Micro–nano surface engineering and property modulation: insights from black silicon for advanced material applications

Abstract

The advancement of micro–nano fabrication techniques has enabled the precise modification of material surfaces, allowing for unique physical and chemical properties to emerge, catering to diverse application requirements. Black silicon, a well-studied material compatible with CMOS technology, provides an excellent model for exploring the overall properties of micro–nano surface morphologies. This review systematically investigates the morphology fabricated by five mainstream synthesis methods for black silicon, offering an in-depth analysis of the underlying mechanisms and the impact of key parameters on surface morphology. In addition to structural features, this review discusses the formation of surface chemical bonds and their roles in modulating surface energy, wettability, and electronic passivation. Through this exploration, we examine how specific morphologies yield distinct physical and chemical properties that drive a wide range of applications, from photodetection to sensing and biomedical technologies. Key challenges of black silicon are discussed, including cost, mass production, and defect control, etc. This review aims to serve as a roadmap for researchers, guiding further advancements not only in black silicon but also in the broader field of nanomaterials, paving the way for breakthroughs across various domains of modern nanotechnology.

---

Junling Lv

Junling Lv received her bachelor's degree in Materials Science and Engineering from Sichuan University in 2023. She is continuing to pursue her PhD degree at Sichuan University. Her research interests are mainly focused on the alternating current photovoltaic effect and its applications.

Haiyang Zou

Dr Haiyang Zou is currently a Professor at the College of Materials Science and Engineering at Sichuan University. Holding a PhD degree from the esteemed Georgia Institute of Technology, USA, he furthered his expertise through postdoctoral research there. Dr Zou's academic journey has led him to specialize in a wide array of areas including nanomaterials, nano-systems, and nano-devices, as well as photonics, piezo-electronics, piezo-phototronics, triboelectric nanogenerators, and flexible electronics. He was recognized as one of the world's most influential researchers in his field based on the databases of standardized citation indicators (the world's top 2% scientists named by Stanford University).

Zhong Lin Wang

Zhong Lin Wang is the Hightower Chair in Materials Science and Engineering, Regents' Professor, and College of Engineering Distinguished Professor at the Georgia Institute of Technology. He is now the director of the Beijing Institute of Nanoenergy and Nanosystems. Prof. Wang has made original and innovative contributions in the fields of oxide nanobelts and nanowires, piezotronics and piezo-phototronics as well as their applications in energy sciences, electronics, optoelectronics and biological science. Dr Wang was elected as a member of the Chinese Academy of Sciences and European Academy of Sciences, an academician of the Academia of Sinica (Taiwan), and an international fellow of the Canadian Academy of Engineering.

---

1. Introduction

The rapid progression of micro–nano fabrication technologies has ushered in transformative capabilities for tailoring material surfaces to enhance or modify physical, chemical, and electronic properties.^1–3 These advancements play a vital role in developing materials for applications that demand high specificity, sensitivity, and adaptability. Among various materials, silicon has long been a cornerstone of the electronic and optical device industries, primarily due to its compatibility with CMOS technology, robust mechanical properties, and well-understood electronic characteristics. As the demand for more efficient and versatile devices grows, the limitations of conventional silicon structures become increasingly apparent.^4–7

Black silicon is typically structured with nanometer- to micrometer-scale features,⁸ such as holes, needles, and columnar formations, and it exhibits properties not found in conventional silicon surfaces. The advent of black silicon, an optically and chemically modified form of silicon, has expanded silicon's applicability beyond conventional electronics. Black silicon demonstrates exceptional photon capture and confinement abilities, owing to its significantly increased active surface area and quantum confinement effects.^9–11 This transition from a traditional, smooth surface to one rich with nanoscale and microscale structures fundamentally transforms silicon's interaction with light, fluids, and biological elements. Its unique surface morphology, often achieved through tailored micro–nano structuring, imbues black silicon with extraordinary properties such as enhanced light absorption,^12,13 superhydrophobicity,^14,15 and augmented chemical reactivity.^16–18 These characteristics make black silicon particularly suitable for a range of applications in photodetection,^19–21 solar cells,^22,23 sensors,^24–26 biomedical devices,^27–32 and environmental monitoring.³³ However, realizing these properties and applications hinges significantly on the precise control of black silicon's morphology,^34,35 which can vary substantially depending on the synthesis method employed. Therefore, it is crucial to study the surface morphology of black silicon and explore how morphological features influence various material and device performances.

In recent years, there has been a growing trend of research articles and reviews focusing on black silicon.^8,9,36–40 Existing reviews have primarily focused on fabrication innovations and specific applications. Liu et al. summarized the fabrication methods and photoelectric properties of black silicon, as well as the applications and improvements in solar energy utilization.⁸ Martin et al. provided guidelines for optimising the relevant parameters for better optoelectronic applications.⁴¹ Zhao et al. offered an in-depth review of the morphology and mechanisms of laser etching to fabricate black silicon, with a focus on its applications and future directions in infrared detection.³⁹ Other reviews have explored topics such as biomaterials,⁴² battery anodes^43,44 and antibacterial surfaces.⁴⁰ However, a comprehensive analysis of black silicon's surface morphology remains insufficient, especially its evolution under different fabrication strategies and its direct influence on material properties (e.g., physical, chemical, and biological properties). The surface architecture, including micro- and nanoscale features such as spikes, pores, and trenches, critically affects light trapping, carrier dynamics, and interfacial behaviors. Moreover, the formation of specific chemical bonds (e.g., Si–H, Si–O, and Si–F) during fabrication also modulates surface energy and states, leading to various properties. Therefore, a systematic review that unifies the morphological characteristics of black silicon with its fabrication-dependent chemical and physical properties is essential. Such a perspective is crucial for understanding its fundamental mechanisms and for guiding the rational design of high-performance black silicon-based devices across diverse applications.

In this review, we aim to provide a comprehensive roadmap for future advancements in micro–nano surface engineering. By examining the relationship between morphology and functionality, we seek to deepen the understanding of how surface morphology and chemical states can be tailored to achieve desired material properties and meet specific application requirements. This paper begins by tracing the evolution of fabrication methods, followed by a detailed analysis of key fabrication techniques, such as chemical etching, reactive ion etching (RIE), metal-assisted chemical etching (MACE), laser etching, and plasma immersion ion implantation etching (PIIIE), to elucidate the mechanisms and factors influencing the morphology of black silicon. This review also investigates how surface chemical bonding states formed during fabrication, in conjunction with morphological features, collectively influence surface energy, surface physicochemical properties, and interfacial electronic characteristics. The interplay between morphology and surface chemistry is critical for optimising light trapping, carrier dynamics, and environmental stability. By correlating fabrication methods, surface morphology, chemical terminations, and material properties, this review provides key insights for the rational design of black silicon devices and identifies both the challenges and opportunities for their broader application. The scope and structure of this review are depicted in Fig. 1. This comprehensive analysis serves as a valuable resource for researchers and engineers seeking to navigate the complexities of black silicon fabrication and offers practical guidance for overcoming the barriers to its broader application. Moreover, our analysis not only applies to black silicon but may also inform research on other nanostructured materials, ultimately contributing to the broader field of nanotechnology and its myriad applications.

Fig. 1 The wide applications of black silicon: photodetectors,⁴⁵ solar cells,⁴⁶ PTE conversion,⁴⁷ luminescence,⁴⁸ photocatalysis,⁴⁹ THz emission,⁵⁰ SERS,⁵¹ and antibacterial.³⁴

2. Micro–nano engineering strategies for morphology control

The fabrication history of black silicon can be traced back to 1959.⁵² At that time, silicon was immersed in an etchant solution consisting of hydrofluoric acid (HF), nitric acid (HNO₃), and water (H₂O) at 25 °C to study the kinetics of silicon etching under acidic conditions. In 1995, a significant milestone was achieved when RIE was employed to fabricate black silicon. The term “black silicon” was coined to describe the distinctly darkened surface of the etched silicon.⁵³ The resulting morphology consisted of grass-like nanostructures formed on the silicon surface. Based on chemical etching, the metal particles were introduced to accelerate the etching process, and more regular structures could be achieved by this method, which was named MACE.⁵⁴ It wasn’t until 1998 that Eric Mazur et al. prepared sulfur-doped black silicon by femtosecond laser irradiation under a sulfur hexafluoride (SF₆) atmosphere,⁵⁵ expanding the light absorption range beyond 1100 nm. After doping sulfur to form a supersaturated doped silicon, the introduced impurities formed a new energy band structure, effectively broadening the light absorption range. Another doping method to fabricate black silicon, PIIIE appeared in 2011,²² the needle-like surface reduced the average reflectance to 1.79% in the range of 300–1100 nm. Some important improvement milestones of the five fabrication methods are shown in Fig. 2 and discussed as follows.

Fig. 2 The development timeline of various black silicon fabrication methods, including several key milestones.^22,52–61

Porous silicon was first discovered at Bell Laboratories in 1956 through electrochemical etching,⁶² but initially garnered little interest due to its defective surface properties. A decade later, in 1966, Memming et al. investigated the factors influencing silicon dissolution, such as electrode potential, crystal doping, and HF concentration,⁵⁶ although the technique did not produce a regular surface morphology at the time. In 1959,⁵² etching with pure acid solution was introduced, marking a new approach. In 2005, the development of molten salts’ electrochemical etching enabled the direct electrolytic reduction of solid silicon dioxide (SiO₂) to silicon, resulting in more uniformly structured hexagonal silicon columns.⁵⁹ This method proved to be a more economical and efficient approach for fabricating black silicon.

RIE was first applied to silicon by Jansen in 1995,⁵³ demonstrating its versatility in creating nanograss and enabling the formation of vertical walls. This capability allows for precise control over silicon nanostructures through parameter adjustments. In 2012, sputtered silver nanoparticles (AgNPs) were introduced as micromasks,⁶³ replacing the SiO₂ layer that was previously formed by oxygen (O₂) and served as a random micro-mask. This substitution reduced the need for O₂ in the process, thereby minimizing damage and prolonging the lifespan of black silicon. Researchers also examined the effects of AgNPs and etching duration on surface characteristics. Nguyen et al. later conducted an in-depth study on how RIE parameters affect the surface morphology of black silicon,⁶⁴ establishing a relationship between optical properties and the aspect ratio of micro–nano structures, which further refined the controllability of RIE in black silicon fabrication. To create more regular microstructures, masks have been employed to control the spatial distribution of the etched features, enabling the development of hierarchical micro–nano hybrid structures that further enhance light absorption.⁶⁵ In recent advancements, black silicon has been successfully fabricated on curved surfaces,⁶⁶ and ultrabroadband absorption extending to 20 µm was achieved in 2023,⁶⁷ highlighting the promising future of RIE in black silicon technology.

MACE was developed by adding metal nanoparticles as catalysts during chemical etching, mostly noble metal nanoparticles. The initial MACE formed porous silicon by depositing aluminum (Al) thin films in 1997,⁵⁴ then the influence of different types of metal nanoparticles, silicon types, and doping levels was investigated by Li et al.⁶⁸ Koynov et al. demonstrated that black silicon prepared by MACE is independent of surface orientation, making the process applicable to various types of silicon, including monocrystalline, polycrystalline, amorphous, and thin films.⁶⁹ Similar to RIE, masks were also widely employed in MACE.⁷⁰ Metal nanoparticles were deposited on the unmasked regions, so micro–nanostructures are selectively formed in these areas. This approach facilitates the creation of regular and controllable nanowire arrays, enhancing the uniformity of the etched structures. The influence of annealing temperature on the morphology and optical properties of black silicon has also been concluded,⁷¹ showing clear benefits for the performance of photovoltaic devices. A recent study reported a gas-phase programmed MACE technique,⁶¹ which effectively addresses the long-standing issue of large-area etching nonuniformity and provides a new strategy for the further development of MACE technology.

One advantage of femtosecond and nanosecond laser irradiation over other methods is its ability to incorporate doping during preparation, effectively broadening the light absorption range in black silicon. Since the pioneering work of Mazur's group,⁵⁵ this technique has attracted significant research interest. Wu et al. expanded the light absorption range of femtosecond-laser-fabricated black silicon to 0.25 µm while maintaining over 90% absorption, and they further provided insights into the underlying infrared absorption mechanism.⁵⁸ Subsequent studies have explored various factors affecting laser etching, including background gas composition,^72,73 laser pulse numbers,⁷⁴ laser fluence and scanning speed,⁷⁵ as well as annealing and substrate temperatures.⁷⁶ In 2011, a direct femtosecond laser surface structuring method was developed to create a velvet-like surface morphology with outstanding antireflective properties.⁷⁷ Recognizing that thermal annealing can deactivate sub-bandgap light absorption, laser annealing has emerged as a promising alternative. It preserves high crystallinity while reactivating sub-bandgap absorption.⁶⁰ Additionally, combining femtosecond laser irradiation with other fabrication methods has been shown to significantly enhance the performance of black silicon.^78,79

PIIIE is another innovative method to fabricate black silicon. Although the mechanism was reported as early as 1996,⁵⁷ its application to silicon was not demonstrated until 2011.²² Since then, numerous studies have explored the factors affecting the process.^80,81 Lim et al. optimized the process by using environmentally friendly plasmas instead of SF₆ and O₂, which were commonly used in the past.⁸² This approach not only reduced environmental impact but also passivated surface defects and enabled the production of ultra-low reflectivity black silicon.

The subsequent sections will provide a detailed examination of each fabrication method, including the mechanisms involved and the various factors that influence surface morphologies.

2.1. Chemical etching

Chemical etching is an appealing technique widely used for black silicon fabrication, offering control over surface morphology and optical properties by selectively removing silicon through chemical reactions. This method includes several types, each with distinct characteristics, etching depths, and applications. In general, pure chemical etching in acidic or alkaline solutions produces relatively shallow texturing at the nanometer to submicronmeter scale, primarily forming random porous or pyramidal structures. Electrochemical etching, in contrast, enables the fabrication of porous silicon with well-defined pore networks that can extend to several micrometers in depth, making it more suitable for sensing and catalytic applications. Electrochemical etching in molten salts represents a more recent development and can produce even deeper and more uniform columnar structures through the high-temperature electrolytic reduction of SiO₂. These differences in etching depth and morphology directly influence the light-trapping ability and optoelectronic performance of black silicon. Therefore, it is important to distinguish the modalities when evaluating their potential applications.

2.1.1. Pure chemical etching. Pure chemical etching involves the use of acidic or alkaline solutions to etch the silicon surface without applying an external electric field. The influencing factors for pure chemical etching mainly include the types of etching solution, solution concentration, temperature, etching time, etc. For polysilicon etching, acidic solutions etch silicon isotropically, while alkaline solutions exhibit anisotropic behavior. This difference arises from whether the corrosion rate of each crystal direction is consistent.

In acidic chemical etching, HF and nitric acid (HNO₃) are commonly used. The etching process is a redox reaction,⁵² where HNO₃ oxidizes the silicon surface and HF subsequently dissolves the resulting oxides. The overall reaction for acidic etching can be expressed by eqn (2.1)

3Si + 12HF + 4HNO₃ → SiF₄ + 4NO + 8H₂O (2.1)

The acid etching approach produces a porous and uneven structure, characterized by micro- and nanoscale pits across the surface, which enhances light absorption by creating multiple reflective surfaces (Fig. 3(a)).⁸³ The balance between HF and HNO₃ concentrations dictates the etching dynamics. When HF concentration is high and HNO₃ concentration is low, the oxidation rate becomes the rate-limiting step, often resulting in the formation of porous silicon. In contrast, when HF concentration is low and HNO₃ concentration is high, the removal of oxides by HF governs the etching rate, making this regime more suitable for silicon surface polishing and the removal of surface damage.⁸ In addition, the SiO_x formed during the oxidation process can serve as an antireflection layer, further improving the antireflection effect.

Fig. 3 SEM images of black silicon morphologies produced by different chemical etching methods. (a) HF/HNO₃ acidic etching.⁸³ Reproduced with permission from ref. 83. Copyright 2009, Elsevier. (b) Alkaline KOH/IPA etching.⁸⁴ Reproduced with permission from ref. 84. Copyright 2021, MDPI. (c) Electrochemical etching.⁸⁵ Reproduced with permission from ref. 85. Copyright 2012, AIP Publishing. (d) Silicon columns obtained by molten CaCl₂ etching at 1123 K.⁵⁹ Reproduced with permission from ref. 59. Copyright 2005, IOP Publishing.

For alkaline chemical etching, solutions containing sodium hydroxide (NaOH), potassium hydroxide (KOH), isopropanol (IPA), or their mixtures are commonly used. The mechanism can be expressed by eqn (2.2) and (2.3) as follows. Furthermore, the etching rate and the resulting surface morphology can be effectively controlled by adjusting the concentrations of the alkaline components.

Si + 4OH⁻ → SiO₄²⁻ + 2H₂ (2.2)

SiO₄²⁻ + 4H₂O → Si(OH)₄ + 4OH⁻ (2.3)

Etching in an alkaline solution of KOH produces a more uniform and smoother surface than acidic etching, with structures that exhibit a consistent orientation. This etching process encourages the formation of pyramid-like structures due to the anisotropic action of KOH on silicon. These pyramidal features effectively reduce light reflection, making them ideal for photovoltaic applications. The alkaline solution enables the creation of two primary types of pyramidal structures on the silicon surface: upright pyramids, which form naturally through KOH etching (Fig. 3(b)),⁸⁴ and inverted pyramids, which generally require the addition of metal nanoparticles to guide their formation.⁸⁶ Vazsonyi et al. investigated the etching process in an inorganic alkaline solution containing low concentrations of NaOH and IPA.⁸⁷ The density and size of the upright pyramids are affected by the etching rate of silicon along the 〈100〉 direction and the anisotropy factor of the solution (i.e., the quotient of the etching rate in the 〈100〉 to 〈111〉 direction). Prolonged etching durations generally lead to an increase in the average pyramid height and contribute to enhanced structural uniformity. On the other hand, alkaline etching without IPA enables the thinning of silicon wafers without forming micro–nano structures,⁸⁸ which is beneficial for the fabrication of ultra-thin flexible devices.

To facilitate a concise quantitative comparison between acidic and alkaline chemical etching, representative characteristics are summarized in Table 1. Acidic etching is generally associated with higher, largely isotropic etching rates, giving rise to porous or pit-dominated surfaces with relatively high roughness and broad feature-size distributions. In contrast, alkaline etching is characterized by lower but strongly anisotropic etching behavior, typically exhibiting etching rate ratios between 〈100〉 and 〈111〉 directions exceeding one order of magnitude. This pronounced anisotropy enables the formation of well-defined pyramidal morphologies with improved surface uniformity and reduced optical reflectance, which are particularly advantageous for photovoltaic applications.

Table 1 Comparison of acidic and alkaline etching

Metric	Acidic chemical etching	Alkaline chemical etching
Etching anisotropy	Isotropic	Strongly anisotropic
Crystal-plane dependence	Weak	Strong (〈100〉 ≫ 〈111〉)
Characteristic morphology	Porous/pit-dominated	Upright or inverted pyramids
Feature size scale	Micro- to nanoscale mixed	Micrometer-scale
Surface roughness	Relatively high	Relatively low
Typical applications	Light trapping, sensing	Photovoltaics, wafer thinning

---

Despite the advantages of pure chemical etching, including its simplicity, good selectivity, and suitability for mass production, this method struggles to produce highly regular microstructures. As a result, its applications are somewhat constrained. To overcome these limitations, pure chemical etching is often combined with metal nanoparticles to achieve more precise structural control. This process will be discussed in detail in Section 3.3.

2.1.2. Electrochemical etching. Electrochemical etching involves the direct corrosion of silicon wafers, with a high-electrode-potential metal serving as the cathode and HF as the electrolyte. At the start of the process, silicon at the anode partially dissolves, following the reactions in eqn (2.4) and (2.5):

Si + 2H₂O → SiO₂ + H₂ + 2H⁺ + 2e⁻ (2.4)

SiO₂ + 6HF → 2H⁺ + SiF₆²⁻ + 2H₂O (2.5)

Electrochemical etching produces a porous silicon structure with well-defined pores and an interconnected network. The porosity and aperture can be tuned by controlling the etching parameters, such as the applied current and electrolyte composition. The resulting morphology is characterized by a high surface area, making it advantageous for applications requiring enhanced surface interactions, such as sensing and catalysis.

As the H⁺ concentration increases in the HF solution, the silicon oxidation rate decreases, allowing fluorine ions to etch the silicon surface layer by layer.⁸⁹ Additionally, the doping concentration and dopant type of silicon significantly influence the etched microstructure: antimony (Sb)-doped wafers produce high-density, tall, and narrow conical structures with deep clefts, resulting in a blacker surface appearance than phosphorus (P)-doped wafers, which create thicker cones with sharper tips.⁹⁰

Etching potential also influences silicon's etching characteristics.^56,91 At lower potentials, oxide formation is slower than oxide dissolution, thus preventing the formation of an oxide film. With an increase in potential, porous silicon formation accelerates, particularly at higher HF concentrations. At even higher potentials, oxidation dominates, and once the potential reaches peak current, a uniform oxide film coats the surface. This initiates a continuous cycle of oxide formation and dissolution, leading to uniform electropolishing across the silicon surface.

Conventional chemical etching typically produces irregular surfaces; however, by pre-treating specific areas on the silicon surface, the etching process can be controlled to form regular microstructures. In a study by Ao et al.,⁸⁵ a hexagonal pattern of inverted pyramidal pits was formed on n-type silicon by photolithography to guide the growth of holes. The aperture was adjustable by varying the current density at the silicon/HF interface for a given interpore distance and under fixed etching conditions. Scanning electron microscope (SEM) images of these structured etchings are shown in Fig. 3(c). Compared to pure chemical etching, electrochemical etching facilitates the formation of regular microstructures more easily and achieves a higher etching rate.

2.1.3. Molten salts electrochemical etching. Molten salts electrochemical etching is a more recent approach that enables the direct electrolytic reduction of the solid SiO₂ to silicon, offering advantages such as low environmental impact, cost-effectiveness, and high efficiency.⁹² This etching method produces more regular, columnar structures, resulting in hexagonally arranged silicon columns (as shown in Fig. 3(d)). This morphology is more uniform and structured than that formed by other chemical etching methods, with each column displaying a distinct vertical alignment. This regularity and columnar formation enhance light-trapping capabilities, rendering it suitable for high-performance optoelectronic applications.

Yasuda et al. investigated the mechanism of direct electrolytic reduction of SiO₂ in molten calcium chloride (CaCl₂),⁵⁹ where the total reaction can be represented in eqn (2.6):

SiO₂ + 4e⁻ → Si + 2O²⁻ (2.6)

In their experiment, SiO₂ was reduced to silicon in molten CaCl₂ at 1123 K under a voltage of 1.10 V for 1 hour, resulting in the formation of hexagonal silicon columns aligned perpendicularly to the silicon/SiO₂ reaction interface. The formation mechanism of these silicon columns was attributed to a volume contraction occurring during the SiO₂-to-Si reduction, which induced cracking. As the reduction process continued, additional SiO₂ was converted to Si, with cracks continuing to propagate perpendicularly to the reaction interface, segmenting the silicon into distinct columns. The overall reaction rate was governed by the diffusion of O²⁻ ions through the vacant spaces in molten CaCl₂, which were generated by the volume reduction during conversion. Using this approach, an anti-reflection coating was fabricated by coating the black silicon surface with titanium dioxide (TiO₂), resulting in an extremely black surface with a reflectivity of about 0.1%.⁹³ This study highlights the potential of molten salts electrochemical etching as an effective method for fabricating advanced black silicon structures with enhanced optical properties.

In summary, chemical etching offers versatile routes for black silicon fabrication, with each modality presenting unique advantages and limitations. Rather than a universally optimal approach, the choice of etching method enables a balance between structural uniformity, scalability, and optical performance. These distinctions not only guide the selection of suitable techniques for specific applications but also point to the potential of hybrid strategies for achieving more advanced black silicon architectures.

2.2. Reactive ion etching

RIE is a plasma-based etching method that enables the formation of various silicon surface morphologies by precisely controlling plasma parameters. RIE is typically performed within inductively coupled plasma (ICP) or capacitively coupled plasma (CCP) systems, shown in Fig. 4.⁹⁴ This technique can produce a range of structures, including isotropic surfaces, reverse tapers, and vertical walls.

Fig. 4 The equipment and mechanisms of RIE.⁹⁴ Reproduced with permission from ref. 94. Copyright 2017, AIP Publishing. (a) Schematic of the ICP and CCP system used for the RIE process. (b) Schematic view of RIE in SF₆/O₂ plasma.

In the initial application of RIE on silicon by Jansen et al., silicon wafers were etched to create grass-like microstructures using a plasma mixture of SF₆, O₂, and trifluoromethane (CHF₃), each gas playing a specific role in the etching process.⁵³ The primary functions include chemical etching, passivation, and selective layer removal, which together enable the formation of intricate micro- and nanostructures. The following is a breakdown of these functions and the associated reactions: (1) SF₆ produces fluorine radicals (F*), which chemically etch silicon by forming volatile sulfur tetrafluoride (SF₄). (2) O₂ generates oxygen radicals (O*), which passivate the silicon surface with a SiO_xF_y layer that forms at cryogenic temperatures and volatilizes at room temperature. This SiO_xF_y layer prevents the silicon from further chemical corrosion, although it can be dissociated by heat or plasma ion bombardment, enabling the formation of controlled, intricate surface structures. (3) CHF₃ produces CF_x⁺ ions that selectively etch the SiO_xF_y layer in one direction, creating volatile CO_xF_y compounds. (4) SF_x⁺ ions aid in removing the oxyfluoride layer by forming volatile SO_xF_y gases. The overall etching process follows these reactions in eqn (2.7)–(2.10):⁹⁵

Si + 4F* → SiF₄↑ (2.7)

SiF_y⁺ + xO* → SiO_xF_y (2.8)

SiO_xF_y + CF_x⁺ → CO_xF_y↑ (2.9)

SiO_xF_y + SF_x + → SO_xF_y↑ (2.10)

These functions work in concert to achieve the directional control, precision, and surface morphology adjustments that RIE offers, enabling the formation of diverse nanostructures. Maintaining a precise balance between chemical etching, passivation, and selective removal is essential for managing the etching profile and attaining a structure with a high aspect ratio.

The key influencing factors of RIE include radio frequency (RF) powers, etching time, gas composition, chamber pressure, and gas flow rate, each affecting the properties of black silicon in distinct ways.⁹⁶ ICP or RF power directly impacts surface morphology, with higher power levels increasing the frequency of ion bombardment, which results in taller structures; however, there is an upper limit to this effect.⁹⁷ The etching time has been studied under a constant working pressure of 2 Pa with SF₆ and O₂ flows at 60 sccm, as shown in Fig. 5(a).⁹⁵ Over time, the black silicon surface becomes progressively rougher; after 1.5 min, independent etching pits begin to appear, marking the onset of anisotropic etching. As etching continues, vertical pores deepen, widen, and connect, with interconnection nearly complete after approximately 3 min. By the 5-minute mark, the entire surface exhibits interconnected pores, effectively reducing reflectivity. Extending the etching time further increases pore width and depth, though pore length is ultimately limited by the aspect ratio-dependent etching effect.⁹⁷

Fig. 5 SEM images showing the effects of different RIE parameters on black silicon morphology. (a) Etching time. (b) Gas composition. (c) Pressure.⁹⁵ Reproduced with permission from ref. 95. Copyright 2014, AIP Publishing. (d) Substrate temperature. (e) Bias voltage.⁶⁴ Reproduced with permission from ref. 64. Copyright 2013, AIP Publishing.

By adjusting the concentrations of SF₆ and O₂, various surface structures can be tailored to meet specific profile requirements.⁹⁵ Higher SF₆ concentrations or lower O₂ levels promote more isotropic structures, while increasing O₂ levels result in positively tapered, conical profiles. Plasma gas composition plays a crucial role in determining surface morphology, as shown in Fig. 5(b), which illustrates SEM images for O₂/SF₆ gas flow ratios of 80 : 60, 60 : 60, 60 : 90, and 60 : 129. As the SF₆ concentration increases, the sidewalls of the microstructure show more fraying due to the formation of shallow nanopores. Conversely, O₂ contributes a passivation effect, and higher O₂ content reduces both the width and depth of the apertures. Additionally, process pressure significantly impacts nanoscale morphologies by modulating the energy and directionality of ions impinging on the surface, providing further control over the etching process.

The impact of process pressure on etched morphologies is significant, influencing the depth and sidewall profiles of the resulting structures.⁹⁵ As shown in Fig. 5(c), the morphologies at pressures of 1 Pa and 2 Pa are similar, characterized by slender structures with nanopores on the sidewalls. However, as the pressure increases to 3 Pa and 4 Pa, the depth of the structures decreases, and the sidewalls become less steep. This change can be attributed to the reduced mean free path of ions at higher pressures, which limits their acceleration and suppresses anisotropic etching. Instead, isotropic etching is enhanced, resulting in a more uniform etching profile. Consequently, the positive sidewall effect preserves more etch pores, leading to shorter and smaller nanostructures.

Substrate temperature significantly impacts surface morphology by affecting the formation of the SiO_xF_y passivation layer.⁶⁴ As shown in Fig. 5(d), at −80 °C, the surface appears rough but does not exhibit a blackened appearance. With decreasing temperature, conical formations begin to develop, becoming more uniform and densely packed at −110 °C. At −130 °C, a distinct anisotropic, isolated, and homogeneous “penguin-like” structure emerges, highlighting the temperature's effect on the etching process.

The bias voltage, as depicted in Fig. 5(e), influences the surface morphology of black silicon by controlling the energy and direction of ions reaching the silicon surface. Without any bias, no distinct surface structure forms, and only surface roughening occurs. At a bias of −10 V, sharp cone structures begin to form. However, with further increases in voltage, there is minimal change in the surface structure, indicating that the effect of bias voltage on the morphology and reflectance of black silicon is relatively limited beyond this point.

The mask plays a crucial role in forming regular structures, and common masking methods include lithography, deposition layer, SiO₂, anodic aluminum oxide (AAO), and polystyrene (PS) sphere masks.^28,98–101 Zhang et al. utilized a two-step RIE process to etch silicon,⁶⁵ and the fabrication process is illustrated schematically in Fig. 6(a). Initially, mask-assisted RIE created micrometer-scale hollow cylinders, followed by maskless RIE to etch nanopores on the top and bottom surfaces. Fig. 6(b) displays an SEM image of a chimney-like array with a diameter of 8 µm, a central circular hole of 2 µm, and a height of 16 µm, with some nanopores on the structures.

Fig. 6 Mask utilization in RIE. (a) Schematic illustration of masked and maskless RIE. (b) SEM image of chimney-like arrays.⁶⁵ Reproduced with permission from ref. 65. Copyright 2020, John Wiley & Sons. (c) AAO membrane transfer and mask etching process. (d) AAO membrane/PMMA/silicon substrate. The blue sections signify the PMMA sandwiched in the gaps. (e) Diameter variation of AAO nanopores and silicon nanopores at different pore-widening times; the inset SEM images show the corresponding AAO membrane and the etched silicon nanopores.⁹⁸ Reproduced with permission from ref. 98. Copyright 2015, Royal Society of Chemistry. (f) Process of PS nanospheres for mask-RIE.²⁸ Reproduced with permission from ref. 28. Copyright 2023, Elsevier.

AAO masks are typically prepared through a two-step anodization process of high-purity aluminum foils.^98,102–104 In the first anodization, aluminum is electrochemically oxidized in an acidic electrolyte such as oxalic, sulfuric, or phosphoric acid, leading to the formation of a porous alumina layer with a disordered arrangement. This layer is then chemically removed, leaving a textured aluminum surface. A second anodization under the same conditions produces a highly ordered hexagonal array of nanopores, whose diameter and interpore distance can be tuned by the choice of electrolyte, anodization voltage, and temperature. Subsequent pore-widening treatment in phosphoric acid allows further adjustment of pore size. The resulting AAO template can serve as a cost-effective, self-assembled nanomask for pattern transfer onto silicon during black silicon fabrication. Lin et al. transferred an AAO mask onto the surface of a silicon wafer spin-coated with a polymethyl methacrylate (PMMA) film.⁹⁸ The PMMA interlayer eliminates gaps to make good contact between the AAO film and the wafer, also avoiding unneeded etching in the shielded area, achieving high-precision aperture control, as shown in Fig. 6(c)–(e). The diameters of silicon nanopores are very close to AAO nanopores.

For PS spheres, as depicted in Fig. 6(f),²⁸ usually large-diameter PS spheres are self-assembled onto the wafer surface, and then ICP or RIE with O₂ plasma is applied to shrink the diameter of the spheres and enlarge the inter-sphere spacing, which allows fine control over the mask geometry. After this step, the exposed regions of the wafer are etched with RIE, transferring the periodic arrangement of the PS spheres into ordered micro–nanostructures on the silicon surface. By tuning the initial sphere size, plasma treatment time, and etching parameters, the resulting structure size and periodicity can be flexibly adjusted. This bottom-up lithography approach is attractive due to its low cost and simplicity; however, issues such as sphere agglomeration, limited large-area uniformity, and difficulty in achieving defect-free ordering remain challenges for practical applications.

In addition to cryogenic RIE, black silicon can also be fabricated through the Bosch process, also known as deep reactive ion etching (DRIE), which alternates between etching and passivation steps to form high-aspect-ratio nanostructures.¹⁰⁵ A typical cycle consists of three sequential steps: anisotropic etching of silicon by SF₆ plasma, conformal deposition of a fluorocarbon passivation film by C₄F₈, and selective removal of the passivation layer at the trench bottom to enable vertical progression of the etch front. The iterative alternation of these steps suppresses lateral etching while maintaining verticality, compared with conventional RIE that often suffers from sidewall tapering and limited aspect ratios. As a result, aspect ratios exceeding 10 : 1 can be readily achieved, making the Bosch process indispensable for microelectromechanical systems (MEMSs),¹⁰⁶ microfluidics, and nanostructured silicon fabrication.

By optimizing the balance between etching and passivation, the Bosch process can generate nano-jungle structures with very high aspect ratios on silicon.¹⁴ These nanostructured black silicon surfaces offered enhanced light trapping, improved wettability control, and superior mechanical robustness compared to conventional nanograss black silicon.

Furthermore, nanoparticle-assisted Bosch processes have been demonstrated to be an effective route for engineering needle-like silicon nanowires with dual functionalities, including broadband antireflection and superhydrophobicity.¹⁰⁷ In this approach, PS nanospheres or other colloidal nanoparticles serve as a self-assembled etch mask, introducing local variations in the etching front. During the cyclic Bosch process, the presence of nanoparticles promotes the anisotropic evolution of high-aspect-ratio protrusions, resulting in dense silicon nanowire arrays with needle-like morphology. These structures integrate both optical and surface advantages: the elongated nanowires provide strong light scattering and multiple internal reflections to reduce reflectance, while their sharp, high-density features lower the solid–liquid contact area, thereby enhancing water repellency.

Hydrogen-containing plasmas, particularly CF₄/H₂ mixtures, have been demonstrated as an effective maskless route for black silicon fabrication. Vassallo et al. showed that by applying RF powers of 200–280 W for 20–30 min, well-defined nanostructures with lateral dimensions of 50–300 nm and depths of 100–300 nm were uniformly produced on silicon surfaces.¹⁰⁸ These nanoscale pillars substantially suppressed reflectance to below 5% across the visible to near-infrared range. Importantly, the study introduced a hydrogen-based passivation mechanism, in which hydrogen replaces oxygen to form volatile HF while simultaneously promoting CF_x passivation layer deposition. A related mechanistic study confirmed that in CF₄/H₂ plasma, fluorine radicals act as the primary etchants,¹⁰⁹ whereas H₂ serves both as a scavenger of fluorine and as a passivating agent, thereby enabling the self-organized formation of high-aspect-ratio nanostructures without lithographic masks. Together, these results highlight hydrogen plasma as a cost-effective and scalable alternative for black silicon fabrication with excellent optical performance.

RIE demonstrates broad applicability across various silicon wafer types and offers precise control over the etching process. Unlike other methods, RIE offers the advantage of fabricating black silicon not only on flat substrates but also on thin films and curved surfaces.⁶⁶ When a highly curved hemispherical silicon lens was placed in the etching chamber, the electric field perpendicular to the surface bent, allowing ions to accelerate along the gradient and impinge on the sample surface nearly vertically. This process results in microstructures aligned parallel to the surface normal. However, the etching must occur in a vacuum chamber at low temperatures to minimize surface damage and prolong the carrier lifetime.⁹⁷

2.3. Metal-assisted chemical etching

MACE is an advanced etching technique that evolved from conventional chemical etching, utilizing the catalytic properties of metal nanoparticles to achieve faster etching rates and greater control over surface structures.¹¹⁰ MACE preparation methods are typically categorized into one-step or two-step processes, based on how metal nanoparticles are introduced onto the silicon surface. In the one-step method, a mixed solution containing HF, HNO₃ (or hydrogen peroxide (H₂O₂)), and silver nitrate (AgNO₃) or chloroauric acid (HAuCl₄) is used.^110,111 In this process, metal nanoparticles are formed in situ on the silicon surface and simultaneously act as catalysts for localized etching. This method is simple and efficient, but the distribution and size of the nanoparticles are less controllable. In the two-step method, metal nanoparticles are first deposited onto the clean silicon wafer surface using techniques such as sputtering, evaporation, or solution reduction. The wafer is then immersed in an etchant solution (typically HF and an oxidant), where redox reactions at the metal-silicon interface induce anisotropic dissolution beneath the metal sites. Compared with the one-step method, this approach allows finer control over particle density, morphology, and etching uniformity, thereby enabling more reproducible nanostructure fabrication.

Although the precise mechanism of MACE is not yet fully understood, it is widely accepted that localized redox reactions occur at the metal-silicon interface after nanoparticle deposition. These reactions selectively etch the silicon beneath the metal particles, creating highly controlled surface morphologies. The whole process is illustrated in Fig. 7. The etching mechanism of MACE in HF/H₂O₂ solution is represented by the equations listed in the left section of Table 2, including two half-reactions at the cathode and anode.^38,112 In contrast, the reactions occurring in HF/HNO₃ solution are shown in the right section.^54,113 Key factors influencing MACE include the type and distribution of metal particles, etching time,^12,114 crystal orientation, and composition of etchant.¹¹⁵

Fig. 7 Fabrication mechanism of MACE.

Table 2 Possible reaction mechanism of MACE

Etching solution	HF + H2O2^38	HF + HNO3^54,113
Cathode	H2O2 + 2H^+ + 2e^− → 2H2O	HNO3 + 3H^+ → NO + 2H2O + 3h^+
		HNO3 → HNO2 → NO → HNO → H2N2O2 →N2O
Anode	Si + 2H2O + 4h^+ → SiO2 + 4H^+	Si + 2H2O + nh^+ → SiO2 + 4H^+ + (4 − n)e^−
	SiO2 + 6HF → H2SiF6 + 2H2O	SiO2 + 6HF → H2SiF6 + 2H2O
	Si + 6F^− + 2H^+ + 2h^+→ SiF6^2− + H2
Overall		Si + 6HF + NO3^− → NOx + H2SiF6 + 2H2O
Note	• n = 2, silicon dissolves directly to form SiF6^2−
	• When the number of cavities at the Si/electrolyte or metal/electrolyte interface increases, dissolution occurs in the tetravalent system	• n is the average number of holes required to dissociate one Si atom
	• n = 2 in the divalent system, n = 4 in the tetravalent system, and n = 3 when both systems are present simultaneously	• HNO3 decomposes throughout the process

---

2.3.1. Metal catalysts. Various metal nanoparticles can be used as catalysts for the MACE process, with Ag being one of the most commonly utilized.¹¹⁶ For Ag deposition, methods such as HF/AgNO₃ mixed solutions, Tollens’ reagent,¹¹⁷ or a liquid-phase chemical reduction are widely employed.¹¹⁸ A novel technique using a two-dimensional non-close-packed silica colloidal crystal has also been developed, enabling the deposition of Ag nanoparticles to fabricate ordered silicon nanowire arrays through nanosphere lithography combined with MACE.¹¹⁹ The type of metal catalyst also plays a major role in MACE because different metals exhibit distinct catalytic ability and stability. For example, platinum (Pt) and palladium (Pd) exhibit higher etching rates than Au,⁶⁸ while Au is more stable than Ag,¹²⁰ making it more suitable for precise control of the etching location without deviation. A recent study has shown that ruthenium (Ru) can achieve pore-size control comparable to that of Au when used as a catalyst,¹²¹ and Ru is compatible with semiconductor CMOS processes. Annealing temperature is another crucial factor that influences the surface morphology of black silicon by affecting the shape and dispersion of nanoparticles.^71,122,123 In two-step MACE, annealing occurs after the deposition of metal nanoparticles and the subsequent etching step, and the former plays the dominant role in determining the final morphology. At low temperatures, the metal nanoparticles are elongated. As the temperature rises, they become more spherical and disperse further apart, eventually forming non-uniform structures at higher temperatures. The reflectance curves will initially decline with rising temperature, reaching a minimum before increasing again.

Table 3 provides a comparative overview of representative metal catalysts used in MACE, highlighting how differences in catalytic behavior translate into distinct etching regimes and morphology evolution pathways. Rather than ranking individual metals, the comparison emphasizes trade-offs among etching efficiency, structural controllability, and process robustness. Such a perspective is particularly relevant for tailoring black silicon architectures toward specific functional targets, where catalyst selection and post-deposition treatments collectively define the accessible parameter space for nanostructure engineering.

Table 3 Comparison of metal catalysts used in the MACE of silicon

Metal catalyst	Typical deposition method	Etching characteristics	Morphology control	Stability during etching	Representative features
Ag (one-step)	HF/AgNO3 mixed solution; Tollens’ reagent; chemical reduction	High etching rate; strong catalytic activity	Good directional control; favors nanowire and porous structures	Moderate; prone to dissolution and migration	Most widely used catalyst; high-aspect-ratio nanostructures; ordered Ag nanoparticle arrays via patterning
Ag (two-step)	Ag nanoparticle deposition + thermal annealing	Etching rate depends on nanoparticle shape and spacing	Strongly tunable via annealing temperature	Moderate	Annealing alters Ag nanoparticle geometry, significantly affecting reflectance behavior
Au	Thermal evaporation; sputtering; chemical deposition	Moderate etching rate; relatively uniform hole injection	High precision in etching location; pore-size controllable	High chemical stability	More stable than Ag; suitable for precise and localized etching, expensive
Pt	Physical or chemical deposition	Very high catalytic activity; fast etching	Enables ultra-high-aspect-ratio nanostructures	High	Higher etching rates than Au; excellent stability but limited by high cost
Pd	Physical or chemical deposition	High catalytic efficiency	Nanowires or nanoporous layers	High	Comparable etching rates to Pt; relatively expensive, limiting large-scale use
Ru	Thin-film deposition	Moderate etching rate	Pore-size control comparable to Au	High	CMOS-compatible catalyst; good stability with precise morphology control

---

2.3.2. Crystal orientation and doping level of the substrate. In MACE, the silicon substrate is a key factor alongside the catalyst and etchant, and its doping type and concentration strongly influence both the etching kinetics and the final morphology.¹²⁴ Systematic studies have shown that higher dopant concentrations accelerate the vertical etching rate of SiNWs, while the porosity of the etched structures is also closely correlated with the doping level. Typically, lightly doped wafers exhibit slower etching rates and produce denser, smoother nanostructures, whereas heavily doped wafers lead to faster etching and generate nanowires with higher porosity and rougher sidewalls.

The crystallographic orientation of the substrate further governs the etching direction. In a study, the influence of p-Si (100) and (111) wafers on etching behavior in an acidic solution was analyzed,¹¹⁷ revealing significant differences in the resulting surface morphology. This finding suggests that wafer orientation impacts the etching process dynamics. Chen et al. further demonstrated that (100)-oriented silicon wafers exhibit a higher average etching rate in alkaline solutions compared to (110) and (111)-oriented wafers.⁸⁶ Consequently, most black silicon studies focus on (100)-oriented wafers due to their superior etching characteristics. Kim et al. provided a microscopic electrochemical explanation for this dependence.³⁷ During etching, H₂O₂ is rapidly consumed near the reaction front, establishing a steep concentration gradient that oscillates periodically due to diffusion-driven replenishment. These fluctuations modulate the etching directions: at low H₂O₂ concentration, etching proceeds preferentially along the vertical 〈100〉 orientation, whereas high concentrations promote etching along non-〈100〉 directions. Temperature also couples with this effect. At elevated temperatures, the accelerated decomposition of H₂O₂ generates more holes, rendering densely packed crystal planes with higher back-bond density, such as {111}, unstable against oxidation. Under such conditions, these planes are also etched, whereas under low H₂O₂ concentration, only the less back-bonded {100} planes are selectively attacked.

2.3.3. Etching time. The effect of etching time is presented in Fig. 8(c),¹¹⁴ showing that the height of the SiNW arrays increases from 0.3 µm to 9 µm as etching time progresses. The SEM micrograph in Fig. 8(a) depicts SiNW arrays with an average diameter of 150 nm and a length of 3 µm, while the corresponding TEM image in Fig. 8(b) reveals that the SiNWs grow along the [100] direction on a (100) Si wafer.

Fig. 8 Images of SiNWs of different etching times.¹¹⁴ Reproduced with permission from ref. 114. Copyright 2010, Optica Publishing Group. (a) SEM image of the SiNW array formed on n-type (100) Si substrates. (b) TEM images of SiNWs with an average diameter of 150 nm. The lattice-resolved image and the corresponding FFT pattern confirm that the SiNWs are oriented along the [100] direction. (c) Side-view SEM images of SiNW arrays with different lengths of 0.3 µm, 1 µm, 2 µm, 3 µm, 5 µm, and 9 µm from (c₁) to (c₆), respectively.

2.3.4. Composition of etchant. MACE has also been applied to textured pyramidal silicon. The nanowires formed on the (111)-oriented pyramid facets are typically shorter than those on planar (100) surfaces due to the intrinsically slower etching rate of the (111) plane. This crystallographic effect enables the fabrication of pyramid-based black silicon with ultra-low reflectivity. The composition of the etchant has a significant impact on the MACE process, as illustrated in Fig. 9.¹²⁵ Taking an etching solution containing AgNO₃ and HF as an example, the left section shows the evolution of black silicon structures as a function of AgNO₃ concentration. As the AgNO₃ concentration increases, the surface becomes rougher. At low Ag⁺ concentrations, SiNWs do not form on the pyramid structures. However, when the Ag⁺ concentrations are too high, the pyramid textures are excessively etched, leading to dispersed and incomplete structures, which significantly increase reflectivity. The influence of HF is shown in the right section. When HF concentration is low, the etching depth is insufficient, resulting in incomplete SiNWs on the pyramid surface. Conversely, at high HF concentrations, the etching becomes too aggressive, causing the surface to flatten, leaving only scattered and incomplete SiNW structures. The best surface morphology for minimizing reflectance is achieved at an optimal concentration ratio of AgNO₃ and HF.

Fig. 9 Influence of etchant composition on the formation of black silicon.¹²⁵ Reproduced with permission from ref. 125. Copyright 2022, Springer Nature. (a)–(f) SEM images of samples etched in solutions with different AgNO₃ concentrations. (g)–(l) SEM images of samples etched in solutions with different HF concentrations.

HF solution generally exhibits a higher etching rate but offers limited controllability. Replacing HF with ammonium fluoride (NH₄F) significantly slows down the reaction kinetics, thereby improving controllability and reducing pore formation in the etched structures.¹²⁶ At the same molar concentration, the weight loss of silicon etched in HF solution is higher than that in NH₄F solution (2.88 g vs. 2.78 g after 5 min), confirming the lower etching efficiency of NH₄F. The resulting nanograss on textured silicon becomes denser and more vertically aligned, often presenting petal-like features at the pyramid tips due to the preferential vertical growth. Moreover, the reduced height of the nanostructures leads to a slightly lower effective refractive index compared with conventional HF-etched black silicon. It should be noted that NH₄F is still a hazardous fluoride salt: although less percutaneously penetrating and systemically toxic than aqueous HF, it is corrosive and can release HF under acidic conditions. Therefore, appropriate protective equipment and fume-hood operation are essential when handling NH₄F.

2.3.5. Lithography or mask technology. In addition to the direct preparation of nanowire arrays, MACE can be combined with lithography or masking techniques to prepare more regular nanowire arrays.^{37,70,127,128} Huang et al. used AAO templates to deposit chromium/gold (Cr/Au) nanodots onto silicon wafers.⁷⁰ Then the AAO template was removed, and a thin layer of Au catalyst was evaporated onto the sample surface. Au-assisted chemical etching was performed in a mixed solution of deionized (DI) water, H₂O₂, and HF to produce uniform SiNWs. The fabrication process is depicted in Fig. 10(a)–(d). SiNWs with average pore diameters of 40, 50, 60, 70, and 80 nm, obtained by adjusting the aperture of the AAO template, are shown in Fig. 10(e)–(i). PS spheres can also serve as masks for MACE.^129,130 An ordered monolayer of PS nanospheres serves as a self-assembled mask, defining periodic patterns on the silicon surface. After mask formation, a thin metal layer is deposited, and the wafer is immersed in the etching solution, producing highly uniform and vertically aligned nanostructures that replicate the periodicity of the PS mask. This process enables the formation of clean and well-defined nano-morphologies. Compared with optical lithography, the use of PS spheres is simpler and more cost-effective, though it has lower resolution; it offers an efficient route for producing ordered nanostructures without the need for photomasks or exposure systems. Its fabrication procedure is comparable to that of RIE, yielding uniform and well-defined nanoscale features.

Fig. 10 Schematic and SEM images of AAO template-assisted-MACE.⁷⁰ Reproduced with permission from ref. 70. Copyright 2010, American Chemical Society. (a)–(d) Fabrication process of the AAO-template-assisted MACE. (e)–(i) SEM images of the AAO templates and the corresponding SiNWs with average diameters of 40, 50, 60, 70, and 80 nm, respectively. Insets show the top-view SEM images of the SiNWs. The scale bars are 200 nm in all SEM images and 100 nm for all the insets.

2.3.6. Other shapes. MACE is a highly versatile technique that goes beyond the traditional fabrication of straight-walled nanowires. It offers the capability to create a wide variety of nanostructures, providing researchers with the flexibility to tailor the morphology and properties of the materials to meet specific application requirements. For instance, Chen et al. demonstrated silicon zigzag nanowires (Si zigzag NWs) by adjusting the etchant composition and etching duration, thereby controlling diffusivity and etching direction.¹³¹ These zigzag NWs are particularly beneficial for applications such as self-cleaning and anti-reflection. SEM images of various nanowires are shown in Fig. 11(a).

Fig. 11 SEM images of representative nanostructures fabricated by direct MACE. (a) Nanowires with varying morphology obtained by tuning the etchant composition and etching time.¹³¹ Reproduced with permission from ref. 131. Copyright 2017, American Chemical Society. (b)–(d) Different 3D spiraling shapes fabricated via MACE.¹³² Reproduced with permission from ref. 132. Copyright 2012, American Chemical Society. (e) Periodically kinked nanowire arrays. (f) Zigzag nanowires. (g) H₂O₂ molar-ratio modulated nanowires.⁶¹ Reproduced with permission from ref. 61. Copyright 2023, AIP Publishing.

Moreover, MACE facilitates the production of 3D nanostructures within a single etching cycle.¹³² The etched profile is influenced by the metal nanoparticles’ shape, and by tailoring the catalyst's geometry, diverse 3D structures can be achieved, such as spirals, as depicted in Fig. 11(b)–(d). Although this flexibility allows the fabrication of complex nanostructures, achieving structural uniformity and large-area scalability remains challenging due to manual control of etching parameters. To overcome these limitations, vapor-phase MACE (VP-MACE) has been developed.¹³³ VP-MACE delivers HF/H₂O₂ vapors to the substrate surface, where they condense into a nanometer-thin liquid layer (<2 nm). Etching occurs only beneath the catalyst, yielding feature resolution comparable to liquid-phase MACE while eliminating bubble formation and preventing the generation of micro- and mesoporous silicon. Catalyst motion is governed solely by near-surface forces, enabling 3D trajectory control and the formation of helical nanostructures. VP-MACE thus improves uniformity, prevents stiction, and enhances resolution in silicon nanofabrication. Building on VP-MACE, programmable approaches introduce independent control over etchant flow rate, pulse time, chamber pressure, substrate temperature, and optional light illumination.⁶¹ By synchronously or asynchronously tuning these parameters, catalyst motion and etch profiles can be precisely engineered, enabling reproducible fabrication of nanowire arrays, point-etched structures, and complex 3D geometries, as illustrated in Fig. 11(e)–(g). This programmability enhances versatility, scalability, and process automation, making VP-MACE a powerful platform for high-aspect-ratio and multifunctional silicon nanostructures.

Another One-step MACE method was used to form nano/micro-hybrid structural arrays.¹³⁴ Single-crystal Si substrates coated with catalytic metal films and photoresist were patterned via photolithography and ion-beam etching to pre-form micron-scale structures, and metal nanoparticles were deposited. After a solution reaction, small nanostructures were formed on the micron-scale structures. This method can also form nano/micro hybrid structure arrays of different shapes, such as circles, rectangles, and triangles, “pentagram”-shaped, and “peanut”-shaped.

In conclusion, MACE offers significant advantages over other nanofabrication methods, making it a highly promising technology for a wide range of applications. One of its key strengths is its broad applicability to various silicon surfaces, including monocrystalline silicon, polycrystalline silicon, and thin-film silicon, demonstrating its versatility across silicon substrates.⁶⁹ MACE also stands out for its scalability, simplicity, and adaptability. It can be seamlessly integrated with other techniques, such as lithography and masking techniques, to achieve precise and directional preparation of surface morphologies. Additionally, its low cost and compatibility with existing manufacturing processes make it ideal for large-scale industrial applications, while offering both efficiency and cost-effectiveness. These advantages position MACE as a cutting-edge solution for fabricating advanced nanostructures across a wide range of industries.

2.4. Femtosecond and nanosecond laser etching

Laser irradiation enables the fabrication of black silicon by precisely ablating and doping silicon surfaces in a single step. This process is independent of the silicon grain orientation.¹³⁵Fig. 12(a) shows a schematic diagram of this fabrication process.¹³⁶ In this setup, cleaned silicon wafers were placed in a vacuum chamber, which is mounted on a 3D stage to control the laser's position along the XY plane, while the Z-axis adjusts the laser spot size focused on the wafer, thereby determining the laser fluence. A neutral density filter was used to ensure the laser fluence remains at optimal levels.

Fig. 12 (a) Schematic of the femtosecond laser irradiation setup.¹³⁶ Reproduced with permission from ref. 136. Copyright 2016, IEEE. (b) SEM image of sharp conical spikes oriented 45° from the surface normal. (c) SEM image of sharp conical spikes oriented parallel to the surface. (d) SEM image of the structures formed in vacuum.⁵⁵ Reproduced with permission from ref. 55. Copyright 1998, AIP Publishing.

In contrast to the smooth holes, trenches, and cavities created by femtosecond irradiation before,¹³⁷ arrays of sharp conical spikes were also formed by irradiating silicon surfaces with 500 laser pulses of 100 fs duration and a fluence of 10 kJ m⁻² in 500 Torr SF₆, as first reported in 1998.⁵⁵ As shown in Fig. 12(b) and (c), the conical spikes are always oriented in the direction of the incident light and independent of gravity, and the morphology in chlorine (Cl₂) was consistent with SF₆. The spherical caps implied that they originated from the dissolution of liquid silica droplets before the liquid wetted the spikes. However, it formed irregular blunt spikes in vacuum (Fig. 12(d)), N₂ and helium (He) gas, indicating there were chemical reaction between SF₆ or Cl₂ to form sharp spikes.

In femtosecond laser etching, ultrashort pulses create localized, high-temperature plasma on the silicon surface, inducing rapid melting, vaporization, and resolidification, which collectively lead to the formation of nanostructures such as spikes or pores.¹³⁸ This process minimizes heat diffusion, ensuring precise material modification and reducing thermal damage. In contrast, nanosecond laser etching involves longer pulses that generate broader heat-affected zones,¹³⁹ leading to micro- and nanostructure formation through thermal ablation, melt expulsion, and capillary instabilities. Nanosecond laser irradiation tends to produce smoother surfaces due to longer melting times and slower rates, while femtosecond laser irradiation results in sharper, more distinct features. Typically, structures formed by femtosecond lasers are about 8 µm in height with ∼4 µm spacing, whereas structures formed by nanosecond lasers can reach ∼40 µm in height and are spaced ∼2 µm apart. Representative morphologies produced by femtosecond laser irradiation (Fig. 13(a) and (c)) and nanosecond laser irradiation (Fig. 13(b) and (d)) are shown in Fig. 13.¹⁴⁰

Fig. 13 SEM images of structures produced by femtosecond and nanosecond laser irradiation.¹⁴⁰ Reproduced with permission from ref. 140. Copyright 2004, AIP Publishing. (a) Morphology generated by femtosecond laser irradiation with a laser wavelength of 800 nm, a pulse duration of 100 fs, and a fluence of 10 kJ m⁻². (b) Morphology produced by nanosecond laser irradiation with a laser wavelength of 248 nm, a pulse duration of 30 ns, and a fluence of 30 kJ m⁻². (c) Enlarged SEM image corresponding to (a). (d) Enlarged SEM image corresponding to (b).

2.4.1. Doping. The notable advantage of femtosecond laser irradiation is its ability to perform doping during the ablation process. The source of impurity atoms depends on the processing atmosphere and pre-deposited layers. Dopants may originate from surrounding gases (e.g., SF₆, Cl₂) or from intentionally applied precursor films on the silicon surface. The mechanism of doping, illustrated in Fig. 14(a), operates as follows:¹⁴¹ the laser pulses melt the silicon surface, allowing dopant atoms to diffuse from the surface into the molten silicon. During femtosecond laser irradiation, hyperdoping occurs through several key processes. Solute rejection or trapping during resolidification occurs when the dopant diffusion into molten silicon outpaces or lags behind the solid–liquid interface velocity, thereby determining the extent of dopant incorporation. Rapid resolidification traps excess dopants, enabling hyperdoped silicon with enhanced electronic and optical properties. The center of the laser-irradiated region becomes single-crystalline, while an amorphous ablative ring forms around it, indicating a crystal phase transition, as shown in Fig. 14(b) and (c).

Fig. 14 Melting and doping processes during laser etching.¹⁴¹ Reproduced with permission from ref. 141. Copyright 2015, AIP Publishing. (a) Illustration of the doping process during silicon melting. (b) Normalized Raman intensity curves of molten silicon. The inset shows the corresponding optical microscope image. (c) Bright-field TEM image of the laser irradiation center. The inset shows the corresponding FFT pattern.

2.4.2. Gas atmosphere. The effect of different ambient gases during femtosecond laser irradiation was studied by Mazur et al. in 2003.⁷²Fig. 15 reveals distinct differences between corrosive and inert gases in shaping black silicon morphology. In corrosive gases like SF₆ and Cl₂, as shown in Fig. 15(a) and (b), sharp conical spikes with dendritic structures are formed on the sidewalls. In contrast, in the inert gases such as N₂ and air, as shown in Fig. 15(c) and (d), the microstructures are blunter. Air produces coarser structures with fewer, less dense dendritic nanoparticles than N₂.

Fig. 15 SEM images of black silicon in different ambient gases.⁷² Reproduced with permission from ref. 72. Copyright 2003, AIP Publishing. (a) SF₆. (b) Cl₂. (c) N₂. (d) Air.

The presence of sulfur promotes the formation of sharp conical spikes; as sulfur concentration in the environment increases, the tip area of each microstructure decreases. Pre-coating the silicon wafer with the doping layer has been shown to further reduce reflectivity in ambient air.¹⁴² When coated silicon is ablated by the laser, impurity atoms on the surface are trapped in the molten state and doped during the formation of micro- and nanostructures.

2.4.3. Laser fluences. For laser fluences,⁷⁵ as shown in Fig. 16(a), with the laser fluence increasing, the silicon surface undergoes four stages of transformation, from clear nanoripples (a₁), to nanoripples and nanoholes (a₂), then to micro-spikes around nanopores (a₃), and finally to separate micro-spikes (a₄). This phenomenon occurs when the laser fluence approaches the ablation threshold, where molten silicon begins to form nanoripples. Upon femtosecond laser irradiation of the silicon substrate, a large density of surface plasma is generated. The surface plasma wave interacts with the incident laser, producing interference that induces a periodic modulation of the laser energy on the silicon surface. This modulation results in a periodically structured ablation pattern on the silicon surface. As the laser fluence continues to increase, the molten silicon accumulates on the nanoripples. Due to its inherent instability, the molten material flows under the influence of surface tension, eventually forming nanoholes.

Fig. 16 SEM images of various influencing factors of femtosecond laser irradiation. (a) Laser fluences. (b) Scanning velocities.⁷⁵ Reproduced with permission from ref. 75. Copyright 2014, Elsevier. (c) Pulse numbers.⁵⁸ Reproduced with permission from ref. 58. Copyright 2001, AIP Publishing.

2.4.4. Scanning velocities. A similar trend is observed with varying laser scanning velocities. Under the same laser flux, a higher laser scanning speed results in less energy accumulation on the silicon surface, leading to the formation of clear nanoripples. As shown in Fig. 16(b), when the scanning speed decreases, the surface changes from nanoripples to nanospikes and nanoholes. However, excessively slow scanning speeds cause severe surface damage.

2.4.5. Pulse numbers. Increasing the number of laser pulses leads to a gradual morphological evolution on the silicon surface.^58,74 As shown in Fig. 16(c), the surface changes from nanoripples to sharp conical spikes, which become progressively denser and taller as the pulse number increases. Eventually, the spike height approaches a steady state.

2.4.6. Temperature. Annealing and substrate temperature are two other important factors. Generally, thermal annealing will decrease the light absorption sharply in the NIR range due to the inactivation of subband absorption,¹⁴³ though this can be reactivated by laser annealing.⁶⁰ Subband absorption allows semiconductor materials to absorb light with photon energies below the bandgap. This means that the material can respond to lower energy photons, thus extending the photoresponse range. Nanosecond laser annealing smooths the sharp conical surfaces at the nanoscale while preserving the cone's structure, resulting in a monocrystalline appearance. Substrate temperature influences the required laser fluence because surface ablation needs a specific temperature to occur; lower substrate temperatures generally require higher laser fluence to reach the critical ablation threshold.⁷⁶

2.4.7. Etching solution. An alternative to vacuum chamber etching is the fabrication of black silicon in acidic or alkaline solutions.^144,145 For instance, Meng et al. achieved black silicon with jungle-like microstructures by using a femtosecond laser to irradiate the silicon surface in an alkaline solution.¹⁴⁵ This method eliminates the need for costly vacuum equipment and dangerous gases such as SF₆ and Cl₂. Moreover, for a silicon-on-insulator substrate, femtosecond laser scanning can produce large-area and flexible black silicon on the thin silicon layers, achieving 97% absorption in the visible spectrum. This approach offers promising potential for applications such as photodetection and solar-thermal systems.¹⁴⁶

In conclusion, femtosecond laser irradiation presents a powerful technique for black silicon fabrication, offering key advantages such as efficient doping, broadened absorption spectra, and precise control of surface morphology. However, challenges remain, including high production costs, potential environmental concerns during fabrication, and lengthy processing times.

2.5. Plasma immersion ion implantation etching

Ion implantation is a surface modification technique where ions are injected into a solid material to alter the physical and chemical properties of its surface without affecting its bulk characteristics. However, the doping concentration declines rapidly with increasing depth in the sample.¹⁴⁷ Plasma immersion ion implantation is often used for semiconductor doping. The principle was summarized by Cheung in 1996.⁵⁷ Xia et al. first used this method to produce needle-like structures on the silicon surface in 2011,²² and the forming mechanism is represented in Fig. 17(a).

Fig. 17 Mechanism and SEM images of PIIIE-fabricated black silicon.²² Reproduced with permission from ref. 22. Copyright 2011, Elsevier. (a) Mechanism of the PIIIE process. (b) SEM image of the black silicon surface (top view). (c) SEM image of the black silicon surface (side view, viewed at 30° to the surface normal).

In this process, plasma was generated at 900 W radio frequency power. SF₆ plasma gas provided the F⁺, SF⁺, SF₃⁺, and SF₅⁺ ions, and O₂ plasma generates O* radicals. With –500 V pulses applied to the sample stage, the SF_x⁺ (x ≤ 5) and F⁺ were injected into the silicon substrate. These reactive ions reacted with silicon and formed a uniform microstructure on the silicon surface after the volatile SiF₄ was formed and escaped. After that, the O* radicals reacted with the fresh surface rapidly to form SiO_x. The SiO_x layer then interacted with the SF_x⁺ (x ≤ 5) and F⁺ ions to produce the Si_xO_yF_z, which partially escaped under the effect of ion bombardment and were further etched by the SF_x⁺ (x ≤ 5) and F⁺ ions. Finally, the porous or needle-like microstructures of black silicon were formed (Fig. 17(b) and (c)). This process reduced the reflectance to 1.79%, significantly lower than 12.07% for textured silicon (NaOH/IPA etching) and 30.01% for polished silicon.

The PIIIE process is a gas-phase etching technique; the key influencing factors include gas composition, RF power, bias voltage, gas flow rate, etching time, pressure, and temperature.

The influence of the SF₆/O₂ gas flow ratio was studied by Xia et al.,⁸⁰ as shown in Fig. 18(a). With an increasing SF₆/O₂ ratio (Fig. 18(a2)–(a4)), the surface morphology of black silicon changed from a porous structure to a needle structure, and the reflectivity decreased sharply. However, when only SF₆ was used as the plasma gas (Fig. 18(a1)), the O and F concentrations on the surface were very low, and the surface could only become rough with high surface reflectivity.

Fig. 18 SEM images of black silicon prepared by PIIIE under different processing parameters. (a) Composition of reactive gases.⁸⁰ Reproduced with permission from ref. 80. Copyright 2012, Elsevier. (b) Bias voltage. (c) H₂ flow rate.⁸² Reproduced with permission from ref. 82. Copyright 2018, Elsevier.

In Lim's research,⁸² H₂ and Ar feedstocks were used in the plasma etching process, with RF power and bias voltage serving as crucial factors in converting neutral gases into reactive species. The influence of bias voltage is illustrated in Fig. 18(b). At a low bias of −25 V, large and random nanohumps and clusters form. Increasing the bias to −50 V reduces the cluster size, filling them with smaller or larger nanohumps. At −75 V, high-density and well-arranged nanocones emerge, whereas further increasing the bias to −100 V causes the nanocones to break, resulting in thin and fragmented nanobranches. A low bias voltage produces a weak sputtering effect, leading to agglomeration of sputtered atoms, while a higher bias promotes the formation of sharp, well-defined cones.

In addition, the influence of H₂ flow was investigated. Increasing the H₂ flow leads to the discharge quenching and reduces the density of active substance in the plasma etching process, thus affecting the surface structure. A denser discharge can be obtained at higher RF power. As shown in Fig. 18(c), as H₂ flow increases, the structure evolves from dense and sharp nanograss tips to wider nanocones and eventually pyramid-like structures, with increasing nanostructure size and decreasing density. The reflectivity of the samples decreased gradually in this process, implying that, in addition to an appropriate aspect ratio, the nanostructure density also has an effect on the reflectance of light.

PIIIE is an advanced method for fabricating black silicon, offering high controllability and doping capability. However, ion implantation concentration diminishes rapidly with depth, and the need for expensive equipment limits its wider application.

2.6. Comparative insights: process-structure-state correlations

The morphology of black silicon differs significantly due to the distinct mechanisms involved in materials removal and structure formation. Chemical etching techniques, including acidic chemical etching, electrochemical etching, and molten salts electrochemical etching, typically result in isotropic structures, while alkaline chemical etching leads to anisotropic structures due to its crystal-plane-selective etching behavior. Pure chemical etching produces irregular pits or porous morphologies, while electrochemical methods provide greater control, forming uniform pores or porous layers with tunable dimensions. Molten salts electrochemical etching, on the other hand, enables deeper and more defined porous structures but requires elevated temperatures and specific chemical environments. RIE forms well-aligned micro- and nanoscale structures, such as needle- or cone-like surfaces, through plasma-induced anisotropic etching, offering precise control over morphology but at the expense of higher equipment complexity and cost. MACE produces highly controlled vertical structures, such as nanowires or nanopores, by leveraging catalytic reactions at the metal-silicon interface. This method is cost-effective and compatible with large-scale production. Femtosecond and nanosecond laser etching methods create a wide variety of morphologies, from nanoripples to needle-like or hierarchical structures, by adjusting laser parameters such as fluence and pulse duration. Femtosecond lasers offer exceptional precision with minimal thermal damage, while nanosecond lasers are more accessible but lead to broader heat-affected zones. Finally, PIIIE combines plasma etching and doping, enabling the formation of needle-like or cone-shaped structures with controllable doping levels (∼10¹⁴–10¹⁸ ions cm⁻²).¹⁴⁸ However, the process is limited to sub-micrometer depths due to the restricted ion energy and mean free path, and its relatively high cost further constrains large-scale applications.

2.6.1. Silicon consumption. During black silicon formation, the nanostructures are typically etched below the original wafer surface due to continuous silicon removal. Depending on the fabrication method, the top of the nanostructures generally lies about 1 µm (for ICP-RIE or MACE) to several micrometers (for femtosecond laser etching) beneath the initial wafer surface.¹⁴⁹ This silicon consumption varies with etching conditions and mask dimensions—smaller pattern openings in MACE lead to deeper etching as a result of the loading effect caused by an imbalance between reactant supply and consumption. In contrast, ICP-RIE produces more uniform etch depths with weaker pattern dependency. These factors must be considered in device design and process integration, since the entire active region is no longer at a constant level. If doping is performed before black silicon etching, doped regions may be partially or completely removed; conversely, post-etch doping can lead to poor electrical contact between the doped black silicon and the planar surface. Furthermore, when black silicon is fabricated on thin epitaxial layers, excessive silicon consumption may cause the nanostructures to penetrate the epilayer, thus determining the minimum thickness requirement for stable device operation.

2.6.2. Mask and pattern. A comparative analysis between blank and patterned black silicon provides a clearer understanding of how mask control impacts the resulting morphology and optoelectronic performance. Blank black silicon is typically formed during laser irradiation, plasma etching, or MACE processes without predefined masks. Its morphology often features randomly distributed nanocones or nanopores, leading to broadband optical absorption but limited spatial uniformity. In contrast, patterned black silicon incorporates additional lithographic or templating steps—such as photolithography, colloidal lithography, or nanoimprint techniques—to define ordered or semi-ordered surface architectures before the texturing step. This enables precise control of feature dimensions, periodicity, and distribution, resulting in enhanced reproducibility and integration compatibility with device layouts. Consequently, patterned black silicon tends to offer improved light-trapping efficiency, directional scattering, and electrical uniformity compared with blank black silicon.

Each method offers distinct advantages tailored to specific applications, with trade-offs in cost, scalability, and structural control. For example, Tan et al. investigated the difference between the dry etching method (ICP-RIE) and the wet etching method (Ag-MACE), and found that wet etching produces longer nanostructures, but the morphology is less organized.¹⁵⁰ For wet etching, the surface structure of black silicon can minimize the reflectivity within 300–1100 nm. Moreover, the needle tip part has better light absorption than the pit part, while the absorption efficiency of dry-tip (95%) is about 2% lower than that of wet-tip (97%), and the influence of needle tip and pit structure is not large in dry etching. More regular patterns often require lithography, while laser etching does not require an etching mask. RIE can form black silicon with high alignment and pattern transfer accuracy without affecting optical and electrical properties.¹⁴⁹ The edge roughness of MACE patterns is often higher than that produced by RIE, and there is a noticeable contrast gradient near the edges, which may indicate uneven etching contours. The solutions used in the MACE process, such as HF and HNO₃, limit the choice of photoresist. The differences among chemical etching, RIE, MACE, femtosecond and nanosecond laser irradiation, and PIIIE are summarized in Table 4.

Table 4 Comparison of different fabrications of black silicon

	Method	Reaction condition	Main influencing factors	Morphology
Chemical etching	Pure chemical etching	Acidic solution: HF + HNO3	1. Types of silicon	Acid solution: isotropic, porous
			2. Solution composition and concentration	Alkaline: anisotropic, pyramid
			3. Temperature	Higher HF: porous
		Alkaline solution: NaOH/KOH+IPA	4. Etching time	Higher HNO3: polish and remove damage
	Electrochemical etching	HF solution	1. Type and concentration of doping atoms	Mainly cones
			2. Solution	Higher potential and HF concentration: polish silicon
			3. Potential
	Molten salts electrochemical	Molten salt solution	1. Oxygen diffusion in the solid phase	Columns
			2. Charge transfer
			3. O^2− diffusion in liquid phase
RIE		ICP or CCP system with SF6/O2 plasmas	1. RF and ICP power	Higher power: taller structure within limits
			2. Etching time	Longer time: increase the width and depth of pores
			3. Gas composition	Higher SF6: shallow nanopores on side walls
			4. Pressure	Higher pressure: gentle surface
			5. Substrate temperature	Higher temperature: uniform and dense conical
			6. Bias voltage	Little effect after microstructure formation
			7. Mask	Desired nanostructures (penguin-like, hollow cylinders, etc.)
MACE		1. Deposit nanoparticles	1. Types of metal particles	Different catalytic activity and stability
			2. Etching time	Longer time: higher height
			3. Composition of etchant	Higher AgNO3 and HF: from forming nanowires to excessively etched
		2. Reaction: HF + H2O2/HNO3, take H2O2 as example	4. Types and orientation of silicon	Form different junctions and (100) orientation better
			5. Temperature	Higher temperature: stronger reactivity
			6. Mask	Desired nanowires (zigzag, spiraling shapes, etc.)
Fs and ns laser irradiation		Femtosecond laser and vacuum chamber with reaction gas	1. Femtosecond or nanosecond laser	Femtosecond: sharp conical spikes
				nanosecond: smoother
			2. Gas composition	Corrosive gas: sharp conical spikes
				Inert gas: blunter structure
			3. Parameters of laser (fluence, scanning velocity, pulse number)	Higher fluence, pulse number, and lower scanning speed: nanoripples to nanopores to sharp conical spikes
			4. Temperature	Help the silicon ablation
PIIIE		Vacuum chamber with reaction gas, most SF6 and O2	1. Gas composition	SF6/O2 ratio increase: porous to needle structure
			2. Bias voltage and RF power	Voltage/power increase: nanohumps and clusters to nanocones to nanobranches
			3. Gas flow rate	Rate increase: bigger nanostructure size and lower density

---

3. Micro–nano surface architectures of black silicon

The surface architecture of black silicon is a critical determinant of its multifunctional capabilities, influencing light absorption, charge carrier behavior, and interfacial properties. This section systematically reviews the interplay between morphological features and surface chemical states in shaping device performance. It first classifies the structural diversity introduced through micro- and nanoscale manufacturing technologies, highlighting how various morphologies (such as spikes, pores, and trenches) contribute to application-specific enhancements in light trapping, charge transport, and surface reactivity. Subsequently, it discusses the surface states and chemical terminations (e.g., Si–H, Si–O, Si–F) introduced during processing, emphasizing their influence on recombination behavior and doping. Finally, recent advances in surface passivation techniques are examined, with a focus on their effectiveness in enhancing the stability and functional efficiency of black silicon across optoelectronic applications.

3.1. Structural diversity of surfaces

Black silicon refers to silicon surfaces that have been engineered to exhibit extremely high light absorption, typically resulting in a distinctive dark appearance. The morphology—meaning the specific shape, structure, and texture of the silicon surface—is instrumental in defining how black silicon interacts with light, heat, and other external factors. Different preparation techniques yield varied morphologies,^77,151 each imparting distinct properties that are advantageous for specific applications such as photovoltaics, optoelectronics, sensors, and biomedical devices. There are several morphologies of black silicon (as shown in Fig. 19), each with distinctive characteristics.

Fig. 19 Different morphologies of black silicon. (a) Pyramid. (b) Nanowires. (c) Porous. (d) Columns. (e) Conical spikes. (f) Chimney-like array.

Pyramid structures,¹⁵² as shown in Fig. 19(a), are commonly used in the solar cell industry as “textured surfaces” to maximize light absorption and thus increase photovoltaic conversion efficiency. These pyramids are influenced by their aspect ratio and grain orientation, which play a role in the reaction dynamics during formation.

Another morphology is silicon nanowires, as shown in Fig. 19(b),¹⁵³ which are appreciated in semiconductor processing due to their high specific surface area. Their diameter and length can be precisely controlled to meet specific functional requirements, making them adaptable for a wide variety of applications, from sensing to energy storage. Recent advancements in fabrication have allowed even finer control over SiNWs' dimensions, resulting in intricate structures with enhanced electrical and optical properties. SiNWs are often produced through methods such as chemical vapor deposition or metal-assisted etching, where factors like catalyst size and etching time dictate the wire dimensions, enabling a high degree of customization.

Porous silicon is characterized by a highly irregular surface texture composed of interconnected pores,¹⁵⁴ the morphology is shown in Fig. 19(c). The pore sizes typically range from nanometers to micrometers. This morphology is typically formed through electrochemical or chemical etching processes that dissolve parts of the silicon structure, creating a network of voids. The high porosity of this morphology results in an extensive surface area that is especially advantageous in applications requiring high reactivity or rapid diffusion, such as in sensors, photocatalysis, and supercapacitors. Additionally, porous silicon is a good candidate for drug delivery applications due to its capacity to carry and release molecules in a controlled manner, owing to the tunable size and distribution of its pores.

Columns (Fig. 19(d))¹⁵⁵ often formed when a mask is used. Their high aspect ratio simultaneously enhances optical absorption and preserves desirable photoelectric properties, making them highly promising for optoelectronic applications.

Another distinctive morphology found in black silicon is that of sharp conical spikes (Fig. 19(e)).¹⁵⁶ This structure is typically formed by irradiating the silicon surface with lasers in the presence of reactive gases, such as SF₆/O₂. In this process, high laser-induced temperatures promote chemical reactions that incorporate sulfur dopants into the silicon, resulting in hyperdoping and the formation of sub-bandgap states that extend optical absorption into the infrared. Moreover, the formation of sharp conical spikes significantly enhances light trapping and suppresses surface reflectance. The combined effects of sulfur hyperdoping and conical microstructure make this architecture highly promising for broadband optoelectronic applications.

Similarly, needle-like structures or “nanograss” result from plasma-based processes where high-energy ions create a densely textured surface with vertical needle features.^14,157 This morphology offers versatile applications, including energy harvesting, self-cleaning materials, sensors, and biomedical devices.

When the micro- or nanostructures are precisely controlled through mask, lithography, or other patterning techniques, more precisely fabricated morphologies can be obtained,^158–160 such as the chimney in Fig. 19(f). The light absorption can be enhanced further, and the specific morphology could achieve additional novel functionalities.

Each morphology offers unique advantages and can be customized through specific fabrication techniques, influencing the performance of black silicon in diverse applications. The selection of morphology and the tuning of its characteristics, such as size, aspect ratio, and orientation, are essential considerations in the design of high-performance silicon-based devices.

3.2. Surface states induced by micro–nano engineering

Beyond geometric morphology, the surface states introduced during micro–nano engineering play a pivotal role in shaping the optical and electronic behavior of black silicon. Processes such as anisotropic etching, plasma treatment, and chemical modification inevitably generate a variety of surface states, including structural defects, chemical terminations, and interfacial bonding irregularities. These states not only govern carrier trapping and recombination but also affect charge accumulation, band alignment, and long-term device reliability. This section explores the origins and functional impact of such surface states, focusing on defect-mediated carrier dynamics, the role of Si–H, Si–O, and Si–F terminations, and strategies for surface passivation. Understanding and controlling these factors is essential for optimizing black silicon performance across photovoltaic, sensing, and photodetection applications.

3.2.1. Role of surface defect states. During the micro- and nanofabrication of silicon, processes such as etching and irradiation often damage the crystal lattice, resulting in the formation of amorphous regions and/or various structural defects on the silicon surface and its near-surface layers. These defects include dangling bonds, interstitials, vacancies, and lattice dislocations induced by local stress, some typical defects are shown in Fig. 20(a).¹⁶¹ These collectively constitute what are also known as surface states or trap states, which have a profound impact on carrier behavior. Even trace amounts of impurities and defects can drastically alter the physical and chemical properties of semiconductor materials, significantly degrading device performance.¹⁶² By disrupting the periodic potential of the crystal lattice, the impurities and defects introduce shallow or deep energy levels within the bandgap that act as carrier recombination centers,¹⁶³ severely suppressing the lifetime and mobility of photo-generated charges. Compared to shallow-level impurities (small ionization energies), which mainly affect carrier concentration and conductivity type, deep-level impurities (larger ionization energies) are generally present in smaller quantities and have deeper energy levels, exerting a stronger influence on carrier recombination. For instance, Au is a well-known deep-level trap that facilitates fast carrier recombination, which can be beneficial in high-speed switching applications. Additionally, the density and energy distribution of these defect states play a decisive role in determining surface reactivity and adsorption characteristics, which are critical for applications in gas sensing and biosensing.

Fig. 20 Defects and SRH recombination in silicon. (a) Dangling bonds, impurity, and vacancies in silicon. (b) SRH recombination.

Carrier trapping and recombination via defect states. Surface and subsurface defect states, such as dangling bonds and dislocations, introduce localized energy levels within the silicon bandgap that act as non-radiative recombination centers.^163,164 These trap states capture photo-generated carriers and facilitate indirect recombination through phonon interactions, leading to significant losses in carrier lifetime and diffusion length. This effect is particularly detrimental in optoelectronic devices such as photodetectors and solar cells, where efficient charge collection is critical under low-light intensity conditions.^165,166

At the atomic scale, surface defects in black silicon primarily arise from the abrupt termination of the silicon lattice, resulting in unsatisfied valence electrons known as dangling bonds.¹⁶⁷ Dangling bonds exhibit high reactivity and electronic trap properties due to their inability to form sufficient bonds with other atoms. These defects introduce localized energy levels within the bandgap, acting as recombination centers that capture carriers. The density of such bonds can reach ∼10¹⁵ cm⁻².¹⁶⁸ Even under ultra-high vacuum (10⁻¹⁰ Torr), silicon surfaces rapidly adsorb ambient species such as oxygen, forming a native SiO₂ layer that passivates most of the dangling bonds.¹⁶⁹ However, due to the lattice mismatch between silicon and SiO₂, some interface states persist, enabling carrier exchange with the bulk and promoting Shockley-Read-Hall (SRH) recombination. The mechanism is shown in Fig. 20(b),¹⁷⁰ when free charge carriers encounter defect levels, non-radiative recombination occurs and energy is released. Furthermore, fabrication processes including high-temperature treatment, ion implantation, and irradiation can introduce additional defect states that enhance carrier trapping and recombination. Lattice mismatch between silicon and SiO₂ leaves some bonds unsaturated, allowing charge exchange with the bulk. Additional surface states can also result from fabrication-induced damage, such as high temperatures, ion implantation, or irradiation (e.g., X-rays, UV lasers).¹⁷¹

Therefore, controlling the density and nature of surface defect states is essential for minimizing recombination losses and enhancing the performance of silicon-based electronic and optoelectronic devices. A stable, low-defect surface is a prerequisite for achieving high sensitivity, low dark current, and strong photoresponse in advanced device architectures.

Defect-induced modulation of band structure and electric fields. In addition to serving as recombination centers, surface and interface defect states can significantly modulate the band structure and electric field distribution in semiconductor devices. These defects often act as charged centers by capturing carriers, forming fixed charge layers near the surface. Electron-trapping states typically lead to the accumulation of negative charges on the order of 10¹² cm⁻²,^172–175 while hole-trapping states may contribute to positive fixed charges.^176,177 The spatial distribution and density of these charges influence the surface potential, induce band bending, and affect the width of space charge regions such as depletion layers in p–n junctions or Schottky interfaces.

The modulation of band bending caused by defect-induced fixed charges directly impacts the built-in electric field, which governs carrier drift and separation dynamics.^178,179 Dangling bonds acting as charge traps increase carrier capture and recombination, contributing to leakage current and threshold voltage fluctuations in MOS devices. Over time, the accumulation of these defects can degrade device stability and lifetime. When the defect-induced electric field aligns with the internal field, it can aid charge extraction and improve performance. In contrast, an opposing field reduces charge collection, causes response hysteresis,¹⁸⁰ saturation drifts, and threshold voltage shifts—particularly critical in low-light or high-speed optoelectronic applications.

While these phenomena are generally viewed as parasitic, they can also be intentionally exploited in certain device designs. In MOSFETs or thin-film transistors, for instance, surface states and interface dipoles can be engineered to adjust threshold voltages or modulate channel conduction.^181,182 Nonetheless, in high-sensitivity detectors and high-speed photonic devices, the presence of such electrically active states is detrimental. Consequently, surface passivation strategies—such as thermal oxidation, ALD of high-k dielectrics, or self-assembled monolayers—are commonly employed to minimize trap densities and stabilize electrostatic conditions. Ultimately, understanding and managing the electrostatic influence of surface and interface traps is essential for optimizing both the static and dynamic performance of modern semiconductor devices.

Defect-enhanced surface reactivity. Beyond their electronic implications, surface defect states significantly enhance the chemical reactivity of silicon surfaces by acting as highly active adsorption and reaction sites.^183–186 Defects such as dangling bonds, vacancies, and low-coordination atoms create localized electronic states with elevated local density of states, which facilitate stronger interactions with gas molecules and functional chemical species. These unsaturated sites exhibit high surface energy and chemical potential, making them preferential sites for adsorption via chemisorption or charge-transfer interactions.

In gas sensing applications, this defect-induced reactivity plays a pivotal role in determining both sensitivity and selectivity. In nanostructured black silicon, for instance, the combination of high surface area and abundant defects synergistically enhances gas-molecule adsorption and electronic modulation.^{25,26,187–189} Electron-acceptor molecules such as O₂ and NO₂ preferentially adsorb at these defect sites, leading to local charge transfer and the formation of ionized species. As shown in Fig. 21(a) and (b), in ambient air, O₂ molecules adsorb on the surface of p-type black silicon, capturing electrons and forming negatively charged species. These species induce upward band bending and increase the surface depletion region. Upon exposure to NO₂ (Fig. 21(c) and (d)), these adsorbed oxygen species undergo further reactions with NO₂ molecules, extracting additional electrons and intensifying hole accumulation in the surface region. This results in a further increase in hole conductivity and a more pronounced change in the electrical response of the device. The adsorption process modifies the surface potential and band structure, often resulting in measurable changes in conductivity or surface barrier height. A higher density of active defect sites leads to more pronounced electrical signal changes upon exposure to analytes. This enhancement improves the response magnitude and lowers the detection limit (down to 1 ppm) while maintaining a rapid response time of approximately 12–14 seconds.

Fig. 21 Sensing mechanism of black silicon gas sensors for air and NO₂.²⁶ Reproduced with permission from ref. 26. Copyright 2025, Royal Society of Chemistry. (a) and (b) Air. (c) and (d) NO₂.

Similarly, in biosensing applications, surface defects facilitate the immobilization of functional biomolecules such as antibodies, aptamers, or enzyme-linked probes.^190–193 These molecules tend to anchor more effectively at high-energy defect sites due to favorable binding affinity, allowing for stable and dense functional layer formation. This not only enhances target capture efficiency but also improves signal transduction, as biomolecular interactions near electronically active defects can induce local charge redistribution, further amplifying the output signal.

From a materials engineering perspective, the ability to modulate surface defect density provides a means of tuning reactivity for application-specific needs. Techniques such as plasma treatment, laser-induced surface reconstruction, and controlled chemical etching can be used to tailor the distribution and density of active sites. However, the enhancement in surface reactivity must be balanced against potential drawbacks such as increased carrier recombination. Therefore, rational defect engineering is crucial for optimizing multifunctional interfaces in silicon-based sensors and bioelectronic platforms.

Challenges and prospects in defect state control. While a moderate density of surface defect states can enhance chemical reactivity and functionalization capabilities, excessive defects often compromise electronic performance by increasing carrier recombination and instability. This inherent trade-off presents a core challenge in the functional tuning of micro- and nanostructured silicon: achieving an optimal balance between surface reactivity and electronic quality. To address this, various strategies have been explored to precisely control the type, distribution, and energy levels of defect states. For example, fine-tuning etching parameters,^94,194–196 such as etchant concentration, temperature, and reaction time—can mitigate undesired lattice damage during structuring. Post-fabrication treatments such as ALD of dielectric coatings or surface reconstruction are effective in passivating dangling bonds and smoothing out surface disorder,^197,198 thereby reducing non-radiative recombination while retaining sufficient active sites for surface interactions.

Despite these advances, achieving defect control at the atomic scale remains technically demanding, particularly for large-area and complex three-dimensional architectures. Moreover, the long-term stability of passivated states under environmental exposure poses additional reliability concerns for real-world applications.

3.2.2. Interfacial bonding and doping. In addition to structural defects, the surface and interfacial chemistry of micro- and nanostructured silicon is fundamentally altered during fabrication, introducing a variety of chemical terminations and doping-related effects that critically influence electronic and interfacial behavior. Chemical terminations such as Si–H, Si–O, and Si–F bonds are commonly formed during black silicon fabrication processes, and they significantly impact surface passivation, reactivity, and charge distribution. For example, Si–H bonds are known for effectively passivating dangling bonds and reducing surface recombination,¹⁹⁹ while Si–O species enhance long-term stability but may introduce fixed charges that shift the energy band alignment. Si–F terminations confer enhanced hydrophobicity and chemical resistance to the silicon surface. Furthermore, micro–nano structuring often leads to unintentional dopant redistribution or activation, which may result in the formation of unintended junctions or non-uniform carrier profiles. The summary and comparison of chemical bonds is provided in Table 5. These interfacial phenomena, governed by local bonding configurations and dopant behavior, collectively modulate surface band bending, energy barriers, and charge accumulation at the silicon interface—thereby playing a decisive role in the electronic performance, environmental stability, and functional integration of silicon-based optoelectronic devices.

Table 5 Summary and comparison of chemical bonds in black silicon

Chemical bonds	Si–H	Si–O	Si–F
Origin	Hydrogen plasma treatment, HF etching	Thermal oxidation, ALD, PECVD	Dry etching (e.g., SF6, CF4 plasmas)
Passivation mechanism	Chemical passivation (saturation of dangling bonds)	Chemical + field-effect passivation	Hydrophobic termination, weak chemical passivation
Passivation performance	Good	Good, enhanced by hydrogen	Poor
Thermal stability	Moderate	Excellent	High
Chemical stability	Moderate	High	Very high
Surface wettability	Neutral/hydrophobic	Hydrophilic	Hydrophobic
Interface defect control	Good	Moderate (requires additional treatment)	Poor
Post-treatment compatibility	Good (suitable for further functionalization)	Moderate (requires careful processing)	Poor (inhibits further modification)

---

Surface terminations and their functional roles. Surface terminations play a vital role in determining the interfacial electronic behavior and chemical stability of micro- and nanostructured silicon.

Si–H terminations are typically formed after HF treatments or wet chemical etching processes.^200–202 These terminations are highly effective at passivating surface dangling bonds by saturating unpaired electrons, thereby significantly reducing surface recombination velocities.²⁰³ As a result, Si–H terminated surfaces often exhibit improved carrier lifetimes and enhanced electronic performance.²⁰⁴ For instance, depositing a 50 nm-thick a-Si: H film on pyramidal silicon samples at a substrate temperature of 180 °C and a gas ratio of [H₂]/[SiH₄] = 3 achieved a maximum effective minority carrier lifetime of 2.3 ms, compared with 1.6 ms obtained at [H₂]/[SiH₄] = 0. This comparison clearly demonstrates the significant improvement in surface passivation and hydrogenation efficiency under optimized conditions. However, this passivation is thermally unstable; Si–H bonds can be easily displaced by ambient oxygen or moisture, leading to oxidation and a gradual degradation of surface quality over time.²⁰⁵

Si–O terminations arise from natural oxidation in air or deliberate exposure to oxygen plasma. These oxygen-containing species enhance the long-term chemical stability of the silicon surface by forming a thin native oxide or interfacial SiO_x layer. Due to the significantly enlarged surface area and high-density surface states introduced during micro-/nanostructure formation, black silicon exhibits a faster oxidation rate and generally develops a thicker SiO₂ layer than planar silicon under identical oxidation conditions.^206,207 The presence of Si–O bonds gives rise to fixed charges at the Si/SiO₂ interface.²⁰⁸ These charges are immobile within the oxide, originating from excess silicon ions near the interface. Under specific oxidation conditions, the fixed positive charge density at the Si/SiO₂ interface varies with crystallographic orientation, following an approximate ratio of 3 : 2 : 1 for (111) : (110) : (100) surfaces.²⁰⁹ Such fixed charges modify the band bending and surface potential, thereby influencing the electronic properties of the silicon surface. In many cases, oxygen termination enhances surface depletion and alters carrier accumulation dynamics by expanding the space charge region. While beneficial in some device contexts, such charge effects must be carefully managed to avoid unintentional electrostatic barriers or drift. H-terminated silicon surfaces are hydrophobic,²¹⁰ exhibiting contact angles (θ) near 90°, whereas sufficiently thick SiO₂ layers (≥ 3 nm) are hydrophilic, with θ approaching 0°. For intermediate oxide thicknesses, θ falls between these extremes. The hydrophilicity is likely due to surface-bound hydroxyl (–OH) groups, which serve as active sites for water adsorption.

Si–F terminations, on the other hand, are commonly introduced during fluorine-based dry etching processes such as SF₆ or CF₄ plasma treatments. F atoms can bond with surface silicon atoms, resulting in enhanced hydrophobicity and chemical inertness. These properties are particularly advantageous for environments requiring corrosion resistance or reduced surface contamination. Si–F bonds on the silicon surface also facilitate hole capture and electron release, promoting surface charge-transfer processes.²¹¹ However, high F coverage may interfere with subsequent chemical functionalization steps, including molecular grafting, wet oxidation, or surface doping, potentially limiting integration flexibility in post-fabrication processing.

Overall, the type and stability of these surface terminations must be carefully considered when designing high-performance optoelectronic devices, as they directly affect surface electronic states, band alignment, and chemical compatibility. For example, the newly etched surfaces of porous silicon are predominantly terminated with hydrides, which are highly reactive in aqueous and atmospheric environments and thus prone to oxidation. To improve chemical stability and tailor the surface functionality, freshly etched silicon is often subjected immediately to post-treatment strategies such as thermal oxidation, silanization, carbonization, or functionalization via covalent bonding of organic groups.^212,213 To address this issue, Guo et al. demonstrated that thermal processing of porous silicon with acetylene gas yields two types of stabilized terminations: hydrogen-terminated thermally hydrocarbonized porous silicon (THCpSi) and hydroxyl-terminated thermally carbonized porous silicon (TCpSi).²¹⁴ THCpSi is formed by heating etched porous silicon in N₂/acetylene at 525 °C, while TCpSi is generated through an additional high-temperature step at 800 °C, resulting in a carbon-rich and hydroxyl-terminated surface. These modified surfaces retain the original nanostructural morphology, including pore size, depth, and distribution. However, they provide large surface areas and tunable chemical properties. Both THCpSi and TCpSi enable rapid charge transport and present versatile surface chemistry suitable for the grafting of diverse functional groups, depending on the application. Contact-angle measurements reveal that freshly etched porous silicon exhibits a contact angle of 90 ± 4°, whereas THCpSi is markedly hydrophobic with a contact angle of 122 ± 4°, and TCpSi becomes highly hydrophilic with a contact angle of only 31 ± 4°. Notably, TCpSi shows enhanced charge-transfer efficiency due to its higher Si–C composition, along with improved chemical stability attributed to increased Si–O content, which helps resist surface oxidation. These findings highlight the importance of engineering surface terminations not only for passivation but also for achieving optimal chemical robustness and interfacial behavior in functional nanostructures.

Doping and dopant redistribution effects. In addition to chemical terminations, micro- and nanostructuring processes can significantly alter the spatial distribution and activation states of dopants in silicon, often in unintended or uncontrolled ways. During high-energy fabrication techniques such as RIE, femtosecond-laser processing, or localized thermal annealing, dopants can diffuse, segregate, or become reactivated in surface and subsurface regions. This redistribution may result in localized shifts in carrier concentration, unintended p–n junction formation, or gradient doping profiles that deviate from the original design.

Unintentional dopant activation or diffusion near the silicon surface can significantly modulate the built-in electric field and carrier transport behavior across semiconductor interfaces. For instance, during surface restructuring or high-energy processing, donor-type impurities such as titanium (Ti) can be introduced into silicon.²¹⁵ Typically, a thin Ti film is first deposited onto the silicon surface using electron-beam evaporation. Subsequent nanosecond laser irradiation melts the silicon surface, allowing Ti atoms to diffuse into the molten silicon lattice and become electrically activated. Ti atoms form bonds with Si and act as donors, imparting n-type conductivity to the resulting Si: Ti layer. High concentrations of Ti impurities cause severe deformation of the silicon lattice during nanosecond-laser melting, accompanied by Ti over-doping, leading to changes in the Si band structure of the Si:Ti layer. These structural changes alter the electronic band structure, enabling photo-induced carrier transitions between the valence band, conduction band, and an intermediate impurity band introduced by Ti. As a result, the Si:Ti layer exhibits enhanced carrier density and a broadened photoresponse range, extending from the ultraviolet to the near-infrared spectrum (200–2500 nm). This modulation of the band structure through hyperdoping presents opportunities for broadband optoelectronic applications but also highlights the need for precise control of impurity levels and their electronic activation near the surface.

Moreover, inhomogeneous dopant profiles can introduce spatially varying band bending and localized carrier accumulation or depletion, which in turn compromise device uniformity, induce nonlinear responses, and increase recombination losses in specific regions. This issue is particularly pronounced in laser-etched black silicon,²¹⁵ where the process often introduces extremely high dopant concentrations (usually >5.9 × 10¹⁹ cm⁻³) into the outer layers of the tip regions of micron-sized cones. Such excessive doping can trigger Auger recombination and lead to poor dopant activation due to structural defects such as dislocations, vacancies, or amorphous phases formed during processing. These defects serve as recombination centers or trap states, impeding charge carrier mobility and reducing device performance. Furthermore, non-uniform doping may cause low shunt resistance and elevated leakage currents,²¹⁶ which are especially detrimental in high-performance photodiodes or three-dimensional integrated systems. In these advanced applications, even nanometer-scale dopant inhomogeneities at the surface can lead to substantial variations in device behavior, increased electrical noise, and operational instability under long-term bias stress.

To mitigate these challenges, dopant redistribution must be carefully controlled through process optimization.^188,217,218 This includes limiting thermal budgets during fabrication, employing conformal passivation layers to suppress out-diffusion, and utilizing low-damage etching techniques to preserve dopant profiles. In some cases, intentional doping via ion implantation or in situ techniques may be used to compensate or realign carrier concentrations at critical interfaces.^219–221 Ultimately, precise dopant management is essential not only for preserving the desired electronic structure but also for ensuring predictable and repeatable performance in micro/nano-scale silicon devices.

3.2.3. Passivation. To mitigate the detrimental influence of surface and subsurface defect states, as well as unstable chemical terminations, various passivation strategies have been developed to improve the electronic and chemical stability of black silicon.^222,223 These strategies are critical for enhancing the long-term stability, reproducibility, and reliability of black silicon devices, especially in photovoltaic and photodetector systems where surface recombination must be minimized. Together, the interplay between structural morphology and surface chemical states governs the multifunctional behavior of black silicon. A holistic understanding and control of both aspects are essential for maximizing performance and expanding the range of viable applications.

The effectiveness of passivation arises from two key mechanisms:^{48,204,224–226} chemical passivation, which neutralizes dangling bonds, and field-effect passivation, which repels minority carriers through built-in electric fields induced by fixed charges. Taking SiO₂ and HfO₂ as examples, the passivation mechanisms are shown in Fig. 22(a).²²⁷ Together, they significantly enhance carrier lifetime and reduce non-radiative recombination. While both mechanisms contribute to carrier lifetime enhancement, chemical passivation is generally considered more effective when targeting defect-state-induced recombination. The common passivation methods and comparisons are shown in Table 6. Thermal oxidation, which forms stable SiO₂ layers, and plasma-enhanced treatments (e.g., nitridation) are conventional methods that offer improved chemical stability and interface control. However, more advanced strategies such as ALD have demonstrated superior results, particularly for nanostructured surfaces. ALD enables the conformal deposition of high-k dielectrics like Al₂O₃ or HfO₂, which provide not only chemical passivation but also strong field-effect passivation due to their favorable fixed charge characteristics. ALD's intrinsic benefits—such as atomic-level thickness control, excellent step coverage, low defect density, and high uniformity over large areas, make it particularly suitable for black silicon architectures with complex surface morphologies.

Fig. 22 Passivation mechanisms of black silicon and double-layer ALD passivation. (a) Chemical passivation and field effect passivation of black silicon.²²⁷ Reproduced with permission from ref. 227. Copyright 2023, Elsevier. (b) Low-magnification TEM image of Al₂O₃/TiO₂ double-layer on SiNWs. (c) High-magnification TEM image of Al₂O₃/TiO₂ double-layer on SiNWs.²²⁸ Reproduced with permission from ref. 228. Copyright 2015, American Chemical Society.

Table 6 Common passivation methods and comparisons

Passivation technique/material	Passivation mechanism	Advantages	Disadvantages	Applicability to black silicon	Typical deposition method
Si–H surface hydrogenation	Chemical passivation	Effectively saturates dangling bonds; simple process	Thermally unstable; easily desorbed	Useful for temporary or auxiliary passivation	HF treatment, H2 plasma passivation
Hydrogenated silicon nitride (SiNx: H)	Field-effect + partial chemical passivation	Low-temperature process; negative fixed charges; also acts as an anti-reflection layer	High interface defect density; often requires an SiO2 interlayer	Commonly used in stacked passivation structures	PECVD
Thermal oxidation (SiO2)	Chemical + field-effect passivation	Reliable and well-established; high interface quality; excellent thermal stability	Requires high temperatures (>800 °C); may damage black silicon nanostructures	Effective if black silicon morphology is preserved	Dry/wet thermal oxidation
Al2O3	Strong field-effect + chemical passivation	High negative fixed charge density; excellent for p-type black silicon; good interface passivation	Poor conformity on high-aspect-ratio surfaces; post-deposition annealing needed	Research hotspot; excellent passivation performance	ALD, thermal decomposition, spray coating
HfO2	Field-effect passivation	High dielectric constant; strong field effect; good thermal stability	Complex process; interface defect density needs improvement	Emerging material with potential for high-performance devices	ALD
SiO2/SiNx bilayer structure	Combined chemical + field-effect passivation	Combines the advantages of both layers; widely used in commercial solar cells	Interface control complexity; interlayer stress issues	Mature and reliable; suitable for photovoltaic devices	Thermal oxidation + PECVD

---

Chemical passivation reduces interface trap density by saturating surface dangling bonds, thereby suppressing non-radiative recombination. Wong et al. studied methyl-terminated (CH₃–) Si(111) surfaces prepared via halogenation/methylation,²²⁹ the CH₃– groups covalently bond to surface silicon atoms. While structurally similar to H-terminated surfaces, CH₃–Si(111) exhibits superior resistance to oxidation in air and aqueous environments. Although both Si–H and Si–C bonds show comparable thermal stability in vacuum, CH₃– provides a higher kinetic barrier against oxidation. Unlike H–Si(111), whose passivation degrades under ambient conditions, CH₃– maintains electronic passivation more stably. Furthermore, alkyl groups induce a surface dipole, modulating the electron affinity and charge transfer at the silicon interface. As a result, CH₃–Si(111) demonstrates improved carrier lifetime and surface recombination properties.

In parallel, field-effect passivation provides an additional suppression mechanism by introducing fixed charges within the passivation layer, typically through oxide or dielectric films that establish a built-in electric field. This field repels minority carriers away from the interface, lowering their concentration at the surface and effectively reducing the recombination rate.

More advanced methods include the ALD deposition of high-k dielectrics such as Al₂O₃ or HfO₂, which not only passivate the surface but also contribute to favorable band alignment and field-effect passivation. Among ALD materials, Al₂O₃ is widely used due to its intrinsic negative fixed charges (10¹²–10¹³ cm⁻²)^173,230,231—originating from oxygen interstitials or aluminum vacancies—provide strong field-effect passivation. These negative charges repel minority carriers in p-type or p⁺-type silicon, thereby suppressing surface recombination. For n-type silicon, ALD HfO₂ is a more suitable alternative owing to its positive fixed charge and higher dielectric constant (∼25).²³² Its higher refractive index (∼2.1) also enables simultaneous antireflection and passivation, making it promising for efficient black silicon solar cells.

To suppress charge injection into defect states within Al₂O₃, a thin interfacial layer of ALD-grown SiO₂ can be introduced. This interlayer effectively reduces the generation of negative fixed charges within the Al₂O₃ layer.^233,234 The SiO₂ layer functions as an electron tunneling barrier, preventing charge injection into trap states within Al₂O₃. When the interlayer is sufficiently thick, the overall fixed charge density in the SiO₂/Al₂O₃ dielectric stack can even shift to positive values due to the combined effect of fixed and bulk charges in the SiO₂ film. This SiO₂/Al₂O₃ stack not only alters the effective charge polarity but also enhances both optical and electrical performance by reducing mid-gap interface defects, partly due to hydrogen diffusion during post-deposition annealing. In addition, the TiO₂ layer has a high positive charge density,²²⁸ which can adjust the negative charge of Al₂O₃ to make the double layer structure appear positive and reduce the defect density. The TEM images are shown in Fig. 22(b) and (c).

The influence of various plasma treatments during the ALD process of Al₂O₃/SiO₂ multilayers on silicon surface passivation has been systematically investigated.²³⁵ Among them, the most significant improvement was achieved by applying H₂ plasma treatment after SiO₂ deposition, reducing the surface recombination parameter to as low as 0.35 fA cm⁻² on (100)-oriented n-type silicon with a resistivity of 10 Ω cm—over five times lower than the value obtained from Al₂O₃ single layers without plasma treatment. Different plasmas (H₂, N₂, O₂, N₂/H₂ mixtures, and Ar) were compared, with particular emphasis on interface modification and dipole formation at Al₂O₃/SiO₂ interfaces, which strengthen field-effect passivation. Post-deposition annealing further enhances passivation by reducing interface defects. At annealing temperatures below 450–500 °C, the effective minority carrier lifetime increases markedly with temperature. At higher temperatures, the lifetime tends to saturate, and the difference between treated and untreated samples becomes more pronounced. Notably, prolonged H₂ plasma exposure yields the most substantial enhancement in surface passivation.

In MACE-fabricated black silicon, post-etch cleaning and passivation are vital for improving surface quality and electronic performance. The standard RCA cleaning protocol,²²⁵ particularly SC-1, effectively removes porous silicon with high defect density by etching fragile nanostructures while preserving larger features, thereby smoothing the surface—an effect comparable to thermal annealing.²³⁶ Despite the increase in surface defect density post-etching (up to ∼7-fold), SC-1 treatment can reduce this to ∼3-fold and significantly enhance minority carrier lifetime by tenfold. Partial passivation is also achieved via residual Si–H and Si–O bonds, with trapped hydrogen remaining stable over months. For further passivation, ALD-deposited Al₂O₃ provides a uniform, conformal coating even on complex nanostructures. Its high density of negative fixed charges induces strong field-effect passivation, depleting minority carriers at the interface and mitigating the impact of remaining defects. Thus, Al₂O₃ offers both chemical and field-effect passivation, making it a preferred choice for stabilizing MACE-processed black silicon.

Beyond Al₂O₃ and HfO₂, other ALD-deposited oxides like ZnO, TiO₂, and Sc₂O₃ offer promising results. Song et al. demonstrated that ALD ZnO (∼4 nm) could effectively passivate defect states on nanostructured silicon and modulate hole transport through enhanced surface band bending.²³⁷ As the ZnO thickness increased, Si–O coordination likely contributed to further chemical passivation at the Al/nanostructured-Si interface, resulting in ∼24% improvement in power conversion efficiency via the formation of a passivating tunnel contact. In addition, ALD-deposited dielectric oxides such as Al₂O₃, TiO₂, HfO₂, and Sc₂O₃ have shown promise for both optical and electrical optimization.^238,239 Finite-difference time-domain (FDTD) simulations revealed improved anti-reflective behavior across a broad spectrum, emphasizing the importance of selecting optimal film thicknesses.

In large-scale applications, SiO₂/SiN_x bilayer passivation has enabled black silicon solar cells with higher performance.^{186,240–242} SiO₂ and SiN_x are generally prepared by thermal oxidation and plasma-enhanced chemical vapor deposition (PECVD). The ratio of Si to N in SiN_x has a significant impact on performance.²⁴³ Si-rich SiN_x films exhibit increased Si–Si and Si–H bonds, enhancing chemical and bulk passivation by saturating dangling bonds and facilitating hydrogen diffusion during annealing. The reduction in N–H bonds and nitrogen back-bonding lowers fixed positive charge density, weakening field-effect passivation. In contrast, N-rich SiN_x films possess higher fixed charge density and stronger field-effect passivation but suffer from increased interface defect density, limiting chemical passivation. For p-type silicon, SiN_x induces strong inversion, whereas SiO_x yields only depletion or weak inversion.

In summary, effective surface passivation is essential for mitigating surface recombination and unlocking the full optoelectronic potential of black silicon. From chemical cleaning and hydrogen termination to advanced dielectric coatings such as Al₂O₃, SiO₂, and SiN_x, each strategy offers distinct advantages rooted in chemical and field-effect mechanisms. The interplay between fixed charges, interface defects, and hydrogen dynamics defines the overall passivation performance. As black silicon technologies continue to evolve toward high-efficiency and low-cost applications, optimizing passivation schemes tailored to specific device architectures and fabrication routes will remain a critical research focus.

4. Tunable properties and functional applications enabled by micro–nano surface engineering: physical, chemical, and biological

The micro- and nanostructures on the surface of black silicon fundamentally alter its material properties, enabling a wide range of advanced functionalities. These structures enhance light absorption by inducing multiple internal reflections within the micro- and nanostructures, significantly reducing surface reflectivity. When doped with impurities that introduce new energy levels, black silicon's absorption spectrum extends into the infrared region, further broadening its spectral response. The morphological modifications not only enhance optical properties but also improve optoelectronic efficiency, field emission characteristics, and photocatalytic activity.²⁴⁴ Moreover, black silicon's unique surface structures have enabled applications in antimicrobial coatings,²⁴⁵ superhydrophobic surfaces, and surface-enhanced Raman scattering (SERS) devices, demonstrating its versatility across various fields.

4.1. Optical properties

The most prominent feature of black silicon is its greatly enhanced light absorption, which is a direct result of its unique surface morphology. The micro- and nanostructures on the silicon surface facilitate multiple light-scattering and trapping effects,^99,246–248 significantly reducing surface reflectance and enabling near-total absorption across a wide wavelength range. Beyond this, the altered surface also introduces a high defect density that promotes carrier recombination, making black silicon an effective photoluminescent material. These surface-induced changes not only improve its optical properties but also extend its potential for diverse applications. Extensive research has focused on understanding the relationship between the surface morphology of black silicon and its optical behavior.

4.1.1. Optical absorption. The nanopores and micropores on black silicon surfaces play a critical role in enhancing its optical properties by altering light behavior and improving absorption efficiency.¹⁵⁹ On a flat silicon surface, as shown in Fig. 23(a), light follows a single absorption path, leading to significant reflection losses. In contrast, nanopores with dimensions comparable to optical wavelengths (Fig. 23(b)) induce multiple refractions and transmissions, effectively reducing surface reflectivity and trapping more light. Micropores, which are larger than the incident light wavelength (Fig. 23(c)), enhance light absorption through geometric-optics effects. Their vertical sidewalls facilitate continuous reflection along the depth of the pores, allowing more light to re-enter the silicon. The optical path direction in the optical absorption layer will shift due to the surface microstructure. At the same time, the reflection probability of each interface will increase due to the angular movement. This effect increases the optical path length and improves light-matter coupling.²⁴⁹ Additional nanopores at the bottom of the microstructure (Fig. 23(d)) absorb residual light and tilt the reflected beams, further extending the optical path. The combination of micro- and nanoscale pores creates an ordered–disordered multiscale structure that synergistically increases light trapping and absorption, resulting in significantly enhanced optical performance.^65,159 Zhang et al. used a two-step RIE process to achieve nanopores on chimney-like microstructures.⁶⁵ The reflection spectra in Fig. 23(e) reveal that silicon samples with chimney and nanostructures exhibit lower reflectance than planar silicon. This micro–nano design effectively reduces reflectance in the UV-visible range but not in the NIR range.

Fig. 23 Optical absorption properties of black silicon. (a)–(d) Light absorption by different micro- and nano structures in black silicon. (a) Polished silicon. (b) Nanostructured silicon. (c) Microstructured silicon. (d) Nano and micro hybrid black silicon. (e) Reflection spectra of planar and textured silicon surface.⁶⁵ Reproduced with permission from ref. 65. Copyright 2020, John Wiley & Sons. (f) Polarization insensitivity of black silicon.⁹⁰ Reproduced with permission from ref. 90. Copyright 2018, Elsevier. (g) Reflection spectra of black silicon processed at different annealing temperatures for two-step MACE.¹²² Reproduced with permission from ref. 122. Copyright 2014, Springer Nature. (h) Reflection spectra of black silicon with/without QDs.²⁵⁰ Reproduced with permission from ref. 250. Copyright 2020, Elsevier.

Doping during the laser ablation process improves the light absorption properties of black silicon. The optical absorption of black silicon prepared in corrosive gases reaches 95% for photon energies above the bandgap, whereas samples processed in inert gases exhibit absorption of 90%.⁷³ NIR broadband light absorption is achieved when the sulfur concentration exceeds 0.6% (about 10²⁰ atoms cm⁻³). For wavelengths exceeding 1100 nm, the absorption of black silicon fabricated in an SF₆ atmosphere decreases slightly (by ∼5%). In contrast, in samples prepared in other gas environments, the absorption drops sharply, even below that of flat silicon, indicating very poor infrared absorptivity. Sulfur impurity doping is confirmed as the key to enhanced infrared absorption.^58,72,251 Other doping elements, such as tellurium (Te), selenium (Se), N, Au, etc., and their effects will be described in Section 4.3.

Simulation studies show that by adjusting the doping level and surface morphology, the light absorption range of black silicon can be extended to wavelengths as long as 20 µm.⁶⁷ Black silicon with a high aspect ratio (∼30 : 1) and heavy doping forms an efficient conical metasurface that suppresses reflection across a broad spectral range. The absorption exceeds 99.5% in the range of 1–8 µm and remains beyond 90% up to 20 µm.

The light absorption of black silicon depends on the incident wavelength and is not sensitive to factors such as light polarization.⁹⁰ Combined with its low reflectivity and high sensitivity, these features make black silicon a promising material for broadband optoelectronic applications. As shown in Fig. 23(f), the surface roughness enhances scattering and strengthens the interaction among light waves with different polarization states, enabling the structure to behave effectively as a homogeneous medium.

4.1.2. Light–matter interactions enhanced by structure/nanoparticle hybridization. Metal nanoparticles integrated with the microstructures on black silicon enhance light interaction through localized surface plasmon (LSP) effects, significantly boosting plasmon resonance and optical performance, the absorption rate and absorption range can be further increased.^{13,65,252,253} When incident light strikes the metal nanoparticles, free electrons oscillate collectively with the electric field, causing the electron cloud to shift away from the nucleus. The coulomb force then restores the displaced electrons, generating localized surface plasma oscillations. When the frequency of the incident light matches the LSP frequency, light absorption by the metal nanoparticles is greatly amplified. The LSP frequency depends on the size, shape, and arrangement of the metal nanoparticles, as well as the dielectric constant of the surrounding material. Additionally, interactions between adjacent nanoparticles can lower the resonance frequency, leading to a redshift in the plasmonic response.²⁵⁴ These effects collectively enhance the light-trapping capabilities and optical performance of black silicon.

Zhang et al. investigated the effect of Ag nanoparticles on the near-infrared light absorption of black silicon fabricated by two-step MACE.¹²² Ag nanoparticles were obtained by rapid thermal annealing of Ag films, with the increase of annealing temperature, the size of the nanoparticles grew and the interparticle spacing increased, thereby affecting the reflectivity, as shown in Fig. 23(g). After plasma treatment, black silicon exhibited a maximum absorption of 93.6% within the near-infrared spectrum (820–2500 nm) and maintained an average absorption of 91.8% across the range of 250–2500 nm. The extensive absorption in the near-infrared region is primarily attributed to plasmon resonance frequency shifts due to particle size, high refractive index, and interparticle coupling. The uneven distribution of Ag nanoparticles results in the plasma-enhanced absorption peak in the NIR range being a composite of multiple peaks. As the incident wavelength lengthens, the decay rate of free electrons in the dielectric constant of Ag accelerates, shortening the plasma lifetime and widening the resonance bandwidth.

Both single-metal and bimetallic nanoparticles can significantly enhance anti-reflective performance.¹³ Bimetallic nanoparticles exhibit superior anti-reflective performance across a broader spectral range compared to single-metal nanoparticles. This is attributed to their wider size and shape distribution, higher surface roughness, and multiple surface plasmon resonance effects. To investigate the anti-reflective effects of metal nanoparticles, researchers modified laser-textured silicon surfaces with Ag, Ag–Au bimetallic, and Cu–Ag–Au trimetallic nanoparticles via thermal annealing. Within the wavelength range of 300 to 1200 nm, the average reflectance of the laser-textured Si surface was 8.3%, while the average reflectance of surfaces decorated with Ag–Au and Cu–Ag–Au nanoparticles decreased to 6.9% and 5.5%, respectively. The decoration with ternary metal NPs resulted in better broadband anti-reflective performance.

In addition, quantum dots play an important role in improving light absorption. For instance, silicon quantum dots (SiQDs)/black silicon hybrid nanostructures reduce reflection across a broad spectrum (300–1000 nm),²⁵⁰ and the reflection spectrum is shown in Fig. 23(h). The SiQDs with an average diameter of 1.8 ± 1.1 nm can convert the UV light into visible light, reducing reflection from 9.9% to 6.5% at 600 nm, with even greater reductions at longer wavelengths. In another research, SiC colloidal quantum dots were used as down-conversion materials to perform the same role as SiQDs,²⁵⁵ converting UV light into visible light. Additionally, quantum dots coated on the surface of SiNWs increase the photoconductivity effect, reducing the surface recombination rate.

4.1.3. Luminescence. The structural defects and impurities on the black silicon surface provide an ideal platform for charge carrier recombination, resulting in varying luminescence properties depending on the fabrication method. Black silicon fabricated via femtosecond-laser irradiation is reported to emit light at 600 nm and 680 nm, corresponding to orange and red light, respectively.²⁵⁶ As the laser scanning speed increases (Fig. 24(a)), the photoluminescence (PL) intensity decreases, accompanied by a redshift in the emission peak. The influence of laser fluence on luminescence can be divided into two stages, as shown in Fig. 24(b). Initially, the PL intensity increases with fluence, and the PL peak appears near 680 nm. Beyond an optimal fluence level, the PL intensity declines sharply. SiO_x is formed in silicon microstructures by laser irradiation. The oxidation process of silicon can be intensified either by increasing the number of laser pulses focused on the same area or by boosting the laser fluence. The modulation of luminescence characteristics by laser scanning speed and fluence is primarily governed by changes in the oxygen-silicon (O/Si) ratio. As the scanning speed increases and the laser fluence decreases, the O/Si ratio decreases. The orange and red PL bands arise from different mechanisms—defect states and quantum confinement, respectively. As illustrated in Fig. 24(d) and (e), under low-power excitation, microstructures generated at high scanning speeds or high fluence conditions predominantly exhibit orange PL. In contrast, structures fabricated under low scanning speeds or low fluence conditions tend to emit red band PL. Notably, under 1 mW excitation power, the PL spectrum is characterized solely by a red emission band. Moreover, orange-band PL is highly susceptible to quenching under 532 nm illumination, whereas red-band PL remains stable due to its distinct underlying mechanism.

Fig. 24 PL spectra of black silicon under different femtosecond-laser processing parameters. (a) Different scanning velocities. (b) Different laser fluences.²⁵⁶ (c) Unannealed sample and samples annealed at different temperatures.²⁵⁷ (d)–(e) O/Si atomic ratio for different scan velocities and laser fluence.²⁵⁶ (f) PL spectrum of laser-treated black silicon before and after KOH etching.²⁵⁷ (a), (b), (d) and (e) Reproduced with permission from ref. 256. Copyright 2011, AIP Publishing. (c) and (f) Reproduced with permission from ref. 257. Copyright 2013, Optica Publishing Group.

The black silicon fabricated by RIE has PL bands in both the visible and infrared regions.²⁵⁸ As the temperature drops from room temperature to 10 K, the visible light bands show a large blue shift from 696 nm to 560 nm. In the infrared region, PL bands centered at 1.14 µm and 1.53 µm are observed, with the former attributed to band-to-band carrier recombination in silicon and the latter to defect-related luminescence.

PL signals were observed only in thermally annealed samples, as shown in Fig. 24(c), highlighting the critical role of annealing in luminescence devices.²⁵⁷ Structural defects and impurities induce a non-radiative recombination process that quenches the PL signal through mobility-induced band-tail state quenching.²⁵⁹ As the annealing temperature increases, the PL intensity rises significantly. This is attributed to strain during annealing, which alters impurity and defect energy levels and can even modify the bandgap structure of black silicon. Fig. 24(f) shows the PL spectrum of laser-treated black silicon before and after KOH chemical etching. After etching, sulfur impurities on the surface layer are removed, and the PL intensity increases, indicating that the luminescence of black silicon is not caused by impurities. Additionally, high-temperature annealing reduces non-radiative defect channels, improving optical performance by allowing more carriers to occupy shallow energy levels near the surface, thereby enhancing luminescence.

The surface plasma treatments can also enhance the luminous efficiency.^252,260 Black silicon and Ag nanoparticles were employed to improve the electroluminescence (EL) intensity of silicon nanocrystals (Si-ncs) in light-emitting diodes (LEDs).²⁶⁰ The black silicon was formed by a two-step MACE, and then the Si-ncs were deposited on it. Samples etched for 5 minutes had the lowest reflectivity and the highest surface roughness. Ag nanoparticles generate surface plasmons that enhance the local electromagnetic field, increasing the excitation probability of Si-ncs and thereby boosting EL intensity. Simultaneously, the textured surface of black silicon enhances charge injection efficiency and light emission by reducing surface reflection and improving light trapping. Specifically, the EL intensity of Si-ncs LEDs increased by 3.2 times with black silicon etched for 5 min and by 4.9 times with Ag nanoparticles annealed at 200 °C. Combining black silicon and Ag surface plasmons achieved a maximum EL enhancement of 8.5-fold, showcasing the synergistic effect of morphology optimization and plasmonic enhancement.

4.2. Electrical properties

Black silicon exhibits distinctive electrical characteristics that are crucial for assessing its performance across various applications. Its key properties include parameters such as carrier concentration, carrier mobility, activation energy, and impurity concentration. These attributes determine the material's efficiency of electrical conduction and photoelectric interactions, directly impacting its performance in photovoltaic solar cells, photodetectors, and sensors.

The surface morphology of black silicon significantly affects its electrical properties. Increased surface roughness leads to greater carrier scattering, reducing carrier mobility. Additionally, the high surface area introduces more defect states that trap carriers, further hindering carrier migration. Heavily doped black silicon films with supersaturated sulfur impurities demonstrate low carrier activation energies, resulting in a weak temperature dependence of carrier concentration and mobility.^261,262 Moreover, black silicon generally has a shorter carrier lifetime than conventional silicon. This reduction is mainly attributed to the quantum confinement effect—where electron and hole wave functions overlap in a confined volume and the high defect density in the surface microstructure.¹⁶⁴ These nanostructures increase carrier recombination rates at the surface, influencing the minority carrier lifetime.

By carefully controlling surface morphology, researchers can optimize the electrical properties of black silicon to align with the specific needs of diverse applications. Thermal annealing is a common method employed to extend the carrier lifetime, but it often induces metal contamination, dopant diffusion, and other adverse effects. As a result, the lifetime of silicon after laser radiation cannot be restored solely through thermal annealing. Chemical acidic etching after laser treatment is an effective method for eliminating laser damage. This process can significantly increase the minority carrier lifetime. The carrier lifetime of black silicon samples with different laser fluences (Fig. 25(a)) was measured by quasi-steady-state photoconductance (QSS-μPCD).⁷⁹Fig. 25(b) shows the minority lifetime maps of the samples. The carrier lifetime of black silicon gradually increases with prolonged etching time, indicating that the laser-induced defects are confined near the surface, and the lifetime remains unchanged after 30 s. A smaller laser fluence leads to a longer carrier lifetime, but extended etching times may reduce light absorption efficiency. A more intuitive result is shown in Fig. 25(c). Therefore, it is necessary to find an optimal balance between these factors. An etching duration of 10 s achieves a high absorption rate of 90.5% and a carrier lifetime of 2 ms. Similarly, significant surface damage caused by pattern transfer during plasma etching can also be eliminated through appropriate damage/defect removal etching,^72,263 thus effectively extending carrier lifetime.

Fig. 25 The change of lifetime with different etching and passivation conditions. (a) A photograph of a laser-treated silicon wafer, with dark circles in the upper-left, upper-right, and lower-left indicating laser fluences of 5.4, 4.3, and 3.2 kJ m⁻², respectively. (b) Lifetime maps of QSS-μPCD with varying etching time in (a). (c) Variation of maximum absorptance (250–1100 nm) and effective minority-carrier lifetime at an excess carrier density of 1.0 × 15 cm⁻³ as a function of etching time.⁷⁹ Reproduced with permission from ref. 79. Copyright 2022, IEEE. (d) SEM image of black silicon after Nafion solution passivated. (e) Minority carrier lifetime of black silicon before and after Nafion passivation.⁴⁶ Reproduced with permission from ref. 46. Copyright 2022, John Wiley & Sons.

Surface passivation has also been demonstrated as an effective method for extending the carrier lifetime of black silicon. While long nanowires provide high specific surface areas, they introduce numerous dangling bonds that can impair optical and electronic properties. Shorter nanostructures, in contrast, are easier to passivate, leading to superior surface passivation and enhanced device performance.²⁶⁴ Chen et al. used a Nafion solution to passivate black silicon nanowires fabricated by MACE.⁴⁶ The nanowires were 160–320 nm long with pore widths of 50–200 nm. Passivation was performed by spin-coating the Nafion solution, which filled most of the nanopores and reached the bottom of the nanowires (Fig. 25(d)), most of the outer-shell defects were passivated. After passivation, as shown in Fig. 25(e), the effective minority carrier lifetime increased from 3.8 µs to 1.1 ms for nanowires with a length of 160 nm and to 0.6 ms for nanowires with a length of 320 nm on a planar wafer. When exposed to oxygen, the lifetime further increased from 1.1 ms to 2.4 ms. A similar phenomenon was observed on pyramid-textured wafers, although the carrier lifetime was shorter. With increasing Nafion concentration, the minority carrier lifetime of long nanowires that were difficult to infiltrate initially increased and then decreased. Seref Kalem used acid vapor to smooth black silicon nanowires, effectively reducing and passivating surface defects.¹⁶⁴ During this process, the oxides on the surface were removed and new oxides were generated, but the larger defects, such as kinks and interface defects, were eliminated. This study also proves that surface recombination is more important than recombination in the bulk cores and interfaces. Other passivation layers, such as Al₂O₃, SiN_x, TiO₂, are also widely used for black silicon,^{94,125,265–267} providing lower reflectivity and higher carrier lifetime.

4.3. Optoelectronic properties

Black silicon exhibits remarkable optoelectronic properties, making it a valuable material in advanced photonic and electronic applications. The optoelectronic performance of black silicon is intricately linked to its surface morphology, including the size, shape, and spatial distribution of nanostructures, which dramatically alter light absorption, carrier generation, and transport characteristics.^268–270 By precisely engineering the morphology of black silicon, researchers can enhance its optoelectronic properties to suit specific applications.

4.3.1. Photodetector. At present, the development of silicon-based detectors has been relatively mature; however, they continue to face challenges such as low response rates and limited spectral response range.^271–274 The introduction of black silicon provides an effective pathway to overcome these limitations. Its nanostructures significantly enhance light absorption by reducing reflection and improving sensitivity across a broad spectral range, from UV to NIR. And the Shockley–Queisser (SQ) limit can be broken by carrier multiplication processes inside silicon nanostructures.²⁷⁵ These properties make black silicon ideal for high-efficiency photodetectors, particularly in applications requiring broadband operation or high sensitivity under low-light conditions.

Doping is a highly effective method to extend the absorption range of black silicon. By introducing impurities, additional new energy levels are created within the silicon bandgap, enabling black silicon to absorb photons with lower energy, including those in the infrared range. This expanded absorption spectrum enhances the overall light-harvesting efficiency, making the material more suitable for applications like infrared photodetectors and solar cells. However, superdoping will lead to a high impurity concentration (5 × 10¹⁹ cm⁻³) in silicon,^45,276 which will produce a large number of background-free carriers, and strong free carrier absorption does not support photoelectric conversion. At the same time, impurity atoms will also cause strong ionized impurity scattering, resulting in a lower light absorption rate. Zhao et al. developed an N⁺–N⁻ photodetector by laser-modified silicon and ordinary silicon,⁴⁵ achieving a high response in the NIR range. The absorptance curves are shown in Fig. 26(a), and the inset is the image of the N⁺–N⁻ photodetector. The responsivities of different reverse biases are shown in Fig. 26(b), and the device exhibited lower responsivity compared to commercial silicon photodetectors under zero bias. However, as the reverse bias increases, the responsivity will be enhanced, attributed to the photoconductive gain. Au has high solid solubility when doped into silicon and can be uniformly distributed with a stable valence, effectively reducing the lattice mismatch and dislocation centers.²⁷⁷ After annealing at 800 °C, the black silicon has an absorption of 95% across 200–1100 nm range. The intermediate energy level introduced by over-doping still maintains a 75% absorption in the 1100–2500 nm range. The photodetectors’ photoelectric response and on/off ratio in the infrared band are more than 10 times higher than those of unprocessed silicon.

Fig. 26 Characterization of photoelectric properties of black silicon. (a) Absorptance of pure Si substrate, laser-modified silicon (M-Si) samples before and after annealing. The inset is the schematic and real image of the M-Si photodetector. (b) Responsivity of M-Si photodetector at different reverse biases, the commercial silicon-PD is shown inset for reference.⁴⁵ Reproduced with permission from ref. 45. Copyright 2018, IEEE. (c) Influence of RTA and FTA, the inset shows the changes of peak responsivity with temperature.¹⁹⁹ Reproduced with permission from ref. 199. Copyright 2020, John Wiley & Sons. (d) Schematic of Nafion passivated black silicon solar cells. (e) J–V curve with and without Nafion passivated planar black silicon solar cell.⁴⁶ Reproduced with permission from ref. 46. Copyright 2023, John Wiley & Sons. (f) J–t curve (HER) of silicon pyramid array/CoS₂ photocathode at 0 V, the inset is a schematic of the photocatalysis process.²⁷⁸ Reproduced with permission from ref. 278. Copyright 2020, Royal Society of Chemistry.

Adjusting the type and concentration of doped atoms is also one way to improve the performance of black silicon photodetectors. Sulfur and nitrogen impurities co-doped black silicon were successfully fabricated using a femtosecond laser in mixed N₂/SF₆ atmospheres,¹³⁶ exhibiting excellent infrared absorption stability. Increasing the nitrogen content reduces the sheet carrier density because nitrogen predominantly forms neutral pair defects. The resultant photodetector achieved a responsivity of 58 mA W⁻¹ at 1.31 µm. Te-hyperdoped black silicon photodiodes were also fabricated,⁷⁶ and the influence of substrate temperature was studied. Increasing the substrate temperature leads to a higher doping concentration gradient, improved lattice quality, and reduced defect density, which together help suppress the dark current of the photodiode. The device exhibited significantly enhanced responsivity compared with commercial silicon photodetectors, particularly in the near-infrared region.

The intrinsically high recombination rate associated with black silicon's textured surface can be mitigated by depositing passivation layers.²⁷⁹ The SiN_x/Al₂O₃ stacks have emerged as particularly effective owing to the chemical hydrogenation of the strong Si–O coordination and hydrogenation passivation at the silicon/Al₂O₃ interface.²⁸⁰ Zhao et al. reported that the black silicon prepared by sulfur-doped femtosecond laser irradiation exhibited absorptance exceeding 85% in the range of 0.2–2.5 µm due to impurity-level-induced sub-bandgap absorption.²⁸¹ Then the SiO₂ passivation layer is introduced to reduce the background free carrier concentration and structural defects, resulting in prolonging carrier lifetime and enhancing device stability, the responsivity of black silicon device reaches 367 mA W⁻¹ at a bias voltage of 10 V and a wavelength of 1030 nm, which is significantly higher than that of planar silicon (47 mA W⁻¹) under the same conditions.

Annealing is another effective method to reform the recombination rate. Huang et al. fabricated a photodetector based on the black silicon prepared by femtosecond laser irradiation in SF₆ environment,¹⁹⁹ by applying rapid thermal annealing (RTA) and hydrogenated amorphous silicon passivation, the device achieved photoelectric response up to 1600 nm. The peak responsivity (∼1100 nm) reached 215.69, 497.51, 679.84, 956.62, and 1097.60 A W⁻¹ under reverse bias voltages of 1.5, 5, 8, 15, and 20 V, respectively. The authors claimed an improvement of nearly three orders of magnitude, with particularly high responsivity maintained in the telecommunication wavelength range (1260–1600 nm) compared to commercial Si and Ge photodetectors under comparable bias (below 1 A W⁻¹ at similar wavelengths). Fig. 26(c) illustrates the performance differences between conventional furnace thermal annealing (FTA) and RTA, with the latter demonstrating superior results. RTA was found to inhibit the diffusion of sulfur atoms while activating the carrier to repair structural defect damage, and passivation can effectively inhibit the high dark current caused by suspension bond defects on the material surface. Su et al. also used RTA to improve the performance of metal-semiconductor–metal photodetectors in N₂ ambient,²⁸² the tensile stress and point defects are eliminated, leading to a huge improvement of carrier mobility, conductivity, and carrier concentration, thus yielding almost 3 orders of magnitude increase in responsivity, reaching 76.8 A W⁻¹ for the sample annealed at 673 K, compared with 0.032 A W⁻¹ for the unannealed device at a wavelength of 600 nm. Laser annealing provides an alternative route, offering localized heating that yields highly crystalline, stress-free phases while preserving the strong absorptance characteristic of black silicon.^60,283 Kearney et al. treated sulfur-hyperdoped black silicon with ultrashort laser pulses,²⁸⁴ demonstrating that ultrafast laser heating can reactivate sub-bandgap absorption that had been quenched by prior thermal annealing, while simultaneously retaining high crystallinity and stable rectifying behavior.

DRIE and PIII are combined to produce black silicon by Zhong et al.,²⁸⁵ the improved doped sulfur element introduces multiple defect levels and form impurity bands on the silicon surface, reducing the silicon bandgap from 1.12 eV to ∼1.0 eV, and the light absorption was enhanced, especially in the NIR range (800–2000 nm), the maximum absorption rate reached 83%. The silicon-PIN photodetector, featuring microstructured black silicon on its backside, demonstrated excellent device performance, achieving a responsivity of 0.53 A W⁻¹ at 1060 nm. Li et al. employed a dual-step strategy combining ion implantation with pulsed-laser reactivation to prepare Ar-doped black silicon photodetectors.²⁸⁶ Pulsed laser irradiation after ion implantation can not only serve as a post-annealing step to improve the crystal quality of the implanted layer, but also enable further Ar superdoping to enhance infrared absorption. The resulting dual-contact photodiode achieved responsivities of 0.975 A W⁻¹ at 1.31 µm and 1.28 A W⁻¹ at 1.55 µm.

4.3.2. Solar cells. In recent years, the growing prominence of environmental and energy issues has led to an intensified focus on clean and renewable energy solutions within scientific research.^287–291 Black silicon has superb light absorption across the visible to infrared spectrum has emerged as a promising material for solar cells.^{198,292–295}

The solar cell made of black silicon prepared by PIIIE for the first time in 2011 (device area: 125 × 125 mm²) has a conversion efficiency of 15.68% and a fill factor (FF) of 0.783,²² then Zhong et al. prepared nanostructured surfaces of different heights by PIIIE,⁸¹ a maximum conversion efficiency of 15.99% was achieved at a height of 300 nm (device area: 156 × 156 mm²), which was the competition result between reduced reflectivity and loss of internal quantum efficiency (IQE). It remains a disadvantage in contact with the electrode and black silicon, but the conversion efficiency can be further improved by surface passivation and preparation of better contact.

Ye et al. reported the first perovskite/silicon tandem cell using industrial-grade black silicon and tunnel oxide passivation contacts.¹⁹⁸ The silicon substrate was textured via MACE and then smoothed by alkaline treatment, which reduces the surface area and removes a large portion of defects. Black silicon undergoes surface reconstruction to provide passivation while retaining the broadband light capture. Due to its unique low polarity and high dispersion components, the reconstructed black silicon improves perovskite wettability and enhances adhesion between subcells. Perovskite grown on the reconstructed black silicon exhibits highly crystalline properties and achieves vertically aligned grain boundaries through nanoconfinement, thereby reducing unwanted carrier recombination losses and promoting charge collection at both contacts. These advancements significantly improve J_SC and FF in series, ultimately resulting in a certified PCE of 28.2%.

Sun et al. mitigated surface recombination in black silicon solar cells by incorporating silane-based additives into the PEDOT: PSS organic layer.²⁹⁶ Through covalent Si–O–Si bonding, both interfacial adhesion and overall electrical contact between PEDOT:PSS and the nanostructured silicon were improved. By adjusting the composition ratio of silane chemicals in PEDOT:PSS, a balance was achieved between light capture and surface-to-volume ratio, resulting in a maximum V_oc value of 640 mV and a PCE of 14.1% (substrate area: 15 × 15 mm²). Additionally, the lifetime of silicon minority carriers was enhanced.

In another study, blade-coated wide-bandgap perovskite films were applied to double-textured silicon at the submicron scale to achieve high-efficiency perovskite/silicon tandem solar cells.²⁹⁷ In this configuration, the perovskite cell “flattens” the textured silicon surface by filling the valleys between the pyramids. This texturing extends the path length of light within the silicon, achieving a matched J_sc exceeding 19 mA cm⁻² and an efficiency of 26%.

Surface passivation has been further shown to significantly enhance device performance. In 2015, Savin et al. prepared nanostructures approximately 800 nm in height and 200 nm in width using low-temperature DRIE,²³ and passivated the surface with a 20 nm Al₂O₃ thin film to fabricate a thick interdigital back contact (IBC) solar cell with a device area of 9 cm². The IBC cell demonstrated an open-circuit voltage (V_oc) of 665 mV, a short-circuit current density (J_sc) of 42.2 mA cm⁻², an FF of 78.7%, and a power conversion efficiency (PCE) of 22.1%. This advancement marked a breakthrough in mitigating surface recombination in black silicon and achieving high conversion efficiency without damaging light absorption, further expanding the applicability of black silicon in solar energy.

In Zhou's research, organic-passivated black silicon solar cells were fabricated on both planar and pyramid silicon wafers.⁴⁶ The device structure and corresponding I–V curves are shown in Fig. 26(d) and (e). On a planar silicon wafer, Nafion passivation significantly enhanced device performance, the V_oc, J_sc, FF, and PCE after Nafion passivation were 661.2 mV, 38.9 mA cm⁻², 82.6% and 21.2% (device area: 142 cm²), respectively, compared with the unpassivated counterparts (V_oc = 500.3 mV, J_sc = 2.6 mA cm⁻², FF = 19.4% and PCE = 1.0%). A similar improvement was observed for pyramid black silicon cells. In another study, Hsu et al. employed SiN_x and Al₂O₃ to passivate MACE-processed black silicon on a pyramidal substrate, achieving an optimal conversion efficiency of 22.9% and a remarkably low surface recombination velocity of 42 cm s⁻¹.²⁶⁴

A core–shell heterojunction structure consisting of SiNW arrays and carbon quantum dots (CQDs) was reported by Xie et al.²⁹⁸ The barrier height of this structure was 0.75 eV, providing a good rectifying effect. When used as a solar cell, the device exhibited a V_oc of 0.51 V, a J_sc of 30.09 mA cm⁻², a FF of 0.593, and a PCE of 9.10%. Graphene quantum dots (GQDs) were used as down-conversion materials in the interface layer of black silicon solar cells prepared via MACE to achieve good band alignment between graphene transparent-conductive electrodes and silicon.²⁹⁹ By doping graphene transparent-conductive electrodes, passivating the silicon back surface, and introducing GQDs into the interface layer, the GQD layer functioned not only as a hole-transport layer but also as an electron-blocking layer, thereby reducing carrier recombination within the device. The maximum PCE reached 13.66% (device area: 9 mm²). The cell also exhibited excellent long-term stability, maintaining 84% of its original PCE after being exposed to air for 30 days.

4.3.3. Photoelectric catalysis. Semiconductor metal oxide photoelectrodes are extensively used in the field of water splitting for hydrogen production, but there are certain limitations due to the low light absorption. Black silicon has emerged as a promising material for photocatalytic water splitting and extracting hydrogen.^{17,85,278,300} When used as a photoanode, the performance of black silicon can be significantly improved by integrating it with other materials, such as nanoparticles, nanocrystals, coating layers, and quantum dots. These integrated components effectively alleviate the intrinsically slow charge-transfer kinetics of black silicon and enhance its catalytic activity.

The inverted silicon pyramid array (SiIP) photocathode can generate a photogenerated voltage of 440 mV and exhibits a high photocurrent density.²⁷⁸ Moreover, its quasi-hydrophilic feature allows for a higher photocurrent density under large reverse bias. The intrinsically slow charge-transfer kinetics of the SiIP photocathode can be further enhanced with the incorporation of cobalt disulfide (CoS₂) nanocrystals. As shown in Fig. 26(f), the initial potential of the SiIP/CoS₂ photocathode was 0.22 V, and the saturation photocurrent density reached 10.4 mA cm⁻². The water-splitting performance can be further improved by using higher-quality silicon wafers and other high-performance catalysts. Yu et al. used a TiO₂ coating layer to passivate the black silicon surface,¹⁷ and a saturated photocurrent density of 32.3 mA cm⁻² was generated by a black silicon/TiO₂/cobaltous hydroxide (Co(OH)₂) heterostructured photoanode. This combination extended the lifetime of black silicon to four hours, markedly improving the system stability during water-splitting reactions. Gu et al. compared the application of columnar and sinuous nanoporous black silicon in the photoelectrochemical hydrogen production process.³⁰¹ Due to the damaged diffusion channels of oxygen on the surface of the sinuous nanoporous black silicon, silicon oxide formation is suppressed in an oxygen-deficient environment, leading to higher stability than the columnar black silicon under different conditions, suggesting the feasibility of controlling the subsurface morphology to extend the stability of the photoelectrode.

Molybdenum disulfide (MoS₂) is a widely studied two-dimensional material with strong catalytic activity at its edges. Molybdenum sulfide quantum dots (MoS_x) have highly exposed edges due to their nanoscale size, which can be combined with black silicon to achieve higher catalytic activity. Wang et al. modified the black silicon surface prepared by MACE with amorphous MoS_x.³⁰² This modification not only introduced more active sites to enhance electrocatalytic activity but also served as a surface passivation layer, mitigating the etching effect of the acidic electrolyte during operation. The resulting device exhibited excellent catalytic performance, with an onset potential of 0.255 V_RHE, a high J_sc of 12.2 mA cm⁻², and a hydrogen evolution rate of 226.5 µmol h⁻¹ cm⁻². It also demonstrated strong operational stability over 8 hours of continuous operation.

Beyond hydrogen evolution, black silicon shows a distinctive advantage for the chemical synthesis of ammonia (NH₃), a major focus in catalysis research.^303–305 Wang et al. reported a photoelectrochemical N₂ reduction reaction route combining black silicon and AgNPs to synthesize NH₃,³⁰⁶ the NH₃ evolution reaction activity was enhanced, achieving an NH₃ yield of 2.87 µmol h⁻¹ cm⁻². Muataz et al. described a solar-driven nanostructured photoelectrochemical cell for the conversion of atmospheric N₂ to ammonia,⁴⁹ exhibiting a yield of 13.3 mg m⁻² h⁻¹ under 2 suns illumination. The ammonia yield of unetched planar silicon was only 11% of the gold nanoparticle/black silicon/chromium structure. The markedly enhanced performance results from the large surface area, low reflectivity, and improved hole transport provided by the black silicon nanostructures.

4.4. THz/Field emission characteristics

In contrast to planar silicon with a relatively large penetration depth, black silicon significantly reduces the light penetration depth to the submicron scale due to its microstructured surface. The enhanced multiple light scattering and the confinement of photogenerated carriers within these nanostructures lead to significant modulation of the local potential. As a result, black silicon exhibits surface anisotropy, which can induce polarization effects that produce terahertz (THz) emission.⁵⁰ The enhanced THz emission is illustrated in Fig. 27(a),³⁰⁷ where the THz polarization direction is perpendicular to the interface. The polarization amplitude is directly proportional to the local intensity of the laser beam. However, oblique incidence leads to a linear phase gradient along the X-axis, resulting in a local pulse delay. The frequency range of the THz wave is 0.1–10 THz (1 THz = 10¹² Hz), corresponding to wavelengths from 0.03 mm to 3 mm, falling between the millimeter-wave and infrared regions of the electromagnetic waves. Fig. 27(b) shows the experimental setup of THz emission. THz emission is observed only from damaged or rough surfaces, whereas polished silicon does not exhibit measurable THz emission. Fig. 27(c) further demonstrates that black silicon produces significantly higher THz emission than either damaged or unpolished silicon surfaces.⁵⁰

Fig. 27 Emission characteristics of black silicon. (a) The enhancement mechanism of THz emission.³⁰⁷ Reproduced with permission from ref. 307. Copyright 2017, Optica Publishing Group. (b) Schematic of THz emission setup. (c) The THz electric field for black silicon, damaged surface, unpolished silicon, and polished silicon.⁵⁰ Reproduced with permission from ref. 50. Copyright 2008, AIP Publishing. (d) Time-domain waveform of silicon wafer and SiNWs with lengths from 0.3 µm to 9 µm. (e) Intensity of THz emission with different SiNWs length and pump-beam power.¹¹⁴ Reproduced with permission from ref. 114. Copyright 2010, Optica Publishing Group. (f) Field emission I–V curves of three silicon samples: type A (black silicon with 20 µm pillars), type B (black silicon with 8 µm pillars), and type C (planar black silicon).³⁰⁸ Reproduced with permission from ref. 308. Copyright 2016, AIP Publishing.

Blumröder et al. studied the THz emission of black silicon fabricated by RIE,³⁰⁷ the pulse width of the THz wave is on the order of picoseconds, so the carrier capture occurring on this timescale can significantly enhance the THz emission. The micro-morphology of the black silicon surface can effectively increase the THz emission, and its amplitude depends on the structure geometry and the depth of the silicon nanostructure, with deeper microstructures leading to stronger emission and slightly wider pulses. This enhancement is attributed to the increased number of carriers produced directly within the silicon needle, where the resulting THz radiation is effectively outcoupled. A similar phenomenon was observed by Jung et al., who used MACE to fabricate SiNW arrays and tested the THz emission performance,¹¹⁴ as shown in Fig. 27(d) and (e), longer SiNW arrays and higher pump-beam power resulted in stronger THz emission.

In addition to THz emission, black silicon can also be used to prepare field-emission devices.^308–310 Langer et al. prepared a black silicon field emission column array on a p-type silicon substrate by RIE.³⁰⁸ As shown in Fig. 27(f), samples with a column height of 20 µm (type A) exhibited a low turn-on field of 6.4 V µm⁻¹, with a field enhancement factor of 800 and an emission current reaching 8 µA at an applied field of 20 V µm⁻¹. Moreover, taller silicon columns improved the emission stability, indicating that such structures can serve as promising alternatives to traditional silicon-tip field-emission arrays.

4.5. Surface-enhanced Raman scattering effect

SERS is a special Raman scattering phenomenon that occurs on some special substrates consisting of noble-metal nanostructures (Au, Ag, Cu, etc.).^191,311 The Raman scattering intensity of molecules can be significantly enhanced due to the localized surface plasmon resonance (LSPR) effect, which greatly improves the specificity and sensitivity of Raman detection.³¹² Nanoparticles stabilized in colloidal solutions are commonly used in SERS analysis, which provides strong signal enhancement in Raman signals. However, they are vulnerable to environmental changes and contamination, and the disordered arrangement of nanoparticles results in poor reproducibility.^313,314 Other approaches, such as direct production or in situ growth of noble-metal nanostructures, are not cost-effective.^315–317 Black silicon has been used as a SERS-active substrate in recent years due to its low cost and reusable features,^128,318 and its SERS performance can be further boosted through precise engineering of both the black-silicon nanostructure and the deposited metal layer.³⁵ The enhancement factor (EF) is calculated as follows:³¹⁹where I_SERS and I_RS are intensities of the SERS and normal Raman bands, and N_SERS and N_RS are the numbers of molecules to be measured on black silicon and bare silicon wafers, respectively.

It is generally believed that SERS arises from the synergistic effect of electromagnetic enhancement, resulting from the synergistic contributions of electromagnetic enhancement, arising from the amplification of the local electromagnetic field near the substrate surface, and chemical enhancement, which results from interactions between adsorbed molecules and the substrate. The Raman signal intensity is proportional to the fourth power of the local field EF.³¹²

The thickness of the surface metal nanolayer has a great influence on the SERS signal. As shown in Fig. 28(a),²⁶⁷ no Raman peaks are observed in the absence of metal nanoparticles. Black silicon exhibits higher SERS intensity than textured silicon, and the maximum spectral enhancement is achieved at an optimal Ag nanoparticle size. Here, the nanoparticle size is determined by the deposited Ag mass thickness during thermal evaporation. For instance, mass thicknesses of 3 nm, 5 nm, and 10 nm result in Ag nanoparticles with average diameters of ∼10 nm, ∼18 nm, and ∼25 nm, respectively. With increasing size, the localized surface plasmon resonance peak undergoes a red shift and broadening, since higher-order multipole modes are excited in addition to dipolar scattering, which in turn affects light-trapping efficiency. In another study, the SERS substrates were fabricated by first preparing silicon inverted pyramids via Ag MACE, subsequently decorated with AgNPs using electroless deposition and radiofrequency sputtering.³²⁰ Compared with electroless deposition, radiofrequency sputtering produces smaller and denser nanoparticles on the black silicon surface, thereby yielding stronger SERS enhancement.

Fig. 28 SERS characteristics of black silicon. (a) SERS spectra of Rhodamine 6G recorded on black silicon with different Ag thickness.²⁶⁷ Reproduced with permission from ref. 267. Copyright 2013, Royal Society of Chemistry. (b) Schematic illustration of the SERS process, including structure fabrication, measurement procedures, and applications. (c) Raman spectra of a 4-MBA monolayer on different substrates.⁵¹ Reproduced with permission from ref. 51. Copyright 2020, American Chemical Society. (d) Storage stability of the 4-MBA-functionalized black silicon/Au substrate over a period of 20 months. (e) Raman spectra of the substrate after mild oxygen plasma treatment as a surface regeneration process over ten reuse cycles.³²¹ Reproduced with permission from ref. 321. Copyright 2023, American Chemical Society.

4-Mercaptobenzoic acid (4-MBA) is commonly used as a test standard for SERS due to its well-defined self-assembled single-layer structure, with well-known surface density and Raman spectrum. The process of the SERS test is shown in Fig. 28(b). A SERS spectrum of 4-MBA on the black silicon/Au substrate was obtained, as shown in Fig. 28(c).⁵¹ The 4-MBA band is not visible on the SiO₂/Au substrate but is very clearly observed on the black silicon/Au substrate, with an EF of 2 × 10⁸.

However, the substrate may become contaminated during the testing process, thereby limiting its potential for reuse. According to a study by Golubewa's group,³²¹ as shown in Fig. 28(d), the black silicon/Au substrate exhibited long-term stability when stored in a well-sealed environment. The intensity of the characteristic Raman band decreased by less than 5% after 120 days and by 40% after 20 months. The contaminants on the substrate can be completely removed by treating it with a mild oxygen plasma at room temperature for 35 minutes. The substrate is then rinsed with DI water and dried. After this cleaning process, the substrate can be reused, and the corresponding reuse spectrum is shown in Fig. 28(e). After ten reuse cycles, the Raman signal of 4-MBA molecules covalently bonded to the Au coating is reduced by a factor of four compared with that of the original substrate, indicating good long-term reusability.

Trau et al. designed a digitized nanopillar SERS platform capable of real-time counting of individual cytokines and dynamic monitoring of immune toxicity in patients undergoing immunotherapy.³²² The silicon nanopillars were fabricated by RIE, and the SERS nanotags were obtained by modifying the surface of gold–silver alloy nanoboxes with different Raman molecules and corresponding target antibodies. By capturing individual cytokines through discrete pillar arrays, highly specific and sensitive cytokine detection down to the attomolar level can be achieved.

4.6. Bactericidal properties

The nanostructures on black silicon surfaces are well suited for use as antibacterial surfaces and have been widely studied.^34,323–325 The application of surface nanostructures for bactericidal and antibiofouling purposes is inspired by the nanocolumn structures on the surface of insect wings,^326–328 which can kill bacteria through mechanical action. The physical bactericidal effects of black silicon were reported by Ivanova et al. in 2013.^323,329 They compared the bactericidal effect of dragonfly wings with the surface nanostructures of black silicon and concluded that the bactericidal effect was caused by mechanical and structural deformation stress of the surface nanostructures. When bacteria attach to the nanostructured surface, the adhesion force causes the cell membrane to stretch between neighboring nanopillars. Once the stress exceeds a critical threshold, the cell membrane is damaged, resulting in cell death. Remarkably, the antibacterial activity of black silicon surface is comparable to that of natural biologically nanostructured surfaces, such as cicada and dragonfly wings, and it can be further optimized by adjusting the nanostructural parameters.^28,330 Even small modifications to the nanostructure can significantly alter its performance as a bactericidal surface. Hence, many researchers have explored the relationship between nanostructural parameters and bactericidal efficacy.^{325,330–332}

Ivanova and colleagues fabricated nanopillar arrays on silicon wafers by deep-UV immersion lithography followed by plasma etching.³²⁵ The nanoarrays with heights of 220 nm, 360 nm, and 420 nm were prepared by extending the RIE time. They proved that the mechanical cleavage of bacterial cells is related to the elastic restoring force of the nanopillars and the extent of nanopillar clustering. As shown in Fig. 29(a) and (b), bacterial cells appeared collapsed or elongated when trapped between the nanocolumns due to the adhesion force. As the aspect ratio increases, the mechanical energy stored in the columns increases, enhancing the bactericidal effect. However, beyond a certain limit, the nanopillars become more flexible during bacterial interactions and tend to aggregate, forming stable nanopillar bunches with a larger surface area, thereby changing the surface nanostructure and reducing bactericidal efficiency. As the nanopillar height increased, as depicted in Fig. 29(c), the mortality rate of P. aeruginosa cells on black silicon with heights of 220, 360, and 420 nm were 53%, 95%, and 89%, respectively, while those of S. aureus cells were 58%, 83%, and 77%, respectively. Su et al. prepared black silicon microneedles with excellent antimicrobial properties and enhanced drug loading capacity.²¹¹ The drug loading efficiency was more than twice that of ordinary silicon microneedles in the same area.

Fig. 29 Antibacterial properties of black silicon surfaces. (a) Viability of P. aeruginosa and (b) S. aureus bacterial cells on black silicon samples with different nanostructure heights. (c) Quantification of cell attachment on black silicon.³²⁵ Reproduced with permission from ref. 325. Copyright 2020, National Academy of Sciences. (d) Bactericidal mechanism on black silicon, including mechanical injury and apoptosis-like death.³⁴ Reproduced with permission from ref. 34. Copyright 2022, American Chemical Society.

In another research, the authors fabricated a black silicon surface by RIE.³³⁰ They found that the bactericidal activity of the nanostructured surface is related to specific parameters such as geometry and surface wettability. When the height of the microcolumn array exceeds 1000 nm, the density is lower than 8 tips per µm², and the structures are poorly dispersed, the antibacterial performance of the black silicon surface decreases significantly. The nano-morphology parameters of a single surface cannot be directly related to the change of bactericidal activity, but the highest bactericidal efficiency can be achieved by combining different parameters. The exact relationship among these factors is worth investigating deeply.

The synergistic mechanism combining physical and chemical actions has been presented recently.^333,334 Zhao et al. posited that mechanical injury alone does not immediately kill bacteria, as the inner plasma membrane often remains intact, allowing some bacteria to survive.³⁴ They explained that an apoptosis-like death is necessary after the initial mechanical injury. Furthermore, they noted that when mechanical stress is removed, damaged cells undergo post-stress apoptosis-like death, which is facilitated by reactive oxygen species (ROS) that accumulate in a non-stressed environment. This indicates that the mechano-bactericidal actions have prolonged physiological impacts on bacteria. According to their findings, illustrated in Fig. 29(d), the sterilization process begins with the mechanical destruction of the bacterial cell wall and outer membrane, followed by a delayed apoptosis-like death, eventually leading to bacterial death through a self-destructive process.

4.7. Biomedical properties

In recent years, black silicon has been increasingly explored in biomedical contexts, not as a chemically functionalized biomaterial, but as a structurally engineered interface capable of interacting with biological systems across multiple length scales. Owing to its intrinsic compatibility with micro–nano fabrication and device integration, black silicon provides a unifying material platform for regulating cell behavior, enabling biosensing and bioimaging, and achieving localized drug delivery. In these applications, the biological performance of black silicon is determined less by its elemental composition and more by how its nanostructured geometry mediates physical interactions with biomolecules, cells, and tissues.

4.7.1. Nano-bio interfaces as the foundation of biomedical functionality. At the core of black silicon-enabled biomedical functionality lies the formation of well-defined nano-bio interfaces, where biological entities interact directly with nanoscale silicon features.^42,335,336 Unlike planar silicon surfaces, nanostructured and porous silicon creates spatially confined and curvature-rich environments that fundamentally alter molecular adsorption, transport, and interfacial dynamics. These effects are intrinsic to the geometry of the nanostructure and do not require additional biochemical modification to become effective.

At the molecular level, confinement within nanoscale pores and between adjacent nanostructures allows proteins, nucleic acids, enzymes, and small-molecule therapeutics to be immobilized with reduced conformational disruption. The local physical environment provided by porous silicon stabilizes biomolecular structure and enables predictable release behavior governed by diffusion pathways, pore connectivity, and interfacial interactions. As a result, porous silicon functions as an active host matrix that regulates molecular exchange rather than serving as a passive carrier.

Importantly, these nano-bio interfacial effects are highly sensitive to geometric parameters such as pore size distribution, aspect ratio, surface roughness, and structural ordering. Subtle variations in morphology can significantly alter adsorption capacity, release kinetics, and interfacial transport, thereby determining biological performance. This morphology-driven nature of nano-bio interactions establishes structural design as a primary control knob in biomedical applications of black silicon, forming the mechanistic basis for subsequent regulation of cellular responses, sensing, imaging, and drug delivery.

4.7.2. Morphology-regulated cellular responses at silicon biointerfaces. Cellular responses to black silicon nanostructures are predominantly governed by morphology-induced nano-bio interactions rather than chemical composition alone. Compared with planar substrates, vertically oriented nanostructures such as nanoneedles and nanowires create highly localized and three-dimensional cell-material interfaces, where geometric parameters including feature height, aspect ratio, tip sharpness, spacing, and porosity dictate how cells adhere, spread, and sense the underlying surface.³³⁷ When the characteristic spacing of nanostructures approaches or falls below the length scale of focal adhesion complexes, continuous adhesion formation is disrupted, leading to altered cytoskeletal organization and reduced cell spreading, even in the absence of surface chemical modification.

A key consequence of nanostructure geometry is the induction of membrane curvature, which plays a central role in regulating intracellular uptake. Vertically aligned silicon nanowires and nanoneedles impose localized curvature on the plasma membrane, lowering the energetic barrier for membrane wrapping and promoting endocytic pathways. The membrane-wrapping and endocytic internalization process induced by silicon nanostructures is schematically illustrated in Fig. 30(a), with representative confocal fluorescence images showing nanowire internalization in Fig. 30(b). This effect has been directly observed in studies on unlabeled independent silicon nanowires, where spontaneous cellular internalization, active intracellular transport, and eventual perinuclear accumulation were reported, enabling the probing of intracellular biophysical dynamics.³³⁸ The efficiency of uptake strongly depends on nanostructure diameter and tip radius, with sharper and smaller features facilitating membrane deformation and enhancing endocytosis.

Fig. 30 Black silicon nanostructures serve as versatile biointerfaces. (a) Schematic illustration of SiNW internalization. (b) Confocal fluorescence micrograph of SiNW internalization (blue scattering).³³⁸ Reproduced with permission from ref. 338. Copyright 2016, American Association for the Advancement of Science. (c) Topical delivery behavior of SiNWs.³³⁹ Reproduced with permission from ref. 339. Copyright 2024, American Chemical Society. (d) Working principle of silicon nanoneedles for ocular drug delivery. (e) Variation of drug dose with the surface porosity of the nanoneedle array. (f) Variation of drug dose with the length of the nanoneedle array.³⁴⁰ Reproduced with permission from ref. 340. Copyright 2022, American Association for the Advancement of Science. (g) SEM image of 500 µm tall nanoneedles.³⁴¹ Reproduced with permission from ref. 341. Copyright 2023, Elsevier.

Beyond uptake processes, nanostructured silicon interfaces exert broad influence over cell fate and collective behavior.³³⁷ The protruding features of nanoneedle arrays provide abundant anchoring sites for filopodia, enhancing adhesion strength while restricting excessive cell spreading and reducing detachment. During stem cell differentiation, curvature-sensitive membrane proteins can be preferentially localized at these protrusions, promoting protrusive growth and lineage-specific differentiation. Moreover, nanostructure density and spatial arrangement further modulate adhesion, migration, and proliferation: high-density nanoneedle arrays generally reduce effective adhesion area and migration velocity while maintaining or even enhancing proliferation, as revealed by high-content live-cell imaging.³³⁹ Conversely, patterned nanostructured silicon surfaces can generate cytophilic and cytophobic regions that guide cell alignment and directional migration,³⁴² demonstrating that geometry alone can act as a physical cue for spatial regulation of cell behavior.

Collectively, these findings demonstrate that cellular behavior on black silicon is tightly coupled to nanoscale morphology. By tailoring nanostructure geometry, porosity, and spatial distribution, black silicon platforms enable precise control over cell-material interfaces, intracellular delivery efficiency, and long-term cellular functions, providing a versatile physical strategy for regulating cell behavior in biomedical applications.

4.7.3. Multifunctional biomedical platforms enabled by nanostructured silicon. Beyond regulating molecular and cellular interactions, black silicon and related nanostructured silicon systems provide multifunctional platforms that integrate sensing, imaging, and delivery within a single material architecture. The high surface area, controllable porosity, and vertically oriented nanostructures enable intimate coupling between physical transducers and biological environments, allowing black silicon to act simultaneously as a structural interface and a functional signal mediator. Importantly, these capabilities arise from morphology-engineered interfaces rather than extensive biochemical functionalization.

In bio-sensing applications, vertically aligned nanostructures establish tight electrical and mechanical coupling with cells, enabling direct access to intracellular and membrane-associated signals. For example, biomimetic nano-transistor arrays fabricated on nanostructured silicon substrates have demonstrated the ability to simultaneously record electrical activity and mechanical forces at the single-cell level, achieving label-free, sub-millisecond resolution and scalable readout.³⁴³ Such platforms demonstrate how nanostructured silicon bridges electrical, mechanical, and biological domains, providing advantages for tracking and distinguishing cell states in drug studies.

Nanostructured silicon platforms have also enabled advanced bio-imaging and molecular analysis by enhancing signal transduction and spatial resolution. Gold nanoparticle-assisted black silicon substrates fabricated via RIE have been successfully applied in surface-assisted laser desorption/ionization mass spectrometry imaging, enabling metabolite mapping from fingerprints and biological tissues with high sensitivity and dual-polarity detection.²⁷ These results demonstrate that black silicon nanostructures can function as efficient interfaces for label-free bio-imaging and chemical mapping at biologically relevant length scales.

In parallel, black silicon-based nanoneedle and microneedle platforms enable localized and minimally invasive drug delivery across multiple biological barriers.^344–347 Biodegradable porous silicon nanoneedles embedded in adhesive patches conform closely to the skin of living animals, causing minimal discomfort. They are effective for tissue-level nanoinjection,³³⁹ the mechanism is shown in Fig. 30(c), providing uniform, sustained, and localized delivery while maintaining structural integrity. The nanoneedles are fabricated using RIE to define the conical shape, followed by MACE, in which varying the etchant composition allows precise control over the surface porosity of the nanostructures.

Biodegradable silicon nanoneedles integrated with tear-soluble contact lenses have been shown to achieve two-stage ocular drug delivery.³⁴⁰ As shown in Fig. 30(d), when the dissolvable contact lens conforms to the eye, silicon nanoneedles are mechanically inserted into the corneal epithelial layer. The contact lens subsequently dissolves in tear fluid within approximately one minute, triggering immediate drug release, while the residual silicon nanoneedle network degrades slowly, enabling prolonged drug delivery. Standard photolithography and anisotropic DRIE define vertically ordered silicon micropillar arrays, which are then shrunk in 10% KOH at 25 °C for 10 min and further modified by MACE to control surface porosity. For instance, as shown in Fig. 30(e) and (f), immunoglobulins 647 dosage increased from 10.17 ± 0.70 to 17.44 ± 0.74 µg as silicon nanoneedle surface porosity increased from 0 to 60% at a fixed length of 60 µm; similarly, dosage increased from 1.41 ± 0.16 to 13.88 ± 0.14 µg as length increased from 10 to 60 µm at 30% porosity. Hollow silicon microneedle arrays (Fig. 30(g)) fabricated by combined wet and dry etching allow both transdermal drug infusion and interstitial fluid sampling,³⁴¹ supporting integrated therapeutic and diagnostic functions. These arrays deliver fluids up to several mL at 30 µL min⁻¹ and withdraw 1 µL of interstitial fluid via capillary action.

Collectively, these studies demonstrate that black silicon serves as a versatile and scalable bio-platform in which morphology-engineered nanostructures unify sensing, imaging, and delivery functions. By leveraging the same structural parameters that govern optical and electronic performance, black silicon enables multifunctional biomedical systems capable of interacting with biological environments across molecular, cellular, and tissue length scales.

4.8. Others

Hydrophobicity. The superhydrophobic surface exhibits excellent anti-contamination, self-cleaning, and anti-frosting performance.^348,349 Yue et al. investigated the frosting behavior of black silicon and found that the superhydrophobic surface delayed the freezing time and prevented the accumulation of ice. However, the anti-icing ability decreased as the surface temperature dropped. The combination of double lithography technology and DRIE enabled the creation of silicon nanostructures with excellent regularity and uniform coverage,³⁵⁰ as well as three-dimensional nano/nano dual-scale structures with large surface areas and hydrophobic properties. Han et al. studied the effect of PIIIE-prepared silicon microstructures on wettability.³⁵¹ As the SF₆/O₂ ratio increased, surface roughness increased sharply and then slightly declined. However, the area ratio of the samples continued to increase, leading to a gradual rise in contact angle, with the rate of increase slowing down.

Beyond self-cleaning and anti-frosting functionalities, the superhydrophobic nature of black silicon also profoundly affects interfacial fluid dynamics. Specifically, nano/micro hierarchical structures created by DRIE or combined DRIE-MACE processes can modulate the slip behavior at the air–water interface by introducing trapped air pockets that effectively reduce liquid–solid contact. Choi and Kim reported that nanoengineered superhydrophobic surfaces exhibited substantial slip lengths,³⁵² approximately 20 µm for water and 50 µm for 30 wt% glycerin, confirming the critical role of surface morphology in drag reduction and flow enhancement at micro- and nanoscales. Similarly, Lee and Kim demonstrated that microstructured silicon arrays fabricated via DRIE achieved pronounced slip flow behavior,³⁵³ which can be attributed to the formation of stable air layers and minimized viscous dissipation at the solid–liquid boundary. These findings highlight the potential of black silicon and DRIE-MACE-derived surfaces not only for superhydrophobicity but also for manipulating interfacial hydrodynamics, offering new strategies for drag reduction and efficient microfluidic transport.

Thermal annealing has been demonstrated to increase the hydrophilicity of planar silicon. As the annealing temperature increases, the static contact angle decreases. However, for black silicon, thermal annealing increases its hydrophobicity, which is attributed to the rough surface with numerous microstructures.³⁵⁴ Annealing optimizes the morphology and chemical composition of black silicon, thereby reducing the contact area and interaction between water droplets and the surface. The hydrophobic property of black silicon can also be used in microfluidic device design,³⁵⁵ which involves fabrication by DRIE through alternating SF₆ etching and C₄F₈ passivation cycles. In this process, black silicon nanostructures are directly formed on the inner surfaces of the silicon microchannel before sealing, as the DRIE step etches the exposed silicon regions defined by the SU-8 or SiO₂ hard masks. When liquids in a microfluidic channel converge and merge, bubble encapsulation usually occurs at the point of convergence. Using black silicon on the inner surface of the pipe makes it difficult for liquid in the microchannel to penetrate the nanostructured layer, thus creating a ventilation channel between the liquid sidewalls and preventing the formation of bubbles.

Gas sensors. The special surface structure of black silicon provides abundant active sites for gas adsorption and reaction, enabling more gas molecules to be captured and involved in the sensing reaction.^356–358 Take a p-type black silicon sensor as an example, when it is exposed to ambient air, oxygen molecules are adsorbed onto the surface and ionized into chemisorbed species (O₂⁻, O⁻, and O²⁻) by extracting electrons from silicon, thereby creating a hole accumulation layer (HAL).¹⁸⁷ Upon exposure to NO₂, the HAL expands further because NO₂ molecules—with a much higher electron affinity (2.27 eV) compared to oxygen (0.43 eV)—withdraw additional electrons and also interact with pre-adsorbed oxygen ions. This enhanced electron extraction significantly increases the hole concentration, thereby decreasing the resistance and improving the conductivity of p-type black silicon sensors. The nanostructured morphology of black silicon, with its extremely high density of surface states, further amplifies this charge-transfer process, enabling high sensitivity and fast response to NO₂.

Wang et al. prepared black silicon with various doping levels via femtosecond laser irradiation in SF₆ and NF₃ mixed gases to study its sensing ability in NO₂ gases at room temperature.¹⁸³ Different responses varied by changing the doping ratio of S and N, as different ratios resulted in different superdoping, and further led to different morphology and defects of black silicon. Higher sulfur content produced sharper and more separated spikes, whereas higher nitrogen content led to blunter features. This method offered high selectivity and yielded a response of approximately 3955% to 20 ppm NO₂ when the SF₆/NF₃ ratio was 56/14. Zhuang et al. fabricated N-hyperdoped NO₂ gas sensors and studied the impact of annealing temperature on their performance.¹⁸⁴ Low-temperature annealing had minimal influence on the electrical properties of black silicon sensors. With increasing temperature, the resistance first slightly decreased, then increased sharply, and eventually dropped significantly. These changes were attributed to variations in the number of nanoparticles within the microstructure, which altered the surface states of black silicon. At an optimal annealing temperature of 873 K, the sensor achieved a detection limit of 7 ppb and an ultra-wide dynamic range spanning approximately five orders of magnitude.

Photo-thermal-electric conversion. The photo-thermal-electric (PTE) conversion is another area where black silicon shows great potential.³⁵⁹ The high density of surface defects in black silicon enhances carrier recombination, reducing its efficiency in converting light energy into electrical energy. However, these same non-radiative recombinations allow for the direct conversion of light energy into thermal energy. Ryosuke et al. explored this effect by coating black silicon with gold colloidal nanoparticles,⁴⁷ finding a notable enhancement in the Seebeck effect. Specifically, 50-nm-diameter gold nanoparticles with an optimal density coating on black silicon resulted in a 50% increase in thermoelectric conversion efficiency. The black silicon used in this study was fabricated via RIE, and the Au nanoparticles further enhanced light absorption.

Following this, Xu et al. prepared black silicon using MACE and applied a polystyrene coating onto its surface to minimize heat loss,³⁶⁰ enabling stable PTE operation. This study further demonstrates that black silicon, with its near-unity broadband absorptance and efficient light-to-heat conversion characteristics, is well-suited for PTE applications.

Additionally, Cheng et al. assessed the photothermal conversion performance of black silicon prepared by RIE.³⁶¹ They observed that under increased solar irradiation, the surface temperature of both black silicon and bare silicon gradually rose, with the black silicon heating up more quickly and reaching higher temperatures at the same intervals. Zhang's study echoed these findings, further affirming black silicon as a promising material for PTE conversion applications.¹⁵⁹ After 1 hour of solar-driven evaporation, the evaporated mass changes for pure water in a Teflon beaker without silicon (control group), water containing disordered nanoporous silicon, and water containing order–disorder hybrid silicon reached 1.35, 2.43, and 2.50 g, respectively. The nanostructured silicon achieved nearly double the water evaporation rate.

4.9. Application-specific morphology of black silicon

Across the applications discussed in this review, the functionality of black silicon is closely linked to its surface morphology, which is ultimately determined by chemistry-driven fabrication processes. High-aspect-ratio nanowires, nanocones, and hierarchical textures are predominantly employed in photodetectors and solar cells to enhance broadband light absorption and suppress surface reflection. Porous and defect-rich morphologies, often generated through chemical etching or MACE, are particularly effective for surface-sensitive applications such as gas sensing, photocatalysis, and SERS, where abundant active sites and enhanced interfacial interactions are required.

For THz devices and field emission, sharp-tip or periodically structured morphologies enable strong local electric-field enhancement, facilitating efficient carrier emission and electromagnetic coupling. In contrast, applications involving wettability control and antibacterial activity primarily rely on multiscale roughness and high-aspect-ratio features, which regulate solid–liquid or cell–surface interactions through physical mechanisms. In biomedical-related applications, nanoscale roughness, curvature, and surface states play a decisive role in modulating protein adsorption, cellular responses, and light-matter-biological interactions. Collectively, these observations highlight morphology engineering as a unifying strategy that bridges chemical fabrication routes and the diverse advanced material applications of black silicon. These application-specific morphology–function relationships are systematically summarized in Table 7, which correlates representative black silicon morphologies with their dominant interaction mechanisms and targeted application domains.

Table 7 Application-oriented morphology of black silicon

Application	Dominant morphology characteristics	Functional role
Photodetectors	High-aspect-ratio nanowires/nanocones	Broadband absorption and efficient carrier generation
Solar cells	Hierarchical micro–nano textures	Reflection suppression and optical coupling
Photocatalysis	Porous, defect-rich structures	Increased active surface and reaction sites
Terahertz devices/Field emission	Sharp-tip or periodic nanostructures	Local electric-field enhancement
SERS	Nanoscale roughness with hierarchical features	Electromagnetic hot-spot formation
Antibacterial surfaces	High-aspect-ratio and mechanically sharp features	Physical disruption of bacterial membranes
Gas sensors	Porous or spike-like morphologies	Enhanced adsorption and surface charge transfer
PTE conversion	Broadband-absorbing nanostructures	Efficient light-to-thermal or light-to-electric conversion
Biomedical applications	Controlled nanoscale roughness and curvature	Regulation of nano-bio interfacial interactions

---

5. Emerging directions, challenges, and outlook

The journey from traditional materials to those with micro- and nanostructured surfaces signifies a paradigm shift in materials science and device engineering. As micro- and nanotechnologies continue to advance, the development of materials with precisely engineered micro- and nanostructures—such as nanopillars, nanowires, and porous surfaces—has led to significant improvements in device performance across various domains. These structures, by enhancing the surface area, light absorption, and overall sensitivity, facilitate the realization of highly efficient, compact, and robust devices. Such advancements have profound implications for fields such as optoelectronics, sensors, and energy harvesting.

There is no doubt that black silicon holds great promise for the future. With continued scientific research and advancements in the semiconductor industry, its applications are expected to expand even further. The desired performance can be achieved through the precise control of material and process parameters. Despite its numerous advantageous properties, there are still many problems to be solved in its application process. The existing problems and future development trends of black silicon are shown in Fig. 31.

Fig. 31 Outlook and future trend for black silicon.

Mass production challenges

Techniques such as RIE demand low temperatures to reduce surface damage, while the mechanisms of MACE remain partially understood. Furthermore, managing the doping concentration and precision in femtosecond laser irradiation is challenging, and PIIIE shows a rapid decrease in doping concentration with depth. These factors emphasize the need for more research into scalable and controllable production processes.

Surface morphology optimization

The nanostructured surface of black silicon, though beneficial for light absorption, creates challenges in carrier transport, recombination, and electrode contact due to the irregularity and gaps between nanostructures. This limits device performance. Therefore, more uniform or novel structures should be investigated for better performance. On the other hand, with the development of new technologies and in-depth research and improvement of existing technologies, more fascinating improvements will appear.

Expanding light absorption range

Current methods struggle to extend black silicon's light absorption beyond 1100 nm. Developing advanced methods to broaden this range, especially to achieve high responses at optical communication wavelengths such as 1060, 1330, and 1550 nm, remains a key research goal for enhancing the capabilities of black silicon-based photodetectors.

Exploring unique properties

Black silicon's properties, such as near-complete light absorption, enhanced catalytic reactivity, and beneficial surface defects for luminescence (though detrimental for photovoltaic performance), provide vast potential for further exploration. Innovations like exploiting the quantum confinement effect in nanostructured black silicon could lead to new electronic and optical behaviors with tunable band gaps. Integrating metal nanoparticles to induce plasmonic resonance effects could also enhance light–matter interactions, broadening the applications in photovoltaics and beyond. Recent work by Wang et al. has further demonstrated the power of advanced characterization techniques in probing such unique properties, providing valuable insights into spin-dependent transport phenomena in semiconductor materials.³⁶²

Surface chemical bond engineering

The various surface chemical bonds formed during black silicon fabrication—such as Si–H, Si–O, and Si–F—play a crucial role in determining surface energy, passivation quality, and carrier recombination rates. However, these bonds can also introduce trap states or affect environmental stability, posing challenges for device reliability. Future research should focus on precise control and modification of these chemical terminations through advanced surface treatments or novel passivation techniques. With ongoing technological progress and a deeper understanding of surface chemistry, significant improvements in electronic and optoelectronic performance are anticipated.

Surface defects and passivation techniques

Surface defects significantly impair black silicon's device performance by causing recombination and reducing carrier lifetimes. Developing advanced passivation techniques using materials like SiN_x, Al₂O₃, or organic passivants is crucial. Additionally, new technologies that effectively utilize surface defects could revolutionize black silicon's applications,²⁸⁷ such as alternating current photovoltaic (AC PV) effect.

Contact optimization

Despite progress in passivation technologies, achieving efficient electrical contact between black silicon's nanostructured surfaces and electrodes remains difficult. This limitation restricts performance enhancement, especially in devices with nanostructured surfaces. Enhancing contact design and integration is vital for unlocking the full potential of black silicon devices.

Uniform morphology for scalability

Achieving uniform nanostructures remains a key challenge for large-scale production and performance consistency, and ancillary processes such as photolithography and masks are commonly used to form ordered micro–nanostructures. The inhomogeneity of current fabrication methods leads to variations in optical and electrical properties, limiting the reliability and scalability of black silicon applications. The development of fabrication processes that ensure morphological homogeneity is critical for the widespread use of black silicon in commercial devices.

Specialized applications

Black silicon has shown potential in a wide range of applications, from infrared detectors to solar cells to antimicrobial surfaces. However, further research is needed to explore its potential in specialized applications, such as biomedicine and advanced sensing technologies. By leveraging its unique properties, black silicon could become a key material for next-generation devices.

Biological research

The biological applications of black silicon, such as its bactericidal properties and potential in drug delivery, remain underexplored. Future research should focus on understanding the relationship between its nanostructures and biological effects, promoting its integration into biomedical fields.

The journey of black silicon from a niche material to a mainstream technology hinges on overcoming these challenges. By addressing production scalability, enhancing surface properties, and leveraging its unique optoelectronic and biological characteristics, black silicon can play a pivotal role in next-generation device technologies.

Author contributions

Junling Lv: writing – original draft, writing – reviewing & editing, investigation, visualization; Lihong Jiang: visualization, investigation; Xinlin Liu: methodology; Gaojie Li: visualization; Mingrui Qian: visualization, investigation; Mingxin Tang: methodology; Xinao Cheng: investigation; Lan Lu: investigation; Xiarong Ren: investigation; Xueling Zhang: investigation; Haiyang Zou: conceptualization, original draft, project administration, resources, funding acquisition, writing – reviewing and editing, supervision; Zhong Lin Wang: conceptualization, writing – reviewing and editing, supervision.

Conflicts of interest

There are no conflicts to declare.

Data availability

No primary research results, software or code have been included, and no new data were generated or analyzed as part of this review.

Acknowledgements

This work was supported by National Natural Science Foundation of China (Grant No. 52573253), the Fundamental Research Funds for the Central Universities (YJ202293), the TCL Science and Technology Innovation Fund, and the Distinguished Young Scholars of the National Natural Science Foundation of China (Overseas) (00301054A1091), Beijing Key Laboratory of High-Entropy Energy Materials and Devices, Beijing Institute of Nanoenergy and Nanosystems (No. GS2025ZD010).

References

1. Y. Li, C. Liu, H. Zou, L. Che, P. Sun, J. Yan, W. Liu, Z. Xu, W. Yang, L. Dong, L. Zhao, X. Wang, G. Wang and Z. L. Wang, Cell Rep. Phys. Sci., 2023, 4, 101191.
2. S. Gupta, A. Knoepfel, H. Zou and Y. Ding, Sens. Actuators, B, 2023, 392, 134030.
3. H. Zou, L. Guo, H. Xue, Y. Zhang, X. Shen, X. Liu, P. Wang, X. He, G. Dai, P. Jiang, H. Zheng, B. Zhang, C. Xu and Z. L. Wang, Nat. Commun., 2020, 11, 2093.
4. P. K. Bondyopadhyay, Proc. IEEE, 1998, 86, 78–81.
5. W. Yangyuan, H. Ruqi, L. Xiaoyan and K. Jinfeng, Beijing, China, 1998.
6. S. Zeng, C. Liu and P. Zhou, Nat. Rev. Electr. Eng., 2024, 1, 335–348.
7. Y. Zhang and G. Chen, Light: Sci. Appl., 2024, 13, 87.
8. X. Liu, P. R. Coxon, M. Peters, B. Hoex, J. M. Cole and D. J. Fray, Energy Environ. Sci., 2014, 7, 3223–3263.
9. J. Lv, T. Zhang, P. Zhang, Y. Zhao and S. Li, Nanoscale Res. Lett., 2018, 13, 1–10.
10. Z. Zhao, Z. Zhang, J. Jing, R. Gao, Z. Liao, W. Zhang, G. Liu, Y. Wang, K. Wang and C. Xue, APL Mater., 2023, 11, 021107.
11. A. J. Bett, J. Eisenlohr, O. Höhn, P. Repo, H. Savin, B. Bläsi and J. C. Goldschmidt, Opt. Express, 2016, 24, A434–A445.
12. S. Uddin, M. R. Hashim and M. Z. Pakhuruddin, Opt. Mater., 2022, 134, 113111.
13. J. Yang, F. Luo, T. S. Kao, X. Li, G. W. Ho, J. Teng, X. Luo and M. Hong, Light: Sci. Appl., 2014, 3, e185.
14. L. Meng, M. Awashra, S. M. Mirmohammadi, S. M. Mousavi, J. Vapaavuori, V. P. Jokinen and S. Franssila, Nanoscale, 2025, 17, 2871–2877.
15. Y. He, C. Jiang, H. Yin, J. Chen and W. Yuan, J. Colloid Interface Sci., 2011, 364, 219–229.
16. Z. Fekete, Á. C. Horváth, Z. Bérces and A. Pongrácz, Sens. Actuators, A, 2014, 216, 277–286.
17. Y. Yu, Z. Zhang, X. Yin, A. Kvit, Q. Liao, Z. Kang, X. Yan, Y. Zhang and X. Wang, Nat. Energy, 2017, 2, 17045.
18. E. Mitsai, A. Kuchmizhak, E. Pustovalov, A. Sergeev, A. Mironenko, S. Bratskaya, D. P. Linklater, A. Balčytis, E. Ivanova and S. Juodkazis, Nanoscale, 2018, 10, 9780–9787.
19. Y. Yang, Z.-Y. Ren, C. Li and J.-H. Zhao, Opt. Quantum Electron., 2023, 55, 259.
20. S. Huang, G. Deng, X. Jin, Y. Lu, G. Song, H. Huang, P. Zhao, C. Zhang, J. Yao, Q. Wu and J. Xu, Opt. Mater., 2020, 110, 110474.
21. A. Alsolami, H. Hussain, R. Noor, N. AlAdi, N. Almalki, A. Kurdi, T. Tabbakh, A. Zaman, S. Alfihed and J. Wang, Appl. Sci., 2024, 14, 9841.
22. Y. Xia, B. Liu, J. Liu, Z. Shen and C. Li, Sol. Energy, 2011, 85, 1574–1578.
23. H. Savin, P. Repo, G. von Gastrow, P. Ortega, E. Calle, M. Garín and R. Alcubilla, Nat. Nanotechnol., 2015, 10, 624–628.
24. G. Seniutinas, G. Gervinskas, R. Verma, B. D. Gupta, F. Lapierre, P. R. Stoddart, F. Clark, S. L. McArthur and S. Juodkazis, Opt. Express, 2015, 23, 6763–6772.
25. Y. Zhao, X.-L. Liu, S.-X. Ma, W.-J. Wang, X.-J. Ning, L. Zhao and J. Zhuang, Sens. Actuators, B, 2021, 340, 129985.
26. M. Aleem, R. Vishnuraj and B. Pullithadathil, RSC Appl. Interfaces, 2025, 2, 220–229.
27. S. A. Iakab, P. Ràfols, M. Tajes, X. Correig-Blanchar and M. García-Altares, ACS Nano, 2020, 14, 6785–6794.
28. F. Tian, M. Li, S. Wu, L. Li and H. Hu, Colloids Surf., B, 2023, 222, 113092.
29. Y. Wang, Y. Yang, Y. Shi, H. Song and C. Yu, Adv. Mater., 2020, 32, 1904106.
30. D. H. K. Nguyen, V. T. H. Pham, M. Al Kobaisi, C. Bhadra, A. Orlowska, S. Ghanaati, B. M. Manzi, V. A. Baulin, S. Joudkazis, P. Kingshott, R. J. Crawford and E. P. Ivanova, Langmuir, 2016, 32, 10744–10751.
31. D. P. Linklater, S. Juodkazis, S. Rubanov and E. P. Ivanova, ACS Appl. Mater. Interfaces, 2017, 9, 29387–29393.
32. W. Xu, C. Pang, C. Song, J. Qian, S. Feola, V. Cerullo, L. Fan, H. Yu and V.-P. Lehto, Acta Biomater., 2022, 152, 473–483.
33. C. Nataraj, K. Mohanta and G. P. Badhirappan, J. Mater. Chem. C, 2025, 13, 5868–5879.
34. S. Zhao, Z. Li, D. P. Linklater, L. Han, P. Jin, L. Wen, C. Chen, D. Xing, N. Ren, K. Sun, S. Juodkazis, E. P. Ivanova and L. Jiang, Nano Lett., 2022, 22, 1129–1137.
35. Y.-L. Deng and Y.-J. Juang, Biosens. Bioelectron., 2014, 53, 37–42.
36. Z. Fan, D. F. Cui, Z. X. Zhang, Z. Zhao, H. M. Chen, Y. Y. Fan, P. L. Li, Z. D. Zhang, C. Y. Xue and S. B. Yan, Nanomaterials, 2021, 11, 41.
37. Z. Huang, N. Geyer, P. Werner, J. de Boor and U. Gösele, Adv. Mater., 2011, 23, 285–308.
38. C. Huo, J. Wang, H. Fu, X. Li, Y. Yang, H. Wang, A. Mateen, G. Farid and K.-Q. Peng, Adv. Funct. Mater., 2020, 30, 2005744.
39. J. H. Zhao, X. B. Li, Q. D. Chen, Z. G. Chen and H. B. Sun, Mater. Today Nano, 2020, 11, 100078.
40. A. Kumar, M. Devi, M. Kumar, A. Shrivastava, R. Sharma, T. Dixit, V. Singh, K. Shehzad, Y. Xu, K. Singh and H. Hu, Sens. Actuators, A, 2022, 347, 113912.
41. M. Otto, M. Algasinger, H. Branz, B. Gesemann, T. Gimpel, K. Füchsel, T. Käsebier, S. Kontermann, S. Koynov, X. Li, V. Naumann, J. Oh, A. N. Sprafke, J. Ziegler, M. Zilk and R. B. Wehrspohn, Adv. Opt. Mater., 2015, 3, 147–164.
42. L. Chen, S. Zhang, Y. Duan, X. Song, M. Chang, W. Feng and Y. Chen, Chem. Soc. Rev., 2024, 53, 1167–1315.
43. S. A. Ahad, T. Kennedy and H. Geaney, ACS Energy Lett., 2024, 9, 1548–1561.
44. S. Wei, X. Xia, S. Bi, S. Hu, X. Wu, H.-Y. Hsu, X. Zou, K. Huang, D. W. Zhang, Q. Sun, A. J. Bard, E. T. Yu and L. Ji, Chem. Soc. Rev., 2024, 53, 6860–6916.
45. J. H. Zhao, C. H. Li, X. B. Li, Q. D. Chen, Z. G. Chen and H. B. Sun, IEEE Trans. Electron Devices, 2018, 65, 4905–4909.
46. X. Zhou, L. Wan, H. Li, X. Yang, J. Chen, K. Ge, J. Yan, C. Zhang, Q. Gao, X. Zhang, J. Guo, F. Li, J. Wang, D. Song, S. Wang, B. S. Flavel and J. Chen, Small, 2023, 19, 2205848.
47. R. Komatsu, A. Balčytis, G. Seniutinas, T. Yamamura, Y. Nishijima and S. Juodkazis, Sol. Energy Mater. Sol. Cells, 2015, 143, 72–77.
48. C. Ramos, Y. Verduzco-López, C. Álvarez-Macías, A. Sanchez-Juarez, A. Dutt and G. Santana, Mater. Today Commun., 2019, 20, 100556.
49. M. Ali, F. Zhou, K. Chen, C. Kotzur, C. Xiao, L. Bourgeois, X. Zhang and D. R. MacFarlane, Nat. Commun., 2016, 7, 11335.
50. P. Hoyer, M. Theuer, R. Beigang and E.-B. Kley, Appl. Phys. Lett., 2008, 93, 091106.
51. L. Golubewa, R. Karpicz, I. Matulaitiene, A. Selskis, D. Rutkauskas, A. Pushkarchuk, T. Khlopina, D. Michels, D. Lyakhov, T. Kulahava, A. Shah, Y. Svirko and P. Kuzhir, ACS Appl. Mater. Interfaces, 2020, 12, 50971–50984.
52. H. Robbins and B. Schwartz, J. Electrochem. Soc., 1959, 106, 505.
53. H. Jansen, M. d Boer, R. Legtenberg and M. Elwenspoek, J. Micromech. Microeng., 1995, 5, 115.
54. D. Dimova-Malinovska, M. Sendova-Vassileva, N. Tzenov and M. Kamenova, Thin Solid Films, 1997, 297, 9–12.
55. T.-H. Her, R. J. Finlay, C. Wu, S. Deliwala and E. Mazur, Appl. Phys. Lett., 1998, 73, 1673–1675.
56. R. Memming and G. Schwandt, Surf. Sci., 1966, 4, 109–124.
57. N. W. Cheung, Mater. Chem. Phys., 1996, 46, 132–139.
58. C. Wu, C. H. Crouch, L. Zhao, J. E. Carey, R. Younkin, J. A. Levinson, E. Mazur, R. M. Farrell, P. Gothoskar and A. Karger, Appl. Phys. Lett., 2001, 78, 1850–1852.
59. K. Yasuda, T. Nohira, K. Amezawa, Y. H. Ogata and Y. Ito, J. Electrochem. Soc., 2005, 152, D69.
60. B. Franta, D. Pastor, H. H. Gandhi, P. H. Rekemeyer, S. Gradečak, M. J. Aziz and E. Mazur, J. Appl. Phys., 2015, 118, 225303.
61. L. L. Janavicius, J. A. Michaels, C. Chan, D. J. Sievers and X. Li, Appl. Phys. Rev., 2023, 10, 011409.
62. A. Uhlir, Bell Syst. Tech. J., 1956, 35, 333–347.
63. Y. Bi, X. Su, S. Zou, Y. Xin, Z. Dai, J. Huang, X. Wang and L. Zhang, Thin Solid Films, 2012, 521, 176–180.
64. K. N. Nguyen, P. Basset, F. Marty, Y. Leprince-Wang and T. Bourouina, J. Appl. Phys., 2013, 113, 194903.
65. Z. Zhang, T. Martinsen, G. Liu, M. Tayyib, D. Cui, M. J. de Boer, F. Karlsen, H. Jakobsen, C. Xue and K. Wang, Adv. Opt. Mater., 2020, 8, 2000668.
66. D. Schmelz, T. Käsebier, Z. Shi, Q. Cheng, N. Sergeev, D. Schelle and U. Zeitner, Mater. Sci. Semicond. Process., 2023, 165, 107646.
67. S. Sarkar, A. A. Elsayed, Y. M. Sabry, F. Marty, J. Drévillon, X. Liu, Z. Liang, E. Richalot, P. Basset, E. Nefzaoui and T. Bourouina, Adv. Photonics Res., 2023, 4, 2200223.
68. X. Li and P. W. Bohn, Appl. Phys. Lett., 2000, 77, 2572–2574.
69. S. Koynov, M. S. Brandt and M. Stutzmann, Appl. Phys. Lett., 2006, 88, 203107.
70. J. Q. Huang, S. Y. Chiam, H. H. Tan, S. J. Wang and W. K. Chim, Chem. Mater., 2010, 22, 4111–4116.
71. N. A. Md Noor, S. K. Mohamad, S. S. Hamil, M. Devarajan and M. Z. Pakhuruddin, Mater. Sci. Semicond. Process., 2019, 91, 167–173.
72. R. Younkin, J. E. Carey, E. Mazur, J. A. Levinson and C. M. Friend, J. Appl. Phys., 2003, 93, 2626–2629.
73. M. Sheehy, L. Winston, J. Carey, C. Friend and E. Mazur, Chem. Mater., 2005, 17, 3582–3586.
74. B. R. Tull, J. E. Carey, E. Mazur, J. P. McDonald and S. M. Yalisove, MRS Bull., 2006, 31, 626–633.
75. Y. Ma, J. Si, X. Sun, T. Chen and X. Hou, Appl. Surf. Sci., 2014, 313, 905–910.
76. Z. X. Jia, Q. Wu, X. R. Jin, S. Huang, J. Z. Li, M. Yang, H. Huang, J. H. Yao and J. J. Xu, Opt. Express, 2020, 28, 5239–5247.
77. A. Y. Vorobyev and C. Guo, Appl. Surf. Sci., 2011, 257, 7291–7294.
78. M. Kovalev, I. Podlesnykh, A. Nastulyavichus, N. Stsepuro, I. Mushkarina, P. Platonov, E. Terukov, S. Abolmasov, A. Dunaev, A. Akhmatkhanov, V. Shur and S. Kudryashov, Materials, 2023, 16, 2350.
79. X. Liu, B. Radfar, K. Chen, E. Pälikkö, T. P. Pasanen, V. Vähänissi and H. Savin, IEEE Photonics Technol. Lett., 2022, 34, 870–873.
80. Y. Xia, B. Liu, S. Zhong and C. Li, J. Electron Spectrosc. Relat. Phenom., 2012, 184, 589–592.
81. S. Zhong, B. Liu, Y. Xia, J. Liu, J. Liu, Z. Shen, Z. Xu and C. Li, Sol. Energy Mater. Sol. Cells, 2013, 108, 200–204.
82. J. W. M. Lim, S. Huang, L. Xu, Y. Y. Lim, Y. X. Loh, C. S. Chan, K. Bazaka, I. Levchenko and S. Xu, Sol. Energy, 2018, 171, 841–850.
83. Y. H. Liu, X. K. Wang, J. B. Luo and X. C. Lu, Appl. Surf. Sci., 2009, 255, 9430–9438.
84. M. Y. Arafat, M. A. Islam, A. W. B. Mahmood, F. Abdullah, T. S. Kiong and N. Amin, Crystals, 2021, 11, 881.
85. X. Y. Ao, X. L. Tong, D. S. Kim, L. B. Zhang, M. Knez, F. Müller, S. L. He and V. Schmidt, Appl. Phys. Lett., 2012, 101, 111901.
86. W. Chen, Y. Liu, L. Yang, J. Wu, Q. Chen, Y. Zhao, Y. Wang and X. Du, Sci. Rep., 2018, 8, 3408.
87. E. Vazsonyi, K. De Clercq, R. Einhaus, E. Van Kerschaver, K. Said, J. Poortmans, J. Szlufcik and J. Nijs, Sol. Energy Mater. Sol. Cells, 1999, 57, 179–188.
88. R. Liu, J. Wang, T. Sun, M. Wang, C. Wu, H. Zou, T. Song, X. Zhang, S.-T. Lee, Z. L. Wang and B. Sun, Nano Lett., 2017, 17, 4240–4247.
89. M. Kopani, M. Mikula, D. Kosnac, P. Vojtek, J. Gregus, E. Vavrinsky, M. Jergel and E. Pincik, Appl. Surf. Sci., 2018, 461, 44–47.
90. J. Müllerová, Ľ. Scholtz, J. Ďurišová, E. Pinčík, M. Solanská and D. Pudiš, Appl. Surf. Sci., 2018, 461, 182–189.
91. X. G. Zhang, S. D. Collins and R. L. Smith, J. Electrochem. Soc., 1989, 136, 1561.
92. T. Jiang, X. Xu and G. Z. Chen, J. Energy Chem., 2020, 47, 46–61.
93. P. R. Coxon, M. Coto, E. Juzeliunas and D. J. Fray, Prog. Nat. Sci.: Mater. Int., 2015, 25, 583–590.
94. M. M. Plakhotnyuk, M. Gaudig, R. S. Davidsen, J. M. Lindhard, J. Hirsch, D. Lausch, M. S. Schmidt, E. Stamate and O. Hansen, J. Appl. Phys., 2017, 122, 143101.
95. M. Steglich, T. Käsebier, M. Zilk, T. Pertsch, E.-B. Kley and A. Tünnermann, J. Appl. Phys., 2014, 116, 173503.
96. J. Xu, A. D. Refino, A. Delvallée, S. Seibert, C. Schwalb, P. E. Hansen, M. Foldyna, L. Siaudinyte, G. Hamdana, H. S. Wasisto, J. Kottmeier, A. Dietzel, T. Weimann, J. K. Prüssing, H. Bracht and E. Peiner, Appl. Phys. Rev., 2024, 11, 021411.
97. A. D. Refino, N. Yulianto, I. Syamsu, A. P. Nugroho, N. H. Hawari, A. Syring, E. Kartini, F. Iskandar, T. Voss, A. Sumboja, E. Peiner and H. S. Wasisto, Sci. Rep., 2021, 11, 19779.
98. N. Lim, Y. Pak, J. T. Kim, Y. Hwang, R. Lee, Y. Kumaresan, N. Myoung, H. C. Ko and G. Y. Jung, Nanoscale, 2015, 7, 13489–13494.
99. S. Wu, Q. Chen, L. Zhang, H. Ren, H. Zhou, L. Hu and C. S. Tan, Nanoscale, 2023, 15, 4843–4851.
100. V. R. Manfrinato, F. E. Camino, A. Stein, L. Zhang, M. Lu, E. A. Stach and C. T. Black, Adv. Funct. Mater., 2019, 29, 1903429.
101. S. Kim, G. S. Jeong, N. Y. Park and J.-Y. Choi, Micromachines, 2021, 12, 119.
102. Z. P. Li, Z. M. Xu, X. P. Qu, S. B. Wang, J. Peng and L. H. Mei, Nanotechnology, 2017, 28, 095301.
103. M. Jung, J.-H. Kim and Y.-W. Choi, J. Nanomater., 2018, 2018, 6249890.
104. R. Xu, Z. Zeng and Y. Lei, Nat. Commun., 2022, 13, 2435.
105. H. Zhong, N. Ilyas, Y. Song, W. Li and Y. Jiang, Nanoscale Res. Lett., 2018, 13, 316.
106. F. Laermer and A. Urban, Plasma Processes Polym., 2019, 16, 1800207.
107. Z. Zhang, G. Liu and K. Wang, Micromachines, 2021, 12, 1009.
108. E. Vassallo, M. Pedroni, S. M. Pietralunga, R. Caniello, A. Cremona, F. Di Fonzo, F. Ghezzi, F. Inzoli, G. Monteleone, G. Nava, V. Spampinato, A. Tagliaferri, M. Zani and G. Angella, Thin Solid Films, 2016, 603, 173–179.
109. F. Ghezzi, M. Pedroni, J. Kovač, F. Causa, A. Cremona, M. Anderle, R. Caniello, S. M. Pietralunga and E. Vassallo, ACS Omega, 2022, 7, 25600–25612.
110. H. M. Branz, V. E. Yost, S. Ward, K. M. Jones, B. To and P. Stradins, Appl. Phys. Lett., 2009, 94, 231121.
111. H. Zhong, A. Guo, G. Guo, W. Li and Y. Jiang, Nanoscale Res. Lett., 2016, 11, 322.
112. S. Surdo and G. Barillaro, Small, 2024, 20, 2400499.
113. C.-Y. Chen and C.-P. Wong, Nanoscale, 2015, 7, 1216–1223.
114. G. B. Jung, Y. J. Cho, Y. Myung, H. S. Kim, Y. S. Seo, J. Park and C. Kang, Opt. Express, 2010, 18, 16353–16359.
115. Y. Jeong, C. Hong, Y. H. Jung, R. Akter, H. Yoon and I. Yoon, Sci. Rep., 2019, 9, 15914.
116. J. Soueiti, R. Sarieddine, H. Kadiri, A. Alhussein, G. Lerondel and R. Habchi, Nanoscale, 2023, 15, 4738–4761.
117. X. Geng, M. Li, L. Zhao and P. W. Bohn, J. Electron. Mater., 2011, 40, 2480–2485.
118. K. Nishioka, T. Sueto and N. Saito, Appl. Surf. Sci., 2009, 255, 9504–9507.
119. K. Peng, M. Zhang, A. Lu, N.-B. Wong, R. Zhang and S.-T. Lee, Appl. Phys. Lett., 2007, 90, 163123.
120. M. O. Williams, D. Hiller, T. Bergfeldt and M. Zacharias, J. Phys. Chem. C, 2017, 121, 9296–9299.
121. A. Mallavarapu, P. Ajay, C. Barrera and S. V. Sreenivasan, ACS Appl. Mater. Interfaces, 2021, 13, 1169–1177.
122. P. Zhang, S. Li, C. Liu, X. Wei, Z. Wu, Y. Jiang and Z. Chen, Nanoscale Res. Lett., 2014, 9, 519.
123. S. Uddin, M. R. Hashim and M. Z. Pakhuruddin, Mater. Sci. Semicond. Process., 2021, 133, 105932.
124. M. Li, Y. Li, W. Liu, L. Yue, R. Li, Y. Luo, M. Trevor, B. Jiang, F. Bai, P. Fu, Y. Zhao, C. Shen and J. M. Mbengue, Mater. Res. Bull., 2016, 76, 436–449.
125. J.-J. Huang, C.-S. Huang, J.-Y. Wang, P.-Y. Chang, C.-N. Chen and S.-L. Ou, Appl. Phys. A, 2022, 128, 485.
126. S. Ray, A. Mondal and U. Gangopadhyay, Appl. Phys. A, 2020, 126, 399.
127. Z. Huang, H. Fang and J. Zhu, Adv. Mater., 2007, 19, 744–748.
128. C. Song, B. Ye, J. Xu, J. Chen, W. Shi, C. Yu, C. An, J. Zhu and W. Zhang, Small, 2022, 18, 2104202.
129. F. J. Wendisch, M. Rey, N. Vogel and G. R. Bourret, Chem. Mater., 2020, 32, 9425–9434.
130. D. Brodoceanu, R. Elnathan, B. Prieto-Simón, B. Delalat, T. Guinan, E. Kroner, N. H. Voelcker and T. Kraus, ACS Appl. Mater. Interfaces, 2015, 7, 1160–1169.
131. Y. Chen, C. Zhang, L. Li, C.-C. Tuan, F. Wu, X. Chen, J. Gao, Y. Ding and C.-P. Wong, Nano Lett., 2017, 17, 4304–4310.
132. O. J. Hildreth, A. G. Fedorov and C. P. Wong, ACS Nano, 2012, 6, 10004–10012.
133. O. J. Hildreth and D. R. Schmidt, Adv. Funct. Mater., 2014, 24, 3827–3833.
134. X. Zhang, Y. Liu, C. Yao, J. Niu, H. Li and C. Xie, Nanoscale Adv., 2023, 5, 142–152.
135. D. A. Zuev, O. A. Novodvorsky, E. V. Khaydukov, O. D. Khramova, A. A. Lotin, L. S. Parshina, V. V. Rocheva, V. Y. Panchenko, V. V. Dvorkin, A. Y. Poroykov, G. G. Untila, A. B. Chebotareva, T. N. Kost and M. A. Timofeyev, Appl. Phys. B, 2011, 105, 545–550.
136. X. Y. Yu, Z. H. Lv, C. H. Li, X. Han and J. H. Zhao, IEEE Sens. J., 2016, 16, 5227–5231.
137. H. Huang, L. M. Yang and J. Liu, Opt. Eng., 2014, 53, 051513.
138. S. K. Sundaram and E. Mazur, Nat. Mater., 2002, 1, 217–224.
139. K. C. Phillips, H. H. Gandhi, E. Mazur and S. K. Sundaram, Adv. Opt. Photonics, 2015, 7, 684–712.
140. C. H. Crouch, J. E. Carey, J. M. Warrender, M. J. Aziz, E. Mazur and F. Y. Génin, Appl. Phys. Lett., 2004, 84, 1850–1852.
141. M.-J. Sher, N. M. Mangan, M. J. Smith, Y.-T. Lin, S. Marbach, T. M. Schneider, S. Gradečak, M. P. Brenner and E. Mazur, J. Appl. Phys., 2015, 117, 125301.
142. Y. Ma, H. Ren, J. Si, X. Sun, H. Shi, T. Chen, F. Chen and X. Hou, Appl. Surf. Sci., 2012, 261, 722–726.
143. T. Gimpel, S. Winter, M. Boßmeyer and W. Schade, Sol. Energy Mater. Sol. Cells, 2018, 180, 168–172.
144. J. Meng, H. Song, X. Li and S. Liu, Appl. Phys. A, 2015, 118, 1197–1203.
145. J. Meng, H. Song, X. Li and S. Liu, J. Laser Appl., 2015, 28, 012005.
146. H. Mei, C. Wang, J. Yao, Y.-C. Chang, J. Cheng, Y. Zhu, S. Yin and C. Luo, Opt. Commun., 2011, 284, 1072–1075.
147. S. Sarkar, E. Nefzaoui, G. Hamaoui, F. Marty, P. Basset and T. Bourouina, Appl. Phys. Lett., 2022, 121, 231703.
148. S. Qin, M. I. Current, S. B. Felch, N. W. Cheung and J. Ziegler, Ion Implantation Applications, Science and Technology, 2012, 15, pp. 2–41.
149. X. Liu, B. Radfar, K. Chen, O. E. Setälä, T. P. Pasanen, M. Yli-Koski, H. Savin and V. Vähänissi, IEEE Trans. Semicond. Manuf., 2022, 35, 504–510.
150. X. Tan, Z. Tao, M. Yu, H. Wu and H. Li, Sci. Rep., 2018, 8, 7863.
151. W. Meng, H. Zhong, W. Hou, Y. Gou, Y. Song and W. Li, IOP Conf. Ser.: Mater. Sci. Eng., 2017, 250, 012015.
152. S. Zhou, H. Fan, S. Wen, R. Zhang, Y. Yin, C. Lan, C. Li and Y. Liu, Adv. Opt. Mater., 2024, 12, 2301982.
153. P. Yu, J. Wu, S. Liu, J. Xiong, C. Jagadish and Z. M. Wang, Nano Today, 2016, 11, 704–737.
154. J. Oh, T. G. Deutsch, H.-C. Yuan and H. M. Branz, Energy Environ. Sci., 2011, 4, 1690–1694.
155. S.-X. Zhang, J.-C. Wang, Y.-M. Zhao, Y.-L. Han, A.-J. Ming, F. Wei and C.-H. Mao, Rare Met., 2023, 42, 2447–2456.
156. X. Wang, B. Zhao, J. Nivas, X. Zhao, W. Du and S. Amoruso, Opt. Express, 2024, 32, 36362–36380.
157. J.-X. Wang, B. Cui, C. Salmean, X. Chen, X. Yan, Y. Mao and S. Yao, Nano Energy, 2024, 125, 109560.
158. C. Yao, Y. Liu, J. Niu, C. Lu, H. Li and C. Xie, Appl. Surf. Sci., 2024, 655, 159641.
159. Z. Zhang, Y. Wang, P. A. Stensby Hansen, K. Du, K. R. Gustavsen, G. Liu, F. Karlsen, O. Nilsen, C. Xue and K. Wang, Nano Energy, 2019, 65, 103992.
160. A. Rahman, A. Ashraf, H. Xin, X. Tong, P. Sutter, M. D. Eisaman and C. T. Black, Nat. Commun., 2015, 6, 5963.
161. U. K. Mishra and J. Singh, Semiconductor device physics and design, Springer, 2008.
162. S. M. Sze and K. K. Ng, Physics of Semiconductor Devices, John Wiley & Sons, 2006.
163. S. Carapezzi, A. Castaldini, F. Mancarella, A. Poggi and A. Cavallini, ACS Appl. Mater. Interfaces, 2016, 8, 10443–10450.
164. S. Kalem, J. Mater. Sci.: Mater. Electron., 2023, 34, 724.
165. F. Guo, B. Yang, Y. Yuan, Z. Xiao, Q. Dong, Y. Bi and J. Huang, Nat. Nanotechnol., 2012, 7, 798–802.
166. Y. Fang, A. Armin, P. Meredith and J. Huang, Nat. Photonics, 2019, 13, 1–4.
167. P. Scherpelz and G. Galli, Phys. Rev. Mater., 2017, 1, 021602.
168. S. Olibet, E. Vallat-Sauvain and C. Ballif, Phys. Rev. B, 2007, 76, 035326.
169. C. R. Helms and B. E. Deal, The Physics and Chemistry of SiO2 and the Si-SiO2 Interface, Springer Science & Business Media, 1988.
170. J.-H. Yang, L. Shi, L.-W. Wang and S.-H. Wei, Sci. Rep., 2016, 6, 21712.
171. L. E. Black and K. R. McIntosh, IEEE Trans. Electron Devices, 2010, 57, 1996–2004.
172. S. L. Pain, L. Wilkins, A. Yadav, Y. Han, R. Beanland, N. E. Grant and J. D. Murphy, Sol. Energy Mater. Sol. Cells, 2025, 282, 113439.
173. N. E. Grant, S. L. Pain, E. Khorani, R. Jefferies, A. Wratten, S. McNab, D. Walker, Y. Han, R. Beanland, R. S. Bonilla and J. D. Murphy, Appl. Surf. Sci., 2024, 645, 158786.
174. K. H. Min, J.-M. Hwang, C. Chen, W.-J. Choi, V. D. Upadhyaya, B. Bounsaville, A. Rohatgi and Y.-W. Ok, Sol. Energy Mater. Sol. Cells, 2024, 273, 112922.
175. A. Damm, M. Bories, J. Benick, M. Hanser, A. Richter, A. Liu, Z. Yang, S. Lange, P.-T. Miclea and J.-I. Polzin, Sol. Energy Mater. Sol. Cells, 2025, 285, 113542.
176. H. F. Haneef, A. M. Zeidell and O. D. Jurchescu, J. Mater. Chem. C, 2020, 8, 759–787.
177. D. Spassov, A. Paskaleva, T. A. Krajewski, E. Guziewicz and G. Luka, Nanotechnology, 2018, 29, 505206.
178. X. Yue, J. Fan and Q. Xiang, Adv. Funct. Mater., 2022, 32, 2110258.
179. Z. Zhang, Y. Wang, T. Guo and P. Hu, Catalysts, 2025, 15, 651.
180. P. Wiśniewski and B. Majkusiak, Materials, 2022, 15, 2733.
181. J. H. Park, F. H. Alshammari, Z. Wang and H. N. Alshareef, Adv. Mater. Interfaces, 2016, 3, 1600713.
182. Y. S. Rim, H. Chen, B. Zhu, S.-H. Bae, S. Zhu, P. J. Li, I. C. Wang and Y. Yang, Adv. Mater. Interfaces, 2017, 4, 1700020.
183. W. Wang, H. Li, X. Liu, S. Ma, Y. Zhao, B. Dong, Y. Li, X. Ning, L. Zhao and J. Zhuang, Sens. Actuators, B, 2023, 382, 133473.
184. W. Wang, S. Ma, X. Liu, Y. Zhao, H. Li, Y. Li, X. Ning, L. Zhao and J. Zhuang, Sens. Actuators, B, 2022, 354, 131193.
185. B. Dong, W. Wang, X.-L. Liu, H. Li, Y. Li, Y. Huang, X. Ning, L. Zhao and J. Zhuang, Opt. Express, 2024, 32, 13384–13395.
186. Y. Liu, T. Lai, H. Li, Y. Wang, Z. Mei, H. Liang, Z. Li, F. Zhang, W. Wang, A. Y. Kuznetsov and X. Du, Small, 2012, 8, 1392–1397.
187. G. Ayvazyan, K. Ayvazyan, L. Hakhoyan and A. Semchenko, Phys. Status Solidi – Rapid Res. Lett., 2023, 17, 2300058.
188. X.-L. Liu, S.-W. Zhu, H.-B. Sun, Y. Hu, S.-X. Ma, X.-J. Ning, L. Zhao and J. Zhuang, ACS Appl. Mater. Interfaces, 2018, 10, 5061–5071.
189. A. Y. Mironenko, M. V. Tutov, A. A. Sergeev, E. V. Mitsai, A. Y. Ustinov, A. Y. Zhizhchenko, D. P. Linklater, S. Y. Bratskaya, S. Juodkazis and A. A. Kuchmizhak, ACS Sens., 2019, 4, 2879–2884.
190. R. Boukherroub, Curr. Opin. Solid State Mater. Sci., 2005, 9, 66–72.
191. Y. Padrez and L. Golubewa, Biosensors, 2024, 14, 453.
192. F. A. Harraz, Sens. Actuators, B, 2014, 202, 897–912.
193. V. Myndrul and I. Iatsunskyi, Materials, 2019, 12, 2880.
194. Y. Jiang, H. Shen, T. Pu and C. Zheng, Appl. Phys. A, 2018, 124, 341.
195. T. P. Pasanen, H. S. Laine, V. Vähänissi, K. Salo, S. Husein and H. Savin, IEEE J. Photovolt., 2018, 8, 697–702.
196. S. Takamura, T. Aota, H. Iwata, S. Maenaka, K. Fujita, Y. Kikuchi and Y. Uesugi, Appl. Surf. Sci., 2019, 487, 755–765.
197. F. Yang, J. A. Aguiar, M. Fairchild, W. Vakki, S. Younan, Y. Zhou, L. Zhuo and J. Gu, Adv. Mater. Interfaces, 2019, 6, 1802085.
198. Z. Ying, Z. Yang, J. Zheng, H. Wei, L. Chen, C. Xiao, J. Sun, C. Shou, G. Qin, J. Sheng, Y. Zeng, B. Yan, X. Yang and J. Ye, Joule, 2022, 6, 2644–2661.
199. S. Huang, Q. Wu, Z. Jia, X. Jin, X. Fu, H. Huang, X. Zhang, J. Yao and J. Xu, Adv. Opt. Mater., 2020, 8, 1901808.
200. H. Neergaard Waltenburg and J. Yates, Chem. Rev., 1995, 95, 1589–1673.
201. R. J. Hamers and Y. Wang, Chem. Rev., 1996, 96, 1261–1290.
202. K. A. Perrine and A. V. Teplyakov, Chem. Soc. Rev., 2010, 39, 3256–3274.
203. Z. Zhao, B. Zhang, P. Li, W. Guo and A. Liu, Int. J. Photoenergy, 2014, 2014, 683654.
204. E. Özkol, P. Procel, Y. Zhao, L. Mazzarella, R. Medlin, P. Šutta, O. Isabella and M. Zeman, Phys. Status Solidi – Rapid Res. Lett., 2020, 14, 1900087.
205. J. A. Schaefer, Phys. B, 1991, 170, 45–68.
206. S. Zare Pakzad, M. Nasr Esfahani and B. E. Alaca, Mater. Today Commun., 2023, 34, 105002.
207. M. Nasr Esfahani, S. Zare Pakzad, T. Li, X. Li, Z. Tasdemir, N. Wollschläger, Y. Leblebici and B. E. Alaca, ACS Appl. Nano Mater., 2022, 5, 13276–13285.
208. R. S. Bonilla, I. Al-Dhahir, M. Yu, P. Hamer and P. P. Altermatt, Sol. Energy Mater. Sol. Cells, 2020, 215, 110649.
209. T. Hattori and T. Suzuki, Appl. Phys. Lett., 1983, 43, 470–472.
210. Y. Matsui and S. Adachi, J. Appl. Phys., 2013, 113, 173502.
211. W. Cheng, X. Wang, S. Zou, M. Ni, Z. Lu, L. Dai, J. Su, K. Yang and X. Su, J. Funct. Biomater., 2023, 14, 245.
212. G. Korotcenkov and E. Rusu, Phys. Status Solidi A, 2019, 216, 1900348.
213. F. Tian, D. F. Taber and A. V. Teplyakov, J. Am. Chem. Soc., 2011, 133, 20769–20777.
214. K. Guo, A. Sharma, R. J. Toh, E. Alvárez de Eulate, T. R. Gengenbach, X. Cetó, N. H. Voelcker and B. Prieto-Simón, Adv. Funct. Mater., 2019, 29, 1809206.
215. G. Y. Dong, H. W. Yang, S. J. Zeng, Z. Q. Shi, Y. J. Ma, C. Wen and W. B. Yang, Opt. Laser Technol., 2023, 164, 109517.
216. W. C. Hsu, Y.-S. Lu, J.-Y. Chyan and J. A. Yeh, Int. J. Photoenergy, 2012, 2012, 197514.
217. E. C. Garnett, Y.-C. Tseng, D. R. Khanal, J. Wu, J. Bokor and P. Yang, Nat. Nanotechnol., 2009, 4, 311–314.
218. N. Chery, M. Zhang, R. Monflier, N. Mallet, G. Seine, V. Paillard, J. M. Poumirol, G. Larrieu, A. S. Royet, S. Kerdilès, P. Acosta-Alba, M. Perego, C. Bonafos and F. Cristiano, J. Appl. Phys., 2022, 131, 065301.
219. Z. Shen, Y. Xia, B. Liu, J. Liu, C. Li and Y. Li, Chin. Phys. B, 2014, 23, 118801.
220. L.-X. Yang, Y.-M. Chao, L. Jia and C.-B. Li, Appl. Therm. Eng., 2016, 99, 253–261.
221. G. von Gastrow, P. Ortega, R. Alcubilla, S. Husein, T. Nietzold, M. Bertoni and H. Savin, J. Appl. Phys., 2017, 121, 185706.
222. W.-C. Wang, C.-W. Lin, H.-J. Chen, C.-W. Chang, J.-J. Huang, M.-J. Yang, B. Tjahjono, J.-J. Huang, W.-C. Hsu and M.-J. Chen, ACS Appl. Mater. Interfaces, 2013, 5, 9752–9759.
223. S. P. Muduli and P. Kale, Mater. Sci. Semicond. Process., 2023, 154, 107202.
224. C.-H. Hsu, S.-M. Liu, W.-Y. Wu, Y.-S. Cho, P.-H. Huang, C.-J. Huang, S.-Y. Lien and W.-Z. Zhu, Arabian J. Chem., 2020, 13, 8239–8247.
225. M. Algasinger, J. Paye, F. Werner, J. Schmidt, M. S. Brandt, M. Stutzmann and S. Koynov, Adv. Energy Mater., 2013, 3, 1068–1074.
226. S. Wang, X. Wu, F. J. Ma, D. Payne, M. Abbott and B. Hoex, IEEE J. Photovolt., 2021, 11, 897–907.
227. X.-Y. Zhang, J. Han, Y.-T. Wang, Y.-J. Ruan, W.-Y. Wu, D.-S. Wuu, J. Zuo, F.-M. Lai, S.-Y. Lien and W.-Z. Zhu, Sol. Energy Mater. Sol. Cells, 2023, 257, 112384.
228. W.-C. Wang, M.-C. Tsai, J. Yang, C. Hsu and M.-J. Chen, ACS Appl. Mater. Interfaces, 2015, 7, 10228–10237.
229. K. T. Wong and N. S. Lewis, Acc. Chem. Res., 2014, 47, 3037–3044.
230. W. Fan, H. Shen, B. Liu, L. Zhao, X. Zhang and H. Pan, Energies, 2023, 16, 6963.
231. J. Xu, W. Zhang, Y. Li, C. Liu, G. Zhang, Z. Shen, P. Verlinden and X. Yuan, Sol. Energy Mater. Sol. Cells, 2025, 289, 113658.
232. G. Y. Ayvazyan, A. V. Aghabekyan, L. A. Hakhoyan, M. V. Katkov and M. S. Lebedev, Passivation Properties of Atomic-Layer-Deposited Hafnium Oxide on Black Silicon Surface, Nis, Serbia, 2021.
233. T. Pasanen, V. Vähänissi, N. Theut and H. Savin, Energy Procedia, 2017, 124, 307–312.
234. A. Kabalan, Appl. Sci., 2019, 9, 1804.
235. A. Richter, H. Patel, C. Reichel, J. Benick and S. W. Glunz, Adv. Mater. Interfaces, 2023, 10, 2202469.
236. G. Ayvazyan, A. Vaseashta, F. Gasparyan and S. Khudaverdyan, J. Mater. Sci.: Mater. Electron., 2022, 33, 17001–17010.
237. J.-W. Song, Y.-H. Nam, M.-J. Park, B. Yoo, J.-S. Cho, R. B. Wehrspohn and J.-H. Lee, Opt. Express, 2016, 24, A1224–A1233.
238. M. V. Katkov, G. Y. Ayvazyan, V. R. Shayapov and M. S. Lebedev, J. Contemp. Phys. (Arm. Acad. Sci.), 2020, 55, 16–22.
239. K. G. Ayvazyan, M. V. Katkov, M. S. Lebedev and G. Y. Ayvazyan, Odessa, Ukraine, 2021.
240. X. Tan, R. Chen and F. E. Rougieux, IEEE J. Photovolt., 2021, 11, 1380–1387.
241. K. H. Min, J.-M. Hwang, E. Cho, H.-e Song, S. Park, A. Rohatgi, D. Kim, H.-S. Lee, Y. Kang, Y.-W. Ok and M. G. Kang, Prog. Photovoltaics Res. Appl., 2021, 29, 54–63.
242. M. Sui, Y. Chu and R. Zhang, J. Phys.: Conf. Ser., 2021, 1907, 012026.
243. A. Soman and A. Antony, Mater. Chem. Phys., 2017, 197, 181–191.
244. J. Wang, T. Y. Lai, H.-T. Lin, T.-R. Kuo, H.-C. Chen, C.-S. Tseng, C.-W. Tung, C.-Y. Chien and H. M. Chen, Angew. Chem., Int. Ed., 2024, 63, e202403333.
245. G. Hazell, P. W. May, P. Taylor, A. H. Nobbs, C. C. Welch and B. Su, Biomater. Sci., 2018, 6, 1424–1432.
246. C. Zhou, Q. Mei, L. Huang, T. Mao, S. Li, Z. Wang, H. Wan, H. Gu and K. Han, ACS Appl. Mater. Interfaces, 2024, 16, 26153–26166.
247. M. A. Juntunen, J. Heinonen, V. Vähänissi, P. Repo, D. Valluru and H. Savin, Nat. Photonics, 2016, 10, 777–781.
248. T. E. Scheul, E. Khorani, T. Rahman, M. D. B. Charlton and S. A. Boden, Sol. Energy Mater. Sol. Cells, 2022, 235, 111448.
249. A. Tooghi, D. Fathi and M. Eskandari, Sci. Rep., 2020, 10, 18699.
250. M. Rashid, N. M. Ahmed, N. A. M. Noor and M. Z. Pakhuruddin, Mater. Sci. Semicond. Process., 2020, 115, 105113.
251. C. H. Crouch, J. E. Carey, M. Shen, E. Mazur and F. Y. Génin, Appl. Phys. A, 2004, 79, 1635–1641.
252. Y. Song, T. Liu, S. Liu, J. Huang, J. Li, C. Tian, T. Yu, Y. He, Y. Liu and Z. Zhong, J. Mater. Sci.: Mater. Electron., 2020, 31, 4696–4701.
253. K. Wang, S. Liu, J. Li, S. Wu, J. Xia, J. Chen, C. Tian, Y. Liu and Z. Zhong, J. Mater. Sci.: Mater. Electron., 2021, 32, 11503–11510.
254. H. Mertens, J. Verhoeven, A. Polman and F. D. Tichelaar, Appl. Phys. Lett., 2004, 85, 1317–1319.
255. A. O. Olaoye, S. Muhammad, M. H. Hussin, M. Z. Pakhuruddin, M. R. Hashim and M. Rashid, Mater. Res. Bull., 2025, 181, 113072.
256. T. Chen, J. Si, X. Hou, S. Kanehira, K. Miura and K. Hirao, J. Appl. Phys., 2011, 110, 073106.
257. Q. Lü, J. Wang, C. Liang, L. Zhao and Z. Jiang, Opt. Lett., 2013, 38, 1274–1276.
258. S. Kalem, P. Werner, Ö. Arthursson, V. Talalaev, B. Nilsson, M. Hagberg, H. Frederiksen and U. Södervall, Nanotechnology, 2011, 22, 235307.
259. A. Serpengüzel, A. Kurt, I. Inanc, J. Cary and E. Mazur, J. Nanophotonics, 2008, 2, 021770.
260. J. R. Chen, D. C. Wang, D. S. Ren, M. Lu and M. J. Peng, Appl. Phys. A, 2021, 127, 154.
261. R. A. Sinton and A. Cuevas, Appl. Phys. Lett., 1996, 69, 2510–2512.
262. G. Kumaravelu, M. M. Alkaisi, A. Bittar, D. Macdonald and J. Zhao, Curr. Appl. Phys., 2004, 4, 108–110.
263. R. Khandelwal, U. Plachetka, B. Min, C. Moormann and H. Kurz, Microelectron. Eng., 2013, 111, 220–223.
264. C. H. Hsu, S. M. Liu, W. Y. Wu, Y. S. Cho, P. H. Huang, C. J. Huang, S. Y. Lien and W. Z. Zhu, Arabian J. Chem., 2020, 13, 8239–8247.
265. K. Chen, T. P. Pasanen, V. Vähänissi and H. Savin, IEEE J. Photovolt., 2019, 9, 974–979.
266. Y. Jiang, H. Shen, W. Yang, C. Zheng, Q. Tang, H. Yao, A. Raza, Y. Li and C. Huang, Appl. Phys. A, 2018, 124, 95.
267. Y. Wang, Y. P. Liu, H. L. Liang, Z. X. Mei and X. L. Du, Phys. Chem. Chem. Phys., 2013, 15, 2345–2350.
268. C. Lan, H. Zou, L. Wang, M. Zhang, S. Pan, Y. Ma, Y. Qiu, Z. L. Wang and Z. Lin, Adv. Mater., 2020, 32, 2005481.
269. B. Dai, G. M. Biesold, M. Zhang, H. Zou, Y. Ding, Z. L. Wang and Z. Lin, Chem. Soc. Rev., 2021, 50, 13646–13691.
270. F. Priolo, T. Gregorkiewicz, M. Galli and T. F. Krauss, Nat. Nanotechnol., 2014, 9, 19–32.
271. W. Peng, X. Wang, R. Yu, Y. Dai, H. Zou, A. C. Wang, Y. He and Z. L. Wang, Adv. Mater., 2017, 29, 1606698.
272. X. Wang, W. Peng, R. Yu, H. Zou, Y. Dai, Y. Zi, C. Wu, S. Li and Z. L. Wang, Nano Lett., 2017, 17, 3718–3724.
273. H. Zou, X. Li, W. Peng, W. Wu, R. Yu, C. Wu, W. Ding, F. Hu, R. Liu, Y. Zi and Z. L. Wang, Adv. Mater., 2017, 29, 1701412.
274. Y. Zhang, J. Y. Y. Loh and N. P. Kherani, Adv. Sci., 2022, 9, 2203234.
275. M. Garin, J. Heinonen, L. Werner, T. P. Pasanen, V. Vähänissi, A. Haarahiltunen, M. A. Juntunen and H. Savin, Phys. Rev. Lett., 2020, 125, 117702.
276. S. Huang, A. Wu, G. Song, J. Cao, J. Yao, Q. Wu, W. Gao and J. Xu, Chin. Opt. Lett., 2024, 22, 113801.
277. W. Tao, J. Sun, Y. Deng, Y. Lian, Z. Chen and L. Jiang, ACS Appl. Mater. Interfaces, 2024, 16, 65710–65724.
278. S. Zhao, G. Yuan, Q. Wang, W. Liu, R. Wang and S. Yang, Nanoscale, 2020, 12, 316–325.
279. R. A. Myers, R. Farrell, A. M. Karger, J. E. Carey and E. Mazur, Appl. Opt., 2006, 45, 8825–8831.
280. V. Naumann, M. Otto, R. B. Wehrspohn and C. Hagendorf, J. Vac. Sci. Technol., A, 2012, 30, 04D106.
281. X. Zhao, K. Lin, B. Zhao, W. Du, J. J. J. Nivas, S. Amoruso and X. Wang, Appl. Surf. Sci., 2023, 619, 156624.
282. Y. Su, S. Li, Z. Wu, Y. Yang, Y. Jiang, J. Jiang, Z. Xiao, P. Zhang and T. Zhang, Mater. Sci. Semicond. Process., 2013, 16, 619–624.
283. J.-H. Zhao, Y. Yang and C. Li, Opt. Quantum Electron., 2020, 52, 390.
284. P. Mc Kearney, S. Schäfer, S. Paulus, M. Roser, F. Piermaier, I. Lebershausen and S. R. Kontermann, J. Appl. Phys., 2023, 133, 013102.
285. H. Zhong, N. Ilyas, Y. Song, W. Li and Y. Jiang, Nanoscale Res. Lett., 2018, 13, 1–8.
286. C. Li, J. H. Zhao, X. H. Liu, Z. Y. Ren, Y. Yang, Z. G. Chen, Q. D. Chen and H. B. Sun, IEEE Trans. Electron Devices, 2023, 70, 2364–2369.
287. H. Zou, G. Dai, A. C. Wang, X. Li, S. L. Zhang, W. Ding, L. Zhang, Y. Zhang and Z. L. Wang, Adv. Mater., 2020, 32, 1907249.
288. L. Jiang, X. Liu, J. Lv, G. Li, P. Yang, Y. Ma, H. Zou and Z. L. Wang, Energy Environ. Sci., 2024, 17, 3700–3738.
289. K. Lozano Montero, R. Calvo Guzman, A. Tewari, H. Zou, Z. L. Wang, M. Mäntysalo and M.-M. Laurila, Adv. Funct. Mater., 2024, 34, 2312881.
290. Z. Zhou, J. Lv, C. Tan, L. Yang and Z. Wang, Adv. Funct. Mater., 2024, 34, 2316175.
291. C.-Y. Liu, D.-L. Li, Z.-H. Wang, Y. Li, S.-Y. Zhou, L. Xu, G.-J. Zhong, H.-D. Huang and Z.-M. Li, Macromolecules, 2023, 56, 1481–1491.
292. A. Vaseashta, G. Ayvazyan, S. Khudaverdyan and L. Matevosyan, Phys. Status Solidi – Rapid Res. Lett., 2023, 17, 2200482.
293. I. R. Putra, J.-Y. Li and C.-Y. Chen, Appl. Surf. Sci., 2019, 478, 725–732.
294. G. Ayvazyan, F. Gasparyan and V. Gasparian, Opt. Mater., 2023, 140, 113879.
295. S. Ray, S. Mitra, H. Ghosh, A. Mondal, C. Banerjee and U. Gangopadhyay, J. Mater. Sci.: Mater. Electron., 2021, 32, 2590–2600.
296. S. Wu, W. Cui, N. Aghdassi, T. Song, S. Duhm, S.-T. Lee and B. Sun, Adv. Funct. Mater., 2016, 26, 5035–5041.
297. B. Chen, Z. J. Yu, S. Manzoor, S. Wang, W. Weigand, Z. Yu, G. Yang, Z. Ni, X. Dai, Z. C. Holman and J. Huang, Joule, 2020, 4, 850–864.
298. C. Xie, B. Nie, L. Zeng, F.-X. Liang, M.-Z. Wang, L. Luo, M. Feng, Y. Yu, C.-Y. Wu, Y. Wu and S.-H. Yu, ACS Nano, 2014, 8, 4015–4022.
299. C. W. Jang, D. H. Shin and S.-H. Choi, J. Alloys Compd., 2021, 877, 160311.
300. K. Oh, L. Joanny, F. Gouttefangeas, B. Fabre, V. Dorcet, B. Lassalle-Kaiser, A. Vacher, C. Mériadec, S. Ababou-Girard and G. Loget, ACS Appl. Energy Mater., 2019, 2, 1006–1010.
301. M. Patrick, F. Yang, W. Vakki, J. A. Aguiar and J. Gu, Adv. Funct. Mater., 2021, 31, 2008888.
302. B. Wang, H. Wu, G. Xu, X. Zhang, X. Shu, J. Lv and Y.-C. Wu, ACS Sustainable Chem. Eng., 2019, 7, 17598–17605.
303. W. Qiu, Y. Liu, M. Xie, Z. Jin, P. Li and G. Yu, EES Catal., 2024, 2, 202–219.
304. H. Li, S. Li, R. Guan, Z. Jin, D. Xiao, Y. Guo and P. Li, ACS Catal., 2024, 14, 12042–12050.
305. K. Liu, H. Li, M. Xie, P. Wang, Z. Jin, Y. Liu, M. Zhou, P. Li and G. Yu, J. Am. Chem. Soc., 2024, 146, 7779–7790.
306. B. Wang, L. Yao, G. Xu, X. Zhang, D. Wang, X. Shu, J. Lv and Y. C. Wu, ACS Appl. Mater. Interfaces, 2020, 12, 20376–20382.
307. U. Blumröder, M. Zilk, H. Hempel, P. Hoyer, T. Pertsch, R. Eichberger, T. Unold and S. Nolte, Opt. Express, 2017, 25, 6604–6620.
308. C. Langer, C. Prommesberger, R. Ławrowski, R. Schreiner, P. Serbun, G. Müller, F. Düsberg, M. Hofmann, M. Bachmann and A. Pahlke, J. Vac. Sci. Technol., B, 2016, 34, 02G107.
309. R. M. Young, H. C. Nathanson, R. S. Howell, E. J. Stewart, B. A. Nechay, T. T. Braggins, E. M. Graves, S. D. Van Campen, R. C. Clarke, S. B. Miserendino and J. Hawk, J. Vac. Sci. Technol., B, 2010, 28, 1060–1065.
310. C. Langer, V. Bomke, M. Hausladen, R. Ławrowski, C. Prommesberger, M. Bachmann and R. Schreiner, J. Vac. Sci. Technol., B, 2019, 38, 013202.
311. Z. Xu, J. Jiang, M. R. Gartia and G. L. Liu, J. Phys. Chem. C, 2012, 116, 24161–24170.
312. E. C. Le Ru and P. G. Etchegoin, Principles of Surface-Enhanced Raman Spectroscopy, Elsevier, 2009.
313. X.-X. Zhou, R. Liu, L.-T. Hao and J.-F. Liu, Talanta, 2021, 221, 121552.
314. S. Cong, Z. Wang, W. Gong, Z. Chen, W. Lu, J. R. Lombardi and Z. Zhao, Nat. Commun., 2019, 10, 678.
315. C. Matricardi, C. Hanske, J. L. Garcia-Pomar, J. Langer, A. Mihi and L. M. Liz-Marzán, ACS Nano, 2018, 12, 8531–8539.
316. G. A. Vinnacombe-Willson, Y. Conti, A. Stefancu, P. S. Weiss, E. Cortés and L. Scarabelli, Chem. Rev., 2023, 123, 8488–8529.
317. W. Yan, L. Yang, J. Chen, Y. Wu, P. Wang and Z. Li, Adv. Mater., 2017, 29, 1702893.
318. L. Golubewa, H. Rehman, Y. Padrez, A. Basharin, S. Sumit, I. Timoshchenko, R. Karpicz, Y. Svirko and P. Kuzhir, Materials, 2023, 16, 1948.
319. C. J. Orendorff, A. Gole, T. K. Sau and C. J. Murphy, Anal. Chem., 2005, 77, 3261–3266.
320. J. Liu, Y. Yan, Z. M. Zhang, Y. C. Liu, J. Ge and Z. S. Guan, Materials, 2023, 16, 3634.
321. L. Golubewa, A. Klimovich, I. Timoshchenko, Y. Padrez, M. Fetisova, H. Rehman, P. Karvinen, A. Selskis, S. Adomavičiūtė-Grabusovė, I. Matulaitienė, A. Ramanavicius, R. Karpicz, T. Kulahava, Y. Svirko and P. Kuzhir, ACS Appl. Nano Mater., 2023, 6, 4770–4781.
322. J. Li, A. Wuethrich, A. A. I. Sina, H.-H. Cheng, Y. Wang, A. Behren, P. N. Mainwaring and M. Trau, Nat. Commun., 2021, 12, 1087.
323. E. P. Ivanova, J. Hasan, H. K. Webb, G. Gervinskas, S. Juodkazis, V. K. Truong, A. H. F. Wu, R. N. Lamb, V. A. Baulin, G. S. Watson, J. A. Watson, D. E. Mainwaring and R. J. Crawford, Nat. Commun., 2013, 4, 2838.
324. D. P. Linklater, H. K. D. Nguyen, C. M. Bhadra, S. Juodkazis and E. P. Ivanova, Nanotechnology, 2017, 28, 245301.
325. E. P. Ivanova, D. P. Linklater, M. Werner, V. A. Baulin, X. M. Xu, N. Vrancken, S. Rubanov, E. Hanssen, J. Wandiyanto, V. K. Truong, A. Elbourne, S. Maclaughlin, S. Juodkazis and R. J. Crawford, Proc. Natl. Acad. Sci. U. S. A., 2020, 117, 12598–12605.
326. E. P. Ivanova, J. Hasan, H. K. Webb, V. K. Truong, G. S. Watson, J. A. Watson, V. A. Baulin, S. Pogodin, J. Y. Wang, M. J. Tobin, C. Löbbe and R. J. Crawford, Small, 2012, 8, 2489–2494.
327. D. W. Green, G. S. Watson, J. Watson and S. J. K. Abraham, Adv. Healthcare Mater., 2012, 1, 140–148.
328. R. W. van Nieuwenhoven, A. M. Bürger, L. L. E. Mears, P. Kienzl, M. Reithofer, A. Elbe-Bürger and I. C. Gebeshuber, Appl. Nanosci., 2024, 14, 531–541.
329. J. Hasan, H. K. Webb, V. K. Truong, S. Pogodin, V. A. Baulin, G. S. Watson, J. A. Watson, R. J. Crawford and E. P. Ivanova, Appl. Microbiol. Biotechnol., 2013, 97, 9257–9262.
330. C. M. Bhadra, M. Werner, V. A. Baulin, V. K. Truong, M. A. Kobaisi, S. H. Nguyen, A. Balcytis, S. Juodkazis, J. Y. Wang, D. E. Mainwaring, R. J. Crawford and E. P. Ivanova, Nano-Micro Lett., 2018, 10, 36.
331. P. W. May, M. Clegg, T. A. Silva, H. Zanin, O. Fatibello-Filho, V. Celorrio, D. J. Fermin, C. C. Welch, G. Hazell, L. Fisher, A. Nobbs and B. Su, J. Mater. Chem. B, 2016, 4, 5737–5746.
332. S. Ghosh, S. Niu, M. Yankova, M. Mecklenburg, S. M. King, J. Ravichandran, R. K. Kalia, A. Nakano, P. Vashishta and P. Setlow, Sci. Rep., 2017, 7, 17768.
333. S. Wang, Z. Liu, L. Wang, J. Xu, R. Mo, Y. Jiang, C. Wen, Z. Zhang and L. Ren, ACS Appl. Mater. Interfaces, 2023, 15, 723–735.
334. J. Butler, R. D. Handy, M. Upton and A. Besinis, ACS Nano, 2023, 17, 7064–7092.
335. Y. He, C. Fan and S.-T. Lee, Nano Today, 2010, 5, 282–295.
336. H. Alhmoud, D. Brodoceanu, R. Elnathan, T. Kraus and N. H. Voelcker, Prog. Mater. Sci., 2021, 116, 100636.
337. A. Zhang, J. Fang, X. Li, J. Wang, M. Chen, H.-J. Chen, G. He and X. Xie, Nanoscale Adv., 2022, 4, 1844–1867.
338. J. F. Zimmerman, R. Parameswaran, G. Murray, Y. Wang, M. Burke and B. Tian, Sci. Adv., 2016, 2, e1601039.
339. C. Wang, C. Gu, C. Popp, P. Vashisth, S. A. Mustfa, D. A. Martella, C. Spiteri, S. McLennan, N. Sun, M. Riddle, C. R. Eide, M. Parsons, J. Tolar, J. A. McGrath and C. Chiappini, ACS Nano, 2024, 18, 14938–14953.
340. W. Park, V. P. Nguyen, Y. Jeon, B. Kim, Y. Li, J. Yi, H. Kim, J. W. Leem, Y. L. Kim, D. R. Kim, Y. M. Paulus and C. H. Lee, Sci. Adv., 2022, 8, eabn1772.
341. C. O'Mahony, R. Sebastian, F. Tjulkins, D. Whelan, A. Bocchino, Y. Hu, J. O'Brien, J. Scully, M. Hegarty, A. Blake, I. Slimi, A. J. P. Clover, A. Lyness and A.-M. Kelleher, Int. J. Pharm., 2023, 637, 122888.
342. P. Premnath, A. Tavangar, B. Tan and K. Venkatakrishnan, Exp. Cell Res., 2015, 337, 44–52.
343. H. Gao, F. Yang, K. Sattari, X. Du, T. Fu, S. Fu, X. Liu, J. Lin, Y. Sun and J. Yao, Sci. Adv., 2022, 8, eabn2485.
344. D. Fan, E. De Rosa, M. B. Murphy, Y. Peng, C. A. Smid, C. Chiappini, X. Liu, P. Simmons, B. K. Weiner, M. Ferrari and E. Tasciotti, Adv. Funct. Mater., 2012, 22, 282–293.
345. M.-A. Shahbazi, P. V. Almeida, A. Correia, B. Herranz-Blanco, N. Shrestha, E. Mäkilä, J. Salonen, J. Hirvonen and H. A. Santos, J. Controlled Release, 2017, 249, 111–122.
346. T. Tieu, M. Alba, R. Elnathan, A. Cifuentes-Rius and N. H. Voelcker, Adv. Ther., 2019, 2, 1800095.
347. J. Shi, C. Sun, E. Liang and B. Tian, Adv. Funct. Mater., 2022, 32, 2107997.
348. X. Yue, W. Liu and Y. Wang, Surf. Coat. Technol., 2016, 307, 278–286.
349. T. Mouterde, G. Lehoucq, S. Xavier, A. Checco, C. T. Black, A. Rahman, T. Midavaine, C. Clanet and D. Quéré, Nat. Mater., 2017, 16, 658–663.
350. Z. Ma, C. Jiang, W. Yuan and Y. He, Nano-Micro Lett., 2013, 5, 7–12.
351. S. Han, R. Yang, C. Li and L. Yang, Appl. Sci., 2019, 9, 566.
352. C.-H. Choi and C.-J. Kim, Phys. Rev. Lett., 2006, 96, 066001.
353. C. Lee and C.-J. C. Kim, Langmuir, 2009, 25, 12812–12818.
354. F. Alexander, M. AlMheiri, P. Dahal, J. Abed, N. S. Rajput, C. Aubry, J. Viegas and M. Jouiad, Sol. Energy Mater. Sol. Cells, 2018, 180, 236–242.
355. Y. Arango, Y. Temiz, O. Gökçe and E. Delamarche, Sci. Adv., 2020, 6, eaay8305.
356. Y. Li, G. Feng, X. Liu, H. Li, B. Dong and L. Zhao, Chin. J. Phys., 2025, 93, 381–388.
357. J. H. Bang, M. S. Choi, A. Mirzaei, W. Oum, S. Han, S. S. Kim and H. W. Kim, Ceram. Int., 2020, 46, 604–611.
358. M. Akbari-Saatlu, M. Procek, C. Mattsson, G. Thungström, H. E. Nilsson, W. J. Xiong, B. Q. Xu, Y. Li and H. H. Radamson, Nanomaterials, 2020, 10, 2215.
359. J. Y. H. Chai, B. Wong and J. Sunarso, Photonics, 2023, 10, 565.
360. G. J. Xu, S. Q. Cheng and B. Cai, Mater. Express, 2018, 8, 294–298.
361. P. Cheng, H. Wang, B. Müller, J. Müller, D. Wang and P. Schaaf, ACS Appl. Mater. Interfaces, 2021, 13, 1818–1826.
362. Z. Wang, I. Kulikov, T. Mustafa, S. Schott, R. L. Carey, J. Behrends and H. Sirringhaus, Phys. Rev. Lett., 2025, 135, 166301.

---