Enhanced Stability of MAPbI3 Perovskite Films with Zirconium Phosphate‐Phosphonomethylglycine Nanosheets as Additive

Methylammonium lead iodide perovskite (MAPbI3) is today considered the most promising component for highly efficient third generation solar cells. However, the lifetime of the solar devices is strongly affected by the stability of the MAPbI3 films toward humidity, UV irradiation, and temperature. The search for efficient protective additives to be used for building up composite perovskite films with enhanced stability is a topic of great interest in the scientific community. In the present paper, a layered zirconium phosphate‐phosphonate based on N,N‐phosphonomethylglycine, exfoliated in thin nanosheets (NS), as additive for the stabilization of MAPbI3 crystalline films toward humidity, UV‐irradiation, and temperature changes is applied. Notably, the additive is extremely efficient in preventing degradation of the perovskite film, preserving the optical and structural properties, and avoiding the phase transitions normally observed due to temperature increase.


Introduction
Today more than ever is essential for the mankind to search for renewable energy sources to fulfill the global energy demands and mitigate climate changes. In this sense, solar energy is one of the most promising renewable energy sources. In the last 10 years, perovskite solar cells (PSCs) have attracted much attention as a promising innovative photovoltaic technology, possibly contributing to a large scale solar energy production, thanks to their high efficiency and low production costs. [1] Despite their attractive potential, the main drawback of PSCs concerns the durability of the perovskite material, due to instability toward many external factors, such as humidity, oxygen, UV light, and temperature fluctuations, [2,3] thus resulting in lifetime of PSCs of barely 1 year, which do not stand a comparison with silicon solar cell, whose lifetime reaches easily 30 years. [1,4] Therefore, the stabilization of the perovskite material is clearly one of the most crucial issues to be addressed in order to facilitate the commercialization of this technology. More specifically, the aim is to intervene to reduce the chemical instability limiting the formation of defects present in the perovskite, both of which can speed up the process of degradation induced by external factors. [5] Several different strategies have been tuned up to meet this goal. One of these focuses on the incorporation of additives directly in the perovskite precursors solution to obtain final materials with enhanced properties. These include, for example, polymers, [6,7] cyclodextrins, [8] fullerenes, [9,10] and inorganic acids. [11,12] In a recent work, Xie et al. reported the use of some phosphonic acids, like 3-Phosphonopropionic acid, as additives and fabricated a PSC with a power conversion efficiency (PCE) of about 21% and enhanced device stability due to the presence of PO(OH) 2 group of the additive, which can establish strong hydrogen bonds, passivating shallow point defects and possibly blocking ion migration. [13] A different approach consists in coating the perovskite layer with specific materials, to protect it from external factors while maintaining unchanged its properties. Specifically, a coating layer deposited over the perovskite can act as a barrier by Methylammonium lead iodide perovskite (MAPbI 3 ) is today considered the most promising component for highly efficient third generation solar cells. However, the lifetime of the solar devices is strongly affected by the stability of the MAPbI 3 films toward humidity, UV irradiation, and temperature. The search for efficient protective additives to be used for building up composite perovskite films with enhanced stability is a topic of great interest in the scientific community. In the present paper, a layered zirconium phosphatephosphonate based on N,N-phosphonomethylglycine, exfoliated in thin nanosheets (NS), as additive for the stabilization of MAPbI 3 crystalline films toward humidity, UV-irradiation, and temperature changes is applied. Notably, the additive is extremely efficient in preventing degradation of the perovskite film, preserving the optical and structural properties, and avoiding the phase transitions normally observed due to temperature increase.
restricting the diffusion of oxygen and moisture and possibly it should protect also from UV light. For example, an ultrathin layer of Al 2 O 3 has been deposited on top of the perovskite layer by atomic layer deposition (ALD), prolonging PSCs lifetime with almost no negative effect on device performance, with 90% of PCE preserved after storage in air for 24 days. [14] Interestingly, when these materials are deposited as nanosheets (NS), they rely on the phenomenon of quantum tunneling to maintain good electron transport properties, while possibly providing moisture protection to the perovskite layer depending on their hydrophobic nature. Therefore, the choice of an additive or a coating agent capable of improving the stability of the perovskite against detrimental environmental conditions with deposition of thin layers to ensure the electronic conduction by tunnel effect, is a highly desirable goal.
Zirconium phosphonates are a class of compounds already used for a large number of applications, ranging from intercalation, [15] proton conductivity, [16] fillers for nanocomposites, [17] and catalysis. [18] In particular, a mixed phosphate-phosphonate based on N,N-bis(phosphonomethyl)glycine (or glyphosine) (L) and with formula Zr 2 (PO 4 )(L) 2 •H 2 O (hereafter ZPGly) [19] was recently used as support for metal nanoparticles in heterogenous catalyst. [17][18][19][20][21][22] The prolonged contact with amine solution (such as propylamine, PrNH 2 ) leads to facile exfoliation of the compound thus forming NS of about 3 nm, which is compatible with the stacking of 2 or 3 single layers. [20] Given such desirable properties of ZPGly, in this work, we investigate the possible stabilization of a MAPbI 3 perovskite, deposited either on glass or silicon substrate, both by coating ZPGly NS as thin layer and by including them as additives directly in the perovskite precursors solution.
The detailed structure of ZPGly was already reported elsewhere. [19] It consists in packing of complex hybrid layers of about 15 Å of thickness, built from the connection of ZrO 6 octahedra, PO 4 tetrahedra, and PO 3 C phosphonate tetrahedra belonging to the glyphosine ligands (Figure 1).
The compound also owns three acid protons per unit formula attributable to the carboxylic groups and to POH groups. Due to the presence of such acidic protons, the calculated ion exchange capacity (IEC) of the compound is 3.67 meq g −1 . Moreover, the carboxylic and amino groups can efficiently coordinate transition metal ions, having a high affinity toward them: [19] in this way, the presence of COOH and PO 3 H 2 free moieties can result in reduced toxicity associated to lead release, coordinating Pb 2+ ions when perovskites degrade.
ZPGly NS stable dispersion are obtained upon exfoliation of the ZPGly crystalline precursor in N,N-dimethylformamide (DMF) and then used as coating agent for crystalline MAPbI 3 perovskite deposition or mixed with perovskite precursors solution before crystallization. The samples are tested throughout several experimental techniques, i.e., X-ray powder diffraction (XRPD), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), scanning electron microscopy (SEM), UV-vis, and luminescence, before and after fast aging treatments in terms of temperature, UV irradiations, relative humidity (RH), and addition of an I − /I 3 − electrolyte. The addition of Zr phosphonate NS results to strongly enhance the MAPbI 3 stability in all the experimental conditions.

Results and Discussions
Two different procedures were carried out in order to build the composite by using ZPGly NS gel both as coating agent or by using it as a physical mixture with the MAPbI 3 solution (see the Experimental Section). Samples, hereinafter, are named as follow: the crude MAPbI 3 , PVK, MAPbI 3 coated with ZPGly NS, PVK@ZPGly, and MAPbI 3 mixed with ZPGly gel, MIX.
XRPD patterns of the PVK, of the coverage PVK@ZPGly, and of the MIX are shown in Figure 2 The presence of the ZPGly phase does not interfere with the structure and crystallinity of perovskite crystals. The presence of weak reflections in the PVK@ZPGly sample, at ≈5° is evidence of the presence of exfoliated ZPGly (see XRD pattern in S1, Supporting Information) Figure 3 shows the SEM images and the energy-dispersive X-ray spectroscopy (EDX) of the samples. PVK a) has grown in the form of highly aligned fibers made by crystals with tetragonal structure. The growth in form of fibers is consistent with the data already reported in literature. [23] Figure 3b shows a section of PVK@ZPGly composite and from EDX analysis the interface between the two materials can be easily seen. On the contrary, in the MIX sample ( Figure 3c) the perovskite and ZPGly are mixed evenly, as confirmed by EDX analysis which show a homogeneous distribution of the Zr and Pb in the sample. Additional SEM images are also reported ( Figures S2-S4, Supporting Information) TEM images were also collected by dispersing the three samples in EtOH before dropping them on the copper grid precoated with a Formvar film. Despite the dispersion in the solvent leads to an overall separation of the two components, the ZPGly nanosheets are clearly visible as thin platelets of 20-200 nm diameter. From TEM it was not possible to measure the lateral thickness of the NS. Perovskite prismatic nanocrystals are also visible as second component both in MIX and in the PVK@ZPGly systems ( Figure S5, Supporting Information). After preparing all the samples, the stability of the protected perovskites against some factors known for threaten the stability of the perovskite alone, [24] namely humidity, UV radiation and temperature was tested. In addition, we conducted tests to check the stability toward redox-active electrolyte I − /I 3 − as reported in the Supporting Information. IR spectra of PVK, ZPGly, PVK@ZPGly, and MIX were also collected and reported in Figure S6 (Supporting Information). When ZPGly meets PVK (PVK@ZPGly and MIX) the PO stretching is not detected, but remarkable peaks are observed in the 1750 and 2750 cm −1 region, which could be tentatively assigned to OPOR modes. The appearance of these signals supports the establishment of interactions between ZPGly and PVK. More importantly, differently from what observed for neat PVK, the OH stretching (3489 cm −1 ) is not observed for PVK @ZPGly, confirming the ability of Zr 2 (PO 4 )(L) 2 ·H 2 O to protect the perovskite layer from the humidity adsorption.

Relative Humidity (RH) Stability
The samples were placed in a desiccator at 75% RH and were monitored through recording of XRD spectra over time.  The degradation of PVK, as a result of the interaction with humidity, is visible to the naked eye (see pictures of the samples in Figure 4A) by the evident color changes.
The evident degradation of the PVK sample after 1 day in the desiccator can be observed. XRPD patterns of the sample through time are also shown in the figure. The trend over time of the patterns shows the decrease of the intensity of the peak belonging to the pristine perovskite, notably the (001) and (002) at 14.02° and 28.34° 2θ, respectively. On the other hand, some other peaks appeared and increased over time, specifically those at 7.93°, 8.42°, 10.46°, and 16.01°: these peaks cannot be assigned to either the MAPbI 3 or PbI 2 , but possibly to an intermediate phase. Notably, several studies have been reported in literature about the mechanism of degradation of perovskite due to humidity resulting in the formation of PbI 2 . According to these studies, [25,26] the degradation starts with the formation of a hydrate phase of the perovskite with structural formula (CH 3 The structure of this phase was characterized by Vincent et al. and it consists of an assembly of PbI 6 4− octahedra interspersed with CH 3 NH 3 + cations and water molecules, with each water molecule forming two hydrogen bonds with two CH 3 NH 3 + cations. [27] (CH 3 NH 3 ) 4 PbI 6 ⋅2H 2 O is a pale yellow crystalline solid and its XRPD pattern is in line with those we show in Figure 4A. Interestingly, at this level the degradation of perovskite is somewhat reversible: by removing the PVK sample from the desiccator, it is possible to see the yellow sample turning black again. To verify the reversibility of the reaction, the PVK sample was placed outside the desiccator after 42 days of storage and left at ambient conditions for 10 days. From the XRPD spectra ( Figure S7, Supporting Information), we can see that the peaks belonging to the perovskite, that are basically absent from the spectrum of the sample just took out of the desiccator (black line), grow over time, as it is expected since the RH is lower in ambient conditions. Whereas, the peaks of the hydrate phase decrease, till the point of being completely absent in the case of the sample kept at ambient humidity for 10 days (blue line). Therefore, the PVK seems to regenerate but the intensity of the peaks is lower probably due to a partial amorphization.
In contrast to the PVK, the sample PVK@ZPGly did not show any color change, and it could indicate that the sample has not undergone any degradation. As a confirmation, XRPD pattern are shown in Figure 4B. From the XRPD patterns, the crystalline structure of the perovskite was preserved and no other peaks associated with the hydrate phase or with PbI 2 appeared. All the intensities of peaks are comparable meaning that no amorphization occurred, contrary to the case of PVK.
Also, in the case of MIX, in images ( Figure 4C) the sample does not seem to have suffered any degradation and the XRPD patterns recorded over time confirm it.

UV Irradiation Stability
It has been extensively reported in literature that UV radiation induces degradation of MAPbI 3 perovskite resulting in the formation of PbI 2 . [27] The test of stability was then carried out, following a previous work by Kajal et al., and exposing the samples to direct UV irradiation at a wavelength of 392 nm (where perovskite crystals absorb, see below) with an intensity of ≈350 mW cm −2 . [28] During the experiment, temperature and humidity were maintained at ambient conditions. The samples were monitored for 12 h by recording XRPD patterns. Figure 5A (right box) shown mere PVK at various time of UVradiation exposure.
It is evident from both the images and the spectra that the sample had undergone severe degradation after only 12 h. Peaks of the perovskite, in XRPD patterns ( Figure 5A, left box), decrease gradually with the prolonged time of irradiation, instead only after 1 h the peaks belonging to PbI 2 , particularly the characteristic peak at 12.64°, become visible and increase. After 12 h of exposition to UV radiation, the PVK crystals appeared almost completely degraded to PbI 2 .
Contrary to the mere PVK, the PVK@ZPGly remains stable for all the 12 h as it is shown from both the images and XRPD patterns ( Figure 5B) except for the formation of the hydrate phase (pattern c, Figure 5B) which disappear after 3 h.
In XRPD pattern reported in Figure 5A, the appearance of peaks characteristic of PbI 2 can be detected and their increase over time indicates degradation of PVK. But comparing the degradation undergone in this sample with the degradation observed in the neat PVK sample, a different level of degradation can be hypothesized. While the PVK is almost completely degraded, the MIX seems to be degraded in a lower degree, as yellow color is observed only on the external part of the crystals. This observation is also confirmed by the variation of relative intensity of the peaks belonging to PVK and PbI 2 ( Figure 5C, right). After an initial decrease after 4 h, the intenity ratio between the two phases remains unchanged over the entire time of measurement (20 h).

Temperature Stability
The thermal behavior of the mere and protected perovskite was studied through TGA and DSC. First, the TGA heating curves of PVK, ZPGly@PVK, and MIX are shown, see Figure 6.
The DSC curve of PVK shows a peak around 55 °C corresponding to the exothermic phase transitions from tetragonal to cubic phase. [29] Instead, the DSC curves of the protected perovskite do not show the phase transition peak suggesting that ZPGly NS, either as coating or mixed phase, might stabilize the tetragonal phase, since the XRPD patterns confirm that the perovskites, mere and protected, were in tetragonal phase before the DSC analysis. Avoiding the phase transitions, and therefore structural instability and lattice distortion caused by temperature fluctuation, can lead to an enhancement of performance of the perovskite [30] (Figure 6, left box).
In order to get more insight in this hypothesis, we collected variable temperature XRPD (VT-XRPD) of the three samples from 25 to 70 °C (5° min −1 heating and 10 min equilibration) and back to 25 °C. VT-XRPD patterns of the 20-30° 2-theta region are shown in Figure S8 (Supporting Information). While the PVK undergoes the reversible phase transition from cubic to tetragonal over 60 °C, detected by the disappearance (and the successive appearance upon cooling) of the (211) reflection, [31] the MIX does not undergo any transition as the (211) peak remained present all over the heating/cooling cycle. Unfortunately, we do not report the results for ZPGly@PVK because the phosphonate coating concealed the peaks of perovskite and the phase transition could not be evidenced.
By comparing the TGA curves ( Figure 6, right box), no remarkable differences can be seen and the pattern is basically the same for all three samples. The major weight losses are at around 220 and 430 °C and could be assigned to the loss of MAI and PbI 2 , respectively. [27] There is only a little weight loss around 125 °C of about 10% already present in the curve of the PVK sample and absent in the other two curves.
This weight loss can be assigned to the DMF loss [29] and its absence in the curves of the composites means that ZPGly could facilitate the evaporation of DMF during the synthesis, while in the PVK sample a residue of DMF remains.

Optical Properties of Perovskite Film
UV-vis reflectance spectra of perovskite film in absence and presence of ZPGly (prepared with the two methods described below), has been collected immediately soon after the preparation and after 24 h later exposure to atmospheric conditions (Figure 7). The optical properties of the PVK@ZPGly film is not modified after 24 h exposure to air, while PVK and the MIX film present a detectable alteration in the 250-550 nm region.
The spectra of the samples have been registered after taking the samples in environment with the 75% of RH for 35 days (Figure 8). The PVK films show an impressive alteration of the electronic properties, as evidenced by the appearance of absorption bands centred at 380 and 425 nm (see black line in Figure 8), attributed to the formation of PbI 3 − and PbI 4 2− , respectively. [24,32,33] This observation documents the alteration of the optical and electronic behaviors of the neat PVK films due to its chemical instability caused by humidity. The samples with ZPGly (PVK@ ZPGly and MIX) do not show such an evident modification in the UV-vis spectra, confirming the ability of the inorganic coating to preserve the optical properties of the perovskite, protecting it from damages caused by the chemical transformations induced by humidity.

Conclusion
A layered zirconium phosphonate based on glyphosine as building block was employed as additive to MAPbI 3 crystalline films deposited on silicon or glass support. The advantage of using such material resided in its high versatility to be obtained as stable dispersion of NS exposing free carboxylic group, having good affinity toward Pb ions and being able to efficiently protect the perovskite deposition without interfering with its structural and optical properties. A second synthetic method was also employed and it was based on mixing the Zr phosphonate gel with the perovskite precursors in order to obtain a physical mixture of the two components. Both methods resulted in a neat improvement of the PVK stability toward humidity, temperature, and UV irradiation. XRPD data and UV-vis spectra confirmed the absence of several degradation products normally observed for unprotected PVK crystals when aged for more than a month at high RH or exposed for several hours to UV irradiation.
Notably, the coating or mixing with ZPGly prevents the tetragonal to cubic phase transition of MAPbI 3 observed around 60 °C which is normally responsible of a general loss of PCE in solar devices due to several reason (density changes causing ruptures, changes in absorption properties, etc.). Future developments will be devoted to the use of the composites in PSC devices in order to test the efficiency and eventually to verify an improved stability according to the results reported in the present paper.

Experimental Section
Analytical and Instrumental Procedures-PXRD: PXRD patterns were collected in reflection geometry in the 4-40° 2θ range, with a 40 s step −1 counting time and with a step size of 0.016° on a PANalytical X'PERT PRO diffractometer, PW3050 goniometer, equipped with an X'Celerator detector by using the Cu Kα radiation. The long fine focus ceramic tube operated at 40 kV and 40 mA. The XRD graph was plotted with Origin Pro 8.5.
TGA: TGA was performed using a Netzsch STA490C thermoanalyzer under a 20 mL min −1 air flux with a heating rate of 10 °C min -1 .
DSC: DSC analysis was conducted using an automatic DSC1 (Mettler Toledo Switzerland). Aluminum pans with opened lids were employed in the experiments for all the samples and an empty pan was used as a reference. Samples of 3-5 mg were weighted directly into the aluminum pans and the thermal analyses were conducted at a heating rate of 10 °C min −1 from 35 to 120 °C. All runs were carried out under a nitrogen flow of 50 mL min −1 .
UV-VIS Spectroscopy: The reflectance spectra were collected by the UV-vis Varian (Cary 4000), equipped with an external integrated sphere DRA900 (diameter of 150 mm), coated by PTFE. A plate of barium sulfate was used as reference.
SEM: Sample morphology was evaluated using a SEM placing the sample on an aluminum stub previously covered with a graphite conductive adhesive. FEG LEO 1525 ZEISS instrument was used.
IR Spectra: The IR spectra in the 4000-400 cm −1 frequency range were acquired with an ALPHA compact FT-IR BRUKER, each spectrum was the average of 30 scans with a 2 cm −1 resolution, using circular windows of Thallium Bromo-Iodide (KRS5).
TEM images were obtained using a Philips 208 transmission electron microscope. The samples were prepared by putting one drop of an ethanol dispersion of the sample powder on a copper grid precoated with a Formvar film and dried in air.
Synthesis of N,N-Bis(phosphonomethyl)glycine (Glyphosine): Glyphosine was prepared following the Moedritzer-Irani method; [34] glycine (5 g, 67 mmol) were dissolved in 6 m HCl (50 mL),together with H 3 PO 3 (11 g, 133 mmol). This mixture was heated to reflux, and paraformaldehyde (8 g, 266 mmol) dispersed in water (10 mL) were slowly added within 2 h. After the last addition of paraformaldehyde, the solution was refluxed for 1 more hour and then the solvent was evaporated. The raw mixture was treated with 2-propanol, yielding a white solid that was filtered under vacuum and dried in an oven at 60 °C; 13.6 g of glyphosine was recovered (Yield = 74%).
Preparation of ZPGly Gel (DMF): The ZPGly dispersion was ultracentifugated to obtain the exfoliated ZPGly, then it was washed in DMF to exchange the water with DMF. ZPGly gel (DMF) containing 5%wt of ZPGly was obtaining.
Preparation of Perovskite: The synthesis of MAPbI 3 (PVK) was carried by slow evaporation in DMF. The 1:1 equivalents of MAI (0.954 g, 6 mmol) and PbI 2 (2.766 g, 6 mmol) were mixed in solvent (5 mL), then the solvent was slowly evaporated in an oven at 70 °C. During the solvent evaporation initially a yellow solid was formed and then turned black. [28] Preparation of Perovskite/ZPGly Composites: Method one (PVK@ZPGly) a film of ZPGly was prepared by deposition of the solution of exfoliated ZPGly in DMF above a silicon support. When the film was completely dry perovskite grow above the film by drop casting method [34] and following evaporation of the solvent at 70 °C.
Method two (MIX) the perovskite precursors solution were mixed together with the ZPGly gel: 1:1 equivalents of MAI (0.954g, 6 mmol), PbI 2 (2.766 g, 6 mmol), and ZPGly gel (DMF) (0.5 g, containing 5%wt of ZPGly) were mixed in solvent (5 mL), were mixed together to obtain a homogeneous solution. After that the solution is deposited on a glass or silicon support and placed at 70 °C to let the DMF evaporate. As a result of the slow evaporation, the film of perovskite and ZPGly is formed with the following composition. 0.8 mg g −1 of ZPGly.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.