Electronic structure of Cu₃N films studied by soft x-ray spectroscopy

Abstract

Soft x-ray emission spectroscopy was used to characterize the electronic structure of seven copper nitride films, one synthesized with atomic layer deposition (ALD) and six grown with chemical vapor deposition (CVD) at different preparation temperatures. Interpretation of the x-ray emission spectra was supported by calculations of the electronic structure for bulk pure Cu(3)N and Cu(3)N with: an excess of Cu atoms, oxygen or carbon impurities, and N vacancies. The calculations are shown to describe the experimental spectra quite well. Analysis of the x-ray spectra suggests that films grown in copper rich environments and above a cut-off temperature of approximately 360 °C have a growing fraction of copper enriched areas, while films prepared below this temperature do not have these areas with excess copper.

Observation of the Direct Energy Band Gaps of Defect-Tolerant Cu3N by Ultrafast Pump-Probe Spectroscopy

“…However, we did not observe any PL from the Cu 3 N between 77 and 300 K, and quantization would require confinement by a barrier. It has been shown by Modin et al 91 that Cu 3 N containing an excess of Cu atoms, O 2 impurities, and N vacancies has a tendency to form Cu-rich areas that will surround Cu 3 N. Considering that the work function of Cu 3 N is ∼0.9 eV larger than that of Cu 92 and that Cu 3 N has a work function of 5.04 eV, 93 We ought to mention in closing that the maxima observed by UPPS match very well with the energies extracted from an analysis of the steady-state UV−vis spectra shown in Figure 2. However, UPPS can be used as an effective tool for understanding even more about Cu 3 N at a fundamental level, for example, finding how the direct energy band gap changes with strain, reduced dimensionality and size, and lifetimes, which in turn will provide deeper insight in conjunction with theoretical calculations as to how this novel material may be used for the realization of solar cells.…”

“…However, we did not observe any PL from the Cu 3 N between 77 and 300 K, and quantization would require confinement by a barrier. It has been shown by Modin et al that Cu 3 N containing an excess of Cu atoms, O 2 impurities, and N vacancies has a tendency to form Cu-rich areas that will surround Cu 3 N. Considering that the work function of Cu 3 N is ∼0.9 eV larger than that of Cu and that Cu 3 N has a work function of 5.04 eV, we expect that the Cu/Cu 3 N metal–semiconductor interface will not lead to quantum confinement in Cu 3 N as shown Figure a. Similarly, the oxidation of Cu 3 N and the formation of Cu 3 N/Cu 2 O and Cu 3 N/CuO p–n heterojunctions will have small barriers and will not result in strong confinement as shown in Figure b.…”

Impact of Oxygen on the Properties of Cu3N and Cu3–xN1–xOx

“…Evidently, the spectral features observed by UPPS in the case of Cu, Cu 2 O, and CuO shown in Figure a–c, respectively, do not coincide with the duo of higher energy maxima at 2.46 and 2.3 eV in Figure a,b, which are directly related to the strained Cu 3 N, which has larger direct and indirect energy band gaps compared to bulk-relaxed Cu 3 N. Before elaborating further, we ought to mention that the occurrence of strained Cu 3 N at the surface of Cu 3 N layers is consistent with the findings of Majumdar et al who observed that surface oxidation gave rise to strained Cu 3 N. This is consistent with the fact that CuO has a monoclinic crystal structure and a lattice constant a = 4.6837 Å, which is larger than the lattice constant of bulk-relaxed cubic Cu 3 N, i.e., a = 3.8 Å, and the CuO/Cu 3 N heterojunction will be strained; similarly, Cu 2 O has a cubic crystal structure and a lattice constant of 4.27 Å, so the Cu 2 O/Cu 3 N heterojunction is also expected to be strained. In contrast, Cu and Cu 3 N have lattice constants of 3.6 and 3.8 Å, respectively, so the Cu/Cu 3 N metal–semiconductor junction is not expected to be strained to the same extent . Nevertheless, for such large lattice mismatches, the critical layer thickness is expected to be low and the Cu 3 N will be relaxed rather than strained when its thickness exceeds a few tens of nanometers.…”

Copper interstitial recombination centers in Cu3N

“…There have been several prior density functional theory studies of Cu 3 N in which the indirect bandgap values of Cu 3 N were significantly underestimated as 0.5 eV [7] and 0.23 eV [8], using the generalized-gradient approximation (GGA) and local-density approximation (LDA), respectively. A reasonable indirect bandgap value of 1.0 eV was obtained by Zakutayev [3] when combining the GGA functional with an additional onsite Coulomb interaction term U d (Cu) = 5 eV and GW quasiparticle energy calculations.…”