Full text loading...
(a) Schematic of scanning thermoelectric microscopy (SThEM) setup with room temperature (T0) probe tip in contact with heated (T1) sample, consisting of the cleaved surface of a GaAs p-n junction. (b) Cross-sectional scanning tunneling microscopy of a p-n junction, with junction location defined by the black dashed line and SThEM tip trajectory defined by the white dotted line. (c) Dashed lines represent the targeted dopant profiles, assuming full activation of ZnGa and SiGa; solid lines represent the redistributed dopant profiles following diffusion during GaAs growth at 580 °C (seeRefs. 9–12 ).
(a) Position-dependence of free electron (black) and hole (blue) concentrations calculated using the 1D Poisson equation (see Ref. 13 ) for the diffused dopant profile in Fig. 1(b) . (b) Position-dependence of the measured thermoelectric voltage, VSThEM, with the junction at x = 0. For comparison, the Seebeck coefficient computed using the free carrier concentration profile was used to determine the voltage profiles using the δT(r), 3D network, and quasi-3D conversion methods, labeled VδT(r), V3D, and Vquasi-3D. The most significant variations in VSThEM occur within 100 nm of the p-n junction interface, consistent with the estimated depletion width of 200 nm in (a). The general agreement between the position-dependences of Vquasi-3D and V3D suggests that the quasi-3D matrix approximation can be used in lieu of the 3D network simulation.
Effective circuit diagram of the tip-sample contact point with position-independent conductances, G, and voltage sources connected in three directions perpendicular and two directions parallel to the p-n junction interface. For isotropic G, the voltage, Vi, is reduced to the average of V in the five directions.
Position-dependence of Squasi-3D, SδT(r), and SComp. Squasi-3D is in general agreement with SComp, while the magnitude of SδT(r) is generally smaller than SComp, especially in the vicinity of the interface.
Article metrics loading...