Enhanced Efficiencies of Perovskite Solar Cells by Incorporating Silver Nanowires into the Hole Transport Layer

In this study, we incorporated silver nanowires (AgNWs) into poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) as a hole transport layer (HTL) for inverted perovskite solar cells (PVSCs). The effect of AgNW incorporation on the perovskite crystallization, charge transfer, and power conversion efficiency (PCE) of PVSCs were analyzed and discussed. Compared with neat PEDOT:PSS HTL, incorporation of few AgNWs into PEDOT:PSS can significantly enhance the PCE by 25%. However, the AgNW incorporation may result in performance overestimation due to the lateral charge transfer. The corrosion of AgNWs with a perovskite layer was discussed. Too much AgNW incorporation may lead to defects on the interface between the HTL and the perovskite layer. An extra PEDOT:PSS layer over the pristine PEDOT:PSS-AgNW layer can prevent AgNWs from corrosion by iodide ions.

Compared with 0-D metal nanoparticles, 1-D silver nanowires (AgNWs) display superior electrical conductivity due to the fast electron transport on their axial direction. Applying AgNWs in transparent conductive electrodes show not only excellently flexible properties but also a very high ratio of DC conductivity to optical conductivity, indicating AgNWs are one of the best potential replacements for indium tin oxides [34]. AgNW networks have been improved extremely in terms of the electrical conductivity, optical properties, and manufacturing [35][36][37][38][39]. Therefore, some studies have used AgNWs to fabricate the top electrodes [40][41][42] and bottom window electrodes [43,44] of PVSCs, revealing feasibility of AgNW electrodes on PVSCs. In this study, we incorporated AgNWs into the PEDOT:PSS as a HTL for inverted PVSCs. The effect of AgNWs on the formation of the perovskite layer, charge transfer resistance, and PCE of PVSCs were studied with respect to varying the concentration of AgNWs. Compared with neat PEDOT:PSS HTL, incorporation of few AgNWs into PEDOT:PSS significantly enhances the PCE by one-fourth. The corrosion of AgNWs with a perovskite layer was discussed.

Device Fabrication
Typically, the ITO-coated glass substrate was masked by PI tape in a T shape (2 × 1.5 cm) and then the unmasked portion of the ITO was etched with zinc powder and diluted HCl (<6 M). After few seconds, the residual zinc powder was removed by washed with ethanol. Then the mask was removed and cleaned three times with detergent, ethanol, and acetone. AgNWs were synthesized as reported previously [45,46], with some of the parameters modified slightly to obtain fine AgNWs (see supporting information). The mixture of PEDOT:PSS and AgNWs EG solution (0, 0.17, 0.51, and 0.85 wt%) in the weight ratio of 20:1 was coated on ITO glass at 3000 rpm for 60 s and then dried by heating at 120 • C for 10 min. MAI and PbI 2 were mixed in DMSO/DMF in the volumetric ratio of 4:6 at 60 • C to obtain a 1.2-M MAPbI 3 solution. The solution was coated on the HTL under the conditions of 1000 rpm for 10 s, followed by 5000 rpm for 20 s and heating at 100 • C to complete the film formation. 10 mg of PCBM and 0.8 mg of CTAB were dissolved in 500 µL of chloroform with ultrasonication and then spin-coated onto the perovskite layer at 1250 rpm for 30 s. Finally, near 100-nm thick silver was thermally deposited on the ETL layer by a thermal evaporator (Kao-Duen Tech, Taiwan) under a vacuum of 5 × 10 −6 torr and a plating rate of 0.8-0.9 Å/s. The structure of the PVSC is shown schematically in Figure S1.

Measurements and Characterization
The crystalline phase of AgNWs and MAPbI 3 were characterized through X-ray diffraction (XRD) using an X-ray diffractometer (Miniflex II, Rigaku, Tokyo, Japan) and Cu Kα radiation (wavelength 0.15418 nm) with a fixed operating voltage of 30 kV and a fixed current of 15 mA. The sheet resistances of the PEDOT:PSS incorporated with AgNWs were measured using a four-pin probe meter (Loresta-GP, Mitsubishi Chemical, Tokyo, Japan) with an MCP-T610 probe. The morphology of AgNWs, MAPbI 3 , and the devices were examined using a field-emission scanning electron microscope (SEM) (JSM-7401F, JEOL, Tokyo, Japan). The length of AgNWs was determined using an optical microscope (OM) (M835, M&T Optics, Taiwan). The absorption and emission spectra of the ITO/PEDOT:PSS-AgNW/MAPbI 3 were determined using a UV-Vis spectrometer (V770, Jasco, Tokyo, Japan) and a fluorescence spectrometer (LS-55, Perkin Elmer, Waltham, MA, USA), respectively. The photocurrent density-voltage (J-V) characteristics were measured under irradiation of 100 mW·cm -2 using a solar simulator (MFS-PV, Hong-Ming Technology, Taiwan) equipped with a source meter (Keithley 2400, Keithley Instruments, Cleveland, OH, USA). Electrochemical impedance spectra (EIS) were measured over the frequency range of 50 mHz-100 kHz with a potential perturbation of 10 mV using an electrochemical workstation (Zennium, Zahner Elektrik, Kronach, Germany).

Results and Discussion
During the AgNW synthesis process, both length and diameter of AgNWs increase with reaction time. In order to prevent the AgNWs from exceeding HTL in thickness to conduct charge recombination or current leakage, we shortened the reaction time to make finer ones in our experiments. The as-prepared AgNWs had lengths of~20 µm and diameters of 20-30 nm (Figure 1), confined by the thickness of the HTL. Although PEDOT:PSS is electrically conductive in nature, the sheet resistance of the PEDOT:PSS layer really decreases with the increase of AgNW content ( Figure S2), revealing the AgNW incorporation can improve the conductivity of the HTL.

Results and Discussion
During the AgNW synthesis process, both length and diameter of AgNWs increase with reaction time. In order to prevent the AgNWs from exceeding HTL in thickness to conduct charge recombination or current leakage, we shortened the reaction time to make finer ones in our experiments. The as-prepared AgNWs had lengths of ~20 μm and diameters of 20-30 nm (Figure 1), confined by the thickness of the HTL. Although PEDOT:PSS is electrically conductive in nature, the sheet resistance of the PEDOT:PSS layer really decreases with the increase of AgNW content ( Figure  S2), revealing the AgNW incorporation can improve the conductivity of the HTL. Observing in the XRD patterns (Figure 2), neither an obvious change on the crystallinity of MAPbI3 for the PEDOT:PSS-AgNW HTL nor a peak attributed to PbI2 (12.7°) was found, implying that AgNW incorporation did not lead to PbI2 separated out from MAPbI3. Note that the XRD patterns did not display the main peak at 37.8° corresponding with the (111) plane of AgNWs because the AgNWs are few and sheltered by the MAPbI3 layer. Figure 3 shows the top-view SEM images of the morphology of the MAPbI3 layer upon the HTL with various AgNW contents. The average crystal size of MAPbI3 in the range of 200-220 nm is not strongly dependent on the AgNW contents in the HTL. However, the PL intensity of MAPbI3 decreases with increasing AgNW amount in the HTL but increases with further increasing AgNW amount (Figure 4). The fluorescence intensity results from the electron-hole recombination. Therefore, the PL analysis implies that moderate AgNW incorporation can reduce charge recombination rate. The AgNW incorporation also induces the slight increase on the absorption in the range of 300-400 nm ( Figure 5), which may arise from surface plasmon resonance absorption of AgNWs [47]. Observing in the XRD patterns (Figure 2), neither an obvious change on the crystallinity of MAPbI 3 for the PEDOT:PSS-AgNW HTL nor a peak attributed to PbI 2 (12.7 • ) was found, implying that AgNW incorporation did not lead to PbI 2 separated out from MAPbI 3 . Note that the XRD patterns did not display the main peak at 37.8 • corresponding with the (111) plane of AgNWs because the AgNWs are few and sheltered by the MAPbI 3 layer. Figure 3 shows the top-view SEM images of the morphology of the MAPbI 3 layer upon the HTL with various AgNW contents. The average crystal size of MAPbI 3 in the range of 200-220 nm is not strongly dependent on the AgNW contents in the HTL. However, the PL intensity of MAPbI 3 decreases with increasing AgNW amount in the HTL but increases with further increasing AgNW amount (Figure 4). The fluorescence intensity results from the electron-hole recombination. Therefore, the PL analysis implies that moderate AgNW incorporation can reduce charge recombination rate. The AgNW incorporation also induces the slight increase on the absorption in the range of 300-400 nm ( Figure 5), which may arise from surface plasmon resonance absorption of AgNWs [47].       Table 1. The experimental results show that the PCEs of the PVSCs increase with increasing AgNW content in the HTL, reached a maximum at PVSC-2 (0.51 wt% AgNWs), and decreased with further increasing AgNW content (PVSC-3). The order is reciprocal to the PL intensity. Because the decrease of the fluorescence intensity (quenching) indicates the decrease of electron-hole recombination rate, the enhancement on PCE caused by the AgNW incorporation may result from the rapid hole transfer away through AgNWs to suppress the charge recombination. The PCE can be enhanced by approximately 25% compared to the pristine cell (PVSC-0). Figure 7 displays the Nyquist plots of the EIS for the PVSCs with various AgNW incorporations in the HTL. As listed in Table 1, the external (R1), interface (R2), and charge recombination resistances (R3) appears sequentially from high to low frequency, respectively. With the increase of AgNW content, R2 decreases significantly, whereas R3 reaches a maximum at 0.51 wt% AgNW incorporation. The results indicate that the AgNW incorporation decreases the interface resistance and the rate of electron-hole recombination, which is consistent with PL and J-V analyses. However, too much AgNW incorporation (0.85 wt% AgNWs) leads to high PL intensity, low PCE, high interface resistance, and low resistance of charge combination. Figure S3 shows the monochromatic incident photon-to-electron conversion efficiency (IPCE) spectrum of PVSC-2 (the best cell). The calculated JSC from integration of the IPCE over the AM 1.5 G is 20.6 mA/cm 2 , which is smaller than the value of 24.3 mA/cm 2 obtained from J-V curve listed in Table 1. The discrepancy in the two methods may arise from different light intensity, mismatch in the absorption spectrum, or degradation of device [48,49]. However, the difference in JSC values obtained from the IPCE and J-V curve is much larger than that reported in literature. To realize the cause of the discrepancy, we measured the performance of the solar cells with a mask. The JSC and PCE led to a 5-10% reduction after applying a mask. Therefore, the enhancement on JSC and PCE by incorporating AgNWs may result partially from the lateral charge transport, which overestimates the performance.   Table 1. The experimental results show that the PCEs of the PVSCs increase with increasing AgNW content in the HTL, reached a maximum at PVSC-2 (0.51 wt% AgNWs), and decreased with further increasing AgNW content (PVSC-3). The order is reciprocal to the PL intensity. Because the decrease of the fluorescence intensity (quenching) indicates the decrease of electron-hole recombination rate, the enhancement on PCE caused by the AgNW incorporation may result from the rapid hole transfer away through AgNWs to suppress the charge recombination. The PCE can be enhanced by approximately 25% compared to the pristine cell (PVSC-0). Figure 7 displays the Nyquist plots of the EIS for the PVSCs with various AgNW incorporations in the HTL. As listed in Table 1, the external (R 1 ), interface (R 2 ), and charge recombination resistances (R 3 ) appears sequentially from high to low frequency, respectively. With the increase of AgNW content, R 2 decreases significantly, whereas R 3 reaches a maximum at 0.51 wt% AgNW incorporation. The results indicate that the AgNW incorporation decreases the interface resistance and the rate of electron-hole recombination, which is consistent with PL and J-V analyses. However, too much AgNW incorporation (0.85 wt% AgNWs) leads to high PL intensity, low PCE, high interface resistance, and low resistance of charge combination. Figure S3 shows the monochromatic incident photon-to-electron conversion efficiency (IPCE) spectrum of PVSC-2 (the best cell). The calculated J SC from integration of the IPCE over the AM 1.5 G is 20.6 mA/cm 2 , which is smaller than the value of 24.3 mA/cm 2 obtained from J-V curve listed in Table 1. The discrepancy in the two methods may arise from different light intensity, mismatch in the absorption spectrum, or degradation of device [48,49]. However, the difference in J SC values obtained from the IPCE and J-V curve is much larger than that reported in literature. To realize the cause of the discrepancy, we measured the performance of the solar cells with a mask. The J SC and PCE led to a 5-10% reduction after applying a mask. Therefore, the enhancement on J SC and PCE by incorporating AgNWs may result partially from the lateral charge transport, which overestimates the performance.     The SEM image reveals that some defects were found in the HTL for high AgNW incorporation (Figure 8a). We infer that too many AgNWs might have protrude out of the HTL and contacted the MAPbI3 solution when the subsequent perovskite layer was coated. The iodide ion in the MAPbI3 might react with AgNWs to form defects [50]. A protection layer may prevent AgNWs from reacting with halogen ions to form nonconducting silver halide, deteriorating the optoelectronic properties [51,52]. In the previous study [45], we proved that PEDOT:PSS can prevent AgNWs from corrosion by the iodine electrolyte. In order to verify the argument, we spin-coated extra PEDOT:PSS layers upon the pristine PEDOT:PSS-AgNW layer to prevent AgNWs from touching the MAPbI3 solution.
After coating an extra one, two, or three PEDOT:PSS layers, no defects were found in the interface The SEM image reveals that some defects were found in the HTL for high AgNW incorporation (Figure 8a). We infer that too many AgNWs might have protrude out of the HTL and contacted the MAPbI 3 solution when the subsequent perovskite layer was coated. The iodide ion in the MAPbI 3 might react with AgNWs to form defects [50]. A protection layer may prevent AgNWs from reacting with halogen ions to form nonconducting silver halide, deteriorating the optoelectronic properties [51,52].
In the previous study [45], we proved that PEDOT:PSS can prevent AgNWs from corrosion by the iodine electrolyte. In order to verify the argument, we spin-coated extra PEDOT:PSS layers upon the pristine PEDOT:PSS-AgNW layer to prevent AgNWs from touching the MAPbI 3 solution. After coating an extra one, two, or three PEDOT:PSS layers, no defects were found in the interface between the MAPbI 3 layer and the PEDOT:PSS-AgNW layer (Figure 8b-d). The PCE was increased from 12.14% to 13.04% with one extra PEDOT:PSS layer, whereas too much PEDOT:PSS (spin-coating PEDOT:PSS two or three times) makes the PCE decline ( Figure S4 and Table 2). Figure 9 shows the Nyquist plots of the EIS for the PVSCs with extra PEDOT:PSS layers. The resistance of charge recombination (R 3 ) increases with one covering PEDOT:PSS layer (PVSC-3-1L) to prevent against iodine-mediated corrosion. However, too much PEDOT:PSS coating decreases the resistance of charge recombination, which increases the probability of electron-hole recombination due to the increase of distance of charge transport and thereby decreases the PCE.

Conclusions
We prepared the inverted PVSCs with incorporating AgNWs into PEDOT:PSS as a HTL. The AgNW incorporation increased the rate of charge transport and suppressed the electron-hole recombination. Incorporation of 0.51 wt% AgNWs revealed a maximum improvement, being a 25% enhancement on the PCE. However, the AgNW incorporation may overestimate performance due to

Conclusions
We prepared the inverted PVSCs with incorporating AgNWs into PEDOT:PSS as a HTL. The AgNW incorporation increased the rate of charge transport and suppressed the electron-hole recombination. Incorporation of 0.51 wt% AgNWs revealed a maximum improvement, being a 25% enhancement on the PCE. However, the AgNW incorporation may overestimate performance due to the lateral charge transfer. Too much AgNW incorporation may result in the contact of AgNWs with the MAPbI 3 solution, thereby forming defects on the interface between the HTL and the MAPbI 3 layer. An extra PEDOT:PSS layer over the pristine PEDOT:PSS-AgNW layer can prevent AgNWs from corrosion by iodide ions, whereas too much PEDOT:PSS protection may increase the distance of charge transport and may thus decrease the PCE.

Conflicts of Interest:
The authors declare no conflict of interest.