the three bands at 2093.8, 2101.5, and 2134.8 cm-1 remained nearly unchanged. On the other hand, the second 120-min immersion led to a change from spectrum b to spectrum c, in which the intensity of the 2083.6-cm-1 band remained unchanged, whereas that of the 2071.4-cm-1 band increased, and those of the 2093.8, 2101.5, and 2134.8 cm-1 bands decreased. The results are in harmony with the aforementioned conclusion that the Si etching occurs in 7.6 M HI, though they also show that the etching proceeds in a rather complicated manner. A newly appearing band at about 2087 cm-1 was assigned in previous work19,20 to Si-H (terrace) modified by an interaction with Si-I (terrace) produced at adjacent positions. The origins of bands at about 2094 and 2102 cm-1 are unknown. It should be noted also that no absorption peaks, assignable to vibrations of etching products, SiHxI4-x, were observed probably owing to their existence in a small amount or an x value of zero (x ) 0). Further investigations are necessary to get effective information on the etching from FTIR spectra. The next question is why aligned rod-shaped clusters are formed for the immersion at 5 °C, in sharp contrast to the cases of immersion at 63 and 81 °C. This will be the most important question from the point of view of the nanostructuring or self-organization of molecular systems. A number of mechanisms have been reported to explain the formation of oriented nanowires on cleaned solid surfaces by deposition under UHV conditions. They can be divided into two groups: One is a thermodynamicsbased mechanism and the other is a kinetics-based one. For example, Tersoff and Tromp explained the formation of rod-shaped Ge clusters on the Si(111) surface in terms of thermodynamics, i.e., by considering the minimization of the cluster energy including the elastic relaxation.26 Similarly, Chen et al. explained the formation of oriented ErSi2 wires on Si(001) by considering the minimization of the strained energy.27 On the other hand, Mo et al. explained the formation of Si wires on Si(100)28 in terms of kinetics, i.e., by considering a difference in the accommodation coefficient between the end and the side of clusters. To clarify which mechanism is working in the present case, we examined what happened when the rod clusters, formed by the immersion at 5 °C, were placed in the HI solution of 63 °C. Experiments showed that the rod clusters changed into the dot clusters at 63 °C. Inversely, when the dot clusters, formed by the immersion at 63 °C, were placed in the HI solution of 5 °C, no change occurred. These results strongly suggest that the formation of the dot clusters at a high temperature is controlled by (21) Jacob, P.; Chabal, Y. J. J. Chem. Phys. 1991, 95, 2897. (22) Morin, M.; Jacob, P.; Levinos, N. J.; Chabal, Y. J.; Harris, A. L. J. Chem. Phys. 1992, 96, 6203. (23) Raghavachari, K.; Jacob, P.; Chabal, Y. J. Chem. Phys. Lett. 1993, 206, 156. (24) Jacob, P.; Chabal, Y. J.; Raghavachari, K.; Christman, S. B. Phys. Rev. B 1993, 47, 6839. (25) Hines, M. A.; Chabal, Y. J.; Harris, T. D.; Harris, A. L. Phys. Rev. Lett. 1993, 71, 2280. (26) Tersoff, J.; Tromp, R. M. Phys. Rev. Lett. 1993, 70, 2782. (27) Chen, Y.; Ohlberg, D. A. A.; Ribeiro, G. M.; Chang, Y. A.; Williams, R. S. Appl. Phys. Lett. 2000, 76, 4004. (28) Mo, Y.-W.; Swartzentruber, B. S.; Kariotis, R.; Webb, M. B.; Lagally, M. G. Phys. Rev. Lett. 1989, 63, 2393.

4608

Langmuir, Vol. 20, No. 11, 2004

Figure 7. The surface (I/Si) atomic ratio, as a function of the immersion time, for H-terminated Si(111) surfaces immersed in 7.6 M HI at 5, 63, and 81 °C.

thermodynamics, whereas that of the rod clusters at a low temperature is controlled by kinetics. We mentioned earlier that the rod clusters formed at 5 °C are aligned in the particular directions, i.e., in the 〈1 h1 h 2〉 direction and equivalents. This fact also strongly suggests that the formation of the rod clusters is controlled by kinetics. The formation of oriented needles (or their aggregates) is well-known for diffusion-limited aggregation (DLA).29,30 A typical example is crystals of snow. It is thus reasonable to assume that the formation of the oriented rod clusters at 5 °C is caused by a similar diffusion-controlled mechanism. This assumption is also supported by the following argument. We mentioned earlier that the density of the nanoclusters in the AFM images decreased with the immersion time (Figures 2-4). Figure 7 plots the atomic (I/Si) ratio for the Si surface after the HI immersion as a function of the immersion time. The atomic (I/Si) ratio, obtained from XPS spectra, increased with the immersion time but showed a tendency of saturation in a range of the long immersion time. The results clearly indicate that the rate of Si etching with HI (and thus the rate of formation of the nanoclusters) decreased with the immersion time. The decreased etching rate will in turn lead to the diffusion-limited condition for the cluster growth and hence to the formation of the oriented rod clusters, though they can keep their shape only at a low temperature such as 5 °C. (29) Witten, T. A. Jr.; Sander, L. M. Phys. Rev. Lett. 1981, 47, 1400. (30) Witten, T. A. Jr.; Sander, L. M. Phys. Rev. B 1983, 27, 5686. (31) Flidr, J.; Huang, Y. C.; Hines, M. A. J. Chem. Phys. 1999, 111, 6970. (32) Roche, J. R.; Ramonda, M.; Thibaudau, F.; Dumas, Ph.; Mathiez, Ph.; Salvan, F.; Allongue, P. Microsc. Microanal. Microstruct. 1994, 5, 291. (33) Pietsch, G. J.; Kohler, U.; Henzler, M. J. Appl. Phys. 1993, 73, 4797.

Imanishi et al.

Why does the etching rate decrease with the immersion time? We can explain this fact in the following. The Si(111) surface has two types of steps, dihydride (SiH2) steps and monohydride (Si-H) ones, as can be easily seen from Figure 5a and Figure 6, spectrum a. The SiH2 steps are more reactive, and they are etched at a much higher rate than the Si-H steps.26 In addition, the etching at the SiH2 steps will convert the SiH2 steps to stable Si-H steps, as indicated, for example, by decreases in the intensities of the 2093.8, 2101.5, and 2134.8 cm-1 bands originating from SiH2 (step) and SiH (terrace) interacting with it and an increase in the intensity of the 2071.4-cm-1 band assigned to Si-H (step) in a change from spectrum b to spectrum c of Figure 6. Thus, the density of the SiH2 steps decreases with the immersion time, which leads to a decrease in the etching rate. The fact that the rod clusters formed at 5 °C are aligned specifically in the 〈1 h1 h 2〉 direction and equivalents implies that the structure of the Si(111) surface affects the cluster formation. Further details are now under investigation. Finally, let us consider a difference in the dot clusters formed by the immersion at 63 °C and those formed at 81 °C. We investigated in detail the growth process for the clusters in a range of immersion temperature from 63 to 81 °C. AFM inspection showed that the changes of the size and shape of the dot clusters occurred abruptly at about 70 °C, with no observation of a transition state in which both the types of the dot clusters coexisted. A possible explanation for the abrupt change in the cluster size and shape, observed at 70 °C, may be given by taking into account a kind of phase transition in the nanosized dot clusters. On the other hand, a transition state in which the dot and rod clusters coexisted was observed at around 30 °C, suggesting that the critical temperature for a change from the kinetics- to thermodynamics-controlled mechanism is about 30 °C. Conclusion The formation of nanosized clusters on nearly atomically flat, H-terminated Si(111) surfaces, when immersed in 7.6 M HI in a temperature range of 5-81 °C, was investigated by AFM, XPS, and FTIR observations. AFM inspection has revealed that long rod clusters, aligned in the directions of the 3-fold symmetry of the Si(111) surface, are formed in the immersion temperature below about 30 °C, whereas circular dot clusters were formed in the immersion temperature above 30 °C. It is discussed that the formation of the oriented rod clusters at a low temperature is caused by a diffusion-controlled mechanism, whereas that of the dot clusters is caused by a thermodynamics-based mechanism. LA035749C