Electronic Structure of CaSnN$_2$: a sustainable alternative for blue LEDs

The electronic band structure of CaSnN2 in the wurtzite-based Pna21 structure is calculated using the Quasiparticle Self-consistent (QS)GW$^{BSE}$ method, including ladder diagrams in the screened Coulomb interaction W$^{BSE}$ and is found to have a direct gap of 2.680 eV at Γ, which corresponds to blue light wavelength of 463 nm and makes it an attractive candidate for sustainable blue light-emitting diodes (LEDs), avoiding Ga and In. The valence band maximum has a1 symmetry and gives allowed transitions to the conduction band minimum for light polarized along the c-direction. The valence band splitting is analyzed in terms of symmetry labeling, and the effective mass tensor is calculated for several bands at Γ. The optical dielectric function, including electron-hole interaction effects is also reported, and the excitons are analyzed, including several dark excitons.

💡 Research Summary

This paper presents a comprehensive first‑principles investigation of CaSnN₂, a previously unexplored II‑IV‑N₂ semiconductor, as a sustainable alternative to Ga‑ and In‑based blue light‑emitting diodes (LEDs). The authors focus on the wurtzite‑derived orthorhombic Pna2₁ structure (space group 33) reported in the Materials Project, rather than the high‑pressure rock‑salt phase previously synthesized. Using the full‑potential linearized muffin‑tin orbital (FP‑LMTO) implementation in the Quest‑aal suite, they first relax the crystal with the PBEsol functional and then compute the quasiparticle band structure with the quasiparticle self‑consistent GW (QS‑GW) method. Crucially, they go beyond the random‑phase approximation by evaluating the screened Coulomb interaction W with ladder diagrams via the Bethe‑Salpeter Equation (BSE), yielding the QS‑GW BSE (also denoted QS ĜW) approach. This self‑consistent scheme includes off‑diagonal self‑energy elements and electron‑hole interaction effects, providing a highly accurate description of both the quasiparticle energies and optical excitations.

The calculated lattice parameters (a = 6.124 Å, b = 7.719 Å, c = 5.619 Å) reveal a pronounced distortion compared with an ideal wurtzite cell: the a/b ratio (0.793) and c/a ratio (1.521) are significantly smaller than the ideal values, reflecting the large size mismatch between Sn⁴⁺ (0.55 Å) and Ca²⁺ (1.00 Å). This structural anisotropy strongly influences the crystal‑field splitting of the valence band maximum (VBM).

Band‑structure results show a direct gap at the Γ point. Within different levels of theory the gaps are: GGA = 1.352 eV, G₀W₀ = 2.438 eV, QS‑GW RP‑A = 2.776 eV, and QS‑GW BSE = 2.680 eV. The inclusion of electron‑hole ladder diagrams reduces the gap by about 0.1 eV relative to the RP‑A result, a magnitude comparable to zero‑point renormalization and possible lattice‑constant errors. The direct gap of 2.680 eV corresponds to a photon wavelength of 463 nm, placing CaSnN₂ squarely in the blue region of the visible spectrum.

Symmetry analysis at Γ identifies the VBM as an a₁ irreducible representation, while the conduction‑band minimum (CBM) is also a₁ (s‑like). Consequently, optical transitions are allowed only for electric‑field polarization parallel to the c‑axis (E∥c). The next valence band (VBM‑1) is b₁ (x‑like) and couples to the CBM for E∥a, whereas deeper valence states (a₂, b₂) are progressively more split by the crystal field. The authors list the energies and symmetries of the three lowest conduction bands and the six highest valence bands, showing that the two higher conduction states lie within 0.02 eV of each other, a feature that influences effective masses.

Effective‑mass tensors are computed from k·p perturbation theory. The CBM exhibits nearly isotropic, light electron masses (m* ≈ 0.22–0.25 mₑ). In contrast, the valence bands display strong anisotropy and even negative inverse masses, indicating saddle‑point character. For example, VBM‑2 (a₂) has an extremely flat dispersion (inverse mass ≈ ‑48 mₑ), implying a very heavy hole, while VBM‑4 (b₂) shows a negative mass along the y direction, consistent with its symmetry. These mass tensors explain why the VBM has its largest dispersion along Γ‑Z (the c direction), matching the a₁ symmetry of the p_z operator.

Projected density of states (PDOS) analysis shows that states up to ~4 eV below the VBM are dominated by N‑p orbitals from both N sites (N_Ca and N_Sn). Deeper states acquire increasing N‑s character and hybridize with Sn‑s and Ca‑d. The CBM is primarily Sn‑s, while Ca‑d contributions appear around 6 eV above the CBM and as bonding states ~0.5 eV below the VBM.

Optical properties are evaluated both in the independent‑particle approximation (IPA) and with the BSE. The dielectric function ε₂(ω) is anisotropic, with three independent components (ε₂^a, ε₂^b, ε₂^c). The BSE calculation, performed with a fine broadening (η = 0.005 Ry) and including the top 15 valence and bottom 13 conduction bands, reveals distinct excitonic peaks below the quasiparticle gap. The lowest exciton appears for E∥c, consistent with the a₁‑a₁ transition, and has a binding energy of roughly 0.05 Ry (≈ 0.68 eV). A second exciton for E∥a lies slightly higher but still below the gap, while the E∥b exciton is already above the gap, reflecting its origin from a deeper valence band. Several dark excitons (optically forbidden) are also identified, which could act as non‑radiative recombination centers in devices.

Overall, the study demonstrates that CaSnN₂ possesses a direct blue‑gap, light electron effective mass, and strong excitonic effects, making it a promising candidate to replace scarce Ga and In in blue LEDs. However, the authors acknowledge that experimental synthesis of high‑quality single‑crystal thin films, reliable p‑type and n‑type doping, and inclusion of electron‑phonon renormalization remain open challenges. Future work should target growth techniques (e.g., molecular‑beam epitaxy or metal‑organic vapor phase epitaxy), carrier‑concentration control, and device fabrication to validate the theoretical predictions.