Efficient conversion of light to charge and spin in Hall-bar microdevices

We report an experimental study of the direct conversion of light into electrical signals in GaAs/AlGaAs Hall-bar microdevices. Our approach, based on different modulation frequencies of the intensity and polarization of the laser beam, allows us to disentangle the charge- and spin-dependent parts of the induced electrical signal and to link them to the incident light intensity and polarization, respectively. We demonstrate that the efficiency of the light to spin conversion in our electrical polarimeter is strongly enhanced by adding a drift component to the transport of the spin-polarized photocarriers, as compared to a purely diffusive transport regime of the device. For a micron-size focused laser beam, the experiments demonstrate that the light to charge and spin conversion efficiency depends on the precise position of the light spot, reflecting the spatially dependent response function of the Hall cross.

Figure 1

(a) The experimental setup with its two modulators: the chopper modulator of the light intensity at reference frequency $f_1=300$ Hz and the photoelastic modulator that switches the ${\sigma }_{+}/{\sigma }_{-}$ circular polarization at $f_2=42$ kHz. (b) Sketch of the device design with Hall crosses on samples A and B. Indicated dimensions are described in the text. Inset: The spatial coordinate system.

Figure 2

Diffusive spin currents in sample A. (a) Simulated Gaussian profile of the optically injected spin polarization density $S_z$ with FWHM $=2\mu \mathrm{m}$ (black curve) and the corresponding diffusive spin current proportional to $\nabla S_z$ (red [gray] curve). (b) Experimental Hall voltages measured at $f_2$ and HC 1 (red [light gray] curve) and HC 2 (blue [dark gray] curve) as a function of a one-dimensional (1D) displacement of the light spot (FWHM $\approx 2\mu \mathrm{m}$ and $P=15\mu \mathrm{W})$. (c) The Hall voltage measured at $f_2$ and HC 1 with respect to the two-dimensional (2D) position of the spot [same parameters as in panel (b)]. (d) A typical dependence of the Hall voltage at HC 1 on the drift current applied along the transport channel for ${\sigma }_{+}/{\sigma }_{-}$ helicities and for power $P=700$ nW.

Figure 3

Drift-dominated spin currents in sample B. The sample is biased at $I=10\mu \mathrm{A}$ and scanned with a light spot of size 5 $\mu \mathrm{m}$ $(2.5\times$ the width of the channel) and power $P=80$ nW. (a) Variation in the longitudinal voltage $V_{xx}$ measured at chopper reference frequency $f_1$. (b) The Hall voltage $V_{xy}$ sensed by the HC 1 at PEM frequency $f_2$. (c) The longitudinal voltage $V_{xx}$ measured at PEM frequency $f_2$ between HC 1 and 2. Insets: Sketches of polarities of Hall voltages with respect to the spot position (red [gray] circles).

Figure 4

Response functions in sample B. (a) The Hall voltage $V_{xy}$ sensed at reference frequency $f_1$, related to photoinduced changes in conductivity. (b) The Hall voltage $V_{xy}$ sensed at PEM frequency $f_2$. Data in panels (a) and (b) are measured with the spot of diameter $1\mu \mathrm{m}$ and power 80 nW; the channel is biased with $I=10\mu \mathrm{A}$. (c,d) Simulations of Hall voltages at frequencies $f_1$ and $f_2$ for the same size of the HC and laser spot.