Robust Aluminum Nitride Passivation of Silicon Carbide with Near-Surface Quantum Emitters for Quantum Computing and Sensing Applications

Abstract

Silicon carbide (SiC) hosts a number of point defects that are being explored as single-photon emitters for quantum applications. Unfortunately, these quantum emitters lose their photostability when placed in proximity to the surface of the host semiconductor. In principle, a uniform passivation of the surface's dangling bonds by simple adsorbates, such as hydrogen or mixed hydrogen/hydroxyl groups, should remove detrimental surface effects. However, the usefulness of atomic and molecular passivation schemes is limited by their lack of long-term chemical and/or thermal stability. In this first principles work, we use aluminum nitride (AlN) to passivate SiC surfaces in a core-shell nanowire model. By using a negatively charged silicon vacancy in SiC as the proof-of-principle quantum emitter, we show that AlN-passivation is effective in removing SiC surface states from the band gap and in restoring the defect's optical properties. We also report the existence of a silicon vacancy-based defect at the SiC-AlN interface, which displays distinct spin and optical properties as compared to the other well-studied defects in SiC.

Figures (7)

• Figure 1: Aluminum nitride passivated silicon carbide nanowire (NW). (a) SiC/AlN NW (cross-sectional view), (b) Hydrogen (-H) terminated SiC/AlN NW to mimic a thicker AlN shell (cross-sectional view). Passivation of SiC core with AlN shell creates two distinct defect sites (S1 and S2) relative to the interface between SiC core and AlN shell. (c) Franck-Condon picture used for interpreting $\Delta$SCF results.
• Figure 2: Electronic structure properties of defect-free 2H-SiC NWs. (a) Total and projected density of states (DOS) for unpassivated/bare SiC/AlN NW, showing the narrowing of the bandgap in the core-shell structure to 2.17 eV due to surface and interface states (highlighted in blue). (b) DOS projected onto the threefold and fourfold coordinated Al and N atoms in the AlN shell, showing their contribution to the states at the band edges of the bare core-shell NW. (c) DOS projected onto the Si and C atoms at the interface between the core-shell NW, showing contributions from the interfacial carbons to the DOS at the valence band edge. (d) The surface and interface states are removed by hydrogenation of the core-shell NW. (e) DOS for SiC-H NW for comparison.
• Figure 3: Formation energies ($\Delta E_{form}$) for $\mathrm{V_{Si}}$ in different charged states as a function of the Fermi energy level (electronic chemical potential) in real hydrogen passivated NWs with the monovacancy at the (a) S1 (interior) site within the core-shell SiC/AlN-H NW, (b) S2 (interface) site of the SiC/AlN-H NW, and (c) interior site of the SiC-H NW. In each plot, the Fermi energy level is given with respect to the valence band maximum. The formation energies are calculated assuming carbon-rich conditions. The favorable doping conditions at which the negatively charged state ($\mathrm{q}=-1$) of the defect is stable, are highlighted.
• Figure 4: SiC/AlN-H NW with a $\mathrm{V_{Si}^{-1}}$ defect in the interior (S1 site). (a) Total density of states (TDOS) in thick gray line, along with the DOS contributed by the SiC core (in gold), AlN shell (thin black line) and the four NN carbon-atoms (in blue). (b) Spin density [$\Delta \rho^{spin}$] isosurface plot for the monovacancy at the S1 site, showing that the most of the spin-3/2 of the defect comes from the unpaired electrons in the $sp^{3}$-hybridized $2s$ and $2p$ orbitals of the carbons surrounding the defect. Here, $\Delta \rho^{spin} = \rho^{\uparrow}-\rho^{\downarrow}$ is the difference in the spin-up and spin-down charge densities. (c) Energy-level diagram, showing the positions of the defect states in majority spin (spin-up) and minority spin (spin down) channels at the $\Gamma$-point. The optically-active defect states in the minority spin channel are well-separated from the valence band (VB) and conduction band (CB) states. (d) The charge density plots of the optically-active empty and filled state for SiC/AlN-H NW at the $\Gamma$-point. Yellow (blue) color corresponds to positive (negative) isovalues.
• Figure 5: SiC/AlN-H NW with a $\mathrm{V_{Si}^{-1}}$ defect at the core-shell interface (S2 site). (a) Total density of states (TDOS), along with the DOS contributed by the nearest neighboring (NN) atoms surrounding the defect. The latter include contributions from the $2s$ and $2p$ orbitals of the three $\mathrm{C_{NN}}$ (in blue) and $\mathrm{N_{NN}}$ (in green). (b) Energy-level diagram, showing the positions of the in-gap defect states in majority spin (spin-up) and minority spin (spin down) channels at the $\Gamma$-point. The lowest defect states, which are resonant with the valence band (VB), mix with the VB states and can no longer be uniquely identified. They are included for the sake of completeness. (c) Spin density plot for the monovacancy at the S2 site. Pink (green) color corresponds to positive (negative) isovalues. Also shown is a close-up of the spin density plot (side-view), consisting of the three $\mathrm{C_{NN}}$'s and the $\mathrm{N_{NN}}$-atom, along with the Al atoms bonded to $\mathrm{N_{NN}}$. (d) Spin density isosurface plot for the negatively charged $\mathrm{N_{C}V_{Si}}$-center in the H-SiC NW [at the same site as the defect in (c)]. The close-up (side-view) shows $\mathrm{C_{NN}}$'s and the $\mathrm{N_{NN}}$-atom, along with the Si atoms bonded to $\mathrm{N_{NN}}$.