Passivation of the grain boundaries of CH₃NH₃PbI₃ using carbon quantum dots for highly efficient perovskite solar cells with excellent environmental stability

Abstract

Organic–inorganic hybrid perovskites are prone to defect formation due to iodine and methylamine ion/defect migration, leading to the formation of lots of defects at the perovskite surface and grain boundaries. Passivation of the defects is an effective method to improve the power conversion efficiency (PCE) and stability of perovskite solar cells (PSCs). To achieve stable passivation, the interaction between the perovskite and additive materials should be taken into consideration. In this work, we for the first time introduced carbon quantum dots (CQDs) as an additive for the stabilization of MAPbI₃via passivation of the grain boundaries of the perovskite. Because the carboxylic groups, hydroxyl groups and amino-groups on the edge of CQDs can bond with the uncoordinated Pb in MAPbI₃, strong and stable interactions between the perovskite and CQDs can be generated, inducing a lower trap-state density and better optoelectronic properties. The typical PCE of the PSCs based on CQD modified MAPbI₃ films increases from 17.59% to 18.81% and the PCE of the optimized champion PSCs reaches 19.38%. Furthermore, the hydrophobic CQD molecules can block the contact between water and MAPbI₃, and even if the CQD modified perovskite is kept under ambient atmosphere without controlling the humidity for 4 months, the MAPbI₃ film still retained its original black color.

---

Introduction

Due to the superior optoelectronic properties of perovskite materials and unremitting efforts of researchers from the fields of materials, chemistry and physics, rapid progress has been made in organic–inorganic hybrid perovskite solar cells (PSCs). The certified best research-cell power conversion efficiency (PCE) reaches 23.3%,¹ which is even higher than that of commercialized solar cells such as multicrystalline Si (22.3%), CdTe (22.1%), and CIGS (22.6%). Although the efficiency question has been answered, there still remain big issues in stability and cost that must be addressed before commercialization of the PSCs.

A high-quality perovskite film with great crystallization and a low trap-state density is a primary requirement for high efficiency PSCs. Since the all-solid-state PSCs reported by Nam-Gyu Park et al. in 2012,² a lot of methods such as solution spin-coating,^3–5 vapor assisted deposition^6–8 and an anti-solvent assisted process^9–11 have been developed to improve the crystallization of the perovskite. However, organohalide perovskites are prone to defect formation, and lots of defects are found at the perovskite surface and grain boundaries, due to the iodine and methylamine ion/defect migration.¹² Defects on the perovskite surface and at grain boundaries inevitably introduce localized trap states, which could act as a recombination center of the photogenerated carriers, leading to unsatisfactory photovoltaic performance. Therefore, passivation of the defect is an effective method to improve the PCE and stability of PSCs.

Excess PbI₂ or CH₃CH₃I (MAI) has been used to passivate or heal the defects at the grain boundaries for suppressing nonradiative charge recombination.^13–18 Besides the passivation by introducing intrinsic ions of perovskites, fullerene based electron acceptors deposited on the perovskites can simultaneously passivate the charge trap states and transport electrons.¹⁹ Recently, contact-passivation by using a chlorine-capped TiO₂ nanocrystal film has also been performed to reduce the interfacial recombination and to enhance the interface binding for large dimensional PSCs.²⁰ Lewis bases thiophene and pyridine are used to treat the perovskite crystal surfaces, and a significant reduction of the non-radiative electron–hole recombination can be observed due to the electronic passivation of the uncoordinated Pb atoms within the crystals.²¹ As a guidance for this approach, various semiconductor polymers containing electron-rich atoms, such as S, N, and O, have been chosen to passivate perovskite films, and this has been proven to be a promising strategy to improve the performance of PSCs.^22–24

To achieve stable passivation, the interaction between the perovskite and additive materials should be taken into consideration. CQDs exhibit unique quantum confinement and surface and edge effects, which can be easily controlled by tailoring the size, shape and heteroatom doping and modifying the surface and edge of CQDs.^25,26 Various functional groups such as carbonyl, hydroxyl, carboxyl, and epoxy groups can be found on the surfaces of CQDs due to the solution based chemical synthesis routes, which makes it easy to functionalize them with various organic, polymeric, or biological species.^27,28 In PSCs, an ultrathin CQD layer or a graphene quantum dot layer has been inserted between the perovskite and TiO₂.^29,30 By adding a CQD layer, the electron mobility and the electron extraction ability of the electron transport layer can be greatly improved. CQDs have also been used as sensitizers for all-inorganic CsPbBr₃ inverse opal PSCs, which enhance light absorption and facilitate charge transfer.³¹

In this work, we for the first time introduced CQDs as an additive for the stabilization of MAPbI₃via passivating the grain boundaries of the perovskite. Strong and stable interactions between the uncoordinated Pb atoms in MAPbI₃ and the carboxylic groups, hydroxyl groups and amino-groups on the edge of CQDs were detected, which could induce a lower trap-state density and better optoelectrical properties. Furthermore, the hydrophobic CQD molecules can attach on the MAPbI₃ surface to block the contact between water and MAPbI₃, thus enhancing the stability of the perovskite. The PSCs with an appropriate concentration of CQDs exhibit more superior photovoltaic performance and long-time stability.

Experimental section

Materials

Fluorine doped tin oxide (FTO) glass (15 Ω sq⁻¹) was purchased from Wuhan Geao company (China). 75% titanium (diisopropoxide) bis(2,4-pentanedionate) (TIPD) isopropanol solution, dimethyl sulfoxide (DMSO, extra dry 99.7+%) and N,N-dimethylformamide (DMF, extra dry 99.7+%) were purchased from Alfa Aesar. Extra dry isopropyl alcohol, ethanol and terpineol were bought from Acros Organics. 30 nm particle paste of mesoporous TiO₂ was bought from Dyesol. PbI₂, CH₃NH₃I, 2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenylamine)-9,9-spirobifluorene (spiro-OMeTAD), 4-tertbutylpyridine (TBP) and lithium bis(trifluoromethanesulfonyl)imide (Li-TFSI) for perovskite film and hole transport layer fabrication were purchased from Xi'an Polymer Light Technology Corp.

Synthesis of CQDs

The CQDs were synthesized according to our previously reported solvothermal treatment method.²⁵ Firstly, 100 mg of citric acid and 20 mg of 1,5-diaminonaphthalene (DAN) were dissolved in 15 mL ethanol, and the solution was then transferred into a 25 mL poly(tetrafluoroethylene) (Teflon)-lined autoclave and heated at 200 °C for 4 h for the formation of CQDs. After the reaction, the reactors were naturally cooled down to room temperature, the solution was diluted with deionized water, and the supernatant was collected by centrifugation. Finally, the solution was subjected to dialysis for two days in order to obtain the CQDs.

Precursor preparation

The pure PbI₂ precursor was obtained by dissolving 1.3 M PbI₂ in a DMSO : DMF mixed solvent for 5 hours. For the PbI₂ precursor of the CQD modified MAPbI₃ film, different amounts of CQDs were added to the PbI₂ precursor to obtain solutions with different concentrations of CQDs. MAI was dissolved in isopropanol with a concentration of 50 mg mL⁻¹.

Device fabrication

The patterned FTO glass was ultrasonically cleaned twice in succession with detergent, water, deionized water and ethanol. The cleaned FTO glass was blown dry with nitrogen and then heated to 450 °C. Then, a compact TiO₂ (c-TiO₂) layer was fabricated by spraying 10 mL of TIPD/ethanol (1 : 9) solution onto the heated FTO glass and sintered at 450 °C for 30 minutes. 100–150 nm mesoporous TiO₂ (mass ratio of TiO₂ : ethanol : terpineol is 1 : 7.5 : 0.75) was spin-coated on c-TiO₂ at 5000 rpm for 30 s, the substrate was dried at 125 °C for 10 min and then the organic part was removed at 500 °C for 30 min. The pure PbI₂ precursor and different CQD added PbI₂ precursors were spin-coated on the TiO₂ substrate at 3000 rpm for 30 s after the substrate cooled down to room temperature. Subsequently, the MAI solution was spin-coated on the as-prepared PbI₂ film in a glove-box and then annealed at 100 °C for 30 min in atmospheric environment with a RH less than 30% to form a MAPbI₃ film. The spiro-OMeTAD solution (72.3 mg of spiro-OMeTAD, 29 μL of TBP and 17.5 μL of Li-TFSI (520 mg mL⁻¹ in acetonitrile) dissolved in 1 mL of chlorobenzene) was spin-coated on the perovskite film at 3000 rpm for 30 s. Finally, 100 nm Ag was thermally evaporated on spiro-OMeTAD at a pressure of 3 × 10⁻⁴ Pa.

Instrumentation and characterization

TEM images were obtained by JEOL JEM 2100 transmission electron microscopy. SEM was conducted using a Hitachi S4800 at acceleration voltages of 2.5 kV and 5 kV. XRD was measured using a D8 Advance (Bruker) diffractometer with an X-ray tube for Cu Kα radiation (λ = 0.1542 nm). The PL of the MAPbI₃ film was acquired using a FluoroLog-3 modular spectrofluorometer from Horiba Scientific and the samples for the PL test were excited using a 500 nm laser. A steady/transient state fluorescence spectrometer F900 (Edinburgh Instruments) was used for TRPL measurements. UV-vis absorption spectra of the perovskite films were recorded on a Shimadzu UV-2450. FTIR spectra were recorded using a Nicolet 380 spectrograph. The pristine MAPbI₃ film, CQD modified MAPbI₃ film and CQD added PbI₂ films for FTIR test were deposited on the KBr substrate. J–V curves of the PSCs were obtained using a computer-controlled Keithley 2400 source meter under standard one sun AM1.5G (100 mW cm⁻²) illumination. The active area of the device is 4 mm². The EQE of the PSCs was tested using QE-R systems (Enli Tech.).

Density functional theory (DFT) calculations

The structure of the CQDs was firstly optimized using Gaussian 09 based on Hartree-Fock (HF) with basis sets 6-31G*. Then first-principles calculation was performed using a Vienna ab initio simulation package (VASP) based on DFT.^32,33 The generalized-gradient approximation (GGA)-Perdew–Burke–Ernzerhof (PBE) functional of the augmented plane wave was applied. A cutoff energy of 500 eV for the plane-wave basis set and a Monkhorst–Pack mesh of 1 × 3 × 2 for the Brillouin zone integration were employed. A vacuum layer of 10 Å was adopted in the calculation. The structure is fully relaxed until the maximum Hellmann–Feynman forces acting on each atom is less than 0.01 eV Å⁻¹. The binding energy E_binding between CQDs and MAPbI₃ is defined as E_binding = E_{(MAPbI3+CQDs)} − E_(MAPbI3) − E_(CQDs), where E_{(MAPbI3+CQDs)} is the total energy of MAPbI₃ with absorbed CQDs and E_(MAPbI3) and E_(CQDs) are the total energies of pristine MAPbI₃ and isolated CQDs, respectively.

Results and discussion

The CQDs are synthesized according to our previous literature report,²⁵ and the structure of the CQDs is illustrated in Fig. 1a. The diameter of the CQDs is 3–5 nm according to the transmission electron microscopy (TEM) image (Fig. 1b). To investigate the effect of the CQDs on the morphology of the MAPbI₃ film, scanning electron microscopy (SEM) of the MAPbI₃ films with different concentrations of CQDs is performed. Fig. 1c shows the SEM image of the pristine MAPbI₃ film, where big gaps between the perovskite grains can be observed. On adding 0.02 mg mL⁻¹ of CQDs into the PbI₂ precursor solution, the obtained MAPbI₃ film (Fig. 1d) demonstrates much more compact crystal grains and the gaps between the grains become much smaller than that of the pristine MAPbI₃ film. As the concentration of the CQDs increases to 0.05 mg mL⁻¹ (Fig. 1e), the gaps between the perovskite grains disappear and the grain boundaries can be clearly identified, which can also be verified from the cross-section SEM image as shown in Fig. 1h. Pristine MAPbI₃ and MAPbI₃ with 0.02 mg mL⁻¹ and 0.05 mg mL⁻¹ CQDs show similar grain size distribution. On further increasing the concentration of CQDs to 0.1 (Fig. 1f) and 0.25 mg mL⁻¹ (Fig. 1g), the grain size of MAPbI₃ is sharply decreased. The crystallization of MAPbI₃ is obviously damaged when the concentration of CQDs reaches 0.25 mg mL⁻¹ (Fig. 1g). To investigate the reason for the bad morphology of the perovskite film with excess CQDs, UV-visible (UV-vis) absorption spectroscopy and X-ray diffraction (XRD) characterization of the PbI₂ film with different concentrations of CQDs were conducted. As shown in Fig. S1,† the intensity of the absorption shoulder (around 490 nm) and XRD peaks (12.7°, 25.6°, 38.7°, and 52.4°) of PbI₂ are significantly decreased as the concentration of CQDs increases to 0.1 and 0.25 mg mL⁻¹, implying that the PbI₂ crystal growth is greatly inhibited. The reduction of the XRD peak of PbI₂ is caused by the interaction between CQDs and PbI₂, which can be validated using the Fourier transform infrared spectroscopy (FTIR) spectra of CQDs and the PbI₂ film with CQDs (10 mg mL⁻¹). As shown in Fig. S2,† the CQDs show C=O stretching vibrations at 1713 cm⁻¹, while they shifts to 1700 cm⁻¹ for the PbI₂ film with the CQD additive. The shift of the C=O stretching vibrations indicates that the PbI₂ film interacts with CQDs. Besides, from the SEM images of PbI₂ films with different concentrations of CQDs (Fig. S3†), we can see that the number of PbI₂ particles increases and the size of the particles decreases with higher concentrations of CQDs (0.1 and 0.25 mg mL⁻¹). These phenomena indicate that the addition of excessive CQDs in the PbI₂ film would lead to greater and faster nucleation of PbI₂ and inhibition of PbI₂ crystal growth. It has been reported that PbI₂ acts as a framework and provides “nucleation” centers for the subsequent crystallization of the perovskite film.^6,34 Thus, the excess nuclei in the PbI₂ film with 0.1 and 0.25 mg mL⁻¹ CQDs accelerate the nucleation of the perovskite and hinder the growth of the perovskite grain, resulting in the bad morphology of the perovskite film.

Fig. 1 (a) Molecular structure and (b) TEM image of CQDs: top-view SEM images of (c) pristine MAPbI₃ film and MAPbI₃ films with (d) 0.02, (e) 0.05, (f) 0.1, and (g) 0.25 mg mL⁻¹ CQDs deposited on the FTO/c-TiO₂/mp-TiO₂ substrate; (h) cross-section SEM images of the MAPbI₃ films with 0.05 mg mL⁻¹ CQDs deposited on the FTO/c-TiO₂/mp-TiO₂ substrate.

To further study whether CQDs can affect the crystal structure of the MAPbI₃ perovskite, XRD characterization was conducted. The XRD patterns of the MAPbI₃ film and MAPbI₃ films with different concentrations (<0.1 mg mL⁻¹) of CQDs are presented in Fig. 2a. The strong diffraction peaks of the perovskite at 14.1°, 28.4°, 31.9° and 40.7° can be observed in the XRD patterns of the MAPbI₃ films without CQDs.^35,36 The same perovskite peaks can also be identified for MAPbI₃ films with low concentrations (<0.1 mg mL⁻¹) of CQDs, and there are no extra peaks in the XRD curves of the MAPbI₃ films with CQDs, which means that the CQDs do not affect the perovskite crystal structure and the CQDs present in the MAPbI₃ films are in the amorphous state. Fig. 2b illustrates the normalized XRD curves from 13.7° to 14.4°. We can see that the XRD patterns of the pristine MAPbI₃ film and MAPbI₃ films with 0.02 and 0.05 mg mL⁻¹ CQDs exhibit almost equivalent full width at half maximum (FWHM) values, while the FWHM becomes broader in the XRD patterns of the MAPbI₃ films with 0.1 mg mL⁻¹ CQDs. The wider FWHM indicates that the grain size is smaller in the perovskite film with 0.1 mg mL⁻¹ CQDs, which is in good agreement with the result of the SEM images.^37,38

Fig. 2 (a) XRD patterns of the pristine MAPbI₃ film and MAPbI₃ films with 0.02, 0.05 and 0.1 mg mL⁻¹ CQDs; (b) normalized XRD patterns ((110) lattice planes) of the pristine MAPbI₃ film and MAPbI₃ films with 0.02, 0.05 and 0.1 mg mL⁻¹ CQDs.

According to the results of XRD and SEM, we can conclude that the CQD additive does not change the perovskite crystal structure but affects perovskite growth. Recently, Wang et al. proposed that a single-walled carbon nanotube with an octadecylamine functional group incorporated in perovskite film fabrication can restrict the evaporation of the solvent, thus enhancing the perovskite crystal and grain size.³⁹ In our work, due to the interaction between the functional group of CQDs and the perovskite, the CQDs are absorbed on the perovskite grain boundaries and impede the evaporation of the solvent during film formation. Besides, the interaction between PbI₂ and the functional group of CQDs can also retard the perovskite growth and improve crystallization.⁴⁰ So, more compact and well-crystallized perovskite films are obtained with a small amount of the CQD additive (less than 0.05 mg mL⁻¹). However, compared to the positive effect on perovskite crystallization induced by CQDs, when excess CQD additive was introduced in perovskite film fabrication, the negative effect induced by excess PbI₂ “nuclei” is more pronounced. Therefore, an appropriate amount of the CQD additive is crucial to form a compact perovskite film with enhanced crystallization.

Besides the morphology and crystallization, the photophysical properties of the perovskite film are also critical for high-performance PSCs. To investigate the photophysical properties of the MAPbI₃ films with CQDs, steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) of the MAPbI₃ films were measured. According to the SEM and XRD results, the optimized concentration of CQDs is fixed at 0.05 mg mL⁻¹ to simplify the process of analysis. As shown in Fig. 3a and b, the CQD modified MAPbI₃ film shows an increased PL intensity compared to the pristine MAPbI₃ film. The increased PL intensity should be attributed to the decreased nonradiative recombination caused by the defect,⁴¹ indicating that the CQDs induce passivation of the perovskite and greatly reduce the defect in the CQD modified MAPbI₃ film. The same information is also reflected by the TRPL curves (Fig. 3b), and the PL decay of the pristine MAPbI₃ film is faster than that of the CQD modified MAPbI₃ film. By fitting the decay curves with a three-component exponential, the average PL lifetime of pristine and CQD modified MAPbI₃ films are calculated as 154.35 ns and 170.48 ns, respectively (detailed fitting parameters are listed in Table S1†). To compare the trap-state density of the pristine MAPbI₃ film and the CQD modified MAPbI₃ film, space-charge-limited current (SCLC) devices with the structure of FTO/TiO₂/perovskite/PCBM/Ag were fabricated. The current density–voltage (J–V) characteristic curves of the SCLC devices were tested in a dark environment. The trap-state density (n_t) can be calculated using the following equation:

n_t = (V_TFLεε₀)/(eL²)

where e is the electron charge, L is the thickness of the perovskite film and ε₀ is the vacuum permittivity. The trap-filled limit voltage (V_TFL) is obtained from Fig. 3c and the relative dielectric constant (ε) of MAPbI₃ is 32.⁴² Finally, the trap-state densities of the MAPbI₃ film without and with 0.05 mg mL⁻¹ CQDs are calculated as 5.12 × 10¹⁵ and 2.95 × 10¹⁵ cm⁻³, respectively, which are in good agreement with the trap-state density reported for the perovskite polycrystalline film (10¹⁵–10¹⁶ cm⁻³).^42,43

Fig. 3 (a) PL and (b) TRPL curves of the pristine MAPbI₃ film and the MAPbI₃ film with 0.05 mg mL⁻¹ CQDs deposited on the quartz substrate; (c) J–V curves of devices with the structure ITO/c-TiO₂/perovskite (300 nm)/PCBM/Ag; (d) XRD patterns of the pristine MAPbI₃ film and the CQD modified MAPbI₃ film after exposure to ambient atmosphere for 300 hours (RH: 50%–60%); (e) XRD patterns of the initial and aged perovskite films (12.4° to 13.4°); (f) UV-vis absorption spectra of the pristine MAPbI₃ film and the MAPbI₃ film after exposing to ambient atmosphere without controlling the humidity for 4 months; the insets show the corresponding photographs of the films.

In addition to the passivation of defects and improvement of photophysical properties, the addition of the CQDs can significantly improve the stability of the MAPbI₃ film. After aging in ambient environment with a relative humidity (RH) of 50–60% for 300 hours, the CQD modified MAPbI₃ film (0.05 mg mL⁻¹) exhibits much less degradation compared to the pristine MAPbI₃ film. The obvious difference appears in their XRD patterns (Fig. 3d). The strong peaks related to PbI₂ (12.7°) can be observed in the XRD pattern for the aged pristine MAPbI₃ film, while this signal of PbI₂ cannot be detected in the XRD pattern for the aged MAPbI₃ film with CQDs. Meanwhile, the intensities of the XRD peaks related to the MAPbI₃ perovskite (14.1°, 28.4°, 31.9° and 40.7°) are obviously reduced for the aged pristine MAPbI₃ film, revealing that many MAPbI₃ crystals are decomposed into PbI₂. Fig. 3e shows the XRD patterns of the initial and aged perovskite films with the 2θ range from 12.4° to 13.4°. In contrast to PbI₂ significantly increasing in the pristine MAPbI₃ film, the XRD curves of the initial and aged MAPbI₃ films with 0.05 mg mL⁻¹ CQDs show almost no difference. The UV-vis absorption spectra also indicate that the MAPbI₃ film with CQDs shows superior ambient stability (Fig. 3f). Even when the sample is left under ambient atmosphere without controlling the humidity for 4 months, the MAPbI₃ film with 0.05 mg mL⁻¹ CQDs still retains its original black color and its UV-vis spectra show no obvious decrease in the intensity.

To explore the reasons for the passivation and improved stability of the CQD modified perovskite film, the interaction between the CQDs and MAPbI₃ is detected by FTIR measurement. The FTIR spectra of CQDs, MAPbI₃, and CQD modified MAPbI₃ films are shown in Fig. 4a and b. The perovskite films for the FTIR test are deposited on the KBr substrate and the concentration of CQDs is increased to 10 mg mL⁻¹ to enhance the signal. The characteristic peak of the C=O bond stretching vibrations in CQDs locates at 1713 cm⁻¹, while this characteristic peak shifts to 1705 cm⁻¹ in the CQD modified perovskite film. The shift of this peak indicates a strong interaction between CQDs and MAPbI₃, which should be caused by the lone pair electrons on the carboxylic group of CQDs delocalized to the empty orbit of Pb²⁺ in MAPbI₃.⁴¹ In addition, the interaction between CQDs and MAPbI₃ is also calculated by density functional theory (DFT). The optimized structures of CQDs and the MAPbI₃ perovskite simulated with Gaussian 09 are shown in Fig. 4c. Because the ammonium groups co-exist in MAPbI₃ and CQDs, it is difficult to distinguish the FTIR absorption features of the ammonium groups of MAPbI₃ and CQDs. Hence, the interactions between uncoordinated Pb atoms and the amino groups of CQDs are also taken into consideration in DFT calculations. To study the effect of CQDs on MAPbI₃, in the initial structure, the CQDs are put near the MAPbI₃ surface, the distances between Pb₁ of MAPbI₃ and O₁ (O of the carboxyl group) of CQDs (Fig. 4d), Pb₃ of the MAPbI₃ and O₂ (O of the hydroxyl group) of CQDs (Fig. 4f), Pb₂ of the MAPbI₃ and N₂ (N of the amino-group) of CQDs (Fig. 4h) and H₁ of the MAPbI₃ and N₁ (N of the amino-group) of CQDs (Fig. 4j) are set to 3.3, 3.0, 2.9 and 2.1 Å, respectively. During the optimization processes, the CQDs are found to move closer to the perovskite surface (Fig. 4c); eventually, the distances between Pb₁ and O₁, Pb₃ and O₂, Pb₂ and N₂, and H₁ and N₁ are found to be 2.5, 2.6, 2.7 and 1.8 Å, respectively. According to the sum of van der Waals radii of ca. 3.05 Å of Pb and oxygen atoms,⁴⁴ the distances of 2.5 and 2.6 Å are deemed to be close enough to form a stable Pb–O bond (Fig. 4d). The binding energies of Pb₁ and O₁, Pb₃ and O₂, Pb₂ and N₂, and H₁ and N₁ are calculated to be −2.04, −0.90, −1.16 and −1.45 eV, respectively. The first adsorption model (Pb₁–O₁) is the most stable adsorption pattern owing to its lowest binding energy among these four models. The negative adsorption energy indicates that the adsorption process is exothermic, i.e. the adsorption of the organic molecules on the perovskite surface is beneficial for the system stability. Because the uncoordinated Pb atoms always act as electronic trap states,⁴⁵ the CQDs bonding with uncoordinated Pb atoms can passivate the trap states. Furthermore, according to Edoardo Mosconi's simulations,^44,46 MAPbI₃ undergoes a rapid solvation process in a liquid water environment; therefore, when multiple CQD molecules are attached on the MAPbI₃ surface, they would form a barrier layer and that helps to block the contact between water and MAPbI₃, which makes MAPbI₃ more stable in the moisture environment.

Fig. 4 (a, b) FTIR spectra of CQDs, MAPbI₃ and CQD modified MAPbI₃ films; (c) the structures of CQDs and MAPbI₃ optimized with Gaussian 09; initial and optimized structures of CQDs on MAPbI₃ based on the (d, e) Pb₁ and O₁ model, (f, g) Pb₃ and O₂ model, (h, i) Pb₂ and N₂ model, and (j, k) H₁ and N₁ model.

Due to the passivation and promoted optoelectronic properties of the CQD modified MAPbI₃ film, the PSCs based on appropriate CQD modified MAPbI₃ films exhibit better photovoltaic performance. The typical J–V curves of the PSCs based on the MAPbI₃ film with different concentrations of CQDs are presented in Fig. 5a, and the detailed photovoltaic parameters are listed in Table 1. The photovoltaic performance distribution of 15 individual PSCs based on the pristine MAPbI₃ film and 0.02, 0.05, and 0.1 mg mL⁻¹ CQD modified MAPbI₃ films are shown in Fig. S4.† The PSCs made of pristine MAPbI₃ films show a PCE of 17.59% for reverse scan with a short-circuit current density (J_sc) of 21.88 mA cm⁻², an open-circuit voltage (V_oc) of 1.103 V and a fill factor (FF) of 72.90% (Table 1), while the PSCs based on CQD (0.02 mg mL⁻¹) modified MAPbI₃ films show an increased PCE of 18.22%. When the concentration of CQDs increases to 0.05 mg mL⁻¹, the typical PCE is boosted to 18.81% with a J_sc of 22.19 mA cm⁻², a V_oc of 1.134 and an FF of 74.75% (reverse scan). The improved photovoltaic performance of devices (0.05 mg mL⁻¹) can be attributed to the sufficient passivation of the perovskite. Finally, the PSCs with 0.05 mg mL⁻¹ CQDs achieve a highest PCE of 19.38% (J_sc = 22.30 mA cm⁻², V_oc = 1.142 V and FF = 76.1%) as shown in Fig. 5b. However, on further increasing the concentration of CQDs to 0.1 mg mL⁻¹, the device performance is obviously deteriorated, which can be ascribed to the bad morphology and crystallization of the perovskite film. Compared to the referenced device (16.31%), the typical forward scan PCE of the devices with 0.02 and 0.05 mg mL⁻¹ CQDs increases to 16.99% and 17.66%, respectively, and then decreases to 15.21% for devices with 0.1 mg mL⁻¹ CQD additive. The hysteresis for the 0.05 mg mL⁻¹ CQD modified devices is the smallest among the devices. Because the defects in the perovskite layer could act as traps for charges and aggravate hysteresis,⁴⁷ the reduced trap density of the 0.05 mg mL⁻¹ CQD modified perovskite layer could be the reason for the decrease of hysteresis. Fig. 5c shows the external quantum efficiency (EQE) curves and integrated current density (J_int) from the EQE test. The J_int of PSCs without and with 0.02, 0.05 and 0.1 mg mL⁻¹ CQDs are 20.89, 21.14, 21.82 and 20.79 mA cm⁻², respectively, which are in good agreement with the J_sc from the J–V test (Fig. 5a). In addition, due to the lower trap state and more efficient charge extraction of the CQD passivated perovskite film, the photoresponse of 0.05 mg mL⁻¹ CQD modified PSCs is much faster than that of the PSCs without CQDs, according to the steady-state output current density at the maximum power point shown in Fig. 5d. However, the PSCs with excess CQDs show worse photoresponse in that the excess CQDs increase the device series resistance, and thus hinder the charge extraction.

Fig. 5 (a) Forward and reverse J–V curves of PSCs without and with 0.02, 0.05 and 0.1 mg mL⁻¹ CQDs under AM 1.5G one sun (100 mW cm⁻²) illumination; (b) J–V curve of the champion device with 0.05 mg mL⁻¹ CQDs; (c) EQE curves and J_int of PSCs without and with CQDs; (d) steady-state output current density at the maximum power point of PSCs without and with CQDs.

Table 1 Detailed photovoltaic parameters of typical PSCs based on the MAPbI₃ film and 0.02, 0.05, and 0.1 mg mL⁻¹ CQD modified MAPbI₃ films

	Scan	J sc (mA cm^−2)	V oc (V)	FF (%)	PCE (%)
W/O	Forward	22.00	1.082	68.51	16.31
	Reverse	21.88	1.103	72.90	17.59
0.02 mg mL^−1 CQDs	Forward	22.08	1.095	70.26	16.99
	Reverse	22.04	1.128	73.29	18.22
0.05 mg mL^−1 CQDs	Forward	22.01	1.125	71.31	17.66
	Reverse	22.19	1.134	74.75	18.81
0.1 mg mL^−1 CQDs	Forward	21.88	1.076	64.55	15.21
	Reverse	21.92	1.121	67.25	16.52

---

Finally, the long-term stability of PSCs based on the pristine MAPbI₃ film and the 0.05 mg mL⁻¹ CQD modified MAPbI₃ film is investigated, and the PCE of samples stored in a N₂-filled glove box and under ambient atmosphere (RH: 50%–60%) are determined. Fig. 6a and b exhibit the normalized PCE of the samples. After aging in the glove box for 1500 hours, the PCE of the PSCs based on the pristine MAPbI₃ film retains 74.1% of the original value, while the PCE of PSCs fabricated with the 0.05 mg mL⁻¹ CQD modified MAPbI₃ film retains 90.1% of the original PCE. For samples stored under ambient atmosphere with a RH of 50–60% for 216 hours, the CQD modified PSCs are also more stable than the PSCs without CQDs. Since the hydrophilic salt in spiro-OMeTAD exposed to wet air would accelerate the destruction of the perovskite film, the PCE of PSCs without CQDs only retains 17.0% of the original PCE, while the CQD modified PSCs still retain 70.5% of the original PCE. Fig. 6b insets show the photographs of the corresponding devices before and after aging for 216 hours under ambient atmosphere. The yellowish PbI₂ can be easily identified for the device based on the MAPbI₃ film without CQDs, while there is almost no change for the device based on the CQD modified perovskite film.

Fig. 6 Normalized PCE of unencapsulated PSCs based on the pure MAPbI₃ film and the 0.05 mg mL⁻¹ CQD modified MAPbI₃ film stored in (a) a N₂-filled glove box and (b) under air atmosphere (RH: 50%–60%).

Conclusion

In summary, efficient and stable PSCs are demonstrated by passivating the grain boundaries of CH₃NH₃PbI₃via bonding with CQDs. The typical PCE of PSCs based on the CQD modified MAPbI₃ film increases from 17.59% to 18.81% and the PCE of the optimized champion PSCs reaches 19.38%. An appropriate concentration of CQDs in the MAPbI₃ film shows negligible influence on the MAPbI₃ crystal structure but can passivate the grain boundaries and decrease the trap-state density of the MAPbI₃ film. The passivation of CQDs should be attributed to the carboxylic groups, hydroxyl groups and amino-groups of CQDs bonding with the uncoordinated Pb ions in the perovskite crystal, which is proved by the FTIR and DFT results. Moreover, the bonding between CQDs and MAPbI₃ leads to a stable absorption of CQDs on the MAPbI₃ surface, forming a protective layer to prevent the perovskite from coming in contact with water, thereby enhancing the stability of PSCs. Our results indicate that the bonding between the additive and the perovskite crystal is crucial to achieve stable passivation for efficient and stable PSCs.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51573042, 51873007, 21835006, and 51602102), and the Fundamental Research Funds for the Central Universities in China (2018MS032, 2016MS50, 2017MS027, 2018ZD07, and 2017XS084).

References

1. Best research-cell efficiencies, NREL. https://www.nrel.gov/pv/assets/images/efficiency-chart.png.
2. H. S. Kim, C. R. Lee, J. H. Im, K. B. Lee, T. Moehl, A. Marchioro, S. J. Moon, R. Humphry-Baker, J. H. Yum, J. E. Moser, M. Grätzel and N. G. Park, Sci. Rep., 2012, 2, 591.
3. J. H. Im, I. H. Jang, N. Pellet, M. Gratzel and N. G. Park, Nat. Nanotechnol., 2014, 9, 927–932.
4. W. Nie, 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.
5. C. Li, Q. Guo, W. Qiao, Q. Chen, S. Ma, X. Pan, F. Wang, J. Yao, C. Zhang, M. Xiao, S. Dai and Z. A. Tan, Org. Electron., 2016, 33, 194–200.
6. Q. Chen, H. Zhou, Z. Hong, S. Luo, H. S. Duan, H. H. Wang, Y. Liu, G. Li and Y. Yang, J. Am. Chem. Soc., 2014, 136, 622–625.
7. Q. Guo, C. Li, W. Qiao, S. Ma, F. Wang, B. Zhang, L. Hu, S. Dai and Z. A. Tan, Energy Environ. Sci., 2016, 9, 1486–1494.
8. C. Li, F. Wang, J. Xu, J. Yao, B. Zhang, C. Zhang, M. Xiao, S. Dai, Y. Li and Z. A. Tan, Nanoscale, 2015, 7, 9771–9778.
9. N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. I. Seok, Nat. Mater., 2014, 13, 897–903.
10. N. Ahn, D. Y. Son, I. H. Jang, S. M. Kang, M. Choi and N. G. Park, J. Am. Chem. Soc., 2015, 137, 8696–8699.
11. M. Xiao, F. Huang, W. Huang, Y. Dkhissi, Y. Zhu, J. Etheridge, A. Gray-Weale, U. Bach, Y.-B. Cheng and L. Spiccia, Angew. Chem., Int. Ed., 2014, 126, 10056–10061.
12. J. M. Azpiroz, E. Mosconi, J. Bisquert and F. De Angelis, Energy Environ. Sci., 2015, 8, 2118–2127.
13. C. Roldan-Carmona, P. Gratia, I. Zimmermann, G. Grancini, P. Gao, M. Graetzel and M. K. Nazeeruddin, Energy Environ. Sci., 2015, 8, 3550–3556.
14. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J. P. Correa Baena, J. D. Decoppet, S. M. Zakeeruddin, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, Sci. Adv., 2016, 2, 1501170.
15. D. Y. Son, J. W. Lee, Y. J. Choi, I. H. Jang, S. Lee, P. J. Yoo, H. Shin, N. Ahn, M. Choi, D. Kim and N. G. Park, Nat. Energy, 2016, 1, 16081.
16. M. Yang, Y. Zhou, Y. Zeng, C. S. Jiang, N. P. Padture and K. Zhu, Adv. Mater., 2015, 27, 6363–6370.
17. Q. Jiang, Z. Chu, P. Wang, X. Yang, H. Liu, Y. Wang, Z. Yin, J. Wu, X. Zhang and J. You, Adv. Mater., 2017, 29, 1703852.
18. Q. Chen, H. Zhou, T. B. Song, S. Luo, Z. Hong, H.-S. Duan, L. Dou, Y. Liu and Y. Yang, Nano Lett., 2014, 14, 4158–4163.
19. Y. Shao, Z. Xiao, C. Bi, Y. Yuan and J. Huang, Nat. Commun., 2014, 5, 5784.
20. H. Tan, A. Jain, O. Voznyy, X. Lan, F. P. García de Arquer, J. Z. Fan, R. Quintero-Bermudez, M. Yuan, B. Zhang, Y. Zhao, F. Fan, P. Li, L. N. Quan, Y. Zhao, Z.-H. Lu, Z. Yang, S. Hoogland and E. H. Sargent, Science, 2017, 355, 722–726.
21. N. K. Noel, A. Abate, S. D. Stranks, E. S. Parrott, V. M. Burlakov, A. Goriely and H. J. Snaith, ACS Nano, 2014, 8, 9815–9821.
22. C.-C. Zhang, M. Li, Z.-K. Wang, Y.-R. Jiang, H.-R. Liu, Y.-G. Yang, X.-Y. Gao and H. Ma, J. Mater. Chem. A, 2017, 5, 2572–2579.
23. J. Jiang, Q. Wang, Z. Jin, X. Zhang, J. Lei, H. Bin, Z.-G. Zhang, Y. Li and S. F. Liu, Adv. Energy Mater., 2018, 8, 1701757.
24. T. Niu, J. Lu, R. Munir, J. Li, D. Barrit, X. Zhang, H. Hu, Z. Yang, A. Amassian, K. Zhao and S. F. Liu, Adv. Mater., 2018, 30, 1706576.
25. F. Yuan, Z. Wang, X. Li, Y. Li, Z. A. Tan, L. Fan and S. Yang, Adv. Mater., 2017, 29, 1604436.
26. F. Yuan, T. Yuan, L. Sui, Z. Wang, Z. Xi, Y. Li, X. Li, L. Fan, Z. A. Tan, A. Chen, M. Jin and S. Yang, Nat. Commun., 2018, 9, 2249.
27. F. Yuan, L. Ding, Y. Li, X. Li, L. Fan, S. Zhou, D. Fang and S. Yang, Nanoscale, 2015, 7, 11727–11733.
28. F. Yuan, S. Li, Z. Fan, X. Meng, L. Fan and S. Yang, Nano Today, 2016, 11, 565–586.
29. H. Li, W. Shi, W. Huang, E. P. Yao, J. Han, Z. Chen, S. Liu, Y. Shen, M. Wang and Y. Yang, Nano Lett., 2017, 17, 2328–2335.
30. Z. Zhu, J. Ma, Z. Wang, C. Mu, Z. Fan, L. Du, Y. Bai, L. Fan, H. Yan and D. L. Phillips, J. Am. Chem. Soc., 2014, 136, 3760–3763.
31. S. Zhou, R. Tang and L. Yin, Adv. Mater., 2017, 29, 1703682.
32. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
33. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758.
34. L. Tanghao, H. Qin, W. Jiang, C. Ke, Z. Lichen, L. Feng, W. Cheng, L. Hong, J. Shuang, R. Thomas, Z. Rui and G. Qihuang, Adv. Energy Mater., 2016, 6, 1501890.
35. S. R. Raga, M. C. Jung, M. V. Lee, M. R. Leyden, Y. Kato and Y. Qi, Chem. Mater., 2015, 27, 1597–1603.
36. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643–647.
37. C. Li, Q. Guo, H. Zhang, Y. Bai, F. Wang, L. Liu, T. Hayat, A. Alsaedi and Z. A. Tan, Nano Energy, 2017, 40, 248–257.
38. Z. Wang, T. Cheng, F. Wang, S. Dai and Z. A. Tan, Small, 2016, 12, 4412–4420.
39. V. T. Tiong, N. D. Pham, T. Wang, T. Zhu, X. Zhao, Y. Zhang, Q. Shen, J. Bell, L. Hu, S. Dai and H. Wang, Adv. Funct. Mater., 2018, 28, 1705545.
40. D. Bi, C. Yi, J. Luo, J.-D. Décoppet, F. Zhang, S. M. Zakeeruddin, X. Li, A. Hagfeldt and M. Grätzel, Nat. Energy, 2016, 1, 16142.
41. L. Zuo, H. Guo, D. W. deQuilettes, S. Jariwala, N. De Marco, S. Dong, R. DeBlock, D. S. Ginger, B. Dunn, M. Wang and Y. Yang, Sci. Adv., 2017, 3, 1700106.
42. Q. Dong, Y. Fang, Y. Shao, P. Mulligan, J. Qiu, L. Cao and J. Huang, Science, 2015, 347, 967–970.
43. T. Bu, X. Liu, Y. Zhou, J. Yi, X. Huang, L. Luo, J. Xiao, Z. Ku, Y. Peng, F. Huang, Y. B. Cheng and J. Zhong, Energy Environ. Sci., 2017, 10, 2509–2515.
44. J. M. Grevy, F. Tellez, S. Bernés, H. Nöth, R. Contreras and N. Barba-Behrens, Inorg. Chim. Acta, 2002, 339, 532–542.
45. W. J. Yin, T. Shi and Y. Yan, Appl. Phys. Lett., 2014, 104, 063903.
46. E. Mosconi, J. M. Azpiroz and F. De Angelis, Chem. Mater., 2015, 27, 4885–4892.
47. H. J. Snaith, A. Abate, J. M. Ball, G. E. Eperon, T. Leijtens, N. K. Noel, S. D. Stranks, J. T. Wang, K. Wojciechowski and W. Zhang, J. Phys. Chem. Lett., 2014, 5, 1511–1515.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8nr08295b

‡ These authors contributed equally to this work.

---