Repetition Rate Locking of Mutually Injected Monolithic Passively Mode-Locked Semiconductor Quantum Dot Lasers

Monolithic passively mode-locked (PML) semiconductor lasers are attractive high-speed photonic emitters potentially enabling novel future secure communication schemes [1]. In this contribution, mutual all-optical injection of two PML InAs/InGaAs quantum dot (QD) lasers emitting at 1250 nm with equal active region and waveguide geometry is studied experimentally for the first time and explained theoretically. The active regions of both two-section lasers consist of 5 layers of InAs/InGaAs QDs, and the emission wavelengths are at around 1255 nm. Both cavity lengths amount to 2 mm with a 0.25 mm long absorber section at the back facet that is reverse biased at 4 V. The facets are anti-reflective (front facet) and high-reflective (back facet) coated. The lasers are stabilized at different heat sink temperatures (Laser 1: 15 °C, Laser 2: 24 °C) to ensure optimum wavelength overlap. Both lasers are mutually coupled by a setup, depicted schematically in Fig. 1a). Emission beams of both lasers are individually collimated and overlapped across a distance of approximately 1.65 m. Optical fine-delay is implemented by a retro-reflector mounted on a linear translation stage. Using a 50/50 beam splitter and incorporating two optical isolators (>60 dB isolation), light from both directions can be fiber-coupled and by an optical spectrum analyzer and a fast photo diode in combination with an electrical spectrum analyzer, spectral and radio-frequency (RF) analysis can be performed. Both lasers are biased at 200 mA and the free-running repetition rate (RR) amounts to 20.074 GHz for Laser 1 and 20.087 GHz for Laser 2. We translate the single-path fine-delay across 66 ps in steps of 1.47 ps and record both lasers repetition rates RR for the mutually injection scheme. Obtained results are depicted in Fig. 1b) as a function of the added time delay. We find that both RRs follow a saw-tooth shaped dependence with plateaus which is attributed to the strong external optical self-feedback stemming from the facets of the opposite laser [1]. Furthermore, the well overlapping RRs of both lasers in Fig. 1b) indicate mutual injection locking, as analyzed in [2]. In a second step, both lasers RRs are analyzed in dependence on the gain section biasing of Laser 1 of 100 mA to 250 mA in steps of 5 mA. Fig. 1c) displays the recorded RRs under mutual injection with Laser 2 biased at 200 mA. Additionally, we measure both RRs with the opposite laser unbiased in order to identify the influence of the OFB without the presence of the mutual injection signals (black and green curve in Fig 1c).