Optically-Pumped, Low-Noise, High-Power Semiconductor Optical Amplifiers

The Erbium-doped fiber amplifiers (EDFA) have revolutionized optical fiber communication systems. Although semiconductor optical amplifier (SOA) existed before EDFA, it has so far been limited to niche applications despite continued efforts aimed at improving its performances. Any meaningful advance in SOA technology would require a fundamentally new approach or a paradigm shift.  We propose a new concept of optically pumping SOA (OP SOA) to improve the performances of SOAs over that of EDFA. The most important performance metrics that conventional SOAs lag behind EDFAs are 1) noise figure (NF), 2) saturation power and 3) polarization-independent gain. The noise figure of an amplifier increases with the population inversion factor, internal loss, and fiber coupling loss. For conventional, electrically-pumped SOAs, 1) maximum doping concentration and thermal effects limit injected carrier density and how low the population inversion factor can be; and in the meantime, 2) high doping concentration leads to high internal loss due to free carrier absorption. The electrical injection also leads to undesirable tradeoffs among noise figures, polarization-dependent gain, and gain bandwidth. All of the above constraints, limitations, and tradeoffs of conventional SOAs can be simultaneously and effectively mitigated, if not completely removed by using optical pumping. In OP SOAs, carriers are generated locally and proportional to the pump intensity, which can be increased beyond the level of electrical pumping, improving NF, gain coefficient and gain bandwidth.  Furthermore, since there is no carrier transport, thermal effects will be greatly reduced. The entire heterostructure of the OP SOA is completely undoped, and hence there will be no free carrier absorption in the cladding regions. Without the need for carrier injection and transport, the mode size can be increased without much restriction which reduces fiber-to-waveguide coupling loss and thus the NF and increases the saturation power. The material composition can be chosen independently to achieve both ultrawide gain bandwidth and polarization-independent gain.  We will design the heterostructure and waveguide geometry and then fabricate the OP SOAs paying careful attention to minimizing the waveguide losses. We will then measure the optical gain and noise figure of the devices. We will optimize the designs and the fabrication processed to meet our target goals of > 20 dB of optical gain and a noise figure of  < 5 dB, and a saturation power above 20 dBm in Phase I. In addition, without the constraints of electrical pumping, many different approaches for achieving polarization independent gain will be combined to arrive at a material and waveguide design for < 2dB polarization-dependent gain. We have assembled a strong team with outstanding expertise in photonic devices and optical communication systems, as well as technology transfer and commercialization to carry out the proposed STTR research.