Designing New Indene-Fullerene Derivatives as Electron-Transporting Materials for Flexible Perovskite Solar Cells.

The synthesis and characterization of a family of indene-C60 adducts obtained via Diels-Alder cycloaddition [4 + 2] are reported. The new C60 derivatives include indenes with a variety of functional groups. These adducts show lowest unoccupied molecular orbital energy levels to be at the right position to consider these compounds as electron-transporting materials for planar heterojunction perovskite solar cells. Selected derivatives were applied into inverted (p-i-n configuration) perovskite device architectures, fabricated on flexible polymer substrates, with large active areas (1 cm2). The highest power conversion efficiency, reaching 13.61%, was obtained for the 6'-acetamido-1',4'-dihydro-naphtho[2',3':1,2][5,6]fullerene-C60 (NHAc-ICMA). Spectroscopic characterization was applied to visualize possible passivation effects of the perovskite's surface induced by these adducts.

Introduction

Since milligram-scale syntheses of fullerene (C60) got into gear in the early 1990s,[1−5] many research groups provided evidence of their remarkable structural,[6,7] magnetic,[8] superconducting,[9] electrochemical,[10] photophysical,[11] and biological properties.[12,13] Currently, fullerenes are being utilized in a wide range of applications, including photovoltaics, light-emitting devices, modern antiviral therapies, or space exploration.[14−18] In particular, fullerenes and a wide range of their derivatives have been successfully used as electron transport materials (ETMs), at first in organic solar cells, and more recently in perovskite solar cells (PSCs).[19−21] This family of compounds is characterized by good electron-accepting properties (effective electron extraction from a photoabsorber), low-temperature processing, and suitable energy levels, enabling the role of electron-selective contact in these photovoltaic technologies. In PSCs, C60 and its derivatives, such as phenyl-C61-butyric acid methyl ester (PCBM), were predominantly used in a p–i–n device configuration (so called, “inverted” architecture), yielding high power conversion efficiencies (PCEs).[22−25] The applicability of the pristine fullerene molecules is limited by solubility constraints in the most common solvents.[26] Modifications of their chemical structure are required to solve this drawback. There are two general approaches toward the fullerene functionalization—covalent and non-covalent interactions.[27−29] The covalent-based approach overall gives more options for a chemical modification. It can be realized by taking advantage of the electrophilic nature of the C60 or using conjugated π-electrons (particularly the [5,6] double bonds),[30] which enables the molecule to undergo Diels–Alder reactions ([4 + 2] cycloadditions), acting as a dienophile.[31−37] These two well-known synthetic pathways provided readily accessible, often quite sophisticated, C60-based building blocks for numerous application fields. The fullerene adducts can be categorized in classes depending on the number of carbon atoms of the substituent, as schematically reported in Figure a. The full carbon ring of C60 shows good performance in PSCs because of high electrochemical stability of its negatively charged reduction products.[38,39] The three-membered carbon ring constitutes one of the most investigated types of a fullerene appendage. The C60 fullerenes with a five-membered carbon ring, namely, indano-C60, are still relatively unexplored.[40] The indene-C60 adducts were synthesized by the Diels–Alder cycloaddition.[31,37] High yields can be obtained for this methodology, but the scope of molecules which can be synthesized so far has been limited by availability of appropriate indene derivative precursors. The first indenyl derivative, the fullerene-indene-C60 bisadduct (ICBA), was reported in 2010.[41] It was designed and synthesized as an alternative to PCBM, for use in polymer solar cells.[41] In 2013, Chen was the first to use ICBA as an electron-transporting material (ETM) in a planar heterojunction p–i–n PSC architecture. Despite the promising results obtained for these indene–fullerene adducts, the topic was not further explored due to the aforementioned limitations. Herein, we report a methodology to obtain indene-fullerene adducts based on simple and cheap substrates. We synthesized a group of indene derivatives, equipped with functional groups of different electronic natures, electron-releasing (e.g., −OMe and −NH2) and electron-withdrawing moieties (e.g., −CN). We used phenol derivatives as starting materials and completed the process in five steps, with the final yield of 50–60% (Figure b–e).[42] To the best of our knowledge, this is the first demonstration of this method in the synthesis of indene-fullerene derivatives. Subsequently, the newly made indenes were employed in the synthesis of several indene-fullerene C60 derivatives. We characterized basic photophysical and electrochemical properties of these compounds as a preliminary assessment of their device implementation potential. Lastly, we incorporated selected fullerene derivatives to the inverted PSC architecture as an ETL and compared the effect of different fullerene modifying groups.

Figure 1

Depiction of a general strategy for obtaining substituted indene-fullerene adducts, (a) possible geometric hydrocarbon motifs appended to a 5, 6 ring junction of C60, (b) retrosynthetic breakdown of an indene-fullerene adduct, (c) reaction pathway to obtain substituted indenes, (d) Diels–Alder [4 + 2] cycloaddition of substituted indenes with C60, and (e) synthesis of a free-amine indene-fullerene derivative.

Results and Discussion

We synthesized a series of substituted indenes in a five-step process, which we schematically present in Figure c. The starting material in the reaction sequence was allyl bromide and an appropriate phenol derivative: 2-naphthol, 4-methoxyphenol, 4-cyanphenol, or 4-acetamidephenol. The substituted indene precursors, 6a–6e, obtained were subsequently transformed by a ring-closing metathesis to afford an excellent yield, as shown in Figure b. Full experimental data and spectroscopic characterizations are provided in the Supporting Information, Section 1.3 and 1.4.3.

Finally, indene-fullerene adducts were obtained in a Diels–Alder reaction. First, at the reflux temperature, indene is undergoing [1,2]-hydrogen shift, yielding isoindene, which subsequently reacts in the Diels–Alder [4 + 2] cycloaddition (Figure d).[43] The final products, Benzo-ICMA, MeO-ICMA, CN-ICMA, NHAc-ICMA, and NHAc-Me-ICMA, were synthesized by carrying out this process in ortho-dichlorobenzene (o-DCB), at 180 °C for 48 h. All the final adducts were purified by column chromatography in toluene/ethyl acetate, except for the MeO-ICMA. This product is metastable and decomposed to the indene precursor and C60. For the amine-substituted indene, in the early steps of the proposed synthesis route, the 4-aminophenol can be transformed to undesired byproducts. Therefore, a fullerene derivative with the amino group had to be synthesized from a fullerene acetamide derivative, Figure e. There are many examples of such post-modification of functionalized fullerenes in the literature.[44−46] We obtained this derivative reaching a high yield of 75%. The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) levels of fullerene derivatives are of great importance for considering their use as electron acceptors and hole-blocking materials in PSCs.[47,48] We derived experimental values of these energy levels by cyclic voltammetry (CV) measurements (Figure a–c), which are presented in Table ; the methodology is further elaborated in the Supporting Information, Section 1.2.1. The CV measurements of all the considered fullerene derivatives showed successive reversible reduction steps, typical of fullerene electrochemistry (Figure a). The functionalization of the fullerene core with the indene moiety shifts the onset of the first reduction potential to more negative values when compared to the pristine C60,[49] indicating an electron-releasing effect of the indene moiety (see Figures S4–S10). The cathodic CV trace of PCBM features four reversible redox steps, corresponding to successive single-electron charging of the fullerene’s π-conjugated structure. By comparing the CV traces of the investigated compounds with PCBM, it is possible to discriminate redox peaks of the fullerene core from other signatures, characteristic for each indene-fullerene derivatives.

Figure 2

Cyclic voltammograms of (a) reduction and (b) oxidation processes of fullerene compounds, recorded in a solution of 0.05 M tetrabutylammonium hexafluorophosphate in o-DCB; (c) HOMO and LUMO energies and shapes of the investigated substituted indene-fullerene adducts.

Table 1

Electrochemically and Theoretically Derived Electronic Properties of the Fullerene Derivative Compounds

compound	Eredonset [V]	Eoxonset [V]	EAa,b [eV]	IPa,b [eV]	Egb [eV]	LUMOc [eV]	HOMOc [eV]	H–L gapc [eV]
PCBM	–1.13	1.00	3.97 (3.59)	6.10 (5.89)	2.13 (2.30)	–3.50	–6.00	2.50
Benzo-ICMA	–1.05	0.98	4.05 (3.57)	6.08 (5.85)	2.03 (2.28)	–3.48	–5.96	2.48
MeO-ICMA	–1.05	1.10	4.05 (3.56)	6.20 (5.84)	2.15 (2.28)	–3.47	–5.95	2.48
CN-ICMA	–1.05	0.94	4.05 (3.60)	6.04 (5.90)	1.99 (2.30)	–3.51	–6.00	2.49
NHAc-ICMA	–1.08	0.87	4.02 (3.57)	5.97 (5.84)	1.95 (2.27)	–3.47	–5.95	2.48
NHAc-Me-ICMA	–1.04	0.82	4.06 (3.56)	5.92 (5.83)	1.86 (2.27)	–3.47	–5.94	2.47
NH2-ICMA	–1.02	0.43	4.08 (3.56)	5.53 (5.53)	1.45 (1.97)	–3.47	–5.81	2.34

EA and IP values for the ferrocene/ferrocenium (Fc/Fc+) redox couple (5.10 eV above the vacuum level) as an internal standard.

Eg is calculated from the difference between the EA and IP from experimental data.

Calculated by DFT at the B3LYP/6-311++G** level of theory with o-DCB as an implicit solvent. Theoretical values are given in parentheses, when reported along with experimental data.

In the anodic branch, all the investigated fullerenes oxidize irreversibly (Figure b), displaying onset potentials ranging from 0.43 V (for the NH-ICMA) up to 1.10 V (for the MeO-ICMA) (see Table , Figures S7 and S8). This variation in the oxidation propensity, however, does not correlate with the electron-withdrawing or electron-releasing character of the indene moiety’s functional groups. Except for the MeO-ICMA, all the other investigated compounds oxidize at lower potentials than PCBM, indicating stronger electron communication of the indene pendant with the fullerene host than of the phenyl ring in PCBM. Interestingly, this communication facilitates electron abstraction (oxidation) from the fullerene-indene derivatives when compared to PCBM, which may be puzzling, since an analogous effect was observed for the process of electron introduction (reduction). This might point to a through-space interaction of the π-electrons of the fullerene and π-electrons of the indene units, sterically locked in position and inclined at a smaller angle above the fullerene core than the rotationally unrestricted phenyl pendant of PCBM (see Figure S11, Supporting Information, Section 1.4.1). The comparable oxidation onset potentials recorded for NHAc-ICMA and NHAc-Me-ICMA indicate a marginal inductive electronic effect of the functionalization of the ternary (>CH−) or quaternary (>C<) carbon bridge atoms, which interface the fullerene and phenylene units. Considering the ferrocene/ferrocenium (Fc/Fc+) redox couple (5.10 eV below the vacuum level) as the internal standard, we determined experimentally the electron affinity (EA) and first ionization potential (IP). All the values, including those for the PCBM molecule, are listed in Table . To gain insights into the electronic properties of the fullerene derivates, we also performed density functional theory (DFT) calculations to evaluate the EA, IP, the electrochemical band gap, and the shape of the HOMO and LUMO orbitals. From the Koopmans theorem and approximation,[50] we compared the EA and IP with the calculated HOMO and LUMO values; the data and additional discussion about applied methodology are provided in Figures S12 and S13, Supporting Information, Sections 1.2.8 and 1.4.2. The CV measurements are in line with the theoretical predictions of the delocalization of the LUMOs in the studied set of molecules (see Figures c and S3). Additionally, in Table , we summarize the calculated HOMO values, which display good agreement with the measured IP values. On the other hand, an overestimation of the EA is found when compared to the LUMO values, and this consequentially leads to overestimation of the calculated HOMO–LUMO (H–L) band gap. All the investigated adducts show similar IP values, with the exception of the NH-ICMA. The reason is that the HOMO level of this structure is destabilized and almost completely localized at the appendage, that is, on the aniline fragment of the indene moiety. The peculiar electronic properties of the NH-ICMA are also reflected in its lower IP value taken from CV measurements, compared to the other compounds. This suggests the higher oxidation tendency of this compound. The calculated electrochemical energy gaps are almost the same for all the studied compounds; the only exception is the NH-ICMA molecule, which displays a lower band gap, in line with the experimental findings.

Based on the electronic properties, we selected the following compounds for further evaluation as potential ETMs in PSCs: the amide derivatives NHAc-ICMA and NHAc-Me-ICMA and the amine derivative NH-ICMA (Figure a).[48]

Figure 3

(a) Schematic structures of the indene-fullerene adducts applied in PSCs, (b) AFM 3D surface topography images (scanning range: 25 × 25 μm2) of the perovskite layer; and (c,e,g,i) perovskite coated with indene-fullerene films (NHAc-ICMA, NH-ICMA, NHAc-Me-ICMA, and PCBM). (d,f,h,j) Cross-sectional FIB-SEM images of the perovskite layer coated with different fullerenes prepared on flexible substrates.

Within this set of fullerene derivatives, we wanted to explore a potential passivation effect of the electronic defects present at the perovskite thin-film surface. It was shown that the most effective passivating agents are simultaneously inactivating both negatively and positively charged defects, exhibiting a zwitterionic effect.[51,52] The fullerene moiety can act as a Lewis acid and passivate negatively charged defects (e.g., undercoordinated I ions and Pb–I anti-sites).[53] The amine bearing –NH tail and the amide group –NHAc are Lewis bases and play a significant role in passivating positively charged defects (e.g., undercoordinated Pb2+ and Pb2+ interstitials).[54] To investigate these effects, we first characterized electrical properties of the selected fullerene thin films. We coated layers of these materials on poly(ethylene naphthalate)/indium tin oxide (ITO) substrates, with a pre-patterned narrow trench in the ITO layer. We extracted conductivity values from the current–voltage curves (shown in Figure S2, Supporting Information). We obtained 3.20 × 10–5, 1.84 × 10–5, and 3.22 × 10–5 S cm–1 for NHAc-ICMA, NH-ICMA, and NHAc-Me-ICMA, respectively. For the reference PCBM, we recorded a comparable value of 2.49 × 10–5 S cm–1. These results are consistent with the PCBM conductivity values reported in the literature.[55] Next, we derived electron mobility values of the indene-fullerene adducts from the space charge-limited current measurements of electron-only devices, applying the Mott–Gurney law.[56] We used the following device architecture: ITO/TiO/ETL (varied fullerene derivatives)/TiO/Ag (experimental details are provided in the Supporting Information). We extracted 1.50 × 10–3, 1.21 × 10–3, 2.24 × 10–3, and 3.40 × 10–3 cm–2 V–1 s–1 for NHAc-ICMA, NH-ICMA, NHAc-Me-ICMA, and PCBM, respectively. In the literature, values reported for the PCBM films are typically in the range of 10–3 cm–2 V–1 s–1.[57−60] The group of new indene-fullerene adducts shows slightly worse electron mobilities, which could influence transport characteristics in these films.

We also investigated the layer formation ability of the indene-fullerene derivatives. We fabricated thin films of different fullerene samples by spin coating relevant solutions on top of perovskite films (glass/perovskite substrates). As a perovskite material, we used the composition of mixed cations and mixed halides, Cs0.04(MA0.17FA0.83)0.96Pb(I0.83Br0.17)3, which was deposited using a solvent engineering strategy, following the method reported in our previous work.[24] The layer morphologies were characterized by atomic force microscopy (AFM), and we derived the root-mean-square (RMS) roughness for each sample; for characterization details, see the Supporting Information, Section 1.2.3. The 3D surface topography and cross-sectional focused ion beam scanning electron microscopy (FIB-SEM) images are shown in Figure b–j. The bare perovskite layer and the perovskite/NHAc-Me-ICMA samples show an RMS of 11.9 and 11.3 nm, respectively. The other fullerene derivatives, NHAc-ICMA, NH-ICMA, and PCBM, display RMS values of 5.4, 99.6, and 23.5 nm, respectively. Additionally, we note that NHAc-ICMA and NHAc-Me-ICMA demonstrated significantly higher solubility in o-DCB compared to NH-ICMA. It is also evident from the FIB-SEM cross-sectional images that NHAc-ICMA displays more conformal and uniform coverage over the perovskite surface than the samples with NH-ICMA, NHAc-Me-ICAM, and PCBM. More characterization details are provided in the Supporting Information, Section 1.2.4.

To probe the possible perovskite surface passivation effect by the selected indene-fullerene derivatives, we performed absolute photoluminescence measurements (photoluminescence quantum yield, PLQY) of these samples; experimental details are described in the Supporting Information, Section 1.2.5. We applied perovskite thin films, which were deposited directly on a glass substrate, followed by coating different ETMs on top (PCBM and newly developed derivatives: NH-ICMA, NHAc-ICMA, and NHAc-Me-ICMA). In this way, we could compare the amount of non-radiative recombination losses originating at the perovskite/ETL contact. The results are presented in Figure a. The bare perovskite layer shows the highest PL intensity. The addition of a fullerene layer partially quenches the signal, primarily due to the increased non-radiative recombination at the perovskite/ETM interface.

Figure 4

(a) Summary of the PLQY and calculated QFLS values determined from the spectroscopic measurements. We also include the implied open-circuit voltages determined for the solar cells representing the different fullerenes, (b) current density–voltage characteristics (light and dark) of the best PSCs with different fullerene adducts, and (c) stabilized power output (SPO) measurement of the same devices.

In Figure a, we provide the summary of the PLQY results, together with the calculated quasi-Fermi level splitting (QFLS) values (corresponding to attainable VOC values under given illumination conditions). It is evident that all the studied ETMs, when coated on top of the perovskite layer, decrease its QFLS. The bare perovskite film displayed a PLQY of ∼0.00383% (measured at 0.05 Sun), which corresponds to the QFLS value of 1.16 eV. Compared to the bare perovskite film, the addition of a fullerene layer causes a significant decrease in the photoluminescence efficiency and gives rise to non-radiative recombination losses. The QFLS obtained for NHAc-ICMA, NH-ICMA, NHAc-Me-ICMA, and PCBM samples is 1.13, 1.12, 1.14, and 1.15 eV, respectively. From this, we can infer that in the set of different indene fullerenes, the NHAc-Me-ICMA compound displays the lowest amount of interfacial recombination losses. These losses were even lower for the PCBM reference sample, with a QFLS of 1.15 eV. We also performed time-resolved photoluminescence (TRPL) decay measurements for the same set of samples (Supporting Information 1.2.6). The curves are shown in Figure S1, Supporting Information. The TRPL signal is affected by charge extraction, charge trapping, and recombination processes.[61] Due to low-intensity (5 mW/cm2) and high-energy photons (405 nm) of the excitation beam (laser goes through the fullerene side), the initial decay is likely to be driven primarily by charge trapping and electron extraction.[61,62] All the ETLs show comparable decays, with the PCBM displaying the fastest and NH-ICMA displaying the slowest early time (first 50 ns) quenching. In view of the highest QFLS for the PCBM, this indicates superior electron extraction in this sample. Structural variations in the fullerene derivatives can strongly influence the electronic contact with the perovskite layer, even when the respective energy levels are unchanged. Size effects of the bulky functional groups can also influence intermolecular interactions of the neighboring carbon cages, which in turn could reflect in the reduction of electron mobility values, leading to electron extraction difficulties.[63]

In order to elucidate the relationship between the molecular structure of the indene-fullerene derivatives and their operation as ETMs in PSCs, we fabricated photovoltaic devices of p–i–n configuration; for experimental details and characterization techniques, see the Supporting Information, Sections 1.1 and 1.2.2.[34] In Table , we report average values of the photovoltaic parameters extracted from the current density–voltage (J–V) characterization measurements; the parameters for the champion devices are given in brackets. The respective J–V curves of the best devices for each fullerene derivative are shown in Figure b. We also measured the spectral response of the representative devices for all the ETM variations. The external quantum efficiency spectra with integrated current density values (similar to the values obtained from the J–V measurements) are shown in Figure S2, Supporting Information.

Table 2

Photovoltaic Parameters Extracted from the Current–Voltage Characterization Measurements of the PSCs Fabricated with Different Fullerene Adducts

fullerene	PCEavg±SD [%]	FFavg±SD [%]	VOCavg±SD [V]	JSCavg±SD [mA cm–2]	SPO [%]
NHAc-ICMA	(13.61) 10.71 ± 2.27	(68.14) 60.07 ± 5.55	(1.00) 0.95 ± 0.04	(20.07) 18.59 ± 0.87	9.97
NH2-ICMA	(10.12) 6.99 ± 1.86	(66.10) 53.92 ± 7.08	(0.90) 0.88 ± 0.10	(17.03) 14.58 ± 1.59	2.75
NHAc-Me-ICMA	(11.98) 7.64 ± 2.51	(60.01) 47 ± 7.83	(1.02) 0.94 ± 0.06	(19.54) 16.80 ± 2.07	7.75
PCBM	(13.07) 9.27 ± 2.27	(72.15) 58.0 ± 8.60	(1.03) 0.97 ± 0.05	(17.63) 16.02 ± 1.56	11.86

The trend in VOC values (Table ) between different fullerenes is in good agreement with the variations in QFLS values (Figure a–c), supporting the previous statement that within the set of indene-fullerene derivatives, the NHAc-Me-ICMA cells display reduced recombination at the perovskite/ETL interface. This also points toward the possible passivation effect of the amine group embedded in the NHAc-Me-ICMA molecule. To further investigate possible chemical interactions between different functional groups in the used fullerene derivatives and perovskite’s surface, we applied X-ray photoelectron spectroscopy (XPS). The XPS data are shown in Figure a. It is evident that the binding energy of the Pb 4f core level of a bare perovskite film shifts toward higher values upon deposition of a thin layer of one of the fullerene derivatives. It implies the presence of a more negative charge around the Pb2+ ions. Upon the formation of the perovskite/ICMA contact, electron-donating moieties in the fullerene derivatives (amine or amide groups) could form a dative bond with uncoordinated Pb2+ ions.[64] The largest shift has been observed for the NHAc-Me-ICMA, which contains both oxygen and nitrogen donors. The reference PCBM displayed a smaller shift, as the coordinating ability of the oxygen atom in its ester group is weaker. The interaction of Pb2+ with NHAc-ICMA and NH-ICMA adducts was further evaluated by additional DFT calculations (see the Supporting Information, Section 1.5, for the detailed methodology). The interaction energy, obtained by subtracting the energy of the bare Pb2+ ion and the energy of the adduct from the total energy of the interacting system, yielded values of −2.79 and −2.21 eV for the NHAc-ICMA and NH-ICMA adducts, respectively (see Figure b,c). This implies that the amide group can exhibit stronger coordination to Pb2+ than the amine moieties, with a possibility of an enhanced passivation effect.[51,52] The evaluation of the interaction energy between the NHAc-ICMA adduct and the perovskite surface resulted in −0.83 eV, thus suggesting a stabilizing interaction between the two materials, as we schematically show in Figure d. Therefore, since the electronic properties and the LUMO energies are expected to be very similar for the two adducts (see data in Table ), we hypothesize that the higher VOC values of the NHAc-ICMA adduct are related to its ability to bind with the undercoordinated Pb2+ and consequentially passivate surface trap states. A stronger steric clash caused by the additional methyl group is the main difference between the NHAc-ICMA and NHAc-Me-ICMA compounds, which can induce differences in the layer packing and ensuing passivation effects.

Figure 5

(a) XPS measurements for Pb orbitals. The input X-ray source was Al Kα (1486 eV). Models for the interaction of Pb2+ with (b) NHAc-ICMA, (c) NH-ICMA, and (d) possible mechanism of passivation of undercoordinated Pb2+ by NHAc-ICMA. Interaction energy (in eV) and atomic distances (in Å) are reported.

Within the cells incorporating the newly synthesized fullerene derivatives, the NHAc-ICMA sample delivered the highest PCE, reaching up to 13.61%. The PCE enhancement upon NHAc-ICMA incorporation predominantly originated from the increase in JSC and FF, yielding 20.07 mA cm–2 and 68.14%, respectively. The improved JSC for the NHAc-ICMA-based cells could originate from the improved ETL morphology and more conformal capping of the perovskite surface, as evidenced by the cross-sectional SEM images in Figure d. In the case of the reference device employing solution-processed PCBM, we measured a PCE of 13.07% and JSC of 17.63 mA cm–2. Additionally, we observed that the PSCs with the NH-ICMA molecule delivered a relatively low JSC. This effect could be attributed to the sub-optimal morphology of this ETL, displaying large aggregates and non-complete perovskite coverage (see Figure f). We also note lower solubility of the NH-ICMA compound in the o-DCB, which could result in a higher aggregation tendency and, in turn, less-uniform film morphologies.[24] This is also evidenced by AFM images (high surface roughness, as shown in Figure ). A smooth and uniform ETL morphology is needed to provide an intimate contact with the perovskite film, which, in turn, leads to effective charge extraction and high current densities.

We also performed SPO measurements under continuous illumination. The SPO efficiency decreased over time for the devices with the indene fullerenes, as shown in Figure c and Table . Additionally, for these cells, we recorded larger differences between JV-derived and SPO-derived PCEs than for the reference, PCBM-based samples. This could be related to the non-optimal electron extraction efficacy (space charge region forming at the interface at lower electric fields), which can be influenced by lower electron mobilities in the newly synthesized ETMs.[65] More detailed photophysical characterization of the interface between the given ETM and perovskite is needed for thorough understanding of the origin of these effects.

Conclusions

In summary, we have developed a series of new fullerene derivatives and have presented the synthesis of the indene-fullerene adducts, and carried it out by appending different indene derivatives to the fullerene C60 through the Diels–Alder cycloaddition process. This synthetic methodology provides a novel approach to obtain a wide spectrum of indene-fullerene adducts, which are inaccessible by previous, conventional pathways. We also performed computational simulations and CV measurements, the results of which indicate that groups with a stronger electron donation effect in the indene structure are the key motif in the indene-fullerene adduct toward its application as an ETL. Based on these results, we fabricated flexible PSCs incorporating indene-fullerene derivatives with the amine and amide groups as ETLs. We tested the NH-ICMA compound and its derivatives, the NHAc-ICMA and NHAc-Me-ICMA, applied to devices with large active areas of 1 cm2. We have also provided an insight into the characteristics of the perovskite/ETL interface via spectroscopic methods. Notably, the NHAc-ICMA and NHAc-Me-ICMA derivatives resulted in a decrease in non-radiative recombination losses when compared with the NH-ICMA compound. The NHAc-ICMA-based devices showed the best photovoltaic performance, 13.61% of PCE. We believe that further optimization of the device-processing protocol could result in further improvements of the cells performance values.