Giant microwave-induced $B$-periodic magnetoresistance oscillations in a two-dimensional electron gas with a bridged-gate tunnel point contact

We have studied the magnetoresistance of a quantum point contact fabricated on a high mobility two-dimensional electron gas (2DEG) exposed to microwave irradiation. The resistance reveals giant $B$-periodic oscillations with a relative amplitude $\mathrm{\Delta }R/R$ of up to $700\%$ resulting from the propagation and interference of the edge magnetoplasmons (EMPs) in the sample. This giant photoconductance is attributed to the considerably large local electron density modulation in the vicinity of the point contact. We have also analyzed the oscillation periods $\mathrm{\Delta }B$ of the resistance oscillations and, comparing the data with the EMP theory, extracted the EMP interference length $L$. We have found that the length $L$ substantially exceeds the distance between the contact leads, and rather corresponds to the distance between metallic contact pads measured along the edge of the 2DEG. This resolves existing controversy in the literature and should help to properly design highly sensitive microwave and terahertz spectrometers based on the discussed effect.

Figure 1

(a) A setup of the device; the current flows from contact 1 to contact 2, and the voltage is measured between contacts 3 and 4; a magnetic field $B$ is perpendicular to the 2DEG plane. (b) Geometry of the metallic contact pads (hatched areas), contact leads (areas B), and of the 2DEG (area A); the black vertical element labeled as “g” is the bridged-gate QPC. (c) The longitudinal resistance $R_{1,2;3,4}\equiv R$ at microwave frequency $f=156$ GHz, temperature $T=4.2$ K, the gate voltage $V_g=-1.39$ V, and different values of the microwave power density.

Figure 2

The longitudinal resistance for different gate voltages $V_g$ at MW frequency $f=156$ GHz and temperature $T=4.2$ K. The MW power attenuation is 0 dB.

Figure 3

(a) The magnetoresistance $R\left(B\right)$ for frequencies $f=128$, 132, 148, and 166 GHz, temperature $T=4.2$ K, MW power attenuation 0 dB, and gate voltage $V_g=-1.39$ V. (b) The interference parameter $qL/2\pi$ calculated from Eqs. (1) and (2) (continuous curves) and the oscillation index (symbols) as a function of $B$. The fitting length $L$ is chosen in such a way that the theoretical curves coincided with the experimental positions of the oscillation maxima at integer values of $qL/2\pi$; for details of the fitting procedure, see the text.

Figure 4

The calculated EMP dispersion relation (1) and (2) plotted as the dimensionless wave vector ${[{\omega }_p\left(q\right)/\omega ]}^2=(2\pi n_s{e}^2/m^{*}\overline{\kappa }{\omega }^2)q$ vs dimensionless magnetic field ${\omega }_c/\omega =(e/m^{*}c\omega )B$. The three finite-$d$ curves are plotted for $n_s={10}^{12}$ ${\mathrm{cm}}^{-2},f=156$ GHz, and parameters of GaAs $\left(\kappa =12.8,m^{*}=0.067\right)$.

Figure 5

The fit of the experimental data from (a) Fig. 1(b) of Ref. [9] (sample A, $f=37.5$ GHz, $L_{ab}=1$ mm, $\mathrm{\Delta }R$ curve) and (b) Fig. 1(a) of Ref. [5] $(f=53$ GHz, $L_{ab}=0.5$ mm).