Mid-infrared silicon photonics emerge as the dominant technology to bridge photonics and electronics in multifunctional
high-speed integrated chips. The transmission and processing of optical signals lying at the mid-infrared
wavelength region is ideal for sensing, absorption-spectroscopy and free-space communications and the use of group IV
materials becomes principally promising as the vehicle towards their realization. In parallel, optical forces originating
from modes and cavities can reach to outstandingly large values when sizes drop into the nanoscale.
In this work, we propose the exploitation of large gradient optical forces generated between suspended silicon beams and
optomechanical transduction as a means of converting signals from the mid-infrared to the near-infrared region. A midinfrared
signal is injected into the waveguide system so as to excite the fundamental symmetric mode. In the 2-5μm
wavelength range, separation gaps in the 100nm order and waveguide widths ranging from 300–600nm, the mode is
mostly guided in the air slot between the waveguides which maximizes the optomechanical coupling coefficient and
optical force. The resulting attractive force deflects the waveguides and the deflection is linearly dependent on the midinfrared
A simple read-out technique using 1.55μm signals with conventional waveguiding in the directional coupler formed by
the two beams is analyzed. A positive conversion efficiency (<0dB) is foreseen for waveguides with suspending lengths
up to 150μm. The converter could be ideal for use in sensing and spectroscopy rendering the inefficient mid-infrared
detectors obsolete. The low-index unconventional guiding in mid-infrared could be a key component towards
multifunctional lab-on-a-chip devices.
In this paper an all-optical reservoir computing scheme is modeled, that paves an alternative route to photonic high bit
rate header identification in optical networks and allow direct processing in the analog domain. The system consists of
randomly interconnected InGaAsP micro-ring-resonators, whereas the computation efficiency of the scheme is based on
the ultra-fast Kerr effect and two-photon absorption. Validation of the system’s efficiency is confirmed through detailed
numerical modeling and two application orientated benchmark tests that consists in the classification of 32bit digital
headers, encoded an NRZ optical pulses, with a bitrate of 240Gbps,and the identification of pseudo-analog patters for
real time sensing applications in the analog domain.
Spectral and power characteristics of QD stripe lasers operating in two-state lasing regime have been studied in a wide range of operation conditions. It was demonstrated that neither self-heating nor increase of the homogeneous broadening are responsible for quenching of the ground-state lasing beyond the two-state lasing threshold. It was found that difference in electron and hole capture rates strongly affects light-current curve. Modulation p-type doping is shown to enhance the peak power of GS lasing transition. Microring and microdisk structures (D = 4-9 μm) comprising 1.3 μm InAs/InGaAs quantum dots have been fabricated and studied by μ-PL and NSOM. Ground-state lasing was achieved well above root temperature (up to 380 K). Effect of inner diameter on threshold characteristics was evaluated.