K-clock Interferometer Silicon Photonics LiDAR

Introduction

LiDAR (Light Detection and Ranging) technology is being extensively developed for 3D imaging applications like autonomous vehicles. Traditional LiDAR systems use mechanical beam scanners which are bulky, heavy, and lack robustness. Silicon photonics offers a compelling solution with nonmechanical beam scanning using optical phased arrays (OPAs) or focal plane arrays (FPAs).

One innovative approach is the frequency-modulated continuous-wave (FMCW) LiDAR chip developed by researchers at Yokohama National University. This chip integrates a slow-light grating (SLG) nonmechanical beam scanner made of photonic crystal waveguides along with an on-chip k-clock sampling interferometer.

SLG Scanner & FMCW Ranging

The SLG scanner consists of 32 photonic crystal waveguides with shallow-etched gratings and integrated switch trees to select one waveguide and propagation direction. This allows non-mechanical beam steering by leveraging the slow-light effect in the gratings.

For FMCW ranging, the laser light needs to be frequency modulated. Typically this is done using expensive equipment like single-sideband modulators and arbitrary waveform generators (AWGs). The YNU team's approach is to directly modulate the laser frequency and compensate the nonlinearity using k-clock sampling with an integrated interferometer.

On-Chip K-Clock Interferometer

The key innovation is integrating a Mach-Zehnder interferometer (MZI) with a long delay line on the same silicon chip as the SLG scanner. This acts as the k-clock interferometer for sampling the FMCW waveform.

As shown in Fig. 1a, 10% of the input light is tapped into the MZI which has a tunable coupler and germanium photodetector (Ge PD). The long delay arm uses a low-loss 24 cm quasi-singlemode waveguide folded 57 times to keep the chip compact.

Fig. 1: (a) Schematic and (b) photo of the integrated LiDAR chip with k-clock interferometer

When the laser wavelength is swept, the output at the Ge PD is a clean sinusoidal clock waveform as seen in Fig. 2, with the frequency determined by the delay line length.

Fig. 2: Output from 24 cm Si waveguide k-clock interferometer

This integrated k-clock avoids the need for separate fiber-optic interferometers used previously. The researchers estimate the on-chip interferometer length could be extended to 1 m while still having acceptable 40 dB total optical loss.

System Operation

The overall FMCW LiDAR system is shown schematically in Fig 3a. A wavelength-swept laser source provides the frequency modulated light to the chip. The k-clock from the integrated MZI interferometer is used to sample the beatnote from the ranging target mixed with the laser output at the SLG port.

Fig. 3: (a) Schematic of FMCW LiDAR system and (b) experimental setup*

Interpolation is applied to the sampled waveform to increase the resolution, with an interpolation factor of 16 used in this work. From the sampled and interpolated waveform, the target range can be calculated via Fourier analysis.

By sweeping the laser wavelength and repeating the ranging process while scanning the SLG beam direction, a 3D point cloud image can be constructed, as demonstrated in Fig. 4b for a target object covered with a retroreflector.

Fig. 4: (a) Target object and (b) real-time 3D point cloud image*

The image acquisition speed was limited to 0.032 fps by the 50 nm/s sweep rate of the laser source, but can be improved with a faster swept source in the future.

Conclusion

This work demonstrates the first integrated silicon photonic FMCW LiDAR with on-chip k-clock interferometer. Leveraging slow-light gratings for beam steering and direct laser modulation with k-clock sampling enables a compact, low-cost and robust system compared to traditional LiDAR architectures.

The 24 cm interferometer length allowed ranging with 16x interpolation, and simulations indicate 1 m delay lines could extend the unambiguous range to 100 m. Further increasing the laser sweep speed will enable real-time video-rate 3D imaging for applications like autonomous vehicles.

This silicon photonic LiDAR highlights the potential for dense integrated photonic systems combining multiple functionalities like lasers, detectors, waveguides and gratings on a single chip for advanced applications beyond just communications and interconnects.

Reference

[1] S. Yamazaki, T. Tamanuki, M. Kamata, and T. Baba, "K-clock interferometer-integrated Si photonics SLG FMCW LiDAR," Department of Electrical and Computer Engineering, Yokohama National University, Yokohama, Japan, 2024.