Strategic Insights into Integrated Photonics: Core Concepts, Practical Deployments, and Future Outlook

Abstract

Integrated photonics is a cutting-edge field that merges optics and electronics on a single microchip, revolutionizing how we manipulate and transmit light. Imagine traditional bulky optical systems condensed onto a chip smaller than a fingernail, enabling faster communication, more efficient sensors, and advanced computing. At its core, integrated photonics relies on guiding light through waveguides etched onto semiconductor substrates, analogous to how wires conduct electricity in traditional electric circuits. These waveguides can route, modulate, and detect light signals with unprecedented precision and speed. This technology holds immense promise across various domains. Despite its immense potential, integrated photonics faces challenges, including manufacturing complexities and integration with existing electronic systems. However, ongoing research and advancements continue to push the boundaries, promising a future where light-based technologies seamlessly integrate into our everyday lives, powering a new era of innovation and connectivity.

1. Introduction

Electronic integrated circuits (EICs) have been the foundation of modern technology, enabling the development of a wide range of electronic devices and systems. These circuits manipulate electrical signals using semiconductor materials such as silicon (Si) [1,2]. However, as demands for higher data speeds, lower energy consumption, and increased bandwidth continue to grow, EICs face limitations due to factors like resistive losses and electromagnetic interference. Photonic integrated circuits (PICs) offer a compelling alternative by leveraging light, rather than electrons, for signal transmission [3]. PICs utilize components like optical fibers, waveguides (WGs), and photonic crystals to process and transmit data, offering advantages such as higher speed, lower energy consumption, and reduced signal loss over long distances [4,5,6,7].

Table 1. PICs versus EICs.

Aspect	PICs	EICs
Operational mechanism	Utilizes photons (light) for signal transmission	Relies on electrons for signal transmission
Medium	Optical fibers, WGs, and photonic crystals	Si or other semiconductor materials
Speed	Usually, higher speed	Limited by electron mobility and resistance
Energy efficiency	Generally, more energy-efficient	Subject to resistive losses and heat dissipation
Bandwidth	Higher bandwidth capabilities	Bandwidth limited by material properties
Signal loss	Lower signal loss over long distances	Signal loss over distance in conductors
Interference	Less susceptible to EM interference	Susceptible to EM interference
Integration density	Limited by photonic component sizes	Can achieve high integration densities
Overall cost	Typically, higher costs due to fabrication complexity	Lower cost due to mature manufacturing processes
Applications	Optical communication, computing, sensing, imaging	Computing, data processing, control systems

provides a general overview of the differences between PICs and EICs, highlighting their respective characteristics and typical applications.

Integrated optics refers to the technology and study of integrating multiple optical devices and functions onto a single chip, similar to the way EICs combine multiple electronic components. The term was introduced by Stewart E. Miller in 1969, who envisioned the integration of optical circuits to perform complex functions within a compact and efficient framework [8]. The development of integrated optics was driven by the need for miniaturization and improved performance in optical communication systems. In the realm of modern optics, the need for PICs arises from the ever-growing demand for faster, more efficient, and compact optical systems [9]. PICs offer a revolutionary approach by integrating numerous optical components onto a single chip, mimicking the functionality of traditional EICs [10]. This integration drastically reduces size, weight, and power consumption while enhancing performance and reliability. Integrated photonics traces its origins back to the late 20th century when researchers began exploring ways to scale down and integrate optical components onto a single platform [11]. The basics of integrated photonics lie in leveraging the properties of light to process, transmit, and detect information. It implicates the integration of optical elements such as WGs, modulators, and detectors on a common substrate, typically Si, silicon nitride (Si₃N₄) or lithium niobate (LN) [12,13,14,15,16]. This integration enables the manipulation and control of light at the nanoscale, leading to compact and efficient photonic devices [10,17]. Over the years, evolution in fabrication techniques, materials science, and device design has propelled integrated photonics to the forefront of innumerable technological domains, including telecommunications, computing, sensing, and quantum technologies [18,19]. Today, integrated photonics stands as a basis of contemporary optics and continues to drive innovations that reshape the way we communicate, compute, and interact with light.

PICs exhibit advanced multiplexing capabilities that significantly enhance their functionality in optical communication systems [20]. These circuits can efficiently manage multiple light signals of different wavelengths simultaneously, a technique known as wavelength division multiplexing (WDM) [21]. This allows a single optical fiber to carry multiple channels of data, greatly increasing the bandwidth and data transmission capacity without requiring additional physical infrastructure. Additionally, PICs support space division multiplexing (SDM) [22], enabling multiple spatial channels within a single fiber, and time division multiplexing (TDM), which divides signals into different time slots [23,24]. The integration of these multiplexing techniques within a compact and scalable platform makes PICs highly advantageous for high-speed, high-capacity optical networks, driving advancements in telecommunications, data centers, and various sensing applications [25].

In this review, we probe into the intricate realms of integrated photonics, exploring material platforms, fabrication methods, applications, and the associated challenges in the development of PICs. Furthermore, a forward-looking perspective is offered on the future of integrated photonics, shedding light on potential advancements and avenues for innovation in this rapidly evolving field. Figure 1 serves as the focal point of this dialogue, elucidating its core concepts and providing a visual anchor for the debate.

2. Fundamentals of PICs

Integrated photonics is a field that emphasizes the development and application of photonic devices united onto a single platform, often referred to as a PIC [26]. These devices manipulate light (photons) instead of electrical signals (electrons), leveraging the exceptional properties of light for a variety of applications. The fundamental components of integrated photonics include WGs, modulators, detectors, and light sources, each playing a critical role in the functionality of PICs.

WGs are the backbone of any photonic integrated circuit. They confine and direct light within the chip, much like electrical wires guide electricity [27,28]. The design of WGs involves controlling the index contrast between the core and the cladding materials, ensuring efficient light propagation with minimal losses [29]. Materials commonly used for WGs include Si, Si₃N₄, and indium phosphide (InP), each offering different advantages such as compatibility with existing semiconductor manufacturing processes, low propagation losses, or efficient light emission and detection capabilities [30,31].

Optical modulators are essential for encoding information into a light wave [32]. They function by varying the intensity, phase, or polarization of the light passing through them [33]. In integrated photonics, modulators often exploit the electro-optic (EO) or thermo-optic (TO) effects [34]. The EO effect, observed in materials like LN and Si, allows for rapid modulation of the refractive index with an applied electric field [13]. TO modulators, on the other hand, change the refractive index through temperature variations, though they tend to be slower and consume more power [35]. Advanced modulators are continuously being developed to balance speed, efficiency, and integration compatibility [36,37].

Detectors in integrated photonics convert optical signals back into electrical signals. They are crucial for applications where data need to be read out or further processed electronically [38]. Photodetectors (PDs) in PICs are typically made from materials with high optical absorption coefficients, such as germanium (Ge) integrated on Si, or III-V compounds like indium gallium arsenide (InGaAs) [15]. These materials are chosen based on their sensitivity to specific wavelengths and their ability to integrate seamlessly with other photonic components on the same chip [39].

The development of on-chip infrared cameras represents a significant advancement in imaging technology, integrating infrared sensing capabilities directly onto semiconductor chips [40]. This innovation leverages microelectromechanical systems (MEMS) and advanced fabrication techniques to create compact, low-power, and cost-effective infrared sensors [41,42]. By embedding infrared detection directly on the chip, these cameras offer enhanced performance, higher resolution, and faster response times. This breakthrough is poised to impact various fields, including consumer electronics, healthcare, security, and automotive industries, enabling applications like night vision, thermal imaging, and non-invasive medical diagnostics [43,44].

Light sources, including lasers and light-emitting diodes (LEDs), are pivotal in integrated photonics. Integrating efficient, coherent light sources onto photonic chips has been a significant challenge due to the material incompatibilities and the thermal management issues involved. Hybrid integration techniques, where III-V semiconductor lasers are bonded onto Si chips, have become a popular approach [45,46,47]. Alternatively, on-chip generation of light through Si photonics has made progress with technologies like Si Raman lasers and Si-based light-emitting diodes [48].

The study conducted by Wu et al. demonstrated that transformation optics techniques enable the design and integration of functional components made from planar gradient index materials into photonic circuits [49]. The exceptional design flexibility provided by transformation optics facilitates the creation of innovative devices such as light source collimators, WG adapters, and WG crossings (Figure 2). These devices are highly applicable to integrated photonic chips and are well-suited to existing assembly methods. By means of the finite-difference time-domain (FDTD) method, full-wave numerical simulations were conducted to showcase the exceptional optical performance and efficient integration of these elements within on-chip photonic systems. These elements, requiring only spatially changing dielectric materials without magnetic properties, ensured low-loss, broadband operation in integrated photonic environments.

The amalgamation of these components onto a single platform offers many advantages, such as reduced size, weight, and power consumption, improved performance and reliability, and the potential for large-scale production using semiconductor fabrication techniques. Integrated photonics is revolutionizing various fields, from telecommunications, where it enables high-speed data transmission and processing, to sensing and imaging, where it provides enhanced sensitivity and resolution. Furthermore, as technology advances, the integration of photonics with electronics (electro-photonics integration) promises even greater capabilities, enabling the development of next-generation devices that leverage the best of both worlds. The future progress of integrated photonics hinges on developing compact, reliable optical components that can be seamlessly integrated on a single substrate.

3. Commercial Availability of Photonic Devices

The commercial availability of photonic integrated devices is rapidly expanding, with numerous companies such as Intel Corporation [50], Cisco Systems [51], Infinera Corporation [52], and NeoPhotonics Corporation [53] among others developing and manufacturing PICs for various industries. From established players to startups, a growing ecosystem of suppliers is driving innovation and accessibility in the photonic integrated device market, paving the way for widespread adoption and further advancements in photonics technology. The global PIC market has experienced a remarkable surge in recent years, propelled by a confluence of factors including escalating environmental consciousness, governmental incentives, and relentless technological progress. This burgeoning market not only offers promising prospects for various stakeholders such as optical fiber communication and sensors but also underscores the significance of collaborative efforts between private enterprises and governments. By fostering synergy in policymaking, research and development initiatives, and investment strategies, this collaboration can expedite the growth trajectory of the PICs market. Moreover, the burgeoning consumer demand further amplifies the avenues for market expansion, accentuating the need for innovative solutions and enhanced product offerings. Despite a robust growth of 26.2% in 2021, the World Semiconductor Trade Statistics (WSTS) revised the projection downward to single-digit growth for the global semiconductor market in 2022, amounting to a total valuation of USD 580 billion, reflecting a still notable increase of 4.4% [54]. This adjustment underscores the dynamic nature of the industry and highlights the importance of adaptive strategies in navigating market fluctuations.

4. Applications of Integrated Photonics

The unified integration of photonic elements onto a single chip allows the creation of compact and efficient devices with amplified performance and versatility. In telecommunications, integrated photonics accelerates data transmission rates and expands bandwidth, crucial for meeting the escalating demands of high-speed internet and mobile communication. Additionally, optical computing holds the promise of revolutionizing processing speed and energy efficiency, laying the groundwork for cutting-edge architectures like optical and neuromorphic computing [18,55,56]. In addition to these areas, integrated photonics revolutionizes sensing technologies by enabling the creation of highly sensitive and compact sensors. These advancements are applicable in diverse fields such as environmental monitoring and medical diagnostics [57,58]. Its influence extends to frontier domains like quantum computing, where it plays a vital role in realizing quantum communication and information processing. Overall, integrated photonics emerges as a beacon of innovation, propelling advancements that redefine communication, computation, sensing, and exploration in the modern era [59,60,61]. In this section, we discussed four vital applications that can be executed using PICs.

4.1. Sensing Applications

A key benefit of using integrated photonics in sensing applications is the substantial decrease in both size and power consumption compared to conventional optical systems [60,62]. By integrating multiple photonic components, such as WGs, modulators, and detectors, onto a single chip, the need for bulky optical setups is eliminated. This miniaturization allows for the deployment of sensors in environments where space is at a premium or where traditional sensors cannot be easily installed. For instance, in medical diagnostics, integrated photonic sensors can be used for in vivo monitoring of biological parameters, providing real-time data with minimal invasiveness [57,63]. These sensors can detect minute changes in biological markers, leading to early diagnosis and improved patient outcomes [64]. Integrated photonics also excels in environmental sensing applications [65,66]. Photonic sensors can perceive a wide range of environmental parameters, including temperature, humidity, pressure, and the presence of various chemicals and pollutants [67,68,69,70]. The extraordinary sensitivity and selectivity of photonic sensors make them ideal for monitoring air and water quality. For example, integrated photonic sensors can detect trace amounts of pollutants in water, providing crucial information for maintaining safe drinking water standards. Additionally, their ability to operate in harsh environments makes them suitable for use in industrial settings, where they can monitor conditions in real time and alert operators to potential hazards.

The study conducted by Liu et al. presented an integrated channel WG-based fluorescent immunosensor designed to swiftly and simultaneously identify up to 32 contaminants (Figure 3a) [71]. The utilization of WG tapers by the sensor enhanced the efficiency of both the excitation and collection of fluorescent signals, even in instances of fluorophore bleaching that occur in solid surface bioassays. Figure 3b illustrates a cross-sectional representation of a sensing area, which includes the WG, isolation layer, and surface chemistry. Figure 3c depicts the propagation of light along a WG chip, offering insight into the sensor’s operational mechanism. This pioneering use of WGs in immunosensors marked the first demonstration of the ability to detect microcystin-LR (MC-LR) in lake water with an optimized WG geometry. To achieve this, a WG chip was activated using 3-mercaptopropyl trimethoxysilane/N-(4-maleimidobutyryloxy) succinimide (MTS/GMBS), which enabled a BSA-MC-LR conjugate to be immobilized. The immunosensor successfully detected MC-LR in real lake samples, even at sub-microgram per liter concentrations (e.g., 0.5 μg/L), with recovery rates ranging from 84% to 108%. These results confirmed the sensor’s potential application for measuring MC-LR in real water samples [71].

Another notable application of integrated photonics in sensing is in the field of telecommunications [72]. Photonic sensors are employed to monitor the health of optical networks, detecting issues such as signal degradation and fiber breaks with high precision [73]. This capability is essential for maintaining the reliability and performance of high-speed communication networks. Integrated photonic sensors can be embedded within the network infrastructure, providing real-time data that enables proactive maintenance and rapid response to faults [74]. The use of integrated photonics in sensing is also advancing the development of quantum sensors, which exploit quantum phenomena to achieve unprecedented levels of sensitivity and accuracy. Integrated photonics provides a platform for developing compact and scalable quantum sensors, which can be used in various applications, from gravitational wave detection to navigation and timing. The integration of quantum sensing elements on a photonic chip opens up new possibilities for fundamental research and practical applications, potentially revolutionizing fields such as biosensing [75], metrology [76], and geophysics [77].

In industrial automation, integrated photonic sensors play a crucial role in enhancing precision and efficiency [78]. These sensors can be integrated into manufacturing processes to monitor and control various parameters with high accuracy. For example, in the semiconductor industry, integrated photonic sensors are used for lithography and inspection processes, ensuring that components meet stringent quality standards. The ability to detect minute defects and variations in materials and products can lead to significant improvements in yield and product quality [79]. Furthermore, integrated photonics can facilitate the development of smart manufacturing systems, where sensors provide continuous feedback to optimize production processes and reduce downtime [80].

4.2. Optical Switches

The optical switch is a crucial component of PICs, finding extensive use in optical communications and networks, optical computing, and sensing applications such as LiDAR [81]. Typically, Si-integrated optical switches utilize the TO [82,83,84,85] or carrier dispersion effect [86,87,88] to achieve reconfigurable signal routing. It is evident that the TO effect is associated with a high level of power consumption. Furthermore, the carrier dispersion effect is characterized by a restricted refractive index alteration. In addition to this, both effects are non-latching, thereby resulting in the continuous utilization of power even when switching is not required. To address these limitations, phase change materials (PCMs) have been integrated into Si-integrated optical switches [89,90,91,92]. Non-volatile (NV) hybrid EO plasmonic switches and innovative architectures for NV combinational and sequential logic circuits are presented by Ghosh et al. [90]. The EO switches in question featured a plasmonic WG coated with a thin layer of PCM, as illustrated in Figure 3d. The optical losses in the WG were managed by modifying the phase of the PCM between its amorphous and crystalline states. The phase transition in the PCM was induced by two methods: either by electrical threshold switching or thermal conduction heating, utilizing external electrical heaters, or the plasmonic WG metal itself as an integrated heater. Figure 3e,f illustrate the field distributions of the plasmonic slot WG in both the amorphous and crystalline phases of Ge₂Sb₂Te₅ (GST). The implementation of logic gates, a half-adder circuit, and sequential circuits utilized these plasmonic switches as active components. The plasmonic switches and logic operations were demonstrated to have minimum extinction ratios (ERs) exceeding 20 dB, compact structures, low operating power, and high-speed performance. By integrating photonics, plasmonics, and electronics on the same platform, an efficient architecture for logic operations was created [90].

4.3. All-Optical Modulators

All-optical modulators based on PICs represent a pivotal advancement in optical communication and signal processing [93,94]. These modulators leverage the inherent advantages of PICs to manipulate light signals directly without converting them to electrical signals, thereby reducing latency and power consumption [95]. Utilizing materials such as Si, InP, or LN, all-optical modulators can achieve high-speed modulation rates necessary for modern data transmission [96]. They operate by altering the properties of the optical signal—such as its phase, amplitude, or polarization—using mechanisms like the EO effect, TO effect, or carrier injection. The integration of these modulators within a photonic circuit allows for compact, scalable, and energy-efficient designs, which are crucial for the development of advanced optical networks, efficient computing, and emerging technologies like quantum communication. The seamless integration of all-optical modulators with other photonic components on a single chip further enhances their potential, paving the way for ultrafast, high-capacity, and resilient communication systems.

The potential for graphene as a material in ultrafast optoelectronics has been demonstrated in applications such as all-optical modulators. The atomic thinness of graphene is a significant factor in its suitability for this purpose. However, its limited light absorption capacity results in graphene-based modulators often having low modulation efficiencies or requiring high switching energies. Plasmonic enhancement offers a potential solution to these challenges, although the insertion loss (IL) associated with plasmon-enhanced devices can be significant. Alaloul et al. proposed a plasmon-enhanced graphene all-optical modulator designed for integration into the silicon-on-insulator (SOI) platform [97]. The schematic of the proposed device and the electric field distribution are shown in Figure 3g and Figure 3h, respectively. The device’s performance metrics included switching energy, ER, IL, and operational speed. Theoretically, it achieved ultrafast switching (<120 fs) with low energy consumption (<0.6 pJ) and operated at an ultra-high bandwidth exceeding 100 GHz. Simulation results demonstrated a high ER of 3.5 dB over a 12 μm length, corresponding to a modulation efficiency of approximately 0.28 dB/μm. Furthermore, the modulator exhibited a record-low IL of ~6.2 dB among plasmon-enhanced graphene all-optical modulators reported to date [97].

4.4. WG Lasers

WG lasers are advanced compact light sources ideal for seamless integration into diverse photonic systems [98]. Innovative techniques in fabricating low-loss WGs within laser crystals of varying geometries have established economical platforms for generating solid-state lasers in compact form factors [99]. By harnessing high intracavity optical intensities, WG lasers achieve remarkably low lasing thresholds while maintaining efficiencies comparable to traditional bulk and fiber laser systems. Recent breakthroughs in photonic structure design have enabled customized beam shaping capabilities, including beam splitting and ring-shaped transformations, thereby significantly enhancing the versatility and applicability of WG lasers [100]. These attributes firmly establish WG lasers as highly promising miniature light sources tailored for PICs [101,102,103].

Figure 3. (a) Graphical illustration of the sensor layout [71]; (b) Cross-sectional view of a sensing patch, illustrating the WG, isolation layer, and surface chemistry location [71]; (c) photographic image showing light propagation along the WG chip [71], (d) schematic of a broadband NV hybrid EO plasmonic switch. The inset shows a cross-sectional view of the plasmonic slot WG coated with GST [90]. Mode profiles of the PCM-coated plasmonic slot WG at a telecom wavelength of 1550 nm: (e) amorphous phase and (f) crystalline phase, calculated for TE polarization. Geometric parameters used for mode profile calculations: metal height “h” = 220 nm, slot width “w” = 180 nm, PCM layer thickness “t” = 20 nm [90], (g) all-optical modulator utilizing a Si rib WG structure [97], (h) E-field profile of the propagating quasi-TM mode within the WG-integrated modulator, with the dashed white line indicating the graphene sheet plane [97], (i) experimental setup schematic for passively Q-switched laser operation in Nd:YVO₄ superficial cladding WG. The inset features a microscope image of the laser-inscribed WG [104], (j) room-temperature fluorescence emission spectra (μPL) depicting transitions from ⁴F_3/2 → ⁴I_9/2 and ⁴F_3/2 → ⁴I_11/2 states of Nd³⁺ ions. Data obtained from the laser-inscribed WG volume is shown in the red solid line, while bulk Nd:YVO₄ crystal emission is represented by the blue solid line. The inset displays the 2D spatial distribution of Nd³⁺ emission intensity captured from the WG’s end-face, corresponding to the region marked within the red dashed square in the microscope image insert (i) [104].

Figure 3. (a) Graphical illustration of the sensor layout [71]; (b) Cross-sectional view of a sensing patch, illustrating the WG, isolation layer, and surface chemistry location [71]; (c) photographic image showing light propagation along the WG chip [71], (d) schematic of a broadband NV hybrid EO plasmonic switch. The inset shows a cross-sectional view of the plasmonic slot WG coated with GST [90]. Mode profiles of the PCM-coated plasmonic slot WG at a telecom wavelength of 1550 nm: (e) amorphous phase and (f) crystalline phase, calculated for TE polarization. Geometric parameters used for mode profile calculations: metal height “h” = 220 nm, slot width “w” = 180 nm, PCM layer thickness “t” = 20 nm [90], (g) all-optical modulator utilizing a Si rib WG structure [97], (h) E-field profile of the propagating quasi-TM mode within the WG-integrated modulator, with the dashed white line indicating the graphene sheet plane [97], (i) experimental setup schematic for passively Q-switched laser operation in Nd:YVO₄ superficial cladding WG. The inset features a microscope image of the laser-inscribed WG [104], (j) room-temperature fluorescence emission spectra (μPL) depicting transitions from ⁴F_3/2 → ⁴I_9/2 and ⁴F_3/2 → ⁴I_11/2 states of Nd³⁺ ions. Data obtained from the laser-inscribed WG volume is shown in the red solid line, while bulk Nd:YVO₄ crystal emission is represented by the blue solid line. The inset displays the 2D spatial distribution of Nd³⁺ emission intensity captured from the WG’s end-face, corresponding to the region marked within the red dashed square in the microscope image insert (i) [104].

Nie et al. presented groundbreaking findings on room-temperature subnanosecond WG laser operation at 1064 nm using a Nd:YVO₄ crystal WG, employing Q-switching facilitated by phase-change nanomaterial vanadium dioxide (VO₂) [104]. VO₂’s capability to transition between insulating and metallic phases provided crucial low-saturation-intensity nonlinear absorption properties, essential for generating subnanosecond pulses. The low-loss WG was fabricated via femtosecond (fs) laser writing with a depressed cladding geometry. Under optical pumping at 808 nm, efficient pulsed laser emission was achieved in the Nd:YVO₄ WG, with a minimum pulse duration of 690 ps and a maximum output average power of 66.7 mW.

Comparative studies involving Q-switched lasers utilizing two-dimensional (2D) materials like graphene and transition metal dichalcogenides (specifically WS₂) coated mirrors revealed that while these materials can also generate 1064 nm laser pulses in the same WG platform, they typically produce shorter pulses lasting around 22 ns, with higher maximum output average powers reaching approximately 161.9 mW. This investigation underscored the distinct lasing properties between PCMs such as VO₂ and 2D materials, emphasizing VO₂’s potential as a cost-effective saturable absorber for subnanosecond laser applications [104].

Figure 3i illustrates the process of generating room-temperature Q-switched WG lasers, accompanied by an optical microscope cross-sectional image of the WG inset. The WG structure comprised an unmodified core encircled by a major arc geometry of parallel low-index tracks, which formed a superficial cladding shape at the sample–air interface. The WG, with a diameter of 50 μm and a lateral separation of 3 μm between adjacent damage tracks, was designed to minimize intracavity losses and support both TM and TE polarizations. This configuration was advantageous for constructing fiber–WG–fiber integrated photonic chips with high coupling efficiency. Figure 3j compares the μ-PL emission spectra of Nd³⁺ ions related to the ⁴F_3/2 → ⁴I_9/2 and ⁴F_3/2 → ⁴I_11/2 transitions from the bulk material and the WG core. The spectra demonstrated similar fluorescence properties, indicating well-preserved Nd³⁺ ion characteristics suitable for integrated laser sources. The spatial distribution map of emission intensity in the inset of Figure 3j shows localized modifications induced by fs laser pulses, forming microstructural amorphized regions and reducing the refractive index to create a low-index cladding that confines light within the WG structure. This map confirmed the efficient preservation of fluorescence in the WG volume [104].

5. Material Platforms for Integrated Photonics

Several material platforms are utilized in the development of integrated photonics, each offering exceptional advantages tailored to specific applications [105]. Si photonics, leveraging the mature semiconductor industry, employs Si as the substrate due to its compatibility with complementary metal oxide semiconductor (CMOS) fabrication processes, enabling cost-effective large-scale integration [106,107]. Si₃N₄ provides low optical loss and compatibility with visible and near-infrared wavelengths, making it suitable for applications requiring high performance and broadband operation [14]. III-V compound semiconductors, such as InP and gallium arsenide (GaAs), offer superior optical properties and are frequently used for light sources, detectors, and modulators in high-speed communication systems. Meanwhile, hybrid integration techniques combine multiple materials, such as Si and III-V materials, to harness the benefits of each, enabling the realization of complex photonic circuits with enhanced functionality and performance [108].

Germanium–Tin (GeSn) alloys have garnered significant attention due to their potential for monolithic integration with CMOS technology, offering promising avenues for electronic and optoelectronic applications [109,110]. Previous research has focused on optimizing growth techniques, such as molecular beam epitaxy (MBE) and chemical vapor deposition (CVD), to achieve high-quality GeSn films with uniform composition and minimized defects [111]. These efforts have encountered several challenges, including managing the high lattice mismatch between GeSn and Si substrates, controlling Sn segregation, and mitigating strain-induced defects. Additionally, maintaining the thermal stability of GeSn during post-growth processing steps is crucial for device performance [112]. Researchers have explored various approaches to overcome these obstacles, such as using buffer layers, optimizing growth parameters, and implementing advanced annealing techniques [113]. Despite these challenges, the continued advancements in GeSn growth techniques underscore its potential as a CMOS-compatible material, paving the way for its integration into next-generation semiconductor devices.

Organic materials and polymers are also explored for their flexibility, low-cost fabrication, and compatibility with flexible substrates, enabling applications in wearable devices and biochemical sensing [114]. Flexible integrated photonics represents photonic devices manufactured on pliable polymer substrates, resilient to mechanical deformations such as bending, folding, rolling, twisting, stretching, or compression, all while upholding their optical competency [115]. The hallmark of these devices lies in their mechanical adaptability, unlocking a realm of pioneering applications. The applications of this technology include flexible imaging and display arrays, short-reach optical links, wearable photonic textiles, solar cells, broadband tunable photonic devices, strain gauges, and optical systems that are capable of seamless incorporation with curved surfaces or biological tissues. Moreover, the incorporation of flexible substrates potentially revolutionizes manufacturing, paving the way for cost-effective, large-scale device production through roll-to-roll (R2R) processing [116]. Overall, the diverse range of material platforms in integrated photonics permits tailored solutions to meet the demands of various applications, spanning telecommunications, computing, sensing, and beyond. The information presented in

Table 2. General comparison of different material platforms.

Material Platform	Si	Si3N4	InP	GaAs	Lithium Niobate (LiNbO3)
Material type	Semiconductor	Dielectric	Semiconductor	Semiconductor	EO crystal
Refractive index	~3.5 [117]	~2 [117]	~3.2 [117]	~3.3 [117]	~2.2 [117]
Transparency range	1.1 µm to 9 µm	0.4 µm to 5 µm	1 µm to 2.6 µm	0.9 µm to 1.8 µm	0.35 µm to 5 µm [118]
Nonlinear coefficient	Low	Low	High	High	Very high [119]
WG loss	Low	Low	Very low	Low	Low
Fabrication method	CMOS-compatible [120]	CMOS-compatible [120]	Epitaxial growth	Epitaxial growth	Bulk crystal growth
Applications	Optical interconnects, sensors [121,122]	Filters, sensors, modulators [12,123,124]	Lasers, detectors [125,126]	Lasers, detectors [127,128]	Modulators, switches [129,130]

offers a general comparison of typical characteristics and applications of different material platforms. However, actual performance may vary depending on specific implementations, fabrication techniques, and device designs.

Integrated photonics has witnessed a surge of interest in 2D materials [131,132,133,134] owing to their remarkable optical characteristics and adaptability due to their extraordinary applications and diverse synthesis methods (Figure 4). Graphene [29,135,136], transition metal dichalcogenides (TMDs) [137], and hexagonal boron nitride (hBN) [138,139] represent some of these materials, offering distinctive benefits for on-chip photonic applications. Their ultrathin structure allows precise manipulation of light–matter interactions and facilitates the control of optical signals at the nanoscale. Moreover, 2D materials exhibit outstanding optical nonlinearities, empowering efficient modulation, switching, and frequency conversion functionalities [140]. Their seamless integration with standard fabrication methods and compatibility with traditional photonic platforms make them particularly appealing for the development of compact and efficient photonic devices.

Tian et al. presented a novel van der Waals PN heterojunction PD, comprising p-type black phosphorous and n-type molybdenum telluride, seamlessly integrated into a Si₃N₄ WG (Figure 5a) [141]. Figure 5b illustrates the energy band diagrams of the black phosphorous/molybdenum ditelluride van der Waals heterostructure in both non-equilibrium (top panel) and thermal equilibrium states (bottom panel). To accurately assess the intrinsic responsivity and internal quantum efficiency of the PD, it was essential to quantitatively extract the absolute light absorption by the black phosphorous/molybdenum ditelluride heterostructure in the waveguiding mode. This was achieved using a Mach–Zehnder interferometer (MZI) based on the Si₃N₄ WG, depicted in Figure 5c. The intrinsic electric field within the PN heterojunction notably reduced the dark current and enhanced responsivity. When a 1 V bias was applied from the n-type molybdenum telluride to the p-type black phosphorous, the dark current remained below 7 nA, which was over two orders of magnitude lower than that of other WG-integrated black phosphorus PDs. This configuration achieved an intrinsic responsivity of up to 577 mA W⁻¹. Notably, the van der Waals PN heterojunction could be further optimized through electrostatic doping, enhancing its rectification properties and boosting the responsivity to 709 mA W⁻¹. Additionally, the PD demonstrated a response bandwidth of approximately 1.0 GHz and maintained consistent photodetection across a broad spectral range, from 1500 to 1630 nm, as verified experimentally [141]. The integration of this van der Waals PN heterojunction PD on a chip, featuring low dark current, high responsivity, and rapid response, offered significant potential for developing efficient on-chip PDs suitable for various photonic integrated circuits based on Si, LN, polymers, and other materials.

Apart from conventional material platforms, the plasmonic platform stands at the forefront of integrated photonics, leveraging the distinct characteristics of surface plasmons to manipulate light at the nanoscale [142,143,144,145]. Through the confinement of light within subwavelength dimensions, plasmonic structures grant unprecedented control over interactions between light and matter, opening pathways for diverse applications in sensing, imaging, and telecommunications [4]. Serving as foundational elements, plasmonic WGs, resonators, and antennas enable the creation of compact photonic circuits capable of efficiently routing and processing optical signals with remarkable speed and efficacy [146]. Also, the capacity to engineer plasmonic materials and structures offers refined tunability, facilitating dynamic adjustments of light properties to accommodate on-demand functionalities. Despite its immense potential, the plasmonic platform for integrated photonics also encounters certain limitations that warrant consideration [147]. One primary challenge lies in the inherent ohmic losses associated with plasmonic materials and structures, stemming from absorption and scattering mechanisms, which can significantly degrade device performance, particularly over longer propagation distances [148]. Likewise, the intricate fabrication processes required for achieving precise control over plasmonic features can be complex and costly, hindering scalability and widespread adoption.

6. Fabrication Techniques and Challenges

Integrated photonics utilizes a range of fabrication techniques to create compact and highly efficient photonic devices as shown in Figure 6 [105,149]. Among these methods, lithography emerges as a fundamental approach, facilitating the precise formation of optical elements on semiconductor substrates. Noteworthy examples include photolithography, electron beam lithography (EBL) [150], and nanoimprint lithography (NIL) [151,152], each offering unique benefits in terms of resolution and scalability. Additionally, techniques such as ion implantation and plasma etching are instrumental in shaping WGs and fine-tuning material properties with exceptional accuracy. Moreover, epitaxial growth methods like MBE [153] and metal–organic chemical vapor deposition (MOCVD) [154] are essential for depositing semiconductor layers with customized optical and electrical properties.

Bonding techniques play a critical role in integrated photonics, facilitating the seamless integration of diverse materials and components onto a single photonic chip. These techniques are essential for ensuring efficient light propagation, minimizing losses, and maintaining alignment precision between various photonic elements [155]. Effective bonding enables the combination of different functional materials, such as Si, InP and LN, each offering unique optical properties [156]. Etching techniques are vital in integrated photonics as they enable the precise structuring of photonic materials, crucial for fabricating components like WGs, resonators, and photonic crystals. These techniques ensure high-resolution definition and smooth sidewalls, which are essential for minimizing optical losses and enhancing device performance. By using various etching methods, such as reactive ion etching (RIE) and wet etching, manufacturers can create intricate photonic structures with high accuracy and repeatability [157,158]. This precision is fundamental for developing complex photonic circuits and devices that are integral to advanced applications in optical communication, sensing, and signal processing.

Integrated photonics holds immense promise for advancing various fields, yet its progression is impeded by several adversities. A significant obstacle is the development of compact, efficient light sources and detectors suitable for integration on a chip scale, essential for enabling intricate photonic functionalities. Fabricating lasers directly on a Si platform has been challenging due to its intrinsic properties. Si, while widely used in electronics, is an indirect bandgap semiconductor, meaning it inefficiently emits light when excited. This property severely limits its ability to generate laser light efficiently. Additionally, integrating diverse photonic elements like WGs, modulators, and detectors demands precise fabrication techniques and material compatibility, often presenting manufacturing complexities. Another obstacle involves minimizing light propagation losses and maintaining high signal fidelity, particularly at the nanoscale. Additionally, issues with standardization and scalability persist, limiting the widespread adoption and commercialization of integrated photonic technologies.

Lowering the cost of PICs poses significant challenges despite their potential for revolutionizing various industries [159,160]. One major obstacle lies in the complexity of the fabrication processes involved, which often demand specialized equipment and expertise, driving up production costs. In addition, the limited availability of materials suitable for photonic applications further complicates cost reduction efforts. Furthermore, achieving economies of scale remains a challenge due to relatively low production volumes compared to conventional electronic integrated circuits. The intricacies of PIC design and the need for stringent quality control add to manufacturing costs. Overcoming these hurdles requires interdisciplinary collaboration across materials science, nanotechnology, photonics, and engineering to fully harness integrated photonics’ potential in transforming computing, communications, sensing, and beyond [107].

The development of low-cost integrated photonic devices has been significantly advanced through the utilization of the sol–gel method combined with dip-coating processes and NIL [161,162]. The sol–gel method allows for the synthesis of high-quality photonic materials at relatively low temperatures, making it cost-effective and versatile [163]. By employing the dip-coating process, these materials can be uniformly deposited onto substrates, ensuring consistent film thickness and quality as shown in Figure 7 [164]. NIL further enhances this approach by providing a precise, high-throughput technique for patterning nanoscale features onto the sol–gel films. This synergy of methods facilitates the production of intricate photonic structures with high precision and scalability, driving down manufacturing costs while maintaining performance and reliability [159,160,165]. Consequently, this integrated approach is paving the way for widespread adoption and innovation in the field of photonics. However, there are still some obstacles that need to be addressed [166].

7. Role of Artificial Intelligence in the Development of Integrated Photonics

Artificial intelligence (AI) was developed to emulate human cognitive functions such as learning, reasoning, and problem-solving, aiming to enhance efficiency, automate repetitive tasks, and solve complex problems across various industries [167,168]. The inception of AI dates back to the mid-20th century, with significant milestones including Alan Turing’s conceptual framework and the development of the first AI programs in the 1950s and 1960s [169,170]. AI’s application in photonics—a field concerned with the generation, manipulation, and detection of light—began gaining traction in the early 21st century. By leveraging AI, researchers have been able to optimize photonic devices and systems, enhance imaging techniques, and improve the design and fabrication processes. This integration has led to advancements in telecommunications, medical imaging, and quantum computing, among other areas.

AI and machine learning (ML) algorithms stand poised to revolutionize integrated photonics by enhancing device design, optimization, and control processes [171]. With its ability to analyze vast amounts of data and complex systems, AI can optimize the performance of photonic devices, leading to unprecedented efficiency and functionality. ML algorithms can aid in designing intricate PICs by swiftly exploring a vast design space and identifying optimal configurations [172]. In addition, AI-driven monitoring and control systems can dynamically adjust parameters in real time, ensuring optimal performance even in varying environmental conditions. This fusion of AI and integrated photonics holds the promise of accelerating the development of next-generation photonic technologies, empowering faster data transmission, improved energy efficiency, and enhanced functionality across various applications, from telecommunications to sensing and computing [56,173].

Artificial neural networks (ANNs) offer a robust approach to photonic design by establishing an implicit relationship between input parameters (such as geometry and materials) and output optical responses, emulating the nonlinear nerve conduction in the human body [174,175,176]. Well-trained ANNs streamline the complex and time-consuming design process, which typically depends on numerical simulations and optimization. ANN models serve two primary functions in photonic design: forward prediction and inverse design. The forward prediction network estimates optical responses from geometric/material parameters, acting as a substitute for comprehensive wave simulations. The inverse design network efficiently identifies the optimal structure based on given optical responses, addressing a critical and challenging aspect of the design process. A key benefit of ANN models is their high processing speed, which is particularly advantageous in situations requiring rapid calculations. As demonstrated by the example of a meta-atom spectrum, a forward prediction model can generate this result in a mere few milliseconds. This is a notable improvement over the processing times typically associated with traditional full-wave simulation methods, such as finite element (FEM) and finite difference time domain (FDTD) approaches [177]. Additionally, ANN models maintain accuracy comparable to rigorous simulations, with mean squared loss of spectrum prediction typically ranging from 10⁻³ to 10⁻⁵ [178]. Furthermore, ANNs can elucidate hitherto unanticipated and non-unique relationships between physical structures and optical responses. This may prompt researchers to devise entirely novel classes of structures. Hammond et al. presented a practical ANN design framework specifically tailored for devices that serve as fundamental components in PICs [172]. The design methodology is presented in Figure 8a. Through experimental validation, the framework’s effectiveness was showcased by training ANNs to model strip WGs and chirped Bragg gratings (BGs). This approach employed a minimal set of straightforward input and output parameters pertinent to photonic circuit designers. Once trained, these ANNs significantly reduce the computational cost, surpassing traditional design methodologies by a factor of 10,000. To demonstrate the capabilities of our innovative design paradigm, both forward and inverse design tools powered by the ANN were demonstrated. Utilizing these tools, several integrated BG devices were designed and fabricated for practical use in photonic circuits. The ANN’s predictions align closely with experimental measurements, achieving high accuracy without the need for any post-fabrication training adjustments. Figure 8b–e provides a comparison between the ANN’s predictions and the measured data, highlighting the accuracy and reliability of the neural network model in predicting the performance of the fabricated devices [172].

8. Potential Challenges in the Development of PICs

The development of integrated photonic devices presents several significant challenges, including material compatibility, fabrication precision, thermal management, and scalability. Addressing these issues is crucial for realizing the full potential of PICs and ensuring their widespread adoption in various applications. One of the primary challenges in integrated photonics is material compatibility [179]. Photonic devices often require materials with specific optical properties, such as high refractive indices or efficient light emission. Si, the material of choice in the semiconductor industry, exhibits excellent electronic properties but has limitations in optical applications, particularly in light emission. To overcome this, researchers are exploring hybrid integration techniques that combine Si with other materials like III-V semiconductors (e.g., indium phosphide and gallium arsenide), which are efficient light emitters [180,181,182,183]. The hybrid approach of Si with III-V semiconductors leverages the mature Si fabrication infrastructure while incorporating the superior optical properties of III-V materials [184]. Advancements in bonding techniques and the development of new material systems, such as Si₃N₄ and polymers, also contribute to addressing material compatibility challenges [185].

Achieving the required precision in the fabrication of photonic devices is another significant hurdle [186]. Photonic structures, such as WGs and resonators, require sub-wavelength accuracy in their dimensions to ensure proper light confinement and propagation [187]. Variations at the nanoscale can lead to significant performance degradation. To address this, the photonics industry is investing in advanced lithography and etching techniques, including deep ultraviolet (DUV) and EBL, which offer the necessary resolution and accuracy [188,189]. Additionally, process control and metrology tools are being enhanced to monitor and correct deviations during fabrication. The integration of AI and ML in the manufacturing process can further improve precision by predicting and compensating for process variations.

Thermal management is a critical challenge in integrated photonics, especially for devices that operate at high power or require precise temperature control [190]. The TO effect, where temperature changes affect the refractive index, can lead to performance instability. Effective thermal management strategies are essential to maintain device performance and reliability [191]. Solutions include the use of heat sinks and thermoelectric coolers to dissipate heat, as well as designing photonic circuits to minimize thermal sensitivity. Researchers are also exploring materials with lower TO coefficients and developing innovative packaging techniques that enhance thermal conductivity while protecting the delicate photonic components [192,193].

Scalability and the seamless integration of photonic and electronic components remain formidable challenges [194,195]. As the demand for complex and multifunctional PICs grows, integrating a large number of photonic devices on a single chip becomes increasingly difficult. This integration requires not only advanced design tools that can handle the complexity but also fabrication processes that can produce high yields. To overcome these challenges, the industry is adopting standardized photonic platforms and design automation tools that streamline the development process. Si photonics, for instance, benefits from leveraging CMOS fabrication techniques, enabling large-scale production and integration with electronic circuits [196]. The development of standardized photonic building blocks and libraries also facilitates scalable design and manufacturing [197].

Optical loss and fiber coupling present significant challenges in the integration of PICs with current optical fibers used in telecommunications [198]. One primary issue is the mismatch between the mode field diameters of optical fibers and the much smaller WGs in PICs, which can lead to high coupling losses [199]. Material compatibility also poses challenges, as the refractive index differences between Si-based PICs and glass optical fibers can cause reflection and scattering losses at the interface [200]. Additionally, the physical alignment precision required to couple fibers to PICs efficiently is stringent, necessitating advanced packaging and alignment techniques to minimize insertion loss. Ensuring thermal stability and mechanical robustness in the coupling interface is crucial to maintaining performance over the operational lifespan [201].

Continuous investment in research and development is crucial, as is the adoption of advanced simulation and modeling tools that can predict the behavior of photonic devices with high accuracy. The integration of AI and ML in the design and manufacturing processes can further enhance efficiency and performance, enabling the rapid development of next-generation photonic devices [202,203,204]. Furthermore, establishing industry standards and fostering collaboration across the photonics ecosystem will drive the adoption of best practices and accelerate technological advancements. By addressing these challenges through innovation and collaboration, the development of integrated photonic devices can progress, unlocking new opportunities and applications in telecommunications, computing, sensing, and beyond.

9. Concluding Remarks

The future of integrated photonics is poised to be transformative, with significant advancements anticipated across a range of high-impact areas. As the technology continues to mature, we can expect even greater integration of photonic and electronic components, leading to more compact, energy-efficient, and efficient devices. Innovations in materials science, such as the development of new photonic materials and advanced fabrication techniques, will enhance the functionality and scalability of PICs. In telecommunications, integrated photonics will support the exponential growth of data traffic and the deployment of next-generation networks like 6G, enabling faster and more reliable connectivity. In computing, photonic chips will revolutionize data centers and efficient computing by overcoming the bandwidth and thermal limitations of electronic interconnects. As the ecosystem around integrated photonics expands, including more widespread adoption and standardization, it will pave the way for innovative applications and drive digital transformation across industries, ultimately shaping a smarter and more connected world.