Ultra-Low Power CMOS-compatible Integrated Photonic Platform for Terabit-Scale Communications

Free-space laser links used for satellite communications are currently limited in modulation speeds and modulation formats due to the high power-per-bit consumption of the optical transceivers. In this project we take advantage of recent breakthrough 3D monolithic integration of photonic structures, particularly high-speed graphene-silicon devices on CMOS electronics in order to create CMOS-compatible high-bandwidth transceivers for ultra-low power terabit-scale optical communications. The design space of the envisioned platform for this project is shown in Fig. 1 in which the target signal integrity and end-to-end energy-per-bit are set as the figure of merits for the inter-satellite free-space link. The parameters subject to optimization in the design space are the modulator’s bandwidth, thermal control of optical devices, and the tradeoff between required net coding gain (NCG) and modulation overhead.
The graphene modulator’s operation bandwidth is not limited by the carrier transport in graphene, but it is bottlenecked by the parasitic components of the modulator circuit. These parasitics are the capacitor formed by the graphene-dielectric-graphene stack, graphene-sheet resistance, and graphene-metal contact resistance. While it is possible to design the device footprint in such a way to reduce the capacitance and the sheet resistance, and simultaneously optimize both modulator performance and operation bandwidth, the contact resistance is fundamentally related to how the metal-graphene interface is formed. Previously, surface contacts have been employed to interface the metal and the graphene, where the metal is deposited at the graphene’s top surface in order to form contacts. The metal and graphene interaction in the surface contacts approach occurs perpendicularly to the graphene molecular two-dimensional (2D) plane. Since graphene’s vertical orbital hybridization is weak, surface contacts are fundamentally incapable of maximizing metal and graphene interaction. In this project our team is aiming at exploring novel ways to break through this barrier and achieve ultra-high bandwidth graphene-based (and other 2D material-based) silicon modulators.
The prevalent optical modulation scheme used in today’s long haul communication network is amplitude modulation (ASK), where the digital data (one/zero) is encoded in the ON/OFF of optical signal. To increase the bandwidth of this binary modulation scheme, ASK is used in conjunction with wavelength division multiplexing (WDM), mode division multiplexing (MDM) and polarization division multiplexing (PDM). However, coherent optical modulation using phase shift keying (PSK) has remained an elusive technology, since the bulk materials used for optical PSK not only manipulate the phase of light, but also induces significant optical absorption/loss. This optical loss occurs because conventional silicon phase modulators (such as PN, PIN, MOS and III -V on silicon) rely on changes in the carrier concentration, which changes not only the real part of the refractive index but also the imaginary part. Today, this optical loss is reduced by decreasing the carrier concentration required for modulation using long devices, thereby limiting the scalability of these systems. Phase modulators based on either thermo-optic or induced electro-optic effects have low optical losses, but suffer from either high electrical power consumption and low speed or a large device footprint and complex fabrication, respectively. Here, we demonstrate a platform that enables phase-only modulation in the near infrared (NIR) region of the electromagnetic spectrum based on semiconductor monolayers integrated on passive silicon nitride waveguides.