Transport measurements and analytical modeling of extraordinary electrical conductance in Ti-GaAs metal-semiconductor hybrid structures

We present a comprehensive study of a phenomenon, extraordinary electroconductance (EEC), in microscopic metal-semiconductor hybrid (MSH) structures at room temperature. Our artificially designed MSH structure shows highly efficient external electric field sensing properties not exhibited by bare semiconductor structures. The microscopic device is fabricated from a GaAs epitaxial layer with a Ti/Au shunt subject to an external electric field and gives a maximum 5.2% EEC effect corresponding to an external electric field resolution of 3.05 V/cm at a bias field of 2.5 kV/cm. Moreover, the study reveals a strong dependence of the transport properties on the geometry of the MSH. An analytical two-layer model is developed which provides good agreement with the experimentally observed data. We propose that scaled down nanoscopic EEC sensor arrays can be used as an imaging technique for the charge distribution on a single cell surface in real time.

Figure 1

(Color online) Panel (a) A three-dimensional (3D) schematic of the EEC van der Pauw structure. A concentric metal shunt of radius $50\mu \text{m}$ is in a direct contact with the GaAs mesa of radius $100\mu \text{m}$. Four leads are deposited on the periphery of the mesa surface and lead 5 is directly connected to the shunt. Panel (b) An SEM image of an EEC device without the shunt metal on top. Panel (c) A cross-sectional view of the EEC multilayer structure. The shunt is composed of two 50-nm-thick metal thin films Ti and Au/Ge. A pair of parallel plates, as shown in Fig. 2, is incorporated to apply an external field with a $1\mu \text{m}$ ${\text{Si}}_3{\text{N}}_4$ dielectric between the top plate and the shunt.

Figure 2

(Color online) A schematic diagram of the experimental setup for measuring the four-point resistance using lock-in method. The EEC device is inside a parallel plate capacitor configuration and the electric field ($E=V/d$, where $d$ is the separation between the top and the bottom plates) is directly proportional to the voltage applied across.

Figure 3

(Color online) Main panel: the room-temperature $I\text{−}V$ characteristic of the Schottky interface. The inset (a) shows the linear dependence between the ideal factor $n$ and the current as $I<16\mu \text{A}$. The inset (b) shows the quasiexponential dependence between the current and the voltage as $0<I<0.5\mu \text{A}$. Lines are the curve fittings.

Figure 4

(Color online) The four-lead resistance of samples with $60\mu \text{m}$ in mesa radius and different values under direct bias voltage.

Figure 5

(Color online) Main panel (a) Four-point resistance of an EEC device measured with respect to an external electric field. Inset (b) shows the device sensitivity at the corresponding shunt biases. Panel (c) EEC effects in four different devices with $\alpha =1/16,5/16,10/16,14/16$.

Figure 6

(Color) Panel (a) A 3D schematic of the two-layer structure when the EEC device is under a reverse bias. The red cylinder in the top layer represents the depletion region. The blue arrows in both layers correspond to the in-plane current considered in calculating the resistance of the EEC device. The dashed red arrow in the top layer shows the nonlaminar current that flows from the top annulus region to the bottom layer underneath the depletion region. Panel (b) The top view of the upper layer in the two-layer model. The center inhomogeneity has the same radius as the metal shunt and the same thickness as the depletion width.

Figure 7

(Color online) The comparison of the curvature between the measured four-point resistance of the EEC device of $\alpha =7/16$ and the values calculated from the two-layer analytical model for different $\beta$’s. The shaded area corresponds to the lines for different values of $\beta$ ranging from 0 (top) to 0.2 (bottom). We have added offset values of $-12.02$ and $+1\Omega$ from the measured resistance and $\beta =0.2$ lines, respectively, so that the plotted lines converge at the zero voltage $\left(V=0\right)$.

Figure 8

(Color online) The comparison between theory and experiment on the four-lead resistance of samples with $60\mu \text{m}$ in mesa radius and a full spectrum of values under direct reverse bias. The solid lines are theoretical predictions from the two-layer model with the adjustable parameter $\left[16\alpha ,R_C\left(\alpha \right)\right]=\left(1,10.92\Omega \right)$, $\left(2,20.45\Omega \right)$, $\left(3,12.69\Omega \right)$, $\left(4,6.09\Omega \right)$, $\left(5,11.23\Omega \right)$, $\left(6,13.63\Omega \right)$, $\left(7,12.02\Omega \right)$, $\left(8,9.39\Omega \right)$, $\left(9,7.81\Omega \right)$, $\left(10,6.91\Omega \right)$, $\left(11,13.22\Omega \right)$, $\left(12,9.16\Omega \right)$, $\left(13,10.97\Omega \right)$, and $\left(14,13.39\Omega \right)$.