Enhancing Wide-Bandgap Triple-Halide Perovskites for Tandem Solar Cells by 0.5% Formate and Zn(II) Doping

TL;DR

Adding small amounts of zinc and formate improves the performance of a type of perovskite material used in solar cells.

Contribution

The study introduces a 0.5% dual doping strategy with Zn(II) and formate to enhance perovines for solar cells.

Findings

Doping with Zn(II) and formate increases photoluminescence quantum yields up to 14%.

Photoluminescence lifetimes reach up to ~7 µs with the optimal doping concentration.

Abstract

The intrinsic properties of wide-bandgap, triple-cation, triple-halide perovskites, like (Cs0.21FA0.74MA0.05)Pb(I0.81Br0.14Cl0.05)3, can be improved by simultaneous doping with zinc cations (0.5%) as well as formate anions (0.5%). Photoluminescence quantum yields (up to 14% on quartz), as well as photoluminescence lifetimes (up to ~7 µs on quartz), indicate improved optical properties. Based on photoluminescent properties, the optimal total doping concentration of zinc ions (Zn(II)) and formate anions (Fo−) is determined to be 1% relative to Pb(II).

Figures

Full text

1. Introduction

Silicon dominates the commercial solar cell market, yet the power conversion efficiency (PCE) of single-junction cells is capped at 33.7% by the Shockley–Queisser (SQ) limit [1] and, considering Auger recombination [2], realistically limited to 29.4% for silicon (Si). Tandem solar cells (TSCs) offer a pathway to exceed these limits by stacking cells with different bandgaps, theoretically enabling PCEs up to ~45% [3]. In such tandem systems, perovskite solar cells (PSCs), known for their tunable bandgap, can be utilized as the top cell to absorb high-energy photons. Silicon, with its suitable bandgap of 1.12 eV, serves as a good bottom cell, capturing the lower-energy photons transmitted through the top layer. To reach these elevated levels of efficiency, optimizing the perovskite (ABX_3_) layer emerges as a crucial strategy (A = mono-cation, B = di-cation, and X = anion).

The current certified world record obtained by LONGi stands at 35.0% PCE (see NREL-chart). Tandem cell research has markedly progressed with the introduction of triple-halide perovskite compositions, characterized by their endurance in extreme conditions, such as temperatures over 200 °C [4] and radiation exposure [5]. Further emphasizing this potential, Mariotti et al. [6] achieved a certified efficiency of 32.5% in triple-halide perovskite–silicon tandem solar cells in 2022. Research by Valenzano et al. [7] on the integration of organic gelators into perovskite films leads to improved longevity without compromising photovoltaic efficiency. Similarly, Xu et al. [8] underscore the critical role of material engineering, especially the stabilization effects of chloride in triple-halide wide-bandgap perovskites. These developments demonstrated the critical importance of targeted material modifications for the advancement of solar cell performance.

Doping with cationic and anionic components has been identified as a pivotal strategy in optimizing perovskite solar cells. Muscarella et al. [9] highlight the effectiveness of Zn(II) as a B-site dopant in improving film stability, photoluminescence, and longevity. Similarly, Jeong et al. [10] demonstrated that the addition of 2% (Fo^−^) formate anions significantly increases cell efficiency to 25.2%. Hoeksma et al. [11] showed that synergy between an anionic and a cationic dopant can make single junction materials thermally more phase-stable with a longer charge carrier lifetime. Uddin et al. [12] demonstrated that the incorporation of these dopants into perovskite films leads to a significant enhancement in photoluminescence quantum efficiency and a decrease in deep trap density.

This work aims to advance triple-halide perovskite–silicon tandem solar cells by exploring the introduction of perovskite compositions simultaneously doped with zinc and formate. This dual-doping method is designed to synergistically enhance optical properties, charge carrier dynamics, electronic properties, and thermal stability, while concurrently reducing defect densities.

While perovskite–silicon tandem solar cells (PSTSCs) have achieved high PCE, there remains significant room for improvement towards the theoretical limit of 44%. The gap between this value and the current “state of the art” (35.0% PCE) implies plenty of options for contributions on the road to a practical value [3] of ~41% PCE. Research will concentrate on precise doping modulation, specifically targeting zinc and formate additives. Next to doping of perovskite materials, the ITO (indium tin oxide) interface layer can be functionalized with a self-assembled monolayer (SAM) as the HTL, (hole transport layer) specifically Me-4PACz [13] and 2PACz [14]. Recent developments [15] show that SAM formation and perovskite deposition can be established in one procedure. An asymmetrically substituted carbazole, with an electron-donating alkoxy group, can lead to more densely packed monolayers that have better HTL properties [16], resulting in a Si tandem cell with 34.58% PCE (certified).

Zinc doping. Since the first use of Zn(II) as an iso-valent B-site dopant in 2016 [17], this d^10^ ion has seen a marked increase in its application in this field, and Zn(II) is, for instance, present in the 2024 record for mini-modules [12]. In 2019, a review [18] was carried out; more recent work is compiled here. Saidaminov [19] and colleagues have correlated the incorporation of Zn(II) with lattice strain relaxation. Performance increase upon zinc incorporation [20,21], as well as an increase in the photoluminescence (PL) lifetime of mixed Pb-Sn perovskite materials [22], have been found. It is now generally accepted that Zn(II) ions take the place of Pb(II) in the bulk of the crystals and result in lattice contraction, strain relaxation, an increase in chemical bonding, and larger crystal domains [18]. Zn(II) has become a standard ingredient in perovskites, of which the presence and function are not even discussed anymore (and thus generally accepted) [23].

Next to Zn(II), one of the most used B-site dopants is Mn(II), a d^5^ species with a paramagnetic nature used in nanoparticles [24]. The synthesis of FAZnFo_3_ as well as FAMnFo_3_ has been reported [25] and could be considered as alloying agents.

Formate doping. The anion of formic acid, formate, has made a strong impact in perovskite research as its presence at 2% (relative to Pb) as X-site dopant in single junction devices has produced an impressive 25.2% PCE [10,11]. More recent work on formate doping has shown similar effects for blade-coated mini-module device efficiencies [12], as well as for PL increase. Formate salts of alkali metals [26] have also been used (at 2.5%), leading to reduced charge recombination and improved charge transport. Next to formamidinium formate (FAFo) [27], CsFo [28] and KFo [29] have also been applied.

It is important to note that one of the most recent single junction top devices (with 26.95% PCE) contains an anionic additive (heptafluorobutyrate) [30] and the current world record (27.2% PCE) [31] contains potassium binoxalate (K^+^[HC_2_O_4_]^−^, 0.7 mol%, 1.5 mg/mL) as well as formamidinium formate ([CH(NH_2_)]^+^[HCOO]^−^), 0.65 mg/mL; 0.007 M; 0.4 mol% vs. 1.8 M Pb) as additives.

The 2D/3D perovskite materials using alkylammonium ions that form a thin 2D layer on the edges of the 3D crystals are essential for the advancement of the field [32,33,34]. Grain boundary interface engineering is a very popular approach to the improvement of perovskite properties [35]. One of the most recent promising alkylammonium compounds is hexyl-ammonium bromide [36] (HABr) that can strongly enhance the PL lifetimes by using a post deposition treatment. It is worth noting that two recent (certified) high-efficiency single-junction devices were based on materials containing HABr, specifically 25.2% in 2021 [37] and 23.2% PCE in 2019 [38]. Other approaches with specific targeted molecules are also promising [35]. Liquid crystal materials (at 4 mol% relative to Pb) can be used to influence crystal formation, leading to 25.6% PCE devices [39].

In 2009, the field commenced [40] with MAPbI_3_ and MAPbBr_3_. Formamidinium (FA) became more important, the breaking of the 20% PCE boundary with (FA_0.8_MA_0.2_)Pb(I_0.8_Br_0.2_)3 as a disruptive exemplar [41]. However, for tandem cells, Cs incorporation has become very important. Nowadays, triple-halide materials containing iodide, bromide, and chloride are the more stable, key materials [34]. We can re-write the composition used by Mariotti et al., (Cs_0.22_FA_0.78_)Pb(I_0.85_Br_0.15_)3 + 5% mol MAPbCl_3_, into a more common description: we get (Cs_0.21_FA_0.74_MA_0.05_)Pb(I_0.81_Br_0.14_Cl_0.05_)3, thereby assuming that all MAPbCl_3_ is incorporated into the perovskite matrix, and no substantial decomposition into HCl and methylamine is taking place (which would lead to excess PbI_2_). It can be noted that indeed [6] their work indicates that PbI_2_ is present in the layers (See Supplementary Materials of Mariotti et al., Figure S5).

As optimization of radiative charge recombination is correlated to the SQ limit and the V_oc_, the photoluminescence quantum yields (PLQY) as well as the emission lifetimes are indicative of the suitability of perovskite materials [42], thus selecting the optimum condition [10]. On a microscopic scale, properties can be slightly different for every crystal in the film, and “hot spots” can be found that are optimally suited for charge generation [18,43,44]. Monitoring the emission of a perovskite material as a function of time after pulsed laser excitation can provide important information. Next to interactions with electron transport layers and hole transport materials, the reduction in the amount of defect states (in the bulk or at the surface) can also be investigated. However, thermally activated de-trapping of charges can also result in long lifetimes [45]; therefore, care has to be taken with the interpretation of time-resolved photoluminescence (TR-PL). There is an imbalance between mathematical models that are often used to fit the TR-PL data (from mono- [6] to multi-exponential and stretched exponential [9,11] models), and the pure solid-state semiconductor photo-physics approach [45,46,47].

Here, we report the fabrication and characterization of wide-bandgap lead halide perovskite materials as thin films on quartz. Photophysical characterization of these layers indicates improved properties after optimal doping with zinc and formate.

2. Results and Discussion

Based on the procedures [6] reported by Mariotti et al., wide-bandgap lead halide perovskite thin films were produced on quartz, but now under ambient conditions. A representative layer and X-ray diffraction (XRD) are presented in Figure 1 and Figure 2. Figure 1 shows the typical smooth dark brown film, with a slight orange appearance upon transmissive view. The XRD pattern obtained is compared with the data reported by Mariotti and shows good resemblance (see also next sections for more details on the PbI_2_ peak). This indicates that the incorporation of Zn and formate (at 1% total doping level) does not change the perovskite lattice structure (see Supplementary Materials for tabulated XRD-data, see Table S1).

Films with different doping levels were obtained, and the relative PL intensity as well as the PL lifetime histograms were determined as a function of total doping level. Both metrics show their highest PL performance at a total doping level of 1%.

As reported earlier for FAPbI_3_-based materials [11] for single-junction devices, the total doping concentration is relatively low (less than 3%). Combining all data for the wide-bandgap materials clearly indicates that for this composition, a total doping concentration of only 1% (expressed relative to Pb(II)) is optimal (i.e., doping with Zn(II) (0.5 mol%), as well as for Fo^−^ (0.5 mol% vs. Pb(II))). Figure 3 shows the relative quantum yield (or emission intensity at identical absorption) for various doping degrees for comparative films. On the right y-axis, the maximum of the lifetime distribution of the emission of a film is shown, as obtained with confocal laser scanning microscopy (see Supplementary Materials for details, Figure S1).

Based on the steady-state PL measurements, further optimization was conducted for the 1% doping levels. Compositional optimization was directed at eliminating the excess PbI_2_ (see Supplementary Materials: Experimental Details). As reported [6] by Mariotti et al., the perovskite material in their tandem cells contains PbI_2_ in the layer. This information was obtained by using GIWAX. Their GIWAXS measurements on perovskite indicate that PbI_2_ is preferentially enriched near the surface of the perovskite layer. PbI_2_ is a non-photoactive material that does not produce photovoltage or photocurrent. Therefore, the work reported here focuses on perovskite films that do not contain PbI_2_.

As also reported by Janssen et al., optimal stoichiometry [48] is needed to obtain optimal properties. This also agrees with our previous work on Zn(II) doping of FA- and MA-based materials [9], where PbI_2_ became apparent in XRD at higher doping levels, reducing emission levels. As displayed in Figure 4, PbI_2_ needles were observed in the top layer of our (non-optimized) films, as produced with and in accordance with the method reported by Mariotti et al. [6].

Clearly, next to GIWAX, simple optical microscopy can also be used to probe the presence of PbI_2_ in films. This has also been reported for FAPbI_3_-based materials. In optical microscopy, PbI_2_ in perovskite films can be identified by white material surrounded by a darker perovskite phase (in black and white images). In colored images, the clear yellow color of PbI_2_ can be easily discerned [11]. This implies that, next to GIWAX and SEM, optical microscopy is a simple method that can be used to probe the presence or formation of PbI_2_ in thin solid films. Figure 4 shows the comparison between optical microscopy and SEM, exemplifying this effect. More SEM and optical images are given in the Supplementary Materials (Figures S2–S5). The perovskite crystals have a dimension of ca. 400 nm (see Figure 4—left). By comparing the unoptimized and optimized films, optical microscopy and SEM images show that the optimized perovskite films do not exhibit visible PbI_2_-related features at the surface, demonstrating the effectiveness of the compositional optimization strategy.

It can be noted that the digitized (black) XRD pattern (Figure 2) is obtained for a lower bulk part of the layer, and therefore does not show PbI_2_ signals (as it is extracted from the GIWAX data). The PbI_2_ is present in the top part of the film, at the top interface.

By maintaining the optimal doping level and slightly adjusting the precursor composition by using compositional engineering (see SI: Experimental details), PbI_2_ could be eliminated from the layers, thereby optimizing the PL performance for the doped layers.

Figure 5 and Figure 6 show the determination of the PLQY as well as the TR-PL of our best optimally doped layers that do not show PbI_2_ needles in their top layer. By using calibrated absolute photoluminescence quantum yield equipment, a maximum PLQY of 14% was obtained for wide-bandgap perovskite on quartz (see Supplementary Figure S6 for UV-Vis and PL). It has to be noted that Mariotti et al. [6] reported 3.75% as PLQY for their films on quartz (and a PL lifetime of ~2 µs, by tail-fit). If all other components are not influenced by the additives, this could imply that the doping strategy reported here could induce higher efficiency in tandem cells.

Here we compare the time-resolved photoluminescence decay of our best optimal and doped (Cs_0.21_FA_0.74_MA_0.05_)Pb(I_0.81_Br_0.14_Cl_0.05_)3 film with previously reported Zn– and Zn–formate–doped FA/MA-based perovskites (Figure 6). On a log–log scale, the red trace shows our 2025 material with a lifetime of 6.84 ± 1.1 µs, compared to 3.65 ± 0.8 µs (2024, blue) and 1.57 ± 0.1 µs (2019, green). This trend reflects progressively longer carrier lifetimes achieved through Zn(II) and Zn–formate additive strategy.

Violin plots of all PLQY and TR-PL data of 1% doped optimized films, as well as experimental details, are provided in the Supplementary Materials (See Supplementary Materials: Figure S7).

Laminated films of the 1% doped optimized wide-bandgap perovskite show very good thermal stability (no yellowing after more than 1 year under dark, or indoor light, ambient conditions, not in a glove-box). However, laser irradiation with 532 nm through a pinhole or a slit (=partial illumination) showed clear signs of permanent photodegradation after 500 hrs. Laser power was varied from 1 to 5 suns. Photobleaching is presented in Figure S8, setup in Figure S9 (see Supplementary Materials). The methylammonium content (~5%) as well as halide migration is likely to play an important role. MAPbI_3_ has been reported to be intrinsically photo-labile [49]. Ion migration is one of the important steps in the photodegradation of hybrid halide perovskites [50,51,52]. Mariotti et al. reported 25% loss after 478 h of illumination (for “undoped” material) [6]. Degradation upon partial shading (= partial illumination) has been reported for ((Cs_0.05_FA_0.81_MA_0.14_)Pb(I_0.9_Br_0.1_)3 + 1 mol% RbCl) [51], a similar wide-bandgap perovskite material. These reports are in accordance with our results on photodegradation.

3. Experiment

All materials and reagents were used without any purification, including dimethyl sulfoxide (DMSO, ≥99.7%) from Acros Organics (Geel, Belgium); dimethyl formamide (DMF, 99.8%) from Thermo Scientific (Waltham, MA, USA); acetone and isopropanol from VWR Chemicals (Radnor, PA, USA); formamidinium iodide (FAI, 99.99%), lead bromide (PbBr_2_, ≥98.0%), lead chloride (PbCl_2_, >99.0%), and lead iodide (PbI_2_, ≥98.0%) from TCI (Tokyo, Japan); methylammonium chloride (MACl, 98%), cesium iodide (CsI, 99.9%), and anisole (99.7%) from Sigma-Aldrich (St. Louis, MO, USA); and zinc formate (ZnFo_2_, 98%) from Alfa Aesar (Haverhill, MA, USA). Ultra-flat glass substrates coated with a 20 nm synthetic quartz layer were obtained from Ossila (Sheffield, UK).

Perovskite (Cs_0.21_FA_0.74_MA_0.05_)Pb(I_0.81_Br_0.14_Cl_0.05_)3 precursor solutions (1.4 M) were prepared by dissolving FAI, MACl, CsI, PbI_2_, PbBr_2_, and PbCl_2_ in 600 µL DMF and 150 µL DMSO according to the stoichiometric ratios, with ZnFo_2_ and ZnI_2_ added at different concentrations as an additive. The solutions were stirred at 60 °C and filtered through 0.2 µm PTFE syringe filters. The substrates were cleaned with detergent, followed by ultrasonic cleaning in water, acetone, and isopropanol (15 min each), then treated with UV–ozone for 2 h. Anisole (= methoxybenzene) was used as the anti-solvent. The perovskite films were spin-coated at 5000 rpm for 45 s (5 s acceleration), with 300 µL anisole dropped at 25 s after the start of spin-coating, and annealed on a hot plate at 100 °C for 20 min.

Scanning electron microscopy (SEM) images were obtained using a Verios 460 microscope (FEI, ThermoFischer: Waltham, MA, USA). Bright-field optical microscopy was performed to examine the surface morphology. X-ray diffraction (XRD) was measured using a Bruker D2 Phaser diffractometer (Billerica, MA, USA)with Cu Kα radiation. Photoluminescence quantum yield (PLQY) was measured using an Edinburgh Instruments FS5 fluorimeter (Livingston, UK) equipped with an SC-30 integrating sphere at 450 nm excitation. Long-lifetime measurements were recorded at AMOLF using 485 nm excitation under low-fluence conditions. Additional TRPL data for the composition series were collected using a confocal microscope (PicoQuant: Berlin, Germany) equipped with a 488 nm Chameleon laser (Coherent: Saxonburg PA, USA, 100 kHz) and a TCSPC detection module.

More experimental details are given in the Supplementary Materials.

4. Conclusions

The concept [11] of improving photophysical properties (PLQY and PL lifetime) of perovskite materials by simultaneous doping with Zn(II) (0.5%) and formate anions (0.5%, in agreement with ref. [31] for narrow bandgap) has been shown to be applicable to wide-bandgap materials such as Cs_0.21_FA_0.74_MA_0.05_Pb(I_0.81_Br_0.14_Cl_0.05_)3. Optimized photophysical properties on quartz are observed at 1% total doping (relative to Pb(II)). A PLQY of 14% has been obtained and a PL lifetime of 6.84 µs. XRD is in agreement with the composition reported by HZB [6], and in non-optimized compositions, PbI_2_ needles can be observed in the top layer of the thin films with optical microscopy (and SEM). Compositional engineering can eliminate the PbI_2_ needles in the top part of the layer. Prolonged laser irradiation (532 nm, 500 h) of laminated films leads to photodegradation.