Electrical Characterization Technique Of HgCdTe Infrared Photovoltaic Dtectors

Mercury cadmium telluride(Hg Cd Te) is the leading semiconductor material for the fabrication of high performance infrared photon detectors, due to its ascendant photoelectric property. In recent two decades, with the development of third-generation Hg Cd Te infrared photodetector, demands of stable process, high yield and uniform distribution of device parameters are rapid growing. Properly designed microelectronic test structures and data analysis techniques are needed to evaluate semiconductor materials, to determine and monitor process centring parameters, to identify and quantify yield-limiting defects and to assess processing tool performance. Therefore, several microelectronic test structures, gated diode, metal-insulator-semiconductor(MIS) capacitor and transmission line test structure, were exploited in this thesis and main content are summarized as follows.1. Gated diode and surface leakage current. This thesis investigates the effect of surface fields on the dynamic resistance of a planar Hg Cd Te mid-wavelength infrared photodiode from both theoretical and experimental aspects, considering a gated n-on-p diode with the surface potential of its p-region modulated. Theoretical models of the surface leakage current are developed, where the surface tunneling current in the case of accumulation is expressed by modifying the formulation of bulk tunneling current, and the surface channel current for strong inversion is simulated with a transmission line method. With proper values of the modeling parameters such as flat-band voltage VFB, surface trap density Nst and channel electron mobility μc, the theoretical R0 A product and corresponding dark current calculated from the proposed model as function of the gate voltage demonstrate good consistency with the measured values.2. Metal-insulator-semiconductor(MIS) Capacitors. The Hg Cd Te/Insulator interfacial characteristics are systemically investigated. Poisson’s equation of space charge region of Hg Cd Te MIS devices is built by taking carrier degeneracy and conduction band non-parabolicity. Based on this, space-charge density Qs, ideal high/low frequency C-V curves and flat-band voltage for Hg Cd Te MIS capacitors are obtained. To interpret a discrepancy of minimum high-frequency capacitance between experimental data and theoretical calculation, which is closely related to surface treatment processes, equivalent high-frequency trapped charges were assumed and the quality of Hg Cd Te MIS devices can be characterized by equivalent high-frequency traps of density Dh,it. The theory and the analysis method were applied to the Hg Cd Te MIS devices and the values of relevant physical parameters have been extracted: fixed charge density, insulator traps of density, minimum interface traps of density, equivalent high-frequency traps of density.3. Nonlinear metal-semiconductor contacts characterization technique. A generalized transmission line model(TLM) and a compatible numerical method have been proposed in this thesis, taking account of the current transport mechanism of metal-semiconductors. A fitting program was established by introducing the genetic algorithm mode to the generalized transmission line model. The model and the program were applied to the analysis Au/Sn/p-Hg Cd Te contact, the values of relevant physical parameters, including Schottky barrier height, ideality factor n, specific contact resistance and sheet resistance under the metal, were extracted by fitting the experimental R-V data.4. Passivation layer resistivity. The current-voltage characteristics and resistivity of Cd Te/Zn S films deposited by e-beam evaporation technique were studied. The typical resistivity of the passivation layer was varying. The fourprobe technique is widely used in the characterization of electrical properties of solids and thin films. To investigate the influence of finite size probes with non-planar contact on the standard four-probe method, an image method has been proposed to simulate the potential distribution within the specimen. We have found that disregarding the probe size brings a less dominate error than that introduced by the approximate formula, when the sample’s thickness is close to the inter-electrode spacing.