Lead (II) fluoride additive modulating grains growth of water-processed metal halide perovskites for enhanced efficiency in solar cells

Perovskite solar cells (PSCs) have emerged as a promising candidate for next-generation photovoltaic technology due to their remarkable efficiency and low-cost fabrication processes [1–3]. In comparison to conventional silicon solar cells, PSCs manifest distinct advantages, including high efficiency achieved with low manufacture cost by solution processing techniques, inherent flexibility, and lightweight characteristics [3]. Such merits render PSCs not only conducive to conventional solar panels but also extend their applicability to a diverse array of domains such as portable and wearable electronics as well as building-integrated photovoltaics [4–7]. As the commercialization of PSCs is on the edge, concerns regarding their stability have been raised alongside a significant focus on the environmental impact of the manufacturing process [8]. The use of large quantities of toxic solvents in the solution-processing manufacture of PSCs could impede the widespread production of PSCs. For example, Europe has officially banned the use of N,N-Dimethylformamide (DMF), a critical solvent in solution processing of PSC, in manufacture due to concern about its effect on human health [9]. In response to this challenge, alternative synthesis routes and less toxic solvents must be developed to address the issue. Among the different fabrication methods [10–13], the water-based solution processing approach has emerged as particularly noteworthy due to its environmental sustainability [14–17]. This approach offers a more eco-friendly, cost-effective solution to the manufacturing process of PSCs, contributing to a more sustainable future for solar energy technologies.

The water-based synthesis method of perovskite thin films for use in PSCs was pioneered by Hsieh et al in 2015, achieving a power conversion efficiency (PCE) of 12.5% without the use of traditional toxic DMF solvents [14]. The poor PCE was attributed to inadequate control of the morphology of perovskite layer stemming from its dissolution and recrystallization during the transformation process from Pb(NO₃)₂ to MAPbI₃. This challenge arises from the very slow perovskite conversion, which is strongly influenced by diffusion dynamics of ions [15, 18, 19]. Modulating crystal growth, particularly accelerating the Pb(NO₃)₂-to-perovskite conversion, is critical for improving the performance of water-processed PSCs. Various strategies have been undertaken to address this challenge. One effective approach is the implementation of a multi-cycle procedure [15], which entails dividing the prolonged incubation period into multiple shorter immersion-wash-dry cycles while upholding the total immersion time of the Pb(NO₃)₂ film in the methylammonium iodide (MAI) solution, thereby enabling a more effective interfacial reaction to control crystal growth, leading to improved perovskite morphology and enhanced PSCs performance [15, 20]. The light radiation spectrum of ambient light during perovskite fabrication and its intensity have been found to influence perovskite crystallization made from Pb(NO₃)₂. Zhai et al demonstrated that varying light intensity can control perovskite growth in water-based processing, enabling fully covered films with uniform grains and achieving a PCE of 23.7% [16]. Furthermore, utilizing additive engineering to modulate the crystallization process of perovskite material has proven to be highly effective [21–24]. Specifically, The introduction of nanosized PbI₂ seeds and halide-free lead salts (PbCO₃, Pb(HCOO)₂, Pb(SCN₂) into the Pb(NO₃)₂ solution were reported to enhance grain formation, with PbCO₃ achieving a PCE of 23.95% [22, 23]. Additionally, methylammonium chloride (MACl) can improve the diffusion kinetics of cations and anions, accelerating the Pb(NO₃)₂-to-perovskite conversion and resulting in a film with more uniform grain distribution. The enhancement is attributed to the smaller size and higher electronegativity of chloride anions compared to iodide anions, which facilitates faster ion transport and crystallization [25]. Nonetheless, despite these advancements, the performance, adaptability, reproducibility, and scalability of PSCs made from the water-based processing technique still lag significantly behind those based on the traditional solution processing using hazardous organic solvents such as DMF and dimethyl sulfoxide (DMSO), primarily due to high defect densities in the fabricated perovskite films [17]. A major challenge remains in regulating crystal growth in the water-processed perovskite fabrication while simultaneously passivating defects in the fabricated perovskite films.

Our work employs PbF₂ as an additive, leveraging fluoride anions to enhance perovskite formation, improve crystallization, and reduce defects, ultimately increasing the stability and efficiency of water-processed PSCs. Given that fluoride anion (F⁻) possesses the highest electronegativity, its incorporation can effectively passivate both anion and cation vacancies [26, 27]. A study on the passivation effects of crown ethers with varying coordination abilities toward Pb²⁺ by Zhang et al has revealed the influence of both enthalpy (the electron-donating ability of the host molecule toward Pb²⁺) and entropy (the interaction distance and flexibility of the host molecule) in determining the effectiveness of passivating ligands additives in perovskites. An effective passivating agent should have high enthalpy or strong electron-donating ability. This further suggests the ability of fluoride (F⁻) anion in passivating defects in perovskite materials [28].Incorporating fluorinated molecules into PSCs made from conventional solution process has demonstrated the ability to increase device efficiency and stability [29–31]. The beneficial potentials of incorporating PbF₂ in perovskite materials have been theoretically predicted [32]. However, to the best of our knowledge, the use of PbF₂ as an additive and its effects on crystallization kinetics of perovskite made from water -based process and on the corresponding device performance have not yet been experimentally investigated. This represents a critical gap in research related to green solvent processing for PSCs and the application of fluoride chemistry in perovskite materials.

Herein, we introduce PbF₂ as an additive to control the growth of perovskite crystals in the water-based processing of PSCs. The study shows that the inclusion of PbF₂ not only accelerates the transformation of the Pb(NO₃)₂ film to perovskite but also enhances the preferred crystal orientation of the resulting perovskite material. The fluoride anion (F⁻), with its high electronegativity, effectively passivates the surface of perovskite grains, thereby improving both the optical and electrical characteristics of the perovskite. As a result, a significant improvement in the performance of the devices is observed, with the PbF₂-treated device achieving a PCE of 18.1%, surpassing the pristine device's PCE of 16.3%. Moreover, the device exhibits prolonged operational and environmental lifetimes.

2.1. Materials

2.2. Water processed fabrication of PSCs

2.3. Characterization

2.4. Density functional theory (DFT) computations

DFT calculations were performed using the Vienna Ab initio Simulation Package with the projector-augmented wave pseudopotential [33, 34]. The electron exchange-correlation interactions were described by the Perdew–Burke–Ernzerhof functional within the generalized gradient approximation level [35]. A plane-wave cutoff of 520 eV was applied to expand the smooth part of the wave function. The convergence criteria for the electronic self-consistent loop and the Hellmann–Feynman forces were set to less than 10⁻⁵ eV and 0.01 eV Å⁻¹, respectively. The Brillouin zones were sampled using the Monkhorst–Pack mesh with a reciprocal space resolution of 2π × 0.04 Å⁻¹ for all calculations. Gaussian smearing was employed with a smearing width of 0.05 eV. A vacuum layer of more than 15 Å was applied in the z-direction to prevent artificial interactions between periodic cells. Grimme's D3 correction was adopted to describe the long-range van der Waals interactions [36–40]. To model the FAPbI₃ (100)/PbF₂ heterostructure, a (1 × 1) unit cell for the FAPbI₃ (100) surface and a (√2 × √2) supercell for the PbF₂ (001) surface are used. For constructing the FAPbI₃ (110)/PbF₂ heterostructure, a (3 × 1) supercell for the FAPbI₃ (110) surface and a (2√5 × √5) supercell for the PbF₂ (001) surface are used, respectively.

The interlayer potential (φ, eV Å⁻²) is defined as the interlayer binding energy per unit area for a specific interface, which is calculated using the following equation:

$$\begin{align}\phi = \frac{{{E_{{\text{tot}}}} - {E_{{\text{FAPb}}{{\text{I}}_{\text{3}}}}} - { }{E_{{\text{Pb}}{{\text{F}}_{\text{2}}}}}}}{{{A_{{\text{surf}}}}}}\end{align} \tag{ 1 }$$

where E_tot is the total energy of the heterostructure, ${E_{{\text{FAPb}}{{\text{I}}_{\text{3}}}}}$ and ${E_{{\text{Pb}}{{\text{F}}_{\text{2}}}}}$ represent the structural energies of the FAPbI₃ and PbF₂ surface, respectively. A_surf is the surface area of the heterostructure.

The perovskite film was fabricated utilizing water-based techniques. We investigated the effect of PbF₂ additives on phase transformation, morphology, optical and electrical properties of the perovskite films. We prepared a series of Pb(NO₃)₂ aqueous solution, with the incorporation of different amounts of PbF₂ (ranging from 0.5 mg ml⁻¹ to 2.0 mg ml⁻¹) into the Pb(NO₃)₂ solution. The solubility of PbF₂ in water is notably limited (0.67 mg ml⁻¹ at 20 °C). Hence, upon introducing PbF₂ at concentrations exceeding 1 mg ml⁻¹, there is a noticeable presence of colloidal PbF₂ particles that are insoluble and dispersed within the solution, as evidenced by dynamic light scattering measurements (figure S1, supporting information). It should be noted that the utilization of PbF₂ as an additive was unfeasible in the conventional organic solvent synthesis, due to the insoluble nature of PbF₂ in organic solvents such as DMF, DMSO, or isopropyl alcohol (IPA).

The impact of the PbF₂ additive on the morphology of the Pb(NO₃)₂ film can be clearly seen in the SEM image in figure S2. The initial spin-coated Pb(NO₃)₂ film displays a porous morphology, which is attributed to the inadequate wetting properties between the Pb(NO₃)₂ solution and the SnO₂ substrate [14]. In contrast, the film containing PbF₂ (1 mg ml⁻¹) exhibits a more continuous coverage. Apparently, the introduction of PbF₂ seeds has influenced the nucleation and growth of Pb(NO₃)₂. This could be because the PbF₂ seeds serve as nucleation sites, which lower the energy barrier for nucleation, facilitating the development of Pb(NO₃)₂ crystals that are more evenly dispersed throughout the film. The water contact angle measurement demonstrates that adding PbF₂ (1 mg ml⁻¹) reduces the contact angle between the Pb(NO₃)₂ aqueous solution and the SnO₂/ITO substrate (without UV-ozone treatment) from ∼57° to ∼53° (figure S2(b)). This improvement in wettability enhances the interaction between the solution and the substrate, ultimately promoting more uniform nucleation. This improved uniformity and continuity of the Pb(NO₃)₂ films can benefit the subsequent perovskite formation step.

Conversion of Pb(NO₃)₂ to the perovskite phase was achieved by immersing the film in an precursor solution of MAI and methylammonium chloride (MACl) dissolved in isopropanol (IPA) for three cycles as described in the experimental section (supporting information). It is noticeable that the conversion rate from Pb(NO₃)₂ film to perovskite was significantly accelerated in the presence of PbF₂. Specifically, the inclusion of 1 mg ml⁻¹ of the PbF₂ additive led to a10 s faster conversion rate, compared to the pristine film without PbF₂ additive. The digital images showing the colour evolution of Pb(NO₃)₂ film immersed in the MAX (I, Cl) precursor solution for varying durations are shown in figure 1(a). It can be seen, that the initially opaque white Pb(NO₃)₂ film (figure S3) transforms to yellow colour after being immersed in the MAX solution for 5 s, indicating the formation of PbI₂. As the incubation period is extended, a dark brown substance, attributed to MAPbI_3−xCl_x, gradually spreads across the substrate. The pristine film transitions to a complete brown film after 30 s of incubation in the MAX solution. In contrast, in the presence of PbF₂ additive, the perovskite formation is achieved at approximately 20 s.

Zoom In Zoom Out Reset image size

We further utilized SEM measurements to observe the morphology evolution of the films. As shown in figure S4, after immersing the Pb(NO₃)₂ film in the incubation solution of MAX in IPA solution for 5 s, the film was transformed into a mixture of aggregates. This can be attributed to the formation of PbI₂ [16]. These aggregates evolved into larger crystals, improving the film's coverage after an additional 10 s of soaking. Concurrently a visible colour change was observed with the film with the appearance of numerous brown areas on the film's surface, indicating the initial formation of the perovskite phase. Accompanying the colour change, the film with PbF₂ exhibited a morphology consisting of larger perovskite crystals, providing better coverage. After 20 s, the PbF₂-containing film displayed the typical morphology of a perovskite layer, with good compactness and uniform grain size. In contrast, the film without PbF₂ still showed numerous pinholes.

An in-situ UV–vis absorption kinetic measurement was employed to investigate the perovskite conversion rate by monitoring the absorbance of Pb(NO₃)₂ films immersed in an MAX solution at 750 nm. The absorption fraction indicates the phase conversion fraction, as shown in figure 1(b). The results clearly demonstrate that the films containing PbF₂ exhibited significantly faster transformation compared to the control films. The conversion fraction increased sharply, reaching approximately 0.6 within 1 min, before slowing down. This deceleration suggests that, at this stage, the transformation process is governed by diffusion kinetics [15]. In contrast, the control film required over 4 min to achieve a similar conversion fraction of 0.6. The crystallization kinetics in this system can be described using the Avrami nucleation and growth theory, which models isothermal solid-state transformation reactions based on nucleation and growth kinetics [41].

According to the Avrami equation: f = 1–exp(−ktⁿ); where k is the rate constant, t is time and n is the Avrami exponent. The rate constants (k) were determined to be 0.20 for the control film and 0.62 for the film containing PbF₂ (figure S5). These values highlight the role of PbF₂ in facilitating nucleation and growth kinetics of the perovskite film in the presence of the PbF₂ additive. It is possible that the presence of F− anions (1.36 Å), which are smaller and more electronegative than iodide anions (2.2 Å), facilitates faster ion transport and crystallization. This is consistent with previous studies which demonstrated that introducing small Cl− anions (1.8 Å) accelerates the conversion of Pb(NO₃)₂ to perovskite by improved diffusion kinetics and Pb²⁺ complexation [16, 25]. Similarly, in our case, the F⁻ anions promote superior diffusion kinetics and enhance the reaction between Pb(NO₃)₂ and MAI.

The final FA_1−xMA_xPbI_3−yCl_y perovskite film was obtained through a cation exchange strategy [42]. To be specific, a solution of FAI in IPA was spin-coated on the MAPbI_3−xCl_x film. Figures 1(c) and (d) present the SEM images of the final perovskite film with and without PbF₂. The pristine perovskite film displays a morphology that lacks homogeneity in terms of grain size, as evidenced by the varying sizes of grains ranging from a few hundred nanometres to over 1 μm. It is also noticed that there is a significant amount of bright particles on the surface of the film, which can be assigned to the presence of residual PbI₂ [43]. In contrast, the film containing PbF₂ exhibits a denser morphology composed of more uniform, larger grains surpassing 1 μm in size with less residual PbI₂. It is possible that the presence of PbF₂ seeds retards the crystal growth during the FA-MA cation exchange stage, thereby facilitating the development of more uniformly structured perovskite crystals with larger grain sizes. We expect that the extremely low solubility of PbF₂ in IPA solvent ensures that the PbF₂ nanoparticles persist and disperse evenly throughout the solid MAPbI_3−xCl_x film, forming physical barriers that slow down the cation exchange, particularly the diffusion dynamic of FA⁺ cations into the film to replace the MA⁺ cations. It also slows down the solvent evaporation processes during the thermal annealing, thus retard the formation of the final FA_1−xMA_xPbI_3−yCl_y perovskite films to achieve films with improved grain size [44]. This is consistent with what was observed in the grain refinement strategy reported previously [23]. Atomic force microscopy (AFM) measurement was conducted to further evaluate the morphology of the fabricated perovskite films. The AFM result (figure S6) also reveals a more uniform film with larger gain size of the film containing PbF₂ additive than the control sample. Furthermore, roughness analysis shows that the film with PbF₂ additive is smoother with the profile roughness factor R_a of 23.3 nm which is significantly lower than that of the control perovskite film (R_a = 29.8 nm). The improvement in the surface smoothness is beneficial for the subsequent deposition of the transporting layer in fabrication of PSCs. Energy-dispersive x-ray spectroscopy measurement was conducted to validate the composition of the perovskite film. The result reveals a notable presence of fluoride anions in the film containing PbF₂, as shown in figure S7. It is worth pointing out, when a higher concentration of PbF₂ (2 mg ml⁻¹) is introduced, the uniformity of the film is reduced, with a significant number of small grains formed on the surface of the film, as shown in figure S8. This phenomenon is attributed to the accumulation of an excessive quantity of insoluble large PbF₂ particles, which consequently damages the morphology of perovskite films.

We then assess the optical properties of the resultant perovskite films by UV–vis absorption and PL spectroscopy. As shown in figure 1(e), there is no noticeable change in the absorption onset and PL peak, indicating the unchanged band gap of the perovskite film in the presence of PbF₂ additive. The optical bandgap was calculated to be ∼1.53 eV based on the Tauc plot (figure S9). It is also observed that the intensity of UV–vis absorption of the perovskite film containing PbF₂ is slightly higher. This can be related to the perovskite morphology with increased compactness and large grain sizes as seen in the SEM images (figure 1(d)) or possibly increased thickness of the film. Surprisingly, the PL emission of the film with PbF₂ additive is much higher, implying that the target film has more efficient radiative recombination or less defects [45], which is in agreement with the improved film morphology. Once more, upon the introduction of an excessive quantity of PbF₂ additive, a noteworthy decline in the PL intensity was detected (figure S10), owing to the existence of numerous defects in the film (figure S8) that can function as non-radiative recombination centres.

The TRPL spectroscopy was employed to investigate the charge carrier recombination dynamic of the perovskite films with and without PbF₂. The TRPL decay curve was fitted with a triexponential function to extract the lifetime of charge carriers, (figure 1(f)). The details of the charge carrier lifetime derived from the fitting are shown in table S1. The control perovskite film exhibited an average charge carrier lifetime of approximately 129.4 ns, while the perovskite film containing PbF₂ had a notably prolonged lifetime of around 147.3 ns. The longer lifetime of charge carriers implies less non-radiative recombination or fewer trap states within the perovskite film containing PbF₂ additive.

The XRD patterns (figure 2(a)) of the perovskite films with and without PbF₂ were obtained to investigate the crystal structure of the materials. The results show that both materials show the cubic phase FA-based perovskite with the preferred orientation of (100) [46]. In addition, no significant peak shift was observed in the XRD pattern, ruling out the possibility of F⁻ anions entering the perovskite crystal lattice. The huge size mismatch between I ion (r_ionic ∼ 206 pm) and F ion (r_ionic ∼ 119 pm) probably prevents the latter from entering the perovskite lattice, as it would induce too much strain. However, F ions can stay on the perovskite surface or grain boundaries through hydrogen and ionic bonding with perovskite's cations [47]. Upon closer examination of the XRD pattern, a notable enhancement in the intensity of the (100) direction peak was observed (figure 2(a)—inset), whereas the intensity of the (210) direction peak remained relatively unchanged (figure S11(a)). This observation suggests that PbF₂ promotes a more favourable growth of the perovskite crystal along the (100) direction. Given that this crystal orientation is highly favourable to charge carrier diffusion, we believe that the enhancement in crystal growth orientation will positively impact charge transport in the resulting PSC devices [48]. It is also noticeable that the full width at half maximum of the (100) peak is decreased with the addition of PbF₂, which is consistent with the increase in the grain size observed in the SEM measurement. A closer examination of the PbI₂ peak at around 12.5° in the XRD pattern confirms a reduction in PbI₂ content in the perovskite film containing PbF₂ (figure S11(b)). Rietveld refinement of the XRD pattern indicates that the PbI₂ content in the perovskite film without PbF₂ is approximately 2.32%, whereas in the film with PbF₂, it is reduced to ∼1.20%. This aligns well with the SEM observations in figure 1(c).

Zoom In Zoom Out Reset image size

XPS was employed to examine the chemical state of the elements on the surface of the perovskite materials. The distinctive XPS signals of N1s, Pb 4 f, and I 3d associated with the perovskite material are identified in the XPS scan spectrum of both the control and the samples with PbF₂ additive (figure S12). A detailed analysis of the Pb 4f XPS spectrum (figure S13) reveals peaks at 138.7 eV and 143.6 eV, corresponding to the spin–orbit splitting of the Pb 4f_7/2 and 4 f_5/2 components in the perovskite material [49]. Additional peaks at 137.8 eV and 142.6 eV are attributed to the presence of lead(II) oxide (PbO) [49]. In the perovskite film with PbF₂ additive, the Pb 4f_7/2 and 4f_5/2 peaks shift to higher binding energies (138.8 eV and 143.7 eV), accompanied by a significant reduction in the PbO peaks (figure S13). Note that the binding energy calibration was performed using the C–C peak at 284.8 eV. The shift to higher binding energy suggests that the PbF₂ additive induces a change in the surface bonding of Pb. It is plausible that the higher electronegativity of F^- compared to I^- leads to a stronger ionic bonding with the Pb²⁺ cation, consequently causing the XPS peak to shift towards a higher energy state [31]. Additionally, there are new peaks appear at 139.5 eV and 144.3 eV, which align with the characteristic Pb²⁺ peaks of PbF₂, confirming presence of PbF₂ on the perovskite surface [50]. Consistent with that, the perovskite film with PbF₂ also shows significant amount of F⁻ in the F 1 s XPS spectra (figure S13). The reduction in PbO suggests that PbF₂ acts as a protective barrier, mitigating lead oxidation and improving film stability.

As a result, a noticeable change was also observed in the energy band position of the perovskite as revealed by the UPS spectrum. The presence of PbF₂ leads to an increase in the cut-off energy band, while maintaining the maximum energy state of the valence band (figure 2(b)). An energy band diagram was constructed based on the combined UPS and UV–visible absorption measurement (figure 2(c)). It is found that the energy band of the perovskite film containing PbF₂ was elevated by 0.11 eV to a more favourable position for electron transport to the adjacent SnO₂ layer.

ToF-SIMS was utilized to examine the distribution of fluoride cation within the devices. The profiles of cations and anions in the perovskite/ITO layers are presented in figures 2(d)–(f). The result indicates the absence of F across the pristine perovskite film (figure 2(d)), while a notable presence of F in the perovskite film made from a small amount of PbF₂ (0.5 mg ml⁻¹) (figure 2(e)). Furthermore, an increase in the F fragment signal at the top of the perovskite layer suggests a higher F concentration on the perovskite surface compared to the bulk perovskite layer. Interestingly, the perovskite film containing a higher PbF₂ additive concentration (1 mg ml⁻¹) showed a substantial rise in the F anion content at the bottom of the perovskite film adjacent to SnO₂/ITO substrate (figure 2(f)). It indicates an accumulation of PbF₂ at the perovskite/substrate interface. This phenomenon is reasonable due to the pronounced affinity between F and SnO₂ in the ITO, which can enhance the charge transport between the perovskite and SnO₂ layer, as demonstrated in our previous study [51].

To clarify the role of the PbF₂ additive in the FAPbI₃ thin film crystallization, DFT calculations were conducted to assess the interlayer potentials (φ, eV Å⁻²) of FAPbI₃/PbF₂ heterostructure with different facets (figures 3(a) and (b)). The calculated results reveal that the φ value for the FAPbI₃ (100) phase/PbF₂ is −0.103 eV Å⁻², which is significantly larger compared to the FAPbI₃ (110)/PbF₂ interface (−0.028 eV Å⁻²). This much lower interlayer potential indicates a preferential interaction between the PbF₂ and the FAPbI₃ (100) phase, thereby facilitating the growth along the [100] direction. This is consistent with the XRD patterns of the perovskite containing PbF₂ as shown above (figure 2(a)).

Zoom In Zoom Out Reset image size

Based on the above analysis, we propose that the presence of PbF₂ in the perovskite induces an anchoring effect to regulate the crystallization of perovskite and improve the interface between the SnO₂ and the perovskite layer [52]. As illustrated in figure 3(c). The F⁻ anions, characterized by their high electronegativity, accumulate at the SnO₂ surface (as confirmed by ToF-SIMS measurements). These F⁻ anions can form strong hydrogen bonds with the ammonium groups in the FA⁺ cations [53]. The analysis of Fourier transform infrared spectroscopy of FAI solutions in water (figure S14), with and without PbF₂, confirms the significant shifts in several characteristic FAI peaks in the presence of PbF₂ including NH₂ rock (1073 cm⁻¹ → 1078 cm⁻¹; 1073 cm⁻¹ → 1078 cm⁻¹; 1341 cm⁻¹ → 1344 cm⁻¹), NH₂ bend (1619 cm⁻¹ → 1630 cm⁻¹) and NCN stretch (1719 cm⁻¹ → 1725 cm⁻¹) while the NH₂ stretch is almost unchanged (3360 cm⁻¹ → 3361 cm⁻¹). The shift to a higher wavenumber suggests the formation of stronger bonds involving the ammonium group in the FA⁺ cation, which are attributed to hydrogen bonding interactions with F⁻ anions [53]. As a result, the F⁻ anions anchor onto the FA⁺ cations during the crystallization of the perovskite, facilitating the assembly of [PbI₆]⁴⁻ octahedra along the (100) direction due to the lower interlayer potential. This anchoring mechanism accelerates the growth of well-crystallized FAPbI₃ perovskite with a preferred orientation, offering substantial advantages for interlayer charge diffusion [54].

We then assessed the performance of these films in PSCs which were assembled with the architecture of ITO/SnO₂/Perovskite/Spiro-MeOTAD/Au (figure 4(a)). A SEM image of the cross-sectional view of the PSC device is shown in figure 4(a). Both the perovskite films with and without PbF₂ additive exhibit similar thicknesses of approximately 370–400 nm (figure S15). However, the film containing PbF₂ demonstrated a more uniform thickness and larger grain sizes, with single grains extending from the electron transport layer to the hole transport layer. This improved grain connectivity is likely to enhance charge transport within the device [55].

Zoom In Zoom Out Reset image size

We optimized the concentration of the PbF₂ additive by fabricating PSCs using the perovskite films containing various concentrations of PbF₂ (0.5 mg ml⁻¹, 1 mg ml⁻¹, and 2 mg ml⁻¹). The statistical performance parameters of these devices (figure S16) shows that a concentration of 1 mg ml⁻¹ PbF₂ yielded the best device performance, which is consistent with improved morphology and radiative recombination shown above (figure 1). Further analysis of the characteristic performance parameters shows that, while the PbF₂ additive has a slight impact on improving short-circuit current density (J_sc), it has a more pronounced effect on enhancing the open-circuit voltage (V_oc) and the fill factor (FF) of the devices. The maximum PCE of the control device without PbF₂ was 16.39%, with a V_oc of 1.03 V, a J_sc of 22.53 mA cm⁻², and an FF of 70.79%. In contrast, the device with 1 mg ml⁻¹ PbF₂ additive produced a PCE of 18.10%, with a V_oc of 1.08 V, a J_sc of 22.17 mA cm⁻², and an FF of 75.61%. Additionally, the device with PbF₂ showed a reduction in hysteresis (figure S17), indicating a more balanced charge extraction at the two critical interfaces of perovskite/SnO₂ and perovskite/spiro-MeOTAD.

The water-based PSC fabrication process is a promising approach for scalable manufacturing due to its low-cost and eco-friendly nature. To demonstrate the scalability of the process, we fabricated PSCs with PbF₂ additive on a 2.5 × 2.5 cm² substrate using conventional spin-coating method (figure S18). Device performance was measured over a 1.0 × 1.0 cm² active area using a shadow mask. It is found the device performance was significantly lower than that of devices with smaller active area, with the highest efficiency of 9.93%. The lower device performance is due to the significantly reduced V_oc and FF compared to smaller-area counterparts (table S3). This is likely due to non-uniform crystallization of the perovskite film over the larger area, which negatively impacted performance. To address this limitation, further optimization of processing conditions and the development of scalable deposition techniques, such as slot-die coating or blade coating, are necessary. These improvements can be the focus of future work to enhance the performance of large-area water-processed PSCs.

The enhanced performance of the PSC devices containing PbF₂ owing to the significant improvement in the V_oc and FF can be attributed to the improved morphology of the perovskite film, coupled with a decrease in trap states within the material, the improved crystal orientation as well as energy band alignment as shown above. The values of integrated J_sc derived from the external quantum efficiency (EQE) spectra of the corresponding PSC devices are 21.22 mA cm⁻² and 20.98 mA cm⁻² for the pristine and target devices, respectively (figure 4(c)). These results align well with the J_sc values obtained from the J–V measurements of the devices. Additionally, measurement of stabilized power output (SPO) also shows a higher SPO for the PSC that incorporates PbF₂ additive (figure 4(d)) with a stable output of 17.8% after 100 s, whereas the SPO of the pristine device reaches 16.0% within the same timeframe.

To obtain a deeper understanding of the mechanism observed in the PSC devices with PbF₂ additives, we investigated the ideality factor (m) of the devices, which is obtained by fitting the plot of V_oc of the devices as a function of light intensity (figure 4(e)) according to the equation of ${V_{{\text{oc}}}} = \frac{{mkT}}{q}{\text{ ln}}({I_{{\text{light}}}}/{I_0}$), in which, k represents the Boltzmann constant, T is temperature, and q is the elemental charge. The result reveal that the m values for the device are 1.23 and 1.14 for the pristine and target devices, respectively. A m value closer to unity (m = 1) indicates the recombination is dominated by the contact interfaces rather than the local recombination in the bulk absorber layer, while a higher value of m is accounted for the nonideal behaviour in the current–voltage due to the recombination in the bulk absorber layer [56]. The decrease in m indicates a reduction in trap-induced recombination within the perovskite layer. This is in agreement with the above PL and TRPL measurement (figures 1(e) and (f)) showing reduced nonradiative recombination in the perovskite layer containing PbF₂.

In order to further validate the impact of the PbF₂ additive on property of trap defects of the perovskite films, we determined the defect density in the perovskite film through space charge limited current measurements using an all-electron transport device structure composes of ITO/SnO₂/Perovskite/C₆₀/Bathocuproine (BCP)/Ag (figure 4(f)—Inset). The defect density was determined by fitting I–V plot of the device under dark conditions (figure 4(f)) according to the equation N_t = $2\varepsilon {\varepsilon _0}\frac{{{V_{{\text{TFL}}}}}}{{e{L^2}}}$; where represents the relative dielectric constant of the perovskite (∼3.2, for organic material), ₀ denotes the vacuum permittivity, e is the electron charge, L is the thickness of the perovskite layer, and V_TFL represents the trap filled limit voltage. The V_TFL acquired for the samples with PbF₂ additive is 0.48 V, notably lower than that of the pristine device (0.62 V). The corresponding density of trap states N_t is determined to be 1.04 × 10¹⁷ (cm⁻³) for the pristine, and 8.02 × 10¹⁶ (cm⁻³) for the target device, respectively. These results further validate the efficacy of PbF₂ additive in reducing the defective traps present in the perovskite film.

Impedance spectroscopy (IS) was utilized to gain insights into the electrochemical processes occurring within the PSC devices. The impedance plots exhibit two distinct arcs (figures 4(g)–(i)), each corresponding to different frequency range. The impedance spectra were fitted using the equivalent circuit shown in the inset of figure 4(g). The series resistance R_s represents the resistance of the transparent conductive oxide, contact layers, and wires. The equivalent circuit also includes two capacitive (C_g and C_s) and two resistive (R₃ and R₁) elements. According to the established interpretation, C_g is associated with the dielectric response of the perovskite layer, while C_s corresponds to surface charge accumulation at the interfaces [57]. The capacitive response at high frequency range (>1 kHz) is dominated by C_g, whereas C_s is prominent at the low-frequency (∼1 Hz) region. The total surface recombination resistance (R_rec) is derived from the series connection of R₁ and R₃ (R_rec = R₁ + R₃). Figure 4(g) shows the IS spectra measured in dark for the pristine and target devices. The corresponding fitted parameters are listed in table S2. These results provide insights into charge accumulation and interfacial processes at the device interfaces. Under dark condition, the surface charge accumulation C_s of the device with PbF₂ additive (9.15 × 10⁻⁹ F) was significantly lower than that of the pristine device (5.74 × 10⁻⁸ F). This capacitance is produced due to accumulation of ions at the contact interface [58]. This lower C_s suggests a more efficient charge extraction at the interface of perovskite and SnO₂ layers.

We further measured the impedance spectrum of the PSCs under different illumination intensity at short circuit conditions (figures 4(h) and (i)). It is observed that the C_s value increases proportionally with light intensity, while the resistance decreases inversely. The low-frequency resistance R₁ showing larger values than the high-frequency R_3, consistent with the electrochemical characteristics of lead halide PSCs [59]. From the fitted parameter from the EIS spectra (table S4), it is evident that both under dark and working conditions, the device with PbF₂ exhibits a significantly higher total recombination resistance than the pristine device. This indicates that recombination of electrons and holes is more suppressed in the PbF₂ based device, allowing more photogenerated charge carriers to be successfully extracted. This behaviour suggests a lower density of defects or trap states in the perovskite or at the interfaces, as these are primary sites for non-radiative recombination. In addition, the device with PbF₂ additive also has significantly lower R_s resistance, which can be assigned to the improved conductivity of the SnO₂ layer when F⁻ anion incorporated and passivated the defects on the surface of SnO₂ as observed in our previous report [51]. This reduced R_s should contribute to the improved FF of the PSC devices (figure 4(b)).

The surface characterization above confirms the accumulation of fluorine on the top surface and the buried interface (bottom surface) of the perovskite film. Fluorine is widely recognized for its inherent hydrophobic properties, which function as a protective barrier shielding the underlying perovskite layer from potential detrimental effects posed by environmental elements such as moisture [60]. To validate this assertion, we measured the water contact angle on the surface of the perovskite layer. As shown in figures 5(a) and (b), the pristine perovskite film demonstrated a commendable contact angle of 77.94°, while the film containing PbF₂ exhibited a further increase of the contact angle to 82.56°. This confirms the enhancement in the hydrophobic nature of the perovskite film in the presence of PbF₂. We further evaluated the resilience of the fabricated PSCs against various environmental factors. Figure 5(c) illustrates the operational stability of the device without encapsulation under 1 sun illumination at a constant bias of 0.8 V. The result shows device incorporating PbF₂ exhibited enhanced stability, with the PCE retention rate of approximately 76% after 320 min of continuous operation. In contrast, the PCE retention rate of the pristine device dropped to around 62%. It is worth noting that throughout the duration of the operational testing, the temperature of the device reached ∼65 °C.

Zoom In Zoom Out Reset image size

As is expected from the contact angle measurement, the device with PbF₂ also exhibited better stability under humid condition. The PSCs device without encapsulation were kept in a climate chamber with a controlled relative humidity of 60% at temperature of 30 °C. A batch of 12 cells for each condition was employed in the stability test. The change in the average PCE of the devices is shown in figure 4(d). The device with PbF₂ retained PCE of ∼39.3% after 120 h of moisture soaking, which surpassed that of the pristine device with retention rate of only 19.2% under the same testing condition. This is consistent with the better water-repellent of the PbF₂-containing perovskite films. Further testing the stability of our water-processed PSCs, including thermal stability and UV radiation stability (figure S20) also show that the PSCs with PbF₂ additive exhibit better thermal stability and UV radiation than the pristine device. Over the testing duration of 72 h at temperature of 65 °C, the devices containing PbF₂ retained above 75% of their initial PCE on average while the pristine devices only retained ∼60% of the original PCE. In the UV radiation stability test, over the testing duration of 70 h, the average retention rate of the PSCs with PbF₂ additive was 95.9%, surpassing that of the pristine device with retention rate of 93.4%. The result further confirms the beneficial effect of PbF₂ on enhancing the structural integrity of the fabricated perovskites.

The PSCs also exhibited distinctive visual differences after undergoing the moisture stability test. As shown in figure S19, images of the PSC devices after three days in humid conditions (relative humidity ∼60%) show that the control device without PbF₂ exhibited clear signs of degradation of the perovskite materials, with the appearance of lots of yellow areas after the environmental stress. In contrast, the PSC device with PbF₂ additives still maintained a better coverage of the characteristic brownish perovskite materials, confirming enhanced material stability against moisture.

Based on the aforementioned analysis, along with the evident enhancement in the performance and stability of PSCs, we believe that the PbF₂ in water solution exists in two distinct states: firstly, as solute PbF₂ that dissociates into Pb²⁺ and F⁻ ions, and secondly, as insoluble PbF₂ particles dispersed in the solution and eventually accumulated at the bottom and the top surface of the perovskite layer. The perovskite formation process is illustrated in figure 6. In the process with PbF₂ additive, the PbF₂ particles are dispersed within the Pb(NO₃)₂, functioning as nucleation promoters a higher density growth of Pb(NO₃)₂. While without PbF₂, a discontinuous film of Pb(NO₃)₂ is obtained due to lack of wettability between Pb(NO₃)₂ and SnO₂ substrate. During the crystallization of perovskite, without PbF₂, the perovskite film grows uncontrollably, and finally results in a perovskite film with a significant amount of anion and cation vacancies.

Zoom In Zoom Out Reset image size

With the presence of PbF₂, the additive coordinates the FA⁺/MA⁺ cations through the hydrogen bonds and anchors them to the SnO₂ surface where insoluble PbF₂ accumulates. This anchoring effect promotes an oriented growth of the perovskite grains with reduced defects at the buried interface of the perovskite film [52]. Eventually, the perovskite films with well-oriented and large grains are obtained. Additionally, the solute PbF₂ supplies a significant amount of F anions that passivated the grain boundaries of the perovskites, which is crucial for minimizing recombination losses associated with grain boundaries in the perovskite film. This not only enhances the electronic properties of the perovskite film but also provides a layer of protection to enhance the stability of the perovskite film, leading to more superior PSCs device fabricated using water processing strategy.

In this study, we have demonstrated for the first time the effect of PbF₂ additives in increasing both PCE and stability of water-processed PSCs. Characterisation of the perovskite films with and without PbF₂ has shown that the PbF₂ additives serve multiple purposes including: (1). They provide nucleation sites that lead to uniform grain formation and improved film morphology. (2). The F⁻ anion from PbF₂ helps bridge the contact between the perovskite and the electron transport layer surface, guiding the crystal growth of the perovskite film. (3). This anchoring effect, combined with the passivation provided by F⁻ anions, reduces surface trap states and enhances the optical and electrical properties of the perovskite grains. As a result, the optimized PSC device incorporating PbF₂ additive achieved a PCE of 18.1%, significantly outperforming the reference device without the additive, which achieved a PCE of 16.3−. Additionally, the devices with PbF₂ exhibited enhanced operational stability and greater resilience against moisture. This research highlights the potential of inorganic fluoride salt additives in advancing cost-effective, environmentally friendly processing methods for high-performance PSCs.

The advancement towards environmentally friendly PSC technology requires the development of fabrication process utilizing green, nontoxic solvents. While the use of water as solvents for PSCs fabrication aligns with the goals of sustainable and scalable manufacturing, the performance gap between water-processed and toxic organic solvent-processed PSCs remains a challenge. In this work, our findings demonstrate the potential of fluoride-based additives in overcoming this gap. The additive is effective in not only improving the perovskite crystal quality but also the defect passivation, leading to significant enhancement of water-processed PSCs performance. We expect further breakthroughs in water-based perovskite fabrication can be obtained through the exploration of diverse halide-based additives and other hybrid anions to precisely control crystallization dynamics and surface passivation. Additive engineering will be a key to mitigate issues such as redissolution and recrystallization during water-based processing of PSCs, thereby improving film quality and reproducibility. Additionally, integrating advanced surface treatments methods and incorporating small molecular organic ligands into the process could further enhance the efficiency and long-term stability of water-processed PSCs. These advancements will pave the way for the development of green, industrial manufacturing of PSCs within the next decade.

This work was supported by the Australian Research Council (ARC) Linkage Project (LP210100217). The acknowledgement goes to the technical support from Central Analytical Research Facility (CARF) at Queensland University of Technology (QUT).

The authors declare no conflict of interest.

Minh Tam Hoang: Conceptualization, experimental design, validation and data analysis, manuscript writing; Junxian Liu: DFT modelling, data analysis and manuscript writing; Yang Yang: ToF-SIMs measurement, solar cells testing and data analysis, Maciej Klein: TRPL, IPCE measurement, Wei-Hsun Chiu: SEM measurement and data analysis; Yongyue Yu: data analysis; Ngoc Duy Pham: Experiment design, and manuscript writing; Paul Moonie: project supervision; Ajay Pandey: TRPL, IPCE measurement supervision and manuscript writing; Liangzhi Kou: Theoretical modelling supervision and manuscript writing; Hongxia Wang: Research concept initiator, experiment design, data analysis, project supervision and manuscript writing.