Second order optical phenomena in gallium phosphide microdisks on silicon – ORPHEUS

Despite the economic and technological interest of the monolithic integration of III-Vs on silicon, the growth of semiconductors such as GaP onto a nonpolar substrate (e.g. Si) remains a real challenge for the development of efficient photonic devices. The nucleation of the III-V material onto Si generates domains of random crystal polarity, called antiphase domains (APDs). These structral defects induce large roughness and they are also commonly considered as recombination centers, detrimental effects for most photonic applications.
But the randomicity of the crystal polarity in GaP/Si layers can also be a major asset for the realization of nonlinear photonic devices for it relaxes the phase matching condition of second order phenomena (the so-called random quasi-phase matching, RQPM). However, this alternative QPM is only achievable under specific conditions of the geometrical distribution of APDs.
The project is organized into four tasks dedicated to solve the technological issues preventing the realization of GaP/Si photonic devices based on RQPM:
i) Are we able to control the APD statistical distribution in GaP/Si layers, to achieve given nonlinear schemes?
ii) What is the best coupling strategy for dual frequency photonic converters?
iii) What is the maximal quality factor in our GaP/Si microdisks, taking into account the structural defects within the material and the induced roughness?
iv) Can we obtain an efficient and broadband OPO in these devices?

Control of the geometrical properties of the antiphase domain distribution in GaP/Si: The research team has developed a unique know-how in the control of nucleation and evolution of APD during the crystal growth. We can now realize GaP/Si layers with emerging APDs of size in the range of 200nm, optimal value for the observation of RQPM in resonant devices.
GaP/Si membranes with sub-nanometer roughness: The occurrence of APDs in GaP/Si induces strong roughness of the layer, detrimental for the realization of photonic devices. We have developed a mechano-chemical polishing process allowing us to decrease the layer roughness down to 1nm RMS, compatible with the realization of high quality factor resonators.
Proposal of a dual-wavelength coupling schemes for photonic converters based on GaP microdisks: Despite the advances in the development of integrated frequency converters, the coupling to a resonant device remains a technological lock for efficient devices. We compared several coupling strategies for the coupling of a microdisk designed for a specific conversion scheme, in the framework of 3D photonic architectures.
Proposal for frequency conversion using surface phonon-polaritons in photonic devices based on GaP. Frequency conversion within the telecom band enabled by the interaction with optical phonons is an exploratory pathway of the ORPHEUS project. We evidenced the specific selection rules appearing when strong coupling occurs in the phonon-photon interaction, which leads to quasi-particles called surface phonon-polaritons, with unique confinement properties within photonic devices.

By the end of 2019, the first GaP/Si microdisks benefiting from APD distribution control and roughness management have been realized. Their characterization will first allow to assess the residual losses induced by structural defects within the material, for the whispering gallery modes in the spectral ranges of interest. This assessment is crucial in order to decide the exploration pathway for nonlinear phenomena in GaP/Si microdisks. Quality factors larger than 5000 for microdisks containing APDs is the mark we set to investigate RQPM in these devices for several nonlinear schemes like second harmonic generation, parametric down conversion or difference frequency generation. These experiments will open the path towards more exploratory research on photon entanglement on chip. IF we cannot realize good enough GaP/Si devices with APDs, we will concentrate on domain-annihilated microdisk for more standard QPM schemes. The integration of microdisks will also go on with both the investigation of GaP growth on SOI and the bonding of GaP/Si microdisks on host substrates and their encapsulation into a photonic circuit.

Peer-reviewed scineitific journals :
1. Ida Lucci et al. Universal description of III-V/Si epitaxial growth processes, Phys. Rev. Mat., 2, 060401, 2018
2. Alejandro Lorenzo Ruiz et al. Dual wavelength evanescent coupler for nonlinear GaP-based microdisk resonators (submitted to OSA Continuum)

International conferences:

1. Alejandro Lorenzo Ruiz et al. THz surface phonon polariton generation in GaP photonic waveguides, International conference on infrared, millimeter and THz waves, 2019, Paris, France
2. C. Cornet et al. Universal growth mechanism of III-V semiconductors on silicon for photonics and solar hydrogen production. SPb Photonic, optoelectronics & energy materials 2019, St Petersburg, Russia (invited)
3. Ida Lucci et al. A universal mechanism to describe III-V epitaxy on Si, 20th European molecular beam epitaxy Workshop, 2019, Lenggries, Germany
4. C. Cornet et al. GaP-based materials on silicon for photonics and energy, Smart NanoMaterials 2018: Advances, innovation and applications, 2018, Paris, France (invited)
5. Lin Chen et al. Photoluminescecnce of 2D-vertical In-rich APBs embedded in InGaP/SiGe/Si, 20th International conference on Molecular beam epitaxy, 2018 Shanghai, China
6. Lin Chen et al. Excitons bounded around In-rich antiphase boundaries, 34th international conference on the Physics of semiconductors, 2018, Montpellier, France

National communication:
1. Alejandro Lorenzo Ruiz et al. Dual wavelength vertical coupling for second harmonic generation in GaP microdisks, 39ème journées nationales d’optique guidée, 2019, Palaiseau, France
2. C. Cornet et al. Universal growth mechanism of III-V/Si: using antiphase boundaries for devices, Réunion plénière du GDR Pulse, 2019, Clermont Ferrand, France
3. Alejandro Lorenzo Ruiz et al. , Slit waveguide coupling system for integrated microresonators, 16èmes journées Nano, micro et opto-électronique, 2018, Cap Esterel, France

Massive growth of data flux and energy sustainability progressively push the information technologies to their limits. The ITRS now motivates researchers to look for hybrid computing solutions both on the hardware and conceptual sides. In this field, photonic integrated circuits (PICs) are particularly promising due to the large bandwidth and high power efficiency they feature, and also because light is an attractive medium to develop alternative computation paradigms such as quantum computing. In this view, the demonstration of specific photonic building blocks is crucial. Such devices should have a potential for large scale integration, CMOS compatibility and at the same time they should offer innovative functionalities for alternative computation paradigms.
Far from the microscale world of PICs, optical parametric oscillators (OPOs) have for long demonstrated their potential as wavelength converters, amplifiers, entangled photon sources etc. Scaling these optical devices down to the chip is a vivid research field; LiNbO3 microresonators integrated on Si and III-V nonlinear components are presently being studied in this goal. The gallium-phosphide on silicon (GaP/Si) platform shows many assets in this field: It shows a larger second order susceptibility than LiNbO3 and its optical properties also allows for more compact devices and optimal wavelength range. GaP can be integrated monolithically on silicon (an expertise domain of FOTON laboratory) and can be processed easily, guaranteeing low cost and large scale integration. GaP/Si integration can even lead to alternative and broadband quasi-phase-matching solutions for 2nd order processes (e.g. random QPM) by use of native defects called anti-phase domains (APDs) in the GaP layer. In spite of this, very few results have been published so far on second order nonlinear processes in GaP microresonators and still less on monolithically integrated GaP/Si ones.
The goal of the ORPHEUS project is thus to demonstrate a broadband GaP microdisk resonator OPO monolithically integrated on silicon. Different operation regimes of this OPO will be under scrutiny:
- the generation of signal and idler modes around 1550nm, a prerequisite for telecom wavelength conversion and photon entanglement,
- the generation of a signal in the telecom band from a pump in the datacom first window (800-900nm) in order to investigate the possibilities of off-chip (850nm) to on chip (1550nm) optical data transfer.
- the configuration where both pump and signal are in the telecom band and where the 2nd order process is mediated by optical phonons.

To do so, the ORPHEUS project is organized in 4 tasks that will address four scientific hypotheses that are at its basis:
i) Are we able to control the statistical distribution of APDs appearing in GaP/Si epilayers?
ii) What are the best solutions to achieve integrated dual wavelength optical coupling into semiconductor microdisks?
iii) What is the limit Q factor of GaP/Si microdisks?
iv) Can we obtain efficient and broadband OPO processes in GaP/Si microdisks?

The scientific diversity of these hypotheses highlights the multidisciplinary character of the ORPHEUS project, which is at the frontier between material science, technological science and advanced optics. The use of the native defects of GaP/Si epilayers in order to magnify nonlinear optics phenomena in GaP/Si photonic devices is the strongest force of the project and a real factor for innovation. This said, the ORPHEUS project also features many intermediate objectives that are by themselves original contributions to the research field of integrated photonics and that will progressively open the path towards the development of more complex photonic architectures based on the GaP/Si technological platform.