Reduced Defects and Enhanced Performance of (FAPbI3)0.97(MAPbBr3)0.03-Based Perovskite Solar Cells by Trimesic Acid Additives.

A high-quality organolead trihalide perovskite film with large-sized crystalline grains and smooth surfaces is required to obtain efficient perovskite solar cells (PSCs). Herein, high-quality (FAPbI3)0.97(MAPbBr3)0.03 perovskite films were fabricated using trimesic acid (TMA) additives in a halide perovskite precursor solution to obtain efficient PSCs. The X-ray diffraction analysis and scanning electron microscopy of the films revealed that the TMA had a significant effect on the roughness of the films by acting as a surface link, thus reducing the surface defects and recombination at the grain boundaries. In addition, with the addition of the TMA additive, a smooth perovskite film with a flat surface and no pinholes was obtained. The perovskite film was used to fabricate a PSC device, and the device exhibited a high power conversion efficiency of 17.26%, which was higher than that of the control device (15.15%) under the same conditions. This study demonstrates a facile method to passivate defects on the perovskite layer via surface modification.

Introduction

Recently, organic–inorganic hybrid lead perovskite solar cells (PSCs) have attracted tremendous attention due to their splendid optoelectronic properties and their facile and cheap fabrication process.[1−5] FAPbI3 perovskite is considered an ideal candidate for the fabrication of efficient and stable PSCs owing to its thermal stability and a low band gap (1.45 eV), which enables the broader absorption of ions of the solar spectrum.[6−8] FAPbI3 consists of two phases: pure-phase black α-FAPbI3 and photoinactive yellow δ-FAPbI3, which is stable at room temperature. The yellow δ-FAPbI3 reduces the crystallinity of FAPbI3 films, thus hindering electron transport and reducing the efficiency of PSCs. The quality of perovskite films, which can be influenced by the preparation method, makes a significant contribution to the device capacity of PSCs.[9−11] However, owing to the difficulties in controlling these perovskite films including their organic components, rough and inhomogeneous films, which are sensitive to air, are formed.[12−14] The modification of the perovskite surface by enhancing the crystallinity, suppressing the surface defects, and reducing the number of pinholes has emerged as a powerful method to improve the film quality of perovskites. In the previous research, our group modified the perovskite surface by inserting an ultrathin NiO@C interfacial layer on the top of the perovskite layer, achieving a high efficiency of 15.78% in the planar n–i–p structure.[15,16]

Tremendous attempts have been devoted to modifying the specific characteristics of perovskite films by developing new preparation processes, such as one- or two-step solution methods, additive engineering, and antisolvent treatment, to obtain efficient and stable PSCs.[1,17−20] For example, halide perovskite films have been widely synthesized by solution-based bottom-up methods for precisely controlling the morphology of the films with large surface areas and high crystalline grain boundaries.[12,21−23] However, these unstable large-sized grain perovskite cuboids are highly sensitive to moisture, thus resulting in the quick degradation of the performance of PSCs. To overcome this disadvantage, the additive-based approach has emerged as an effective strategy to passivate surface defects. This approach improves the efficiency and stability of PSCs by improving the morphology and optoelectronic properties of mixed-perovskite devices. Recently, trimesic acid (TMA) has been used in a PbI2 solution to improve the thermal stability and air stability of bulk MAPbI3 perovskite films.[24] Trap-state passivation by nonvolatile small molecules was investigated with carboxylic acid groups by benzoic acid, p-phthalic acid, and trimesic acid to improve the performance of perovskite solar cells to 15.81%.[25] To our knowledge, there were no reports investigating the TMA additive directly in the hygroscopic organic such as methylammonium (MA+) and formanidium (FA+), which could affect the surface of the perovskite layer.

Herein, we employed an optimal amount of TMA to investigate the functions of additives on the surface of perovskite films. We found that the presence of a TMA additive played an important role in controlling the grain size and quality of the perovskite films.[18,26,27] In addition, the TMA additive induced the growth of a uniform perovskite on the perovskite film, thus suppressing surface defects and reducing carrier recombination at the grain boundaries. Furthermore, the optimal TMA concentration for obtaining a high-quality perovskite film was 10 wt %. The high-quality perovskite film was used to fabricate a PSC, and the device achieved a maximum power conversion efficiency (PCE) of 17.26%, which is higher than that of the control PSC 15.15%.

Results and Discussion

To accurately investigate the effect of the TMA additive on the PSCs, a batch of 15 devices was fabricated under various TMA conditions. Herein, mesoscopic PSCs with a structure of fluorine-doped tin oxide (FTO)/cp-TiO2/mp-TiO2/(FAPbI3)0.97(MAPbBr3)0.03/spiro-OMeTAD/Au were designed. The schematic illustration of the architecture of the mesoscopic PSCs is shown in Figure a. The thickness of the absorber thin film was approximately 350 nm. A low band gap FAPbI3 was selected to fabricate the PSCs to increase the light-harvesting potential of the device. The three-dimensional (3D) molecular structure of TMA including the benzene ring and three carboxyl groups is shown in Figure b. TMA was added as an additive to the perovskite layer to investigate its effects on the optical and performance properties of PSCs. Figure c shows the schematic illustration of the addition of TMA during the fabrication of the (FAPbI3)0.97(MAPbBr3)0.03-based PSCs by a two-step deposition method.

Figure 1

(a) Architecture of the mesoporous PSCs. (b) 3D chemical structure of the TMA additive. (c) Schematic illustration of the experimental process for the fabrication of the absorber layer using a two-step deposition process.

To understand the interaction between TMA and the perovskite, Fourier transform infrared spectroscopy (FTIR) was carried out to verify the presence of TMA in the control perovskite film and the 10%-TMA additive perovskite film as a representative sample. As shown in Figure , a strong peak can be observed at 1714 cm–1 in the control and modified perovskite films, which corresponded to the strong stretching vibration of C=N from FA+. In addition, two additional peaks were observed at 3402 and 3260 cm–1, which could be attributed to the stretching vibration of N–H.[28] The FTIR spectrum of the TMA perovskite film is similar to that of the (FAPbI3)0.97(MAPbBr3)0.03 film. No new stretching vibration peak was observed during the reaction, indicating that the carboxylic acid group of the TMA was not inserted into the film but affected the perovskite surface.[29]

Figure 2

FTIR spectra of the pure TMA and TMA-additive perovskite films.

The (FAPbI3)0.97(MAPbBr3)0.03 perovskite thin films were fabricated by a two-step spin-coating process, and the size and shape of the crystal grains were significantly influenced by the TMA concentration. Figure a–f shows scanning electron microscopy (SEM) images of the original perovskite film (as control) and the perovskite films with 5%-, 10%-, 15%-, and 20%-TMA additives. As shown in the SEM image, the crystal grains of the pristine (FAPbI3)0.97(MAPbBr3)0.03 films exhibited clear boundaries with an average grain size of approximately 300 nm. After the addition of 5%- and 10%-TMA additives, the perovskite layer exhibited a uniform and smooth surface with a lower surface roughness, which reduced the recombination sites and improved the charge carrier mobility. However, with a further increase in the TMA concentration to 15% and 20%, a poor-quality perovskite surface with pinholes and large cuboids was observed, suggesting that this TMA additive is ineffective for improving the morphology of the perovskite film.

Figure 3

(a–e) Top-view field emission scanning electron microscopy (FE-SEM) images of the perovskite films modified with different concentrations of TMA: 0%, 5%, 10%, 15%, and 20%. (f) Cross-sectional SEM image of the PSC device.

The UV–visible (UV–vis) absorption spectra of the perovskite films are shown in Figure a,b. As shown in the image, the light absorption intensity of the perovskite films increases slightly with increasing TMA concentration owing to the increase in the size of the perovskite grain. The absorption edge of the perovskite films is almost unchanged with the presence of the TMA additive. The absorption of the modified perovskite films is stronger than the absorption of the pristine perovskite film. However, the absorption of the 20%-TMA perovskite film is lower than that of the 15%-TMA perovskite film, which is likely attributed to the defects in the surface of the film. As shown in Figure c, the steady-state photoluminescence (PL) spectra of the perovskite films with the maximum emission peak of the film at 790 nm were observed at an excitation of 530 nm, which is consistent with the reports of a previous study.[30] The luminescence emission peak depicts radiative recombination and nonradiative decay.[31] A further increase in the peak intensity indicates the reduction of surface defects, which could suppress the recombination at the perovskite/HTL interface.[32−36] This is consistent with the SEM images and UV–vis absorption spectra. However, the perovskite films with 15- and 20%-TMA additives exhibited weak-intensity PL peaks, indicating the poor morphology of the films with several pinholes and smaller perovskite grains, which resulted in the reduction of the PCEs. Figure d shows the X-ray diffraction (XRD) patterns of the films with various TMA contents recorded from 10 to 60°. The 2θ peaks at 13.95, 19.83, 24.45, 28.12, 33.84, 34.81, 40.27, and 42.85° in the XRD patterns of the (FAPbI3)0.97(MAPbBr3)0.03 films confirmed the existence of the FAPbI3 perovskite phase in the fabricated films. However, in the perovskite films with 15%- and 20%-TMA additives, the weak-intensity PL peaks explain the poor morphology with more pinholes and smaller perovskite grains, which results in the reduction in PCEs.[37,38] In addition, the peak observed at 12.63° in the XRD pattern of the pristine FAPbI3 perovskite film without the TMA additive corresponded to the incomplete reaction between PbI2 and formamidinium iodide (FAI) prepared using a stoichiometric precursor. With the addition of 10%-TMA, the peak at approximately 12.63° decreased; however, with an increase in the TMA content to 15%, the peak intensity increased. In contrast, the peak intensity at approximately 12.63° decreased with a further increase in the TMA content beyond 20%. These changes in the XRD results confirmed that the PbI2 phase was strongly influenced by the concentration of the TMA additive. Therefore, the performance enhancement by the TMA additive could be attributed to the suppression of the surface defects.[24]

Figure 4

(a) UV–vis absorption spectra of the control FAPbI3 film and the TMA-modified perovskite film; (b) zoom-in view near the absorbance edge; (c) PL spectrum of the perovskite deposited on glass; and (d) XRD patterns of the perovskite film and perovskite films modified with 5%-, 10%-, 15%-, and 20%-TMA additives.

We investigate the control and modified electron-only device’s electronic trap states using the space–charge–limited current (SCLC) technique. Figure a shows the sample structure for this measurement, while Figure b shows the dark current–voltage (I–V) analysis for control and modified perovskite devices. At the low voltage, an ohmic response of the electron-only devices was confirmed by the fit to I = V (linear). At the high voltage, the current shows a significant increase for perovskite films modified with 0%-, 5%-, 10%-, 15%-, and 20%-TMA additives at VTFL = 0.42, 0.32, 0.22, 0.27, and 0.47 V, respectively. The trap-state density is calculated using the following equationwhere e is elementary of the electron, L is the thickness between the two metal contacts, ε is the relative dielectric constant of perovskite (26), ε0 is the vacuum permittivity, and ntraps is the trap-state density. Correspondingly, we found that the control device has a trap density ntraps = 3.6 × 1015 cm–3, while 10%-TMA perovskite film shows a low trap density ntraps = 1.88 × 1015 cm–3.[20,39] Clearly, using a 10%-TMA additive effectively passivates electron traps, leading to the formation of a high-quality perovskite layer.

Figure 5

(a) Sample structure for trap density measurement. (b) Dark I–V curves of the electron-only devices and fabricated using perovskite films modified with different concentrations of TMA: 0%, 5%, 10%, 15%, and 20%.

The effect of TMA on the photovoltaic properties of the PSCs is summarized in Table . Figure a–d shows the statistical distribution of the open-circuit voltage (VOC), short-circuit current density (JSC), fill factor (FF), and PCE of the control and TMA additive-perovskite-based devices. The control device without TMA exhibited a JSC of 21.84 mA·cm–2, a VOC of 0.95 V, a FF of 0.73, and a PCE of 15.15%. The high efficiency of the TMA additive in the perovskite-based device could be related to the excellent quality of the absorber layer. Moreover, an additive of MAPbBr3 was employed in the perovskite layer to enhance the stabilization of the α-FAPbI3 phase. The addition of TMA to (FAPbI3)0.97(MAPbBr3)0.03 perovskite increased the overall efficiency of all of the PSC devices. The perovskite film with a 10%-TMA additive exhibited the champion PCE of 17.26% with a VOC of 1.01 V, a JSC of 23.11 mA·cm–2, and a FF of 0.74, indicating that the optimum TMA content was 10% (Table , Figure e). The enhanced efficiency of the PSC owing to the increased values of VOC and FF could be attributed to the improved optical properties, good quality of the perovskite film, and lower series resistance (Table ). The enhanced FF of the modified perovskite films could be assigned to the surface flatness of the perovskite films, which improved the charge-transfer property.[40] However, with an increase in the TMA concentration to 20%, the PCE values decreased to 15.28%, which could be attributed to the formation of a coarse perovskite film with several pinholes. Figure f shows the current density–voltage (J–V) curves of the optimum PSC device (10%-TMA), which are consistent with those of the discussed results.

Table 1

Photovoltaic Performance of PSCs with Various Amounts of TMA Based on 75 Cells

PSCs	JSC (mA·cm–2)	VOC (V)	FF	PCE (%)	Rs
control	21.84	0.95	0.73	15.15	148
5%-TMA	23.94	0.99	0.69	16.46	142
10%-TMA	23.11	1.01	0.74	17.26	120
15%-TMA	22.49	0.99	0.71	15.91	127
20%-TMA	20.96	1.00	0.72	15.28	139

Figure 6

(a–d) Statistical distribution of the PCE, JSC, VOC, and FF of the control PSC and the TMA-modified perovskite-based PSCs. Data were collected from 75 cells. (e) Current density–voltage curve of the best cell obtained using 10% TMA. (f) Photocurrent density–voltage curve of the TMA-modified FAPbI3-based PSCs at different TMA concentrations.

Figure a exhibits the external quantum efficiency (EQE) spectra of the control and 10%-TMA-based PSCs. The light response from the EQE spectra of the modified devices was approximately 90%, higher than those of the control samples. This could be assigned to the improved optical properties and reduced charge recombination at the hole transport layer/perovskite interface owing to the flat surface of the perovskite layer. The integrated JSC values of the control and 10%-TMA-based PSCs from the EQE spectra were 23.5 and 22.5 mA·cm–2, respectively, which are highly consistent with the measured JSC. The histogram of each device was fitted with a normal distribution model, as shown in Figure b. Both histograms illustrate that the 10%-TMA additive in perovskite-based devices exhibited enhanced solar cell performance.

Figure 7

(a) EQE spectra of the control and 10%-TMA-based PSCs. (b) PCE distribution of the 30 PSC devices.

Conclusions

In conclusion, a modified procedure for absorber layers with a TMA additive at different concentrations was developed to fabricate high-quality perovskite films using a two-step deposition method. By varying the amount of TMA in the perovskite precursor solutions, a significant enhancement of the PCE from 15.15 to 17.26% was obtained at a TMA concentration of 10 wt%. This improvement was based on the improved quality of the perovskite film owing to the flat surface, lack of pinholes, and low series resistance, which enhanced the JSC and VOC. Furthermore, the 10 wt %-TMA additive-based devices exhibited a suppressed hysteresis effect compared to the control device. The improved performance could be attributed to the presence of the TMA additive in the perovskite precursor solution during the perovskite formation process. The findings of this study indicate that the efficiency of mesoscopic PSCs can be further improved by modifying the absorber film via the addition of additives.

Experimental Section

Reagents and Materials

Titanium diisopropoxide bis(acetylacetone) (75 wt % in isopropanol, Sigma-Aldrich), butyl alcohol (99%, Sigma-Aldrich), TiO2 (Dyesol NT-18, Dyesol, Australia), absolute alcohol (≥99.5%, Sigma-Aldrich), lead(II) iodide (99.999% trace metals basic, Sigma-Aldrich), N,N-dimethylformamide (DMF, 99.8%, Sigma-Aldrich), dimethyl sulfoxide (DMSO, ≥99.9%, Sigma-Aldrich), trimesic acid (TMA; benzene-1,3,5-tricarboxylic acid, 95%, Sigma-Aldrich), formamidinium iodide (FAI, greatcellsolar, Australia), methylammonium bromide (MABr, Sigma-Aldrich), methylammonium hydrochloride (MACl, Sigma-Aldrich), 2-propanol (anhydrous, 99.5%, Sigma-Aldrich), chlorobenzene (99.8%, Sigma-Aldrich), acetonitrile (99.93%, Sigma-Aldrich), and 2,2′,7,7′-tetrakis[N,N-di(4-methoxyphenyl)amino]-9,9′-spirobifluorene (spiro-OMeTAD, 99%, Sigma-Aldrich), 4-tert-butylpyridine (98%, Sigma-Aldrich), bis(trifluoromethane)sulfonimide lithium salt (Li-TSFI; ≥99.0%, Sigma-Aldrich) were used. All reagents were used as received without any further purification.

Device Preparation

Briefly, a glass/FTO substrate (7 Ω sq–1) was cleaned consecutively with isopropanol, acetone, distilled water, and ethanol, then dried under a N2 flow, and finally treated with UV ozone for 15 min. Subsequently, a TiO2 blocking layer was spin-coated on the washed FTO substrate at 2000 rpm for 20 s using 0.15 M titanium diisopropoxide bis(acetylacetone) (75% in 1-butanol) in 1-butanol solution, which was heated at 125 °C for 15 min. After cooling to room temperature, the porous layer with a TiO2 paste was diluted in ethanol (1:6, wt/wt), after which the paste was spin-coated on the TiO2 compact layer at 4000 rpm for 20 s, followed by annealing at 480 °C for 30 min. The perovskite layer was deposited by performing a two-step spin-coating deposition method under a N2 atmosphere in a glovebox. First, a PbI2 solution (1.3 M) was prepared by dissolving 600 mg of PbI2 in 900 μL of DMF and 100 μL of DMSO, after which the mixture was stirred at room temperature for 1 day. Subsequently, 50 μL of PbI2 was spin-coated on the mesoporous TiO2 film at 2000 rpm for 20 s (without loading time). The mixture solution of FAI/MABr/MACl (60:6:6, mg) without and with a certain amount of TMA (mass ratio of TMA vs FAI) in 2-propanol was spin-coated onto the PbI2-coated substrate for 20 s (loading time) at 4000 rpm for 20 s and dried at 150 °C for 15 min. Subsequently, 50 μL of the spiro-OMeTAD solution was dropped on the absorber layer at 3000 rpm for 30 s, where 72.3 mg of spiro-MeOTAD was dissolved in 1 mL of chlorobenzene with the addition of 28.8 μL of 4-tert-butylpyridine and 17.5 μL of Li-TFSI solution (52 mg of Li-TSFI in 100 μL of acetonitrile). Finally, 80 nm of gold was evaporated on the spiro-OMeTAD film.

Device Characterization

The UV–vis absorption spectra were measured using an Agilent 8453 UV–vis spectrophotometer (Agilent 8453, Agilent Technologies, Santa Clara, CA). The XRD patterns of the films were obtained using an XRD Rigaku DMAX 2200 system (Rigaku, Tokyo, Japan) with Cu Kα (λ = 0.15406 nm) as the X-ray source. An infrared spectrometric analyzer (Vertex 70, Bruker, Ettlingen, Germany) was used to record the FTIR spectra. The cross-sectional morphology and the surface of the films were characterized using a field emission scanning electron microscope (Hitachi S-4700, Tokyo, Japan) operated at 10 kV. The steady-state PL spectra of the films were determined using a QuantaMaster TM 50 PTI (Birmingham, New Jersey). The J–V curves of PSCs were measured under 1 sun illumination (AM1.5G, 100 mW·cm–2), using a solar simulator (Polaromix K201, Solar simulator LAB 50, McScience K3000, McScience, Gyeonggi-do, Korea). The size of PSCs was calculated with an area of 0.053 cm2 using a metal mask. The EQE was measured on a computer test controlling system (McScience, Gyeonggi-do, Korea) containing a xenon lamp power source.