Carrier gas effects on the SiGe quantum dots formation

(1)

Carrier gas effects on the SiGe quantum dots formation

C.-H. Lee

a

, C.-Y. Yu

a

, C.M. Lin

a

, C.W. Liu

a,b,

*

, H. Lin

c

, W.-H. Chang

c

a_{Department of Electrical Engineering and Graduate Institute of Electronics Engineering, National Taiwan University, Taipei, Taiwan, ROC} b

National Nano Device Laboratories, Hsinchu, Taiwan, ROC

c

Department of Electrophysics, National Chiao Tung University, Hsinchu, Taiwan, ROC

1. Introduction

Self-assembled SiGe quantum dots (QDs) have attracted much interest in the potential applications in nanoelectronics and optoelectronics recently. To meet the requirement of device applications, size and shape uniformity are the two most important parameters to be considered. Inﬂuences of temperature[1], wetting-layer growth[2], and Si capping[3]on dot size and dot shape have been reported in previous works. Carrier gases such as He[4], H2[5], and N2 [6], have been used in the process to control the partial pressure of precursors. However, the comparison of different carrier gases was not reported yet. In this work, morphologies of SiGe QDs grown in He and H2were characterized by atomic force microscopy (AFM) and high-angle annular dark-ﬁeld scanning transmission electron microscopy (HAADF-STEM). Strain relaxation and Ge composition were investigated through Raman spectroscopy, low temperature photoluminescence (PL) measurement, and energy dispersive X-ray spectroscopy (EDS). Moreover, a simple epitaxial model is proposed for He and H2growth.

2. Experimental procedures

SiGe QDs were grown by ultra-high vacuum chemical vapor deposition (UHV/CVD) at 600 8C. The base pressure of our UHVCVD system was ultra-high vacuum of 10 9_{Torr. Pure silane (SiH}

4) and germane (GeH4) were used as reactant gases, and the gas flow ratio of GeH4 and carrier gases (H2 and He) was fixed at 100 sccm/ 100 sccm for QDs growth. Before the epitaxial growth, Si wafers were dipped in 10% HF solution to remove the native oxide. No SiH4 was used during QDs growth. Due to the Si/Ge interdiffusion at 600 8C, Ge layers transformed into SiGe alloys. After the deposition of 60-nm-thick Si buffer layer using the SiH4 flow rate of 100 sccm, 1-layer and 5-layers QDs were grown. Finally, a 15-nm Si cap layer using SiH4flow rate of 100 sccm was grown on top of the SiGe QDs layer to avoid oxidation of the SiGe surface. Note that there were no carrier gases during the Si buffer layer and Si cap layer growth. In this work, we used pure germane for the reactant gas instead of dilute germane in He or H2, and the gas flow maintained at 100 sccm for QDs growth. The Ge partial pressure should be the same for H2and He growth. For the SiGe quantum well growth at 500 8C, the thickness of SiGe film is 3 nm for both He growth and H2growth. The carrier gas effect on the Ge growth rate seems similar in this work. Variations in Ge composition, dot density and strain are due to the surface effect instead of the Ge partial pressure (dose).

Applied Surface Science 254 (2008) 6257–6260

A R T I C L E I N F O

Article history:

Available online 18 March 2008 PACS: 68.65.Ac 68.65.Hb 68.55.-a 66.30.-h Keywords: Carrier gas effect SiGe quantum dot Hydrogen passivation Surface mobility UHVCVD

A B S T R A C T

SiGe quantum dots (QDs) grown by ultra-high vacuum chemical vapor deposition using H2and He carrier gases are investigated and compared. SiGe QDs using He carrier gas have smaller dot size with a better uniformity in terms of dot height and dot base as compared to the H2carrier gas. There is a higher Ge composition and less compressive strain in the SiGe QDs grown in He than in H2as measured by Raman spectroscopy. The Ge content is higher for He growth than H2growth due to hydrogen induced Si segregation and the lower interdiffusivity caused by the more strain relaxation in the He-grown SiGe dots. The photoluminescence also conﬁrms more compressive strain for H2growth than He growth. Hydrogen passivation and Ge–H cluster formation play an important role in the QDs growth.

* Corresponding author at: Department of Electrical Engineering and Graduate Institute of Electronics Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan, ROC. Tel.: +886 2 23635251x515; fax: +886 2 23638247.

E-mail address:[email protected](C.W. Liu).

C o n t e n t s l i s t s a v a i l a b l e a tS c i e n c e D i r e c t

Applied Surface Science

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / a p s u s c

(2)

3. Results and discussion

Morphologies of 1-layer SiGe QDs without Si capping layer grown in He and in H2have been observed by AFM (Fig. 1). Average dot heights are 10 and 12 nm for He and H2growth, while average dot bases are 88 and 89 nm for He and H2growth, respectively. Dot densities are similar (1.1 1010cm 2) for He-grown sample and H2-grown sample. The average dot height indicates that H2-grown sample can have slightly taller dots (12 nm height) than He-grown sample (10 nm height) measured by AFM. The HAADF-STEM also shows the similar trend. Dot height distribution (Fig. 2) and dot base distribution (Fig. 3) are different for H2growth and He growth. For H2-grown sample, the bimodality is clearly evidenced in H2 growth for dot height distribution (Fig. 2(b)), and a broad distribution is observed for dot base (Fig. 3). Standard deviations of dot height are 3.2 and 4.4 nm for He growth and H2 growth,

respectively, and those of dot base are 29 and 41 nm for He growth and H2 growth, respectively. QDs grown in He have better uniformity than in H2. Different morphologies are also found through HAADF-STEM images. The scanning transmission electron microscopy of 5-layers QDs grown in He and H2 are shown in

Fig. 4(a) and (b), respectively. The thickness of the Si layer between the SiGe QDs layers is 60 nm. H2carrier gas creates taller dots than He carrier gas, consistent with AFM results.

Raman spectra of 5-layers SiGe QDs grown in He and H2with 488 nm laser excitation are shown inFig. 5with the resolution of 0.2 cm 1_{. Besides strong bulk Si signal at 520 cm} 1_{, the Si–Ge 2TA} phonon, Ge–Ge phonon, and Si–Ge phonon of dots were also observed. For the peak at 417 cm 1_{, QDs contribute more} intensity of the Raman spectra than wetting layers. The sample with wetting layers only has very weak intensity at 417 cm 1_. The Si–Ge phonon peak (417 cm 1_{) is sensitive to both the strain}

Fig. 1. The 2.5mm 2.5mm AFM image of SiGe QDs grown in different carrier gases: (a) He and (b) H2.

Fig. 2. The dot height distribution of SiGe QDs grown in (a) He and (b) H2. The bimodality is clearly evidenced in H2growth.

Fig. 3. The dot base distribution of SiGe QDs grown in (a) He and (b) H2.

C.-H. Lee et al. / Applied Surface Science 254 (2008) 6257–6260 6258

(3)

and Ge composition[7]. Both more compressive strain and higher Ge content increase the wave number of this peak. The Si–Ge 2TA phonon peak (at 220 cm 1_{) is sensitive to only the Ge} concentration[8,9]. Wave numbers of 2TA mode are 222 and 223 cm 1for He growth and H2growth, respectively. Note that the 1 cm 1_{Raman shift between He-grown and H}

2-grown QDs is well beyond the resolution (0.2 cm 1_{) of the Raman measurement.} The 1 cm 1 _{lower in wave number of 2TA mode indicates the} average Ge concentration of He-grown SiGe layers is 2% higher than H2-grown layers[10]. The 2% larger Ge content leads to a 0.6 cm 1 _{shift of the Raman peak at 417 cm} 1 _[11]_{if lattice} strain are the same in He-grown and H2-grown samples. However, wave numbers of Si–Ge phonon peaks are both at 416.5 cm 1_for He and H2growth. This result indicates that He-grown QDs have less strain (more relaxation) than H2-grown QDs. The larger Ge concentration in SiGe QDs grown in He is partly attributed to the highly relaxed nature of the dots. Note that H2 carrier gas can create slightly taller dots than He carrier gas. The smaller dot height in He-grown sample, which corresponds to a smaller contact angle, leads to more relaxation than H2-grown sample due to the decrease of strain energy[12].

Higher Ge content suggested in the Raman measurement is due to suppressed Ge/Si interdiffusion in He-grown QDs. The stronger Si/Ge interdiffusion due to larger compressive strain[13]in H2 -grown sample leads to less Ge content, which is consistent with the conclusions from Raman measurement. H2 carrier gas, which increases Si segregation during QDs growth [14] can also be responsible for the smaller Ge composition in H2-grown sample.

Note that the Ge concentration difference between He-grown and H2-grown samples cannot be resolved by EDS.Fig. 6shows the low temperature (10 K) PL measurement of 5-layers QDs with He and H2 carrier gases. The PL emission peak of He-grown sample is 13 meV higher than that of H2-grown sample. Ge concentration and strain might be responsible for the emission peak shift. For the effect of Ge content, the emission peak of He-grown sample should be 6 meV lower than that of H2-grown sample due to higher Ge content[15]. However, this is not consistent with the experimental measurement. Strain of H2-grown sample is 0.11% higher than that of He-grown sample based on the Raman result [8]. The emission peak of the He-grown sample would be 12 meV higher than that of H2-grown sample due to less compressive strain in QDs. Considering the effect of Ge content and strain simulta-neously, the net emission peak shift is 6 meV. He-grown sample has higher PL emission peak energy than H2-grown sample, which agrees with the experimental data qualitatively. The error of quantitative analysis (6 vs. 13 meV) is probably due to some measurement error, but is not fully understood.

Carrier gas effect plays an important role in the QDs growth.Fig. 7

shows a simple growth model for H2and He growth. For H2carrier gas, the Si surface is passivated by hydrogen due to the adsorption of H2from the environment[16]. Hydrogen passivation can block the surface sites for dissociative adsorption of germane[17]. H2-rich environment also enhances the Ge–H cluster formation, which increases Ge surface mobility[18]and strengthens surface diffusion. Ge atoms can diffuse on the surface and assemble with other Ge atoms which have been already adsorbed on surface sites.

Never-Fig. 6. The PL spectrum of the SiGe dots grown in He and H2. The 13 meV variation

of the emission peak is caused by the difference in compressive strain. Fig. 5. The Raman spectra of SiGe QDs grown in He and H2.

Fig. 4. The HAADF-STEM images of the regular regions of QDs grown in (a) He and (b) H2.

(4)

theless, for He carrier gas, hydrogen passivation decreases since there are no H2in the environment. The He carrier gas ﬂow can take away the hydrogen which is desorbed on the surface, and surface mobility can also reduce due to the reduction of Ge–H cluster. Therefore, H2carrier gas creates taller dots and worse uniformity in dots distribution than He carrier gas.

4. Conclusion

In conclusion, morphologies, compressive strain and Ge composition of QDs in He and in H2 carrier gases are studied. The He growth has a smaller dot size and a better uniformity in dots height and base width distribution than H2growth. Raman results show higher Ge concentration and more relaxation in 5-layers QDs layer grown in He as compared to those grown in H2 carrier gas. Higher dot density for He growth leads to a larger strain relaxation in SiGe layer as compared to H2 growth. Hydrogen induced Si segregation and slower Si/Ge interdiffusion due to smaller strain are responsible for the lower Ge content in grown sample. PL spectra also conﬁrm smaller compressive in He-grown sample. Hydrogen passivation during the process and Ge–H cluster formation should be responsible for the difference between He growth and H2growth.

Acknowledgements

The authors would like to acknowledge Dr. Yung-Hui Yeh and Mr. H. T. Chen at the DTC/ITRI for the Raman measurement, and D.

J. Lockwood, J. -M. Baribeau and X. Wu at the National Research Council of Canada for the HAADF-STEM measurement. This work was supported by National Nano Device Laboratories and National Science Council of ROC under contract no. 95-2221-E-002-370.

Reference

[1] G. Jin, J.L. Liu, K.L. Wang, Appl. Phys. Lett. 83 (2003) 2847. [2] H.J. Kim, Y.H. Xie, Appl. Phys. Lett. 79 (2001) 263.

[3] Y.Q. Wu, F.H. Li, J. Cui, J.H. Lin, R. Wu, J. Qin, C.Y. Zhu, Y.L. Fan, X.J. Yang, Z.M. Jiang, Appl. Phys. Lett. 87 (2005) 223116.

[4] S.W. Lee, P.S. Chen, T.Y. Chien, L.J. Chen, C.T. Chia, C.W. Liu, Thin Solid Films 508 (2006) 120.

[5] B.S. Meyerson, Appl. Phys. Lett. 48 (1986) 797.

[6] P. Meunier-Beillard, M. Caymax, K. Van Nieuwenhuysen, G. Doumen, B. Brijs, M. Hopstaken, L. Geenen, W. Vandervorst, Appl. Surf. Sci. 224 (2004) 31. [7] M. Stoehr, D. Aubel, S. Juillaguet, J.L. Bischoff, L. Kubler, D. Bolmont, F. Hamdani, B.

Fraisse, R. Fourcade, Phys. Rev. B 53 (1996) 6923.

[8] P.H. Tan, K. Brunner, D. Bougeard, G. Abstreiter, Phys. Rev. B 68 (2003) 125302. [9] M.H. Liao, C.-H. Lee, T.-A. Hung, C.W. Liu, J. Appl. Phys. 102 (2007) 053520. [10] J.S. Lannin, Phys. Rev. B 16 (1977) 1510.

[11] J.-M. Baribeau, X. Wu, D.J. Lockwood, J. Vac. Sci. Technol. A 24 (2006) 663. [12] H.T. Johnson, L.B. Freund, J. Appl. Phys. 81 (1997) 6081.

[13] G. Xia, M. Canonico, J.L. Hoyt, International SiGe Technology and Device Meeting (ISTDM), 2006.

[14] E. Rudkevich, F. Liu, D.E. Savage, T.F. Kuech, L. McCaughan, M.G. Lagally, Phys. Rev. Lett. 81 (1998) 3467.

[15] J.-P. Noe¨l, N.L. Rowell, D.C. Houghton, D.D. Perovic, Appl. Phys. Lett. 57 (1990) 1037.

[16] P. Bratu, W. Brenig, A. Groß, M. Hartmann, U. Ho¨fer, P. Kratzer, R. Russ, Phys. Rev. B 54 (1996) 5978.

[17] R. Larciprete, S. Cozzi, E. Masetti, M. Montecchi, G. Padeletti, Thin Solid Films 315 (1998) 49.

[18] A. Nayfeh, C.O. Chui, K.C. Saraswat, T. Yonehara, Appl. Phys. Lett. 85 (2004) 2815. Fig. 7. The growth model of SiGe dots grown in (a) He and (b) H2. The major difference is the hydrogen induced H-passivation on the surface.

C.-H. Lee et al. / Applied Surface Science 254 (2008) 6257–6260 6260