Thermodynamics and mechanism of metal-induced crystallization in immiscible alloy systems: Experiments and calculations on Al/a-Ge and Al/a-Si bilayers

The mechanism of metal-induced crystallization (MIC) in immiscible alloy systems has been explained on a unified thermodynamic basis. Interface thermodynamics has been shown to play a decisive role for the occurrence of MIC. The thermodynamic predictions agree excellently with the corresponding experimental observations obtained in this project for the $\mathrm{Al}∕\mathrm{Ge}$ and $\mathrm{Al}∕\mathrm{Si}$ layer systems, which show two distinctly different types of MIC behaviors. As a result, a model has been developed that rationalizes and predicts MIC processes in immiscible alloy systems.

Figure 1

(Color online) AES concentration-depth profiles of the (a) Al/a-Ge and (c) $\mathrm{a}\text{−}\mathrm{Ge}∕\mathrm{Al}$ specimens upon annealing for temperatures and times as indicated. Corresponding XRD patterns are shown in (b) and (d) for Al/a-Ge and $\mathrm{a}\text{−}\mathrm{Ge}∕\mathrm{Al}$, respectively.

Figure 2

(Color online) AES concentration-depth profiles of the (a) Al/a-Si and (c) $\mathrm{a}\text{−}\mathrm{Si}∕\mathrm{Al}$ specimens upon annealing for temperatures and times as indicated. Corresponding XRD patterns are shown in (b) and (d) for Al/a-Si and $\mathrm{a}\text{−}\mathrm{Si}∕\mathrm{Al}$, respectively.

Figure 3

SEM micrographs (recorded in the Auger microscope) of the surfaces of the (a) Al/a-Ge, (c) $\mathrm{a}\text{−}\mathrm{Ge}∕\mathrm{Al}$, and (e) $\mathrm{a}\text{−}\mathrm{Si}∕\mathrm{Al}$ specimens annealed for temperatures and times as indicated. The corresponding AES analyses (derivative spectra) of small areas (size: $20\mathrm{nm}$) indicated in the micrographs are presented in (b), (d), and (f), respectively.

Figure 4

[(a) and (b)] The evolution of the Ge(111) diffraction peak upon in situ annealing of the (a) Al/a-Ge and (b) $\mathrm{a}\text{−}\mathrm{Ge}∕\mathrm{Al}$ specimens. An explosivelike, full crystallization of the a-Ge layer is observed at $150°\mathrm{C}$ for both specimens. [(c) and (d)] The evolution of the Si(111) diffraction peak upon in situ annealing of the (c) Al/a-Si and (d) $\mathrm{a}\text{−}\mathrm{Si}∕\mathrm{Al}$ specimens. For both specimens, a gradual crystallization of a-Si is observed with increasing temperature and annealing time.

Figure 5

(Color online) Calculated (a) crystallization energies, (b) surface energies, and (c) interface energies related to the $\mathrm{Al}∕\mathrm{Si}$ and $\mathrm{Al}∕\mathrm{Ge}$ layer systems. The calculation methods have been discussed in detail in the text.

Figure 6

(Color online) (a) Energetics of the Al grain boundary wetting by a-Ge and a-Si. A positive driving force ${\gamma }_{\mathrm{Al}}^{GB}-2{\gamma }_{⟨\mathrm{Al}⟩∕\left\{\mathrm{Ge}\text{or}\mathrm{Si}\right\}}^{\mathit{interf}}$ is evident for both Ge and Si to diffuse into the Al grain boundaries (wetting). (b) Energetics of the nucleation of crystalline Ge (and Si) at the Al GBs and at the Al/a-Ge and Al/a-Si interfaces. Note that the thicknesses of the free Ge or Si layers are about 2 ML at the interfaces with Al and $\sim 4$ ML at the Al GBs (both these thicknesses are shown by gray, horizontal lines in the figure). It follows that c-Ge can nucleate both at the Al GBs and Al/a-Ge interfaces (above $50°\mathrm{C}$), whereas c-Si can nucleate only at the Al GBs (above $140°\mathrm{C}$).

Figure 7

(Color online) (a) Energetics of the continued diffusion of Ge and Si atoms into the Al layer after completing the initial nucleation of the crystallization process at the Al GBs. A positive driving force is present for both Ge and Si for continued diffusion into the $⟨\mathrm{Al}⟩∕⟨\mathrm{Ge}⟩$ and $⟨\mathrm{Al}⟩∕⟨\mathrm{Si}⟩$ boundaries, respectively. (b) Energetics for continued lateral grain growth of c-Ge and c-Si in the original Al layer (perpendicular to the original Al GBs). Continued grain growth, rather than repeated nucleation of crystallization, is favored for both Ge and Si. The formation of new crystalline nuclei is impossible.

Figure 8

(Color online) (a) Energetics of Ag and Au grain boundary wetting by a-Si. It follows that a-Si can wet both Ag GBs and Au GBs. (b) Energetics of the nucleation of crystalline Si at Ag and Au GBs and at Ag/a-Si and Au/a-Si interfaces. It follows that c-Si can nucleate at the Ag GBs at (above) $\sim 400°\mathrm{C}$. For the Au/a-Si system, it follows that c-Si cannot nucleate directly at Au GBs and at the Au/a-Si interface. MIC in Au/a-Si is, however, mediated by the formation of metastable ${\mathrm{Au}}_3\mathrm{Si}$ nucleated at Au GBs (see text).