The Photochemical Stability of PbI₂ and PbBr₂: Optical and XPS and DFT Studies

Abstract

We investigated the photochemical stability of PbX₂ (X = I and Br) halides by optical and X-ray photoelectron spectroscopy (XPS). The optical absorbance displayed a strong reduction for PbI2 with light soaking and permanent behavior for PbBr₂. The XPS survey spectra showed a sharp drop in the I:Pb ratio for PbI₂ from 1.63 to 1.14 with exposure time from 0 to 1000 h while for PbBr2, it remains practically unchanged (1.59–1.55). The measurements of the XPS Pb 4f and Pb 5d spectra have shown the partial photolysis of PbI₂ with the release of metallic lead whereas PbBr₂ demonstrated remarkable photochemical stability. According to the density functional theory (DFT), calculations of the metal and iodide vacancy formation energies for PbBr₂ are higher than for PbI₂ which confirms the better stability to light soaking. The high photochemical stability of PbBr₂ means that it can be used as excess under MAPbBr₃ perovskite synthesis to improve not only the power conversion efficiency but also stability to light soaking.

1. Introduction

Lead halide perovskites are considered the most promising materials for photovoltaics [1,2,3]. The power conversion efficiency (PCE) of perovskite solar cells has rapidly increased from 3.8 to 25.6% [4,5,6] and has reached the level of silicon cells. On the other hand, the low resistance of halide perovskites to light, temperature, and humidity hinders their commercialization [7]. To solve this problem, numerous studies have been performed over the last years to analyze the causes and mechanism of degradation of halide perovskites. Based on these studies, the main efforts have been directed at improving the surface morphology [8], grain size [9], halogen ratio [10], the composition of cations [11], charge transport layers [12,13], and electrodes [14,15]. It is found that the introduction of excess PbI₂ into CH₃NH₃PbI₃ (MAPbI₃) precursors reduced the defect density and increased the carrier lifetime and efficiency in a thin PbI₂-rich perovskite film [16,17,18]. Time resolved photoluminescence (TRPL) measurements have shown that excess PbI₂ results in a longer carrier lifetime in MAPbI₃ [19,20]. Scanning electron microscopy (SEM) and (X-ray diffraction) XRD measurements showed that the addition of a 10 mol % excess of PbI₂ to MAPbI₃ did not lead to obvious changes in morphology and crystal structure but was accompanied by a significant increase in PCE [16]. However, subsequent studies have shown that despite the benefits of increasing PCE, the implementation of excess PbI₂ does not improve the long-term photochemical stability of perovskites since simultaneously formed PbI₂ crystals easily decompose to metallic Pb and gaseous I₂ through photolysis [21,22,23,24]. In this regard, the search for other additives that perform the same useful role as excess PbI₂ but without photolysis is relevant and in demand [25]. One of these additives can be PbBr₂, which like PbI₂ passivates the defects at grain boundaries and interfaces improving the quality of perovskite films [26,27,28]. It was found in Ref. [29] that if MAPbI₃ perovskite films decomposed with the formation of PbI₂ upon irradiation with visible light, then the introduction of PbBr₂ into a mixed halide solid solution caused more structural defects and grain boundaries and an increased defect density ensured the photochemical stability of these perovskite materials. In connection with this, we have undertaken a comparative study of the photochemical degradation of PbI₂ and PbBr2 with the help of X-ray photoelectron spectroscopy (XPS) which has proven itself in the study of the resistance of halide perovskites currents to light soaking [30].

2. Experimental and Calculation Details

The PbI₂ precursor solution was prepared by adding 1.35 M PbI₂ in 1 mL anhydrous dimethyl sulfoxide (DMSO) (Sigma-Aldrich, Taufkirchen, Germany) and stirring at 60 °C for 12 h after which 7.5 mg of 2-TC Am was added to the precursor solution and then centrifuged at 4000× g rpm min for 30 s. The spin-coating films were annealed at 70 °C on a hot plate for 5 min. The successful formation of PbI₂ and PbBr₂ films was examined by X-ray diffraction study (XRD) (Billerica, MA, USA). Only peaks of lead halides were detected (Figure 1), which indicates the high quality of the original films.

For photochemical aging, an LG sulfur plasma lamp (LG, Seoul, Republic of Korea) was used as a light source which provides a good approximation to the AM1.5G solar spectrum. Wavelengths below 350 nm were cutoff using an additional UV filter. The light power on the samples was ~70 mW/cm² at a temperature of 45 ± 2 °C. The absorption spectra were measured with an AvaSpec-2048-2 UV-VIS fiber spectrometer built into the glove box.

XPS was used to measure core level and VB spectra with the assistance of a PHI XPS 5000 VersaProbe spectrometer (ULVAC-Physical Electronics, Chanhassen, MN, USA) equipped with a spherical quartz monochromator and an energy analyzer working in the range of binding energies from 0 to 1500 eV. The energy resolution was ΔE ≤ 0.5 eV.

The first principles calculations were carried out with the Vienna Ab initio Simulation Package [31,32]. The generalized gradient approximation was chosen in the PBE parametrization [33]. For an accurate calculation of the total energies and forces, the kinetic energy cutoff for the plane waves was set to 500 eV. The total energy convergence was set to 10⁻⁵ eV and atomic positions were relaxed with force convergence of 0.01 eV/Å per atom. The Monhkorst–Pack k point mesh of 2 × 2 × 1 divided the reciprocal space.

3. Results and Discussion

The photochemical degradation of PbI₂ and PbBr₂ films was studied with the help of optical spectroscopy. The obtained absorption spectra of PbI₂ films irradiated for different times are compared with the spectra of PbBr₂ (Figure 2a,b). From the data obtained, it follows that after 200 h of irradiation, a significant decrease in the optical absorption of PbI₂ is observed; this indicates the photochemical degradation of this halide. This causes the film to decompose into Pb⁰ and iodine gas, thereby reducing the film thickness and hence the absorption. On the other hand, the optical spectra of PbBr₂ show no change until 1000 h of exposure to visible light.

In Figure 3, the XPS survey spectra of PbI₂ and PbBr₂ halides irradiated with visible light are presented.

Table 1. Surface composition (in at.%).

Sample	C	O	I	Br	Pb	Na	Ca	I(Br):Pb
PbI2 as prepared	44.9	8.5	27.5	-	168	2.3	-	1.63
photo 200 h	53.4	11.0	20.3	-	15.3	-	-	1.32
photo 500 h	56.3	15.4	12.6	-	10.8	4.9	-	1.16
photo 1000 h	50.0	19.2	9.7	-	8.5	8.4	4.2	1.14
PbBr2 as prepared	61.1	11.2	-	15.8	9.9	2.0	-	1.59
photo 1000 h	69.0	10.2	-	12.0	7.7	1.1	-	1.55

shows the surface composition determined from these spectra. As follows from these data, the I:Pb ratio drops sharply for PbI₂ from 1.63 to 1.14 with exposure time from 0 to 1000 h, while for PbBr₂ it remains practically unchanged (1.59–1.55). Such a reduction in the I:Pb ratio is usually associated with the photolysis of lead iodide which is accompanied by decomposition into metal and gaseous iodine. For a more complete analysis of the photolysis process, we measured the XPS Pb 4f and Pb 5d spectra of the core levels of PbI₂ and PbBr₂ as a function of the irradiation dose with a high energy resolution which is presented in Figure 3. For comparison, the spectra of metallic lead taken from [34] are shown in this figure. Their analysis shows that the contribution of metallic lead to the PbI₂ halide consistently increases with exposure time, which indicates an increase in the corresponding decay product and is in complete agreement with the results of measurements of XPS survey spectra. On the other hand, as follows from Figure 4, no traces of the appearance of metallic lead are observed when PbBr₂ halide is irradiated for 1000 h which again coincides with the invariance of the I:Br ratio determined from the XPS survey spectra. Thus, the data obtained show a significant difference in the photochemical stability of PbI₂ and PbBr₂ binary halides and make it possible to justify the choice of an excess of PbB_r2 in the synthesis of halide perovskites as having a much higher resistance to photolysis.

To demonstrate the relative stability of the PbBr₂ with respect to PbI₂, we calculated the cohesive energy and defect formation energy of the PbBr₂ monolayer. For this aim, the monolayer with the 6 × 6 × 1 supercell and vacuum size of 23 Å was constructed for 1T PbBr₂ structure. The relaxed monolayer has a hexagonal honeycomb structure formed by the Br edge-sharing octahedra with parameters that are in good agreement with the earlier data of Ref. [35]. The Pb-Pb distance is 4.49 Å and a monolayer thickness is 3.33 Å, which assumes an octahedron squeezing along the direction perpendicular to the slab. The Pb-Br distance is 3.08 Å and Br-Pb-Br angles are 93.6 and 86.4 for the Bromium atoms in the same and opposite Br surfaces, respectively.

Figure 5 shows the relaxed crystal structures with the Pb and Br vacancies (left and right panels, respectively). In the V_Pb structure (Pb vacancy), the most distant from the vacant octahedron is almost unchanged: the Pb-Br distances deviate from ideal by the order of mÅ and the angles are modified by less than half of a degree. This guarantees that the supercell size was properly chosen to minimize the vacancy interaction. The Pb-V_Pb distance is 4.36 Å and the thickness of the monolayer in this place, 3.29 Å, is about one percent smaller than in the ideal case. The V_Pb-Br distance is 3.2 Å. Therefore, the lead atoms near the vacancy move to each other trying to close the hole but the surrounding Br₆-octahedra are distorted in such a way that the Br-Br distance in the surface regular triangle is 4.76 Å (this distance is 4.49 Å in the pristine structure). In the case of the V_Br structure, the most distant octahedron is also almost unchanged with respect to the pristine structure. At the same time, the Pb-Pb distance for atoms close to the vacancy is 4.15 Å, which is even smaller than the corresponding one in the V_Pb structure. The Br-Br distance in the surface regular triangle is 4.43 Å, which is smaller not only with respect to the V_Pb structure but with a pristine one also. So, in the case of the V_Br structure, all atoms move towards the empty space created by Br vacancy.

The cohesive energy and defect formation energies of the Pb- and Br- vacancies (V_Pb and V_Br) were calculated as described in Refs. [36,37]. These results are presented in

Table 2. The cohesive and defect formation energies in charge neutral state for the PbBr₂ compared to PbI₂ (from Wang et al. [38]). All values are in eV.

	1T PbBr2			1T PbI2
	Pristine	VPb	VBr	Pristine	VPb	VI
Ecoh, eV	3.07	2.95	2.99	2.63	2.59	2.62
Edef, Br/I-rich		1.2	2.3		1.1	2.2
Edef, Br/I-poor		4.3	1.3		4	0.9

. We did not consider a charged defect since in the case of a charged slab, the total energy grows linearly with the distance from the slab; hence, it cannot be used to evaluate relative energies. The chemical potentials for the Br-rich and Pb-rich limits correspond to the Br₂ molecule and Pb bulk solid, respectively. The Br-poor and Pb-poor limits can be obtained from the formula for an equilibrium state for the compound under consideration,

μ_PbBr2 = μ_Pb + μ_Br2,

(1)

therefore, we take the values of the μ_Br from −3.04 eV to −1.5 eV and μ_Pb from −6.66 eV to −3.59 eV in the following discussions. One can clearly see that the all evaluated values for the PbBr₂ are larger than its corresponding PbI₂ counterparts [38]. These findings unambiguously indicate that PbBr₂ has higher stability with respect to the single vacancy formation of any type in comparison to the PbI₂.

Thus, the conducted experimental and theoretical studies show a much higher photochemical stability of PbBr₂ compared to PbI₂ halide. This suggests that PbBr₂ halide can be preferable for use as an excess additive for the improvement of properties of MAPbBr₃ perovskite.

4. Conclusions

The results of this work unambiguously indicate a significant photochemical degradation of lead iodide compared to lead bromide. It has been shown by XPS that significant degradation of PbI₂ starts already at 200 h, while PbBr₂ remains stable up to 1000 h of light soaking. In accordance with the DFT calculations, the formation energies of lead and halogen vacancies turn out to be higher in PbBr₂ than in PbI₂, which confirms our experimental results. The obtained results indicate that the use of PbBr₂ as an excess additive in the development of stable hybrid organometallic perovskites is preferable.