Silicon photonic LiDAR systems have emerged as a promising technology for high-resolution, compact, and cost-effective depth sensing in applications such as autonomous vehicles, robotics, and augmented reality. Leveraging the mature CMOS fabrication ecosystem, these systems integrate optical components with electronic drivers, enabling scalable production. Three primary architectures dominate silicon photonic LiDAR: phased arrays, frequency-modulated continuous-wave (FMCW), and time-of-flight (ToF) systems. Each approach has distinct advantages and trade-offs in beam steering, coherence requirements, and performance under ambient light conditions.

Phased array LiDAR systems utilize optical phased arrays (OPAs) to steer laser beams without mechanical components. These arrays consist of multiple optical antennas controlled by phase shifters, which adjust the phase of emitted light to form a directed beam through constructive interference. Silicon photonic OPAs typically employ thermo-optic or electro-optic phase shifters, with the latter offering faster response times but requiring higher drive voltages. Beam steering resolution depends on the number of antenna elements and their spacing, with larger arrays providing narrower beam divergence and higher angular resolution. A key challenge in OPA-based LiDAR is achieving wide-angle steering, as grating lobes and side lobes can degrade beam quality. Mitigation strategies include non-uniform antenna spacing and multi-wavelength operation. Coherence length requirements are moderate, as phased arrays rely on coherent interference but do not demand ultra-narrow linewidth lasers.

Frequency-modulated continuous-wave LiDAR offers superior range resolution and velocity measurement capabilities by modulating the frequency of the laser source. The system emits a chirped optical signal, and the reflected light is mixed with a reference beam to generate a beat signal. The frequency difference encodes both distance and relative velocity of the target. FMCW LiDAR requires highly coherent lasers with long coherence lengths, typically exceeding hundreds of meters, to maintain phase stability during frequency sweeps. Silicon photonic implementations integrate narrow-linewidth lasers, such as hybrid III-V/silicon distributed feedback (DFB) lasers, with low-loss waveguides and balanced photodetectors. Beam steering in FMCW systems often combines OPAs or micro-electromechanical systems (MEMS) mirrors with wavelength tuning, leveraging the relationship between wavelength and beam angle in grating-based couplers. A major advantage of FMCW is its inherent resistance to ambient light interference, as the coherent detection scheme rejects incoherent noise. However, nonlinearities in frequency chirping can degrade range accuracy, necessitating precise laser control circuits.

Time-of-flight LiDAR measures the round-trip delay of short laser pulses to determine distance. Silicon photonic ToF systems use high-speed modulators, such as Mach-Zehnder or electro-absorption modulators, to generate nanosecond or picosecond pulses. Single-photon avalanche diodes (SPADs) or time-gated detectors capture the returning signals with high timing resolution. Beam steering is commonly achieved through MEMS mirrors or diffractive optical elements, though solid-state alternatives like liquid crystal metasurfaces are under development. ToF systems have relatively relaxed coherence requirements, as they rely on pulse timing rather than phase coherence. However, they face challenges in ambient light rejection, particularly in outdoor environments where sunlight can saturate detectors. Narrowband optical filters and time-domain gating improve signal-to-noise ratios, but dynamic range limitations persist for long-range applications. Range resolution is determined by pulse width and detector jitter, with sub-centimeter resolution achievable using picosecond pulses and low-jitter timing circuits.

Integration with CMOS drivers is a critical enabler for silicon photonic LiDAR. Co-design of photonic and electronic components optimizes power efficiency, noise immunity, and signal processing capabilities. High-voltage drivers for electro-optic phase shifters, transimpedance amplifiers for photodetectors, and time-to-digital converters for ToF systems must be co-integrated without compromising optical performance. Thermal crosstalk between nearby electronic and photonic components can degrade phase stability, necessitating careful layout and isolation techniques. Heterogeneous integration, such as bonding III-V gain materials to silicon waveguides, further enhances performance by combining high-efficiency light sources with low-loss silicon photonics.

Range resolution and ambient light rejection remain key challenges across all architectures. In phased arrays, angular resolution improvements require larger apertures, which conflict with size constraints. FMCW systems achieve high range resolution through wide frequency sweeps but face trade-offs between sweep bandwidth and laser tuning linearity. ToF systems balance pulse width and peak power to maximize resolution without violating eye safety limits. Ambient light rejection strategies vary by architecture: FMCW benefits from coherent detection, while ToF systems rely on spectral and temporal filtering. Advanced signal processing algorithms, such as matched filtering and multi-pulse averaging, further enhance performance in noisy environments.

Future developments in silicon photonic LiDAR will focus on improving scalability, power efficiency, and robustness. Monolithic integration of lasers, detectors, and beam steering elements will reduce packaging complexity and cost. Novel materials, such as silicon nitride for low-loss waveguides, may extend performance beyond traditional silicon photonics. Co-optimization of photonic and electronic components will enable higher modulation speeds and lower power consumption, critical for mobile and battery-powered applications. As the technology matures, standardization of interfaces and testing protocols will facilitate broader adoption across industries.

In summary, silicon photonic LiDAR architectures offer distinct advantages in performance, size, and manufacturability. Phased arrays provide solid-state beam steering, FMCW enables precise velocity measurement, and ToF systems deliver simplicity and scalability. Integration with CMOS electronics unlocks the full potential of these systems, though challenges in range resolution and ambient light rejection require continued innovation. Advances in materials, device design, and signal processing will further solidify silicon photonics as a cornerstone of next-generation LiDAR technology.