Efficient promotion of charge separation and suppression of charge recombination by blending PCBM and its dimer as electron transport layer in inverted perovskite solar cells

Abstract

Organic–inorganic hybrid perovskite solar cells have achieved great success in recent years. Meanwhile, inverted structured device is an important branch in perovskite photovoltaics owing to its peculiar advantages of low-temperature fabrication process and non-hysteresis behavior. In this kind of device, the electron transport material, as well as the interface between perovskite and electron transport layer (ETL), plays a crucial role in photoelectric conversion process. We report that the utilization of the fullerene derivative blend, PCBM and its dumb-belled dimer, as electron transport material in inverted perovskite solar cells, which could significantly enhance the photovoltaic performances. The morphology of ETL can be regulated by changing the admixing ratio of PCBM and its dimer. Moreover, the steady-state/time-resolved fluorescence and transient photovoltage decay measurements indicate that the optimization of perovskite/ETL interface through an appropriate fullerene blend is beneficial to promote charge separation and suppress charge recombination.

---

Introduction

The prominent properties of organic–inorganic halide perovskite, such as excellent light-harvesting ability¹ and long charge diffusion length,^2,3 have aroused a great research enthusiasm of scientists in the photovoltaic field. Since the first attempt of using perovskite as a visible-light sensitizer by Miyasaka et al. in 2009,⁴ the power conversion efficiency (PCE) record of perovskite solar cell (PSC) has been continuously refreshed with the highest value of more than 20%.^5,6 Generally, PSC consists of a perovskite active layer sandwiched by the electron and hole transport layers (ETL and HTL) contacting corresponding electrodes. In the configuration of the conventional mesoporous- and planar-structured perovskite devices,^7,8 ETL is directly prepared on the transparent conductive electrode, while in the so-called inverted planar PSC, it is the HTL fabricated on the transparent electrode.⁹ Though the conventional PSCs exhibit superior photovoltaic performances, the requirement of high temperature for sintering TiO₂ and the obvious hysteresis phenomenon might hinder their future applications.¹⁰ In contrast, owing to the advantages of low-temperature preparation and negligible hysteresis, the inverted perovskite device has drawn an increasing attention recently.^11,12

In the inverted PSC, the traditional transport materials in organic photovoltaic devices, poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) and fullerene derivatives, are respectively utilized as the HTL and ETL.¹¹ After the first inverted PSC with a PCE of 3.9% was reported by Guo et al. in 2013,¹³ amounts of efforts have been made on this intriguing device and the PCE is elevated to more than 18%.^11,14 Expect for the optimization of core perovskite film,^15–17 numerous studies are focused on the electron transport materials and relevant interface engineering. Although some new n-type materials (e.g. ZnO¹⁸) were developed as the ETL in inverted PSC, fullerene derivatives are still the most used and effective electron transport materials, in which [6,6]-phenyl-C₆₁-butyric acid methyl ester (PCBM) is the most representative one.^19,20 To improve the properties of fullerene derivatives and relevant interfaces, buffer layers are often inserted between the fullerene ETLs and the electrodes.^21–24 In fact, fullerene derivatives not only act as electron transport materials, but also are responsible for optimization of their interfaces with perovskite and the electrode. Huang et al. found that thermal annealing of the PCBM layer can effectively passivate the trap states, reduce the interfacial charge recombination rate and further facilitate the carrier transport.²⁵

Therefore, the optimization of fullerene electron transport materials and the exploration of the involved photoelectric conversion process in perovskite film and device systems are essential for further improvement of the perovskite photovoltaic performances. For example, there are reports that PCBM doped with other non-fullerene materials (e.g. graphdiyne²⁶ and oleamide²⁷) could improve the morphology and properties of PCBM film, resulting in an increased electrical conductivity of ETL or a lowered charge carrier recombination rate. Such doping or blending engineering has been used in the polymer photovoltaic devices to optimize the active layer for a better photovoltaic performance.^28–31 Moreover, in inverted PSCs, the blends of polymer (e.g. polystyrene) and fullerene have also been utilized to improve the properties of the ETLs.^32,33

Dumb-belled PCBM dimer (d-PCBM) is a kind of PCBM derivative containing two fullerene spheres in each molecule, as shown in Fig. 1a. The d-PCBM with improved aggregation tendency shows better phase separation against the donor material than PCBM.^34,35 Further research indicates that the replacement of PCBM with d-PCBM in active layer could significantly increase exciton dissociation and reduce non-geminate charge recombination.³⁶ Herein, we report the introduction of d-PCBM as electron transport material into the perovskite photovoltaics. More prominently, d-PCBM and PCBM are mixed at different ratio to modulate the aggregation and film-forming ability of the fullerene blend material. It's the first attempt of directly blending two kinds of fullerene derivative as electron transport material. The morphologies of these different PCBM:d-PCBM blend films are characterized and the fluorescence kinetics of related perovskite/fullerene bilayers are investigated. The inverted PSCs with the configuration of ITO/PEDOT:PSS/CH₃NH₃PbI_3−xCl_x/PCBM:d-PCBM/Ca/Ag, as displayed in Fig. 1b, are fabricated and the transient dynamic measurements are conducted. The results indicate that charge separation and charge recombination could be optimized by mixing two fullerene materials at proper ratio, which further enhances the photovoltaic performance of the devices.

Fig. 1 (a) The chemical structures of d-PCBM and PCBM. (b) Device architecture of the inverted PSC based on fullerene blend as ETL. Statistical RMS roughness data of PCBM:d-PCBM films with various ratios on (c) quartz substrates and (d) perovskite substrates. The models of two films on different substrates are respectively shown in blanks of (c) and (d). AFM images of PCBM film and d-PCBM film on quartz substrate are also shown in (c). Black dot in (d) represents RMS roughness data of perovskite film.

Experimental

Materials and preparation

Perovskite precursor solution (0.75 M) was prepared by pre-mixing methylammonium iodide (MAI, DYESOL, Australia), lead iodide (PbI₂, 99.999%, Alfa) and lead chloride (PbCl₂, 99.999%, Alfa) in anhydrous N,N-dimethylformamide (DMF, 99.8%, ACROS ORGSNICS) solvent with a molar ratio of 4 : 1 : 1, then the solution was stirred overnight at 60 °C. According to our previous work,³⁴ d-PCBM is synthesized by reacting [6,6]-phenyl-C₆₁-butyric acid (PCBA) with neopentyl alcohol, 2,2-dimethyl-1,3-propanediol and 2-(hydroxymethyl)-2-methylpropane-1,3-diol, and the preparation route of PCBA is shown in Scheme S1.† The fullerene blend solution was gained by merging PCBM solution (30 mg ml⁻¹, in chlorobenzene) and d-PCBM solution (30 mg ml⁻¹, in chlorobenzene) with various volume ratios (6 : 1, 4 : 1 and 1 : 1). These solutions were filtered through PTFE filters (0.22 μm) and pre-heated at 70 °C before operation.

Fabrication of films and devices

The indium tin oxide glass (ITO, STN-SI-10 OHM, Nanbo, ∼10 Ω sq⁻¹) was pre-patterned by etching with Zn power and diluted HCl solution before cleaned in an ultrasonic bath for 15 min, in which detergent, deionized water, acetone and ethanol were sequentially used. Then the substrates were subjected to ultraviolet plasma for 20 min to clean off organic residuals. The PEDOT:PSS solution (Clevios P VP AI 4083, Heraeus) was spin-coated at 4000 rpm for 35 s on the substrates and annealed at 145 °C in air for 20 min before moved into an inert glove box. The perovskite precursor solution was then spin-coated at 3000 rpm for 40 s, and then gradually heated to 100 °C for 30 min. After cooling down, the fullerene solutions were then deposited by a consecutive two-step spin-coating process at 1000 and 2000 rpm for 6 and 40 s, respectively. Finally, 8 nm Ca and 100 nm Ag were thermally evaporated, successively (the active layer was fixed to 0.1 cm² for current density–voltage (J–V) measurements). The fullerene blend films for morphology characterization were prepared on quartz or perovskite substrates in the same manipulation process as device fabrication.

Characterization

The roughness of the films was evaluated by atomic force microscopy (AFM) (Dimension Icon, Bruker), and the statistical root-mean-square (RMS) roughness comes from ten sets of data. The morphology images were captured by field-emission scan electron microscope (SEM, JEOL-7401F). X-ray diffraction (XRD) measurement was performed on Shimadzu XRD-7000 using Cu Kα radiation in the 2θ range from 8° to 60° with a scan rate of 4° per min. The UV-vis absorption spectra were acquired by a Shimadzu UV-3600 spectrometer. The steady-state photoluminescence spectroscopy was conducted on Edinburgh FLS980 (excited at 510 nm from the side of spin-coated materials). The time-resolved photoluminescence was measured by streak camera (C5680, Hamamatsu Photonics) and the film samples are excited at 510 nm from the side of spin-coated materials. All the tested samples for optical measurements were prepared on quartz substrates directly following the same craft mentioned above. The J–V characteristics were obtained by exposing tested devices to AM 1.5G sun illumination (100 mW cm⁻²) generated by a sun light simulator (SS150, Zolix Instruments Co., Ltd.). All data were acquired by Keithley 2400 sourcemeter. Incident photon-current conversion efficiency (IPCE) was measured with the photoelectric conversion test system (SCS100-X150-DSSC, Zolix Instruments Co., Ltd.). In transient photovoltage (TPV) measurement, a continuous-wave LED (530 ± 5 nm) irradiated the glass side to provide a steady-state voltage, and weak laser pulses (532 nm, 7 ns) were superimposed and correlated to the background illumination to ensure a small perturbation voltage amplitude (smaller than 5% of the steady-state voltage). Electrical signals were collected by a digital oscilloscope (600 MHz, Lecroy, input impendence 1 MΩ). Corresponding photocurrent measurement was carried out by immediately switching the input impendence to 50 Ω after TPV measurement. The obtained photovoltage and photocurrent decay traces were fitted by exponential functions, and the charge recombination and transport lifetimes were weighted average values.

Results and discussion

The d-PCBM was synthesized according to the route as depicted in Scheme S1.† Expect for the alkyl chain used to covalently link the PCBM, no additional special modification is adopted in d-PCBM molecule. Thus, it is understandable that d-PCBM shows similar LUMO energy level and absorption ability as PCBM.³⁴ To investigate the aggregation properties of PCBM, d-PCBM and their blends, various fullerene blend films with different mixing ratios were prepared on quartz substrates and their roughness were evaluated by AFM. As shown in Fig. 1c and S1,† the pure PCBM film shows relatively smooth film surface with the RMS roughness of only 1.60 nm, while pure d-PCBM film on quartz substrate displays a high RMS roughness of 97.8 nm with many micron-level holes and larger aggregation. These can be attributed to the excellent film-forming ability of PCBM and the enhancement of self-aggregation ability of covalently bonding of two fullerene balls for d-PCBM. Besides, it is clearly seen that the RMS roughness of fullerene films increases monotonously as the proportion of PCBM and d-PCBM varies from 6 : 1 to 4 : 1 and 1 : 1, which indicates that the aggregation extent of fullerene film can be controlled by the addition of d-PCBM.

The different fullerene films were also coated on perovskite films and the corresponding AFM images are demonstrated in Fig. 2, which present the variation of statistical RMS roughness (Fig. 1d). Obviously, although the fullerene films on perovskite layers exhibit rougher surfaces than the fullerenes on quartz substrates, the deposition of fullerene layer significantly decreases the RMS value of perovskite film, except d-PCBM. The decrease of the roughness would be beneficial to the contact with metal electrode in devices.^26,37 As shown in Fig. 2a, the perovskite film exhibits the interconnected crystalline domains with a RMS roughness of 34.3 nm. The XRD pattern of the perovskite film shows two main diffraction peaks located around 14.1° and 28.4° corresponding to (110) and (220) planes of perovskite crystal,^8,38,39 respectively (Fig. S2†). The deposition of PCBM film on perovskite layer shows a rather smooth surface with a decreased RMS roughness of 10.3 nm owing to the excellent dispersity of PCBM (Fig. 2b). With the introduction of d-PCBM into PCBM, the fullerene blend films present increasing RMS values from 11.0 nm to 12.1 nm and 17.4 nm for the PCBM : d-PCBM ratio of 6 : 1, 4 : 1 and 1 : 1, respectively (Fig. 2c–e). Particularly, pin-holes become obvious when the ratio is 1 : 1, indicating the high self-aggregation trend at this point. As expected, the pure d-PCBM layer on perovskite possesses the maximum RMS value of 43.0 nm (Fig. 2f), with the morphology of large island-like aggregation and deep cracks. The top-view morphologies of the perovskite/fullerene films are also visualized by SEM, as shown in Fig. S3.† It could be seen that PCBM and three kinds of PCBM:d-PCBM blend films are uniformly deposited on the perovskite with the similar morphology, while the d-PCBM layer surface presents obvious aggregation with high roughness. These results reveal the control ability of d-PCBM to the aggregation degree and morphology of the fullerene films on perovskite, which would be extremely crucial for the performances of the related perovskite photovoltaics.

Fig. 2 The AFM images (size: 30 × 30 μm) of (a) perovskite film, (b) perovskite/PCBM film, (c) perovskite/PCBM:d-PCBM (6 : 1) film, (d) perovskite/PCBM:d-PCBM (4 : 1) film, (e) perovskite/PCBM:d-PCBM (1 : 1) film and (f) perovskite/d-PCBM film.

The steady-state and the time-resolved fluorescence spectra for perovskite film and five kinds of fullerene films coated perovskite layers are displayed in Fig. 3a and b, respectively. The corresponding fluorescence quenching efficiencies (η_q) calculated from the time-resolved fluorescence spectra are duplicated in Fig. 3c. As seen from the insets of Fig. 3a and b, the intrinsic perovskite film exhibits a strong emission peak at 776 nm. After the fullerene materials are coated, the photoluminescence is dramatically quenched owing to the charge transfer and separation at the perovskite/fullerene interface.²⁶ Besides, the integrated fluorescence spectra within 20 ns (Fig. 3b) show a similar tendency of fluorescence quenching effect with steady-state measurements. As shown in Fig. 3c, when d-PCBM is doped into PCBM with the PCBM : d-PCBM ratio of 6 : 1 and 4 : 1, the η_q values (0.962 and 0.957) are both improved compared with 0.934 of PCBM film. When the blend ratio is 1 : 1, the value of η_q decreases to 0.933, similar to that of PCBM film. Obviously, the quenching efficiency is closely related to the morphology of the films, as well as the perovskite/fullerene interface. The incomplete coverage of perovskite film leads to inefficient charge separation at the perovskite/d-PCBM interface. For the films coated with PCBM:d-PCBM blends with superior morphologies, the quenching effects are remarkably improved. It is reasonable to speculate that introducing a moderate amount of highly self-aggregated d-PCBM could optimize the perovskite/fullerene interface and facilitate the charge transfer.

Fig. 3 (a) The steady-state and (b) time-resolved fluorescence spectra of PCBM film, PCBM : d-PCBM (6 : 1) film, PCBM : d-PCBM (4 : 1) film, PCBM : d-PCBM (1 : 1) film and d-PCBM film coated perovskite layers. The two insets are unabridged figures including spectrum of the intrinsic perovskite film. (c) η_q of five different fullerene films coated on perovskite layers. (d) Fluorescence decay kinetics from the streak camera measurement for six films as mentioned above. The inset of (d) displays the η_CS of five different fullerene films coated perovskite layers.

In addition, it is worth noting that a red shift of emission peaks is observed when fullerene materials are coated on perovskite film (Fig. 3b). From the peak positions listed in Table S1,† it is seen that the red shift is most obvious for PCBM : d-PCBM (6 : 1) film and PCBM : d-PCBM (4 : 1) film with the peak position of 795 nm and 794 nm, respectively. It is well known that peak of about 780 nm (776 nm in our work) corresponds to the band-to-band radiative recombination, while a weak trap state luminescence is also found at the red-edge of the emission,²⁵ shown as the asymmetric Gaussian peak. The band-to-band emission can be quenched more easily than the trap states emission owing to the distinct charge mobility. So it is reasonable to infer that PCBM : d-PCBM (6 : 1) and PCBM : d-PCBM (4 : 1) film can quench the band-to-band radiative recombination more efficiently and completely, leading to the prominence of the weak trap states luminescence.

The conclusion is also supported by the dynamic results in Fig. 3d (acquired from integration of raw data of streak camera in Fig. S4†). The fluorescence dynamic curves^40–42 are fitted by exponential function (Fig. S5†) and the fluorescence decay lifetimes (τ_fl) are shown in Table S2.† The decay lifetimes of 0.737 ns and 0.885 ns for perovskite/PCBM:d-PCBM (4 : 1) film and perovskite/PCBM:d-PCBM (6 : 1) film verify the efficient and rapid fluorescence quenching. Finally, apparent charge separation efficiency (η_CS) is estimated for the various perovskite/fullerene bilayers. Herein, η_CS = (k_p − k₀)/k_p, and k₀ is equal to the reciprocal of τ_fl of intrinsic perovskite film, k_p is equal to the reciprocal of τ_fl of every kind of perovskite/fullerene film. The related parameters are collected in Table S3.† As displayed in the inset of Fig. 3d, perovskite/PCBM:d-PCBM (4 : 1) film presents the most outstanding η_CS of 0.784, with the increase of 14.6% compared with 0.684 of perovskite/PCBM film, indicating the efficient photo-generated carrier separation at optimized perovskite/fullerene interface. What's more, as seen from original integral kinetic curves of Fig. 3d, at the initial moment, all the peak values of the perovskite/fullerene films have a more significant decline than the intrinsic perovskite film. This phenomenon suggests that there exists a faster attenuation phase beyond the nanosecond time scale of streak camera. Considering the low absorption (less than 0.1) for fullerene thin films (Fig. S6a†), the lowest peak values for perovskite/PCBM:d-PCBM (6 : 1) and perovskite/PCBM:d-PCBM (4 : 1) films also manifest a more efficient fluorescence quenching at perovskite/fullerene interface.

Complete PSCs based on different fullerene ETLs were prepared to investigate the photovoltaic performances. The statistical averages of the photovoltaic parameters, including short-circuit current density (J_SC), open-circuit voltage (V_OC), fill factor (FF) and PCE, are shown in Fig. 4a and Table S4.† Example J–V curves and related parameters are shown in Fig. S7a and Table S5,† respectively. It is noteworthy that the devices fabricated are free of obvious hysteresis phenomenon (Fig. S7b†). As expected, the PSC based on PCBM : d-PCBM (4 : 1) obtains the highest average values in J_SC (16.70 mA cm⁻²), V_OC (0.94 V), FF (0.73) and PCE (11.43%). The performance of the PCBM : d-PCBM (6 : 1) based device is worse than the PCBM : d-PCBM (4 : 1) based one. For the PCBM : d-PCBM (1 : 1) based PSC, J_SC and V_OC are obviously better than pure PCBM based PSC, while the final PCE is actually worse. It should be due to the worst FF (0.63) of the device, which is closely related to the worst charge collection to be discussed below. In addition, device based on pure d-PCBM as ETL was also constructed, but the device performance is relatively poor and the J–V characteristic curve is abnormal (Fig. S8†), which could be attributed to the poor coverage of d-PCBM on perovskite layer and the direct contact of perovskite and metal electrode.

Fig. 4 (a) Statistical J_SC, V_OC, FF and PCE data of inverted perovskite devices based on PCBM, PCBM : d-PCBM (6 : 1), PCBM : d-PCBM (4 : 1) and PCBM : d-PCBM (1 : 1) as ETLs from eight devices for every type. (b) IPCE spectra and integrated photocurrent densities for the PCBM, PCBM : d-PCBM (6 : 1), PCBM : d-PCBM (4 : 1) and PCBM : d-PCBM (1 : 1) based PSC with PCEs most approaching average values.

The corresponding IPCE spectra accompanied with integrated current densities are illustrated in Fig. 4b. For all the devices, the IPCE curve shows a broad range covering the whole visible region from 300 nm to 800 nm with an onset around 780 nm ascribed to the typical absorption edge of perovskite.^11,14 The highest IPCE value correlated to device based on PCBM : d-PCBM (4 : 1) ETL approaches 0.80, which gives rise to the superior photovoltaic performance. Correspondingly, the integrated current density decreases from 18.47 to 17.90, 16.95 and 16.59 mA cm⁻² for the PSCs with PCBM : d-PCBM (4 : 1), PCBM : d-PCBM (1 : 1), PCBM : d-PCBM (6 : 1) and PCBM, respectively. These results have the same trend as the J_SC values (Fig. 4a) obtained from the J–V curves.

To further unravel the underlying carrier behaviors in complete photovoltaic devices with various fullerene blends, TPV decay measurement, a broadly used approach to study the charge recombination dynamics of solar cells,^43–45 was performed. An example of original transient voltage dynamics is depicted in Fig. S9.† Charge recombination lifetimes (τ_r) are extracted by exponential fitting of the decay traces (fitting example is displayed in Fig. S10a†), and the τ_r–V_ph dependencies are exhibited in Fig. 5. From a rough view, all of the τ_r evolutions divide into two distinct regions with a demarcation point around 550 mV. In low voltage region, all τ_r keep almost constant in about several milliseconds, irrespective of the variation of fullerene ETLs. However, when the voltage increases beyond 550 mV, τ_r gradually decrease to microsecond magnitude for all tested samples, indicating a faster charge recombination at high voltage region (high quasi-Fermi level).^41,42,46 More importantly, τ_r starts to be discernible in these devices made of various fullerene ETLs. The two distinct regions suggest that there exist two competitive processes. In low voltage region, photo-generated electrons predominately populate in deep trap states, and hopping is the only pathway for a trapped electron/hole to meet or capture another hole/electron.^47–49 As the perovskite active layers of each device are fabricated under the same condition, and the effect of fullerene ETL on charge separation and transfer is negligible in such a deep energy states, it can be conjectured that the hopping properties among these devices are similar. So the almost constant and similar recombination lifetimes are observed in low voltage region. As the voltage increases, the Fermi level gradually rises to approach the conductive band. On this condition, it is much easier for a trapped electron to accomplish interfacial charge separation through conductive band de-trapping. At this time, the impacts of various fullerene ETLs on charge separation and transport start to stand out. Device based on PCBM : d-PCBM (4 : 1) presents a relatively longer recombination lifetime, which could be ascribed to the more efficient and rapid charge separation than other samples, as demonstrated by the fluorescence measurements (Fig. 3). The result indicates that PCBM : d-PCBM (4 : 1) ETL provides a more efficient perovskite/ETL interface, which significantly promotes the charge separation and suppresses the charge recombination.

Fig. 5 The τ_r–V_ph dependencies from TPV measurements of the PCBM, PCBM : d-PCBM (6 : 1), PCBM : d-PCBM (4 : 1) and PCBM : d-PCBM (1 : 1) based PSCs at the semi-logarithmic scale.

Photocurrent decay measurement is conducted by switching the test circuit into short circuit after TPV measurement on the same illumination condition. The corresponding dynamic curves are fitted by multi-exponential function (Fig. S10b†) to get the charge transport lifetime (τ_t) at different voltages. The dependency of τ_t on V_ph from 550 mV to 850 mV is shown in Fig. 6a, which nearly exhibits a constant tendency, similar to the previous reports.^50,51 The τ_t value of PCBM : d-PCBM (1 : 1) based PSC is about 0.45 μs, while the others have similar τ_t values of about 0.35 μs. This means that the PSC with PCBM : d-PCBM (1 : 1) has the worst charge transport ability, and the rest devices exhibit the comparable charge transport behaviors. Furthermore, the charge collection efficiency (η_CC) is estimated from the formula, η_CC = τ_r/(τ_r + τ_t), as depicted in Fig. 6b. Apparently, all the devices possess the constant values of η_CC close to 100% within a V_ph range from 550 mV, and then the η_CC values drop at different V_ph. The η_CC of PCBM : d-PCBM (1 : 1) based PSC begins to decrease at about 660 mV and reaches 0.85 at 850 mV. The η_CC of PCBM based PSC has an obvious decrease at about 750 mV with 0.87 left at 850 mV. The PSCs with the PCBM : d-PCBM ratio of 4 : 1 and 6 : 1 have a better performance with the drop-point of about 800 mV. Especially, the PSC with PCBM : d-PCBM (4 : 1) still holds the η_CC value of 0.98 at 850 mV, which is 12.6% higher than that of the PSC with PCBM. Such excellent charge collection ability could give the credit to the effective charge recombination at the perovskite/ETL interface. Therefore, both of Fig. 5 and 6 quantitatively indicate that the PCBM : d-PCBM (4 : 1) blend ETL behaves best in suppressing charge recombination and elevating charge collection at high voltage, which is crucial for the improvement of the device photovoltaic performance.

Fig. 6 (a) The τ_t–V_ph dependencies from the transient photocurrent measurement at high voltage (>550 mV) for the PCBM, PCBM : d-PCBM (6 : 1), PCBM : d-PCBM (4 : 1) and PCBM : d-PCBM (1 : 1) based PSCs. (b) The η_CC–V_ph dependencies by combining τ_r and τ_t from transient photoelectric measurement.

Conclusion

In conclusion, a novel fullerene derivative, d-PCBM, is used as electron transport material by admixing with PCBM. The fullerene blend with an appropriate mixing ratio could regulate the ETL morphology owing to the self-aggregation ability of d-PCBM, and thus further optimize the perovskite/ETL interface. More importantly, the modified perovskite/ETL interface could effectively promote charge separation and suppress charge recombination revealed by combination of steady-state/time-resolved photoluminescence and TPV measurements. The blend with the PCBM : d-PCBM ratio of 4 : 1 exhibits the best performances in the aspects of photo-electron conversion and charge transport, endowing the corresponding PSC device with the best photovoltaic performance, with 14.6% and 12.6% higher charge separation and collection efficiencies than pure PCBM based device. Our work is expected to provide a new route for the preparation of high-quality electron transport material for high-efficiency perovskite devices. Moreover, the basic dynamic explorations are expected to provide an in-depth understanding of the dynamic mechanism in inverted PSC.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 21133001, 21173266 and 21473250) and the Fundamental Research Funds for the Central Universities, the Research Funds of Renmin University of China (Grant No. 15XNLQ04 and 16XNLQ04).

References

1. C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, Inorg. Chem., 2013, 52, 9019–9038.
2. G. C. Xing, N. Mathews, S. Y. Sun, S. S. Lim, Y. M. Lam, M. Gratzel, S. Mhaisalkar and T. C. Sum, Science, 2013, 342, 344–347.
3. S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza and H. J. Snaith, Science, 2013, 342, 341–344.
4. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050–6051.
5. H. Zhou, Q. Chen, G. Li, S. Luo, T. B. Song, H. S. Duan, Z. Hong, J. You, Y. Liu and Y. Yang, Science, 2014, 345, 542–546.
6. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234–1237.
7. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643–647.
8. M. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395–398.
9. P. Docampo, J. M. Ball, M. Darwich, G. E. Eperon and H. J. Snaith, Nat. Commun., 2013, 4, 2761.
10. H. S. Kim and N. G. Park, J. Phys. Chem. Lett., 2014, 5, 2927–2934.
11. J. H. Heo, H. J. Han, D. Kim, T. K. Ahn and S. H. Im, Energy Environ. Sci., 2015, 8, 1602–1608.
12. J. You, Z. Hong, Y. M. Yang, Q. Chen, M. Cai, T. B. Song, C. C. Chen, S. Lu, Y. Liu, H. Zhou and Y. Yang, ACS Nano, 2014, 8, 1674–1680.
13. J. Y. Jeng, Y. F. Chiang, M. H. Lee, S. R. Peng, T. F. Guo, P. Chen and T. C. Wen, Adv. Mater., 2013, 25, 3727–3732.
14. C. Bi, Q. Wang, Y. Shao, Y. Yuan, Z. Xiao and J. Huang, Nat. Commun., 2015, 6, 7747.
15. P. W. Liang, C. Y. Liao, C. C. Chueh, F. Zuo, S. T. Williams, X. K. Xin, J. Lin and A. K. Jen, Adv. Mater., 2014, 26, 3748–3754.
16. W. Y. Nie, H. H. Tsai, R. Asadpour, J. C. Blancon, A. J. Neukirch, G. Gupta, J. J. Crochet, M. Chhowalla, S. Tretiak, M. A. Alam, H. L. Wang and A. D. Mohite, Science, 2015, 347, 522–525.
17. J. You, Y. Yang, Z. Hong, T.-B. Song, L. Meng, Y. Liu, C. Jiang, H. Zhou, W.-H. Chang, G. Li and Y. Yang, Appl. Phys. Lett., 2014, 105, 183902.
18. J. You, L. Meng, T. B. Song, T. F. Guo, Y. M. Yang, W. H. Chang, Z. Hong, H. Chen, H. Zhou, Q. Chen, Y. Liu, N. De Marco and Y. Yang, Nat. Nanotechnol., 2016, 11, 75–81.
19. T. Salim, S. Sun, Y. Abe, A. Krishna, A. C. Grimsdale and Y. M. Lam, J. Mater. Chem. A, 2015, 3, 8943–8969.
20. L. Meng, J. You, T. F. Guo and Y. Yang, Acc. Chem. Res., 2016, 49, 155–165.
21. C.-H. Chiang, Z.-L. Tseng and C.-G. Wu, J. Mater. Chem. A, 2014, 2, 15897–15903.
22. J. Seo, S. Park, Y. Chan Kim, N. J. Jeon, J. H. Noh, S. C. Yoon and S. I. Seok, Energy Environ. Sci., 2014, 7, 2642–2646.
23. Q. Wang, Y. Shao, Q. Dong, Z. Xiao, Y. Yuan and J. Huang, Energy Environ. Sci., 2014, 7, 2359–2365.
24. X. Liu, M. Lei, Y. Zhou, B. Song and Y. Li, Appl. Phys. Lett., 2015, 107, 063901.
25. Y. Shao, Z. Xiao, C. Bi, Y. Yuan and J. Huang, Nat. Commun., 2014, 5, 5784.
26. C. Kuang, G. Tang, T. Jiu, H. Yang, H. Liu, B. Li, W. Luo, X. Li, W. Zhang, F. Lu, J. Fang and Y. Li, Nano Lett., 2015, 15, 2756–2762.
27. F. Xia, Q. Wu, P. Zhou, Y. Li, X. Chen, Q. Liu, J. Zhu, S. Dai, Y. Lu and S. Yang, ACS Appl. Mater. Interfaces, 2015, 7, 13659–13665.
28. S. H. Liu, P. You, J. H. Li, J. Li, C. S. Lee, B. S. Ong, C. Surya and F. Yan, Energy Environ. Sci., 2015, 8, 1463–1470.
29. Y. Huang, W. Wen, S. Mukherjee, H. Ade, E. J. Kramer and G. C. Bazan, Adv. Mater., 2014, 26, 4168–4172.
30. D. Pathak, T. Wagner, T. Adhikari and J. M. Nunzi, Synth. Met., 2015, 199, 87–92.
31. D. Pathak, T. Wagner, T. Adhikari and J. M. Nunzi, Synth. Met., 2015, 200, 102–108.
32. Y. Bai, H. Yu, Z. L. Zhu, K. Jiang, T. Zhang, N. Zhao, S. H. Yang and H. Yan, J. Mater. Chem. A, 2015, 3, 9098–9102.
33. Y. H. Seo, J. S. Yeo, N. Myoung, S. Y. Yim, M. Kang, D. Y. Kim and S. I. Na, ACS Appl. Mater. Interfaces, 2016, 8, 12822–12829.
34. J. Liu, X. Guo, Y. J. Qin, S. D. Liang, Z. X. Guo and Y. F. Li, J. Mater. Chem., 2012, 22, 1758–1761.
35. J. Ge, J. Liu, X. Guo, Y. J. Qin, H. X. Luo, Z. X. Guo and Y. F. Li, Chem. Phys. Lett., 2012, 535, 100–105.
36. N.-J. Zhao, Z.-H. Lin, W. Zhang, J. Liu, Y.-W. Wang, Y.-J. Qin, X.-C. Ai, Z.-X. Guo and J.-P. Zhang, J. Phys. Chem. C, 2014, 118, 21377–21384.
37. H. Dong, Z. Wu, B. Xia, J. Xi, F. Yuan, S. Ning, L. Xiao and X. Hou, Chem. Commun., 2015, 51, 8986–8989.
38. N. Tripathi, M. Yanagida, Y. Shirai, T. Masuda, L. Han and K. Miyano, J. Mater. Chem. A, 2015, 3, 12081–12088.
39. D. Wang, Z. Liu, Z. Zhou, H. Zhu, Y. Zhou, C. Huang, Z. Wang, H. Xu, Y. Jin, B. Fan, S. Pang and G. Cui, Chem. Mater., 2014, 26, 7145–7150.
40. H. Wang, X. Zeng, Z. Huang, W. Zhang, X. Qiao, B. Hu, X. Zou, M. Wang, Y. B. Cheng and W. Chen, ACS Appl. Mater. Interfaces, 2014, 6, 12609–12617.
41. Z. Liu, M. Zhang, X. Xu, L. Bu, W. Zhang, W. Li, Z. Zhao, M. Wang, Y. B. Cheng and H. He, Dalton Trans., 2015, 44, 3967–3973.
42. J. Cui, C. Chen, J. Han, K. Cao, W. Zhang, Y. Shen and M. Wang, Adv. Sci., 2016, 3, 1500312.
43. Y. Li, Y. Zhao, Q. Chen, Y. M. Yang, Y. Liu, Z. Hong, Z. Liu, Y. T. Hsieh, L. Meng, Y. Li and Y. Yang, J. Am. Chem. Soc., 2015, 137, 15540–15547.
44. R. S. Sanchez, V. Gonzalez-Pedro, J. W. Lee, N. G. Park, Y. S. Kang, I. Mora-Sero and J. Bisquert, J. Phys. Chem. Lett., 2014, 5, 2357–2363.
45. 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, 7081.
46. J. M. Marin-Beloqui, J. P. Hernandez and E. Palomares, Chem. Commun., 2014, 50, 14566–14569.
47. H. Y. Wang, Y. Wang, M. Yu, J. Han, Z. X. Guo, X. C. Ai, J. P. Zhang and Y. Qin, Phys. Chem. Chem. Phys., 2016, 18, 12128–12134.
48. Y. Wang, H. Y. Wang, M. Yu, L. M. Fu, Y. Qin, J. P. Zhang and X. C. Ai, Phys. Chem. Chem. Phys., 2015, 17, 29501–29506.
49. D.-L. Gao, Y. Wang, P. Zhang, L.-M. Fu, X.-C. Ai and J.-P. Zhang, RSC Adv., 2015, 5, 84959–84966.
50. Y. S. Liu, Q. Chen, H. S. Duan, H. P. Zhou, Y. Yang, H. J. Chen, S. Luo, T. B. Song, L. T. Dou, Z. R. Hong and Y. Yang, J. Mater. Chem. A, 2015, 3, 11940–11947.
51. D. Bi, G. Boschloo, S. Schwarzmuller, L. Yang, E. M. Johansson and A. Hagfeldt, Nanoscale, 2013, 5, 11686–11691.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra22023a

---