Silicon Photonics Devices and Integrated Circuits

1. Introduction

The rapid evolution of integrated photonics has ushered in a transformative era for optical communication and information processing systems, with silicon-based optical chips emerging as a cornerstone technology. Building upon the mature infrastructure of complementary metal-oxide-semiconductor (CMOS) technology, these devices leverage silicon and silicon-derived substrate materials to achieve unprecedented synergies between photonic and electronic functionalities [1,2]. Unlike traditional semiconductor chips constrained by electronic interconnects, silicon photonic devices exploit the unique optical properties of silicon to enable high-speed data transmission while maintaining compatibility with existing semiconductor manufacturing processes [3]. This dual advantage positions silicon photonics as a critical enabler for next-generation computing architectures, particularly as conventional electronic systems approach fundamental physical limits in power efficiency and heat dissipation.

The fundamental appeal of silicon photonics lies in its material compatibility and scalability. Silicon’s high refractive index ($n\approx 3.5$ at telecommunications wavelengths) enables strong light confinement in submicron-scale waveguides, facilitating the dense integration of optical modulators [4]. Moreover, the inherent transparency of silicon in the near-infrared spectrum (1.1–1.6 $\mathsf{\mu }$m) aligns perfectly with standard optical communication bands, allowing for seamless integration with existing fiber-optic networks [5]. Recent advances in heterogeneous integration techniques have further expanded the material palette, enabling the incorporation of III-V semiconductors (e.g., InP, GaAs) and germanium-based photodetectors through advanced bonding processes [6]. These developments have transformed silicon photonic circuits from simple passive structures to fully functional systems incorporating lasers, modulators, and detectors on a single chip [7].

Operational efficiency constitutes another critical advantage driving silicon photonics adoption. Photonic interconnects demonstrate orders-of-magnitude improvements in energy-per-bit metrics compared to their electronic counterparts, with experimental silicon photonic links achieving femtojoule per bit efficiencies [8]. This energy advantage becomes particularly pronounced in data-intensive applications such as artificial intelligence accelerators, where optical neural networks leveraging silicon photonic matrix multipliers have demonstrated 10–100× improvements in computational energy efficiency relative to electronic GPUs [9,10]. The inherent parallelism of photonic computing architectures, combined with silicon’s capacity for wavelength-division multiplexing, creates multidimensional processing capabilities that transcend conventional electronic limitations [11].

The quantum information revolution presents another frontier for silicon photonic integration. The strong nonlinearity in silicon material enables the generation and manipulation of entangled photon pairs through spontaneous four-wave mixing processes [12]. Furthermore, the CMOS-compatible fabrication of quantum photonic circuits allows mass production of entangled photon sources and quantum logic gates at scales unattainable with conventional nonlinear crystals [13,14]. These advances position silicon photonics as a critical platform for developing practical quantum communication networks [15,16] and fault-tolerant quantum computers [17].

The purpose of this Special Issue, “Silicon Photonics Devices and Integrated Circuits”, is to present the most recent findings and creative solutions in the field. The following are the original research papers and review papers on integrated photonics gratings, waveguides, filters, demultiplexer and sensors.

2. An Overview of Published Articles

Using silicon nitride waveguides with innovative bent asymmetric coupled structures and partial Euler bends can reduce radiation loss in curved sections. The first paper by Chamorro-Posada [18] finds that shaping the bend with a clothoid region featuring linearly varying curvature allows for smoother transitions between straight and curved segments. By adjusting parameters such as the external waveguide width and the gap between the main and wrapping waveguides, it becomes possible to control mode confinement and minimize radiation losses. A careful modal analysis shows that the principal mode—selected for its effective index closest to that of the isolated main waveguide—can be maintained even under curvature-induced distortions. The addition of an external waveguide proves beneficial, reducing the required lateral offset and further improving field confinement. Trade-offs between radiation loss and device footprint are addressed by tuning the length of the partial Euler bend. Such design flexibility opens up new possibilities for compact, low-loss curved waveguide sections used in Q-enhanced microresonators and integrated polarizers. This approach emphasizes practical design strategies that are compatible with current fabrication processes while enhancing overall photonic integrated circuit performance.

Chirped integrated Bragg gratings in silicon-on-insulator technology offer a versatile means for wavelength-selective filtering and dispersion management. Praena and Carballar [19] examine two classic methods: one that chirps via Bragg period variation, and another that employs a waveguide width modulation. A novel alternative combines these methods to maintain a constant average effective refractive index along the device while varying the coupled Bragg wavelength. The resulting design framework is compatible with CMOS manufacturing processes, and provides enhanced control over spectral features, making it well suited for applications in optical communications, photonic signal processing, and dispersion control.

High-extinction photonic filters based on cascaded Mach–Zehnder interferometer-coupled resonators address the increasing demand for precise wavelength selectivity and pump rejection in integrated photonic circuits. Constructed on a silicon nitride platform using I-line lithography, Chen et al. [20] present the design integrating low-loss bus and microring resonators connected via an interferometric feedback loop. The coupling parameters, resonator gap, and feedback loop length are carefully engineered to influence both the spectral sharpness and extinction ratio of the filter. Simulations reveal that while a single-ring configuration provides modest extinction, cascading two, three, or four resonators linearly boosts the extinction ratio up to values exceeding 36 dB for a four-stage cascade. Experimental results corroborate the simulations, displaying clear, narrow-band filtering within the telecommunication C-band. This cascaded architecture not only enhances filtering performance, but also offers scalability and robust integration for advanced optical signal processing and quantum photonics applications, all without the need for thermal tuning.

Dielectric waveguide sensors offer a unique approach to achieving high sensitivity by enhancing the evanescent field. Designing waveguides in configurations like slot, ridge, or hybrid plasmonic structures allows a portion of the guided light to extend beyond the core, improving interaction with surrounding analytes. Different material platforms such as silicon, silicon nitride, and polymers bring distinct advantages in refractive index contrast, fabrication flexibility, and integration potential. The review article by Butt et al. [21] systematically explores dielectric waveguide-based sensors with enhanced evanescent fields for biosensing and gas sensing. These sensors excel at detecting subtle changes in refractive index or molecular absorption, making them ideal for environmental monitoring, medical diagnostics, and industrial process control. Careful engineering of the waveguide structure leads to improved light–matter coupling and dynamic range, which is critical for achieving low detection limits. Strategies that enhance the evanescent field not only boost performance, but also open up promising avenues for integrated photonic sensor systems that combine high sensitivity with scalability for future applications.

Advanced optical demultiplexers based on silicon nanowires provide a promising solution for high-density datacenter interconnects. The review article by Jeong et al. [22] discusses three silicon-nanowire-based optical demultiplexers. One design features a 16-channel dense wavelength division multiplexing filter with 100-GHz spacing, while another targets eight channels for long-reach applications with 800-GHz spacing. A third configuration uses a four-channel coarse wavelength division multiplexing scheme with 20 nm spacing. A key element is the integration of a polarization splitter-rotator that efficiently converts arbitrarily polarized input signals into a consistent ${\mathrm{TE}}_{00}$ mode for subsequent processing. Techniques combining delayed interferometers with arrayed waveguide gratings help achieve flat-topped spectral responses, low insertion losses, and minimal polarization-dependent losses. Extensive theoretical and experimental evaluations reveal clear channel separation, reliable operation with 32 Gbps NRZ signals, and robust performance under realistic conditions. Advanced lithography techniques ensure high fabrication precision, keeping channel uniformity and low excess loss within tight tolerances. This work offers valuable insights into creating compact, high-performance optical filters that address the growing demands of next-generation data communications.

3. Conclusions

In conclusion, silicon-based optical chips represent a technological nexus where photonics and electronics converge to redefine performance boundaries. The articles in this Special Issue exemplify the rapid progress in silicon photonics, addressing challenges in loss reduction, and spectral control. Innovations in waveguide engineering, hybrid grating designs, and cascaded filter architectures underscore the field’s potential for transformative applications in communications, quantum technologies, and environmental monitoring. Future work will focus on leveraging machine learning for design optimization and 3D integration for multifunctional photonic-electronic systems. The coming years will undoubtedly witness silicon’s transition from an electronic workhorse to a photonic linchpin in the post-Moore era.