Growth of Epitaxial ZnSnxGe1−xN2 Alloys by MBE

ZnSnxGe1−xN2 alloys are chemically miscible semiconductor compounds with potential application as earth-abundant alternatives to InxGa1−xN. Preparation of ZnSnxGe1−xN2 thin-films by reactive RF sputter deposition yield low-mobility, nanocrystalline films. In contrast, the growth of ZnSnxGe1−xN2 films by molecular-beam epitaxy (MBE) on c-plane sapphire and GaN templates is described herein. Epitaxial films exhibited 3D growth on sapphire and 2D single-crystal quality on GaN, exhibiting substantial improvements in epitaxy and crystallinity relative to nanocrystalline sputtered films. Films on sapphire were n-type with electronic mobilities as high as 18 cm2 V−1 s−1, an order of magnitude greater than the 2 cm2 V−1 s−1 average mobility observed in this work for sputtered films. Mobility differences potentially arise from strain or surface effects originating from growth techniques, or from differences in film thicknesses. In general, MBE growth has provided desired improvements in electronic mobility, epitaxy, and crystal quality that provide encouragement for the continued study of ZnSnxGe1−xN2 alloys.


Fabrication Details for ZnSn x Ge 1-x N 2 MBE Films
C-plane sapphire or GaN-template-on-sapphire substrates were loaded without pretreatment into an SVT Associates Nitride MBE chamber. Chamber base pressures were < 2×10 -7 torr. Substrates were heated to 250 °C for 2 hrs. before a nitrogen-plasma cleaning. The nitrogen plasma was delivered to each substrate by an HD-50 Oxford Instruments nitrogen-atom source at approximate chamber pressures of 3 x10 -5 Torr to 5x10 -5 Torr and 300W power. While rotating, each substrate was cleaned for ~10 min before beams from separate elemental Zn, Sn, and Ge Knudsen cells, employed at discrete temperatures, were introduced. Table S1 lists the parameters for the MBE growths, including substrate temperatures, Knudsen cell temperatures, and nitrogen plasma parameters.  300  275 845 1215 51  18  10  21  250  3-5x10 -5  300  275 845 1215 53  22  11  14  *  250  3-5x10 -5  300  275 845 1215 52  17  12  19 a) After growth, the Knudsen cell shutter did not close within 30 seconds of the other sources.
After growth, in-situ RHEED patterns were obtained from the film surfaces. The electron beam was held at 9 kV, 2.5 A filament current, 0.5 A emission current, and ~ 6.3 kV objective voltage. Samples were then transferred to atmosphere and sections were cut off to perform X-ray photoelectron spectroscopy (XPS). Table 1 displays the resulting XPS stoichiometries. The error in XPS stoichiometry determination is generally ± atomic 10%, and the technique only probes the first ~10nm of the film. Adventitious carbon and oxygen signals were not included in the reported results. Although some surface oxidation may be present, the precise amount could not be distinguished from adventitious material. Argon-ion sputter cleaning of samples preferentially sputtered nitrogen from the compound semiconductor, leaving it with more metallic character. Other non-destructive characterization methods, such as Energy Dispersive Spectroscopy (EDS) or X-Ray Fluorescence (XRF), on the <50 nm thin films were inconclusive, as they incorporated oxygen or nitrogen originating from the substrates.
Thru-film oxidation was not assumed because studies of ZnSnxGe1-xN2 films that were made by reactive radio-frequency sputtering and probed by energy dispersive spectroscopy (EDS) at 10kV (~200nm penetration depth b) ) displayed compositional stability and consistent electrical mobility measurements after a year. Only thicker MBE films would facilitate the certainty in the extent of oxidation.
Sample cross-sections analyzed by transmission electron spectroscopy (TEM) used an electron-beam energy of 300 keV and a camera length of 250 mm. A polycrystalline gold standard at the same energy and camera length were used for calibrating the distances and calculating the ZnSnxGe1-xN2 lattice parameters. Lattice parameters were calculated from the TEM selected area diffraction. The Zn-IV-Nitride lattices are orthorhombic, thus geometric calculations were performed to obtain the orthorhombic a and c. The b lattice parameters could not be obtained because the films were oriented.
Hall measurements of ZnSnxGe1-xN2 films on sapphire and GaN were performed using ohmic In contacts. b) http://www.globalsino.com/EM/page4795.html Epitaxial 3D island growth of ZnSnxGe1-xN2 on sapphire was observed. [Reference 11] The epitaxial nature was seen from the Pendellosung oscillations of the (002) ZnSnxGe1-xN2 XRD peak. SI Figure 1b shows the ZnSnxGe1-xN2 RHEED image as spots with similar symmetry as the simulated hexagonal GaN pattern for transmission diffraction. Matching symmetry of diffraction patterns to hexagonal GaN is expected for wurtzite ZnSnxGe1-xN2 3D islands because the electron beam transmits through the peak of the crystal island to create the diffraction pattern. [Reference 11]

Band Gap Tunability
Ellipsometry was performed with a JA Woollam Alpha-SE ellipsometer and the data were modeled with Complete-Ease Software using a wavelength-by-wavelength layer on top of an alumina substrate layer. Using the extinction coefficients, absorption coefficients were calculated to be on the order of 10 5 cm -1 , highlighting good absorption properties, similar to known thin-film absorbers like CdTe. Extinction coefficients were transformed to absorption coefficients with Equation 2, where α is the absorption coefficient, k is the extinction coefficient, and λ is the probing wavelength. The data were then fitted using the Tauc method for direct band gaps. The x-intercepts for squared absorption coefficients were collected and plotted by composition.