Enhancing Conductivity of Nickel Oxide to Achieve Scalable Preparation for High-Efficiency Perovskite Solar Modules

Abstract

NiO_x, prepared via the sputtering method, exhibits low conductivity and energy level mismatch with the perovskite layer, thereby limiting further enhancements in the performance of perovskite solar modules (PSMs). Unlike traditional methods that enhance the performance of NiO_x through reactive sputtering or directly doping NiO_x targets with metal ions, both of which incur high costs and low efficiency, we employ an evaporation method using LiF to achieve efficient and low-cost doping of NiO_x. Compared to the pristine NiO_x, the incorporation of LiF significantly increases the conductivity of NiO_x. Additionally, the incorporation of LiF enhances the quality of the deposited perovskite films, as well as the energy level alignment and symmetry between NiO_x and the perovskite, effectively improving the hole extraction and transport capabilities between NiO_x and the perovskite. As a result, the PSM (active area of 57.30 cm²) fabricated in air achieves an impressive efficiency of 19.54%. Furthermore, the unencapsulated PSM retains 80% of its initial efficiency after 700 h of continuous illumination, whereas the NiO_x-based PSM drops to 80% after only 150 h. This study provides a simple and low-cost method for doping NiO_x, which is of great significance for the further industrialization of PSMs.

1. Introduction

Nickel oxide (NiO_x), as an excellent P-type semiconductor, is widely used as a hole transport layer (HTL) in large-area inverted perovskite solar cells (PSCs) due to its low cost, good stability, and superior scalability [1,2,3]. Various methods for depositing NiO_x films over extensive areas include solution-based techniques such as blade coating, spray coating, and electrochemical deposition, as well as vacuum-based methods like electron beam evaporation, atomic layer deposition, and magnetron sputtering [4,5]. Among these methods, magnetron sputtering stands out as a well-established deposition technology, offering advantages such as high uniformity, repeatability, low-temperature deposition, and excellent film thickness control [6,7]. These characteristics make it particularly suitable for large-area, high-quality deposition of NiO_x layers on both planar and flexible substrates, which is crucial for the industrialization of PSCs [8]. However, despite the widespread use of NiO_x prepared via magnetron sputtering as the HTL in perovskite solar modules (PSMs), its relatively low conductivity, poor geometric symmetry with the perovskite layer, and energy level mismatch significantly hinder interfacial hole extraction and transport efficiency [9,10]. Poor geometric symmetry refers to the roughness of NiO_x affecting the quality of the deposited perovskite film. This ultimately limits the further enhancement of the open-circuit voltage (V_oc) and fill factor (FF), and compromises the long-term stability of the device.

To address this issue, strategies typically involve metal ion doping, including Cs⁺, Cu²⁺, Mg²⁺, and Li⁺⁸ [11,12], as well as surface passivation methods utilizing various organic molecule interlayers [13,14], PTAA [15,16], and self-assembled monolayers [17,18]. Nevertheless, in large-scale industrial applications, organic molecular layers are usually applied using methods such as blade coating. The thickness of these molecular layers typically measures only a few nanometers, making it challenging to control thickness and uniformity with simple blade coating techniques. Furthermore, while organic materials offer significant advantages in terms of photoelectric conversion efficiency, they generally exhibit poor stability, which often leads to suboptimal long-term performance of the devices [18,19]. Therefore, ion doping emerges as an effective method, with Li⁺ being the most commonly used element for doping NiO_x. Typically, there are two methods for doping NiO_x using the magnetron sputtering method: (1) A Ni target and a doped metal target can be employed, introducing argon and oxygen into the chamber, respectively, while adjusting the power of both targets and the gas flow rates of argon and oxygen to achieve metal doping of NiO_x films. However, this method necessitates precise control of power and gas flow rates, complicating the process and reducing repeatability. (2) A NiO_x target pre-doped with metal ions can also be utilized; however, this method requires multiple attempts to produce a doped target with optimal performance. Once the metal doping ratio in the target is established, it cannot be adjusted further, which will undoubtedly increase the overall production costs of the target.

Here, we achieve doping of NiO_x through a low-cost and simple method by depositing LiF onto sputtered NiO_x. X-ray photoelectron spectroscopy (XPS) analysis reveals a significant increase in the ratio of Ni³⁺ to Ni²⁺ in NiO_x, indicating that the successful doping of Li⁺ has enhanced the conductivity of NiO_x. The incorporation of LiF also reduces the contact angle of NiO_x, resulting in improved wetting of the perovskite precursor solution on the NiO_x film, which subsequently enhances the crystallinity of the deposited perovskite films. Combined with ultraviolet photoelectron spectroscopy (UPS) characterization and studies of interfacial carrier transport dynamics, the results indicate that the incorporation of LiF increases hole concentration in NiO_x, achieves better energy level alignment with the perovskite, and enhances the symmetry between NiO_x and the perovskite, thereby effectively improving the V_oc and FF of the device. Consequently, the PSMs fabricated via slot-die coating in air, with an active area of 57.30 cm², demonstrate an efficiency increase from 16.33% to 19.54%. Furthermore, the unencapsulated NiO_x/LiF-based PSM retains 80% of its initial efficiency after 700 h of continuous illumination in air, significantly outperforming the NiO_x-based devices, which only maintained 80% efficiency after 150 h.

2. Materials and Methods

2.1. Materials

Methylammonium bromide (MABr, 99.9%), lead (II) iodide (PbI₂, 99.999%), lead (II) bromide (PbBr₂, 99.99%), and bathocuproine (BCP) were purchased from Xi’an Polymer Light Technology corp,(Xi’an, China). Methylammonium chloride (MACl, 99.9%), formamidine hydroiodide (FAI, 99.9%), and cesium iodide (CsI, 99.999%) were purchased from Advanced Election Technology Co., Ltd, (Liaoning, China). Lithium fluoride (LiF) was purchased from Shanghai Aladdin Biochemical Technology Co.,Ltd, (Shanghai, China). C₆₀ was purchased from Luminescence Technology Corp, (Taiwan, China). N-dimethylformamide (DMF, 99.8%) and 1-methyl-2-pyrrolidinone (NMP, 99%) were purchased from Sigma-Aldrich, (St. Louis, Missouri, USA).

2.2. Device Fabrication

For the 10 × 10 cm² solar module, the FTO glass substrates were sequentially cleaned with acetone, deionized water, and anhydrous ethanol, with each ultrasonic cleaning step lasting 15 min. A 1064 nm red laser was utilized to scribe the FTO substrate, forming 11 strips (P1). The NiO_x layer was prepared by radio-frequency (RF) sputtering onto the pre-cleaned FTO glass substrates—previously treated with ultraviolet ozone for 20 min. This process used a chamber pressure of 0.37 Pa, RF power of 200 W, and an argon flow rate of 100 sccm for 30 min. Following this, LiF was thermally evaporated onto the NiO_x films in a vacuum chamber (<5 × 10⁻⁴ Pa) and then annealed at 300 °C for 20 min under ambient conditions. For the preparation of Cs₀.₀₅MA₀.₁FA₀.₈₅Pb(Br₀.₁I₀.₉)₃ perovskite films, a 1 M perovskite precursor solution was prepared by dissolving FAI, PbI₂, MABr, CsI, PbBr₂, and MACl (0.18 mmol) in a mixed solvent of DMF and NMP (v/v: 6:1). The perovskite films were coated onto the FTO/NiO_x substrate using the slot-die method, maintaining a gap of 200 μm between the slot and the coated substrate, a feed pump speed of 1.2 μL/s, and a substrate movement speed of 7000 μm/s. After coating, the substrate was rapidly transferred to a vacuum for 35 s and then annealed at 100 °C for 30 min to yield dark films. Subsequently, C₆₀ (55 nm) and BCP (8.5 nm) were thermally evaporated sequentially onto the perovskite films in a vacuum chamber (<5 × 10⁻⁴ Pa), and a 532 nm green laser was employed to etch the resulting perovskite/C₆₀/BCP films (P2). Finally, a 200 nm copper electrode was thermally evaporated onto the BCP film in a vacuum chamber (<3 × 10⁻⁴ Pa), and the copper electrode was divided using the same 532 nm green laser with a line width of 350 μm (P3). The active area of the modules manufactured on a 10×10 cm² PSM measures 57.3 cm², with an aperture area of 64.38 cm² and comprising 11 series-connected cells, resulting in a geometric fill factor (GFF) of 89% [20,21].

2.3. Characterization

Ultraviolet (UV)–visible absorption and transmission spectra were recorded using a Shimadzu UV-1900 spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Photoluminescence (PL) spectra and time-resolved photoluminescence (TRPL) measurements were obtained using the FLS980 Series of Fluorescence Spectrometers (Edinburgh Instruments, Edinburgh, UK) with an excitation wavelength of 480 nm. The surface roughness of the films was assessed with the S NEOX 090 (Sartorius, Golsheim, Germany), and the contact angles were measured using the XG-CAMA1 (XG Optoelectronics Technology Co., Ltd, Shenzhen, China). Ultraviolet photoelectron spectroscopy (UPS) spectra were obtained using the Escalab 250Xi (Thermo Fisher Scientific, Massachusetts, USA). Similarly, X-ray photoelectron spectroscopy (XPS) measurements were performed with the same Escalab 250Xi instrument. X-ray diffraction (XRD) analysis was conducted on an Ultima IV X-ray diffractometer (Rigaku Corporation, Tokyo, Japan), utilizing Cu Kα radiation (λ = 1.54 Å) at a scanning speed of 5° min⁻¹. Scanning electron microscopy (SEM) images were collected using a JEOL 7610F field-emission scanning electron microscope (JEOL Ltd., Tokyo, Japan). The J–V characteristics of the perovskite solar modules were measured with a Keithley 2400 source meter (Tektronix, Oregon, USA) under simulated sunlight conditions of 100 mW cm⁻², following the AM 1.5G standard air mass spectrum (Newport, Oriel Class A, 91195A). J–V curves were obtained through reverse scanning (from 13 V to −0.1 V) and forward scanning (from −0.1 V to 13 V). The steady-state power conversion efficiency was derived from measurements of stable photocurrent density at the maximum bias voltage (Vmax). To ensure consistency across measurements, all samples were sourced from the same batch and tested in identical positions. Five samples were evaluated for each characterization type, including PL, TRPL, UV-Vis absorption, UV-Vis transmission, XRD, surface roughness, contact angles, conductivity, and mobility, with the reported data representing the median value of the test results. For XPS, SEM, and UPS, the cutting positions were maintained consistently, and at least five modules are prepared for each condition to perform J–V testing.

3. Results and Discussion

To assess whether the deposition of LiF impacts the transmittance of NiO_x films, we characterized the films using ultraviolet–visible (UV-Vis) transmission spectroscopy. The NiO_x/LiF films exhibit transmission spectra similar to that of pristine NiO_x (Figure S1). Notably, the NiO_x film with 3 nm of LiF displays maximum transmittance in the 400–550 nm range (Figure 1a). This enhanced transmittance is attributed to a reduction in surface roughness (to be elaborated later), which decreases light scattering and reflection loss, ultimately benefiting light absorption by the perovskite layer [22]. To investigate the influence of deposited LiF on the electrical properties of the NiO_x films, we fabricated devices with the structure FTO/HTL/Cu and measured the hole mobility and conductivity of both NiO_x and NiO_x/LiF films using the space-charge-limited current (SCLC) method [23]. We employed the Mott–Schottky equation (Equation (1)) to fit the measured current–voltage (J–V) curves of the devices, enabling the calculation of hole mobility for the films [24].

$$J={\displaystyle \frac{9{\epsilon }_0{\epsilon }_r{V}^2}{8L^3}}$$

(1)

Here, J represents current density, μ represents hole mobility, ${\epsilon }_0$ represents the vacuum permittivity, ${\epsilon }_r$ represents the dielectric constant of NiO_x, L represents the thickness of the NiO_x film, and V represents the applied voltage to the device. As shown in Figure 1b, the hole mobilities of the NiO_x film and NiO_x/LiF film are 1.27 × 10⁻⁴ cm²V⁻¹ s⁻¹ and 1.91 × 10⁻⁴ cm²V⁻¹ s⁻¹, respectively. Clearly, the incorporation of LiF enhances the hole mobility of the NiO_x film, thereby facilitating hole transport. Additionally, we calculated the conductivity of the NiO_x films using Equation (2) as follows [25]:

$$\sigma ={\displaystyle \frac{d}{AR}}$$

(2)

where σ is the conductivity, d is the thickness of the NiO_x film, A is the active area, and R represents the resistance determined from the I–V curves of the NiO_x or NiO_x/LiF films. It can be calculated that the conductivities of the NiO_x and NiO_x/LiF films are 4.77 × 10⁻⁶ S/cm and 1.18 × 10⁻⁵ S/cm, respectively (Figure S2). This enhancement in conductivity is advantageous for reducing recombination losses of electrons and holes.

To investigate the mechanism by which deposited LiF enhances the conductivity and hole mobility of NiO_x films, we conducted XPS analysis on the surfaces of both NiO_x and NiO_x/LiF films to study their elemental chemical states [10,26]. As shown in Figure 1c,d, we performed Lorentz–Gaussian peak fitting on the Ni2p_3/2 peaks of both the NiO_x and NiO_x/LiF films. The peak at 854 eV corresponds to Ni²⁺, while the peak at 855 eV is attributed to Ni³⁺ induced by Ni²⁺ vacancies. Notably, the ratio of Ni³⁺/Ni²⁺ increases from 2.98 to 3.02, which suggests that Li⁺ has been incorporated into the NiO_x matrix, leading to an improvement in conductivity and mobility [27]. Additionally, the X-ray diffraction (XRD) results for both NiO_x and NiO_x/LiF films indicate that they have the same crystalline phase structure (Figure S3). However, the XRD peaks of the NiO_x/LiF film exhibit a blue shift, confirming that Li⁺ incorporates into the lattice of NiO_x²⁰. The relationship between homogeneous and heterogeneous nucleation indicates that the wettability of the substrate influences the nucleation and crystallization of the deposited perovskite, as illustrated in Equation (3) [28,29]:

$$G_{heterogeneous}=G_{homogeneous}\times {\displaystyle \frac{\left(2+\cos \theta \right){\left(1-\cos \theta \right)}^2}{4}}$$

(3)

where $G_{heterogeneous}$ represents the nucleation free energy and θ represents the contact angle of the solid/liquid interface. As illustrated in Figure 1e,f the NiO_x/LiF film surface exhibits slightly lower root mean square roughness (Sq = 0.69 nm) and arithmetic mean roughness (Sa = 0.55 nm) compared to the NiO_x film, which has a root mean square roughness of Sq = 0.75 nm and an arithmetic mean roughness of Sa = 0.60 nm. We also measured the contact angle, and the results exhibit that the contact angle of the NiO_x film is 48.2°, while that of the NiO_x/LiF film is significantly reduced to 37.1°. This reduction in contact angle and roughness indicates that the perovskite precursor solution adheres better to the NiO_x/LiF substrate, which lowers the energy barrier for perovskite nucleation and facilitates the formation of high-quality perovskite films [30].

Therefore, to observe the changes in the prepared perovskite films, we examined the morphology of the perovskite films deposited on NiO_x and NiO_x/LiF substrates using scanning electron microscopy (SEM) [31]. As shown in Figure 2a,b, the perovskite films deposited on both substrates exhibit similar morphologies and have dense polycrystalline grains, ensuring effective light absorption [32]. However, particle size analysis from the SEM images reveals that the average grain size of the perovskite film deposited on the NiO_x substrate is 169.5 nm, while that of the perovskite film on the NiO_x/LiF substrate significantly increases to 269.7 nm. This increase in grain size helps reduce the defect density in the perovskite films, improves their crystallization quality, and thereby enhances their optoelectronic performance and stability [33,34].

The improvement in the crystallization quality of the perovskite films was further confirmed by XRD characterization. As shown in Figure 1c, there are two sharp diffraction peaks at 14.06° and 28.36°, corresponding to the (001) and (002) crystal planes of the perovskite [17]. Compared to the perovskite film deposited on NiO_x, the intensity of diffraction peaks in the films deposited on NiO_x/LiF significantly increases, indicating a substantial enhancement in crystallinity and a higher degree of preferred orientation. Additionally, photoluminescence (PL) and absorption spectra of the perovskite films reveal that, compared to the films deposited on NiO_x, the PL and absorption intensities of the films on NiO_x/LiF significantly increase (Figure 2d and Figure S4), demonstrating improved optical performance attributed to enhanced crystallization quality [35,36]. As shown in Figure 2e,f, the perovskite film deposited on NiO_x/LiF exhibits a significant reduction in roughness compared to the film on NiO_x, with Sq decreasing from 12.96 nm to 9.37 nm and Sa decreasing from 9.85 nm to 7.61 nm. This reduction in roughness indicates that the perovskite films on the NiO_x/LiF substrate are smoother, which helps reduce interface defects and improve charge carrier transport efficiency [37], thereby enhancing the performance of the device.

The efficient transfer of photogenerated charge carriers between the hole transport layer and the perovskite layer is influenced by the energy level offset between their valence bands. This factor also significantly impacts the V_oc of the PSCs. To elucidate the energy level changes in NiO_x and NiO_x/LiF, we analyzed the energy level structures of the films using UPS [38], as shown in Figure 3a. Here, E_cut-off represents the cut-off energy for secondary electrons, and E_onset denotes the energy difference between the valence band maximum (E_VBM) and the Fermi level (E_f). The E_f and E_VBM can be calculated using the following Equations (4) and (5), where 21.22 is the photon energy of HeI in electronvolts (eV) [39].

$$E_f=21.22-E_{cut\text{-}off}$$

(4)

$$E_{VBM}=21.22-(E_{cut\text{-}off}-E_{onset})$$

(5)

Through the UPS spectra, the E_f and E_VBM of NiO_x films can be calculated as 4.81 eV and 5.27 eV, respectively. For the NiO_x/LiF film, the E_f and E_VBM values increase to 4.92 eV and 5.31 eV, respectively. The schematic in Figure 3b illustrates that after the incorporation of LiF, the valence band position of the NiO_x film aligns more closely with that of the perovskite layer. This alignment suggests that the symmetry between the NiO_x and perovskite layers has improved and the energy barrier has been reduced, facilitating rapid hole extraction at the NiO_x/perovskite interface and decreasing energy losses, thereby optimizing the V_oc of the PSCs [40]. Additionally, the energy gap between E_f and E_VBM decreased from 0.46 eV to 0.39 eV, which helps to increase the hole concentration in NiO_x and contributes to improvements in short-circuit current density (J_sc) and FF of the PSCs [41].

Furthermore, we investigated the interface carrier extraction and transport capabilities between the perovskite layer and NiO_x layer using PL and time-resolved photoluminescence (TRPL) characterization [42,43]. Figure 3c,d display the PL and TRPL spectra, with irradiating fluorescence incident from the glass side. Compared to the perovskite film deposited on NiO_x, the perovskite film deposited on NiO_x/LiF exhibits significant PL quenching, indicating that the incorporation of LiF reduces charge carrier recombination at the interface, thereby enhancing hole extraction and transport via NiO_x. To calculate the charge transfer kinetics between NiO_x and the perovskite, a double-exponential function was employed to fit the TRPL curve, as described by the following equation:

$$I(t)=I_0+A_{1}exp({\displaystyle \frac{-t}{{\tau }_1}})+A_{2}exp({\displaystyle \frac{-t}{{\tau }_2}})$$

(6)

Here, A₁ and A₂ represent the decay amplitudes, while ${\tau }_1$ and ${\tau }_2$ correspond to the charge carriers quenched by the HTL and radiation recombination in the perovskite film. The parameters and results obtained from the fitting are presented in Table S1. The formula for calculating the average lifetime is ${\tau }_{ave}=(A_1{{\tau }_1}^2+A_2{{\tau }_2}^2)/(A_1{\tau }_1+A_2{\tau }_2)$ [44,45]. The results indicate that after the incorporation of LiF, ${\tau }_1$ decreases from 2.46 ns to 1.29 ns, ${\tau }_2$ decreases from 99.26 ns to 68.62 ns, and the average lifetime decreases from 95.00 ns to 65.20 ns. These changes demonstrate a significant enhancement in NiO_x’s ability to extract holes from the perovskite, consistent with PL results [46]. This improvement can primarily be attributed to two factors: the enhanced crystallization quality of the perovskite film and the improved P-type performance of the NiO_x film.

As shown in Figure 4a,b, we fabricated a typical inverted-structure PSM based on FTO/NiO_x/perovskite/C₆₀/BCP/Cu, with the perovskite layer prepared in air using a slot-die coating method. The PSM consists of 11 interconnected sub-cells, with an active area of 57.30 cm². The area between sub-cells P1 and P3, termed the “dead area”, does not contribute to the photocurrent and has a width of approximately 800 μm (Figure S5), resulting in a geometric fill factor of 89.68% for the PSM. Figure 4c shows the impact of different LiF thicknesses on device performance, indicating that optimal performance is achieved with a deposition thickness of 3 nm. To verify the reproducibility of the performance enhancement, we fabricated 10 PSMs based on NiO_x and NiO_x/LiF, with the distribution of their photovoltaic parameters statistically presented in Figure 4d and Figure S6. These results demonstrate that after the incorporation of LiF, all parameters exhibit significant increases, reflecting the high reproducibility of device performance. Subsequently, the champion PSMs based on NiO_x and NiO_x/LiF were further measured using forward and reverse scans (Figure 4e). In the forward scan, the devices based on NiO_x/LiF exhibit an increase in V_oc from 10.81 V to 11.59 V, J_sc from 2.15 to 2.19 mA cm⁻², and FF from 67.06% to 75.84%, resulting in an increase in PCE from 15.57% to 19.24%. In the reverse scan, V_oc increases from 11.08 V to 11.67 V, J_sc from 2.15 to 2.20 mA cm⁻², and FF from 68.53% to 76.17%, and the resulting PCE increases from 16.33% to 19.54%. The improvement in PCE is primarily attributed to enhanced conductivity of NiO_x, better energy level alignment with the perovskite layer, and improved crystallization quality of the perovskite film. The hysteresis factor of the devices based on NiO_x/LiF is significantly reduced, mainly due to the considerable enhancement in carrier transport at the NiO_x/perovskite interface [39]. Furthermore, the champion PSM based on NiO_x/LiF maintains maximum output voltage for 800 s, with results showing that the output photocurrent density stabilizes at 1.97 mA cm⁻², corresponding to a stable PCE of 19.23% (Figure 4f). In order to evaluate the impact of LiF incorporation on the stability of the PSM, unencapsulated devices were placed in an environment with 60–70% relative humidity at 45 °C, and maximum power point tracking (MPPT) tests were conducted under white light-emitting diode (LED) illumination (equivalent to one-sunlight intensity) [22]. According to the results presented in Figure 4g, the NiO_x/LiF-based devices maintain 80% of their initial efficiency after continuous operation for 700 h, while the NiO_x-based devices experience a 20% decrease in efficiency after only 150 h under the same testing conditions. The incorporation of LiF significantly improves module stability, primarily due to the enhanced quality of the perovskite film and reduced interface carrier recombination.

4. Conclusions

In summary, we employ a simple and cost-effective method to achieve efficient doping of NiO_x by thermally evaporating LiF onto sputtered NiO_x. After the incorporation of LiF, the ratio of Ni³⁺ to Ni²⁺ in NiO_x increases from 2.98 to 3.02, indicating that the effective doping of Li⁺ has led to an improvement in the conductivity of NiO_x. The reduction in the contact angle of the NiO_x/LiF film enhances the wettability of the perovskite precursor solution on NiO_x, thereby improving the crystallization quality of the deposited perovskite film. Additionally, the incorporation of LiF results in a significant increase in hole concentration in NiO_x and brings its valence band position closer to that of the perovskite layer. This adjustment greatly enhances the efficiency of hole injection from the perovskite layer to the NiO_x and reduces energy loss during transport, improving the V_oc and FF of the devices. Based on the improved P-type performance of NiO_x and enhanced perovskite crystallization quality, the resulting PSM achieves an exceptional PCE of 19.54%. Furthermore, compared to the PSM based on NiO_x, the unencapsulated LiF-incorporated PSM maintains 80% of its initial efficiency after 700 h of continuous illumination, while the efficiency of the NiO_x-based PSM decreases by 20% after only 150 h, demonstrating the significant improvement in the stability of the NiO_x/LiF-based PSM.