988
(Invited) Atomically Controlled Processing for Si and Ge CVD Epitaxial Growth

Monday, 30 May 2016: 11:00
Sapphire 410 A (Hilton San Diego Bayfront)
J. Murota (Tohoku University, Japan), Y. Yamamoto, I. Costina (IHP), B. Tillack (IHP, Technische Universitat Berlin, Germany), V. Le Thanh (Universite Aix Marseille, France), R. Loo, and M. Caymax (imec, Belgium)
Atomically controlled processing has become indispensable for the fabrication of Si-based ultrasmall devices and heterodevices for ultralarge scale integration.  Our concept of atomically controlled processing is based on atomic-order surface reaction control [1-3]. The final goal is the generalization of atomic-order surface reaction processes and the creation of new properties in Si-based ultimate small structures. In the stage of nanometer-scale device fabrication, preparing atomic-order controlled surfaces and interfaces and the introduction of Ge into the Si process are required. In this work, we describe thesegregation of dopants to the surface and surface roughness generation observed during in-situ doping and their suppression by applying an atomic layer doping approach. Furthermore results of nm-order thick Ge epitaxial growth on Si(100) using an ultraclean low-pressure CVD system are presented.

Figure 1 shows the P depth profile in a Si epitaxial film.  Even after switching off the PH3 gas supply, P atoms are incorporated into the Si layer during Si epitaxial growth.  In the steady state, it is estimated that P concentration saturates at around 3x1019cm-3 and P coverage on the Si surface is about 0.1 monolayer (6.8x1013cm-2). These values are calculated from the PH3 partial pressure and Si growth rate shown in Fig. 1 based on the surface reaction model shown in Ref. 3. The P incorporation rate, that is the product of the Si growth rate (0.8nm/min) and P concentration (about 1x1019cm-3), is about 8x1011cm-2/min. If it is assumed that the desorption of P atoms on the Si surface can be neglected [4], 6.8x1013cm-2 P atoms on the surface are incorporated into the Si epitaxial layer with 8x1011cm-2/min even after the PH3 supply has been stopped.  The increase of surface roughness from RMS 0.1nm to 0.8nm during in-situ doping can be explained by the suppression of the SiH4reaction caused by the P atoms which are present on the surface.   Suppression of P segregation and surface roughening are achieved using P atomic layer doping in Si.  This includes the optimization of the capping Si layer deposition conditions and the P atomic amount [5] as shown in Fig. 2. This may result from low temperature Si epitaxial growth on P atoms with high growth rate. That could be demonstrated for P atomic layer doping during Ge epitaxial growth too [6].

In a second part of the paper results for nm-order-thick Ge high quality epitaxial growth on Si (100) are described. Because pure Ge and Si oxide are formed on Si substrate even at 300oC by the reaction between vaporized Ge oxide and Si substrate [7], the suppression of Ge oxide vaporization in the CVD reactor is very important.  It was proposed that Ge island growth in an initial growth stage is caused by the suppression of adsorption and/or decomposition of GeH4 on the H-terminated Si surface [8].  Even on the initially H-free Si surface, Si atoms are H-terminated by the decomposition of GeH4.  In order to decrease the hydrogen desorption temperature from the Si site, nm-thick Si1-xGex buffer layer on Si (100) was proposed [9,10]. The suppression of surface roughening was improved by nm-order-thick Ge epitaxial growth on nm-order-thick Si1-xGex/Si(100). The surface segregation of Si through nm-order-thick Ge layer from the substrate was observed at Ge growth temperature of 350oC.

These results demonstrate the capability of atomically controlled processing for nm-order thick Si and Ge CVD epitaxial growth for ultralarge scale integration.

This work is partially supported by the JSPS Core-to-Core Program, “International Collaborative Research Center on Atomically Controlled Processing for Ultralarge Scale Integration”.

References

[1] J. Murota et al., Jpn. J. Appl. Phys., 33, 2290 (1994).

[2] B. Tillack et al., Thin Solid Films, 369,.189 (2000).

[3] J. Murota et al, Jpn. J. Appl.Phys., 45, 6767 (2006).

[4] Y. Shimamune et al,. Surf. Appl. Phys., 162-163, 390 (2000).

[5] Y. Chiba et al., Thin Solid Films, 518, 5231 (2010).

[6] Y. Yamamoto et al., Solid-State Electron., 83, 25 (2013).

[7] K Minami et al., Solid-State Electron., 110, 40 (2015).

[8] S. Kobayashi et al., J. Cryst. Growth, 174, 686(1997).

[9] J. Murota et al., Adv. Nat. Sci.:Nanosci. Nanotechnol., 6, 015001 (2015).

[10] J. Murota et al., ECS Trans. 67(1), 135(2015).