Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction

SUSC-20778; No of Pages 6

January 14, 2016;

Model: Gulliver 5

Surface Science xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Surface Science journal homepage: www.elsevier.com/locate/susc

3Q2

Bin Leong Ong a,⁎, Sheau Wei Ong b, Eng Soon Tok a a

7

a r t i c l e

8 9 10 11 12 23 24 25 26 27 28 29

Article history: Received 6 November 2015 Accepted 31 December 2015 Available online xxxx

i n f o

R O

Electronic Materials Growth and Interface Characterization (eMaGIC) Laboratory, Department of Physics, National University of Singapore, 2 Science Drive 3, Singapore 117551, Singapore Department of Physics and Yale-NUS College, National University of Singapore, 2 Science Drive 3, Singapore 117551, Singapore

a b s t r a c t

Evidence for the influence of Si(001)-(2 × 1) surface reconstruction on the elongation direction of CoSi2 flat islands is discussed in this paper. Step height analysis of these flat islands shows that flat island heights, HA, follow discrete values of NA such that HA = mNA + c, where NA = 1, 2, 3, …, m is equivalent to the number of monoatomic step height (1.4 Å) of the Si(001) surface, and c is the initial island height when NA = 0. The NA values were found to be correlated to the flat island elongation direction with respect to the (2 × 1) dimer rows. For a given terrace, the preferred elongation direction of these flat islands is parallel to the Si dimer rows. As a result, orthogonally elongated islands are clearly resolved on adjacent terraces, which are separated by monoatomic steps. The endotaxial growth of these flat islands is thus also influenced by the anisotropic adatom diffusion due to (2 × 1) surface reconstruction. © 2016 Published by Elsevier B.V.

P

b

E

C

32

1. Introduction

35 36

To achieve preferential growth and selective control over the morphology of silicon -based heteroepitaxial low-dimensional nanostructures, such as nanowires, requires fundamental insights into the growth dynamics and kinetics [1–3]. Homoepitaxial growth of Si nanowires, for example, on Si(001)-(2 × 1) surface have shown to be influenced by surface reconstruction, where anisotropic Si-adatom diffusion leads to growth of Si islands forming nanowires [4,5]. On the other hand, the formation of heteroepitaxial nanowires, such as rare-earth silicide system [6–9], have been attributed to anisotropic lattice mismatches between islands and substrates. These islands elongate preferentially along the direction with lower lattice-mismatch in order to minimize strain energy. In other cases, the shape transition of compact islands to nanowires driven by strain relaxation have also been reported for material systems such as Ag nanowires [10] and silicide nanowires such as CoSi2 on Si(001) [11–16]. The introduction of nanowires growing epitaxially into the substrate surface (endotaxy) reviewed recently for several systems such as Ni, Co, Pt, and Fe on Si(001), Si(110), and Si(111) surfaces, however, revealed a more complex mechanism for nanowire formation and growth [17–19].

45 46 47 48 49 50 51 52 53

R

N C O

43 44

U

41 42

R

34

39 40

13 14 15 16 17 18 19 20 21 22

T

E

D

Keywords: Scanning tunneling microscopy Surface reconstruction Nanowire growth Shape evolution Silicide Anisotropy

33 31 30

37 38

O

4 5 6

F

2

Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction

1Q1

⁎ Corresponding author. Tel.: +65 6516 7655. E-mail address: [email protected] (B.L. Ong).

Of particular interest is the formation of CoSi2 low-dimensional structures. Recent works have shown that nanowires on Si(001) are formed not only because of strain-induced shape transition [13,16] but is also a consequence of a thermally activated growth due to endotaxy [18,20–24]. It has been reported that CoSi2 forms two types of islands, ridge and flat, on the Si(001) surface [23]. Briefly, ridge islands form preferentially low energy CoSi2{111}//Si{111} interface (Type B) with Si, which appears to show a “twinned” type interface due to a stacking fault. They grow endotaxially because the low energy Type B interface is more energetically favorable than other higher energetic planes along the b110 N directions (CoSi2{111}//Si{115}, CoSi2{115}//Si{111}, and CoSi2{112}//Si{112}). Its preferred growth over other interfacial planes results in wire formation. For flat-type islands, they form Type A, i.e., CoSi2{111}//Si{111}, interface endotaxially with Si. Type A interface, unlike Type B, is “untwinned” where the lattice structure of CoSi2 follows that of Si. These islands are bound by the energetically favorable Type A interface on all sides, resulting in the preferential formation of compact islands. However, growth at low temperature results in wire-like flat islands before compact islands are seen at high temperatures. Although the occurrence of wire-like features have been attributed to the presence of corner barriers [23,25,26], the influence of Si(001)-(2 × 1) surface reconstruction due to the anisotropic diffusion of adatoms along and across dimer rows has not been adequately addressed. Here, we report the anisotropic growth of wire-like flat islands at low growth temperatures between 500 °C and 650 °C. By analyzing the island height with respect to the

http://dx.doi.org/10.1016/j.susc.2015.12.039 0039-6028/© 2016 Published by Elsevier B.V.

Please cite this article as: B.L. Ong, et al., Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction, Surf. Sci. (2016), http://dx.doi.org/10.1016/j.susc.2015.12.039

54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79

115

3. Results and discussions

116

3.1. Flat island height evolution and height analyses

117 118

Details of the structure of ridge-type and flat-type islands and their epitaxial relationships with the Si(001) substrate have been reported previously [23]. Fig. 1a shows a 500 nm × 500 nm global surface morphology of

102 103 104 105 106 107 108 109 110 111 112

119

C

100 101

E

98 99

R

96 97

R

94 95

O

92 93

C

90 91

N

88 89

U

86 87

F

113 114

The STM experiments were carried out in situ in an OMICRON UltraHigh Vacuum (UHV) system with base pressure of 2 × 10− 10 Torr, equipped with an OMICRON Variable Temperature-Scanning Tunneling Microscope (VT-STM). Si samples were cut from Boron-doped P-type singular Si(001) wafers with resistivity less than 0.1 Ω/cm supplied by Virginia Semiconductors. These samples were chemically etched ex situ based on the recipe described by Ong et al. [27,28]. They were then dipped in dilute aqueous hydrofluoric (HF) acid solution to terminate the surface with hydrogen prior to outgassing in the UHV chamber for 8 h at ~ 300 °C. The samples were then progressively annealed in steps of 50 °C to 700 °C and then flashed at 30 s per cycle to 1100– 1150 °C. Upon cooling to room temperature, the clean samples' surface morphologies were verified in situ using a VT-STM. The samples were then deposited at elevated temperatures (530 °C, 560 °C, 590 °C, and 620 °C) with cobalt (Goodfellow, N 99.99 + % purity) using electronbeam evaporation with a rate of 0.1 ML/min for 1 min. Temperatures and deposition rate were determined using an infrared pyrometer and a quartz crystal monitor (QCM), respectively. The surface morphologies were subsequently characterized in situ using the VT-STM. The STM tungsten tips were fabricated using the Omicron Tip-Etching kit and then outgassed in UHV prior to STM scans. All STM images were taken in situ at room temperature using constant-current mode, with a tunneling current at −1.0 nA, sample bias of −2.0 V with acquisition done in a unidirectional mode. Drift in between images were kept to a minimum by allowing longer times for the STM tip to scan the surface. The STM scan direction was kept unidirectional from bottom to top. The STM images were background-corrected (planar) and flattened (“Flatten discarding regions”) using the WSxM software (Nanotec Electronica) [29]. Analyses and measurements from these images were done using the same software.

O

85

along two orthogonal [110] and ½110 directions. Apart from these islands, the surface is also decorated with “holes” (labeled as “h”) surrounding several islands with remnant Si terraces exhibiting (1 × 2) reconstruction (labeled as “t”). Fig. 1b shows the average length to width aspect ratio of the flat islands measured as a function of its growth temperature. The plot represents the shape evolution of the flat islands with increasing growth temperature. Between 530 °C and 650 °C, the average length to width aspect ratio is shown to increase rapidly from 2:1 to 7.5:1. Thereafter, it decreases rapidly back to 2:1 above 710 °C. The increase in the aspect ratio shows the elongation of flat islands at lower growth temperatures. However, interfacial energies of the flat island are reported to be similar on all sides of the island since they are bound at the CoSi2{111}//Si{111} interface. The formation of flat islands energetically favors a compact isotropic shape (e.g., square) rather than a nanowire. The elongation of the flat island therefore suggests other factors which kinetically limit the growth and formation of compact flat islands. We proceeded to examine the flat islands in more detail and subsequently reveal the correlation between the flat island's heights and their elongation direction with respect to the Si dimer rows. Two flat islands from Fig. 1a (labeled as “2a” and “2b”), which appear to elongate parallel and perpendicular to the Si(001)(2 × 1) dimer rows, were then scanned at higher resolution and shown in Fig. 2a and b. The flat island in Fig 2a elongates along the Si dimer rows with respect to the main Si(001)-(2 × 1) terrace labeled “1,” while it is found to elongate perpendicular to the adjacent remnant Si(001)-(1 × 2) terrace labeled “2.” The line profile AB shows that the island height is about 2.4 ± 0.2 Å with respect to terrace “2,” while the line profile CD shows that the height is 3.8 ± 0.2 Å with respect to terrace “1.” Comparing the two line profiles, the height difference between the Si terraces “1” and “2” is thus about 1.4 Å. Fig. 2b shows the second flat island elongating along [110] perpendicular to the Si(001)-(2 × 1) Si dimer rows. The line profiles EF and GH give the island a height of 3.7 ± 0.2 Å with respect to remnant Si(001)-(1 × 2) terrace labeled “2” and 5.1 ± 0.2 Å with respect to the main Si(001)-(2 × 1) terrace labeled “1.” Comparing the two line profiles EF and GH, the height difference between terrace “1” and “2” is also about 1.4 Å. Both flat island measurements suggest that their heights are closely correlated to the monoatomic step height of the Si(001) surface (1.4 Å). The remnant Si terraces in

R O

2. Methods

rows elongates along the ½110 direction on Si(001)-(2 × 1) surface as 121 shown in the figure. The surface morphology consists only of two types 122 of islands: flat type (labeled “F″) and ridge type (labeled “R”), elongating 123

P

84

a Si(001) surface after 0.1 ML of cobalt is deposited at 530 °C. Si dimer 120

T

82 83

Si(001)-(2 × 1) reconstructed surface, we elucidate the apparent orthogonal elongation directions along b110 N of flat nanowires as observed at low growth temperatures—a consequence which we attribute to the influence of surface reconstruction of Si(001).

D

80 81

B.L. Ong et al. / Surface Science xxx (2016) xxx–xxx

E

2

Fig. 1. (a) STM global morphologies of Si(001)-(2 × 1) deposited with 0.1 ML Co at 530 °C (500 nm × 500 nm). Some flat islands and ridge islands are labeled “F″ and “R,” respectively. Features labeled “h” are holes which formed as Si is consumed during cobalt deposition. Consumption of Si results in the remnant Si(001)-(1 × 2) terraces left after cobalt deposition (labeled as “t”). (b) Average length to width aspect ratio of flat islands measured as function of growth temperature.

Please cite this article as: B.L. Ong, et al., Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction, Surf. Sci. (2016), http://dx.doi.org/10.1016/j.susc.2015.12.039

124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164

3

R

E

C

T

E

D

P

R O

O

F

B.L. Ong et al. / Surface Science xxx (2016) xxx–xxx

167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184

both figures suggest consumption of Si during cobalt disilicide island growth. The islands in Fig 2 reveal an interesting observation. When the flat island is 2.4 Å in height with respect to the Si dimer rows, it is elongating perpendicular to them. When its height is 3.8 Å (2.4 + 1.4) with respect to the dimer rows, it is found elongating parallel to them. When the island height is 5.2 Å (2.4 + 2(1.4)), it is found elongating perpendicular again to dimer rows. The increase in island height from 2.4 Å to 3.8 to 5.2 Å appears to be brought about by the consumption of Si terraces surrounding the islands. Motivated by these observations, we extended the height analyses to flat islands formed at three other growth temperatures: 560 °C, 590 °C, and 620 °C. Fig. 3a shows an STM surface morphology of CoSi2 islands formed on Si(001) at 560 °C. The image shows a dominant Si(001)-(2 × 1) surface having Si dimer rows (white lines) extending

U

165 166

N C O

R

Fig. 2. High-resolution STM images of 2(a) and 2(b) identified in Fig. 1. (a) Flat island with height at 2.4 ± 0.2 Å with respect to terrace “2” and 3.8 ± 0.2 Å with respect to terrace “1” (line profiles AB and CD). The island elongates parallel to the Si dimer rows on terrace “1” but perpendicular to those on “2.” (b) Flat island with height at 3.7 ± 0.2 Å with respect to terrace “2” and 5.1 ± 0.2 Å with respect to terrace “1” (line profiles EF and GH). The island is elongating perpendicular to Si dimer rows on terrace “1” but parallel to those on “2.”

along ½110. Consistent with previous observations, we found that the heights and elongation directions of these islands are correlated to the dimers-rows. The measurements, as shown in the table of Fig 3a, islands with heights of 2.4, 3.8, 5.2, and 6.6 Å, were found. It is interesting to note that flat islands of heights 2.4 ± 0.2 Å (blue) and 5.2 ± 0.2 Å (red)

elongate perpendicular to the Si dimer rows whereas those with heights 3.8 ± 0.2 Å and 6.6 ± 0.2 Å elongate parallel to the dimer rows. In other words, flat islands with the same height are always found to be elongating in the same direction, be it along or perpendicular to the Si dimer rows. At higher growth temperatures, the consumption of Si appears more significant thus resulting in the formation of larger flat islands with significantly higher heights with respect to the surface. We measured a total of about 230 flat islands with growth temperatures at 530 °C, 560°, 590 °C, and 620 °C. The height distributions of the flat islands are found to peak at regular intervals with each exhibiting a narrow size-distribution as shown in Fig.3b. The interval or height difference between every peak is found to be 1.4 Å. By labeling each peak with an integer value NA starting with the first peak height as NA = 1, the average peak height HA of the flat islands follows a linear relation with NA as shown in Fig. 3(c), such that H A ¼ mNA þ c

185 186 187 188 189 190 191 192 193 194 195 196 197 198 199

ð1Þ

where c is the y-intercept height value (1.0 ± 0.1 Å) when NA is 0 and m 201 is the slope (1.4 ± 0.1 Å). Based on the height analysis results, three

Please cite this article as: B.L. Ong, et al., Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction, Surf. Sci. (2016), http://dx.doi.org/10.1016/j.susc.2015.12.039

B.L. Ong et al. / Surface Science xxx (2016) xxx–xxx

U

N

C

O

R

R

E

C

T

E

D

P

R O

O

F

4

Please cite this article as: B.L. Ong, et al., Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction, Surf. Sci. (2016), http://dx.doi.org/10.1016/j.susc.2015.12.039

5

R O

O

F

B.L. Ong et al. / Surface Science xxx (2016) xxx–xxx

P

Fig. 4. (a–d) Schematic illustrations describing the formation and elongation of flat islands with respect to the Si(001) dimer rows, with a table detailing the final island elongation direction and height values with respect to the Si(001) domain in d.

214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235

(2 × 1) dimer rows along the ½110 direction. Line profiles IJ and KL across the island's short and long sides indicate that the height of the flat islands is about 1.2 ± 0.2 Å and 1.1 ± 0.2 Å with respect to the surrounding dimer rows. No flat islands of the same height (1.0 Å) are found to elongate perpendicular to the Si dimer rows, suggesting that flat islands may have initially formed and elongated in a direction along the Si dimer rows during cobalt deposition. The anisotropic growth of CoSi2 flat islands is therefore also influenced by the reconstruction of Si(001)(2 × 1) surface. The implication of flat islands elongating preferentially along the Si dimer rows is in agreement with previous studies, which reported that diffusion of Si adatoms along the dimer rows is faster than across them [5] while Co adatoms favor diffusion into the Si(001) subsurface [30,31]. The preferential subsurface diffusion of Co adatoms to the UD site further suggests the preferential diffusion of adatoms under the dimers in the direction parallel to the dimer rows. Preferential diffusion of adatoms along the dimer rows will in turn allow a higher flux of adatoms to be incorporated to the flat islands' longer ends, thereby resulting in the elongation of the islands along the dimer rows. This anisotropy in surface diffusion of adatoms will then result in a slight difference between the fluxes of atoms arriving at each side of the island, which consequently allow the flat islands to grow faster at the ends forming nanowires.

236

elongate parallel the Si dimer rows along ½110 (see Fig. 4b). Thus, the flat islands (red) that are formed on the second layer of Si(001) will be orthogonal in direction to those (yellow) that are formed on the first layer of Si(001). Furthermore, with the consumption of the first layer of Si, flat islands (yellow) that grew along [110] will show an increase in apparent height from 1.0 to 2.4 Å. Further deposition of cobalt leads to the consumption of Si in the second layer exposing the third layer of Si(001). As shown in Fig. 4c, flat islands (yellow) that form on the first layer of Si will now be elongating parallel to the dimer rows along [110] on Si(001)-(1 × 2), whereas those (red) forming on the second layer of Si will be elongating perpendicular to them. The flat islands (yellow) will then have a height of 3.8 Å while perpendicular ones (red) will have a height of 2.4 Å. New flat islands (green) nucleating on the newly exposed third layer will again be elongating along the dimer rows along [110]. These islands will be parallel to the flat islands (yellow) that were formed from the first layer and perpendicular to those (green) formed from the second layer. As subsequent layers of Si get further consumed, more flat islands will nucleate and form, resulting in a surface morphology where flat islands appear to elongate both along and perpendicular to the Si(001) dimer rows (Fig. 4d). Due to the formation of flat islands at different

249 250

D

The apparent observation of flat islands that appear to elongate both along the [110] and ½110 directions on a single Si(001)-(2 × 1) domain can now be explained as illustrated in Fig. 4. When cobalt is deposited on the Si(001)-(1 × 2) surface above 500 °C, it will react with Si to form CoSi2 islands on the substrate surface (see Fig. 4a). Flat islands (yellow) are initially formed with relatively small height (≈ 1.0 Å), and they are oriented parallel to the dimer rows along [110]. During cobalt deposition, the top layers of Si are consumed to form the cobalt silicide islands. The consumption of the first layer of Si will then expose the second layer of Si with Si(001)-(2 × 1) reconstruction. The flat islands (yellow) that initially formed on the first layer will now be elongating perpendicular to the new layer of Si(001)-(2 × 1) dimer rows. New flat islands (red) that nucleate and form on the second layer will

E

T

C

212 213

E

210 211

R

208 209

R

206 207

N C O

204 205

observations can be made. First, the slope value m indicates that flat island heights are separated by the same value of 1.4 Å, which coincides with the monoatomic step height of the Si(001) surface (1.4 Å). Second, flat islands with odd values of NA elongate perpendicular to the Si dimer rows of the Si(001) surface, while flat islands with even values of NA elongate parallel to the Si dimer rows instead. Third, the y-intercept in Eq. (1) suggests that the smallest possible height of the flat islands is about 1.0 ± 0.1 Å when NA = 0, where the island would be elongating parallel to the Si dimer rows at the initial nucleation and growth. Interestingly, the existence of flat islands with height about 1.0 Å is also found, albeit few. Fig 3d shows a high-resolution STM image of one such flat island elongating parallel to the Si(001)-

U

202 203

Fig. 3. (a) Height measurements and identification of flat island elongation directions with respect to the dominant Si(001)-(2 × 1) surface for CoSi2 islands grown at 560 °C. (b) Histogram plot of island heights measured from flat islands formed at 530 °C, 560 °C, 590 °C, and 620 °C. The island heights appear to be centered at discrete values NA. (c) Plot of flat island average heights (Å) as a function of NA, indicating a linear relationship between NA and HA with y-intercept c ≈ 1.0 ± 0.1 Å and slope m ≈ 1.4 ± 0.1 Å. The “⊥” and “//” labels indicate the elongation directions of the flat islands with respect to the Si dimer rows on the dominant surface. (d) STM image of a flat island elongating parallel to the Si dimer rows. Line profiles IJ and KL across the flat island show that the island heights are about 1.2 ± 0.2 Å and 1.1 ± 0.2 Å, respectively. The island's top is terminated with a c(2 × 2) surface reconstruction (inset).

Please cite this article as: B.L. Ong, et al., Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction, Surf. Sci. (2016), http://dx.doi.org/10.1016/j.susc.2015.12.039

237 238 239 240 241 242 243 244 245 246 247 248

251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269

B.L. Ong et al. / Surface Science xxx (2016) xxx–xxx

Competing financial interests

300 301

304

C

298 299

The authors declare no competing financial interests.

E

296 297

[17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33]

R

294 295

[14] [15] [16]

R

289 290

O

287 288

C

285 286

[8] [9] [10] [11] [12] [13]

N

283 284

U

281 282

J.A. Venables, Surf. Sci. 299–300 (1994) 798. R.M. Tromp, J.B. Hannon, Surf. Rev. Lett. 09 (2002) 1565. A. Pimpinelli, J. Villain, Physics of Crystal Growth, Cambridge University Press, 1998. Y.W. Mo, J. Kleiner, M.B. Webb, M.G. Lagally, Surf. Sci. 268 (1992) 275. Y.W. Mo, J. Kleiner, M.B. Webb, M.G. Lagally, Phys. Rev. Lett. 66 (1991) 1998. Y. Chen, D.A.A. Ohlberg, R.S. Williams, J. Appl. Phys. 91 (2002) 3213. D. Lee, D.K. Lim, S.S. Bae, S. Kim, R. Ragan, D.A.A. Ohlberg, Y. Chen, R.S. Williams, Appl. Phys. A Mater. Sci. Process. 80 (2005) 1311. B.Z. Liu, J. Nogami, Nanotechnology 14 (2003) 873. B.Z. Liu, J. Nogami, Surf. Sci. 540 (2003) 136. F.K. LeGoues, M. Liehr, M. Renier, W. Krakow, Philos. Mag. B 57 (1988) 179. J. Tersoff, R.M. Tromp, Phys. Rev. Lett. 70 (1993) 2782. I. Goldfarb, G.A.D. Briggs, Phys. Rev. B 60 (1999) 4800. S.H. Brongersma, M.R. Castell, D.D. Perovic, M. Zinke-Allmang, Phys. Rev. Lett. 80 (1998) 3795. W. Ma, R. Nötzel, H.-P. Schönherr, K.H. Ploog, Appl. Phys. Lett. 79 (2001) 4219. G. Zhou, J.C. Yang, Phys. Rev. Lett. 89 (2002) 106101. J.C. Mahato, D. Das, R.R. Juluri, R. Batabyal, A. Roy, P.V. Satyam, B.N. Dev, Appl. Phys. Lett. 100 (2012) 263117. P.A. Bennett, Z. He, D.J. Smith, F.M. Ross, Thin Solid Films 519 (2011) 8434. Z. He, D.J. Smith, P.A. Bennett, Phys. Rev. Lett. 93 (2004) 256102. S. Liang, R. Islam, D.J. Smith, P.A. Bennett, J. Cryst. Growth 295 (2006) 166. T. George, R.W. Fathauer, Appl. Phys. Lett. 59 (1991) 3249. C.W. Lim, C.-S. Shin, D. Gall, J.M. Zuo, I. Petrov, J.E. Greene, J. Appl. Phys. 97 (2005) 044909. C.W. Lim, I. Petrov, J.E. Greene, Thin Solid Films 515 (2006) 1340. B.L. Ong, W. Ong, Y.L. Foo, J. Pan, E.S. Tok, Surf. Sci. 606 (2012) 1649. G. Kellermann, L.A. Montoro, L.J. Giovanetti, P.C. dos Santos Claro, L. Zhang, A.J. Ramirez, F.G. Requejo, A.F. Craievich, Phys. Chem. Chem. Phys. 17 (2015) 4945. Z. Li, M.K. Singh, E.S. Tok, J.P.Y. Tan, M. Lin, Y.-L. Foo, Appl. Phys. Lett. 90 (2007) 101914. Y. Li, M.C. Bartelt, J.W. Evans, N. Waelchli, E. Kampshoff, K. Kern, Phys. Rev. B 56 (1997) 12539. W.J. Ong, E.S. Tok, Phys. Chem. Chem. Phys. 9 (2007) 991. W. Ong, E.S. Tok, H. Johll, H.C. Kang, Phys. Rev. B 79 (2009) 235439. I. Horcas, R. Fernández, J.M. Gómez-Rodríguez, J. Colchero, J. Gómez-Herrero, A.M. Baro, Rev. Sci. Instrum. 78 (2007) 013705. G.W. Peng, A.C.H. Huan, E.S. Tok, Y.P. Feng, Phys. Rev. B (2006) 74. J. Choi, Y. Kim, D.K. Lim, D.H. Kim, S. Kim, J. Phys. Chem. C 115 (2011) 15467. B. Voigtländer, V. Scheuch, H.P. Bonzel, S. Heinze, S. Blügel, Phys. Rev. B 55 (1997) R13444. V. Scheuch, B. Voigtländer, H.P. Bonzel, Surf. Sci. 372 (1997) 71.

F

303

279 280

[1] [2] [3] [4] [5] [6] [7]

O

302

We have re-visited the endotaxial growth of flat CoSi2 islands between 530 °C and 650 °C by considering the influence of surface reconstruction. Analyses of the flat island heights and elongation directions with respect to the Si(001)-(2 × 1) provided direct evidence that the initial nucleation and growth of flat islands occur preferentially along the Si dimer rows and endotaxial growth of the flat islands proceeds with the consumption of Si from the surface. While ridge islands are not studied in detail in this work, this work suggests that the nucleation and growth of the ridge islands may similarly be influenced by the (2 × 1) surface reconstruction.

277 278

308

References

R O

293

276

P

4. Conclusions

274 275

This work is supported by the National University of Singapore 306 (NUS) grant number R-144-000-322-112. 307

D

292

272 273

305

Acknowledgments

T

291

stages of Si-consumption, flat islands will exhibit height differences that coincides with the height difference between two adjacent Si(001) layers (see table in Fig 4). Flat islands with the largest height with respect to the Si(001)-(1 × 2) surface in Fig 4d will be yellow (5.2 Å). Those with the next largest height will be red (3.8 Å) while those in green will be about 2.4 Å in height. Flat islands (blue), which are newly formed on the surface, will have a height of about 1.0 Å. The apparent increase and difference in flat island heights are due to the consumption of Si from the surface. This implies that flat islands prefer to grow into the Si(001) surface (endotaxy) and elongate along the Si dimer rows forming nanowires at temperatures less than 650 °C. Finally, it should be noted that all flat islands exhibit a rectangular shape and distinct flat-top surface with a c(2 × 2) or (√2 × √2)R45° reconstruction (Fig. 3d, inset), albeit disordered with respect to the Si(001)-(1 × 2). The observed c(2 × 2) reconstruction is consistent with the CoSi2(001)-(√2 × √2)R45° model as proposed by Voigtländer et al. [32]. In these STM images, dark streaks that cut across the Si dimer rows are observed. Scheuch et al. [33] have suggested that these streaks are formed when cobalt atoms disrupt the Si(001) - (2 × 1) reconstruction during deposition, leading to cobalt-related silicon vacancies forming a quasi-periodic (2 × m) reconstruction where m is between 6 and 9.

270 271

E

6

Please cite this article as: B.L. Ong, et al., Endotaxial growth of CoSi2 nanowires on Si(001) surface: The influence of surface reconstruction, Surf. Sci. (2016), http://dx.doi.org/10.1016/j.susc.2015.12.039

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350