Skip to main content

Photonic integrated devices for second order nonlinear optical processes

Periodic Reporting for period 2 - PISSARRO (Photonic integrated devices for second order nonlinear optical processes)

Reporting period: 2020-01-01 to 2021-06-30

Full on-chip photonic integration holds the promise of bringing advanced optical technologies close to the end user by enabling cheap, compact and efficient devices. Having efficient optical nonlinearity on low loss integrated photonic circuits is key for enabling a wide variety of applications. Yet, despite the remarkable development of the field in the last couple of years there still seems to be an undeniable trade-off between available device functionality, fabrication and integration costs.

Silicon nitride (Si3N4), with tuneable material composition and well-developed CMOS-compatible material processing techniques [is heading towards being the bridging material platform for passive devices, nonlinear components, and high-performance electro-optical modulators. Si3N4 is remarkable owing to its excellent linear optical properties, and has also been exploited for its nonlinear properties, almost solely through the use of its third-order material susceptibility since unfortunately, it does not exhibit second-order susceptibility (χ(2)). This is a crippling weakness for such promising platform given the range of applications enabled by χ(2) processes. State of the art equipment therefore performs second order nonlinear mixing processes off the chip, blocking the route towards low cost miniaturized systems.

PISSARRO we aim at unlocking second order frequency mixing processes in CMOS compatible platforms with the efficiency and most importantly the flexibility required by many applications. To that end we are studying optically induced self-organized gratings in waveguides, and enable inducing a χ(2) in silicon nitride. In addition, such optical poling methods provide an all-optical control and reconfigurability essential for simple yet universal designs. We will bring on chip essential, yet missing functionalities such as second-harmonic generation, difference-frequency generation, or spontaneous parametric down conversion, that will further anchor integrated photonics as a key enabling technology. It also represents a necessary step to move beyond laboratory grade equipment towards real products by significantly scaling down the form factor but without sacrificing performance.
From the beginning of the project up to the end of this period, the work has been focus on four main pillars: (I) the investigation of the photogalvanic effect in silicon nitride waveguides, (II) waveguide engineering for the control of optically inscribed gratings, (III) improving the efficiency of the induced second order effects and (IV) applications.

(I) Investigation of the photogalvanic effect in silicon nitride waveguide: we performed studies on how the self-organized inscription of the gratings is influenced by temperature, the length of the waveguide and the optical power used for the inscription. We quantified the decay of the grating with temperature and showed good stability at room temperature. We observed that short waveguides are difficult to pole owing to the necessary built up of seeding second harmonic power while there exists a threshold for the pump intensity to initiate the process. As an important step in understanding of the photogalvanic process, we were able to image the inscribed gratings in a wide range of waveguides with different cross sections using two photon microscopy. This allowed us to establish that the grating inscription takes place from the interference of pump and its second harmonic both on the fundamental waveguide mode, despite the very large refractive index mismatch. As such, the photogalvanic effect can lead to gratings with very short periods, and the second harmonic is automatically generated on the fundamental mode, a great advantage for applications.

(II) Waveguide engineering: while quasi phase matching is automatically satisfied between the pump and its second harmonic, such that the peak of the process can be controlled in wavelength, we also showed that optimizing the dispersion of the waveguide can be used to engineering the bandwidth of the second harmonic process. We performed simulations and validated our findings experimentally. We also investigated the effect of mode mixing due to bends in the waveguide designs. Since long length are beneficial for initiating and improving the photogalvanic effect, the waveguides need to be folded on the chip. We observed that depending on the dimensions of the waveguides and the radius of the bends, mode mixing takes place such that the second harmonic leaks into other modes than the fundamental ones.This line of research results in knowledge necessary for the design and layout of the waveguides.

(III) Improving efficiencies: an important aspect is how efficient the nonlinear process is. We have been investigating ways of increases the current values by self-seeding or externally seeding the process. We are also looking at increasing the length of the waveguides: leveraging the extreme low loss fabrication in collaboration with EPFL colleagues and using resonating structures.

(IV) Applications: in addition to broadband and tunable second harmonic generation, we have shown how SiN waveguides can process femtosecond pulses by combining octave spanning supercontinuum generation and second harmonic generation of the femtosecond pump. This fully silicon nitride scheme allowed for the detection of the frequency comb carrier envelope offset frequency (fceo) with more than 35 dB of optical SNR , sufficient for phase stabilization of the frequency comb. Additional work on other applications are currently underway.
During this first period, we were able to establish the formation rules of the optically induced gratings in silicon nitride waveguide, revealing for the first time important information on the microscopic nature of the induced second order nonlinearity in those waveguides. This crucial information completely confirmed the initial coherent photogalvanic effect hypothesis and established fundamentally new understanding of the phenomena.
We have shown extreme tunability and reconfigurability of a second-order based frequency converter in silicon nitride and we are now pushing further the efficiencies. We believe that by the end of the project we can aim for a record combined tunability and efficiency of such second order nonlinear processes, relying on a standard silicon nitride platform that does not require any additional processing compared to standard waveguides. We foresee impact for quantum application as well as widely tunable light generation on chip.
Image of chip used for second harmonic generation