UC San Diego

Single photon avalanche detectors (SPADs) have attracted increasing attention because of their critical roles in many important applications such as quantum cryptography, optical time-domain reflectometry, time-resolved spectroscopy, non-line-of-sight optical communication, space communication and light detection and ranging. Although Silicon SPADs are limited to wavelength below 1.1um, III-V based SPADs with Separate Absorption and Multiplication (SAM) structure are of special interest in the telecommunication wavelengths, exhibiting superior performance with high single photon detection efficiency and low dark count rate. Conventionally, due to the self- sustaining nature of avalanche, operation of SPADs requires integration of external quenching circuit to achieve both quenching and recovering capabilities. Integration of external quenching circuit with III-V SPADs, however, adds substantial complexity to device fabrication, especially for array detectors when each individual detector requires an external quenching circuit. Here, we present single photon avalanche detectors featuring a Transient Carrier Buffer (TCB) layer to form an energy barrier which can tentatively stop avalanche-generated carriers, demonstrating self-quenching and self-recovering capabilities. For this self-quenching and self-recovering SPAD, the escape rate of those stopped avalanche carriers from the barrier determines the self-recovery time and thus count rate of the single-photon detector. A physical model has been developed to simulate the dynamic characteristics of the detector. The simulation results agree well with the experimental data, and the self- recovery time is found to be reduced with increasing temperature, the magnitude of overbias, the dosage in the charge layer and the barrier height. In addition, thermionic emission shows a stronger dependence on temperature and a weaker dependence on device bias and charge layer dosage than tunneling. The model contains no fitting parameters and therefore can be used to model and predict the device behaviors