1. Introduction

In the past ten years, organic-inorganic hybrid perovskite solar cells (PSC) have shown tremendous enhancements in photovoltaic performance, with the highest certified power conversion efficiency (PCE) of more than 26% [1–6]. However, all inorganic cesium-lead halide PSC with the chemical formula CsPbX₃ (X = I, Br or mixed) have attracted more and more research interest because of their outstanding thermal stability compared with organic-inorganic hybrid batteries. Among them, CsPbIBr₂ perovskite balances the trade-off between stability and efficiency, and is considered a promising representative. At present, researchers have made great efforts to improve the power conversion efficiencies (PCEs) of CsPbIBr₂ PSCs [7–14]. The PCE has reached more than 11%, but it still lags behind the analogues of the same type of high iodide (over 21% for CsPbI₃ and over 17% for CsPbI₂Br) [15,16].

However, compared with organic-inorganic hybrid perovskites, all inorganic perovskites still face many challenges. Among the two most important key problems, one is that the light absorption range is limited by the wide bandgap of the material, and the other is the serious energy loss caused by the poor film quality [17–20]. It is worth mentioning that the hole-free carbon-based inorganic PSCs (C-IPSCs) solves the problem of material hygroscopicity due to the absence of organic components, alleviates the decomposition rate of the device. In addition, carbon materials have hydrophobic and environmentally friendly properties, as hole transport layer and back electrodes, the manufacturing process is simplified and costs are reduced.

For the former, the preparation of bulk-heterojunction (BHJ) layer on perovskites is the current mainstream solution. Our previous work, by dividing poly (3-hexylthiophene-2,5-diyl) and [6,6]-phenyl methyl C61 butyric acid methyl ester (P3HT:PCBM) into thin BHJ layer, constructed a perovskite/BHJ system, successfully expanded the light absorption range, achieved better charge transfer dynamics, suppressed interfacial energy loss and improved long-term stability [21]. Secondly, due to the soft ion crystal nature of perovskite, defects cannot be avoided during the preparation and operation of the film. The interfacial passivation strategy can effectively optimize the grain boundary and passivate the defects on the surface of the perovskite to improve the quality of the film. The methods have been shown to minimize non-radiation composite losses and improve energy loss [22–26]. Yu and colleagues introduced organic dye molecules (VG1-C8) into the CsPbIBr₂ film as a functional layer to passivate the trap state on the perovskite film, and the efficiency of the device reached 12.1%, which was surprisingly improved [27]. Ding and colleagues used polylactic acid (PLA) to passivate the surface of the perovskite and induce grain growth, thereby obtaining a series of inorganic devices with excellent performance [28]. Liu et al. introduced 4-aminodiphenylamine (4-ADPA) interlayer to modify the surface and grain boundary of the perovskite film. Thereby improving the quality of the film, reducing defects, and inhibiting carrier recombination [29]. However, significantly improving the photoelectric performance of CsPbIBr₂ PSC is still a challenge until now.

In this study, benzyltrimethylammonium chloride (BTACl) was introduced as a passivation layer to modify the interface between the perovskite and the carbon layer. The quaternary amine cations and chloride ions in BTACl can synergistically passivate the positive and negative anion defects in the perovskite, and the presence of a hydrophobic benzene ring leads to the surface of the treated perovskite having excellent hydrophobic properties and improving the stability in the air. More importantly, BTACl acts as an electric dipole molecule to form a built-in electric field, which improves the extraction and collection of holes between CsPbIBr₂ and the carbon electrode. In the end, the device structure of FTO/SnO₂/BTACl-CsPbIBr₂/carbon achieves a PCE of 11.30% and unencapsulated devices exhibit better stability in ambient air. This work provides a simple and feasible strategy for efficient and stable C-IPSC.

2. Results and discussion

The device configuration and molecular structure of BTACl used in this article are display in Fig. 1(a). The passivation layer is spin-coated on the CsPbIBr₂ with different concentrations of BTACl/isopropanol solution. According to the defect category of perovskite, a hypothetical model of the passivation mechanism of CsPbIBr₂ is demonstrated (Fig. 1(b)). Unlike organic-inorganic hybrid perovskites, BTACl containing benzene rings and quaternary ammonium halides (NR₄⁺Cl⁻) cannot form a 2D/1D phase on top of the inorganic perovskite membrane through solid or sequential cation exchange. Therefore, we infer that BTACl will stay on the surface of CsPbIBr₂ to passivate the defect [30–32]. More importantly, quaternary amine cations (NR₄⁺) and chloride ions(Cl⁻) can passivate positive and negative anion defects, similar to the synergistic passivation effect of the amino and carboxyl groups of phenylalanine (PAA) on perovskite films, a small amount of Cl^- addition has been widely reported to improve the charge compound life in perovskites [33–35]. We used X-ray photoelectronic spectroscopy (XPS) measurements to prove the presence of BTACl on the perovskite (Fig. S1a-c).The appearance of N and Cl verifies the existence of the passivation layer. Moreover, by analyzing the C1s core-level energy spectrum, in addition to the C–C (284.6 eV) and C–N (286.08 eV) peaks, also found that a peak appeared at the binding energy of 292.7 eV, which is consistent with the π-π bond in the phenyl functional group of the BTA⁺ cation (Fig. 1(c)) [36–38]. More consequentially, the binding energies of Cs 3d, Pb 4f, I 3d and Br 3d increased respectively, indicating that a local charge transfer occurred between the PbX₆^4- frame and the passivation layer (Fig. S2a-d) [39]. Fourier transform infrared spectroscopy (FTIR) was used to further prove the existence of the passivation layer (Fig. S3). the characteristic peaks of the alkenyl C = C stretching vibrations from the aromatic ring of BTA⁺ appears in the range of 1500 to 1400 cm⁻¹ and the characteristic peak that appears at about 1120 cm⁻¹, which corresponds to the C-N stretching vibration peak of the quaternary ammonium group [40].

 

Fig. 1. (a) The device structure and chemical structures of BTACl used for passivation layer, (b) Possible passivation mechanism of the BTACl layer for the perovskite film, (c) C 1s XPS spectra of pristine and passivated CsPbIBr₂ films, (d) XRD patterns and (e) magnified spectra of (200) plane peak of pristine and different passivation concentrations.

Download Full Size | PDF

The influence of the passivation layer on the crystal structure of CsPbIBr₂ was studied by X-ray diffraction (XRD) method. All samples have two main diffraction peaks in the same position (Fig. 1(d)). Since BTA⁺ has a large ion radius and does not exchange with Cs⁺, there is no small-angle diffraction peak, indicating that no 2D covering layer on the surface. Figure 1(e) shows the magnified spectrum of (200) planar peaks. The XRD peak moves to lower angle after passivation, which indicates the expansion of the CsPbIBr₂ lattice [41]. The full width at half maximum (FWHM) of the main peak of XRD (Table S1) shows that the film passivated with 1 mg mL⁻¹ BTACl solution has reduced value, indicating that the crystallinity and grain size increases. When the passivation concentration continues to increase, the XRD peak of the perovskite film moves to a lower position and shows a greater FWHM value, which may be due to a large amount of BTACl reduces the crystallinity of the perovskite. The surface morphology of perovskite was studied by scanning electron microscopy (SEM) (Fig. S4). Pristine films have some pinholes and obvious grain boundaries, while there are no pinholes in the passivation film and the grains are more larger. This may be due to the passivation layer is filled with the pinholes/grain boundaries of the CsPbIBr₂ film, which matches the analysis results of XRD.

In order to study the dipole arrangement and charge dynamics at the perovskite interface. The work function, valence band and conduction band of film were obtained by UPS (Fig. 2(a)) and band gap of film was obtained from Fig. S5. In addition, the UV-vis absorption spectra (Fig. S5) display the passivated sample showed enhanced absorption than the pristine film, indicating better membrane quality, consistent with the XRD and SEM results. Combined with the relevant parameters, it can be calculated that the Fermi energy levels (E_F) of the pristine and 1 mg/ml BTACl passivation CsPbIBr₂ films are -3.68 and -3.54 eV, respectively. The valance band maximum (VBM) is -5.72 and -5.45 eV, respectively, and the CBM is -3.64 and -3.37 eV, respectively. The corresponding energy band diagram of the battery is shown in Fig. 2(b). Therefore, the energy offset (Δµ) between the VBM of the BTACl passivated CsPbIBr₂ and the WF (-5.0 eV) of the carbon electrode is reduced from 0.72 to 0.45 eV. The reduced energy level mismatch is conducive to optimizing the performance of the device. In addition, due to the difference in the work function between CsPbIBr₂ or BTACl passivated CsPbIBr₂ and the carbon electrode material, a Schottky barrier is formed at the interface. The values of upward band bending (ΔΦ) are calculated as 1.32 eV (pristine) and 1.46 eV (1 mg/ml), respectively. The higher ΔΦ in the passivation film not only prevents the back transfer of electrons between the interfaces, but also accelerates the hole extraction rate at the interface [42].

 

Fig. 2. (a) UPS curves of pristine and passivated CsPbIBr₂ films, (b) Energy level diagram, (c) Electro-static potential (ESP) map of BTA⁺ cation, (d) Schematic diagram of the BTACl arrangement on the CsPbIBr₂ lattice, (e) Schematic diagram of interface dipole arrangement and energy band bending in CsPbIBr₂/BTACl/Carbon structure.

Download Full Size | PDF

We speculate that the upward bending of the valence band and the change in vacuum energy level are due to the formation of interfacial dipoles, resulting in an increase in work function and valence band, which is due to the presence of quaternary ammonium salt groups and conjugated benzene rings on the BTACl [43–46]. The electrostatic potential distribution of monomers containing quaternary ammonium salt groups in BTACl is calculated by density functional theory (DFT) to further verify this hypothesis (Fig. 2(c)). The dipole moment formed by the center of positive and negative charge is 5.97 Debye. We know that the interaction between the relevant molecular groups and the surface of the perovskite determines the direction of the surface dipole. According to theory, the quaternary ammonium salt group is located on the surface of the perovskite, the benzene ring and carbon chain are distributed in the outer layer and the schematic diagram of the BTACl arrangement on the lattice is shown in the Fig. 2(d). This arrangement increases the vacuum energy level and bends the valence band upward (Fig. 2(e)). Since E_VB of CsPbIBr₂ is lower than the E_F of the carbon electrode, contact will lead to serious interfacial charge recombination, resulting in a low filling factor (FF) and open circuit voltage (V_oc). When we introduce BTACl electric dipole passivation layer between the two functional layers, will reduce the hole transmission barrier and increase the electron transmission barrier, thereby promoting hole collection and electron blocking [47].

In order to study the changes in charge carrier dynamics of perovskite films, steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) attenuation measurements were used to evaluate the effects of BTACl on charge recombination and exciton life in CsPbIBr₂ films. As shown in Fig. 3(a), the passivated film prepared on the glass substrate produces a higher PL strength, indicating that the non-radiation composite in the CsPbIBr₂ film is suppressed. The TRPL test results are displayed in Fig. 3(b). The carrier life is determined by fitting the dynamics with a two-exponential function (Table S2). The τ_ave of the 1 mg/ml-BTACl/CsPbIBr₂ film is 5.86 ns longer than the pristine film (1.06 ns), indicating that passivated film has a lower trap density, which is conducive to the charge transfer and performance of PSCs. For further explore the influence of BTACl passivation on the defect density of the device, a pure electronic device (ITO/SnO₂/CsPbIBr₂ or BTACl-CsPbIBr₂/Carbon) was made. The passivated device has a more smaller trap filling limit voltage (V_TFL) than pristine one (0.79 V vs 1.17 V) (Fig. 3(c)), and trap state density (${N_t} = \frac{{2{\mathrm{\varepsilon }_0}{\mathrm{\varepsilon }_r}{V_{TFL}}}}{{q{L^2}}}$) can be calculated, [48] The N_t values of the pristine and passivated device is 8.86 × 10⁻¹⁵cm⁻³ and 5.98 × 10⁻¹⁶cm⁻³, respectively, demonstrating that BTACl passivated can significantly reduce the trap density.

 

Fig. 3. (a) PL and (b) TRPL of CsPbIBr₂ perovskite films with or without BTACl, (c) The space-charge limited current (SCLC) measurements of models with ITO/SnO₂/CsPbIBr₂ with or without 1 mg/ml BTACl/Carbon.

Download Full Size | PDF

In order to study the passivation effect of BTACl on photovoltaic devices, the J−V curves with different passivation concentrations and relevant parameters are display in Fig. S6 and Table S3. The optimal performance device is obtained when the concentration reaches 1 mg/ml. When the concentration exceeds 1 mg/ml, the decrease in PCE is mainly due to the excessive BTACl reducing the crystallinity of the perovskite (discussed earlier). Obviously, BTACl passivation significantly improved device performance, The J−V curves of the champion PSC with and without passivation of forward and reverse scan are displayed in Fig. 4(a), and the relevant parameters are listed in Table S4. 1 mg/ml-BTACl passivated CsPbIBr₂ PSC with a PCE of 11.30% (V_oc of 1.32 V, J_sc of 11.79 mA cm⁻², FF of 72.62%), which was significantly higher than the pristine (PCE of 8.43%, V_oc of 1.27 V, J_sc of 10.12 mA cm⁻², FF of 65.61%). Moreover, we evaluated the hysteresis index (HI) of the device. 1 mg/ml-BTACl passivated CsPbIBr₂ PSC exhibits a smaller HI (0.038 vs 0.173), and the slight hysteresis is mainly due to excellent carrier transmission and extraction capabilities. Further, the devices [ITO/SnO₂/CsPbIBr₂ or BTACl- CsPbIBr₂/carbon] measured in the dark are shown in Fig. S7. The results show that BTACl passivated CsPbIBr₂ PSCs have a higher V_oc, which is consistent with the results (Fig. 4(a)). Worth noting that we evaluated from the histogram of 20 individual devices (Fig. 4(b)) that the average efficiency of the pristine and passivated devices was 7.89%±0.23% and 11.17%±0.15%, respectively, proving a good reproducibility. For the pristine and 1 mg/ml-BTACl passivated CsPbIBr₂ devices, the stable photocurrent density of PSC was measured at the maximum power points (MPP) of 0.88 and 1.01 V (Fig. 4(c)). The passivated device observed more stable J_sc (10. 81 mA cm⁻²) and PCE (10.91%) within 400 s than pristine. EQE measurements are performed on the both PSCs (Fig. 4(d)). In the entire test area, the 1 mg/ml-BTACl passivated device showed a higher EQE value than pristine. The integrated J_sc value were 9.85 and 11.51 mA cm⁻² respectively, which is consistent with the J-V results.

 

Fig. 4. (a) J−V curves of the champion PSCs with and without BTACl passivation of forward and reverse scan, (b) The statistic distribution of PCEs from 20 identical devices, (c) The steady-state maximum PCE outputs, (d) EQE spectra and the corresponding integrated J_sc.

Download Full Size | PDF

In addition, several characterization techniques were performed, including transient photocurrent (TPC), transient photocurrent (TPV) and J–V measurements related to light intensity to evaluate the carrier dynamics of the device. As shown in the TPC curve (Fig. S8) and the TPV curve (Fig. 5(a)), the photogenic current life of the BTACl passivated device (2.31µs) is shorter than the pristine one (3.24µs), while the photoelectric voltage life of the passivated device (2.91 ms) is higher than the pristine one (1.42 ms). Shorter charge extraction time and longer charge non-radiative composite life are obtained in the passivated device, indicating that BTACl can inhibit charge non-radiative recombination and increase holes extraction efficiency. The carrier composition in the device is further studied by measuring the light intensity dependence of J_sc and V_oc. As shown in Fig. S9, the values of α (${J_{SC}} \propto {I^\alpha })$ is ideal factor associated with bimolecular recombination, the pristine and 1 mg/ml-BTACl passivation device are 1.096 and 0.998, respectively, indicating that passivated device (close to 1) has fewer non-radiated charge complexes, which explains the reason for the high V_oc. Furthermore, Fig. 5(b) shows the relationship between V_oc and light intensity. the ideality factor (n, ${V_{oc}} = n\frac{{kT}}{q}\ln I + c$) is associated with charge recombination mechanism [49]. 1 mg/ml-BTACl passivation device with n value of 1.39 (closer to the ideal diode) lower than pristine one (2.11), which showing that BTACl can effectively reduce trap-assisted recombination in the PSC. These results explain the reasons for the enhancement of V_oc and PCE of the 1 mg/ml-BTACl passivation device.

 

Fig. 5. (a) TPV curves, (b) V_oc depended on light intensity, (c) dark J-V curves, (d) capacitance-voltage and (e) Mott-Schottky plots of devices with pristine and 1 mg/ml-BTACl passivation devices, (f) Long-term stability of unencapsulated devices with and without BTACl.

Download Full Size | PDF

In order to further explore the reasons for the enhanced performance of the passivation device, the dark current measurement was carried out (Fig. 5(c)). According to the diode formula of V = nKTln(J_SC/J₀)/q (J₀ represents the saturation current density), [50] The 1 mg/ml-BTACl passivation CsPbIBr₂ device has n of 1.84 and J₀ of 3.16 × 10⁻⁷mA cm⁻² lower than the pristine (n is 2.41, J₀ is 4.79 × 10⁻⁵mA cm⁻²), which indicate a smaller leakage current in passivated device, also means that the fewer charge loss during the photoelectric conversion process. Figure 5(d) depicts the capacitance-voltage (C−V) curve, which obtained by fixing the AC frequency at 1kHz and changing the bias voltage. Obviously, within the range of bias voltage changes, 1 mg/ml-BTACl passivation PSC shows a smaller capacitance than pristine, indicating that less charge accumulates at the interface. As further evidence, the Mott-Schottky (M-S) curve was test to study the change of built-in potential (V_bi) in the device (Fig. 5(d)). The results show that the V_bi of 1 mg/ml-BTACl passivation PSC increases from 1.26 V to 1.36 V, which further indicates the formation of interface dipoles. At the same time, the M-S curve trend is consistent with the Voc value obtained from the J–V curve. This reveals that the passivated device with an enhanced driving force for dissociating photogenic carrier, and the extended depletion region keeps low accumulated charge at the interface, which directly helps to increase V_oc and suppress J-V hysteresis [21].

Further study the air stability of both device.Unencapsulated 1 mg/ml-BTACl passivation PSCs can retain about 85% of the initial PCE after 750 hours at aging in ambient air (20°C, ∼60% relative humidity) (Fig. 5(f)), while the PCE of the pristine battery drops below 40% of the original value. In addition, other normalized photovoltaic parameters of passivation PSCs (including V_oc, J_sc, and FF) also show excellent stability (Fig. S10). Studies have shown that under the influence of moisture and oxygen, the phase transition and degradation of inorganic perovskites preferentially start from the defect location of the interface [51]. BTACl passivation inactivates uncoordinated lead and halide ions on the surface, improving the stability under environmental conditions. In addition, water contact angle measurements show that the BTACl passivated perovskite surface has a larger contact angle (84.59°) than pristine (62.13°) (Fig. S11). This may be because the biphenyl ring in BTACl increases the hydrophobicity of the interfacial membrane and slows down the degradation of perovskite, thereby improving long-term stability.

3. Conclusion

In short, we proved that the BTACl as the passivation layer can effectively improve the photovoltaic performance of C-IPSCs. A series of characterizations have verified that the quaternary amine cations and chloride ions in BTACl can synergistically passivate the surface and grain boundary defects in the perovskite, and the hydrophobic benzene ring makes the perovskite surface have excellent hydrophobic properties to improve stability in the air. Moreover, BTACl forms a built-in electric field as an electric dipole molecule, which enhances the transfer of charge. The champion device reached a PCE of 11.30% at the optimal BTACl concentration under above optimization, higher than the pristine one (8.43%). At the same time, the unencapsulated device retains about 85% of the initial PCE in the ambient air (20°C, ∼60% relative humidity). This simple strategy provides a promising way to manufacture efficient and stable C-IPSC.