Unexplored MBE growth mode reveals new properties of superconducting NbN

Accessing unexplored crystal growth conditions often reveals new regimes of physical behavior. In this work, performing molecular beam epitaxy growth of the technologically important superconductor NbN at temperatures greater than 1000 ${}^{\circ }\mathrm{C}$ reveals a growth mode that has not been accessed before. This mode results in persistent reflection high-energy electron diffraction (RHEED) oscillations through the entire growth, resulting in atomically smooth surfaces, normal metal resistivities of $\sim 37\mu \mathrm{\Omega }$ cm. We find that the superconducting critical temperature depends strongly on growth temperature, and report a maximum superconducting critical temperature of 15.5 K. Electron microscopy studies reveal a rich range of crystalline phases that depend on the growth temperature and correlate to the physical properties. Surprisingly, a reversal of the sign of the Hall coefficient from n-type to p-type is observed as the NbN films are cooled, indicating an electronic structure that has not been observed before in this material. In addition to this observation, the crystallinity of the high-temperature epitaxial NbN allows for an ordered Abrikosov vortex lattice to be imaged for the first time in this superconductor.

Figure 1

(a) In situ RHEED specular spot peak intensity as a function of time throughout the first 500s of growth of NbN films on 6H-SiC substrate for different substrate temperatures. (b) Post-growth AFM surface height maps for NbN films grown at different substrate temperatures. The scale bars correspond to 400 nm. (c) In situ post-growth RHEED patterns.

Figure 2

(a)-(d) HAADF-STEM image of an NbN film grown at $1000{}^{\circ }\mathrm{C}$ shows a transition from cubic $\delta$-NbN to hexagonal $\varepsilon$-NbN, a few unit cells away from the substrate. Scale bars in (b)-(d) correspond to 1nm. (e)-(f) HAADF-STEM image of a NbN film grown at $1200{}^{\circ }\mathrm{C}$ reveals a structure consistent with both the $\delta$-NbN phase and the closely related $\gamma \text{−}{\mathrm{Nb}}_4{\mathrm{N}}_3$. Near the 6H-SiC/NbN interface are several layers of NbN that have a lattice orientation rotated ${60}^{\circ }$ about the growth axis relative to the orientation that is dominant in this image.

Figure 3

(a) Symmetric $2\theta /\omega$ XRD scans of 30 nm NbN thin films grown on 6H-SiC substrate. (b)-(d) XRD reciprocal space maps of NbN films grown at (b) $800{}^{\circ }\mathrm{C}$, (c) $1000{}^{\circ }\mathrm{C}$, and (d) $1200{}^{\circ }\mathrm{C}$. The region mapped is centered on the 6H-SiC (1 0 16) RLP. The expected epitaxial relationship of $\delta$-NbN on 6H-SiC places the $\delta$-NbN (4 2 2) RLP near the 6H-SiC (1 0 16) RLP.

Figure 4

(a) The EBSD map of a 100 nm NbN film reveals that the cubic twin grain boundaries form parallel stripes across the surface with width of approximately 550 nm. (b) AFM surface mapping of the same NbN film shows a series of parallel depressions whose separation and orientation correspond with the grain boundaries observed in the EBSD map. (c) A region of the 6H-SiC substrate that was shadowed during growth shows atomic terraces whose separation and orientation do not accurately match the grain boundaries orientation and separation. (d) AFM mapping of a NbN sample grown at $1200{}^{\circ }\mathrm{C}$ at the site of a cubic twin grain boundary shows there is a depression at the site of the boundary approximately 1 nm deep.

Figure 5

(a) Normal state resistance versus temperature of NbN thin films. (b) Resistance versus temperature in the range of the transition to the zero-resistance state. (c) Hall resistance versus temperature. (d) Hall voltage versus field for a single NbN film grown at $1000{}^{\circ }\mathrm{C}$.

Figure 6

(a)-(e) Scanning tunneling microscope differential conductance measurements at 4.7 K of a NbN film grown at ${\mathrm{T}}_{\mathrm{s}}=1200{}^{\circ }\mathrm{C}$: (a) differential conductance versus voltage is well fit using the BCS density of states yielding the energy gap of the NbN film. (b) Surface topography map and (c) differential conductance waterfall plot shows a highly uniform superconducting energy gap across the sample surface. (d) Differential conductance map and (e) surface topography map in the presence of a 1 T magnetic field shows an Abrikosov vortex lattice that is distorted in the vicinity of a cubic twinning grain boundary. Sample bias was 0.1 V and constant current was 10 pA for both (b) and (e), and modulation voltage was 0.5 mV peak-to-peak for (a), (c), and (d).