Introduction

Silicon stands as the second most abundant element on earth, surpassed only by oxygen. Its ample presence and cost-effectiveness has made it one of the most accessible inorganic materials available. Over the past few decades, the global semiconductor-related industries have experienced rapid growth, underlining the significant role of silicon. One mainstream application of silicon in recent technological advancements has been in the realm of energy storage. Compared to graphite, silicon boasts a specific capacity that is more than 10 times greater (4200 mAh/g as opposed to 372 mAh/g) [1]. This makes it a prime candidate for high-capacity battery anode materials. Further studies have shown that when silicon particles are sized below 150 nm, they can mitigate the cracking issues experienced during the repeated lithiation and delithiation processes [2]. Another major application lies in the optical domain. Silicon nanoparticles with sizes ranging from 100 to 200 nm can be used for applications that exhibit structural colors due to Mie resonance [3]. When they are scaled down to the order of 1 nm, it emerges as a photonic material, specifically quantum dots. Intriguingly, the confinement effect dictates that the smaller these quantum dots are, the stronger their luminescent properties.

In this review, we aim to overview the fabrication technology of silicon nanoparticles. Then, we look into the characterization method of fabricated silicon nanoparticles and their applications. We also offer perspectives on future directions of silicon nanoparticle technology.

Fabrication

Fig. 1
figure 1

Various production methods for silicon nanoparticles. a Grinding of silicon lumps in liquid (Top-down). Reprinted from [4] with permission from American Chemical Society. b Laser printing on to the glass substrate (Top-down). Reprinted from [5] with permission from Springer Nature. c Laser pyrolysis of silane (SiH4) (Bottom-up). Reprinted from [6] with permission from Springer Nature. d Plasma pyrolysis of silane (SiH4) (Bottom-up). Reprinted from [7] with permission from American Chemical Society. e Reduction of tetrachlorosilane (SiCl4) (Bottom-up). Reprinted from [8] with permission from John Wiley and Sons. f Carbonthermal reduction of silica (Reduction). Reprinted from [9] (CC BY 4.0). g Magnesiothermic reduction of silica (Reduction). Reprinted from [10] (CC BY 4.0)

Full size image

Among the fabrication methods for silicon nanoparticles, the top-down approach stands out in terms of quality attributes such as purity and uniformity. As the name intuitively suggests, the top-down method involves breaking down bulk silicon into nanoparticles. One of the strategies for this is through the electrochemical etching of silicon wafers [11]. For p-type (100) silicon wafers with resistivity ranging from 1–10 \(\Omega\)-cm, a mixture of HF and H2O serves as the etchant. This procedure can yield colloidal suspensions of silicon particles around 150 nm in diameter. Silicon nanoparticles can also be mechanically produced through grinding, either by utilizing a mechanical milling process or grinding bulk silicon in a blender (Fig. 1a) [4, 12]. While the distribution of the fabricated particles tends to be broad, stratification by size within the solution allows for its minimization, enabling the production of silicon nanoparticles ranging from 100–400 nm in diameter. These methods have the advantage of a relatively simple production process, but a drawback is the reduced sphericity of the resulting silicon nanoparticles. On the other hand, laser ablation offers a method to produce spherical silicon nanoparticles. Similar to the previous method, it subtracts from bulk silicon, utilizing laser irradiation in a liquid to produce a colloidal solution [13,14,15]. While nanosecond, picosecond, and femtosecond lasers are employed, shorter pulses have the advantage of minimizing thermal effects and the interaction between laser pulses and cavitation bubbles, thereby producing smaller silicon nanoparticles [16]. The size distribution of the produced nanoparticles tends to be broad, but particles several nm and larger can be manufactured. Beyond colloidal forms, there’s also a method involving laser printing onto a substrate. By exposing a silicon wafer to a ring-shaped femtosecond laser intensity, molten silicon nanodroplets can be transferred to a glass surface, allowing for the fabrication of silicon nanoparticles larger than 400 nm in diameter [17]. Alternatively, when a Gaussian laser pulse is applied to the top of an SOI wafer, the device layer undergoes rapid heating and localized melting, facilitating its transfer to the upper glass substrate (Fig. 1b) [5]. This produces silicon nanoparticles with diameters exceeding 160 nm. While there are production volume limitations, a notable advantage is the ability to position high-sphericity silicon nanoparticles precisely where desired on the substrate.

Table 1 Advantages and challenges of silicon nanoparticle fabrication method
Full size table
Table 2 Size range and uniformity of silicon nanoparticle fabrication method
Full size table

The bottom-up fabrication method for silicon nanoparticles is advantageous in terms of mass production. Initially, silicon nanoparticles can be produced through the pyrolysis of silane (SiH4), a gaseous precursor, primarily via laser heating or plasma heating. For laser heating, CO2 lasers are predominantly used because silane has an efficient absorption at 10.6 \(\mu\)m wavelength [18]. By introducing a gas flow and heating it with a laser, silicon nanoparticles can be collected on a collector (Fig. 1c) [6, 19, 20]. Not only lasers, but plasmas can also be used to produce silicon particles (Fig. 1d) [7]. Using this method, silicon nanoparticles primarily larger than 10 nm can be produced in large quantities, but a drawback is the difficulty in controlling the particle size. Silicon nanoparticles can also be produced by reducing the liquid precursor silicon tetrachloride (SiCl4). Initially, silicon nanoparticles were produced under high-temperature and high-pressure conditions using Na as a reducing agent [21]. Subsequently, technologies were developed to produce them at room temperature using NaSi and KSi as reducing agents [22]. Recently, similar methods based on organic solutions have been used to produce silicon nanoparticles (Fig. 1e) [8, 23]. Through this method, silicon nanoparticles larger than 10 nm can be produced at a lower cost compared to using silane. However, they still face challenges in size control and the generation of by-products.

Another method to produce silicon nanoparticles is through the reduction of silica nanoparticles. Due to the strong bond between silicon and oxygen atoms, a significant amount of energy is required to reduce silica to silicon [24]. When silica nanoparticles are heated with carbon at high temperatures, a carbothermal reduction reaction occurs, enabling the production of silicon nanoparticles with diameters ranging from 80 to 200 nm (Fig. 1f) [9, 25]. Additionally, using magnesium as a reducing agent for silica, silicon nanoparticles can be produced through the magnesiothermic reduction reaction (Fig. 1g) [26, 27]. Since this process takes place at temperatures above the melting point of magnesium, which is 650 °C, it can be conducted at temperatures lower than the carbothermal reduction of silica particles, which requires temperatures above 2000 °C [28]. Through this method, it is possible to produce silicon nanoparticles at a scale of 10 nm [10]. However, magnesiothermic reduction for silicon production presents a significant limitation due to inevitable incomplete reduction, resulting in unreacted silica or magnesium silicide formation.

The advantages and challenges of each fabrication method explained previously are summarized in Table 1. In addition, the size range and uniformitiy of silicon nanoparticle produced by different fabrication processes are summarized in Table 2.

The fabricated silicon nanoparticles can have their energy levels finely tuned through a doping process. In the fabrication of silicon nanoparticles using a top-down process, the utilization of doped n-type or p-type silicon as the starting material enables the production of doped silicon nanoparticles [29]. In the bottom-up fabrication of silicon nanoparticles, the introduction of silane along with precursor gases enables the production of doped silicon nanoparticles [30]. The silicon nanoparticles can also be doped post fabrication through solution processing and annealing [31]. For boron doping, a boric acid solution is mixed with silicon nanoparticles and then dried to form a powder. Upon annealing the powder at 900 °C in an argon environment, doped silicon nanoparticles are obtained.

Fig. 2
figure 2

Characterization of silicon nanoparticles. Size, shape, mass, crystallinity, and other properties of silicon nanoparticles are measured with a variety of equipments considering whether nanoparticles are in powder form or in suspension

Full size image
Fig. 3
figure 3

a (left) XPS spectras, (center) Bright-field TEM images, (right) XRD patterns. Reprinted from [32] with permission from American Chemical Society. b HRTEM image of a region of milled silicon nanoparticle. Reprinted from [33] with permission from Springer Nature. c Calculated Raman spectra of silicon nanocrystals with various sizes. Reprinted from [34] with permission from AIP Publishing. d (top) Experimental scattering color visible range of selected silicon nanoparticles and (bottom) their corresponding SEM images for different values of diameters from 84 to 220 nm. Reprinted from [4] with permission from American Chemical Society. e UV–Vis spectra of the silicon nanoparticles. Reprinted from [35] with permission from Springer Nature. f FT-IR spectrum of silicon nanoparticle powder measured in diffuse reflection. Reprinted from [36] (CC BY 4.0). g Photoluminescence spectra of silicon nanocrystals of 2.5, 3.8 and 6.2 nm diameter. Reprinted from [37] (CC BY 4.0). H SANS data for a-Si and nano-Si before ball milling with corresponding fits (continuous lines) based on size distributions obtained via a Monte Carlo analysis. Reprinted from [38] (CC BY 4.0)

Full size image

Characterization

The fabricated silicon nanoparticles are stored either in powder or colloid forms and are characterized based on their shape, size, mass, crystallinity, and other properties (Fig. 2). Typically, for nanoparticles with diameters above 100 nm, Scanning Electron Microscopy (SEM) is used, while for those under 100 nm, Transmission Electron Microscopy (TEM) is the method of choice. While SEM and TEM are widely used to analyze both the shape and size of nanoparticles, the number of nanoparticles that can be analyzed through these imaging techniques is limited. Therefore, to analyze the size distribution of particles, Dynamic light scattering (DLS) is predominantly used [39, 40]. For mass, it can be calculated using the known size and density of the silicon particles. Assuming a spherical shape, the size and density measured through Small Angle X-ray Scattering (SAXS) or Differential Centrifugal Sedimentation (DCS) can also be used for calculation [41]. Additionally, single nanoparticle mass can be measured using capillary resonators or solid microchannel resonators [42, 43].

Notably, silicon nanoparticles produced via reduction may not be entirely reduced to silicon, and impurities might be present. The elemental composition of the silicon nanoparticle clusters can be quantified using X-ray Photoelectron Spectroscopy (XPS) spectra (Fig. 3a) [32]. The crystallinity of silicon nanoparticles is another essential metric. Amorphous silicon benefits battery anodes, while crystalline silicon is preferable for optical applications [44, 45]. This can be verified through High-Resolution TEM (HRTEM) imaging, X-ray Diffraction (XRD), and Raman spectroscopy (Fig. 3a–c) [32,33,34]. XRD not only provides information about crystallinity but also shows that as the diameter of the silicon nanoparticles decreases, the peaks become broader, indicating a change in particle size. Also, Raman spectroscopy not only reveals crystallinity but is also influenced by particle size. As the particle size decreases, a redshift and the broadening of the Full Width at Half Maximum (FWHM) occurs. This phenomenon is due to the size effect of silicon nanoparticles where phonon vibrations are trapped in nanoparticles [46].

Silicon nanoparticles exhibit Mie resonance of electric and magnetic dipoles in the visible light spectrum [47, 48]. To analyze this, single silicon nanoparticles are positioned on a glass substrate, and a dark field image captures the Mie scattering characteristics (Fig. 3d). This property is prominent in silicon nanoparticles with diameters between 100 and 200 nm. Various silicon nanoparticle diameters enable Mie resonators to tune vibrant broad-spectrum colors in a dark background to RGB [4]. Colloids of dispersed silicon nanoparticles can be analyzed using Ultraviolet-Visible (UV–Vis) spectroscopy (Fig. 3e) [35, 49]. The optical properties of silicon nanoparticles, sensitive to factors like size, shape, concentration, aggregation state, and refractive index, can be characterized using UV-Vis, and the colloidal stability can be analyzed. Fourier-Transform Infrared Spectroscopy (FTIR) offers insights into bond strength, characteristics, specific functional group band positions, and molecular structure interactions (Fig. 3f) [36]. When the silicon nanoparticles are reduced to a few nm, the quantum confinement-induced photoluminescence energy can be extensively adjusted (Fig. 3g) [50]. High photoluminescence quantum yields, up to 60%, have been reported for silicon nanoparticle colloids [51]. Efforts to expand the tunability of photoluminescence by adjusting nanoparticle size are ongoing. A blue shift in the PL peak is evident as nanoparticle size decreases due to enhanced quantum confinement [37]. Small Angle Neutron Scattering (SANS) measurements not only offer general information about particle size but, through scattering data analysis, allow estimation of the internal surface area per volume unit of the material (Fig. 3h) [38]. This is especially beneficial for materials with chemically altered surfaces, and it can be utilized for characterizing porous structures.

Fig. 4
figure 4

a Lithiation process of silicon nanoparticles. Cracks do not occur below 150 nm in diameter. Reprinted from [2] with permission from American Chemical Society. b (left) Photographs of solutions of size-separated silicon nanoparticles. (right) Silicon nanoparticle-PVP film on PET substrate bent at bending radius of 9 mm. Reprinted from [52] (CC BY 4.0). c (left) The schematic of a structure composed of identical silicon nanoparticles and an ARC on the top. (right) A simple planar structure with silicon thickness identical to left. Reprinted from [53] (CC BY 4.0). d (left) Photograph of SiLED. (right) Schematic view of SiLED stack. Reprinted from [54] with permission from American Chemical Society. e (left) Schematic of a biological synapse. (right) Schematic of a synaptic silicon nanoparticle phototransistor. Reprinted from [55] with permission from Elsevier. f Schematic of the synthesis method of the silicon nanoparticles and their use for rutin detection. Reprinted from [56] (CC BY 4.0)

Full size image

Application

The primary applications of silicon nanoparticles are in energy, specifically as anode materials for lithium-ion batteries, and in optics as Mie resonators. Silicon, as previously mentioned, has an energy capacity that’s over 10 times that of carbon, making it a promising candidate for next-generation anode materials. One of the challenges with silicon is the cracking that occurs during the lithiation and delithiation processes, which can impair performance. However, using silicon nanoparticles smaller than 150 nm can mitigate these cracking issues (Fig. 4a) [2]. Recently, there’s been an interest in combining silicon nanoparticles with amorphous carbon, graphene, and carbon nanotubes [57]. Through this approach, numerous electrical contact points are provided, and the pathway for electron transfer is shortened, enhancing the anode’s electrical conductivity. This mitigates electrical insulation and improves electrode dynamics during cycles. In terms of optical applications, silicon nanoparticles exhibit structural colors due to Mie resonance. Structural colors can also be realized based on the surface plasmon resonance of nanostructures [58]. However, to achieve high color purity over a wide color spectrum, complex structures such as multilayer and asymmetric metal-dielectric nanostructures are required. Silicon nanoparticles offers an angle-insensitive structural color spanning the entire visible spectrum with simple structure, making it a viable candidate for color ink (Fig. 4b) [52]. In order to dye the substrate with a color identical to that of a solution dispersed with silicon nanoparticles, it is necessary to keep the inter-particle distance over 50 nm while solidifying the ink. For this purpose, a methanol solution with dispersed silicon nanoparticles was first mixed with a polyvinylpyrrolidone (PVP) solution to prevent particle aggregation, then the mixture was drop-casted onto the substrate. This allows for painting the substrate with color ink. Furthermore, such Mie resonance, when coupled with a metallic substrate, can realize new resonant modes, enabling the fabrication of nano-antennas [59]. The nano-antenna was fabricated by depositing a monolayer of quantum dots onto the substrate through drop casting, followed by drying in air at room temperature. Finally, a solution with dispersed silicon particles was dropped onto the silicon quantum dot monolayer and dried.

Another energy application involves the fabrication of photovoltaic devices. Ultra-thin silicon solar cells often suffer from low efficiency due to their insufficient thickness for light absorption. By arranging silicon nanoparticles in multilayers, they can be utilized as light-capturing absorbers for ultra-thin solar cells while simultaneously forming a p-n junction (Fig. 4c) [53]. Devices with such a structure can achieve a short circuit current density of up to 26 mA/cm\(^2\), which is promising for monocrystalline silicon ultra-thin solar cells, considering that devices of similar thickness exhibit values below 15 mA/cm\(^2\) [60]. Conversely, silicon nanoparticles can be utilized to fabricate light-emitting diodes (LEDs) (Fig. 4d) [54]. Quantum dots made by reducing the size of silicon nanoparticles to below 5 nm can be used to produce LEDs, with tunability ranging from red to yellow-orange spectrum regions. The LED was fabricated by spin-coating various functional layers onto a glass substrate covered with an ITO structure produced through lithography, followed by a thermal evaporation step for the cathode. These LEDs offer the advantage of a high external quantum efficiency and low turn-on voltages.

Similar to fabricating transistors from silicon, silicon nanoparticles can also be utilized to fabricate transistors. Recent advancements have reported the development of synaptic optical transistors, which can be stimulated by optical and electrical spikes utilizing the optical absorption of silicon nanoparticles (Fig. 4e) [55]. For the fabrication of a synaptic device, a solution with dispersed silicon nanoparticles is drop-casted between electrodes spaced at a specific distance apart, and subsequently baked at 150 °C for 30 min to form a film. These synaptic silicon phototransistors could potentially offer low energy consumption, leading to more efficient devices. Moreover, water-soluble silicon nanopaticles are emerging as compact fluorescent silicon nanomaterials (Fig. 4f) [56]. Compared to conventional fluorescent quantum dots, water-soluble silicon nanoparticles have a simpler composition and are devoid of heavy metal elements. Fluorescent probes can be fabricated by adding a solution dispersed with silicon nanoparticles to PBS solution, mixing evenly, and then adding rutin solutions of various volumes to prepare a series of solutions. Silicon nanoparticles can be metabolized in the body into orthosilicic acid and subsequently excreted through urine, ensuring they’re non-toxic. Crucially, due to their outstanding optical properties, robust stability, and excellent biocompatibility, silicon nanoparticles are well-suited as fluorescent probes.

Perspectives

The future of silicon nanoparticles is heavily contingent upon the mass-production capability with high quality and low cost. With advancements in synthesis techniques allowing for large-scale production, consistent quality control remains challenges. In addition, uniformity in particle size, shape, and properties is essential for both fundamental research and commercial applications. It is evident that a production process that enables aforementioned uniformity efficiently will become a gold standard. Moreover, cost reduction is imperative. As research advances and more innovative production techniques become available, the production cost will certainly drop, making it easier to access silicon nanoparticles across various industrial sectors.

The adoption of in-situ measurement techniques, where silicon nanoparticles are characterized in real-time during synthesis, provides evidences that were previously unattainable. These techniques provide instant feedback, facilitating adjustments in the production process to obtain desired properties. In addition, the push towards high-speed and high-throughput metrologies ensures that vast quantities of silicon nanoparticles can be characterized quickly, facilitating quality control in mass production scenarios.

Conclusion

In this review, we aim to overview silicon nanoparticle technology. We focuses on fabrication methods of silicon nanoparticles, specifically categorizing top-down, bottom-up and reduction approaches. We also look into various characterization techniques crucial for fabricated silicon nanoparticles. While silicon nanoparticles have been widely utilized in energy and optics applications due to their inherent benefits, there still remains room for further improvement, particularly in fabrication and characterization. Advancements in silicon nanoparticles will undoubtedly pave the way for development of new applications.