Agnieszka Paszuk

Controlling the polarity of metalorganic vapor phase epitaxy-grown GaP on Si(111) for subsequent III-V nanowire growth

Nanowire growth on heteroepitaxial GaP/Si(111) by metalorganic vapor phase epitaxy requires the [-1-1-1] face, i.e., GaP(111) material with B-type polarity. Low-energy electron diffraction (LEED) allows us to identify the polarity of GaP grown on Si(111), since (2×2) and (1×1) surface reconstructions are associated with GaP(111)A and GaP(111)B, respectively. In dependence on the pre-growth treatment of the Si(111) substrates, we were able to control the polarity of the GaP buffers. GaP films grown on the H-terminated Si(111) surface exhibited A-type polarity, while GaP grown on Si surfaces terminated with arsenic exhibited a (1×1) LEED pattern, indicating B-type polarity. We obtained vertical GaAs nanowire growth on heteroepitaxial GaP with (1×1) surface reconstruction only, in agreement with growth experiments on homoepitaxially grown GaP(111).

In situ controlled heteroepitaxy of single-domain GaP on As-modified Si(100)

Metalorganic vapor phase epitaxy of III-V compounds commonly involves arsenic. We study the formation of atomically well-ordered, As-modified Si(100) surfaces and subsequent growth of GaP/Si(100) quasisubstrates in situ with reflection anisotropy spectroscopy. Surface symmetry and chemical composition are measured by low energy electron diffraction and X-ray photoelectron spectroscopy, respectively. A two-step annealing procedure of initially monohydride-terminated, (1 × 2) reconstructed Si(100) in As leads to a predominantly (1 × 2) reconstructed surface. GaP nucleation succeeds analogously to As-free systems and epilayers free of antiphase disorder may be grown subsequently. The GaP sublattice orientation, however, is inverted with respect to GaP growth on monohydride-terminated Si(100).

Lattice-engineered Si1-xGex-buffer on Si(001) for GaP integration

XRD techniques determined that 270 nm GaP grown on 400 nm Si0.85Ge0.15/Si(001) substrates by MOCVD is single crystalline and pseudomorphic, but carry a 0.07% tensile strain after cooling down to room temperature due to the bigger thermal expansion coefficient of GaP with respect to Si (Fig. 2). TEM and AFM examinations indicated a closed but defective GaP layer (Fig. 3(a)) with low root mean square of roughness (rms) of 3.0 nm for 1 μm2 surface area (Fig. 3(b)). Although TEM studies confirm the absence of misfit dislocations in the pseudomorphic GaP film, growth defects (e.g. stacking faults, microtwins, and anti-phase domains) are detected, concentrating at the GaP/SiGe interface (Fig. 3(c)-(d), Fig. 4). We interpret these growth defects as a residue of the initial 3D island coalescence phase of the GaP film on the Si0.85Ge0.15 buffer. TEM-EDX studies reveal that the observed growth defects are often correlated with stoichiometric inhomogeneities in the GaP film (not shown here). Finally, ToF-SIMS detects sharp heterointerfaces between GaP and SiGe films with a minor level of Ga diffusion into the SiGe buffer (Fig. 5).

Lattice-engineered Si 1−x Ge x -buffer on Si(001) for GaP integration

XRD techniques determined that 270 nm GaP grown on 400 nm Si 0.85 Ge 0.15 /Si(001) substrates by MOCVD is single crystalline and pseudomorphic, but carry a 0.07% tensile strain after cooling down to room temperature due to the bigger thermal expansion coefficient of GaP with respect to Si (Fig. 2). TEM and AFM examinations indicated a closed but defective GaP layer (Fig. 3(a)) with low root mean square of roughness (rms) of 3.0 nm for 1 μm 2 surface area (Fig. 3(b)). Although TEM studies confirm the absence of misfit dislocations in the pseudomorphic GaP film, growth defects (e.g. stacking faults, microtwins, and anti-phase domains) are detected, concentrating at the GaP/SiGe interface (Fig. 3(c)-(d), Fig. 4). We interpret these growth defects as a residue of the initial 3D island coalescence phase of the GaP film on the Si 0.85 Ge 0.15 buffer. TEM-EDX studies reveal that the observed growth defects are often correlated with stoichiometric inhomogeneities in the GaP film (not shown here). Finally, ToF-SIMS detects sharp heterointerfaces between GaP and SiGe films with a minor level of Ga diffusion into the SiGe buffer (Fig. 5).

MOVPE-preparation of Si(111) surfaces for III–V nanowire growth

We studied the preparation of the clean Si(111) surface in H2 ambient with in situ reflection anisotropy spectroscopy and UHV-based surface science tools after contamination-free transfer. X-ray photoelectron spectroscopy confirmed complete oxide removal after high-temperature annealing. In situ RAS enabled observation of the oxide removal in dependence of process temperature. Monohydride termination was verified by Fourier transform infrared spectroscopy which agrees with a (1×1) surface reconstruction we observed by scanning tunneling microscopy and low energy electron diffraction. By atomic force microscopy analysis of the morphology, we found that wet-chemical pretreatment has an impact on the different silicon surfaces we have prepared, including homoepitaxy and termination of silicon with arsenic.