Skip to main content

Low power consumption silicon optoelectronics based on strain and refractive index engineering

Periodic Reporting for period 4 - POPSTAR (Low power consumption silicon optoelectronics based on strain and refractive index engineering)

Reporting period: 2020-04-01 to 2020-09-30

The performance evolution of telecommunication networks, computing systems and integrated circuits requires increasing communication bandwidth at all interconnect levels. Also, the power efficiency, i.e. the energy required to transfer data, must be considerably decreased. The use of silicon photonics has been well identified in all prospective reports as a means to overcome electrical interconnect bandwidth and power efficiency limitations. This research domain has exhibited a remarkable rate of development, with current advances, which were inconceivable 10 years ago. This evolution is largely based on the vision that silicon as a mature integration platform can bring photonic integrated circuits closest to electronic circuits, driven by the economy of scale of generic wafer-scale integration technologies. However, despite the demonstration of high performance silicon modulators, germanium photodetectors, and III-V lasers on silicon, their integration in a common chip is highly challenging due to the different materials and technologies involved and it is far from being cost-effective. In addition, wideband silicon modulators require bias swings of several volts to achieve good modulation behavior, which result in high power consumption and considerably degrade the global energy impact of the circuit. Furthermore, current Silicon-on-Insulator (SOI) photonic platform exhibits some major drawbacks for the development of reliable and robust technology for a large range of applications: i) No second order nonlinearity because of its centro-symmetry, i.e. no Pockels effect, (ii) high two photon absorption (TPA) leading to free carrier absorption (FCA), increasing the overall optical loss and decrease nonlinear optical efficiency, (iii) Poor efficiency in light detection in the telecom wavelength range.
In this silicon photonics ecosystem, the POPSTAR project will address a new route to make key advances in the development of low power consumption multi-wavelength high-speed communication circuits based on second- and third-order nonlinear optical effects in silicon. The original idea of the project is to generate strains in sub-wavelength silicon photonic nano-structures leading to significant breakthroughs in second-order nonlinearities efficiency (Pockels effect) and in third-order nonlinearities.
The POPSTAR project has been addressing several challenges from the beginning of the project. The first objective was to describe and understand the evolution of the second-order nonlinear optical coefficient as a function of the strain induced into silicon waveguide. A complete physical model has then been developed taking into account the strain gradient induced in silicon by a top stress layer, the optical mode and the evolution of the electric field. Carrier effect, which can be considered as a parasitic effect for the evaluation of Pockels coefficients, has also been studied and was taken into account in this model. It is a pure multiphysics work and the understanding of Pockels effect in strained materials was not an easy task. A new model has been developed and is always under validation. It will paves the way in the optimization of the Pockels effect in silicon and in the reduction of the carrier impact.
The second objective was to demonstrate Pockels effect in silicon waveguide. Numerous samples have been fabricated in the host clean room and tested based on the deposition of silicon nitride on the top of the waveguides. Optical modulation results up to 40GHz have then been obtained, which clearly demonstrated the presence of Pockels effect in silicon. Furthermore, an open eye diagram at 10Gbit/s has also been obtained. This promising result opens the route towards the demonstration of high-speed optical modulators in silicon. These promising measurements have been carried out considering silicon nitride (SiN) as a stress layer. It was also demonstrated that SiN is not the best material to act as a stressor due to some fixed charges at the SiN/Si interface. Then, as a third objective, new materials to induce strain in silicon have been explored. First, we developed the epitaxial growth of YSZ (Yttria-stabilized zirconia) on silicon. YSZ is a crystalline oxide with very interesting properties. The optimization of the epitaxial growth has been performed on Sapphire and silicon substrates and good crystalline quality has been obtained. However, the strain induced in silicon is not strong enough, especially due to the formation of an oxide layer at the Si/YSZ interface. We are also studying new materials including AlN and Chalcogenide in collaboration with Aalto Univ. and MIT to induce nonlinear effects and strain in silicon waveguide. Strip and rib waveguides have been designed, fabricated and characterized leading to promising low propagation loss.
During the project, the PI team has also demosntrated outstanding results on photodetectors and avalanche photodiode for high speed communications.
In parallel, some design, fabrication and characterization have been performed on subwavelength silicon structures. The objective here was to demonstrate the versatility of subwavelength structures for dispersion engineering and mode confinement engineering.
To conclude, all these results confirm the great potential of strained silicon photonics for the development of low power consumption and high-speed photonic circuits.
The POPSTAR project propose to address numerous challenge in Si photonics based on strained silicon photonics. Indeed, the demonstration of Pockels effect in silicon will lead to a new class of chirp-free, ultra-high speed and low power consumption optical modulators. Furthermore, the strain engineering coupled with the refractive index engineering based on sub-wavelength nano-structuration can also have a significant impact in variable applications related on nonlinear optical effects including all optical processing, routing systems using phase switching devices, lab-on-chip circuits for the detection of multi-assays in parallel combining the source and detection in silicon platform. The concepts and the technology developed in the project won’t be limited in the NIR wavelength range and can also be exploited in the mid-IR region where some recent developments have demonstrated a great potential mainly oriented for sensing and free space communications.
Here below are the main results beyond the state of the art obtained during Popstar project:
- Development of a new model describing Pockels effect in strained materials
- Demonstration of high speed Pockels effect in Si waveguide
- Demonstration of reliable SiGe heterojunction for light detection at 40Gbit/s
- Demonstration of 40Gbit/s avalanche Ge receiver
- Demonstration of Si/SiN photonics devices for wavelength and polarization managing.
- 2 octave supercontinuum in CMOS compatible SiN waveguides
- Ultra-high optical gain in erbium doped YSZ on SiN photonics waveguides
- Low loss crystalline YSZ waveguide
- First demosntration of third order nonlinearity in YSZ waveguides including ab-initio simulations.
High-speed optical modulation in strained si waveguides