Equilibrium surface segregation enthalpy of Ge in concentrated amorphous SiGe alloys
Introduction
Numerous works reported that, during the growth of epitaxial Si/Ge layers, the control of Ge concentration is difficult due to the preferential surface segregation of Ge [1], [2], [3]. The driving force of germanium surface segregation in Si–Ge system is due to the difference in the surface tension of the elements [4], [5] and the difference in size between Ge and Si atoms favours also the phenomenon (the atomic radius of these elements are the following rGe=0.137 nm and rSi=0.132 nm). However, the changes of surface composition with temperature have been studied only in few works [6], [7] and to our knowledge only one experimental work is devoted to the study of equilibrium segregation of Ge in SiGe alloys [8], which can be described in terms of phase transformation and characterized by a segregation enthalpy.
In this paper, we present an AES study of Ge surface segregation kinetics in a series of concentrated homogeneous amorphous Si1−xGex thin films in the temperature range of 653–673 K (amorphous Ge crystallizes at about 693 K [9]). The main difficulty of this study was to determine equilibrium data (Xs(Ge) versus Xb(Ge)) from AES kinetics results (i.e. the variations of the peak-to-peak heights of Si and Ge Auger signals versus time). Firstly, since the Ge (1147 eV) peak is not sensitive to the composition of the topmost layer the Si (92 eV) was used to follow the variation of composition in the surface layer. Secondly, because of the high bulk concentration of the segregating specie (Ge) in the alloys, the estimation of the atomic fraction of Ge in the first layer needs a modelisation. Thirdly, long annealing is necessary to restore homogeneity of the alloy after segregation (i.e. to estimate the bulk Ge concentration near the segregated layer).
Section snippets
Materials and techniques
In order to study the composition dependence of segregation, amorphous Si1−xGex thin film alloys with different concentrations of Ge (in the range of 18–58 at.%) were prepared. The general experimental set-up as described in previous work [8] is the following: the samples (thickness of 20 nm) were prepared by dc magnetron sputtering onto a SiO2 substrate (size of mm) at room temperature using a Si target covered by Ge pieces (base pressure in the equipment: 5×10−7 Pa, argon pressure during
AES results
Annealing of the samples are performed at 653 K (Xb(Ge): 0.18, 0.37, 0.55) and 673 K (Xb(Ge): 0.51, 0.58) to avoid crystallisation of the alloys. Typical variations of Si, Ge, and C Auger signals (peak-to-peak height) versus time during annealing are the following:
- 1.
the decrease of the Si (92 eV) peak-to-peak height is important in all experiments while the variation of the Ge (1147 eV) peak-to-peak height is not significant;
- 2.
at the end of the kinetics (for annealing time long enough) constant values
Determination of surface equilibrium data at 653–673 K
To estimate equilibrium data, we assumed the existence of a local equilibrium during the kinetics process [12]. It means that the surface and the bulk (near the surface) concentrations follows the equilibrium isotherm during Ge segregation.
A depletion (due to the local exchange between Si superficial atoms and Ge bulk atoms from the underlying layers) was expected in the bulk layers near the surface at the beginning of the process. Assuming that the annealing time was long enough to restore the
Comparison between experimental and theoretical Ge surface segregation isotherms
The obtained experimental equilibrium data Xs(Ge) versus Xb(Ge) at 653 and 673 K are reported in Table 1. and plotted in Fig. 2. As the Si–Ge system is almost ideal, we have fitted our data obtained for different bulk concentrations using a McLean–Langmuir type isotherm:From Eq. (3), we have determined the experimental segregation coefficient (taking it as a fitting parameter): . Then, supposing Arrhenius-type temperature dependence for the
Conclusion
We have studied by AES the segregation of germanium in series of amorphous Si1−xGex thin film alloys at 653 and 673 K and the main results of this work are the following:
- 1.
The variation of Auger signals during annealing indicates a local exchange between Ge and Si atoms in the uppermost layers: Ge atoms replace the Si surface atoms to minimize the surface free energy as observed in crystalline alloys. For annealing time long enough, equilibrium data have been determined and an experimental
References (13)
- et al.
Surf. Sci.
(1998) - et al.
Surf. Sci.
(1977) - et al.
Surf. Sci.
(1990) - et al.
Surf. Sci.
(2001) - et al.
Surf. Sci.
(1975) - et al.
Appl. Phys. Lett.
(1996)
Cited by (7)
Tin surface segregation, desorption, and island formation during post-growth annealing of strained epitaxial Ge<inf>1-x</inf>Sn<inf>x</inf> layer on Ge(0 0 1) substrate
2014, Applied Surface ScienceCitation Excerpt :The degree of segregation, given by the ratio of Sn surface composition xs to bulk composition x, increases as the annealing temperature increases. Similar to Ge segregation in the Si1−yGey system [31], this thermally activated process is driven by the lower surface energy and larger covalent radius of Sn compared to those of Ge. Preferential segregation of Sn at the surface reduces the strain energy since the Ge0.915Sn0.085 layer is under compressive strain (∼1.2%).
Solid solution strengthening and phase transformation in high-temperature annealed si<inf>80</inf>Ge<inf>20</inf> alloy
2014, Journal of Crystal GrowthCitation Excerpt :On the other hand, the over-annealing treatment (>900 °C) demotes the hardness effect. As the experimental segregation enthalpy of Si–Ge alloy is only 5.3±0.5 kJ mol−1 [21], the precipitation of elemental Ge causes the decayed hardness, after Ge atoms get enough thermal energies to overcome the activation barrier. As displayed in Fig. 3, the broaden Ge peak at 900 °C points to the presence of amorphous Ge, while the raised Ge peak at 1000 °C displays the nanoscaled Ge.
Dissolution kinetics of Si into Ge (111) substrate on the nanoscale
2010, Thin Solid FilmsCitation Excerpt :For the crystalline system we observed that the rate of change of the XPS signals for both species increases (Fig. 1). In case of surface segregation of Ge [21] (e.g. after fast arrival of Ge atoms by short circuit diffusion) or significant broadening of the Si/Ge interface, the Ge atoms would quickly occupy lattice sites on or near the surface. Since the diffusion currents decreases with time, this would lead to a decreasing rate of change in the XPS signal.
Surface roughening in Si<inf>1-x</inf>Ge<inf>x</inf> alloy films by 100 MeV Au: Composition dependency
2006, Solid State CommunicationsCitation Excerpt :The preferential loss of Ge and the consequent enrichment of Si result in spatial variation of concentration of Ge atoms at the surface for neutralizing the excess strain [27]. Prior concentration-gradient-triggered diffusion of adatoms in the vapor deposition process suggests evolution of a structure or mound coarsening [30]. In order to simulate this phenomenon, Mayr et al. [30] have incorporated a nonlinear term in the continuum growth equation, which is mainly concerned with the hindrance of runaway growth of the unstable mode.
Controlling the interface composition of core-shell and axial heterojunction nanowires
2012, Journal of Applied PhysicsAES measurements of Sb mass transport in amorphous Si thin films
2006, Journal De Physique. IV : JP