Development of n-Type, Passivating Nanocrystalline Silicon Oxide Films via Plasma-Enhanced Chemical Vapor Deposition

Abstract

Nanocrystalline silicon oxide (nc-SiOx:H) is a multipurpose material with varied applications in solar cells as a transparent front contact, intermediate reflector, back reflector layer, and even tunnel layer for passivating contacts, owing to the easy tailoring of its optical properties. In this work, we systematically investigate the influence of the gas mixture (SiH₄, CO₂, PH₃, and H₂), RF power, and process pressure on the optical, structural, and passivation properties of thin n-type nc-SiOx:H films prepared in an industrial, high-throughput, plasma-enhanced chemical vapor deposition (PECVD) reactor. We provide a detailed description of the n-type nc-SiOx:H material development using various structural and optical characterization techniques (scanning electron microscopy (SEM), energy dispersive X-ray (EDX), Raman spectroscopy, and spectroscopic ellipsometry) with a focus on the relationship between the material properties and the passivation they provide to n-type c-Si wafers characterized by their effective carrier lifetime (τ_eff). Furthermore, we also outline the parameters to be kept in mind while developing different n-type nc-SiOx:H layers for different solar cell applications. We report a tunable optical gap (1.8–2.3 eV) for our n-type nc-SiOx:H films as well as excellent passivation properties with a τ_eff of up to 4.1 ms (implied open-circuit voltage (iV_oc)~715 mV) before annealing. Oxygen content plays an important role in determining the crystallinity and hence passivation quality of the deposited nanocrystalline silicon oxide films.

1. Introduction

In the past decade, nanocrystalline silicon oxide (nc-SiOx:H) films, as a functional multipurpose material, have garnered significant interest of the photovoltaic community [1,2,3,4]. This is due to its broad range of optical and electrical properties, which can be tuned by varying deposition parameters: i.e., absorption coefficient, refractive index, optical band gap, and p-type and n-type doping, which allow it to make selective contacts. Doping films with phosphorus (n-type) has resulted in further improved conductivities, and hydrogen atoms in the mixture have resulted in effective defect passivation [5]. Nanocrystalline silicon oxide (nc-SiOx:H) consists of crystalline silicon nanoparticles embedded in an amorphous phase composed of silicon oxide (a-SiOx:H) and a-Si:H phases [6,7]. The two-phase structure of nc-SiOx:H materials has been suggested to help achieve a large optical gap while keeping a high conductivity. The oxygen-rich Si-O-Si phase of the material helps to increase the optical gap and decrease the refractive index, while the Si-rich portion of the material, via an intricate three-dimensional network of crystallites, is known to provide high electrical conductivity pathways both in lateral and transverse directions [8].

The research on nc-SiOx:H layers was first developed for single-junction μc-Si:H solar cells. It provided better light trapping and acted as a good window layer [4,9,10,11]. By using a suitable bilayer of two crystalline SiOx:H layers (as n-layer as well as back reflector), solar cell efficiencies improved in p-i-n structure-based single-junction solar cells [12]. Later on, the material was applied to silicon heterojunction (HJT) solar cells on both textured and flat substrates, where the nanocrystalline silicon oxide layer (nc-SiOx:H) has been used as a wide-gap window layer instead of an a-Si:H layer [13,14]. Compared with a-Si:H layers, n-type nc-SiOx:H offers higher transparency, similar electrical conductivity, and a lower refractive index. Mazzarella et al. also studied the use of ~5 nm thick n-type nc-SiOx:H as front contacts (exhibiting high front surface field) in rear-emitter SHJ solar cells, resulting in efficiencies of 22.6% [15]. The use of nc-SiOx as an electron transport layer has subsequently improved the power conversion efficiency to 23.1% (V_oc = 738 mV, J_sc to 39 mA cm⁻²) [16].

Doped nc-SiOx:H layers also have dual application in tandem thin films solar cells: (i) as an intermediate reflector between the n/p recombination junction (i.e., a-Si:H and μc-Si:H) due to better refractive index matching and (ii) as an n-layer of the μc-Si:H bottom cell to reduce parasitic absorption [6,10,17]. The application of n-type and p-type nc-SiOx:H as an FSF layer and rear-emitter layer resulted in a textured rear-emitter SHJ solar cell with an efficiency of 21.4% [8]. Even poly-Si layers in TOPCon solar cells have been replaced by ultrathin nc-SiOx:H to enhance transparency [18,19]. Further, nc-SiOx layers have demonstrated clear advantages for maximizing the infrared response of c-Si bottom cells in combination with perovskite top cells [20]. Nanocrystalline silicon oxide (nc-SiOx:H) layers, in this case, not only offer better carrier transport and field effect passivation but also improve carrier selectivity and reduce parasitic absorption [21]. Depending on the application, these layers fulfil various requirements, such as (i) a doped layer to create the electric field in the diode, (ii) a nucleation layer to promote crystalline growth, or (iii) a suitable recombination junction with low series resistance. Therefore, in this work, we discuss the versatility of nanocrystalline silicon oxide (nc-SiOx:H) material deposited by an industrial, high-throughput plasma-enhanced chemical vapor deposition (PECVD) tool for different photovoltaic applications: front contacts, back reflectors, and intermediate window layers. Parameters like concentration of dilution gases, doping gases, RF power, and total pressure are examined in a detailed manner as illustrated in Figure 1a. During the PECVD deposition of nc-SiOx:H, silane (SiH₄) provides silicon, carbon dioxide (CO₂) supplies oxygen, phosphine (PH₃) introduces phosphorus for doping, and hydrogen (H₂) serves various functions such as influencing film properties and reducing defects. The combination of these precursor gases, along with the plasma environment, allows for the controlled growth of nc-SiOx:H with a tailored composition and properties for specific applications in solar cells.

Through this work, we explore and try to understand the relationship between the optical and structural properties of n-type nc-SiOx:H films and their ability to passivate the surface of crystalline silicon as systematically illustrated in Figure 1b. It should be noted that passivation quality of nc-SiOx:H films is a critical factor influencing the performance and efficiency of solar cells. For n-type c-Si wafers, passivation is particularly crucial as it helps to minimize the surface recombination of minority carriers, such as holes, and nc-SiOx:H films act as passivation layers for the silicon surface. The material properties, including the crystallinity, hydrogen content, and oxygen content, significantly influence the ability of these films to passivate the surface effectively. The presence of hydrogen is essential for passivation, as it helps in the formation of silicon–hydrogen bonds that can passivate dangling bonds and trap charge carriers. By understanding and controlling the deposition parameters, it is possible to tailor the passivation layer to suit the specific requirements of n-type c-Si wafers in solar cell applications. This paper distinguishes itself from previous studies by placing a pronounced emphasis on passivation properties within the realm of nanocrystalline silicon oxide (nc-SiOx:H) films. While prior research has predominantly examined these films through the lens of optical and structural properties, our work takes a pioneering approach by elevating the significance of passivation. The primary focus revolves around the intentional development of nc-SiOx:H films endowed with superior passivating properties, aiming to contribute innovative insights to the field of solar cell technology.

2. Materials and Methods

N-type nc-SiOx:H layers were deposited in an industrial-grade, plasma-enhanced chemical vapor deposition system, PECVD, (IndeoTec Octopus) on 1/4 of 4-inch, n-type FZ c-Si wafers with a resistivity of 1–3 Ω.cm and Corning glass. N-type FZ c-Si wafers were dipped in 5% HF solution for 30 s to remove the native oxide before introducing them in the PECVD chamber. A mixture of silane (SiH₄), dihydrogen (H₂), and carbon dioxide (CO₂) was used as precursor for the deposition as shown in Figure 2. For n-type doping, 1% phosphorus (PH₃) diluted in hydrogen was utilized. Typical deposition parameters were a pressure of 2.5 Torr, a substrate temperature of 180 °C, a deposition time of 10 min, and a radio frequency power of 50 W. The flow rates for gases were as follows: SiH₄, 2–8 sccm; CO₂, 0–4 sccm; and PH₃, 0.5–2 sccm. For all samples discussed in this manuscript, H₂ flow rate was fixed at 500 sccm. Furthermore, the radio frequency (RF) plasma power and pressure were varied within 20 W–80 W and 1.4 Torr–3 Torr, respectively. It should be noted that in all cases, a thin layer of a-Si:H was deposited on a c-Si wafer at 25 W, 1.6 Torr for 30 sec in order to provide a better passivation of the c-Si before deposition of the nanocrystalline silicon oxide layers. The deposition time was fixed at 10 min. Note that we target thin films (in the range of 10 nm–45 nm) relevant for passivation layers in solar cells. Thicker layers could be better crystallized but are not useful because absorption would be too high.

In this work, minority carrier lifetime (τ) is used as a measure of passivation quality because it provides information about the recombination of minority carriers’ (photocarrier) lifetime at the surface of a semiconductor material. It is composed of the bulk lifetime and surface lifetime, which depend on the surface properties of the sample. In low-level injected material, where the number of minority carriers is less than the doping level, the lifetime is determined by the recombination rate (R), as expressed in Equation (1):

$$\tau =\frac{∆n}{R}$$

(1)

where τ is the minority carrier lifetime, and $\Delta n$ is the excess minority carrier concentration [22]. As per Equation (1), longer τ indicates better passivation, suggesting lower recombination rates at the surface. This means there is a higher probability of carriers reaching the electrical contacts before recombining, which contributes to an enhanced charge collection efficiency for solar cells. It should be noted that the terms ‘long lifetime’ and ‘high lifetime’ are used interchangeably when we discuss this. The effective minority carrier lifetime (τ_eff) of the symmetrical samples was determined using the photoconductance decay (PCD) method using Sinton Lifetime Tester. The as-deposited values extracted at a minority carrier injection level of 10¹⁵ cm⁻³ are reported. The PCD measurements were either performed in the transient mode for samples with τ_eff above ~100 μs and in the quasi-steady-state mode using a generalized analysis for samples with low τ_eff < 100 μs [23,24]. Similarly, 1-sun implied open-circuit voltage (iV_oc) was also extracted. Thermal annealing was avoided in this study so that an additional step was evaded for industrial compatibility. Subsequently, the homogeneity of the passivation quality was also determined by photoluminescence (PL) images obtained using a BTi luminescence imaging system at 1 sun. Optical properties of the films were measured via spectroscopic ellipsometry (SE) in the spectral range [1.5 eV–5 eV] using a phase-modulated ellipsometer (UVISEL from HORIBA John-Yvon). We measured films on both c-Si and glass substrates. The ellipsometry data were modeled using Bruggeman effective medium approximation (BEMA) using the DeltaPsi2 software on c-Si substrates. The optical gap of the film, denoted by E₀₄, was estimated as the photon energy at which the absorption coefficient was 10⁴ cm⁻¹. We used the E₀₄ value to provide an estimation of the optical gap because it is simpler and less error-prone than the Tauc method. However, if we carefully estimate the Tauc gap of the film, we can easily find a linear relationship between E₀₄ and the Tauc gap, as used in [25,26]. Raman spectroscopy was used to analyze nanocrystalline and amorphous silicon phases in the films. Raman spectra were obtained with Horiba Raman Microscope in the range of 350 cm⁻¹–1100 cm⁻¹ using a laser with a wavelength of 532 nm on silicon samples. Raman spectra of the n-layers deposited on Si samples were used to identify and distinguish the presence of amorphous silicon and crystalline silicon phases in the nc-SiOx:H films. Finally, scanning electron microscope (SEM) images were obtained using a Zeiss instrument (Merlin compact) at various magnifications, and we also used an energy-dispersive X-ray (EDX) tool to analyze the compositional changes at nanoscale level.

3. Results and Discussion

3.1. Optimization of Oxygen Content (CO₂ Series)

Carbon dioxide (CO₂) is used an oxygen precursor in the PECVD process for silicon oxide deposition, i.e., it provides oxygen for the formation of the silicon oxide film. The interaction of CO₂ in the plasma generates reactive oxygen species that react with silicon precursors to form silicon oxide bonds. In this set, the CO₂ flow rate was varied from 0–4 sccm in order to determine the optimum deposition conditions for the n-type nc-SiOx:H films. The deposition conditions used for these films were fixed as follows: PH₃—0.8 sccm, SiH₄—4 sccm, power—50 W, pressure—2.5 Torr. Figure 3 shows the effect of the CO₂ flow rate on the passivation quality provided by n-type nc-SiOx:H films. The CO₂ flow ratio is defined in Equation (2) as

$$r_{{CO}_2}=\frac{f_{{CO}_2}}{f_{{SiH}_4}+f_{{CO}_2}} \times 100$$

(2)

where $f_{{CO}_2}$ and $f_{{SiH}_4}$ are the CO₂ and SiH₄ gas flow rates, respectively. We started with $r_{{CO}_2}$= 0%, which means we are likely to obtain amorphous Si or n-type nc-Si:H films.

It can be observed that upon introduction of CO₂ gas in the deposition chamber, there is an improvement in lifetime from 879 ± 31 μs to 2412 ± 72 μs. A similar trend was observed in iV_oc values as well, increasing from 674.5 ± 0.5 mV to 710 ± 1 mV. However, too much CO₂ in the chamber, i.e., increasing the gas flow ratio from 33% to 50%, results in deterioration of the lifetime and the passivation quality of the film (τ_eff = 377 ± 5 μs and iV_oc = 649 mV).

In order to understand the dependence of the effective lifetime on the CO₂ fraction, we analyzed the structural properties of the layers. Figure 4 displays the CO₂ gas flow dependence on the material properties of n-type nc-SiOx:H films. EDX spectra show peaks at 0.27 eV, 0.53 eV, and 1.742 eV corresponding to carbon, oxygen, and silicon, respectively. It verifies that increasing the CO₂ flow rate from 0% to 50% results in an augmentation of O content (from 165 counts → 245 counts → 342 counts) but a relative reduction in Si content (10,970 counts → 10,898 counts → 10,825 counts) within the films. We also observed an increase in the thickness of the films from 15.7 nm to 31.6 nm with a rise in $r_{{CO}_2}$. E₀₄ values however, are in a similar ballpark, i.e., 2.16 eV for these samples.

In the plasma, CO₂ splits into O and CO according to the reaction:

CO₂ → CO + O

(3)

As the $r_{{CO}_2}$ in the deposition chamber increases from 0% to 33%, the presence of CO and O is expected to increase in chamber as well as in the films as per Equation (3). According to the etching model of μc-Si:H formation, during the deposition process, atomic H reaching the film-growing surface weakens the Si–Si bond in the amorphous network, which can be replaced by stronger Si–O bonds due to presence of free O atoms/radicals, resulting in a greater likelihood for the Si–O–Si bond formation. Das et al. demonstrated that a higher concentration of atomic oxygen introduces additional structural defects [12], aiding merit to the above hypothesis.

When $r_{{CO}_2}$ is further increased from 33% to 50%, more free oxygen forms stronger bonds with Si, forming more Si–O–Si bonds that are more energetically stable than Si–Si bonds. This reduces the etching potential due to higher hydrogen dilution, which can lead to an increased presence of O in the films, thicker layers due to the reduced etching, lower crystallinity, and, eventually, poor passivation quality, as seen in case of $r_{{CO}_2}$ = 50% in Figure 3. Although most of the C is expected to leave the deposition chamber in the form of CO or CO₂, a small amount is still embedded in the film [8,27]. Even in this case, at $r_{{CO}_2}$ = 50%, C is incorporated in the deposited n-type nc-SiOx:H film (Figure 4), i.e., 144 counts compared to 121 counts in the case of $r_{{CO}_2}$ = 0%. This could also be a reason for lower crystallinity at $r_{{CO}_2}$ = 50%.

Scanning electron microscopy (SEM) images (Figure 5a,b) also confirm visually that the sample with CO₂ flow rate 33% has more grains than its 50% counterpart. It has also been shown by Das et al. that an increase in the CO₂ flow rate leads to a decrease in the grain size as well as the crystallite size in n-type μc-SiOx:H films [12]. This phenomenon is attributed to the elevated partial pressure of CO₂ within the gas mixture, resulting in higher oxygen content in the films. The heightened presence of atomic oxygen introduces extra structural defects, thereby hindering the growth of crystallites and influencing their size [8]. Also, C gets incorporated in these n-type nc-SiOx:H films. On the other hand, Smirnov et al. and Shin et al. illustrated that extremely high O incorporation in the film can potentially cause an obstruction in the growth of crystallites, resulting in a very low dark conductivity [25,28].

The ability to control the O content in n-type nc-SiOx:H films offers a means to tailor their passivation quality and crystallinity, which are critical for enhancing the efficiency of solar cells. By optimizing the O content, it is possible to achieve improved passivation of crystalline silicon wafers, leading to reduced recombination rates at the surface and, consequently, higher τ_eff. This, in turn, can enhance the efficiency of carrier extraction in solar cells, ultimately contributing to improved overall device performance. Higher oxygen content ($r_{{CO}_2}$ = 33%) enhances the passivation quality of films and crystallinity. However, too much oxygen is detrimental for the overall quality of films. This is due to more incorporation of O in the films at even higher $r_{{CO}_2} (50\%).$ We obtain thicker films but with lower crystallinity and poor passivation quality. The relationship between the oxygen content, crystallinity, and passivation quality of n-type nc-SiOx:H films is a key factor in the development of advanced solar cell technologies, offering a pathway to enhance the efficiency and performance of solar energy conversion devices. Therefore, it is important to keep in mind the application of these passivating n-type nc-SiOx:H films while deciding the flow rate of CO₂ gas. Since we are looking at films with high passivation quality, we choose $r_{{CO}_2}$ = 33% for next set of experiments.

3.2. Optimization of SiH₄ Series

Moving forward, we will explore the influence of the silane (SiH₄) flow rate on the growth of our n-type nc-SiOx:H films. SiH₄ serves as the silicon precursor, i.e., it provides the silicon component for the formation of the silicon oxide films. It undergoes dissociation in the plasma, and the resulting silicon species contribute to the deposition of the silicon oxide film. The incorporation of hydrogen from silane can also influence the hydrogenation level in the film, affecting its passivation. We will be using the hydrogen dilution parameter for ease of comparison, as most publications report this parameter. However, we also report the results with regards to the changing silane flow rate. It is generally understood that a higher hydrogen dilution is necessary to obtain nanocrystalline Si films. The hydrogen dilution ratio is described as R = [H₂]/[SiH₄], and the nature of the grown film depends strongly on this parameter. The deposition conditions used for these films were fixed as follows: CO₂—2 sccm, PH₃—0.8 sccm, power—50 W, pressure—2.5 Torr. As far as the passivation properties of the nanocrystalline silicon oxide films are concerned, we observe that upon decreasing the hydrogen dilution, the lifetime increases drastically from 42 μs (R = 250) to 4020 ± 71 μs (R = 62.5) (Figure 6a). Correspondingly, iV_oc also mirrors this trend as it rises from 578 mV to 715.5 ± 0.5 mV, as shown in Figure 6b. The photoluminescence mapping in Figure 6c shows that the effective carrier lifetime is quite homogeneous over the quarter of wafer.

As the silane flow rate increases, we see an increase in thickness of the n-type nc-SiOx:H films from 25.4 nm to 40.9 nm and, hence, the deposition rate of films (Figure 7a). The optical gap of films also decreases steadily from 2.33 eV to 2.04 eV with an increase in the SiH₄ flow rate (lower R) as shown in Figure 7.

Raman spectra of the films reveal c-Si peak at 520 cm⁻¹, as the substrate was c-Si. The broader features observed around 480 cm⁻¹ are attributed to an a-Si phase. Films with a low silane flow rate (2 sccm and 4 sccm), i.e., higher dilution (high R), show no presence of an a-Si phase. Raman data also illustrate that these n-type nc-SiOx:H films develop an amorphous phase when we increase the silane flow rate from 2 sccm to 8 sccm (Figure 8a). The composition of the films obtained from BEMA modelling (Figure 7b) also shows rise in the a-Si fraction from 16% to 60% with a corresponding surge in flow rate from 2 sccm to 8 sccm. SEM imaging on the 2 sccm sample revealed the existence of very small grains. The elemental EDX analysis (Figure 8b) of the n-type nc-SiOx:H films deposited at a SiH₄ flow rate of 4 sccm reveals higher levels of elemental C (148 counts) and O (351 counts) compared to those of its 8 sccm counterpart (C = 121 counts and O = 214 counts). Figure 7b also indicates a higher presence of the SiO₂ phase (4 sccm—48%, 8 sccm—29%) and nc-Si phase (4 sccm—25%, 8 sccm—11%) in the 4 sccm sample compared to its 8 sccm counterpart. In contrast for samples with an 8 sccm flow rate, the a-Si phase is more dominant, i.e., 60% compared to 27% (Figure 9). It implies that relatively higher hydrogen dilution is required for achieving better micro-crystallinity, whereas low hydrogen dilution is needed for producing amorphous nc-SiOx:H films.

Given the high E₀₄ values at high R (low SiH₄ flow rate), these films are well-suited to be used as back reflector or window layer, whereas films at low R (high SiH₄ flow rate) can act as a passivating contact or even recombination layers in solar cells. Since we focus on highly passivating layers, we chose the SiH₄ flow rate 8 sccm for the next set.

3.3. Optimization of RF Power

Radio frequency (RF) power is another important parameter that affects the dissociation of precursors and, therefore, the film composition. It is known that at a given set of deposition conditions, an increase in ion energy on the substrate may lead to defects and be potentially harmful to the interface [29]. In contrast, because of the increased surface mobility of reactive species, a sufficiently low level of ion bombardment is deemed to be beneficial for growth. Therefore, it is imperative for us to understand the role of power in determining passivation properties and doping nanocrystalline thin films. The deposition conditions for this set are CO₂—2 sccm, PH₃—0.8 sccm, SiH₄—8 sccm, pressure—2.5 Torr.

As the RF power is reduced from 50 W to 20 W, a decline in the passivation properties of the films is observed, with a decrease from 3784 ± 99 μs at 50 W to 759 ± 112 μs at 20 W (Figure 9). The lower passivation at 20 W may be attributed to the deposition of films with a higher susceptibility to defects within the initial few nanometers of growth at lower power, considering the film’s thickness is only 6.4 nm (refer to Figure 10a). Concurrently, iV_oc exhibits an increase from 653 ± 11 mV at lower power (20 W) to 714 ± 1 mV when the power is elevated to 50 W. Additional investigations were conducted at power levels exceeding 50 W (specifically, 60 W, 70 W, and 80 W). Nevertheless, the passivation quality, as indicated by both the effective lifetime (τ_eff) and iV_oc, systematically deteriorates beyond 50 W, which could be attributed to excessive ion bombardment energy (plot in Supplementary Section-Figure S3).

Figure 10 shows that as we decrease the power from 50 W to 20 W, the optical gap of the films increases from 1.56 eV to 1.96 eV. Moreover, a decrease in power also results in a significant drop in the crystalline fraction from 54% to 13%, accompanied by an increase in the a-Si fraction from 24% to 70%. This can be attributed to the fact that at low power, the H₂ dissociation and H production required for the nucleation of crystallites is not enough. A similar trend was reported by Juneja et al. based on Raman spectroscopy and SEM images [30]. Higher power promotes the decomposition of silane (SiH₄) and hydrogen (H₂), generating radicals like SiH_x and atomic hydrogen. SiH₃ specifically acts as a film-forming precursor, and atomic hydrogen promotes the nucleation of a crystalline phase and enhances precursor diffusion on the growing surface [31]. It has been reported that higher applied power favors microcrystalline/nanocrystalline growth [25,32].

Overall optimized power results from a tradeoff between the generation of a high-density plasma and low-energy ions, which result in the deposition of a highly crystalline structure. In order to form good nanocrystalline silicon oxide films, we need good passivation, crystallinity, as well as absorption (i.e., high lifetime, high crystalline fraction, and E₀₄ closer to 2 eV). Therefore, we chose a RF power of 30 W for the next set.

3.4. Optimization of Total Pressure

The deposition conditions for this set are CO₂: 2 sccm, PH₃: 0.8 sccm, SiH₄: 8 sccm, power: 30 W. Figure 11 clearly shows a positive correlation between the pressure and passivation quality provided by nanocrystalline silicon oxide films. Higher chamber pressure results in higher lifetime. For deposition pressure values between 1.4 Torr and 1.8 Torr, the passivation quality is rather low (less than 300 μs). It is only at 2 Torr and above that the passivation quality of the samples improves, with the sample deposited at 3 Torr showing the highest passivation quality. However, because some powder formation was observed on the sample at such a high pressure, we preferred 2.5 Torr for sample deposition.

Figure 12 gives us an insight into this. At low chamber pressure (1.4 Torr–1.8 Torr), we observe that very thin films (3.3 nm–7.5 nm) are formed. It is only at 2 Torr that the deposition rate of the films takes off, resulting in about three-times-thicker films (~23 nm–27 nm). This is consistent with the low lifetime values, as well (Figure 11). It is known that silane dissociation by electron impact and the reaction H+SiH₄ = SiH₃+H₂ are the main channels that produce SiH₃ radicals. Below 2 Torr, the deposition rate is attributed to the increase in the dissociation of rate of SiH₄ and H₂. Between 2 and 2.5 Torr, as the deposition pressure increases, we see a decrease in the nc-Si phase from 21.5% to 18.6%. This trend has also been reported in our previous work on p-uc-SiOx:H films [33] and independently by Gabriel et al. [34]. This is because the deposition rate decreases due to SiH₃ radicals and is dominated by the production of H₃⁺ ions and its thermal fluxes according to the equations shown below [35,36,37,38]:

H₂ + e = H₂⁺ + 2e

(4)

H₂ + H₂⁺ = H₃⁺ + H

(5)

At 3 Torr, we obtain a slightly higher E₀₄ (2.1 eV) and nc-Si phase (~23.7%). Zhang et al. reported that 2.5 Torr is the critical gas pressure where the ionization of H₂⁺ meets the energy threshold, so the production H₃⁺ ions decreases even though the collision frequency still increases [38]. Therefore, H₂⁺ ions’ production and their thermal fluxes dominate the H₃⁺ ions’ production (according to Equation (5)) and their fluxes, resulting in a higher crystalline fraction and higher disorder.

Therefore, relatively low power (30 W) and high pressure (2.5 Torr) work best for passivating crystalline films that are suitable for application in silicon heterojunction devices and perovskite/silicon tandem solar cells.

3.5. Optimization of Phosphine Flow Rate (PH₃ Series)

As we know, PH₃ is a dopant precursor and is used when doping of the nc-SiOx:H film with phosphorus is required. The addition of phosphorus can modify the electrical properties of the film, influencing its conductivity or other characteristics. Therefore, we made films with various values of the phosphine doping ratio by varying the PH₃ flow ratio. The PH₃ flow ratio ($r_{{PH}_3})$ is defined in Equation (6) as

$$r_{{PH}_3}=\frac{f_{{PH}_3}}{f_{{SiH}_4}+f_{{PH}_3}}⨯100$$

(6)

where f_PH3 and f_SiH4 are the PH₃ and SiH₄ gas flow rates, respectively. The PH₃ gas was used as a source of phosphor atoms. Please note that PH₃ is diluted at 1% in H₂. The deposition conditions for these films were fixed as follows: CO₂—2 sccm, SiH₄—8 sccm, power—30 W, pressure—2.5 Torr. Usually, the doping atoms are smaller in number, so it is expected that the crystallinity of the deposited film will not change significantly with the addition of PH₃. However, we do see a change in the passivation quality of the films. Figure 13 shows that upon increasing the flow of PH₃ gas, there is an increase in the passivation quality. τ_eff improves from 1673 μs for $r_{{PH}_3}$ = 6% to 2721 μs for $r_{{PH}_3}$ = 20%, and iV_oc follows a similar trend, which can be related to an increased field effect passivation. For $r_{{PH}_3}$ = 20%, an iV_oc of 697 mV is observed. It should be noted that overall, a uniform passivation with a lifetime in the range of 3.5 ms can be observed on the sample with $r_{{PH}_3}$ = 20%, as shown by PL imaging in Figure 14.

The adaptability of hydrogenated nanocrystalline (nc-SiOx:H) material deposited by industrial, high-throughput PECVD for photovoltaic applications has been discussed in this work. In addition to talking about how pressure, power, and precursor gases (SiH₄, CO₂, and PH₃) affect the optical and structural characteristics of deposited n-type nc-SiOx:H films, we also emphasized the significance of passivation qualities when tailoring their development. These n-type nc-SiOx:H films have a tunable optical gap range of 1.8–2.3 eV and excellent passivation properties (τ_eff up to 4.1 ms, iV_oc~715.5 mV) achieved before annealing. Bearing in mind different applications such as front contacts, back reflectors, and intermediate window layers, it is evident that for the application of n-type, nc-SiOx:H films in solar cells, there is a trade-off between electrical (τ_eff and iV_oc), structural (crystalline fraction), and optical properties (optical gap, E₀₄). Meanwhile, n-type nc-SiOx:H films with a high E₀₄ are favorable for window layers or intermediate reflectors, as their passivation properties are poor. On the contrary, films with high passivation properties and a low E₀₄ are more suited for front contact or even recombination junction layers in tandem solar cells. Given the role oxygen plays in controlling the defect structure and absorption in the films, $r_{{CO}_2}$ is a convenient parameter to vary the microstructure as well as the passivation and optical properties of these films. Both the RF power and pressure play an important role in determining the crystalline fraction of the films and control the amorphous and crystallite phases. Relatively low power (30 W) and high pressure (2.5 Torr) works best for passivating crystalline films suitable for application in silicon heterojunction devices and even perovskite/silicon tandem solar cells. Our results suggest that the properties of n-doped nc-SiOx:H films can be conveniently adjusted to fulfil various requirements as a front contact, intermediate layer, and back reflector.

Furthermore, here are practical insights based on our work that researchers and industry professionals can use to tailor the development of n-type nc-SiOx:H layers to meet the specific requirements of different solar cell applications:

• Fine-tune the $r_{{CO}_2}$ ratio to achieve the desired balance between passivation quality and crystallinity, considering the specific requirements of the solar cell application.
• Controlling the dilution rate (R) during deposition allows for tuning of the hydrogen content in the film for effective passivation.
• Tailor the $r_{{PH}_3}$ to achieve the desired doping level for specific solar cell designs, ensuring optimal passivation properties without introducing excessive defects.
• Choose an appropriate plasma property (power and pressure) based on the desired balance between crystallinity and passivation.

4. Conclusions

This work focused on optimizing the deposition conditions for n-type nanocrystalline silicon oxide (nc-SiOx:H) films, aiming at enhancing their potential for applications in silicon heterojunction devices and tandem cells as a front contact, intermediate layer, and back reflector. The study systematically explored various deposition parameters, including carbon dioxide (CO₂) flow, silane (SiH₄) flow, RF power, total pressure, and phosphine (PH₃) doping, to understand their individual and collective impacts on film properties. In the investigation of CO₂ flow, it was observed that a 33% CO₂ flow ratio positively influenced the minority carrier lifetime (τ_eff = 2412 ± 27.5 μs) and open-circuit voltage (iV_oc = 711 ± 1 mV), indicating improved passivation quality. However, increasing the CO₂ flow ratio to 50% resulted in a decline in both τ_eff and iV_oc, suggesting a delicate balance. The introduction of atomic oxygen from CO₂ influenced the film morphology and composition, impacting crystallite growth and overall passivation quality. Silane (SiH₄) flow was then studied, revealing that higher hydrogen dilution (lower R values) led to a transition from microcrystalline to amorphous phases. An increase in the SiH₄ flow (8 sccm) resulted in lower oxygen incorporation, as indicated by a decrease in the optical gap (2.04 eV). This sample also showed a τ_eff as high as 4.1 ms (iV_oc = 715.5 mV) before annealing. Raman spectra supported these findings, showing shifts in the crystalline and amorphous Si phases. Power optimization was explored, highlighting that power in the range of 30–50 W struck a balance between passivation, crystallinity, and absorption. Films deposited at medium powers exhibited better passivation properties, while powers above 50 W led to increased defects and lower crystallinity. Pressure variation demonstrated a positive correlation between higher pressure and improved passivation quality. Films deposited at 2.5 Torr showed the best properties, indicating the importance of deposition conditions in achieving optimal film characteristics. Lastly, phosphine (PH₃) doping was investigated, showing that increasing PH₃ flow rates improved passivation quality, as seen in the increased τ_eff and iV_oc, probably related to enhanced field effect passivation. Overall, the optimal deposition conditions were identified as a 33% CO₂ flow ratio, 8 sccm SiH₄ flow, 30 W power, 2.5 Torr pressure, and 20% PH₃ flow ratio for doping. These conditions aim to strike a balance between passivation quality, crystallinity, and absorption for potential applications in solar cell technology. Further studies on conductivity and crystalline volume fraction are suggested for a more comprehensive understanding of material properties and practical applications in solar cell technology. Overall, this research contributes valuable insights to the ongoing development of efficient thin-film solar cell technology.