The relentless pursuit of sustainable and efficient energy conversion technologies has placed thin film solar panels at the forefront of photovoltaic research. Among the various thin film technologies, copper-based compounds, such as CIGS and CZTS, have demonstrated remarkable potential due to their high absorption coefficients, tunable bandgaps, and excellent stability. The architecture of these high-performance thin film solar panels is a sophisticated multilayer stack, where each layer plays a critical role. The window layer, typically composed of a transparent conductive oxide (TCO) and a high-resistivity buffer, is paramount for maximizing light ingress and facilitating efficient carrier collection. In this context, intrinsic zinc oxide (i-ZnO) films serve as an indispensable component of this window layer. Its primary functions are to prevent electrical shunting at the p-n junction interface, passivate surface states to reduce carrier recombination, and act as a high-quality seed layer for subsequent TCO deposition. Therefore, the precise control over the structural, morphological, and optoelectronic properties of ZnO thin films is directly linked to the ultimate power conversion efficiency of the final thin film solar panels.
The fabrication of ZnO thin films can be achieved through various techniques, including sol-gel processing, pulsed laser deposition (PLD), and magnetron sputtering. For industrial-scale manufacturing of thin film solar panels, magnetron sputtering stands out due to its exceptional advantages: excellent film uniformity and adhesion over large areas, high deposition rates, dense and pinhole-free microstructures, and outstanding process reproducibility. However, a significant challenge in sputter-deposited ZnO films is the inherent incorporation of point defects, primarily oxygen vacancies (VO) and zinc interstitials (Zni), which act as n-type dopants. This often results in carrier concentrations ($N$) that are too high (≥1018 cm-3) for optimal performance as a buffer layer in thin film solar panels. An ideal i-ZnO layer for this application requires a delicate balance: a low carrier concentration (preferably ~1016 cm-3 or lower) to minimize interface recombination and a high carrier mobility ($\mu$) to ensure efficient transport of photogenerated carriers towards the electrodes. This study focuses on systematically tuning the pulsed-DC magnetron sputtering process parameters—specifically substrate temperature and reactive oxygen gas flow—to engineer ZnO thin films with these tailored properties, making them ideally suited for integration into next-generation copper-based thin film solar panels.
Pulsed-DC magnetron sputtering was employed for the deposition of intrinsic ZnO films onto commercially available soda-lime glass substrates. A high-purity ZnO ceramic target was used. The key advantage of the pulsed-DC power mode over standard DC is its effective mitigation of charge buildup on insulating targets, leading to more stable plasma conditions and higher quality films. The base pressure of the deposition chamber was maintained below $3 \times 10^{-3}$ Pa to minimize impurity incorporation. Argon (Ar) served as the primary sputtering gas, and oxygen (O2) was introduced as a reactive gas to control the film stoichiometry. The total working pressure was fixed at 0.8 Pa. The substrate temperature ($T_s$) and the oxygen-to-total gas flow ratio, defined as $R = O_2 / (O_2 + Ar)$, were the two primary variables investigated. A series of films were deposited with $T_s$ varying from 200°C to 350°C while keeping $R$ constant at 66.7%. Another series was deposited with $R$ varying from 0% to 66.7% while keeping $T_s$ constant at 300°C. All other parameters, including pulse frequency, target-to-substrate distance, and deposition time, were held constant to ensure a consistent film thickness of approximately 220 nm. The deposition conditions are summarized in Table 1.
| Parameter | Value / Condition |
|---|---|
| Target | ZnO Ceramic (99.99%) |
| Substrate | Soda-lime glass |
| Base Pressure | $< 3 \times 10^{-3}$ Pa |
| Sputtering Gas | Ar (99.99%), O2 (99.99%) |
| Working Pressure | 0.8 Pa |
| Power Mode | Pulsed DC (100 kHz frequency) |
| Substrate Temperature ($T_s$) | 200°C, 250°C, 300°C, 350°C (Series 1) |
| O2/(O2+Ar) Ratio ($R$) | 0%, 33.3%, 50%, 66.7% (Series 2) |
| Film Thickness | ~220 nm |
The influence of substrate temperature on the crystallographic structure of the sputtered ZnO films was profound. X-ray diffraction (XRD) θ–2θ scans revealed that all films exhibited a polycrystalline structure corresponding to the hexagonal wurtzite phase of ZnO (JCPDS No. 36-1451). A dominant diffraction peak corresponding to the (002) plane was observed near $2\theta \approx 34.4^\circ$, indicating a strong preferential orientation with the crystallographic c-axis perpendicular to the substrate surface. This c-axis texturing is highly desirable for thin film solar panels as it can influence charge transport anisotropy and surface energy. The intensity of the (002) peak increased systematically with rising $T_s$, as shown in the comparative analysis in Table 2. This enhancement signifies a significant improvement in the overall crystallinity and the degree of preferential orientation. The crystallite size ($D$) was estimated from the full width at half maximum (FWHM, $\beta$) of the (002) peak using the Debye-Scherrer formula:
$$ D = \frac{0.89 \lambda}{\beta \cos \theta} $$
where $\lambda$ is the X-ray wavelength (1.5406 Å) and $\theta$ is the Bragg angle. The calculated $D$ values, listed in Table 2, demonstrate a clear trend of grain growth with increasing substrate temperature. This is attributed to the enhanced surface mobility of adatoms at higher temperatures, allowing them to diffuse to lower energy lattice sites and promote Ostwald ripening.
| $T_s$ (°C) | (002) Peak Intensity (a.u.) | FWHM, $\beta$ (degrees) | Crystallite Size, $D$ (nm) | RMS Roughness (nm) |
|---|---|---|---|---|
| 200 | 1,250 | 0.32 | 26.8 | 2.1 |
| 250 | 1,980 | 0.28 | 30.7 | 3.0 |
| 300 | 3,150 | 0.24 | 35.8 | 4.5 |
| 350 | 4,200 | 0.21 | 40.9 | 6.8 |
Atomic force microscopy (AFM) was employed to investigate the surface morphology and roughness evolution. The AFM images (not referenced by number per instruction) corroborated the XRD findings, showing a clear increase in grain size with $T_s$. The films deposited at lower temperatures (200°C) displayed a relatively smooth, fine-grained surface. As $T_s$ increased to 350°C, the surface transformed into a distinct granular structure with well-defined, larger grains. The root-mean-square (RMS) surface roughness, quantified from the AFM data, increased monotonically from ~2.1 nm at 200°C to ~6.8 nm at 350°C, as detailed in Table 2. While excessive roughness can be detrimental for the formation of sharp heterojunctions in thin film solar panels, a moderate level is often acceptable and can even enhance light trapping within the absorber layer.
The electrical properties of the ZnO films, crucial for their role in thin film solar panels, were characterized using Hall effect measurements at room temperature. The results, plotted as a function of $T_s$, reveal critical trends. The carrier concentration ($N$) remained within the order of $10^{16}$ cm-3 across the temperature range, showing no strong systematic variation. This indicates that the oxygen-rich deposition condition ($R=66.7\%$) was effective in suppressing the formation of intrinsic donor defects like oxygen vacancies, keeping the films in a semi-insulating regime suitable for buffer layers. In contrast, the Hall mobility ($\mu$) exhibited a pronounced increase from approximately 8.5 cm²/V·s at 200°C to 18.4 cm²/V·s at 350°C. This enhancement can be directly correlated with the improved crystallinity and larger grain size. Within a polycrystalline film, charge carriers are scattered by grain boundaries, point defects, and ionized impurities. The mobility ($\mu$) is limited by the combined effect of these scattering mechanisms, often expressed through Matthiessen’s rule:
$$ \frac{1}{\mu} = \frac{1}{\mu_{GB}} + \frac{1}{\mu_{defect}} + \frac{1}{\mu_{ion}} $$
where $\mu_{GB}$, $\mu_{defect}$, and $\mu_{ion}$ represent the mobilities limited by grain boundary, defect, and ionized impurity scattering, respectively. The increase in grain size (decrease in grain boundary density) and improved crystal quality (reduction in point defect density) at higher $T_s$ directly reduces the $1/\mu_{GB}$ and $1/\mu_{defect}$ terms, leading to a higher overall $\mu$. A high-mobility buffer layer is vital for thin film solar panels as it minimizes resistive losses during the extraction of photogenerated carriers.
The optical transparency of the window layer is non-negotiable for efficient thin film solar panels. UV-Vis-NIR spectrophotometry confirmed that all ZnO films, regardless of $T_s$, possessed excellent optical properties. The average transmittance in the visible spectrum (400-800 nm) exceeded 90% for every sample. The optical bandgap ($E_g$) was determined from the transmittance data by applying the Tauc relation for direct bandgap semiconductors:
$$ (\alpha h\nu)^2 = A(h\nu – E_g) $$
where $\alpha$ is the absorption coefficient, $h\nu$ is the photon energy, and $A$ is a constant. Plotting $(\alpha h\nu)^2$ versus $h\nu$ and extrapolating the linear region to the x-axis yielded the $E_g$ values. A slight increase in $E_g$ from ~3.28 eV to ~3.30 eV was observed with increasing $T_s$, which can be attributed to the Burstein-Moss effect. Although the carrier concentration is relatively low, the improved crystallinity may reduce band tailing (Urbach energy), leading to a sharper absorption edge and a marginally larger apparent bandgap. This high and stable transparency is ideal for allowing maximum photon flux to reach the absorber layer in thin film solar panels.
While substrate temperature governs kinetics and crystallinity, the reactive gas composition ($R$) directly controls the film’s stoichiometry and defect chemistry. XRD analysis of films deposited at different $R$ values (0% to 66.7%) at a fixed $T_s$ of 300°C showed that all films retained the hexagonal wurtzite structure with c-axis orientation. However, there was no significant variation in peak intensity or FWHM, suggesting that $R$ has a less pronounced effect on grain size and crystallinity compared to $T_s$ within this parameter space.
The electrical properties, however, were dramatically influenced by the oxygen partial pressure. The results are summarized in Table 3. Films deposited in a pure Ar atmosphere ($R = 0\%$) exhibited a high carrier concentration (~$1.2 \times 10^{19}$ cm-3) and a relatively low resistivity, characteristic of highly n-type ZnO due to a high density of oxygen vacancies. As $R$ increased, $N$ dropped precipitously by several orders of magnitude. This is a direct consequence of shifting the deposition process towards more oxidizing conditions. The increased availability of active oxygen species during growth fills oxygen vacancy sites and suppresses their formation, effectively compensating the native n-type doping. At $R = 66.7\%$, $N$ reached a value of ~$4.5 \times 10^{16}$ cm-3}, which is optimal for a buffer layer in thin film solar panels. The carrier mobility showed a complex, non-monotonic behavior with $R$. Initially, as $R$ increased from 0% to 33.3%, $\mu$ increased, likely due to a reduction in ionized impurity scattering as the concentration of ionized oxygen vacancies decreased. At the highest $R$ of 66.7%, $\mu$ stabilized or slightly decreased, which could be due to the introduction of new scattering centers related to slightly off-stoichiometric compositions or interstitial oxygen.
| $R$ [O2/(O2+Ar)] | Carrier Conc., $N$ (cm-3) | Mobility, $\mu$ (cm²/V·s) | Resistivity, $\rho$ ($\Omega \cdot$cm) | Optical Bandgap, $E_g$ (eV) |
|---|---|---|---|---|
| 0% | $1.2 \times 10^{19}$ | 6.2 | $8.4 \times 10^{-2}$ | 3.34 |
| 33.3% | $5.8 \times 10^{17}$ | 12.1 | $8.9 \times 10^{-1}$ | 3.31 |
| 50% | $1.1 \times 10^{17}$ | 14.5 | 3.9 | 3.29 |
| 66.7% | $4.5 \times 10^{16}$ | 15.8 | 8.7 | 3.28 |
The optical bandgap showed a clear trend associated with the Burstein-Moss effect, as seen in Table 3. For the highly degenerate film deposited with $R=0\%$ ($N \approx 10^{19}$ cm-3), the Fermi level lies within the conduction band. The lowest states in the conduction band are filled, blocking optical transitions to these states. Consequently, photons require higher energy to excite electrons from the valence band to the first available empty states in the conduction band, leading to an enlarged apparent bandgap ($E_g \approx 3.34$ eV). As $R$ increases and $N$ decreases, the Fermi level moves down towards the conduction band minimum, reducing this blocking effect and causing the optical bandgap to decrease to its intrinsic value near ~3.28 eV. All films maintained high visible transmittance (>90%), confirming that the electrical tuning via oxygen control does not compromise optical clarity, a key requirement for the front layer in thin film solar panels.
The integration of these findings provides a comprehensive roadmap for engineering ZnO thin films via pulsed-DC magnetron sputtering. The substrate temperature ($T_s$) acts as the primary driver for microstructural evolution. Higher $T_s$ promotes:
1. Enhanced crystallinity and c-axis orientation.
2. Increased grain size.
3. Higher carrier mobility due to reduced grain boundary and defect scattering.
4. A moderate increase in surface roughness.
Its effect on carrier concentration is secondary when depositing under oxygen-rich conditions.
Conversely, the oxygen gas ratio ($R$) serves as the primary knob for controlling electronic structure:
1. It directly and powerfully modulates the carrier concentration over several orders of magnitude by controlling the oxygen vacancy density.
2. It allows for precise tuning of the film from a conductive TCO-like state ($R$ ~ 0%) to a highly resistive buffer layer state ($R$ > 50%).
3. It induces the Burstein-Moss shift in the optical bandgap for highly doped films.
Its effect on crystallinity is relatively weak compared to $T_s$.
For the specific application as the high-resistivity component (i-ZnO) of the window layer in copper-based thin film solar panels, the optimal combination derived from this study involves a relatively high substrate temperature (e.g., 300-350°C) paired with a high oxygen gas ratio (e.g., $R$ = 50-66.7%). This recipe yields a film with:
• Excellent crystallinity for high carrier mobility ($\mu$ > 15 cm²/V·s).
• Low carrier concentration ($N$ ~ 1016 – 1017 cm-3) to minimize interface recombination.
• Outstanding optical transparency (>90%).
• A suitable surface morphology for subsequent layer deposition.
This tailored ZnO film would effectively perform its duties: passivating the underlying buffer/absorber interface, preventing shunting paths, and serving as an ideal foundation for a low-resistance, transparent top TCO contact (like Al:ZnO or ITO). The systematic control demonstrated here underscores the versatility of magnetron sputtering for fabricating functional layers with tightly specified properties. By decoupling the influences of thermal energy and plasma chemistry, manufacturers of thin film solar panels can precisely fine-tune the i-ZnO buffer layer to match the specific requirements of different absorber materials (CIGS, CZTS, etc.), thereby pushing the boundaries of efficiency and stability in this promising class of photovoltaic devices. The ongoing optimization of every layer, including the seemingly simple ZnO window layer, is what propels the continuous advancement of thin film solar panel technology towards higher performance and broader commercialization.