Silicon carbide (SiC) is of interest for electronic devices intended to operate in demanding temperature and radiation environments. For engineers assessing material behavior, irradiation-induced defects matter because they can alter the electronic structure of a crystal. A study of the 3C-SiC divacancy known as the Ky5 center uses first-principles calculations to examine one measurable signature of that defect: zero-field splitting in electron paramagnetic resonance (EPR).
The work focuses on a neutral pair of neighboring vacancies—one at a silicon site and one at a carbon site, written (VSi-VC)0. By calculating the defect’s electron density and the interactions that contribute to its spin-Hamiltonian constants, the researchers explain why the observed splitting is substantial and identify which interaction dominates.
Why the Ky5 defect matters in cubic SiC
Neutron irradiation can create defects in SiC crystals. In neutron-irradiated cubic 3C-SiC, EPR measurements have identified several paramagnetic centers, including Ky5. Earlier studies attributed Ky5 to the neutral silicon-carbon divacancy. The center is a triplet, with spin S = 1, and has an axial zero-field-splitting (ZFS) constant reported in two modifications as 443 × 10−4 cm−1 and 454 × 10−4 cm−1.
ZFS describes the separation of spin energy levels in the absence of an applied magnetic field. Its value reflects interactions within the defect and its surrounding crystal. That makes it useful for investigating defect structure, but calculating it for an extended defect is not straightforward: the electron distribution is spread across the vacancy region and neighboring atoms, rather than confined to a single paramagnetic ion.
How the 3C-SiC divacancy was modeled
To capture that spatial distribution, the study used density functional theory and the WIEN computational package. The calculations applied the full-potential linearized augmented plane-wave method and a spin-polarized, relativistic treatment. Electron density and the potential governing it were determined self-consistently, rather than assumed from a simplified point-charge description of the crystal field.
The model placed a silicon vacancy and a carbon vacancy at their nearest-neighbor separation in a supercell representing cubic SiC. The resulting electron density was evaluated in atomic spheres and in the space between atoms. This approach allowed the calculation to account for the defect’s extended electronic state and to examine spin-up and spin-down density separately.
The calculated density forms a localized state with both s-like and p-like character. Its symmetry axis is parallel to the [111] direction, consistent with an axial crystal field acting on the defect. The density-of-states calculation also places a triplet defect state within the band gap, split by the crystal field into a singlet and a doublet.
Dipole-dipole interaction dominates zero-field splitting
The researchers evaluated two contributions to the spin-Hamiltonian ZFS: magnetic dipole-dipole interaction between electron spins and spin-orbit interaction. Their calculations show that the dipole-dipole term provides the main contribution. It gives an axial ZFS constant of 489 × 10−4 cm−1, reasonably close to the reported experimental value of 454 × 10−4 cm−1, although the calculated value is higher.
By contrast, the calculated spin-orbit contribution is negligible—about 1 × 10−4 G in the reported calculation. The study attributes this small contribution to the defect-state spin-orbit constant calculated using the defect’s own electron density and crystal potential. The authors note that using atomic spin-orbit constants alone would not properly represent this extended defect.
This comparison clarifies the physical origin of the Ky5 EPR signature: for this neutral divacancy in 3C-SiC, the spatial distribution of the defect’s electron spin density is central to the ZFS, while spin-orbit coupling contributes very little under the modeled conditions.
Hyperfine fields and the local defect environment
Separating spin-up and spin-down electron densities also makes it possible to calculate hyperfine fields at surrounding nuclei. These fields arise from interaction between the defect’s unpaired electron spin and nuclear spins, and can contribute additional structure to an EPR spectrum.
Most calculated hyperfine constants for neighboring atoms are small. One notable result is the Fermi hyperfine constant for the three nearest carbon atoms, labeled C9 in the calculation: 38.16 G. The study estimates that the corresponding EPR component would have an amplitude of only about 3% of the total signal, making it difficult to observe. This illustrates why a predicted local interaction does not necessarily produce an easily resolved spectral feature.
What the results mean for SiC materials assessment
For materials and device teams, the study offers a way to connect an EPR measurement with the microscopic structure and interactions of an irradiation-induced defect. It also shows why defect-specific calculations matter: a model suited to a point defect may not accurately describe an extended vacancy pair whose electron density occupies the surrounding lattice.
The results do not establish how Ky5 affects device performance or thermal stability. Rather, they provide a theoretical account of its spin signature and a calculated framework for interpreting the defect’s electronic structure. In practical materials assessment, that distinction is important: identifying and understanding a defect is a step toward evaluating its relevance, not a direct measure of its impact on a particular device.
Conclusion
The Ky5 neutral divacancy in cubic SiC produces an axial EPR zero-field splitting that can be understood through self-consistent modeling of its extended electron density. The calculated dipole-dipole contribution is dominant and broadly agrees with experiment, while the spin-orbit term is negligible. For engineers and technical decision-makers working with SiC materials, the study demonstrates how first-principles defect modeling can help interpret radiation-related EPR signals without relying on point-charge assumptions that may not fit an extended defect.