Fig. 1. (a) Architecture and schematic of the proposed optical receiver; (b) 3D model diagram of the p-i-n PD; (c) cross-sectional schematic of the PD.

Fig. 2. Simulation results for the proposed PD. (a) Responsivity and bandwidth versus the length of the Ge region; (b) S21 curves and energy distribution at a length of 20 µm.

Fig. 3. (a) Equivalent circuit model of PD; (b) bandwidth and peaking versus the inductance induced by wire bonding; (c) wire bonding with a common PCB (left) and a customized PCB (right); (d) frequency responses from co-simulation method; (e) simulated group delay at different EQ strengths; (f) simulated eye diagrams at different EQ strengths.

Fig. 4. Proposed SiPh receiver. (a) Micrograph of two chips on the test board; (b) graph of hybrid integration by co-packaging; (c) experimental setup for the O-E response.

Fig. 5. (a) Measured O-E S-parameters of the PD and the proposed optical receiver; (b) measured I-V curves (including dark current and light current) and responsivity of the PD used.

Fig. 6. Measured eye diagrams. Optical input at (a) 56 Gb/s NRZ and (b) 112 Gb/s PAM-4; electrical output from CH1 at (c) 56 Gb/s NRZ, (d) 80 Gb/s PAM-4; electrical output at 112 Gb/s PAM-4 from (e) CH1, (f) CH2, (g) CH3, and (h) CH4.

Fig. 7. (a) Measured results of BERs versus the input optical power of the PD at 56 Gb/s in NRZ and 112 Gb/s in PAM-4; (b) measured THD of the EIC.

Table 1. Comparison with State-of-the-Art High-Speed Optical Receivers