Electron paramagnetic resonance studies of silicon-related defects in diamond

We report the results of multifrequency electron paramagnetic resonance studies at temperatures between 8 and 300 K on diamonds synthesized by chemical vapor deposition and intentionally silicon doped with isotopes in natural abundance or isotopically enriched. The ${}^{29}\text{S}\text{i}$ hyperfine structure has provided definitive evidence for the involvement of silicon in two electron paramagnetic resonance centers in diamond that were previously suspected to involve silicon: KUL1 and KUL3. We present data that unambiguously identify KUL1 as an $S=1$ neutral silicon split-vacancy ($D_{3d}$ symmetry) defect ${\left(V\text{-Si-}V\right)}^0$, while KUL3 is shown to be ${\left(V\text{-Si-}V\right)}^0$ decorated with a hydrogen atom, ${\left(V\text{-Si-}V:\text{H}\right)}^0$.

Figure 1

The structure of the silicon split-vacancy center $\left(V\text{-Si-}V\right)$. The solid black circles (labeled $i\text{−}n$) represent carbon atoms, the lighter circle indicate the relaxed silicon site, and the white circles indicate the vacancies.

Figure 2

The points indicate the determined positions, as a function of angle in the $\left(1\overline{1}0\right)$ plane of the KUL1 EPR lines at $X$-band frequencies. The solid curves show the best fit to the experimental data and were used to determine the parameters $g$ and $D$ (see text).

Figure 3

A comparison between the KUL1 hyperfine structures observed around line i (see Fig. 2) (a) for a sample with ${}^{29}\text{S}\text{i}$ in its natural abundance and (b) for one enriched to 90% ${}^{29}\text{S}\text{i}$. The spectra were recorded at $X$-band frequencies with $\phantom{\Vert }\underset{̱}{B}\Vert ⟨111⟩$.

Figure 4

KUL1 hyperfine structure observed at $X$-band frequencies with $\theta =80°$ in the $\left(1\overline{1}0\right)$ plane. The plots beneath illustrate the simulated data obtained using the determined spin-Hamiltonian parameters and Eqs. (1, 2). The lowercase labels denote lines that are absent in $Q$-band spectra. (a) Displays the structure around resonance line vi and (b) shows the hyperfine satellites seen around line iv (see Fig. 2).

Figure 5

(a) Splitting of the ${}^{29}\text{S}\text{i}$ hyperfine satellites around KUL1 resonance lines i $(\square )$, ii $(△)$, and iii $(○)$ (see Fig. 2) as a function of angle in the $\left(1\overline{1}0\right)$ plane. (b) Splitting of the ${}^{13}\text{C}$ satellites as a function of angle in the $\left(1\overline{1}0\right)$ plane. The three data sets ($○$, $△$, and $\square$) illustrate the splitting of satellites ${{}^{13}\text{C}}_{\text{A}}/{{}^{13}\text{C}}_{\text{C}}$, ${{}^{13}\text{C}}_{\text{B}}/{{}^{13}\text{C}}_{\text{E}}$, and ${{}^{13}\text{C}}_{\text{D}}$, respectively (see Fig. 4). The error bars are comparable to the size of the data points. The plotted lines were created using the spin-Hamiltonian parameters and Eqs. (1, 2).

Figure 6

The determined positions as a function of angle in the $\left(1\overline{1}0\right)$ plane of the KUL3 EPR lines observed at (a) $X$-band and (b) $Q$-band frequencies. The plotted lines were obtained using the determined parameters for $g$ and $A$ (see Table ).

Figure 7

EPR lines observed around $g=2$ at $X$-band frequencies with $\underset{̱}{B}$ along $⟨111⟩$ for (a) sample A (${}^{29}\text{S}\text{i}$ in natural abundance) and (b) sample B (enriched with 90% ${}^{29}\text{S}\text{i}$). The central resonance line in (b) arises due to the presence of ${\left({\text{N}}_s\right)}^0$ in this sample. (c) Shows a spectrum simulated using the determined parameters for KUL3, given in Table .

Figure 8

The magnitude of the zero-field $\left(D\right)$ interaction for KUL1 as a function of temperature. Note that $1\text{MHz}\sim 0.036\text{mT}$ for $g=2.0023$.