Decoration of growth sector boundaries with nitrogen vacancy centers in as-grown single crystal high-pressure high-temperature synthetic diamond

Large $(>100\mathrm{m}{\mathrm{m}}^3)$, relatively pure (type II) and low birefringence single crystal diamond can be produced by high pressure high temperature (HPHT) synthesis. In this study we examine a HPHT sample of good crystalline perfection, containing less than 1 ppb (part per billion carbon atoms) of boron impurity atoms in the (001) growth sector and only tens of ppb of nitrogen impurity atoms. It is shown that the boundaries between $\left\{111\right\}$ and $\left\{113\right\}$ growth sectors are decorated by negatively charged nitrogen vacancy centers $\left(\mathrm{N}{\mathrm{V}}^{-}\right)$: no decoration is observed at any other type of growth sector interface. This decoration can be used to calculated the relative $\left\{111\right\}$ and $\left\{113\right\}$ growth rates. The bulk (001) sector contains concentrations of luminescent point defects (excited with 488- and 532-nm wavelengths) below ${10}^{11}\mathrm{c}{\mathrm{m}}^{-3}$ $\left({10}^{-3}\mathrm{ppb}\right)$. We observe the negatively charged silicon-vacancy $\left(\mathrm{Si}{\mathrm{V}}^{-}\right)$ defect in the bulk $\left\{111\right\}$ sectors along with a zero phonon line emission associated with a nickel defect at 884 nm (1.40 eV). No preferential orientation is seen for either $\mathrm{N}{\mathrm{V}}^{-}$ or $\mathrm{Si}{\mathrm{V}}^{-}$ defects, but the nickel related defect is oriented with its trigonal axis along the $\langle 111\rangle$ sector growth direction. Since the $\mathrm{N}{\mathrm{V}}^{-}$ defect is expected to readily reorientate at HPHT diamond growth temperatures, no preferential orientation is expected for this defect but the lack of preferential orientation of $\mathrm{Si}{\mathrm{V}}^{-}$ in $\left\{111\right\}$ sectors is not explained.

Figure 1

(a) Schematic drawing of a perfect cuboctahedral crystal where specific growth sectors have been labeled. (b) Shows a slice taken parallel to the (001) growth sector with specific growth sectors indicated.

Figure 2

(a) Cross polarization image of the NDT HPHT sample (polarizers vertical and horizontal) showing a small fracture in the crystal (A) and an inclusion (B). (b) White beam x-ray topograph (400 reflection) with the fracture and inclusion highlighted. (Scale bar: 1 mm).

Figure 3

(a) Low temperature (77 K) cathodoluminescence image (visible light emitted) of the NDT HPHT diamond sample studied. (Scale bar: 200 µm). (b) Low temperature (77 K) cathodoluminescence spectra recorded in the (001) and a $\left\{111\right\}$ growth sector. The free and bound exciton images are shown in the insets. The higher order replica exciton peaks (higher order monochromator artefacts) are indicated with the * symbol.

Figure 4

(a) CL image reported in Fig. 3 but now labeled with four regions (A, B, C, and D) where room temperature confocal PL images have been recorded. (b) A small region of the CL image, as indicated, with the different growth sectors labeled. (c) Room temperature confocal PL image recorded with a 640-nm long pass filter (to pass the emission from the $\mathrm{N}{\mathrm{V}}^{\text{–}}$ defects) in the same region as (b). The $\left\{111\right\}$/$\left\{113\right\}$ growth sector boundaries are decorated with $\mathrm{N}{\mathrm{V}}^{-}$ defects (identity of the $\mathrm{N}{\mathrm{V}}^{-}$ defects confirmed by ODMR and localized luminescence spectroscopy measurements, see text for details). (d), (e), and (f) show room temperature confocal PL images recorded with a 640-nm long pass filter for the locations labeled B, C, and D in (a). It is clear from (d), (e), and (f) that only the $\left\{111\right\}$/$\left\{113\right\}$ growth sector boundaries are decorated with $\mathrm{N}{\mathrm{V}}^{-}$ defects. (Scale bar; 10 µm).

Figure 5

(a) 80 µm × 80 µm confocal image at a depth of 7 µm from the surface of the sample at location A (Fig. 4). Two locations are marked for the XZ slices. (b) and (c) are the resultant planar slices for XZ’ and XZ’’ respectively. The angles labeled are between the vertical ([001] direction) and the line shown: ${\theta }_1={55}^{\circ }\pm 5^{\circ }$, ${\theta }_2={30}^{\circ }\pm 5^{\circ }$, ${\theta }_3=0^{\circ }\pm 5^{\circ }$, and ${\theta }_4={19}^{\circ }\pm {10}^{\circ }$. (Scale bar; 10 µm).

Figure 6

Employing optical filters matched to the emission range of each defect, the spatial distribution of each defect at the $\left\{113\right\}$ to $\left\{111\right\}$ sector interfaced can be mapped. (a) $\mathrm{N}{\mathrm{V}}^0$, (b) $\mathrm{N}{\mathrm{V}}^{\text{–}}$ (c) $\mathrm{Si}{\mathrm{V}}^{\text{–}}$ (with ∼10% $\mathrm{N}{\mathrm{V}}^{\text{–}}$), and (d) 1.40-eV defect. The transmission profile of each filter is given in (e). For the 1.40-eV defect shown in (d), the 532-nm excitation power was increased to 100 mW, c.f. 5 mW for the other defect distribution maps in this figure. (Scale bar; 5 µm).

Figure 7

(a) Optically detected magnetic resonance (ODMR) study at location A (Fig. 4) (Scale bar; 10 µm). (b) Classification histogram of $\mathrm{N}{\mathrm{V}}^{\text{–}}$ centers based on their photon autocorrelation minima with a representative measurement for a single center (inset). (c) The four possible trigonal orientations along $\langle 111\rangle$ for the $\mathrm{N}{\mathrm{V}}^{\text{–}}$ center with the measurement occurrence from ODMR in this study. (d) Single center ODMR spectra showing the measured fluorescence intensity (data points) as a function of microwave frequency for the differently oriented $\mathrm{N}{\mathrm{V}}^{\text{–}}$ defects in a magnetic field applied such that the vector specifying its orientation makes a different angle with each of the four symmetry axes of the individual defects. The solid curves show fits to the spin Hamiltonian for the differently orientated defects.

Figure 8

The effect of laser polarization (vertical or horizontal) on both $\mathrm{Si}{\mathrm{V}}^{-}$, (a) and (b), respectively, and the 1.40-eV defect, (d) and (e), respectively. By taking the difference in intensity due to laser polarization it is clear to see in (c) that $\mathrm{Si}{\mathrm{V}}^{-}$ does not exhibit any preferential orientation, whereas there is a clear correlation in (f) for the 1.40-eV defect to indicate this defect has grown in with a preferred orientation (parallel to the sector growth direction). (Scale bar; 10 µm).