EPR of a defect in CVD diamond involving both silicon and hydrogen that shows preferential alignment

A defect involving both silicon and hydrogen has been characterized using multifrequency electron paramagnetic resonance. The defect, denoted WAR3, was observed in a single-crystal chemical-vapor deposition diamond, which was homoepitaxially grown on a {110}-oriented substrate and doped with isotopically enriched silicon (90% ${}^{29}\text{S}\text{i}$). The obtained data are explained by a silicon divacancy structure which is decorated by a hydrogen atom and is in the neutral charge state, ${\left({\text{Si-V}}_2:\text{H}\right)}^0$ $\left(S=\frac{1}{2}\right)$. The experimentally derived ${}^{29}\text{S}\text{i}$ and ${}^1\text{H}$ hyperfine parameters are in agreement with values calculated using the spin-density-functional technique, ruling out a nonplanar structure. Defects are usually randomly oriented such that there is an equal probability for the symmetry axis of the defect to lie along each of the crystallographically equivalent directions. However, the WAR3 defect shows preferential alignment with respect to the {110} growth plane of the sample. Approximately four times as many WAR3 centers were aligned with their mirror planes lying perpendicular to the growth plane, compared to a statistical distribution. This indicates that the majority of WAR3 defects grew in as units rather than by the diffusion and aggregation of constituents. Analysis of the increase in the WAR3 concentration and the decrease in preferential alignment upon annealing the sample at $1400°\text{C}$ shows that WAR3 can also be created post growth.

Figure 1

$X$-band $\left(\upsilon \sim 9.75\text{GHz}\right)$ WAR3 EPR spectrum recorded at room temperature with $\underset{̱}{B}\parallel \left[111\right]$ for sample A after annealing at $1200°\text{C}$. The ${}^{29}\text{S}\text{i}$ hyperfine splitting is apparent, where the group of lines arising from ${}^{28}\text{S}\text{i}$ and ${}^{30}\text{S}\text{i}$ lies between the ${}^{29}\text{S}\text{i}$ (90% relative abundance) satellite groups. Each group consists of three resonance lines, which are further split into two due to a hyperfine interaction with a ${}^1\text{H}$ atom.

Figure 2

Roadmaps showing the positions of the ${}^{29}\text{S}\text{i}$ hyperfine EPR lines for WAR3, as a function of angle in a {110} plane at (a) $X$-band and (b) $Q$-band frequencies. The solid curves show the best fits to the experimental data and were obtained using the determined Hamiltonian parameters summarized in Table .

Figure 3

NIR absorption spectrum obtained at 77 K for sample A after annealing $2000°\text{C}$ and heat treating for 20 min at $577°\text{C}$ in the dark. Heating under these conditions leads to an increase in the integrated intensity of the 737 nm [1683.5(1) and 1684.3(1) meV] optical band and a decrease of that of the 946 nm [1311.3(4) meV] peak. Conversely, illuminating with a 224 nm laser excitation for $\sim 2\text{min}$ results in the opposite effect. The inset figures show the change in the absorption coefficient, where $\Delta \mu$ is the difference in the absorption coefficient between the cases, where the sample was heated or illuminated. The insets were recorded at different resolutions.

Figure 4

(Color online) Schematics of (a) planar and (b) nonplanar $\left({\text{Si-V}}_2:\text{H}\right)$ complexes showing the ${}^{13}\text{C}$ sites for which hyperfine terms are quoted in Tables . The arrows showing the directions of the theoretical hyperfine tensor elements listed.

Figure 5

WAR3 EPR lines observed around $g=2$ at $X$-band frequencies for sample A with (a) $\underset{̱}{B}\parallel \left[\overline{1}1\overline{1}\right]$ and (b) $\underset{̱}{B}\parallel \left[111\right]$, where a (110) growth plane is assumed. Consequently, the former direction lies in the growth plane while the latter lies out of the growth plane. The difference in the relative intensities of the peaks is attributed to the preferential alignment of the defect. (c) displays the simulated spectrum for WAR3 if it was statistically oriented. The spectra were collected after annealing the sample for 1 h at $1200°\text{C}$.

Figure 6

(Color online) Cross sectional view of the $\left(1\overline{1}0\right)$ mirror plane of (a) (Si-V:H) and (b) WAR3. Vacancies are illustrated as squares. The carbon atoms above and below the $\left(1\overline{1}0\right)$ plane are shown as larger and smaller circles, respectively. The numbered carbon atoms for the WAR3 defect refer to those given in Table and Fig. 4a.