What defines the perovskite solar cell efficiency and stability: fullerene-based ETL structure or film morphology?

Abstract

In this contribution, we report the synthesis and structural characterization of a series of fullerene derivatives and their further systematic investigation as promising ETL (electron transport layer) materials in p–i–n perovskite solar cells (PSCs). The devices fabricated using a set of fullerene derivatives F1–F6 demonstrated high power conversion efficiencies (PCEs), up to 19.0%, compared to the 17.3% obtained with reference cells assembled using the benchmark ETL material, [6,6]-phenyl-C₆₁-butyric acid methyl ester (PC₆₁BM). The improved photovoltaic performance of PSCs incorporating the fullerene derivatives originated from a decreased trap density at the perovskite/ETL interface and full coverage of the perovskite absorber layer, which was revealed by the photoluminescence (PL) spectra and infrared scattering scanning near field optical microscopy (IR s-SNOM). Also, we enhanced our experimental results with theoretical DFT and DFT-MD calculations which gave further insight on the dependence between the structure and electron mobility in these films. Significantly improved operational stability was achieved for non-encapsulated devices using fluorine-loaded fullerene derivative F5 as the ETL, which retained >60% of the initial efficiency after ∼1300 h of continuous illumination (1 sun), whereas the reference cells with PC₆₁BM as the ETL degraded to ∼40% within 200 h under the same aging conditions. Therefore, the obtained results demonstrated that the molecular structure of the fullerene derivatives affects the performance of PSCs, whereas the film morphology plays a crucial role in defining the operational stability of the devices.

---

Introduction

Perovskite solar cells have presently achieved a power conversion efficiency (PCE) of 25.7% as result of the exceptional optoelectronic properties of lead halide based absorber materials; however, the stability of PSCs has become the next crucial challenge which remains the bottleneck to be addressed prior to commercialization.^1,2 Among the various device architectures of PSCs, the p–i–n configuration, with a fullerene-based top electron transport layer (ETL), is one of the most intensely studied^3,4 with a presently attained PCE of 23.7%.⁵ These devices typically use PC₆₁BM as the ETL, and the high performance of PC₆₁BM is explained by its high electron mobility and good solubility in nonpolar solvents.^6–8 Although PC₆₁BM delivers high efficiencies in p–i–n PSCs, this material has significant drawbacks; in particular, it cannot provide long operational stability of PSCs under environmental conditions^9–11 as well as under exposure to light and/or heat.^12,13 There is growing experimental evidence that PC₆₁BM is responsible for facile degradation of p–i–n PSCs.^14,15 The aging pathway was associated with migrating ions such as MA⁺ and I⁻ (in the case of the MAPbI₃ absorber), which diffuse through the PC₆₁BM ETL and then induce corrosion of the top electrode.^15–17 The loss of PCE in PSCs is associated with the presence of deep trap defects as per Shockley–Reed–Hall theory.¹⁸ However, due to strong anharmonic behavior of the material,¹⁹ the nature of these deep-trap defects is still widely debated. Experimental evidence indicates that PC₆₁BM forms a radical product with the iodide anion when introduced into precursor solution.²⁰ This is indicative of its possible involvement in redox chemistry at the perovskite/ETL interface. PC₆₁BM apparently heals the antisite defect I_Pb(PbI₃⁻), which is a recombination trap. It should be noted that this effect seems to be more related to the C₆₀ cage²⁰ than to the functionalizing group and thus this aspect is expected to be common to all functionalized fullerenes. A second factor is that the surface-based I_Pb defect is a source of the Pb²⁺/Pb⁰ and I⁻/I₃⁻ redox pair which we have described previously.²¹ Triiodide is a powerful superhalogen and thus may be expected to be highly corrosive, so PC₆₁BM's ability to prevent its formation would be a major advantage of this ETL. This also explains why this defect is only particularly harmful at the surface: as long as it is trapped within the bulk of the material, it is part of a closed redox cycle²² and thus no material is lost.

Up until now, tremendous efforts have been made to develop new fullerene derivatives with functional molecular design which may improve the stability of p–i–n devices. In the context of devices possessing ambient stability, a binary ETL was reported based on an SFX-PDI2 material ([fluorene-9,90-xanthene] (SFX) core building block with two PDI units) and PC₆₁BM, which delivers a PCE of 15.2%; however, the devices exhibited modest stability for 720 h at 40–50% humidity.²³ In another study, PC₆₁BM was doped with hexamethonium bromide (HMB), which delivered PCE ∼18.0% in PSCs, that demonstrated a shelf life of >1100 h under an inert atmosphere without light.²⁴ Recently, Z. Xing et al. have introduced a new set of pyrrolidinofullerene derivatives as ETLs in p–i–n PSCs and reported a PCE of 20%.²⁵ The designed ETLs enabled long-term ambient stability for ∼1056 h (T₈₀) in air but without light illumination.

In regard to the stability of PSCs under light and thermal stress, there are many contributions following the approach of blending fullerene derivatives and using dual or triple interfacial modification materials such as PPDIN6,²⁶ ZrL3:bis-C₆₀,²⁷ Ti(Nb)O_x,²⁸ BCP/Cr/Cr₂O₃,²⁹ and SnO₂/ZnSnO₂/ITO/LiF.³⁰ The reports indicated that the most stable ETL material/electrode combination enables at best 1000 h of stable PSCs operation under a continuous illumination of 100 mW cm⁻². Thus, insufficient operational stability of p–i–n PSCs limits their commercialization. Overall, it has become clear that all fullerene derivatives as single-component ETLs provide poor PSC stability under continuous illumination and at high temperatures.

Although the majority of efforts thus far have focused on superior efficiency, the stability issue is equally important, so achieving efficient and stable fullerene-based ETL operation in p–i–n PSCs represents an urgent and important task. As such, there is a strong demand for the development of new fullerene-based ETL materials with optimum electronic and physicochemical properties such as interface behavior, film morphology, electron mobility, energy levels, and trap states which can deliver the target efficient and stable p–i–n PSCs.^31–33

Herein, we report the synthesis and structural characterization of a series of fullerene derivatives F1–F6 as promising ETL materials for p–i–n PSCs. All studied materials F1–F6 showed good photovoltaic performance comparable to the PC₆₁BM benchmark. Moreover, the devices assembled with fullerene derivatives F2, F4, and F5 demonstrated improved operational stability under continuous light exposure as compared to the reference cells using PC₆₁BM.

Results and discussion

Fullerene derivatives F1–F6, such as PC₆₁BM, are methanofullerenes, since they contain a cyclopropane fragment attached to the fullerene cage (Fig. 1). Adducts F1–F3 and F6, in contrast to the classical PC₆₁BM molecule, contain no ester fragment in their structure which makes them less polar materials than the reference molecule. In contrast, the phenyl substituent of the developed fullerene derivatives F1–F3 and F6 is modified by introducing additional electron donor alkoxy and trimethylsilyl substituents at different positions. It is known from literature data that the introduction of electron donor groups in close proximity to the fullerene framework significantly affects its electronic properties.^34–39 The introduction of alkoxy substituents in the phenyl moiety for adducts F1 and F3 not only reduces their electron affinity but also has a beneficial effect on the performance characteristics of organic solar cells as compared to the reference PC₆₁BM-based system.⁴⁰ Fullerene F4, as well as PC₆₁BM, contains an ester group (propyl ester of 2-thienylpropanoic acid) but with a longer solubilizing propyl radical. Moreover, instead of the phenyl group, the 2-thienyl substituent is attached to the cyclopropane fragment.⁴¹ It should be noted that the fullerene derivative F5 contains both the ester group with a pentafluorobenzyl radical and phenyl substituent attached to the cyclopropane fragment. The presence of the pentafluorobenzyl substituent should form a hydrophobic surface on the perovskite film, which would improve the ambient stability of the devices. Thus, the presented series of designed compounds covers a significant range of methanofullerenes with varying molecular structures, which makes their study promising as electron transport materials in perovskite solar cells.

Fig. 1 Chemical structures of fullerene derivatives F1–F6 and PC₆₁BM.

The synthesis of the fullerene derivatives consists of the formation of the corresponding tosylhydrazones, which are then introduced into the Hummelen–Wudl reaction with C₆₀.⁴² The details of the synthesis and spectroscopic characteristics of the fullerene derivatives F1, F3 and F4 were reported previously,^40,43,44 whereas the data for the new compounds F2, F5 and F6 are given in the ESI.† It should be noted that all the studied fullerene derivatives were structurally characterized. The X-ray single crystal structures of PC₆₁BM, F1, F3 and F4 were published previously.^40,44,45 Herein, it was possible to grow single crystals of F2, F3, F5 and F6 suitable for X-ray diffraction analysis by slow evaporation of their chloroform solutions (details given in the ESI†). Thus, having a set of structurally characterized fullerene derivatives provided us with a unique opportunity to investigate possible relationships between the molecular structures of these compounds and the performance characteristics of PSCs using them as ETL materials.

The full configuration of p–i–n PSCs was ITO/PTAA/Cs_0.12FA_0.88PbI₃/ETL/Mg/Ag as shown in Fig. 2a.

Fig. 2 (a) Schematic layout of the used p–i–n PSC architecture and energy level diagram; (b) current–voltage (J–V) characteristics and (c) external quantum efficiency (EQE) spectra of PSCs with an Cs_0.12FA_0.88PbI₃ absorber and PC₆₁BM, F2, and F3 used as ETL materials.

While investigating each of the fullerene derivatives, we optimized the ETL film thickness by varying the solution concentration (10–50 mg mL⁻¹) and spinning frequency (from 1000 to 5000 rpm). Under the optimal conditions, the ETLs' film thicknesses were ∼22 nm (F1), ∼36 nm (F2), ∼40 nm (F3), ∼34 nm (F4), ∼21 nm (F5), ∼18 nm (F6) and ∼30 nm (PC₆₁BM).

The current density–voltage (J–V) curves for the representative PSCs using methylammonium-free Cs_0.12FA_0.88PbI₃ perovskite absorber materials and PC₆₁BM, F2, and F3 as ETL components obtained under the simulated AM1.5G 100 mW cm⁻² solar irradiation are presented in Fig. 2b. Fig. S1† (ESI) shows the J–V curves and EQE spectra for the PSCs assembled with fullerene derivatives F1, F4, F5, and F6. Table 1 lists the characteristics of all the solar cells based on Cs_0.12FA_0.88PbI₃ as an absorber layer. Notably, all fullerene derivatives delivered high PCEs approaching 19.0%, whereas the reference cells using PC₆₁BM provided a lower PCE of 17.3%. The external quantum efficiency (EQE) spectra measured for devices with different ETLs (Fig. 2c) reconfirmed the current density (J_SC) values obtained from the J–V curves.

Table 1 Photovoltaic parameters extracted from the current–voltage characteristics of methylammonium-free perovskite solar cells incorporating various ETLs^a. The champion cell parameters are given in brackets

ETL	ETL conc., [mg mL^−1]/coating freq. [rpm]	V OC [V]	J SC [mA cm^−2]	FF [%]	PCE [%]
a The standard deviation was calculated for 16 cells scanned in forward and reverse directions with 10 mVs^−1. b F1 and F6 were processed from CS2-chlorobenzene mixtures with the components taken in a volume ratio of 6 : 4 and 1 : 9, respectively.
PC61BM	30/3000	1.02 ± 0.03 (1.05)	21.9 ± 0.8 (22.7)	62.0 ± 1.3 (72.4)	16.4 ± 0.9 (17.3)
F1	20/3000^b	1.0 ± 0.01 (1.01)	22.0 ± 1.0 (23.1)	70.5 ± 4.0 (73.0)	15.7 ± 1.0 (17.1)
F2	30/2000	1.01 ± 0.04 (1.09)	22.3 ± 0.9 (22.2)	75.4 ± 1.6 (78.6)	17.0 ± 0.9 (19.0)
F3	30/1000	1.01 ± 0.01 (1.0)	21.9 ± 1.1 (23.7)	71.0 ± 2.2 (72.8)	15.7 ± 0.9 (17.3)
F4	30/2000	1.0 ± 0.03 (1.08)	22.4 ± 0.9 (22.2)	70.0 ± 1.5 (72.1)	15.6 ± 0.8 (17.3)
F5	20/3000	1.0 ± 0.02 (1.0)	22.1 ± 0.9 (23.4)	70.6 ± 3.0 (73.7)	15.6 ± 0.9 (17.3)
F6	20/4000^b	1.0 ± 0.02 (1.0)	21.1 ± 1.5 (23.4)	64.3 ± 4.3 (67.4)	13.5 ± 1.3 (15.6)

---

The above results prove that all new fullerene derivatives presented high performances comparable to the PC₆₁BM ETL material in combination with a methylammonium-free Cs_0.12FA_0.88PbI₃ absorber material. The enhancements in PCE might be assigned to the improved perovskite/ETLs interface or better film morphology of the designed ETLs.

It has been previously demonstrated that the chemical structure of fullerene derivatives has a significant impact on the performance of PSCs.⁴⁶ The PV performance of the fullerene derivatives F1–F6 might be connected to their charge transport properties, which are influenced by the molecular packing and distance between the fullerene cages, since the charge is hopping from one cage to another.^47,48 Consequently, we used X-ray diffraction to probe the structure of the fullerene derivatives and packing directly. For this purpose, the distance between the nearest carbon atoms of two fullerene cages in the crystal lattice was chosen as the measured experimental value. Thus, the closest distance between the carbon atoms of two fullerenes in the C₆₀ fullerene crystal is ∼3.27 Å, while for the PC₆₁BM crystal this distance is ∼3.19 Å. An even shorter distance between the nearest carbons of the fullerene frameworks was observed in the crystal of compound F5 (∼3.10 Å). For compounds F2 and F6, the distance between the carbon cages was ∼3.2 Å. In a number of compounds F1, F3 and F4, this distance gradually increases from 3.24 to 3.3 Å, respectively (Table 2).

Table 2 Nearest carbon–carbon distances between the neighbouring fullerene cages in the crystal lattice (experimental) and computationally optimized unit cells, frontier energy levels and experimental electron mobilities of the functionalized fullerenes

Comp.	Sol.^a	Electron mobility, μe [cm^2 V^−1 s^−1]	C–C dist.^b (Å) [Exp]	C–C dist. (Å) [Theo] ^c	HOMO^d (eV)	LUMO^d (eV)
a Solvent molecules incorporated in the crystal lattice in some cases as revealed by structure analysis: tol. states for toluene and CB for chlorobenzene. b Shortest distance between the carbon cages of neighboring molecules. c SCAN+D3. All HOMO/LUMO levels are calibrated to the experimental LUMO level of PCBM.^49 d PBE/6-311G*.
C60	—	—	3.273	3.31	−5.85	−4.22
PC61BM	—	3.13 × 10^−2	3.193	3.18	−5.51	−4.03
F1	—	2.09 × 10^−4	3.236	3.27	−5.39	−3.93
F2	Tol.	4.77 × 10^−2	3.200	3.18	−5.34	−3.93
F3	—	3.25 × 10^−4	3.250	3.29	−5.30	−3.92
F4	CB	3.218 × 10^−2	3.300	—	−5.55	−4.04
F5	CB	1.037 × 10^−3	3.099	—	−5.57	−4.06
F6	—	6.623 × 10^−3	3.197	3.24	−5.49	−4.01

---

There is no direct correlation between the experimental crystal structure of the fullerene derivatives F1–F6 and their performance as ETL materials in PSCs. However, one should keep in mind that some of the fullerene derivatives were isolated in the form of solvates with toluene or chlorobenzene molecules, which strongly influence the crystal lattice. The fullerene derivatives when processed as ETL films could form solvent-free structures (e.g., due to high vacuum treatment during the device fabrication) or be represented by solvates with the molecules of the solvent used for film casting (usually chlorobenzene in this work). Thus, the real structures of the ETL films could deviate from the crystallography data we obtained in this work.

To gain a deeper fundamental understanding of the relation of the material crystal structure and photovoltaic performance,⁵⁰ we performed a variety of density functional theory (DFT) calculations (see the ESI† for details). In the case of the solid state, we list the nearest carbon–carbon distances for the fullerene dimers in the optimized unit cells and also provided the HOMO (highest occupied molecular orbital) and the LUMO (lowest unoccupied molecular orbital) energies of the functionalized fullerenes in Table 2. An analysis of the molecular dynamics (MD) trajectories obtained via the density functional based tight binding (DFTB) results was obtained by calculating the various observables applicable to semiclassical Marcus theory. This analysis revealed that, in the case of PC₆₁BM and the other functionalized fullerenes, charge transfer only occurs when the fullerene cores are in proximity to one another. In contrast, the charge transfer is much lower between the functionalizing organic chain and the fullerene, which implies that charge transfer in these systems is anisotropic and only occurs in planes of aligned fullerene cores. When comparing C₆₀, PC₆₁BM and other functionalized fullerenes, the following observations were made:

The fullerene derivatives have a higher LUMO level when compared to PC₆₁BM (Table 2), which could provide a better match for the perovskite conduction band and thus improve open circuit voltage of PSCs.

- The lambda term (associated with electron-phonon coupling) of all of the functionalized fullerenes is lower than that for PC₆₁BM (Table 2). It is also notable that the pristine fullerene has a lower value than all studied functionalized fullerenes, which indicates that the functionalization is generally unfavourable for this term due to a lower “rigidity”.

- It has been previously reported that the full rotation of a buckyball is suppressed in PC₆₁BM⁵¹ and our molecular dynamics simulations confirm this is the case for PC₆₁BM and the other functionalized fullerenes. The functionalizing group interferes in rotations and this suppresses the appearance of edge–edge alignments.

- The electron transfer integral J_ij is an order of magnitude larger when there is pentagonal face-edge alignment as compared to the situation when there is edge–edge alignment (Fig. S2, ESI†). The latter motif appears in many local minima of the energetic landscape of a 20 ps DFTB-MD simulation. It is notable that this arrangement is likely the first step of excited-state dimerization⁵²via an addition reaction. It should be noted that charge-recombination is not strongly affected by fullerene rotation.

Theoretical calculations predict the shortest C–C distance of 3.18 A between the fullerene cages in the crystal packings of PC₆₁BM and F2 molecules, which both demonstrate the highest electron mobilities in organic field-effect transistors (OFETs, Table 2). Furthermore, F2 actually provides the highest photovoltaic performance in PSCs (Table 1). At the same time, the fullerene derivatives F1 and F3 are predicted to have the longest C–C distances of 3.27–3.29 A in the lattice and both give the lowest charge carrier mobilities. Compound F6, which is predicted to have a cage-to-cage distance of 3.24 A, showed modest electron mobility in OFETs and reasonable photovoltaic performance (Tables 1 and 2). Thus, the computational results agree very well with the experimental charge carrier mobility and photovoltaic performance data.

In addition, DFTB molecular dynamics simulations revealed that the average root-mean-square deviation (RMSD) value of the trajectory and its standard deviation for the F2 film is lower than the corresponding values for F1 and PC₆₁BM (Fig. 3). This result suggests that F2 has a molecular structure facilitating hopping of charge carriers from one fullerene cage to another with much smaller energy barriers than F1 and PC₆₁BM. Thus, the results of theoretical calculations corroborate well with our experimental findings.

Fig. 3 RMSD trajectories of the first 8 ps of DFTB-MD simulations of PC₆₁BM, F1 and F2.

In the context of the trap density at the interface and defect healing, we performed steady-state photoluminescence (PL) measurements to investigate the effect of various fullerene derivatives on the perovskite absorber films. Basically, PL spectra were recorded first for pristine glass/Cs_0.12FA_0.88PbI₃ samples and then remeasured after ETL deposition, which revealed the influence of the deposited fullerene derivatives on the emission properties of the perovskite films in a straightforward way. As shown in Fig. 4a, the PL intensity of Cs_0.12FA_0.88PbI₃ was quenched after PC₆₁BM coating. In contrast, perovskite PL was enhanced after depositing F2 (Fig. 4b) thus indicating that the trap density at the Cs_0.12FA_0.88PbI₃/F2 interface is lower, which translated to the highest PCE of 19.0%. Overall, the normalized PL intensity of perovskite films with F1–F6 (I/I_o) ETLs coated on top is shown in Fig. 4c (the PL spectra are presented in Fig. S3, ESI†). It was revealed that F1–F3 and F6 are healing the defects on the surface of the perovskite films thus enhancing their PL, which could be due to the interaction of e.g. lone electron pairs of O atoms with the undercoordinated Pb²⁺ cations,^53,54 which represent the most common type of defects in lead halide perovskite films. We have also observed reproducible blue shifts of the perovskite emission band maximum after depositing thin films of fullerene derivatives: 5–6 nm for all fullerene derivatives except for F3, which induced just a minor PL peak shift by ∼1 nm. It was reported before that the blue shift of the perovskite emission band is associated with the healing of surface traps, which promote non-radiative recombination.⁵⁶ Therefore, the observed blue shift of the PL bands could be attributed to the defect passivation.

Fig. 4 PL spectra of Cs_0.12FA_0.88PbI₃ films before and after deposition of different fullerene derivatives: (a) PC₆₁BM, (b) F2, and (c) the normalized PL intensity for various glass/Cs_0.12FA_0.88PbI₃/ETL stacks (the PL intensity before ETL deposition is taken as 1).

Thus, PL spectroscopy could be used for narrowing down the range of promising charge transport materials as demonstrated previously,⁵⁵ though the defect healing ability of the fullerene derivatives cannot be directly correlated with the photovoltaic performance since it is affected by multiple other parameters, including the film morphology.

Anther important parameter is the operational stability of the fabricated PSCs. It is known that Cs_0.12FA_0.88PbI₃ represents one of the most stable absorber layer formulations among the diverse range of studied lead halide perovskites.⁵⁷ Therefore, the stability of p–i–n perovskite solar cells using this absorber material is limited by the charge transport interlayers and electrode coatings.

The operational behaviour of PSCs using various ETLs was studied for the devices assembled in ITO/PTAA/Cs_0.12FA_0.88PbI₃/ETLs/Al configuration without encapsulation. The experiments were performed under continuous light soaking (100 mW cm⁻², 45 ± 3) and open circuit conditions in an inert atmosphere, which is considered a model of an ideal encapsulation. The evolution of the normalized device PCEs upon aging is shown in Fig. 5, whereas other normalized parameters are presented in Fig. S4 (ESI).† The reference devices with PC₆₁BM exhibited a rapid degradation within ∼170 h. Fullerene derivatives F1, F3, and F6 also provided a poor stability under the same aging conditions. The other three fullerene derivatives F2, F4, and F5 enabled a considerably enhanced operational stability of PSCs. In particular, PSCs using F5 as the ETL maintained ∼60% of the initial PCE after 1300 h of aging, which is more than twice better than that of the reference cells with PC₆₁BM.

Fig. 5 Volution of the normalized PCE of PSCs with F1–F6 and PC₆₁BM as ETLs under continuous light soaking (100 mW cm⁻², 45 ± 3) in an inert atmosphere without encapsulation.

It is reasonable to presuppose that the film morphology impacted the stability of the devices under light illumination. Therefore, we performed AFM topography and infrared scattering-type scanning near-field optical microscopy (IR s-SNOM)⁵⁸ study of glass/Cs_0.12FA_0.88PbI₃/ETL samples to check the ETL coverage on the top of the Cs_0.12FA_0.88PbI₃ layer as presented in Fig. 6. The IR s-SNOM images (right column in Fig. 6) at the characteristic ETL vibration frequencies (spectra are shown in Fig. S5†) should be understood in such a way that all blue areas with the lowest signal intensity correspond to the pinholes or voids in the ETL layer atop Cs_0.12FA_0.88PbI₃ perovskite. In this context, the obtained results reveal that F4 and F5 fully cover the Cs_0.12FA_0.88PbI₃ layer with uniform and pinhole-free films. In contrast, the thin films of other fullerene derivatives demonstrated pinholes (blue areas in the right column in Fig. 6) thus revealing nonuniform Cs_0.12FA_0.88PbI₃ surface coverage. The highest density of defects in the ETL film was observed for F6, which showed inferior stability in PSCs with Cs_0.12FA_0.88PbI₃.

Fig. 6 Topography (left column) and IR s-SNOM images of glass/Cs_0.12FA_0.88PbI₃/ETL samples recorded at the perovskite (middle) and ETL (right) vibration frequencies.

The most important insight was obtained by scanning the glass/Cs_0.12FA_0.88PbI₃/ETL samples at a characteristic perovskite vibration frequency of 1709 cm⁻¹ (middle column in Fig. 6). These images should be interpreted in such a way that the appearance of high-intensity signals (red spots) reveals the perovskite phase, which is exposed to the surface through the pinholes in the ETL layer. This kind of defects results in a direct contact of perovskite with the top electrode, which is expected to lead to its fast and severe corrosion.

Indeed, the stacks using in particular F6 and to a lesser extent F1 and PC₆₁BM showed these deep pinholes exposing the perovskite phase on the surface, which translated to the poor operational stability of PSCs using these fullerene derivatives as ETLs. Meanwhile, the stacks using fullerene derivative F5 showed no detectable pinholes in the ETL layer and, consequently, the corresponding devices demonstrated the best operational stability. Thus, these results reveal a crucial effect of the nanoscale film morphology and uniformity on the operational stability of p–i–n PSCs. We have recently reported similar findings featuring the effect of hole-transport layer morphology on the efficiency and stability of n–i–p perovskite solar cells.^58,59

Conclusions

In summary, we synthesized and structurally characterized a series of fullerene derivatives F1–F6, which were investigated as promising electron transport materials for p–i–n perovskite solar cells. The device efficiencies obtained using F1–F6 as ETLs were comparable or higher than those of the reference cells assembled using the benchmark fullerene-based material PC₆₁BM. The improvement in the photovoltaic performance of the devices incorporating new fullerene derivatives, most likely, originated from the optimized molecular structure, which we further investigated with DFT and DFT-MD simulations. DFTB molecular dynamics simulations showed that the average RMSD value and its standard deviation for the best-performing fullerene derivative F2 are considerably lower than the corresponding value for PC₆₁BM This result suggests that F2 has an appropriate molecular structure facilitating hopping of charge carriers from one fullerene cage to another with much smaller energy barriers than in the case of PC₆₁BM. Other important finding was the strongly enhanced photostability of PSCs using F2, F4, and F5 as ETLs. In particular, the devices assembled with the fluorine-loaded fullerene derivative F5 maintained >60% of the initial PCE after 1300 h of continuous 1 sun equivalent light illumination under open circuit conditions, which is twice better than that for the reference cells using PC₆₁BM as the ETL material. IR s-SNOM microscopy revealed a good relationship between the operational stability of PSCs using different fullerene derivatives and uniformity of their films. Thus, our study has shown that tuning the molecular structures of the functionalized fullerene derivatives could improve the PSC performance, whereas achieving long-term operational stability requires optimal ETL film morphology. These findings could be considered useful guidelines for designing new charge-transport materials for efficient and stable p–i–n perovskite solar cells.

Conflicts of interest

No conflicts of interest are reported here.

Acknowledgements

This work was supported by the Russian Science Foundation (project no. 22-73-10138). L. G. Gutsev thanks the Louisiana Optical Network Infrastructure (LONI) for the computational infrastructure used to complete this project. L. G. Gutsev acknowledges the Russian Science Foundation Grant No. 21-73-00080 for financially supporting the theoretical portion of this work. X-ray diffraction data for F6 were collected using the equipment of the Center for molecular composition studies of INEOS RAS supported by the Ministry of Science and Higher Education of the Russian Federation (contract/agreement no. 075-03-2023–642).

References

1. S. Gharibzadeh, P. Fassl, I. M. Hossain, P. Rohrbeck, M. Frericks, M. Schmidt, T. Duong, M. R. Khan, T. Abzieher, B. A. Nejand, F. Schackmar, O. Almora, T. Feeney, R. Singh, D. Fuchs, U. Lemmer, J. P. Hofmann, S. A. L. Weber and U. W. Paetzold, Energy Environ. Sci., 2021, 14, 5875–5893.
2. Best Research-Cell, Efficiency Chart | Photovoltaic Research | NREL, https://www.nrel.gov/pv/cell-efficiency.html, accessed 25 March 2023.
3. X. Lin, D. Cui, X. Luo, C. Zhang, Q. Han, Y. Wang and L. Han, Energy Environ. Sci., 2020, 13, 3823–3847.
4. S. Collavini and J. L. Delgado, Sustainable Energy Fuels, 2018, 2, 2480–2493.
5. M. Degani, Q. An, M. Albaladejo-Siguan, Y. J. Hofstetter, C. Cho, F. Paulus, G. Grancini and Y. Vaynzof, Sci. Adv., 2023, 7, eabj7930.
6. L. Wan, W. Zhang, S. Fu, L. Chen, Y. Wang, Z. Xue, Y. Tao, W. Zhang, W. Song and J. Fang, J. Mater. Chem. A, 2020, 8, 6517–6523.
7. M. Zhang, Q. Chen, R. Xue, Y. Zhan, C. Wang, J. Lai, J. Yang, H. Lin, J. Yao, Y. Li, L. Chen and Y. Li, Nat. Commun., 2019, 10, 4593.
8. W. Chen, G. ning Zhang, L. ming Xu, R. Gu, Z. he Xu, H. jiang Wang and Z. bing He, Mater. Today Energy, 2016, 1–2, 1–10.
9. M. Elnaggar, M. Elshobaki, A. Mumyatov, S. Y. Luchkin, N. N. Dremova, K. J. Stevenson and P. A. Troshin, Sol. RRL, 2019, 3, 1900223.
10. N. A. Cooling, E. F. Barnes, F. Almyahi, K. Feron, M. F. Al-Mudhaffer, A. Al-Ahmad, B. Vaughan, T. R. Andersen, M. J. Griffith, A. S. Hart, A. G. Lyons, W. J. Belcher and P. C. Dastoor, J. Mater. Chem. A, 2016, 4, 10274–10281.
11. E. M. Younes, A. Gurung, B. Bahrami, E. M. El-Maghraby and Q. Qiao, Carbon, 2021, 180, 226–236.
12. J. H. Heo, S.-C. Lee, S.-K. Jung, O.-P. Kwon and S. H. Im, J. Mater. Chem. A, 2017, 5, 20615–20622.
13. Y. R. Kim, C. M. Oh, C. J. Yoon, J.-H. Kim, K. Park, K. Lee, I.-W. Hwang and H. Kim, Nanoscale, 2021, 13, 5652–5659.
14. M. Elnaggar, A. G. Boldyreva, M. Elshobaki, S. A. Tsarev, Y. S. Fedotov, O. R. Yamilova, S. I. Bredikhin, K. J. Stevenson, S. M. Aldoshin and P. A. Troshin, Sol. RRL, 2020, 4, 2000191.
15. A. F. Akbulatov, L. A. Frolova, M. P. Griffin, I. R. Gearba, A. Dolocan, D. A. Vanden Bout, S. Tsarev, E. A. Katz, A. F. Shestakov, K. J. Stevenson, P. A. Troshin, A. F. Akbulatov, L. A. Frolova, A. F. Shestakov, P. A. Troshin, M. P. Griffin, D. A. Vanden Bout, S. Tsarev, K. J. Stevenson and E. A. Katz, Adv. Energy Mater., 2017, 7, 1700476.
16. H. Back, G. Kim, J. Kim, J. Kong, T. K. Kim, H. Kang, H. Kim, J. Lee, S. Lee and K. Lee, Energy Environ. Sci., 2016, 9, 1258–1263.
17. H. Lee and C. Lee, Adv. Energy Mater., 2018, 8, 1702197.
18. S. Taheri, A. Ahmadkhan kordbacheh, M. Minbashi and A. Hajjiah, Opt. Mater., 2021, 111, 110601.
19. X. Zhang, M. E. Turiansky, J. X. Shen and C. G. Van De Walle, J. Appl. Phys., 2022, 131, 090901.
20. J. Xu, A. Buin, A. H. Ip, W. Li, O. Voznyy, R. Comin, M. Yuan, S. Jeon, Z. Ning, J. J. McDowell, P. Kanjanaboos, J. P. Sun, X. Lan, L. N. Quan, D. H. Kim, I. G. Hill, P. Maksymovych and E. H. Sargent, Nat. Commun., 2015, 6, 1–8.
21. L. A. Frolova, S. Y. Luchkin, Y. Lekina, L. G. Gutsev, S. A. Tsarev, I. S. Zhidkov, E. Z. Kurmaev, Z. X. Shen, K. J. Stevenson, S. M. Aldoshin and P. A. Troshin, Adv. Energy Mater., 2021, 11, 2002934.
22. Y. Yu, F. Zhang and H. Yu, Sol. Energy, 2020, 209, 408–414.
23. C. Chen, C. Wu, X. Ding, Y. Tian, M. Zheng, M. Cheng, H. Xu, Z. Jin and L. Ding, Nano Energy, 2020, 71, 104604.
24. C. Y. Chang, B. C. Tsai, Y. C. Hsiao, Y. C. Huang and C. S. Tsao, Phys. Chem. Chem. Phys., 2016, 18, 31836–31844.
25. Z. Xing, F. Liu, S. H. Li, Z. C. Chen, M. W. An, S. Zheng, A. K. Y. Jen and S. Yang, Adv. Funct. Mater., 2021, 31, 2107695.
26. X. Meng, C. H. Y. Ho, S. Xiao, Y. Bai, T. Zhang, C. Hu, H. Lin, Y. Yang, S. K. So and S. Yang, Nano Energy, 2018, 52, 300–306.
27. S. Wu, Z. Li, M. Q. Li, Y. Diao, F. Lin, T. Liu, J. Zhang, P. Tieu, W. Gao, F. Qi, X. Pan, Z. Xu, Z. Zhu and A. K. Y. Jen, Nat. Nanotechnol., 2020, 15, 934–940.
28. W. Chen, Y. Wu, Y. Yue, J. Liu, W. Zhang, X. Yang, H. Chen, E. Bi, I. Ashraful, M. Grätzel and L. Han, Science, 2015, 350, 944–948.
29. S. Bai, P. Da, C. Li, Z. Wang, Z. Yuan, F. Fu, M. Kawecki, X. Liu, N. Sakai, J. T. W. Wang, S. Huettner, S. Buecheler, M. Fahlman, F. Gao and H. J. Snaith, Nat, 2019, 571, 245–250.
30. K. A. Bush, A. F. Palmstrom, Z. J. Yu, M. Boccard, R. Cheacharoen, J. P. Mailoa, D. P. McMeekin, R. L. Z. Hoye, C. D. Bailie, T. Leijtens, I. M. Peters, M. C. Minichetti, N. Rolston, R. Prasanna, S. Sofia, D. Harwood, W. Ma, F. Moghadam, H. J. Snaith, T. Buonassisi, Z. C. Holman, S. F. Bent and M. D. McGehee, Nat. Energy, 2017, 2, 17009.
31. E. Castro, J. Murillo, O. Fernandez-Delgado and L. Echegoyen, J. Mater. Chem. C, 2018, 6, 2635–2651.
32. J. Lian, B. Lu, F. Niu, P. Zeng and X. Zhan, Small Methods, 2018, 2, 1800082.
33. L. Jia, M. Chen and S. Yang, Mater. Chem. Front., 2020, 4, 2256–2282.
34. S. Huang, G. Zhang, N. S. Knutson, M. T. Fontana, R. C. Huber, A. S. Ferreira, S. H. Tolbert, B. J. Schwartz and Y. Rubin, J. Mater. Chem. A, 2015, 4, 416–424.
35. F. Matsumoto, T. Iwai, K. Moriwaki, Y. Takao, T. Ito, T. Mizuno and T. Ohno, J. Org. Chem., 2012, 77, 9038–9043.
36. M. Karakawa, T. Nagai, K. Adachi, Y. Ie and Y. Aso, RSC Adv., 2017, 7, 7122–7129.
37. F. Matsumoto, T. Iwai, K. Moriwaki, Y. Takao, T. Ito, T. Mizuno and T. Ohno, ACS Appl. Mater. Interfaces, 2016, 8, 4803–4810.
38. A. V. Mumyatov, F. A. Prudnov, L. N. Inasaridze, O. A. Mukhacheva and P. A. Troshin, J. Mater. Chem. C, 2015, 3, 11612–11617.
39. F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M. Kroon and J. C. Hummelen, Org. Lett., 2007, 9, 551–554.
40. A. V. Mumyatov, A. E. Goryachev, F. A. Prudnov, O. A. Mukhacheva, D. K. Sagdullina, A. V. Chernyak, S. I. Troyanov and P. A. Troshin, Synth. Met., 2020, 270, 116565.
41. P. A. Troshin, E. A. Khakina, M. Egginger, A. E. Goryachev, S. I. Troyanov, A. Fuchsbauer, A. S. Peregudov, R. N. Lyubovskaya, V. F. Razumov and N. S. Sariciftci, ChemSusChem, 2010, 3, 356–366.
42. J. C. Hummelen, B. W. Knight, F. Lepeq, F. Wudl, J. Yao and C. L. Wilkins, J. Org. Chem., 1995, 60, 532–538.
43. P. A. Troshin, D. K. Susarova, E. A. Khakina, A. A. Goryachev, O. V. Borshchev, S. A. Ponomarenko, V. F. Razumov and N. S. Sariciftci, J. Mater. Chem., 2012, 22, 18433–18441.
44. P. A. Troshin, E. A. Khakina, M. Egginger, A. E. Goryachev, S. I. Troyanov, A. Fuchsbauer, A. S. Peregudov, R. N. Lyubovskaya, V. F. Razumov and N. S. Sariciftci, ChemSusChem, 2010, 3, 356–366.
45. M. Casalegno, S. Zanardi, F. Frigerio, R. Po, C. Carbonera, G. Marra, T. Nicolini, G. Raos and S. V. Meille, Chem. Commun., 2013, 49, 4525–4527.
46. O. Fernandez-Delgado, P. S. Chandrasekhar, N. Cano-Sampaio, Z. C. Simon, A. R. Puente-Santiago, F. Liu, E. Castro and L. Echegoyen, J. Mater. Chem. C, 2021, 9, 10759–10767.
47. Y. Zhang, C. R. Zhang, L. H. Yuan, M. L. Zhang, Y. H. Chen, Z. J. Liu and H. S. Chen, Mater. Chem. Phys., 2018, 204, 95–104.
48. N. C. Miller, S. Sweetnam, E. T. Hoke, R. Gysel, C. E. Miller, J. A. Bartelt, X. Xie, M. F. Toney and M. D. McGehee, Nano Lett., 2012, 12, 1566–1570.
49. C. J. Brabec, A. Cravino, D. Meissner, N. S. Sariciftci, T. Fromherz, M. T. Rispens, L. Sanchez and J. C. Hummelen, Adv. Funct. Mater., 2001, 11, 374–380.
50. Z. Xing, S. H. Li, P. Y. Xu, H. R. Tian, L. L. Deng, Y. R. Yao, B. W. Chen, F. F. Xie, M. W. An, D. Q. Yun, S. Y. Xie and L. S. Zheng, Chem. Res. Chin. Univ., 2022, 38, 75–81.
51. L. Chen, X. Wang and S. Kumar, Sci. Rep., 2015, 5, 1–10.
52. I. Ramirez, A. Privitera, S. Karuthedath, A. Jungbluth, J. Benduhn, A. Sperlich, D. Spoltore, K. Vandewal, F. Laquai and M. Riede, Nat. Commun., 2021, 12, 471.
53. M. J. Choi, Y. S. Lee, I. H. Cho, S. S. Kim, D. H. Kim, S. N. Kwon and S. I. Na, Nano Energy, 2020, 71, 104639.
54. T. Sen Su, F. T. Eickemeyer, M. A. Hope, F. Jahanbakhshi, M. Mladenović, J. Li, Z. Zhou, A. Mishra, J. H. Yum, D. Ren, A. Krishna, O. Ouellette, T. C. Wei, H. Zhou, H. H. Huang, M. D. Mensi, K. Sivula, S. M. Zakeeruddin, J. V. Milić, A. Hagfeldt, U. Rothlisberger, L. Emsley, H. Zhang and M. Grätzel, J. Am. Chem. Soc., 2020, 142, 19980–19991.
55. A. Al-Ashouri, E. Köhnen, B. Li, A. Magomedov, H. Hempel, P. Caprioglio, J. A. Márquez, A. B. M. Vilches, E. Kasparavicius, J. A. Smith, N. Phung, D. Menzel, M. Grischek, L. Kegelmann, D. Skroblin, C. Gollwitzer, T. Malinauskas, M. Jošt, G. Matič, B. Rech, R. Schlatmann, M. Topič, L. Korte, A. Abate, B. Stannowski, D. Neher, M. Stolterfoht, T. Unold, V. Getautis and S. Albrecht, Science, 2020, 370, 1300–1309.
56. Y. Shao, Z. Xiao, C. Bi, Y. Yuan and J. Huang, Nat. Commun., 2014, 5, 5784.
57. A. F. Akbulatov, M. I. Ustinova, L. Gutsev, S. A. Tsarev, N. N. Dremova, I. Zhidkov, S. Y. Luchkin, B. R. Ramachandran, L. Frolova, E. Z. Kurmaev, K. J. Stevenson, S. M. Aldoshin and P. A. Troshin, Nano Energy, 2021, 86, 106082.
58. A. F. Latypova, N. A. Emelianov, D. O. Balakirev, P. K. Sukhorukova, N. K. Kalinichenko, P. M. Kuznetsov, Y. N. Luponosov, S. M. Aldoshin, S. A. Ponomarenko, P. A. Troshin and L. A. Frolova, ACS Appl. Energy Mater., 2022, 5, 5395–5403.
59. A. F. Latypova, A. V. Maskaev, L. G. Gutsev, N. A. Emelianov, I. E. Kuznetsov, P. M. Kuznetsov, S. L. Nikitenko, Y. V. Baskakova, A. V. Akkuratov, E. A. Komissarova, L. A. Frolova, S. M. Aldoshin and P. A. Troshin, Mater. Today Chem., 2022, 26, 101218.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 2250001, 2250002 and 2252121. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3se00432e

---