A Photodetector Applied in Optical Communication System and Its Performance Prediction Under Mathematical Model

To improve the transmission speed and operating power of photodetectors, the uni- traveling-carrier photo-diode (UTC-PD) structure with InAIAs/InGaAs and P-type heterostructures is proposed. Compared with InP materials, the higher conduction band and valence band energy of InAIAs materials can effectively block the diffusion of photogenerated electrons to the electron blocking layer, and at the same time reduce the barrier height when the photogenerated holes transit to the P contact layer. On this basis, a dual-die symmetrically-connected p-i-n type photodetector array is prepared. The array can achieve a 3 dB bandwidth of more than 14.6 GHz at an average photocurrent of 5 mA. At the same time, in order to solve the performance degradation prediction of the photodetector construction system under multi-coupling conditions, a mathematical model with the energy domain of photoelectric detection system as the table parameter is proposed. This model uses the digital twin model constructed by the modulation transfer function (MTF) to construct the static physical model of the photoelectric detection system, and uses the Bayesian network to realize the evolution description of the transfer function, thereby describing the dynamic degradation description of the photoelectric detection system. In the experiment, the structure of different heterogeneous UTC-PDs is compared. Using InAlAs/InGaAs type and InP/InGaAs heterostructures, the 3 dB bandwidth with 0 deviation, which is above 20 GHz, can be realized. The proposed dual-die symmetric connection structure can improve the saturation characteristics of its array under the photodetection epitaxial layer structure and device size. The mathematical model is used to receive the various parameters and status data of the photoelectric detection system. By comparing the output image and the standard target image, the predicted data is consistent with the measured data, which proves that the proposed photoelectric detection structure-performance prediction model is credible.