Blue-Shifting Photoluminescence in HFCVD-Deposited Tin-Doped SRO Films

Silicon-rich oxide (SRO) films doped with tin (Sn-SRO) were successfully deposited using hot filament chemical vapor deposition (HFCVD), with tin-doped SBA-15 as the solid source material. SBA-15 acted as a protective layer, ensuring thermal stability and enabling the controlled incorporation of tin into the SRO films. Structural and optical analyses, including Fourier- transform infrared spectroscopy (FTIR), transmission electron microscopy (TEM), and photoluminescence (PL), were performed to evaluate the films. FTIR revealed SiO2 absorption peaks and Si-O-Sn bond formation, confirming the progressive incorporation of tin. TEM demonstrated the formation of silicon nanocrystals (Si-NCs), with decreasing size as tin concentration increased. This size reduction was correlated with enhanced PL intensity and a blue shift, attributed to stronger quantum confinement effects. These analyses provide a comprehensive understanding of the structural and optical behavior of Sn-doped SRO films. These findings highlight the potential of Sn-SRO films for optoelectronic applications, where controlling light emission and crystallinity is essential.

Introduction

Silicon-based systems, particularly those incorporating silicon nanocrystals (Si-NCs) within dielectric matrices such as silicon dioxide (SiO₂), silicon nitride (Si₃N₄), and silicon carbide (SiC) [1]–[3], have garnered significant attention due to their potential applications in optoelectronic and photovoltaic technologies. At the nanometer scale, quantum confinement effects emerge, significantly altering the electronic and optical properties of materials. While bulk silicon remains crucial for the semiconductor industry, its indirect bandgap limits its effectiveness in optoelectronic devices. However, the discovery of room-temperature photoluminescence in porous silicon (Si-P) by Canham [4] demonstrated how silicon’s properties can dramatically change at the nanoscale, paving the way for new silicon-based light emission technologies.

The growing interest in developing efficient, cost-effective optoelectronic devices, such as modulated lasers and light-emitting diodes (LEDs), has accelerated research into nanomaterials due to their unique electronic and optical behaviors. Among these materials, silicon-rich oxide (SRO) has emerged as a promising candidate for constructing nanostructures. Its inherent properties, including strong light emission, optical gain, nonlinear optical responses, high thermal and chemical stability, and compatibility with existing silicon microelectronics, make it an ideal material for both research and industrial applications in semiconductors [5]. The robustness of SRO and its compatibility with silicon-based systems provide a practical platform for integrating nanomaterials into current semiconductor technologies.

Several deposition techniques have been investigated for fabricating SRO films, including Hot Filament Chemical Vapor Deposition (HFCVD) [6]. HFCVD offers the advantage of low substrate temperatures and rapid deposition rates, promoting the formation of films containing embedded Si-NCs [7], [8]. The inclusion of Si-NCs within dielectric matrices like SRO is known to enhance optical properties, making them suitable for advanced material applications. However, conventional solid sources such as porous silicon (Si-P) and quartz, commonly used in SRO deposition, are often expensive, motivating the exploration of alternative sources like SBA-15 [9]–[11].

SBA-15, a mesoporous silica material, is characterized by its high surface area, large pore size, and notable thermal stability, which enable the controlled incorporation of metal dopants. In this study, SBA-15 doped with tin (Sn) was used as a solid source to deposit Sn-doped SRO films. Previous research has demonstrated that Sn doping in various materials can improve crystallinity and modulate the growth of nanocrystals [12], often reducing their size. These findings suggest that similar effects may occur in SRO films when doped with Sn, potentially influencing both their structural and optical characteristics.

Given that SRO’s properties are known to change significantly with the introduction of dopants [13], Sn incorporation may lead to enhanced crystallinity and a reduction in the size of silicon nanocrystals within the SRO matrix. This study will investigate whether Sn doping in SRO yields these effects, providing new insights into how structural modifications influence material behavior.

Additionally, Sn has been shown to induce quantum confinement effects in other doped materials, often linked to enhanced photoluminescence (PL). These effects arise as electron and hole movement becomes restricted within smaller nanocrystals. Therefore, it is expected that Sn-doped SRO films may exhibit similar quantum effects, resulting in measurable changes in PL behavior. Observing such shifts would provide valuable information about the potential optical performance of these materials.

Using SBA-15 as a solid source in the HFCVD process offers the advantage of controlled Sn incorporation into the SRO films. Additionally, SBA-15 serves as a protective layer, preventing unwanted reactions during the high-temperature deposition process, which typically reaches 2000°C in HFCVD. This approach is expected to yield Sn-doped SRO films with tailored structural properties, making them suitable for further exploration in advanced material studies.

Experimental Procedure

Silicon-rich oxide (SRO) films were deposited using a vertical hot filament chemical vapor deposition (HFCVD) reactor (Fig. 1). The preparation of the solid silicon source followed a careful and precise method to obtain high-purity SRO films. First, 100 mg of pre-synthesized and analyzed Sn-SBA-15 powder, which was confirmed to contain Si, O, and Sn atoms [14], was finely ground and compressed using a hydraulic press, forming disks with a diameter of 10 mm and a thickness of 2 mm. This step ensured a homogeneous and uniform solid source for the deposition process.

Fig. 1. Schematic diagram of the experimental setup for the deposition of doped and undoped SRO films.

The deposition procedure was conducted with a carefully optimized setup. A p-type silicon (111) substrate, with a resistivity of 2 Ω·cm − 3 Ω·cm, was placed in the reactor chamber alongside the solid source. During deposition, the tungsten filament was heated to 2000°C, and a steady flow of 100 sccm of hydrogen (H₂) was introduced into the reactor. The substrate temperature was maintained at 800°C, and the deposition was carried out for 10 minutes. The distances between the filament and the solid source, as well as between the solid source and the substrate, were kept at 3 mm to ensure optimal film growth. The system pressure was held constant at 1 atmosphere throughout the process, ensuring stable and controlled conditions (Fig. 1).

In this HFCVD process, molecular hydrogen serves as the precursor gas, entering the reactor through a designated inlet. Upon contact with the tungsten filament, the hydrogen undergoes thermal and catalytic decomposition, yielding atomic hydrogen. This atomic hydrogen etches the solid source material (SBA-15 and Sn-SBA-15), generating radicals that are transported to the substrate for film formation through the CVD process.

The deposited SRO films were designated as Sn-SRO-X, where X corresponds to the nominal Si/Sn molar ratio in the source material (Sn-SBA-15). This labeling allows for direct comparison of films based on their Sn concentration.

Results and Discussion

TEM Characterization of SRO Films Synthesized from SBA-15 and Sn-SBA-15 Precursors

Figs. 2a–c show typical TEM images of undoped SRO films. The lattice fringes, dispersed within the amorphous SiO₂ matrix, indicate the formation of nanocrystals (NCs), with some highlighted by yellow circles.

Fig. 2. TEM analysis of undoped SRO films, highlighting nanocrystal formation and structural properties. (a), (b), and (c) TEM images of undoped SRO films, with Si-NCs highlighted using yellow circles. (d) Inverse FFT of the area enclosed by the red box in (b), showing lattice structure. (e) FFT diffractogram corresponding to the area marked by the box in (b), with the inset displaying the size distribution histogram of the Si-NCs.

The TEM images reveal that the Si-NCs exhibit a quasi-spherical morphology, with a few showing elongated structures. The size distribution histogram of the Si-NCs, shown in the inset of Fig. 2c, was obtained by examining 25 NCs, revealing an average size of 1.9 nm ± 0.63 nm. Using the inverse Fast Fourier Transform (FFT) of the crystalline area marked by a red box in Fig. 2b, the lattice spacing, d, between two planes was estimated, as shown in Fig. 2d. This spacing, present in most Si-NCs, is approximately 0.2 nm.

Figs. 3a and b present high-resolution TEM analysis of Sn-SRO films with a nominal Si/Sn molar ratio of 10. These images show that at lower tin concentrations, the size of the Si nanocrystals (Si-NCs) increased to approximately 8 nm, as observed in the Sn-SRO (10) sample. However, as the Sn concentration increases further, the nanocrystal size decreases. Interplanar distances of 0.16 nm and 0.31 nm were observed for the Si-NCs, corresponding to the (311) and (111) planes of crystalline Si (JCPDS #00-027-1402), respectively, as confirmed by the Fast Fourier Transform (FFT) diffraction patterns of the selected regions in the blue boxes (inset of Fig. 3c). The FFT analysis of the area selected with a red box in Fig. 3b (lower inset) revealed the presence of interplanar distances associated with the (101) plane of tin oxide (SnO₂) (JCPDS #88-0287). Interestingly, overlapping NCs of Si and SnO₂ were observed in several regions of the films, as shown in the FFT diffractogram of Fig. 3d. The (110) plane of SnO₂ and the (311) and (111) planes of Si are clearly visible. The interplanar distances measured from the inverse FFT image are presented in Fig. 3c.

Fig. 3. TEM analysis of Sn-SRO-(10) films demonstrating the structural and crystalline characteristics. (a) and (b) TEM images showing the morphology and nanocrystal distribution. (c) Inverse FFT of the area enclosed by the box in (a), highlighting the lattice structure. (d) FFT diffractogram corresponding to the area marked by the box in (a), with insets showing the FFTs of the selected areas.

The thickness of the films was confirmed through cross-sectional TEM images, showing that the undoped SRO films exhibited a thickness of approximately 70 nm, while the Sn-SRO (10) films had a thickness of around 300 nm. These thicknesses fall within the typical range for films deposited using HFCVD, where both deposition time and hydrogen flow play crucial roles in determining the final film thickness. Furthermore, the increase in the size of silicon nanocrystals (Si-NCs) has been linked to thicker films, as larger nanocrystals promote greater film growth by reducing the number of nucleation sites, leading to an increased layer thickness. Under controlled conditions, HFCVD consistently produces films within this thickness range, which aligns with the results obtained in this study [15], [16]. The thicker films, such as Sn-SRO (10) and Sn-SRO (20), are associated with lower concentrations of Sn, which induces fewer nucleation sites for the formation of Si-NCs [17]. These larger nanocrystals facilitate greater film growth, resulting in thicker layers. In contrast, the thinner films, such as undoped SRO and Sn-SRO (30), are associated with higher Sn concentrations, which lead to a greater number of nucleation sites, producing smaller nanocrystals that limit overall film growth [18].

This correlation between nanocrystal size and film thickness suggests that the introduction of Sn alters both the nucleation process and the crystallinity of the SRO films. As silicon atoms precipitate during nucleation, they tend to form Si-NCs. However, as the available excess silicon becomes insufficient, nucleation sites decrease, and the NC density reaches a peak. The Si-NCs continue to grow until they enter a maturation stage, where larger clusters form at the expense of smaller NCs, leading to a decrease in NC density.

The presence of overlapping Si and SnO₂ nanocrystals in several regions, as observed in the FFT diffractograms, further supports the conclusion that Sn enhances the crystallinity of the films, especially in the thinner samples, where nanocrystal growth is more constrained [19]–[21].

Optical Characterization by Fourier Transform Infrared Spectroscopy (FTIR) of SRO and Sn-SRO-(X)

Fig. 4 presents the FTIR spectra of undoped and Sn-SRO(X) films. The spectra display characteristic absorption peaks of SiO₂, with bands centered at 425, 810, 1074, and 1175 cm⁻¹ (shoulder), corresponding to the stretching vibrations of siloxane groups. The band at 883 cm⁻¹ is attributed to the vibrational modes of Si-OH and Si-H species. Additionally, Si-Si bonds are observed at 600 cm⁻¹, along with the presence of CO₂ at approximately 2257 cm⁻¹. The bands around 667 and 530 cm⁻¹ correspond to the stretching modes of Sn-O and Sn-OH, respectively [22]–[24]. The band at 530 cm⁻¹ is more pronounced in the Sn-SRO (10) sample compared to other Sn-SRO films.

Fig. 4. FTIR Absorbance Spectra: This figure shows the FTIR absorbance spectra of undoped and tin-doped SRO films, with the nominal Si/Sn molar ratio varying from 10 to 30. The spectra illustrate the changes in absorption peaks due to tin incorporation.

The intensity of the 883 cm⁻¹ band tends to decrease as the Sn content in the SRO films increases. This could result from gaseous tin oxide and tin hydroxide species reacting with silanol groups and silicon hydride (Si-H) bonds on the surface of the amorphous SiO₂ film and the Si nanocrystals (Si-NCs), leading to the formation of Si-O-Sn bonds. This interaction may also explain the appearance of the band at 667 cm⁻¹ [25].

It is important to note that the FTIR spectra of the thinner films, such as undoped SRO and Sn-SRO (30), required magnification of X1.5 due to their lower intensity, consistent with their smaller thicknesses. These films, with thicknesses around 70 nm as confirmed by TEM images, exhibit weaker absorption bands, necessitating amplification to clearly observe the characteristic peaks. This observation aligns with previous discussions on the correlation between film thickness and nanocrystal size, where thinner films exhibit smaller Si-NCs, limiting overall film growth.

It is important to note that the FTIR spectra of the thinner films, such as undoped SRO and Sn-SRO (30), required magnification of X1.5 due to their lower intensity, consistent with their smaller thicknesses. These films, with thicknesses around 70 nm as confirmed by TEM images, exhibit weaker absorption bands, necessitating amplification to clearly observe the characteristic peaks. This observation aligns with previous discussions on the correlation between film thickness and nanocrystal size, where thinner films exhibit smaller Si-NCs, limiting overall film growth.

TEM studies from the previous section support that tin oxide species deposited on the silicon film form SnO₂ crystals, as confirmed by the FFT diffractogram of the region indicated by the yellow box in the TEM image of Fig. 3b. Additionally, tin oxide species deposited on the surface of the Si-NCs result in overlapping crystals, as shown in Fig. 3d.

The nature of the shoulder at 1175 cm⁻¹ has been debated for many years. This peak intensifies when the refractive index of the film is low, indicating lower film density and increased structural disorder in the Si-O-Si network [26]. We have associated this peak with the possible presence of less rigid bonds in the Si-O-Si network and the formation of Si-Si-O₃ bonds (Si-O-Si linked to Si) [26]. The structural disorder observed in thinner films, such as those of undoped SRO and Sn-SRO (30), further supports this interpretation, as their lower density and larger surface area contribute to the intensification of this peak [27].

Photoluminescence Properties of SRO and Sn-SRO-(X) Films

The photoluminescence (PL) spectra were recorded with an excitation wavelength of λ = 250 nm.

Fig. 5 presents the PL spectra of SRO and Sn-SRO-(X) films, where X represents the tin concentration. All samples exhibit broad PL emissions in the visible range (400 nm to 700 nm), with peak emissions observed at 460 nm, 450 nm, 490 nm, and 505 nm for SRO, Sn-SRO-(30), Sn-SRO-(20), and Sn-SRO-(10) films, respectively. As shown in Fig. 5, the PL intensity increases with higher Sn cncentrations, accompanied by a notable blue shift as the Sn content increases [28].

Fig. 5. PL spectra of SRO and Sn-SRO films.

At lower Sn concentrations, the PL spectra shift towards longer wavelengths, typically attributed to an increase in the size of silicon nanocrystals (Si-NCs). This shift towards the infrared region is caused by larger nanocrystals having smaller bandgaps, which results in photon emission at lower energies. The growth of larger nanocrystals can be explained by a reduced number of nucleation sites and fewer defect densities, leading to a more extended crystal lattice [29], [30]. As the Si-NC size increases, the quantum confinement effects diminish, resulting in lower PL intensity and a shift towards longer wavelengths.

In contrast, as the Sn concentration increases, the average size of the Si-NCs decreases, causing a shift towards shorter wavelengths (blue shift) in the PL spectra. This behavior is attributed to the increased number of nucleation sites induced by Sn, which accelerates the formation of smaller Si-NCs by limiting their overall growth. Smaller nanocrystals exhibit stronger quantum confinement effects, which increase the energy gap between the valence and conduction bands, thereby shifting the emission to higher energies (shorter wavelengths) [31]. The presence of quantum confinement in smaller Si-NCs restricts the movement of charge carriers, causing the material to behave similarly to quantum dots (QDs), where the size of the nanostructures directly influences their optical properties [32]–[34].

Furthermore, the increase in PL intensity at higher Sn concentrations may also be related to the formation of Sn-related defects or oxides, as observed in the FTIR spectra (Fig. 4) and TEM images (Fig. 3b). Tin oxide is known to emit in the ultraviolet range, suggesting that interactions between Sn oxide species and Si-NCs may introduce new radiative states, further enhancing the PL intensity. These interactions may increase the number of Sn-O bonds on the surface of the Si-NCs, altering their optical properties and potentially introducing additional radiative recombination centers [35].

Although the PL intensity at higher Sn concentrations does not reach the levels observed in undoped SRO, this does not imply that the films are of poor quality. Rather, the reduction in PL intensity may be attributed to the introduction of Sn-related defects, which introduce non-radiative recombination centers. Despite this, the combination of quantum confinement effects and Sn-related defects in these films indicates strong potential for optoelectronic applications, where precise control of optical responses and defect management is critical.

The increased crystallinity induced by Sn doping suggests that the Si nanocrystals are formed in a more controlled manner, with fewer defects in the thinner films. This improvement in crystallinity has been observed in other studies where Sn doping enhances crystallinity, supporting its role in improving material structure [35].

Conclusion

Tin-doped silicon-rich oxide (Sn-SRO) films were successfully deposited using the Hot Filament Chemical Vapor Deposition (HFCVD) technique. Transmission electron microscopy (TEM) analysis revealed the formation of quasi-spherical silicon nanocrystals (Si-NCs). In Sn-SRO samples with low Sn concentrations (Si/Sn molar ratio = 10), an increase in the size of the Si-NCs was observed, reaching approximately 8 nm, compared to the 1.9 nm ± 0.63 nm observed in undoped films. However, as the Sn concentration increased, the nanocrystal size decreased due to a higher number of nucleation sites, resulting in thinner films with a greater number of smaller nanocrystals.

FTIR spectra confirmed the successful incorporation of Sn into the SRO matrix, evidenced by the presence of Sn-O bonds. The FTIR spectra of the thinner films, such as Sn-SRO (30), showed lower intensity due to their reduced thickness, consistent with the TEM results.

Photoluminescence (PL) studies demonstrated an increase in PL intensity with higher Sn concentrations, accompanied by a blue shift in the emission spectrum, attributed to quantum confinement effects, which intensify as the nanocrystal size decreases. However, at higher Sn concentrations, dopant-related defects introduce non-radiative recombination centers, limiting overall PL efficiency, particularly in thicker films.

While complementary characterizations could further enrich the study, the TEM, FTIR, and PL results effectively address the primary objectives of this work. These findings provide critical insights into the structural and optical behavior of Sn-doped SRO films, underscoring their potential for advanced optoelectronic applications, where precise control over light emission and crystallinity is essential.

Finally, the ability to control both nanocrystal size and quantum confinement effects through Sn doping allows for precise tuning of the optical properties. Nevertheless, at higher Sn concentrations, the reduction in nanocrystal size and the introduction of defects limit radiative efficiency, indicating the need for a careful balance between dopant concentration and defect control to optimize the optoelectronic properties of the films.