Atomic composition profile change of SiGe islands ...

Atomic composition profile change of SiGe islands during Si capping F. H. Li, Y. L. Fan, X. J. Yang, Z. M. Jiang, Y. Q. Wu et al. Citation: Appl. Phys. Lett. 89, 103108 (2006); doi: 10.1063/1.2345589 View online: http://dx.doi.org/10.1063/1.2345589 View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v89/i10 Published by the AIP Publishing LLC.

Additional information on Appl. Phys. Lett. Journal Homepage: http://apl.aip.org/ Journal Information: http://apl.aip.org/about/about_the_journal Top downloads: http://apl.aip.org/features/most_downloaded Information for Authors: http://apl.aip.org/authors

Downloaded 26 Sep 2013 to 202.116.1.148. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

APPLIED PHYSICS LETTERS 89, 103108 共2006兲

Atomic composition profile change of SiGe islands during Si capping F. H. Li, Y. L. Fan, X. J. Yang, and Z. M. Jianga兲 Surface Physics Laboratory (National Key Laboratory), Fudan University, Shanghai 200433, China

Y. Q. Wu and J. Zou School of Engineering and Centre for Microscopy and Microanalysis, The University of Queensland, Qld 4072, Australia

共Received 14 April 2006; accepted 11 July 2006; published online 6 September 2006兲 The 6% Ge isocomposition profile change of individual SiGe islands during Si capping at 640 ° C is investigated by atomic force microscopy combined with a selective etching procedure. The island shape transforms from a dome to a 兵103其-faceted pyramid at a Si capping thickness of 0.32 nm, followed by the decreasing of pyramid facet inclination with increasing Si capping layer thickness. The 6% Ge isocomposition profiles show that the island with more highly Si enriched at its one base corner before Si capping becomes to be more highly Si intermixed along pyramid base diagonals during Si capping. This Si enrichment evolution inside an island during Si capping can be attributed to the exchange of capped Si atoms that aggregated to the island by surface diffusion with Ge atoms from inside the island by both atomic surface segregation and interdiffusion rather than to the atomic interdiffusion at the interface between the island and the Si substrate. In addition, the observed Si enrichment along the island base diagonals is attempted to be explained on the basis of the elastic constant anisotropy of the Si and Ge materials in 共001兲 plane. © 2006 American Institute of Physics. 关DOI: 10.1063/1.2345589兴 Self-assembled SiGe islands have been extensively studied both for their potential applications in optoelectronic devices1 and for the fundamental investigation of heroepitaxial growth.2 The knowledge of atomic composition profile of SiGe islands is vitally important for the physical property predication of island-based device as well as for understanding the island growth mechanism. However, most methods used in investigating the atomic composition in SiGe islands are based on diffraction3,4 or spectroscopy5,6 techniques, which give out an average value on composition over a large number of islands. Recently, using 31% H2O2 solution etchant which selectively etches Ge over Si and stops etching SiGe material with the Ge content of 65%, a German research group7,8 systematically investigated the 65% Ge isocomposition profiles in the as-grown islands, many different atomic composition profiles were reported, such as the crosslike composition profile for the pyramid-shaped islands grown at 580 ° C, the ringlike composition profile for the dome-shaped islands grown at 600 ° C, and the moundlike composition profile for the dome-shaped islands grown at 620 ° C. Using another etchant of the mixture solution of 1HF : 2H2O2 : 3CH3COOH 共BPA solution兲,9,10 which selectively etches SiGe alloy over pure Si and stops etching SiGe alloy with Ge concentration of 6%, they observed a “halfmoon” shape isocomposition profile in an annealed island,9 unveiling a lateral movement of an island during in situ annealing. Those observations give us more detailed information on the atomic processes during the island growth, which is certainly very helpful for people to fully understand the island growth mechanism. Following these works, one question naturally comes out: how does a Si capping process, which is necessary for island-based material applications, affects the atomic composition in a buried island? The answer a兲

Author to whom correspondence should be addressed; electronic mail: [email protected]

0003-6951/2006/89共10兲/103108/3/$23.00

to this question is also very important not only for the full understanding of island growth mechanism during Si capping but also for the understanding of the physical properties of island-based materials, since those properties depend much on the microstructures of the capped SiGe islands including the atomic composition distribution in the islands. In this letter, atomic force microscopy and BPA solution etching experiments are used to study the 6% Ge isocomposition profile change in the SiGe islands during initial Si capping. Besides the Si enrichment at four corners of a pyramid base as observed before,7,8 the Si enrichments at the center of the pyramid base as well as along the pyramid base diagonals are also observed in the initially Si capped islands. The Si enrichment inside an island is explained by an atomic exchange model in which the capped Si atoms that aggregated to an island by surface diffusion exchange with Ge atoms from inside the island by both surface segregation and interdiffusion, and the Si enrichment along the pyramid base diagonals is suggested to be explained by the different Young’s moduli along different directions. The samples were grown in a solid source molecular beam epitaxy system. The substrate used was p-type 共001兲 wafer. After oxide removal at 1000 ° C in the growth chamber, the substrate temperature was lowered to 640 ° C, and a 50-nm-thick buffer layer was grown, a two-step growth method11 was adopted in island growth at 640 ° C in order to improve the island size uniformity. A 0.7-nm-thick Ge layer was deposited, followed by a 5 min interruption, a 0.05-nm-thick Ge layer was then deposited. The Ge growth temperature and growth rate were 640 ° C and 0.01 nm/ s, respectively. The Si capping layers with different thicknesses were deposited at 640 ° C with a deposition rate of 0.06 nm/ s. After the Si capping layer deposition, the substrate temperature was immediately cooled to room temperature. The etched samples were prepared by etching samples in BPA solution for a period of time of 90 s. The surface

89, 103108-1

© 2006 American Institute of Physics


103108-2

Li et al.

FIG. 1. 共0.5⫻ 0.5 ␮m2兲 AFM images of SiGe islands before etching 共left column兲 and corresponding etched SiGe islands 共right column兲 which were etched in BPA solution for a period time of 90 s. The islands are capped with 关共a兲 and 共e兲兴 0-, 关共b兲 and 共f兲兴 0.32-, 关共c兲 and 共g兲兴 0.36-, and 关共d兲 and 共h兲兴 0.42-nm-thick Si layer at a temperature of 640 ° C; the insets show the corresponding three-dimensional AFM images of single etched SiGe islands.

morphologies of the island samples before and after etching were characterized by atomic force microscopy in contact mode in air. The reliability of the BPA solution that stops etching at a fixed Ge composition 共6%兲 was checked by etching the samples for different periods of time 共from 90 to 300 s兲, and no significant differences were found in the morphologies of the etched samples with an etching time longer than 90 s. Therefore, by observing the surface morphologies of the etched islands, we can obtain the 6% Ge isocomposition profiles inside individual SiGe islands. The capped islands were also etched in 31% H2O2 solution for 20 min, and no significant difference was observed in the island shape after etching, indicating that the Ge composition on the surface of the capped islands is less than 65%. The left column of Fig. 1 shows the atomic force microscopy 共AFM兲 images of the SiGe islands before etching. As shown in Fig. 1共a兲, before Si capping the SiGe islands are dome-shaped, and this dome shape transforms to a 兵103其faceted pyramid12 at a Si capping thickness of 0.32 nm, as

Appl. Phys. Lett. 89, 103108 共2006兲

shown in Fig. 1共b兲. With increasing Si capping layer thickness, the pyramid base size increases, and the height and the facet inclination of the pyramid decrease as shown in Figs. 1共c兲 and 1共d兲. As suggested by Wu et al.12 these observations indicate that the shape change of SiGe islands during Si capping at 640 ° C is a kinetic rather than a thermodynamic process. The right column of Fig. 1 shows the 6% Ge isocomposition profiles of the SiGe islands corresponding to the left column. Before Si capping, the profile of the as-grown island shows a so-called half-moon structure,9 as illustrated in Fig. 1共e兲, indicating that before Si capping the islands are more highly Si intermixed at one corner of the island base. After capping the islands with Si layer thicknesses of 0.32 and 0.36 nm, this half-moon feature disappears and four mounds are formed at four corners of the pyramid base as shown in Figs. 1共f兲 and 1共g兲, respectively. The Si enrichment at four corners of the capped pyramid base was similar to that as observed previously in the 65% Ge isocomposition profile of the as-grown SiGe pyramids.7,8 However, the most striking features in 6% Ge isocomposition profile are the mounds appearing at the center of pyramid base and even along the pyramid base diagonals for the capped islands with a capping layer thickness of 0.42 nm, as shown in Fig. 1共h兲. Katsaros et al.8 have observed similar profiles to those, as shown in Fig. 1共h兲, in the 65% Ge isocomposition profiles of as-grown SiGe pyramid islands, they proposed a kinetic growth model comprising only surface diffusion processes to explain the profiles. During an island growth, the substrate Si atoms could diffuse out through a very thin wetting layer, and migrate to and mix into a pyramid island by surface diffusion. Since those Si atoms can reach a pyramid island corner from two sides and reach a pyramid edge from only one side, more Si atoms will be mixed at four corners of the pyramid base, resulting in the Si enrichment at four corners of pyramid base. When the pyramid grows up, this Si enrichment evolves to the Si enrichment at four corners of pyramid base as well as along the pyramid base diagonals. But our case is different, before Si capping the island is more highly Si enriched at one corner rather than at four corners of the island base, and the Si enrichment along the pyramid base diagonals is only observed in the capped pyramid islands. Thus, the above kinetic growth model based on the island growth process cannot be simply used to explain the observed Si enrichment inside the capped pyramid islands, especially the Si enrichment along the pyramid base diagonals of a capped island. Two kinds of atomic processes may result in the observed Si enrichment with a content as large as 94% along the pyramid base diagonals in a capped island. One is the bulk atomic interdiffusion at the interface between the SiGe island and Si substrate. However, one previous study8 clearly shows that at the temperature of 620 ° C the Si–Ge bulk interdiffusion at the interface is too small to cause the Si content to be as large as 35% inside a SiGe island. In order to explain to the high Si content inside the island, Si incorporating into the island via surface diffusion during an island growth is suggested. Moreover, Denker et al.9 show that even at a higher temperature of 740 ° C the Si–Ge bulk interdiffusion at the interface is still negligible to the formation of the Si enrichment with a half-moon shape and a Si content as large as 94%, since this half-moon shape remains almost unchanged with in situ annealing at 740 ° C for a time period


103108-3

Appl. Phys. Lett. 89, 103108 共2006兲

Li et al.

FIG. 2. Si 共a兲 and Ge 共b兲 Young’s modulus along the 具100典 and 具110典 directions in 共001兲 plane at different temperatures. The data are obtained according to the data from Ref. 13 by calculation.

varied from 20 min to 10 h. In our case, the islands are capped at 640 ° C which is much lower than 740 ° C, thus the Si–Ge bulk interdiffusion at the interface should be negligible to the Si enriching along the pyramid base diagonals. The other kind of atomic process to result in our observed Si enrichment along the pyramid base diagonals is a little more complicated one, in which Ge atomic surface segregation and associated Si atomic interdiffusion into the island are needed. The capped Si atoms that aggregated to an island by surface diffusion mix at the island perimeter with Ge atoms which segregated from inside the island and then diffuse into the island. This diffusion of Si atoms into the island is facilitated by the surface segregation of the Ge atoms in the island. However, even with this atomic exchange model, we could not yet explain the Si enrichment along the pyramid base diagonals in a capped island. We suggest to attribute this Si enrichment to the elastic constant anisotropy of Si and Ge materials. Figure 2 shows the Si and the Ge Young’s moduli along the 具110典 and 具100典 directions in 共001兲 plane at different temperatures;13 Young’s modulus along 具110典 direction is much larger than that along the 具100典 direction in both Si and Ge materials. With the same strain, the strain energy density along the 具110典 directions 共pyramid base diagonal directions兲 will be much larger than that along the 具100典 directions 共pyramid base edge directions兲. From the viewpoint of strain energy, the strain along the 具110典 direction should be the lowest since Young’s modulus along this direction is the largest in 共001兲 plane,14 that is, the fully relaxed lattice constant along the 具110典 direction should be most close to the lattice constant of Si substrate, which results in the Si enrichment along the pyramid island base diagonals. Actually, Denker et al.14 have observed an anisotropy of the

trench around a SiGe island and attributed this anisotropy to the anisotropy of Young’s modulus in Si 共001兲 plane. In conclusion, we have studied the 6% Ge isocomposition profile change of SiGe islands during initial Si capping. Before Si capping the SiGe islands are more highly Si intermixed at one corner of the island base, while during Si capping at a temperature of 640 ° C, the SiGe islands become to be more highly Si intermixed along the SiGe pyramid base diagonals. This Si enrichment evolution inside the islands during Si capping is caused by the exchange of capped Si atoms that aggregated to islands by surface diffusion with Ge atoms from inside the islands by both atomic surface segregation and interdiffusion, while the Si enrichment along the pyramid base diagonals in a capped island is attributed to the elastic constant anisotropy of Si and Ge materials. This work was supported by the special funds for Major State Basic Research Project No. G2001CB3095 of China and by NSFC 共60425411 and 10321003兲 and Shanghai Science and Technology Commission. This work was also supported by Australian Research Council. 1

S. Tong, J. L. Liu, J. Wan, and K. L. Wang, Appl. Phys. Lett. 80, 1189 共2002兲. F. M. Ross, R. M. Tromp, and M. C. Reuter, Science 286, 1931 共1999兲. 3 T. U. Schülli, J. Stangl, Z. Zhong, R. T. Lechner, M. Sztucki, T. H. Metzger, and G. Bauer, Phys. Rev. Lett. 90, 066105 共2003兲. 4 J. Stangl, A. Daniel, V. Holý, T. Roch, G. Bauer, I. Kegel, T. H. Metzger, Th. Wiebach, O. G. Schmidt, and K. Eberl, Appl. Phys. Lett. 79, 1474 共2001兲. 5 F. Boscherini, G. Capellini, L. Di Gaspare, F. Rosei, N. Motta, and S. Mobilio, Appl. Phys. Lett. 76, 682 共2000兲. 6 G. Capellini, M. De Seta, and F. Evangelisti, Appl. Phys. Lett. 78, 303 共2001兲. 7 U. Denker, M. Stoffel, and O. G. Schmidt, Phys. Rev. Lett. 90, 196102 共2003兲. 8 G. Katsaros, G. Costantini, M. Stoffel, R. Esteban, A. M. Bittner, A. Rastelli, U. Denker, O. G. Schmidt, and K. Kern, Phys. Rev. B 72, 195320 共2005兲. 9 U. Denker, A. Rastelli, M. Stoffel, J. Tersoff, G. Katsaros, G. Costantini, K. Kern, N. Y. Jin-Phillipp, D. E. Jesson, and O. G. Schmidt, Phys. Rev. Lett. 94, 216103 共2005兲. 10 T. K. Carns, M. O. Tanner, and K. L. Wang, J. Electrochem. Soc. 142, 1260 共1995兲. 11 J. Cui, Q. He, X. M. Jiang, Y. L. Fan, X. J. Yang, F. Xue, and Z. M. Jiang, Appl. Phys. Lett. 83, 2907 共2003兲. 12 Y. Q. Wu, F. H. Li, J. Cui, J. H. Lin, R. Wu, J. Qin, C. Y. Zhu, Y. L. Fan, X. J. Yang, and Z. M. Jiang, Appl. Phys. Lett. 87, 223116 共2005兲. 13 Numerical Data and Functional Relationships in Science and Technology, Landolt-Börnstein 共Springer, Berlin 1982兲, Group III, Vol. 17, Part A, 63 and 105. 14 U. Denker, O. G. Schmidt, N. Y. Jin-Phillipp, and K. Eberl, Appl. Phys. Lett. 78, 3723 共2001兲. 2
