Metal-semiconductor-metal photodetectors

MSM photodiodes attracted attention due to their high-speed performance and ease of integration, but this interest has waned recently. This paper endeavors to explore why this occurred and tries to address these issues. MSM photodiodes have a much lower capacitance per unit area than p-i-n photodiodes, and are often transit time limited. MSM photodiodes are comprised of back-to-back Schottky diodes using an interdigitated electrode configuration on top of an active light collection region. The transit time is related to the spacing between these interdigitated electrodes. MSM photodiodes are more easily integrated with pre-amplifier circuitry than p-i-n photodiodes. One reason is that MSM photodiodes do not require doping which eliminates any parasitic capacitive coupling between the photodiode and doped regions within the active transistors. Another reason is that the Schottky electrodes of the MSM photodiodes are essentially identical to the gate metallization of field effect transistors (FET), which might eliminate one photolithography step. But, MSM photodiodes suffer from very low external quantum efficiency (EQE) and high leakage currents. MSM photodiodes exhibit low EQE because the metallization for the electrodes shadows the active light-collecting region. Shadowing can limit the incident light from reaching the active region of the MSM detector and prevent an ideal MSM from achieving high EQE. Transparent conductors have been shown to nearly double responsivity. Leakage currents are determined primarily by the Schottky barrier heights. These can be unreliable. However, thin wide bandgap cap layers can be inserted below the Schottky and different metals used for the anode and cathode to break symmetry and to circumnavigate these concerns.