Views
4 years ago

Bi: Perfect surfactant for Ge growth on Si?111??

Bi: Perfect surfactant for Ge growth on Si?111??

1392 Appl. Phys. Lett.,

1392 Appl. Phys. Lett., Vol. 74, No. 10, 8 March 1999 Schmidt et al. FIG. 1. XSW data obtained after ong>growthong> of 15 BL ong>Geong> on ong>Biong>:Si111-) ) in 111 and 220 Bragg reflection, reflectivity , ong>Biong>-L and ong>Geong> K fluorescence. Explanations regarding basic theory and analysis of XSW can be found in the literature. 13 The output parameters of an XSW analysis are the coherent position c and coherent fraction f c . These are defined as the phase and amplitude of the corresponding Fourier coefficient of the atomic distribution function and ong>forong> ong>Geong> can hence be calculated from A 111 1 111 c 2 argA 111 N1 0 , 2 2N 111 f c A111sin 2 N sin 2 2. If ong>Biong> successfully acts as a ong>surfactantong> then the ong>Biong> XSW results should strongly differ from that of ong>Geong>. For SME ong>growthong>, all ong>Biong> is expected to float on the surface and occupy sites at identical height with respect to the surface, depending on the surface roughness. Conclusively, one expects a large ong>Biong> coherent fraction and a small ong>Geong> coherent fraction. Figure 1 displays ong>Biong> and ong>Geong> XSW data obtained after ong>growthong> of 15 bilayered BL ong>Geong> on ong>Biong>:Si111-)) at 485 °C. The ong>Biong> and ong>Geong> coherent position and fraction are also shown. The results obtained in 111 Bragg reflection are in perfect agreement with the expected values ong>forong> SME ong>growthong>. For a completely relaxed ong>Geong> film (0.0423, 0 0.02, cos(2(1)/8), N16 14 111 Eq. 1 yields c 111 0.34 which is physically equivalent to c 1.34) and 111 f c 0.28 in perfect agreement with the experimental results ong>forong> ong>Geong>. In SME, the ong>Biong> coherent position is expected to shift according to the number of ong>Geong> layers and the difference in 111 lattice plane spacing between ong>Geong> and Si: c 1 111 (N) (N1)00. From measurements at low ong>Biong> coverages (1 ML), we deduce the additional offset due to the ong>Biong>-ong>Geong> bond length to be 00.020.01. 11 Thus, ong>forong> N 111 16 we obtain c 0.690.01, again perfectly reproducing the experimental results. 111 The coherent ong>Biong> fraction f ong>Biong> did not change during ong>growthong>. After deposition of 15 BL ong>Geong> we measured the same value as ong>forong> the starting ong>Biong>:Si111-)) surface, i.e., 111 f ong>Biong> 0.780.03. The difference from unity can be attributed to the presence of incoherent ong>Biong> islands on the surface. These islands are due to excess ong>Biong> on the surface. FIG. 2. CTR data and best fit solid lines after deposition of 11.5 BL and 35 BL ong>Geong> on ong>Biong>:Si111-)), Q1 corresponds to the 111 Bragg reflection. The measurements of crystal truncation rods CTR is a sensitive tool ong>forong> the investigation of film thicknesses and the roughness of interfaces and surfaces 15,16 and has even been successfully employed ong>forong> the characterization of ultrathin buried ong>Geong> structures. 17,18 Within the kinematical approximation, ong>forong> a ong>Geong> film grown by SME, the CTR intensity can be calculated to be IQF SiQ R int Si 1e iQa Si F ong>Geong>Q R int surf iQNa ong>Geong>Rong>Geong> •e ong>Geong> 1e iQa •e ong>Geong> iQaint surf iQa F ong>Biong>Q ong>Biong>Rong>Geong> •e inteiQNaong>Geong>eiQa 2 ong>Biong> . 2 The three terms in the model function reflect the sample geometry, i.e., scattering contributions from the Si substrate, the ong>Geong> film and the ong>Biong> surface layer. The bilayer geometry of the Si111 and ong>Geong>111 planes is found in the factor: FSi,ong>Geong>(Q) f Si,ong>Geong>(Q)2 cos((1/8)QaSi,ong>Geong>), however Fong>Biong>(Q) f ong>Biong>(Q). Here f ong>Biong>,ong>Geong>,Si(Q) denotes the atomic scattering factor of the corresponding element. Nong>Geong> is the number of ong>Geong> layers in the film. For noninteger values of N, we used the weighted average of Nn and Nn1; aSi,ong>Geong> are the lattice spacing of the Si substrate and the ong>Geong> layer, respectively. The lattice spacings at the Si-ong>Geong> interface are denoted as aint . For the lattice spacing between ong>Geong> and ong>Biong>, aong>Biong> is used. The surface coverage of ong>Biong> is given by ong>Biong> . In Eq. 2, both surface and interface roughness are described by a Gaussian y (1/2)q distribution function: Rxe 2 2 2 ax y, with qQ2/a. Figure 2 shows CTR measured after deposition of 11.5

Appl. Phys. Lett., Vol. 74, No. 10, 8 March 1999 Schmidt et al. FIG. 3. STM image after deposition of 10 BL ong>Geong> on ong>Biong>:Si111-)) at 485 °C. The inset shows a zoom of 400400 Å 2 . and 35 BL ong>Geong> on ong>Biong>:Si111-)) at 485 °C. The results ong>forong> the major model parameters obtained from least square fits are also displayed. From the fit we determine the ong>Geong> lattice spacing to be equal to the bulk lattice spacing of ong>Geong> bulk aong>Geong>a ong>Geong> (1.0010.001), even after deposition of only 11.5 BL ong>Geong>. Hence, at these coverages the ong>Geong> film is vertically completely relaxed. We find an interface roughness of 1.6–1.8 Å, which is of the order of the roughness of our starting surface. The surface roughness of the films is found in the order of 3–4 Å, decreasing with film thickness. This effect is due to a rough ong>growthong> mode ong>forong> a ong>Geong> film thickness around the transition from pseudomorphically strained to relaxed film ong>growthong>. This process has been investigated in detail by XSW, CTR, and STM and will be the subject of a separate paper. 19 Figure 3 shows a 20002000 Å2 STM image obtained after deposition of 10 BL ong>Geong> at 485 °C. Three terrace levels are visible, in good agreement with the small surface roughness determined from the CTR measurements. In addition to single steps at the edges of neighboring terraces a periodic surface height modulation is found in the image with an amplitude of 0.5 Å. This feature is even more clearly visible in the inset of Fig. 3 and indicated by superimposing the array of white dots. This frame represents an area of 400 400 Å2 of the same surface. The additional fine structure visible in this image is due to the )) surface structure, resolved here. The valleys of the periodic surface height modulation white dots are aligned in 11¯0] like orientations and the distance between neighboring rows of minima indicated by white dots is approximately 105 Å. These observations indicate the presence of a periodic dislocation net- work at the Si-ong>Geong> interface as also found ong>forong> Sb as ong>surfactantong>. 6 The ong>forong>mation of this dislocation network enables the relaxation of the pseudomorphically strained ong>Geong> film. The evaluation of the periodicity of the dislocation network shows a laterally completely relaxed ong>Geong> film within the experimental uncertainty. Auger electron spectroscopy AES was employed to determine the desorption temperature of ong>Biong> from ong>Geong> films on Si111. At substrate temperatures exceeding 500 °C rapid ong>Biong> desorption was observed, resulting in complete removal of ong>Biong> from the surface within two minutes. On bare Si111 surfaces the same effect is found at 550 °C. Hence a large ong>Biong> flux is required to compensate ong>forong> the ong>Biong> desorption during ong>Geong> deposition to successfully perong>forong>m ong>Biong>-SME. In conclusion, ong>Biong> acts as ong>surfactantong> ong>forong> ong>Geong> ong>growthong> on Si111. A layer of ong>Biong> segregates to the surface during ong>Geong> deposition and alters the ong>Geong> ong>growthong> from a Stranski- Krastanov mode to a layer-by-layer like ong>growthong> mode ong>forong> larger amounts of ong>Geong>. Measurements of XSW, CTR and STM show that completely relaxed ong>Geong> films are found after deposition of more than 10 BL ong>Geong>. The relaxation proceeds by the ong>forong>mation of a highly periodic dislocation network at the Si-ong>Geong> interface. Further ong>Geong> deposition leads to the ong>growthong> of ong>Geong> films with smooth surfaces and interfaces. Most importantly ong>forong> applications, the ong>surfactantong> ong>Biong> can easily be removed from the surface by annealing at temperatures above 500 °C. 1 M. Copel, M. C. Reuter, E. Kaxiras, and R. M. Tromp, Phys. Rev. Lett. 63, 632 1989. 2 M. Horn von Hoegen, F. K. LeGoues, M. Copel, M. C. Reuter, and R. M. Tromp, Phys. Rev. Lett. 67, 1130 1991. 3 J. Falta, M. Copel, F. K. LeGoues, and R. M. Tromp, Appl. Phys. Lett. 62, 2962 1993. 4 J. Falta, T. Schmidt, A. Hille, and G. Materlik, Phys. Rev. B 54, R17288 1996. 5 S. Maruno, S. Fujita, H. Watanabe, Y. Kusumi, and M. Ichikawa, Appl. Phys. Lett. 68, 2213 1996. 6 M. Horn von Hoegen, A. Al Falou, H. Pietsch, B. H. Müller, and M. Henzler, Surf. Sci. 298, 291993. 7 G. Meyer, B. Voigtländer, and N. M. Amer, Surf. Sci. 274, L541 1992. 8 B. Voigtländer and A. Zinner, J. Vac. Sci. Technol. A 12, 1932 1994. 9 H. Minoda, S. Sakamoto, and K. Yagi, Surf. Sci. 372, 11997. 10 J. C. Woicik, G. E. Franklin, C. Liu, R. E. Martinez, I.-S. Hwong, M. J. Bedzyk, J. R. Patel, and J. A. Golovchenko, Phys. Rev. B 50, 12246 1994. 11 T. Schmidt, J. Falta, and G. Materlik unpublished. 12 E. Vlieg, A. E. M. J. Fischer, J. F. van der Veen, B. N. Dev, and G. Materlik, Surf. Sci. 178, 361986. 13 J. Zegenhagen, G. Materlik, and W. Uelhoff, J. X-Ray Sci. Technol. 2, 214 1990. 14 The lateral atom density of a completely relaxed ong>Geong> film compared to that of a pseudomorphically strained ong>Geong> film is smaller by a factor of 1/1.04 2 . Thereong>forong>e N16 ong>forong> 15 BL. 15 I. K. Robinson, Phys. Rev. B 33, 3830 1986. 16 I. K. Robinson, R. T. Tung, and R. Feidenhans’l, Phys. Rev. B 38, 3632 1988. 17 J. Falta, D. Bahr, G. Materlik, B. H. Müller, and M. Horn von Hoegen, Appl. Phys. Lett. 68, 1394 1996. 18 D. Bahr, J. Falta, G. Materlik, B. H. Müller, and M. Horn von Hoegen, Physica B 221, 961996. 19 T. Schmidt, J. Falta, and G. Materlik unpublished. 1393

Growth of Ge hexagonal meshwork films on Si(111)-7 × 7
X-ray Analysis of Si/Ge/Si(001) - Bedzyk Research Group
Effect of carbonization on the growth of 3C-SiC on Si (111) by ...
Inhomogeneous growth of CaF 2 adlayers on Si(111 ... - ScienceDirect
MBE Growth of SiC on Si (111) - IEEE Xplore
Surfactant-assisted epitaxial growth and magnetism - IMDEA ...
Growth of strained Si and strained Ge heterostructures on relaxed ...
Si passivation for Ge pMOSFETs: Impact of Si cap growth conditions
Growth of GaN epilayers on Si(111) substrates using multiple buffer ...
ag on au/si(111) - Synchrotron Radiation Center - University of ...
Heteroepitaxial Growth of Ge Nanowires on Si ... - ResearchGate
LOW-TEMPERATURE Si (111) HOMOEPITAXY AND DOPING ...
Surface alloying of Pb as a surfactant during epitaxial growth on Cu ...
Epitaxial growth of Fe-Si compounds on the silicon (111) face - CNRS
GaAs on Si„111…—crystal shape and strain relaxation in nanoscale ...
Challenges in epitaxial growth of SiGe buffers on Si (111), (110 ...
ž / Epitaxial growth of silver on Br-passivated Si 111 substrates ...
Substrate-step-induced effects on the growth of CaF2 on Si (111)
Lattice-dynamical calculation of phonon scattering at ideal Si–Ge ...
Photonic band engineering in opals by growth of Si/Ge multilayer ...
Growth of Ge on Si(0 0 1) - J. Eymery's Web Page
Interfacial reactions of Ni on Si 1-x Ge x (x= 0.2, 0.3) at low ...
3C-SiC growth on Si substrates via CVD: An introduction - Mansic
Growth of Si-NWs on the thin films of conductive and transparent thin ...
Low-temperature solid-phase heteroepitaxial growth of Ge-rich ...