Lu Renabd,
Lusheng Liangad,
Zhuangzhuang Zhangd,
Zilong Zhangad,
Qiu Xiongabd,
Nan Zhaoc,
Yaming Yuc,
Rosario Scopelliti
e and
Peng Gao
*abd
aCAS Key Laboratory of Design and Assembly of Functional Nanostructures, Fujian Provincial Key Laboratory of Nanomaterials, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, Fujian 350002, China. E-mail: peng.gao@fjirsm.ac.cn
bUniversity of Chinese Academy of Science, Beijing 100049, China
cCollege of Materials Science and Engineering, Huaqiao University, 361021 Xiamen, China
dLaboratory for Advanced Functional Materials, Xiamen Institute of Rare Earth Materials, Haixi Institute, Chinese Academy of Sciences, Xiamen 361021, China
eInstitut des Sciences et Ingénierie Chimiques, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
First published on 19th January 2021
It took only 11 years for the power conversion efficiency (PCE) of perovskite solar cells (PSCs) to increase from 3.8% to 25.2%. It is worth noting that, as a new thin-film solar cell technique, defect passivation at the interface is crucial for the PSCs. Decorating and passivating the interface between the perovskite and electron transport layer (ETL) is an effective way to suppress the recombination of carriers at the interface and improve the PCE of the device. In this work, several acceptor–donor–acceptor (A–D–A) type fused-ring organic semiconductors (FROS) with indacenodithiophene (IDT) or indacenodithienothiophene (IDDT) as the bridging donor moiety and 1,3-diethyl-2-thiobarbituric or 1,1-dicyromethylene-3-indanone as the strong electron-withdrawing units, were deposited on the SnO2 ETL to prepare efficient planar junction PSCs. The PCEs of the PSCs increased from 18.63% for the control device to 19.37%, 19.75%, and 19.32% after modification at the interface by three FROSs. Furthermore, impedance spectroscopy, steady-state and time-resolved photoluminescence spectra elucidated that the interface decorated by FROSs enhance not only the extraction of electrons but also the charge transportation at the interface between the perovskite and ETL. These results can provide significant insights in improving the perovskite/ETL interface and the photovoltaic performance of PSCs.
The device structure of PSCs usually consists of transparent conductive oxides (i.e., fluorine-doped tin oxide FTO and indium tin oxide ITO), electron transport layer (ETL, e.g., TiO2, SnO2, ZnO, C60), ABX3 type perovskite layer (where the A is a cation, e.g., methylammonium (MA+), formamidinium (FA+), Cs+; B is a metal cation, e.g., Pb2+, Sn2+; and X is an anion, e.g., I−, Br−, Cl−, SCN−),11 hole transport layer (HTL, e.g., Spiro-MeOTAD, PTAA, P3HT, CuSCN, CuPc) and metal electrode (e.g., Au, Ag).3,12 The primary function of ETL is to transport electrons while blocking holes from recombination, which is essential for the PCE and lifespan properties of PSCs.13 Due to the need for the high temperature (≥450 °C) annealing process, the application of traditional TiO2 as ETL in energy-efficient PSCs is becoming unfavorable.14,15 On the other hand, low temperature processable (<150 °C) SnO2 with better optical and electric properties than TiO2 is conceptually considered as a superior candidate for highly efficient PSCs.16,17 So far, ETLs based on SnO2 have demonstrated excellent features like high electron mobility, wide band gap,18 good photostability and high transparency,19 and good energy level alignment with the perovskite absorber.20 However, it is worth noting that the defects at the SnO2/perovskite interface vary dramatically depending on the fabrication methods, which will affect the performance and stability of the devices.11,21,22 For example, these defects may cause charge accumulation and non-radiative recombination,3 which will lead to the hysteresis effect, performance loss, and deterioration of device stability.11,23 Researches have indicated that passivation at the SnO2/perovskite interface can significantly inhibit the formation of interface defects.11,16 For example, fullerene derivatives,13,14,24,25 graphene,26 graphene quantum dots (GQDs),27 methylammonium chloride,28 fused-ring organic semiconductors (FROS),29,30 ionic liquids,22,31 and self-assembled monolayers11 are used as passivators to promote the extraction of electrons from the perovskite layer to the ETLs.
The FROS materials have recently become popular non-fullerene acceptors in organic solar cells due to excellent characteristics such as tunable optical band gap and energy level, high electron mobility, excellent photo-thermal stability, etc.32 Normally, such molecules are composed of a ring-fused backbone with two strong electron-withdrawing groups at both ends, which usually contain Lewis-base-type functional groups such as carbonyl groups, cyano groups, and thiocarbonyl groups. Therefore, the application of n-type FROS as a passivator at the interface between ETL and perovskite could effectively passivate the under-coordinated Pb2+ ions and further improve the performance of PSCs.
Herein, we use a series of n-type FROSs called IDT-T, IDT-I, and IDDT-T to passivate the interface between the perovskite layer and SnO2 layer in PSCs, leading to reduced interfacial loss and less hysteresis in hopes of enhancing the performance of PSC devices. Among them, IDT-T and IDT-I are comprised of one IDT backbone and two 1,3-diethyl-2-thiobarbituric and two 1,1-dicyromethylene-3-indanone electron-withdrawing end groups, respectively. In comparison, IDDT-T is composed of an IDDT ore and two 1,3-diethyl-2-thiobarbituric terminal groups.33–35 In the device structure of FTO/SnO2/FROS/perovskite/Spiro-OMeTAD/Au, the effect of the three FROSs on the structure, morphology, and photoelectric properties of the interface are characterized and compared by scanning electron microscopy (SEM), steady-state/time-resolved photoluminescence spectroscopy, electrochemical impedance spectroscopy (EIS), and other techniques. The results of the work showed that the PCE of PSCs modified by the FROSs layer increased from 18.63% without interface modification to 19.37% (IDT-T), 19.75% (IDT-I), 19.32% (IDDT-T). This improvement is due to the larger grain size of the perovskite, shorter decay lifetimes, smaller series resistance, and larger recombination resistance of the device compared to undoped PSCs. The performances (VOC, JSC, and FF) of the PSCs modified by the FROS layer have been improved, but the efficiency does not increase significantly, which may be due to the matching degree between energy levels.
![]() | ||
Fig. 1 (a) Device structure. (b) Chemical structures of IDT-T, IDT-I, and IDDT-T. (c) Calculated ESP profiles of IDT-T, IDT-I and IDDT-T. |
In order to accurately analyze the effect of different acceptor moiety on the molecular configuration and packing in the condensed state, single-crystal X-ray diffraction (SC-XRD) analysis was performed, and single crystals of IDT-T and IDT-I were successfully prepared by slowly diffusing methanol into their solutions in dichloromethane.36 Fig. S18† shows the single-crystal structures of the two molecules observed from two perspectives. The corresponding crystallographic parameters are summarized in Table S2.† It can be seen from Fig. 2 that in a unit cell, two independent IDT-T molecules stack tightly via four O–H hydrogen bonds (symmetry, 2.717 Å and 2.613 Å) and two C–H–π (symmetry, 2.364 Å) noncovalent intermolecular short contacts (Fig. 2(a)). In the case of IDT-I, two independent molecules pack closely by C–H – C–H (2.297 Å) interactions (Fig. 2(c)).37 Notably, both IDT-T and IDT-I exhibit intramolecular S⋯OC short contacts with distances (2.648 Å and 2.620 Å) closer than the sum of the van der Waals radii of the S and O (3.25 Å), which can non-covalently lock the molecular conformation.38–40 This interlocked network will not only provide better molecular coplanarity but also endow closed and ordered molecular packing.36 The torsion angles between the bridging π backbones and the acceptors are 11.83° and 4.38° for IDT-T and IDT-I, respectively, indicating better planarity for IDT-I than IDT-T.
The thermal stability of the three FROSs was checked by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). As shown in Fig. S20,† the thermal stability of IDT-I is higher than IDT-T and IDDT-T. At 5% weight loss, the decomposition temperatures (Td) of IDT-T, IDT-I, and IDDT-T are 349 °C, 375 °C, and 342 °C, respectively. DSC measurement showed that the three FROSs did not show noticeable glass transition temperature (Tg) and melting temperature (Tm). This indicates that the three compounds are very stable within 300 °C without phase transition (Fig. S21†).
The ultraviolet-visible (UV-vis) absorption spectra of the FROSs show a similar profile (Fig. 3(a)) with two characteristic absorption peaks. The IDDT-T has a longer conjugation length than that of IDT-T, so the highest occupied molecular orbit (HOMO) energy level is more destabilized, and the absorption peak is slightly red-shifted. The introduction of the stronger electron-withdrawing dicyano groups stabilized the lowest unoccupied molecular orbital (LUMO) energy level further, thereby also reduces the bandgap of the system and shifts the absorption peak to a longer wavelength. Therefore, IDT-I showed the most red-shifted absorption band (646 nm) comparing to IDT-T (592 nm) and IDDT-T (612 nm). UV-Vis spectra of FROSs in thin-film state red-shifted 19 nm, 46 nm, and 20 nm for IDT-T, IDT-I, and IDDT-T, respectively, compared to those measured in dichloromethane (DCM), Fig. S22† suggesting intermolecular aggregation in the solid state.34 Based on the intersection of the UV-vis spectra and fluorescence spectra (Fig. 3(a)), the optical bandgap (Egopt) of IDT-T, IDT-I, and IDDT-T are calculated as 2.03 eV, 1.83 eV, and 1.95 eV, respectively.
Besides, as shown in the cyclic voltammetry (CV) measurement in Fig. 3(b), all the three FROSs show irreversible oxidation potentials. Based on the onset of the CV curves, we can calculate the corresponding HOMO energy levels as −5.78 eV (IDT-T), −5.76 eV (IDT-I), and −5.69 eV (IDDT-T). The LUMO energy levels can then be subsequently calculated as −3.75 eV (IDT-T), −3.93 eV (IDT-I), and −3.74 eV (IDDT-T). As shown in the energy diagram in Fig. 4(e), the IDT-I has the best matched LUMO energy level with the conduction band of perovskite, favoring the transfer of electrons from the perovskite layer to the ETL. Moreover, DFT calculations are also used to estimate the HOMO and LUMO orbitals of FROSs. As shown in Table 1, a similar trend with the experimental values was achieved.
![]() | ||
Fig. 4 SEM top-view images of (a) the pristine perovskite film and the perovskite films with (b) IDT-T, (c) IDT-I, and (d) IDDT-T optimized treatment conditions. (e) Energy-level diagrams of the corresponding materials of the device.13,45 |
λonset [nm] | Ega [eV] | EHOMOb [eV] | ELUMOc [eV] | Egcal [eV] | EHOMOcal [eV] | ELUMOcal [eV] | |
---|---|---|---|---|---|---|---|
a Eg = 1243/λonset (eV).b EHOMO = −5.1 − (Eox − E1/2(Fc/Fc+)) (eV).c ELUMO = Eg + EHOMO (eV). | |||||||
IDT-T | 612.8 | 2.03 | −5.78 | −3.75 | 2.44 | −5.69 | −3.25 |
IDT-I | 679.8 | 1.83 | −5.76 | −3.93 | 2.24 | −5.63 | −3.39 |
IDDT-T | 638.8 | 1.95 | −5.69 | −3.74 | 2.32 | −5.45 | −3.13 |
Fig. 3(c) shows the transmittance spectra of SnO2 and SnO2/FROS thin films prepared on FTO substrates. Compared with the SnO2 substrate, the light transmittance of SnO2/FROS reduced slightly for IDT-T at 470–630 nm, IDT-I at 470–735 nm, and IDDT-T at 300–800 nm. This indicates that incorporating the FROS interlayer does not induce significant optical losses.41 The electron mobility of different FROSs based SnO2 ETLs were measured using the space charge limited current (SCLC) model.19 The result indicates that the introduction of ultrathin FROSs does not affect electron mobility significantly (Fig. S23 and Table S3†). To check the influence of the FROS layer on the perovskite layer deposited atop, thin-film X-ray diffraction (XRD) was used to characterize the solid structure of the perovskite films, and the results are displayed in Fig. 3(d). The diffraction peaks at 14.1°, 28.3°, and 31.8°, are attributed to the 〈110〉, 〈220〉, and 〈310〉 facets of the perovskite.41–43 The peak at 12.7° results from excessive PbI2, the existence of which has been demonstrated to have a positive effect on grain boundary passivation.11,44 〈220〉 crystal planes of FROSs compared to this from the pristine perovskite is a clear evidence of enhanced perovskite crystallinity. Generally, the addition of FROSs did not destroy the intragranular crystal structure of the perovskite. However, compared to the XRD of pristine perovskite film, the 〈110〉 reflection in the XRD of the perovskite films deposited on FROS modified SnO2 is more intense with a smaller full-width at half-maximum (FWHM). This observation agrees with the enlarged grains shown by scanning electron microscope (SEM) in Fig. 4(a–d) and S24,† confirming the enhanced crystallinity in the presence of the FROS layer (vide infra).
To study the morphology of the perovskite films based on IDT-T, IDT-I, and IDDT-T, SEM top-view images of the pristine perovskite film and the perovskite films deposited on the FROS layer under optimized conditions are shown in Fig. 4. The comparison shows that the crystal grain size of the modified perovskite is significantly larger than that of the pristine perovskite, meaning the FROSs underlayer can help to promote the growth of perovskite crystal grains. The reason behind the increased grain size after FROS modification is linked with the hydrophobic nature of organic FROSs surface, which affects the nucleation and grain growth behavior.46 Larger grains can reduce the defects of the perovskite and hence the interfacial recombination of the photo-generated charges.47 Besides, the cross-sectional SEM of SnO2/FROS-based PSCs (Fig. S24†) shows that the perovskite crystal grains penetrate almost the entire perovskite light-absorbing layer in the longitudinal direction. In contrast, the perovskite on pristine SnO2 shows more grain boundaries, which may be responsible for the higher interfacial recombination. XRD and SEM demonstrated that the addition of the FROS layer enhances the crystallization of perovskite.
Device performances are closely correlated to charge dynamics in perovskite solar cells. The electronic quality and charge transfer at the ETL/perovskite interface was evaluated by steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) measurements.48–50 Fig. 3(e) represents the steady-state PL spectra of the perovskite films deposited on neat SnO2 and SnO2/FROSs, respectively. Significant PL quenching is observed on perovskite films deposited on SnO2/FROS, contrasting to the intensive emission peak of the reference sample at ∼795 nm. This could be caused by the improved charge carrier extraction after the insertion of the FROS and reduced surface charge trapping. The FROS films could facilitate the charge transport between perovskite and ETL.
To further study the effect of the FROS layer on the dissociation and recombination of the charge carriers, TRPL was conducted. Fig. 3(f) exhibits the TRPL spectra of SnO2/perovskite film and the perovskite films deposited on SnO2/FROS. The PL decay curves were fitted by a biexponential decay function, including a fast decay (τ1) and a slow decay (τ2) component.13,51 τ1 can be attributed to the PL quenching via trap states or ETL/perovskite interfacial charge transfer,52,53 and τ2 can be attributed to the PL quenching by the radiative recombination of free charges.44,54 The detailed fitting parameters are listed in Table S4.† The average lifetimes (τavg) of different films were estimated as 105.39, 44.10, 61.02, and 53.03 ns for the SnO2/perovskite, SnO2/IDT-T/perovskite, SnO2/IDT-I/perovskite, and SnO2/IDDT-T/perovskite, respectively. The decay lifetimes of SnO2/FROS/perovskite decreased dramatically compared to that of the SnO2/perovskite. After using the FROS modified SnO2, the electrons can transfer more efficiently from the perovskite active layer to the ETL, and less recombination occurred inside the perovskite layer. The TRPL result is consistent with the steady PL measurement.
In Fig. 5(a), the current density–voltage (J–V) characteristics of the champion devices based on SnO2 and SnO2/FROS substrates were measured under simulated air mass 1.5 global (AM 1.5G) solar irradiation. The corresponding detailed photovoltaic parameters of devices based on pristine SnO2 and SnO2/FROS substrates are summarized in Table 2. The most important observation from the J–V curves was the beneficial role of the FROS decoration of the SnO2 film. The best PCEs from SnO2/FROS based device (SnO2/IDT-T: 19.37%, SnO2/IDT-I: 19.75%, SnO2/IDDT-T: 19.32%) are all enhanced comparing to the neat SnO2 based device (18.63%) in the reverse scanning (RS) direction. To quantify the hysteresis effect, the hysteresis index (HI) was calculated according to the method described in the previous reports.55,56 Compared with SnO2 based device (HI = 7.19%), SnO2/IDT-I based device has lower HI (4.4%), while SnO2/IDT-T and SnO2/IDDT-T based devices show slightly increased HI (9.98% and 7.61%). The effective suppression of hysteresis by the IDT-I treatment leads to optimized efficiency and significantly improved consistency in the extracted photovoltaic parameters under different scan directions.41
Material | Sweep | VOC (V) | JSC (mA cm−2) | FF (%) | Efficiency (%) | HIa (%) |
---|---|---|---|---|---|---|
a HI = (PCERS − PCEFS)/PCERS % 100% | ||||||
SnO2 | RS | 1.029 | 23.61 | 76.70 | 18.63 | 7.19 |
FS | 1.021 | 23.46 | 72.19 | 17.29 | ||
SnO2/IDT-T | RS | 1.023 | 24.00 | 78.91 | 19.37 | 9.71 |
FS | 1.021 | 24.40 | 70.16 | 17.49 | ||
SnO2/IDT-I | RS | 1.051 | 24.02 | 78.22 | 19.75 | 4.40 |
FS | 1.080 | 23.68 | 73.82 | 18.88 | ||
SnO2/IDDT-T | RS | 1.022 | 23.80 | 79.44 | 19.32 | 7.61 |
FS | 1.055 | 23.25 | 72.77 | 17.85 |
Fig. 5(b) shows the statistics of PCE of devices with SnO2 and SnO2/FROS substrates. The average PCEs of the devices with SnO2, SnO2/IDT-T, SnO2/IDT-I, and SnO2/IDDT-T ETLs are 18.27%, 18.92%, 19.22%, and 19.10%, respectively. The corresponding statistic values of VOC, JSC, and FF are listed in Fig. 5(c–e), respectively (Table S5†). Devices based on SnO2/IDT-I ETL has the highest PCE due to the outstanding VOC and FF, which can be attributed to the following reasons: (i) the dicyano group of IDT-I passivates the defect of perovskite film and thereby reduces the interface carrier recombination;57 (ii) the more favorable LUMO level of IDT-I (Fig. 4(e)) benefits the electrons transport from the perovskite layer to the SnO2.
Besides, the EQE spectra and integrated JSC of the best-performance devices with SnO2 and SnO2/FROS substrates are shown in Fig. 5(f). The devices with the SnO2/FROS ETL represented higher quantum efficiency in the range of 300–850 nm than the devices with the SnO2 substrate. The integrated photocurrent values from the EQE spectra of the devices with SnO2, SnO2/IDT-T, SnO2/IDT-I, and SnO2/IDDT-T substrates are 22.90, 23.44, 23.07, and 23.03 mA cm−2. The deviations of the corresponding samples are 3.01%, 2.66%, 3.96%, and 3.24%, respectively, which are within 4% error compared to the corresponding JSC obtained from J–V curves.
The EIS measurement of the devices was conducted to analyze the effect of the FROS layer on the interfacial charge transfer and carrier recombination behavior. Fig. 6(a) depicts the Nyquist plots of the devices with and without the FROS layer and equivalent circuits (inset). The fitted parameters by the equivalent circuits are listed in Table S7.† In the Nyquist plots, there are two different semicircles locates at different frequency ranges. The high-frequency region and low-frequency region correspond to the charge transfer resistance (Rct) and recombination resistance (Rrec), respectively.58,59 It notes that the series resistance in the PSC could be reduced through the introduction of FROS. The Rs of the optimized FROS containing devices are almost half of the Rs of the reference device, indicating reduced contact resistance at the perovskite/ETL interface. Although the Rcts of the optimized devices are slightly higher than that of the reference device, their Rrecs are significantly increased. A higher Rrec suggests reduced non-radiative recombination, so the FROS treated devices all show higher VOCs than that of the reference device, which may be attributed to the increased grain size of perovskite as discussed above (Fig. 4(a–d) and S24†). Therefore, these results also revealed that, after introducing the FROS layer, the carrier transport efficiency is improved and the carrier recombination is suppressed, both of which are beneficial for increasing FF value (Fig. 5(e)).60,61
![]() | ||
Fig. 6 (a) Electrical impedance spectroscopy (EIS), Nyquist plots of the devices with SnO2, SnO2/IDT-T, SnO2/IDT-I, and SnO2/IDDT-T substrates at −0.5 V under dark conditions, respectively (equivalent circuit model for the Nyquist plots.53 Rs: series resistance. Rct: charge transfer resistance. Rrec: resistance of interfacial recombination. C: capacitance). (b) Maximum power output and current density of the device with SnO2, SnO2/IDT-T, SnO2/IDT-I, and SnO2/IDDT-T substrates (at 0.87 V). |
To further study the effect of FROS passivation on perovskite defect, the SCLC measurement based on electronic-only devices (FTO/SnO2/FROS/perovskite/PCBM/Ag) with or without FROS was conducted to obtain the trap density statistics of perovskite films.62,63 The average trap densities were determined from the dark current–voltage characteristics (electronic-only, Fig. S25†) based on eqn (S2), ESI.† The results show that after introducing the FROS, the trap density decreases from 1.19 × 1015 cm−3 for control to 9.82 × 1014 cm−3, 8.16 × 1014 cm−3 and 8.80 × 1014 cm−3 for SnO2/IDT-T, SnO2/IDT-I, and SnO2/IDDT-T samples, respectively. The reduced trap density indicates that FROS materials could passivate the defects such as under-coordinated Pb2+ ions using their Lewis base type functional groups like carbonyl, cyano, and thiocarbonyl groups.
The stability is also crucial to the development of PSCs. Fig. S26† displays the long-term humidity stability of different ETL-based devices. The sample SnO2/IDT-I showed slightly better stability than other samples under 10–20% RH in the atmosphere after 240 h, even though the performance of all samples did not exhibit an apparent deterioration. Fig. 6(b) shows the steady-state power output and photocurrent output of the best devices, measured at its maximum power point (VMPP = 0.87 V) for 300 s. The device with SnO2 substrate exhibits a photocurrent of 20.7 mA cm−2 and a PCE of 18.0%. In comparison, the device with SnO2/IDT-T, SnO2/IDT-I, and SnO2/IDDT-T substrate exhibit photocurrents of 21.9 mA cm−2, 22.2 mA cm−2, and 21.7 mA cm−2 as well as PCEs of 19.1%, 19.3%, and 18.8%, respectively. Compared with the values extracted from the J–V curves (PCEJ–V), the steady-state PCEs (PCESS) of the four device structures lose 3.3%, 1.6%, 2.3%, and 2.5% ((PCEJ–V–PCESS)/PCEJ–V × 100%) for the control, SnO2/IDT-T, SnO2/IDT-I and SnO2/IDDT-T samples, respectively. It indicates that the devices with SnO2/FROS substrate have lower PCE loss and better power output stability than control. This improvement could be ascribed to the reduced defect in the devices. Firstly, the incorporation of FROS between SnO2 and perovskite can passivate the defect at the SnO2/perovskite interface through the interaction between the Lewis-base-groups with under-coordinated Pb2+ ions in perovskite. Secondly, the introduction of FROS improved the morphology of perovskites with enlarged grains, thereby reducing the grain boundaries and the defects therein. One of the SnO2/IDT-I based devices could keep 97.4% of initial PCE after 2400 s measurement at its maximum power point (Fig. S27†). These results prove that the addition of the FROS layer improves device stability.
Footnote |
† Electronic supplementary information (ESI) available: Fig. S1–S27 and Tables S1–S7 provide additional characterization including structure characterization, DFT calculation, thermal properties and device performance. CCDC 2036683 and 2036684. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d1ra00090j |
This journal is © The Royal Society of Chemistry 2021 |