Julolidine functionalized benzimidazoline‐doped fullerene derivatives for efficient and stable perovskite solar cells

Fullerene derivatives are highly attractive materials in solar cells, organic thermoelectrics, and other devices. However, the intrinsic low electron mobility and electrical conductivity restrict their potential device performance, such as perovskite solar cells (PSCs). Herein, we successfully enhanced the electric properties and morphology of phenyl‐C61‐butyric acid methyl ester (PCBM) by n‐doping it with a benzimidazoline derivative, 9‐(1,3‐dimethyl‐2,3‐dihydro‐1H‐benzoimidazol‐2‐yl)‐julolidine (JLBI‐H) via a solution process. We found the n‐doping can not only improve the conductivity and optimize the band alignment but also enable the PCBM to have a constantly strong charge extraction ability in a wide temperature from 173 to 373 K, which guarantees a stable photovoltaic performance of the corresponding PSCs under a wide range of operating temperatures. With the JLBI‐H‐doped PCBM, we improved the efficiency from 17.9% to 19.8%, along with enhanced stability of the nonencapsulated devices following the aging protocol of ISOS‐D‐1.

Perovskite solar cells (PSCs) have reached power conversion efficiencies (PCEs) >26%, approaching the PCEs of state-of-the-art crystalline-silicon solar cells.Inverted (p-i-n structure) PSCs with a sequential deposition of hole transporting (p), intrinsic (i), and electrontransporting (n) layers have gained remarkable attention in recent years due to their low hysteresis effect, superior long-term stability, and absence of high-temperature treatments during fabrication (below 100°C).
In inverted PSCs, the electron transporting layer (ETL), such as C60 and phenyl-C 61 -butyric acid methyl ester (PCBM), not only accepts and transfers the electron from the perovskite active layer but also protects the perovskite active layer from moisture and oxygen invasion.Emerging evidence has shown that the voltage loss as well as the stability of inverted PSCs is governed by the top ETL. [1]For example, solution-processed PCBM has a high energy disorder and relatively low carrier mobility, which may restrict their potential application in high-efficiency PSCs.In addition, the solution fabrication process of perovskite film inevitably contains massive defects at the upper surface, which is in direct contact with the ETL (e.g., C60 or PCBM).Therefore, further improvements to the performance and stability of PSCs will require dedicated management of the ETL and its contact with the perovskite layer.
Promising strategies to improve the electric property and morphology of PCBM have been reported.For example, Huang et al. developed a series of new fluorinated fullerenes with different lengths of fluorocarbon chains to improve the hydrophobicity of the ETL.The new fullerene derivatives not only have higher electric properties but also prohibit water molecule penetration. [2]However, the synthesis of fullerene derivatives is prone to generating isomers, resulting in purification challenges and substantial costs.
Aiming to enhance the electric properties and morphology of PCBM, julolidine was introduced onto 1H-benzimidazolyl as an n-dopant (named 9-(1,3dimethyl-2,3-dihydro-1H-benzoimidazol-2-yl)-julolidine [JLBI-H]) of PCBM.The three tertiary amines in JLBI-H exhibit strong electron-donating capabilities, while the two fused cycloalkyl edges are employed to improve their miscibility with PCBM. [15]Ultraviolet photoelectron spectroscopy (UPS) and space charge limited current (SCLC) were used to study the n-doping effect on energy alignment and carrier mobility.Temperature-dependent steady-state photoluminescence (PL) and time-resolved PL (TRPL) demonstrate that JLBI-H doping enhances the charge extraction properties of PCBM film in a wide range of temperatures in comparison with the pristine PCBM or C60.Solar cells are fabricated and characterized to disclose the effect of the dopant in promoting charge transfer and inhibiting recombination reactions within full devices.In addition, flexible PSCs have been developed to assess the adaptability of this strategy in other types of devices.

| RESULTS AND DISCUSSIONS
All solar cells investigated in this work had a standard p-i-n architecture (Figure 1A) and were fabricated by sequential deposition of components onto glass covered with indium-doped tin(IV) oxide, which served as a cathode in the final device.Thoroughly optimized procedures adopted from the literature and established in our laboratory were employed.The examined solar cells featured a [2-(3,6-Dimethoxy-9H-carbazol-9-yl) ethyl]phosphonic acid (MeO-2PACz) self-assembled monolayers (SAMs) as a hole transporting layer (HTL), a high performing ca.600 nm thick FA 0.95 MA 0.05 PbI 2.85 Br 0.15 perovskite (FA + = CH(NH 2 ) 2 , MA + = CH 3 NH 3 + ) light absorber, a ca.20 nm doped-PCBM electron transporting layer, and a silver counter electrode.As evidenced in Figure 1B, the incorporation of JLBI-H doping results in a more uniform film thickness was attributed to the improved miscibility of PCBM, leading to a good contact to both perovskite and Ag electrode.The formation of the target perovskite was confirmed by X-ray diffraction (XRD) analysis (Supporting Information S1: Figure S1).XRD patterns exhibited a set of sharp diffractions expected for the cubic FA 0.95 MA 0.05 PbI 2.85 Br 0.15 perovskite phase with a mean crystallite size of 152 ± 16 nm (derived from the full width at half maximum of the (100) peak).
The interface band alignment between perovskite and PCBM has been examined by UPS measurements.The highest occupied molecular orbital (HOMO) of pristine and doped PCBM are −6.40 to −6.16 eV, respectively (Figure 1C, and Supporting Information S1: Figure S2).The lowest unoccupied molecular orbital (LUMO) is roughly assessed by subtracting the optical gap from the HOMO, which is derived from the absorption spectrum.After doping, the LUMO shifts from −4.50 to −4.26 eV (Supporting Information S1: Figure S3).The shift of the LUMO confirms a successful n-doping of PCBM by JLBI-H.The up-shift is favorable for minimizing the energy loss during the transfer of electrons from the perovskite layer (E CBM = −4.23 eV) to the ETL. [10]urthermore, we observed a stronger interface dipole at the perovskite|ETL interface after the doping (Figure 1D,E), which is conducive to electron transport from the perovskite layer to the ETL. [13]According to our previous results, the n-doping effect is caused by a hydride or hydrogen atom transfer from PCBM to JLBI-H, thus generating fulleroid anions.In addition, as an N-heterocyclic aromatic unit, julolidine bears two trimethylenes units at the ortho position sharing a nitrogen atom to form two fused six-membered rings.Therefore, the two fused cycloalkyl edges would endow the lipophilicity and increase its affinity with PCBM without much molecular size expansion. [13]e then studied the doping effect on the morphology of PCBM films.The perovskite film shows clear grain boundaries and uniform grains with an average size of 650 nm (Figure 2A).After covering the PCBM ETL, the grains become blurred due to the full coverage of PCBM (Figure 2B).When the PCBM has been doped, the grains of the perovskite become clear (Figure 2C), which is possibly due to a higher conductivity of the doped PCBM. [16]As shown in Figure 2D, the PL intensity of perovskite film is significantly quenched due to a charge transfer from perovskite to PCBM. [17]The quenching of effect was more pronounced after doping by JLBI-H.When the doping level was 0.25 mol.%, the sample showed the lowest PL, which indicates the best charge extraction ability of this ETL.As shown in Figure 2E and Table 1, the TRPL lifetimes exhibit a similar trend, that is, the lifetime of perovskite film is quenched from 882 to ~100 ns.Again, we observed the shortest lifetime in the sample with a doping level of 0.25 mol.%, further confirming the strongest charge extraction capability. [18]hen, SCLC measurements are used to evaluate the electron mobility of PCBM films with successful doping (Figure 2F and Supporting Information S1: Figure S4).1]  We then fabricated solar cells to investigate the benefits of JLBI-H-doping on photovoltaic performance.As shown in Figure 3A and Table 2, the devices based on pristine PCBM had an average PCE of 17.1 ± 0.5%, an open-circuit voltage (V OC ) of 1.04 ± 0.02 V, a short-circuit current density (J SC ) of 22.0 ± 0.4 mA/cm 2 , and a fill factor (FF) of 75.2 ± 2.7%, which are consistent with previous reports. [22]In contrast, JLBI-H-doped devices had a significantly higher PCE of 19.5 ± 0.3%, with a V OC of 1.05 ± 0.01 V, J SC of 24.2 ± 0.3 mA/cm 2 , and FF of 76.6 ± 1.0% (Supporting Information S1: Figure S5).The J-V curves of the most efficient devices are shown in Figure 3A.The devices with doped PCBM have a smaller hysteresis in comparison with the pristine ones (0.025 vs. 0.089), which can be explained by the enhanced charge extraction properties of the ETL. [23,24]The quasi-steadystate power outputs (qSPO) at the fixed bias near the maximum power point are 16.7% and 18.2% for solar cells based on pristine and doped PCBM, respectively.(Figure 3B).The increase in PCE is mainly due to the significant enhancement of J SC , which can be explained by the enhanced incident photoelectron conversion efficiency (IPCE), as shown in Figure 3C.The integrated J SC values of doped devices are much higher than pristine ones (i.e., 23.6 vs. 20.9mA/cm 2 ), particularly at short wavelength regions (300-400 nm).According to the literature, the shorter-wavelength light penetrates the perovskite film shallowly, which makes carriers accumulate at the surface of the perovskite film. [25]Therefore, the enhanced IPCE in the doped device is primarily due to improved interface contact between PCBM and perovskite, which enables a more effective charge carriers extraction and less charge accumulation at the surface of the perovskite film. [16]The influence of doping level on the solar cell performance is shown in Figure 3D and Supporting Information S1: Figure S5.The device efficiency is highest when the doping concentration is 0.25 mol.%, which is consistent with the PL/TRPL results.Additionally, devices with thicker ETL have been fabricated to magnify the critical contribution of n-doping.As shown in Supporting Information S1: Figure S6, when the ETL thickness increases, the device performance continuously decreases, primarily due to poorer J SC and FF.Compared to the pristine device, the doped device exhibits higher J SC and FF at any thickness, which may be due to the effective n-doping in the doped device.Moreover, we applied this ETL to flexible PSCs.The flexible devices based on JLBI-H doped PCBM show a champion efficiency of 16.7%, which is higher than that of undoped devices at 15.2% (Supporting Information S1: Figure S7).
To further understand the source of PCE variation in PSCs, the series resistance (R S ) and ideality factor (m) have been extracted by fitting the dark J dark -V curve (Figure 3E) with Equations ( 1) and (2): (1) where J 0 , q, k B , and T are the reverse saturation current, elementary charge, Boltzmann constant, and cell temperature, respectively.The R S value of the pristine device is 4.84 Ω, and slightly reduced to 4.06 Ω after doping.The doped device also shows lower m-values (1.81 vs. 1.85), indicating a less nonradiation trap-assisted recombination in comparison with the pristine ones. [26,27]he solar cell performance under different light intensities has been performed to study the charge extraction and recombination within full devices.As shown in Figure 3F, the J SC values have a linear relationship with light intensity (I).Fitting the curve with J SC ∝ I α provides the α values.The increased α value from 0.88 to 0.90 T A B L E 1 Parameters used to fit the biexponential function a to TPRL curves (Figure 2E) of perovskite, perovskite deposited with PCBM without doping, and with different doping levels.

Sample
A | 373 indicates that the space charge effect is suppressed and the charges can be more efficiently collected before recombination. [28]The relationship between V OC and light intensity was further studied to reveal the recombination kinetics in the device (Figure 3G).The slope of doped devices is lower than that of the pristine one, that is, 1.22k B T/q, versus 1.42k B T/q, suggesting a suppressed trapassisted recombination loss in the former. [29]ransient photovoltage (TPV) and photocurrent (TPC) decay measurements were further examined to study the recombination and charge-extraction processes. [30,31]he TPV lifetimes of the doped device are longer than the pristine ones at all light intensities (Figure 3H and Supporting Information S1: Figure S8), indicating less charge recombination in the doped device.Moreover, the TPC lifetimes of the doped devices are all shorter than the pristine ones, indicating a more efficient charge extraction in the former (Figure 3I).If we consider that all solar cells are identical except the ETL, any difference in their recombination process is caused by the PCBM layer. [30,32,33]herefore, we can conclude that the JLBI-H-doping not only suppressed the recombination but also enhanced the charge extraction at the perovskite|PCBM interface, thus improving the solar cell performance. [34]e then evaluate unencapsulated solar cells by aging them under ambient conditions with a relative humidity of 35 ± 5% and temperature of 27 ± 3°C in the dark (ISO-D-1). [35,36]The doped devices maintained their initial efficiencies of 91% after 1000 h, while it is only 67% for the pristine ones (Supporting Information S1: Figure S9).The stability improvement is possibly due to the increased hydrophobicity of the ETL as demonstrated by the increased water contact angle after doping (Supporting Information S1: Figure S10). [37] addition to the applications at terrestrial, PSCs have recently found potential applications in other extreme conditions, such as in near-space and polar regions. [38,39][42] As shown in Figure 4, when the temperature decreased, the efficiency of all PSCs exhibited T A B L E 2 Photovoltaic parameters of PSCs based on PCBM and J-PCBM ETLs.

Sample
V OC (V) J SC (mA/cm Note: All measurements were under AM 1.5 G 1-sun irradiation with an aperture of 0.16 cm 2 . Abbreviations: FB, forward-bias; FF, fill factor; PCE, power conversion efficiency; qSPO, quasi-steady-state power output; SC, short-circuit.a Average V OC , J SC , FF, and PCE were derived from the J-V curves (scan rate 100 mV/s) recorded for 18 cells of each type in the FB to SC direction, and for the best-performing devices recorded in both two directions.b qSPO values were derived from the final point of 120 s measurements at fixed voltages corresponding to the maximum power points in the J-V data for the best-performing cells.
an initial rise followed by a decline.Notably, the JLBI-H doped device achieved its peak efficiency at 193 K, primarily attributed to the elevation of both J SC and FF.Moreover, the JLBI-H-doped device consistently surpasses both the pristine and C60-based devices across all temperatures.This observation underscores the enhanced application potential of the doped device, particularly in challenging or extreme conditions.As shown in Supporting Information S1: Figure S11, J-V curves under dark conditions were tested at various temperatures to extract the series resistance (R S ).Although the R S showed little variation with temperature, the R S of the doped device is lower than that of the pristine device at all ranges of temperature, which may also be the reason for the better performance of the doped device. [43]Additionally, we tested the temperature-dependent conductivity of PCBM films.The conductivity of both the pristine and doped PCBM films exhibits a consistent increase with decreasing temperature (Supporting Information S1: Figure S12).Therefore, the change in conductivity should not be the primary factor for the difference in device performance at different temperatures.
To delve deeper into the mechanism behind the variation of FF under lower temperatures, we further explored the carrier dynamics at the interfaces through temperature-dependent PL and TRPL measurements.As shown in Figure 5A-C and Supporting Information S1: Figure S13a, when the temperature decreases from 373 to 173 K, all samples exhibit a redshift in peak positions, a narrowing of full-width at half maximums (FWHMs), and an increase in PL intensity.The increase in perovskite PL strength may be because the photocarriers have less thermal energy and are therefore located closer to the edge of the band at lower temperatures. [44]The increase in strength of the perovskite|PCBM sample may be due to poor charge extraction. [45]By contrast, the perovskite|J-PCBM increases slowly and is even more robust than that of perovskite|C60 samples (Supporting Information S1: Figure S13b).Figure 5D shows PL quenching efficiency (η = (PL perovskite − PL with ETL )/ PL perovskite , PL with ETL and PL perovskite referring to the PL intensities of perovskite with and without ETL, respectively).The quenching efficiency of PCBM drops sharply when the temperature is above 273 K, that is, only 6% at 373 K, which indicates a poor charge extraction capability of PCBM at elevated temperatures.In contrast, the quenching efficiency of doped-PCBM is above 80% in the whole range from 373 to 173 K.When temperatures are above 333 K, it exhibits an even better quenching effect than C60 (Figure 5D).These findings demonstrate the exceptional charge extraction capability of JLBI-H-doped PCBM.
Color plots of normalized PL spectra for (A) perovskite|PCBM, (B) perovskite|J-PCBM, and (C) perovskite|C60 at different temperatures from 173 to 373 K; the black scatter plot shows the corresponding PL peak positions for each temperature.The PL emission peak value is normalized to 1 at each temperature.(D) The temperature-dependent PL intensity quenching efficiency of perovskite films covered with PCBM, C60, and J-PCBM.The temperature-dependent (E) fast decay, and (F) slow decay lifetimes of perovskite films covered with PCBM, C60, and J-PCBM.J-PCBM, JLBI-H-doped PCBM film; P, perovskite; PCBM, phenyl-C 61 -butyric acid methyl; PL, photoluminescence.
In temperature-dependent TRPL measurements, all the attenuation curves conform to the law of double exponential function (Supporting Information S1: Figures S13 and S14, the extracted lifetimes are shown in Supporting Information S1: Tables S1-S4). [46]For bare perovskite, from 373 to 253 K, both lifetimes of τ 1 and τ 2 increase along with the decreased temperature, indicating reduced trap-assisted nonradiative recombination processes both at the surface and in the bulk.Further lowering the temperature from 253 to 173 K leads to a decline in the lifetimes, which is likely due to the transition of perovskites from tetragonal to orthorhombic phases. [44]For the perovskite|PCBM sample, the lifetimes τ 1 and τ 2 basically follow a similar trend with the bare perovskite.In contrast, the lifetimes of perovskite|doped-PCBM and perovskite|C60 remain constant short in the whole temperature range.This result demonstrates that JLBI-H-doping can substantially enhance the electron extraction capacity of PCBM, particularly at extreme temperatures.Hence, the heightened efficiency of doped devices at lower temperatures can be ascribed to their robust charge extraction capability.
Based on the above results, we propose a scheme to explain this phenomenon.In a high energy disordered PCBM film, the charge transport occurs via hopping between localized sites in PCBM.Electrons participate in charge transport only when they are thermally activated above the transport energy level (E T ). [43]At lower temperatures, the thermal energy decreases with decreasing temperature, and less number of electrons could be activated to the E T , resulting in their accumulation at the interfaces rather than transport to the electrode.Thus, we observed an increased PL intensity, reduced quenching efficiency, elongated lifetimes, and thus reduced FF and efficiencies.Doping PCBM with JLBI-H introduces two fused cycloalkyl edges, fostering a favorable affinity with alkylated PCBM and promoting enhanced miscibility without much expansion in the molecular size.This reduces the barrier between electrons and E T , sustaining a consistently high charge extraction capability for PCBM across a broad temperature range.

| CONCLUSION
We conclude by emphasizing that n-type dopants with dedicatedly designed molecular structures for doping PCBM can improve the electron mobility and conductivity of fullerene materials.As a promising example, these benefits signify that JLBI-H doping not only enhances the miscibility of JLBI-H with PCBM and suppresses recombination reactions at the interface but also enhances charge extraction within the charge transport layer.Moreover, this strategy enables the PCBM to have a constant strong charge extraction ability in a wide temperature from 173 to 373 K, which indicates a promising application of this ETL in other similar conditions, such as near-space and polar regions.

Hubei
Provincial Natural Science Foundation of China, Grant/Award Number: 2022CFB1000; Knowledge Innovation Program of Wuhan-Shuguang Project, Grant/Award Number: 2023010201020367; State Key Laboratory of Advanced Technology for Materials Synthesis and Processing (Wuhan University of Technology), Grant/Award Number: 2022-KF-17; Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT, and Future Planning, Grant/Award Numbers: 2019K1A3A1A61091347, 2021M3H4A1A02051234 1 | INTRODUCTION
curves of the champion devices (arrows show the direction of the scan).(B) The qSPO at a fixed bias around maximum power point.(C) IPCE spectra and corresponding integrated current densities.(D) Statistics of the PCE distribution (18 independent cells for each type).(E) J-V curves under dark conditions; inset shows the the dV/dJ versus J −1 plots.(F) J SC versus light intensity.(G) V OC versus light intensity.(H) The TPV lifetimes versus the V OC at different light intensities, and (I) TPC lifetimes versus the light intensity of PSCs based on PCBM and J-PCBM ETLs.IPCE, incident photoelectron conversion efficiency; J-PCBM, JLBI-H-doped PCBM film; PCBM, phenyl-C 61 -butyric acid methyl; PSC, perovskite solar cell; qSPO, quasi-steady-state power output; TPC, transient photocurrent; TPV, transient photovoltage.
b P