UBC Theses and Dissertations
A theoretical and experimental investigation of metal-semiconductor contacts Horita, Robert Eiji
Metal-semiconductor contacts have been known empirically to obey a modified diode equation I = I (exp qV/akT - 1) where the parameter 'a' often took values greater than two. However, previous theories, which are briefly summarized in the thesis, could not simply account for the anomaly where 'a' is greater than two. Previous theories are extended by considering one-dimensional bipolar flow of carriers and neglecting recombination in a homogeneous semiconductor filament with a rectifying and an ohmic contact at opposite ends. The zero-electron-current theory by Borneman et. al. (1955) is extended to high injection levels by using the junction relations of Misawa (1955). Then the non-zero-electron-current theory is developed. This theory shows that 'a' is unity for low Injection Into extrinsic semiconductors and that a = (3b-M)/(b-M) for arbitrary injection into intrinsic semiconductors and, for high injection Into extrinsic semiconductors, where 'M' is the electron to hole current ratio and 'b' is the electron to hole mobility ratio. Thus, 'a' can take any value depending on the magnitude of M/b. To check the non-zero-electron-current theory, experiments were performed only on n-type germanium. Rectifying metal-semiconductor contacts were made by electroplating copper and rhodium on germanium and ohmic contacts were made by alloying antimony-doped gold wire in a nitrogen atmosphere. A side-arm probe adjacent to the plated-contact measured the voltage across the contact. For the rhodium-contact specimen, two side-arms adjacent to the rhodium contact on opposite sides of the germanium filament were used to measure the transverse a-c resistance as a function of longitudinal d-c current. This measurement demonstrated the occurrence of injection and extraction at the contact and that the level of injection was from moderate to high in the vicinity of the contact for longitudinal current density between about 10 mA/cm² and 1000 mA/cm² . Comparison of the I-V measurements of the contacts with the non-zero-electron-current theory indicated that an intermediate level of injection occurred at the rhodium contact for longitudinal current density between about 0.1 mA/cm² and 100 mA/cm² and that high injection occurred, with M = 4.94, at the copper contact for current density between about 1 mA/cm² and 50 mA/cm². Thus, the level of injection can be calculated by comparison of the experimental value of 'a' with the non-zero-electron-current theory.
Item Citations and Data