Anomalous nuclear magnetic resonance spectra in ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ nanowires

We report ${}^{209}\mathrm{Bi}$ nuclear magnetic resonance spectra in nanowires and single crystals of ${\mathrm{Bi}}_2{\mathrm{Se}}_3$. Crystals were powdered to simulate the random orientation of the nanowires, and the spectra are compared with theoretical expectations derived via an orientational study of a single crystal. The nanowire spectra are consistent with the randomly oriented powders; however, we find an unusual suppression of signal intensity as a function of field orientation that may be associated with screening by surface currents.

Figure 1

(a) X-ray diffraction spectrum (red) of nanoribbons grown by the methods reported here. The phase purity of the nanowires is confirmed by the matching of peak locations between experimental data and predictions (blue vertical lines). (b), (c) Scanning electron microscopy images of nanoribbons and nanoplates.

Figure 2

(a) Spectra of ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ single crystal taken as a function of the angle $\theta$ between the magnetic field and the normal to the quintuple layers (inset), at 10 K and 9 T. The single-crystal spectra at nonzero angles reveal intensity reductions as shown also in Fig. 6. (b) Theoretical angular-dependent spectra of ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ predicted using the amplitudes from the spectrum at $\theta =0$ and the formula $\Delta {\nu }_Q=\frac{1}{2}{\nu }_Q(3{cos}^2\theta -1)$, showing the expected increase in center amplitude near ${60}^{\circ }$.

Figure 3

NMR spectra of the ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ nanoparticles at 5 K (red circles) and 60 K (green squares).

Figure 4

Experimental 10 K, 9 T ${}^{209}\mathrm{Bi}$ spectra of ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ ground with a mortar and pestle, showing the spectrum of bulk states. The data was first taken with the powder as made (blue squares), then after strain relief (green circles), then after fixing in epoxy resin in zero field to prevent alignment of grains (red triangles). All three spectra are dominated by two broad peaks, as in the nanoparticle case (Fig. 3).

Figure 5

${}^{209}\mathrm{Bi}$ powder spectra simulated by three separate methods, (a), (b), and (c), as described in the text. (d) Sum of theoretical amplitude spectra with the same number of angle points as the experimental data; this spectrum exhibits bumpiness similar to the solid gold pattern. (e) Experimental spectrum of mortar-and-pestle ground sample.

Figure 6

Total areas under spectrum as a function of angle: experimental (red circles), theoretical expectation (dotted line), and attempts to reproduce powder spectrum as described in text (black, blue, and green lines).

Figure 7

Simulated powder spectra (black, blue, and green lines) and observed spectrum in the mortar-and-pestle sample (red circles) using spectral weighting functions $\beta \left(\theta \right)$ as described in the text.

Figure 8

Spin-spin relaxation rate as a function of angle for a single crystal of ${\mathrm{Bi}}_2{\mathrm{Se}}_3$. The relaxation times are independent of angle above ${18}^{\circ }$, thus ruling out spin-spin decoherence as a cause for the anisotropy of signal intensity.