Dynamic ${}^{14}\mathrm{N}$ nuclear spin polarization in nitrogen-vacancy centers in diamond

We studied the dynamic nuclear spin polarization of nitrogen in negatively charged nitrogen-vacancy (NV) centers in diamond both experimentally and theoretically over a wide range of magnetic fields from 0–1100 G covering both the excited-state level anticrossing and the ground-state level anticrossing magnetic field regions. Special attention was paid to the less studied ground-state level anticrossing region. The nuclear spin polarization was inferred from measurements of the optically detected magnetic resonance signal. These measurements show that a very large (up to $96\pm 2\%$) nuclear spin polarization of nitrogen can be achieved over a very broad range of magnetic field starting from around 400 G up to magnetic field values substantially exceeding the ground-state level anticrossing at 1024 G. We measured the influence of angular deviations of the magnetic field from the NV axis on the nuclear spin polarization efficiency and found that, in the vicinity of the ground-state level anticrossing, the nuclear spin polarization is more sensitive to this angle than in the vicinity of the excited-state level anticrossing. Indeed, an angle as small as a tenth of a degree of arc can destroy almost completely the spin polarization of a nitrogen nucleus. In addition, we investigated theoretically the influence of strain and optical excitation power on the nuclear spin polarization.

Figure 1

Left: (a) the NV center's energy-level structure. Right: The nuclear spin polarization process at the (b) ESLAC and (c) GSLAC. Dashed lines indicate nonradiative transitions and straight lines represent the optical transition, which conserve nuclear spin.

Figure 2

Experimental setup. The top of the image shows the overall experimental scheme. The bottom of the image shows a zoomed-in, detailed scheme of the sample holder and the microwave wire.

Figure 3

(a), (b), (d), (e) ODMR signals at individual magnetic field values. (c) Experimental (blue dots, each representing one ODMR measurement) and theoretical (solid curves) nuclear spin polarization as a function of magnetic field. A more significant change depending on the angle can be seen at the GSLAC region rather than the ESLAC. Pumping rate for theoretical calculations was ${\mathrm{\Gamma }}_p=5\mathrm{MHz}$. The experimental ODMR curve fit determined that the magnetic field angle $\theta =0.2^{\circ }$, which is in good agreement with the calculated curve at magnetic field angle $\theta =0.2^{\circ }$.

Figure 4

(a) Theoretical nuclear spin polarization around the GSLAC for different magnetic field angles $\theta$. The pumping rate for the numerical calculations was ${\mathrm{\Gamma }}_p=5\mathrm{MHz}$. (b) Experimental nuclear spin polarization at the GSLAC for different magnetic field angles $\theta$.

Figure 5

Left: ODMR signals at the GSLAC (1024 G) for different angles between the magnetic field and the NV axis; (a) shows the theoretical ODMR signal, whereas (b), (c), (d), and (e) show experimental (black dots) and fitted (red curve) ODMR signals. Right: representations of the external magnetic field direction with respect to the NV center's axis (f) for no transverse magnetic field or (g) a visually exaggerated transverse magnetic field. (a) corresponds to the geometry shown in (f); (c), (d), and (e) correspond to the geometry shown in (g).

Figure 6

Theoretical nuclear spin polarization around the GSLAC for different transverse strain $N_x$ values. The pumping rate in the numerical calculations was ${\mathrm{\Gamma }}_p=5\mathrm{MHz}$. The excited-state strain was set to $N_x^e=10N_x^g$.

Figure 7

Theoretical nuclear spin polarization around the GSLAC at different pumping rate ${\mathrm{\Gamma }}_p$ values for $\theta =0.2^{\circ }$.