Polymer/Inorganic Hole Transport Layer for Low-Temperature-Processed Perovskite Solar Cells

In the search for improvements in perovskite solar cells (PSCs), several different aspects are currently being addressed, including an increase in the stability and a reduction in the hysteresis. Both are mainly achieved by improving the cell structure, employing new materials or novel cell arrangements. We introduce a hysteresis-free low-temperature planar PSC, composed of a poly(3-hexylthiophene) (P3HT)/CuSCN bilayer as a hole transport layer (HTL) and a mixed cation perovskite absorber. Proper adjustment of the precursor concentration and thickness of the HTL led to a homogeneous and dense HTL on the perovskite layer. This strategy not only eliminated the hysteresis of the photocurrent, but also permitted power conversion efficiencies exceeding 15.3%. The P3HT/CuSCN bilayer strategy markedly improved the life span and stability of the non-encapsulated PSCs under atmospheric conditions and accelerated thermal stress. The device retained more than 80% of its initial efficiency after 100 h (60% after 500 h) of continuous thermal stress under ambient conditions. The performance and durability of the PSCs employing a polymer/inorganic bilayer as the HTL are improved mainly due to restraining perovskite ions, metals, and halides migration, emphasizing the pivotal role that can be played by the interface in the perovskite-additive hole transport materials (HTM) stack.


Introduction
Halide perovskite solar cells (PSCs) have become one of the most investigated research topics in the renewable and sustainable energy community, owing to their effectiveness and high-power conversion efficiency (PCE) [1], recently reaching more than 25% [2].
The halide perovskite used for photovoltaics has a typical AMX 3 crystal structure, where A is a monovalent cation, such as Cs + , Rb + , CH 3 NH 3 + , NH 2 CH = NH 2 + , CH 3 CH 2 NH 3 + ; M = Pb 2 + , Sn 2 + ; and X is a halide, such as Cl -, Bror I - [3]. At the same time, employing a mixed-cation and mixed-halide perovskite composition can dramatically improve the performance and stability of PSCs [4,5]. In a typical PSC, the perovskite layer is deposited between a metal oxide electron transport layer (ETL), with either a mesoporous or planar structure, and a hole-transport layer (HTL). Several hole was used as an absorber layer in our PSC, employing a one-step anti-solvent quenching method, which produced a lower surface roughness and smaller interface contact resistance when compared with the conventional two-step method [53]. After the final adjustment and optimization, a PCE of 15.3% was obtained by employing a polymer/inorganic bilayer (P3HT/CuSCN) strategy. Moreover, the atmospheric life span and thermal stability of non-encapsulated PSCs were significantly improved in P3HT/CUSCN based devices.

Materials and Methods
The fabricated devices in this work were made from FTO/glass substrates (Pilkington, 8Ω cm −1 , 25 mm × 25 mm), etched by a raster scanning laser (Nd:YVO 4 pulsed at 30 kHz average output power, P = 10 W). The patterned substrates were cleaned in an ultrasonic bath by using deionized water, acetone and 2-propanol solution for 10 min intervals each. Then FTO was kept under a UV lamp for 20 min, and the SnO 2 solution (SnCl 2 ·2H 2 O, in ethanol, 23.5 mg mL −1 ) was deposited on the patterned FTO as ETL by two-step spin coating: 500 rpm with an acceleration of 250 rpm/s (for 4 s), followed by 2000 rpm with an acceleration of 1000 rpm s −1 for 20 s. The substrates were then annealed for 10 min at 100 • C, followed by 180 • C for 3 h.
Prior to perovskite deposition, SnO 2 substrates were kept under a UV lamp for 1 h and then immediately transferred to the glove-box. The "mixed perovskite" solution contained FAI ( [3].
The perovskite layer was deposited by a one-step anti-solvent quenching method; the spin coating was performed in two parts, first at a speed of 1000 rpm for 10 s with an acceleration of 250 rpm s −1 , followed by a second step at 6000 rpm for 20 s with 2000 rpm s −1 . Chlorobenzene was dripped by micropipette over the spinning substrate 7 s before the end of the latter. The samples were then annealed at 100 • C on a hot plate for 1 h.
Different concentrations of CuSCN solutions were prepared (15 mg mL −1 , 25 mg mL −1 in dipropyl sulfide). Then 15 mg CuSCN powder was easily dissolved in dipropyl sulfide solvent, while the dissolution of 25 mg CuSCN in the same solvent required double filtration, with a 0.22 µm PTFE filter, followed by 1 h at 100 • C to obtain a clear solution, and finally the cold solution was spin-coated at 6000 rpm for 45 s. The substrates were then kept at 100 • C for 10 min to remove any remaining solvent, and another layer of CuSCN was deposited on top of the previous one and annealed for 10 min at 100 • C. In order to improve the stability of the fabricated PSCs to withstand 85 • C thermal cycling, a P3HT (124 kDa MW) in chlorobenzene solution was prepared following a previously reported procedure [54,55]. The P3HT layer was applied below and above the CuSCN by dynamic deposition following the same procedure as for the CuSCN layer, but without the final annealing step. At the end, 100 nm Au was deposited using thermal evaporation in a high vacuum chamber (10 −6 mbar). All coating and fabrication steps were performed in a glovebox, except for the SnO 2 deposition. Figure 1 shows a schematic overview of the sample preparation process.
The morphology and the grain size of SnO 2 , perovskite and HTM layers were investigated by field emission scanning electron microscopy (FE-SEM, HITACHI (S-4160)). The UV-Vis spectra were obtained by using a BLACK-Comet UV-Vis spectrometer (Dymax EC-5000 lamp, with irradiance power of 225 mWcm −2 ) [56]. Dark J-V measurement and the photocurrent-voltage measurements of the devices were performed using a solar simulator (ABET Sun 2000, class A) under one-sun illumination conditions AM 1.5 (100 mWcm −2 ), calibrated with a certified reference Si cell (RERA Solutions RR-1002). The electrochemical impedance spectroscopy (EIS) measurements were performed with an Eco-ChemieAutolab PGSTAT 302N electrochemical workstation controlled with NOVA software in the frequency range of 0.01 to 100 kHz, under 100 mWcm −2 solar simulator at a bias voltage ranging from 0.7 to 1.0 V. °C thermal cycling, a P3HT (124 kDa MW) in chlorobenzene solution was prepared following a previously reported procedure [54,55]. The P3HT layer was applied below and above the CuSCN by dynamic deposition following the same procedure as for the CuSCN layer, but without the final annealing step. At the end, 100 nm Au was deposited using thermal evaporation in a high vacuum chamber (10 −6 mbar). All coating and fabrication steps were performed in a glovebox, except for the SnO2 deposition. Figure 1 shows a schematic overview of the sample preparation process.

Results
The optical properties of the HTM on the perovskite layer were investigated by UV-Vis absorption spectroscopy ( Figure 2). The absorption edge for all types of samples was observed at 750 nm, as expected. For the sample with P3HT, a clear onset of the P3HT absorption around 650 nm was observed [57]. The morphology and the grain size of SnO2, perovskite and HTM layers were investigated by field emission scanning electron microscopy (FE-SEM, HITACHI (S-4160)). The UV-Vis spectra were obtained by using a BLACK-Comet UV-Vis spectrometer (Dymax EC-5000 lamp, with irradiance power of 225 mWcm −2 ) [56]. Dark J-V measurement and the photocurrent-voltage measurements of the devices were performed using a solar simulator (ABET Sun 2000, class A) under one-sun illumination conditions AM 1.5 (100 mWcm −2 ), calibrated with a certified reference Si cell (RERA Solutions RR-1002). The electrochemical impedance spectroscopy (EIS) measurements were performed with an Eco-ChemieAutolab PGSTAT 302N electrochemical workstation controlled with NOVA software in the frequency range of 0.01 to 100 kHz, under 100 mWcm −2 solar simulator at a bias voltage ranging from 0.7 to 1.0 V.

Results
The optical properties of the HTM on the perovskite layer were investigated by UV-Vis absorption spectroscopy ( Figure 2). The absorption edge for all types of samples was observed at 750 nm, as expected. For the sample with P3HT, a clear onset of the P3HT absorption around 650 nm was observed [57].  The photovoltaic performance of the fabricated PSCs was evaluated under simulated sun AM 1.5 G (100 mWcm −2 ). The PV parameters obtained from the J-V curves of the devices are shown in Figure 3 and Figure S1 of the Supporting Materials. The corresponding results are also given in Table  1. The photovoltaic performance of the fabricated PSCs was evaluated under simulated sun AM 1.5 G (100 mWcm −2 ). The PV parameters obtained from the J-V curves of the devices are shown in Figure 3 and Figure S1 of the Supporting Materials. The corresponding results are also given in Table 1.    The concentration of the precursor solution and the thickness of the HTL seemed to play a key role in the performance assessment of the PSCs. Higher PCE results were obtained with 25 mg mL −1 CuSCN in dipropyl sulfide.

SnO2/Perovskite/CuSCN/Au
Increasing the thickness of the CuSCN layer up to 90 nm, achieved with a double-spin coating, improved the maximum (average) efficiency up to 14.7% (13.2%), as reported in Figure 3 and Figure  S1, respectively. This significant improvement is essentially related to the increase in open-circuit voltage (Voc) and fill factor (FF), as shown in Table 1. Based on the results, by increasing the thickness of the CuSCN layer, the Voc was significantly increased. This improvement can be attributed to a better hole extraction due to a reduction in recombination phenomena. On the other hand, a significant improvement in FF implies a reduction in contact resistance and improved charge transfer processes.
As mentioned before, one of the most important parameters of PSC devices is hysteresis, which is normally investigated by comparing the forward and backward J-V scan. Figure 4a     The concentration of the precursor solution and the thickness of the HTL seemed to play a key role in the performance assessment of the PSCs. Higher PCE results were obtained with 25 mg mL −1 CuSCN in dipropyl sulfide.
Increasing the thickness of the CuSCN layer up to 90 nm, achieved with a double-spin coating, improved the maximum (average) efficiency up to 14.7% (13.2%), as reported in Figure 3 and Figure S1, respectively. This significant improvement is essentially related to the increase in open-circuit voltage (Voc) and fill factor (FF), as shown in Table 1. Based on the results, by increasing the thickness of the CuSCN layer, the Voc was significantly increased. This improvement can be attributed to a better hole extraction due to a reduction in recombination phenomena. On the other hand, a significant improvement in FF implies a reduction in contact resistance and improved charge transfer processes.
As mentioned before, one of the most important parameters of PSC devices is hysteresis, which is normally investigated by comparing the forward and backward J-V scan. Figure 4A  Furthermore, the FE-SEM cross-section image from the P3HT/CuSCN-based device is also presented in Figure 4d. The PV parameters obtained from the J-V curve indicate that the P3HT/CuSCN/Au-based device (Figure 4c and Figure S2) showed a better performance (PCE = 15.3%), compared with the CuSCN/P3HT/Au-based device (PCE = 8.8%).
A summary of the statistical analysis of PV parameters, extracted from J-V curves for 12 independent devices, are given in Figure 4c and Figure S2, while max (average ± standard deviation) PV parameters are reported in Table 2. The stability of the non-encapsulated cells, under environmental and thermal conditions, are shown in Figure 5a,b, respectively. The atmospheric shelf-life of PSCs with a P3HT interlayer ( Figure  5a) was longer than that with only CuSCN. As shown in Figure 5b, the thermal stability of the P3HT/CuSCN/Au structure turned out to be several times better compared with the devices having other arrangements. Planar PSCs using the CuSCN HTL or the CuSCN/P3HT HTL rapidly lose their initial efficiency against 85 °C thermal stress, mainly due to the interaction between CuSCN and perovskite. This phenomenon could be related to a structural change of perovskite surface upon The J-V plots for the PSCs with the different situations of the P3HT layer (perovskite-HTL and HTL-Au interfaces) are shown in Figure 4B. The statistical PCEs are shown in Figure 4C. Furthermore, the FE-SEM cross-section image from the P3HT/CuSCN-based device is also presented in Figure 4D.
A summary of the statistical analysis of PV parameters, extracted from J-V curves for 12 independent devices, are given in Figure 4C and Figure S2, while max (average ± standard deviation) PV parameters are reported in Table 2. The stability of the non-encapsulated cells, under environmental and thermal conditions, are shown in Figure 5A,B, respectively. The atmospheric shelf-life of PSCs with a P3HT interlayer ( Figure 5A) was longer than that with only CuSCN. As shown in Figure 5B, the thermal stability of the P3HT/CuSCN/Au structure turned out to be several times better compared with the devices having other arrangements. Planar PSCs using the CuSCN HTL or the CuSCN/P3HT HTL rapidly lose their initial efficiency against Energies 2020, 13, 2059 7 of 12 85 • C thermal stress, mainly due to the interaction between CuSCN and perovskite. This phenomenon could be related to a structural change of perovskite surface upon heating at 85 • C [58], which could be accelerated by Cu(II) ions through a reaction with halide ions and form copper halide defects. Furthermore, due to the constraints on the CuSCN precursor solvent (requirement of polar solvents), the interaction of such solvents with perovskite surface during deposition of HTM leads to increased defects on the perovskite layer and accelerates the thermal decomposition. On the other hand, by using the P3HT/CuSCN bilayer structure as the HTL, the device could retain more than 80% of its initial efficiency after 100 h of continuous thermal stress. The positive effect of the P3HT layer could be attributed to: (i) polymeric barrier against diffusion of mobile ions of perovskite layer to the back contact [59]; (ii) passivation of perovskite surface against polar solvents of the CuSCN precursor; (iii) reduction in the metal diffusion [59,60]. In particular, the robust hydrophobic nature of P3HT [61], achieved reasonable passivation against atmospheric water and oxygen.
Energies 2020, 13, x FOR PEER REVIEW 7 of 12 heating at 85 °C [58], which could be accelerated by Cu(II) ions through a reaction with halide ions and form copper halide defects. Furthermore, due to the constraints on the CuSCN precursor solvent (requirement of polar solvents), the interaction of such solvents with perovskite surface during deposition of HTM leads to increased defects on the perovskite layer and accelerates the thermal decomposition. On the other hand, by using the P3HT/CuSCN bilayer structure as the HTL, the device could retain more than 80% of its initial efficiency after 100 h of continuous thermal stress. The positive effect of the P3HT layer could be attributed to: (i) polymeric barrier against diffusion of mobile ions of perovskite layer to the back contact [59]; (ii) passivation of perovskite surface against polar solvents of the CuSCN precursor; (iii) reduction in the metal diffusion [59,60]. In particular, the robust hydrophobic nature of P3HT [61], achieved reasonable passivation against atmospheric water and oxygen. To study the intrinsic mechanism ruling the performances of PSCs made with different HTLs, we performed a series of experiments to identify charge recombination and transfer. Regarding the evaluation of the recombination process in the fabricated devices (CuSCN and P3HT/CuSCN as HTL), we used J-V measurement under dark conditions and corresponding Tafel plots as shown in Figure 6a. Generally, the leakage current of the P3HT/CuSCN was one order lower than that for the device with CuSCN only. Although both devices showed almost the same shunt resistance and recombination in the depletion layer, the main difference arose from recombination in the neutral layer, which was mainly governed by trap-assisted recombination mechanisms. To study the intrinsic mechanism ruling the performances of PSCs made with different HTLs, we performed a series of experiments to identify charge recombination and transfer. Regarding the evaluation of the recombination process in the fabricated devices (CuSCN and P3HT/CuSCN as HTL), we used J-V measurement under dark conditions and corresponding Tafel plots as shown in Figure 6A. Generally, the leakage current of the P3HT/CuSCN was one order lower than that for the device with CuSCN only. Although both devices showed almost the same shunt resistance and recombination in the depletion layer, the main difference arose from recombination in the neutral layer, which was mainly governed by trap-assisted recombination mechanisms.
When the applied bias reached~0.8 V, the device containing P3HT/CuSCN shows a sharp trap filling behavior that evidences very high trap-assisted recombination resistance. Furthermore, despite the addition of an extra (P3HT) layer, there was no sign of increment in the depletion layer's recombination, which could be attributed to the intrinsic fast carrier mobility of P3HT and better coverage of the perovskite interface with P3HT.
The EIS was also carried out under dark conditions at different bias potentials to better define the electron transport/recombination process within the PSC. The semicircle in the Nyquist plots ( Figure S4) was related to the impedance of the free charge carriers' recombination at the perovskite/HTM layer (R rec ). The highest R rec indicated the lowest recombination process because of the longer free charge carriers' lifetime and their proper collection [55]. Figure 6B evidences that the recombination resistance significantly increased in the presence of a P3HT/CuSCN/Au-based PSCs, in excess of two orders of magnitude. This improvement could be associated with the uniform coating of the P3HT/CuSCN layer as well as the supporting role of the P3HT layer in blocking the diffusion of ions.
Energies 2020, 13, 2059 8 of 12 In order to calculate the lifetimes of the free charge carriers, constant phase elements (CPE) were extracted from the Nyquist curves ( Figure 6). The highest lifetime was obtained for P3HT/CuSCN/Au-based under intermediate voltages, compared with those cells having different arrangements. This is in full agreement with the data obtained from dark J-V and EIS analysis ( Figure 6). When the applied bias reached ~0.8 V, the device containing P3HT/CuSCN shows a sharp trap filling behavior that evidences very high trap-assisted recombination resistance. Furthermore, despite the addition of an extra (P3HT) layer, there was no sign of increment in the depletion layer's recombination, which could be attributed to the intrinsic fast carrier mobility of P3HT and better coverage of the perovskite interface with P3HT.
The EIS was also carried out under dark conditions at different bias potentials to better define the electron transport/recombination process within the PSC. The semicircle in the Nyquist plots ( Figure S4) was related to the impedance of the free charge carriers' recombination at the perovskite/HTM layer (Rrec). The highest Rrec indicated the lowest recombination process because of the longer free charge carriers' lifetime and their proper collection [55]. Figure 6b evidences that the recombination resistance significantly increased in the presence of a P3HT/CuSCN/Au-based PSCs, in excess of two orders of magnitude. This improvement could be associated with the uniform coating of the P3HT/CuSCN layer as well as the supporting role of the P3HT layer in blocking the diffusion of ions.
In order to calculate the lifetimes of the free charge carriers, constant phase elements (CPE) were extracted from the Nyquist curves ( Figure 6). The highest lifetime was obtained for P3HT/CuSCN/Au-based under intermediate voltages, compared with those cells having different arrangements. This is in full agreement with the data obtained from dark J-V and EIS analysis ( Figure 6).

Conclusions
A hysteresis-free planar n-i-p perovskite solar cell with a PCE of ~15.3% was developed with a

Conclusions
A hysteresis-free planar n-i-p perovskite solar cell with a PCE of~15.3% was developed with a P3HT/CuSCN bilayer as an HTL. The designed structure exhibited better performance by enhancing the electron transfer rate and reducing the recombination processes with respect to the typical CuSCN layer. Therefore, the present work can be considered as an effective and promising way to realize hysteresis-free SnO 2 -based planar perovskite solar cells through the use of P3HT/CuSCN as a cost-effective and suitable HTL alternative to conventional materials, such as Spiro-OMeTAD. It was further proved that the long-term stability of PSCs can be significantly increased by using P3HT as an interlayer between the perovskite and CuSCN. By employing this bilayer strategy, PSCs can maintain more than 80% of their initial PCE after 100 h under the continuous thermal stress conditions at 85 • C. The performance and durability of the PSCs were well improved and the hysteresis effect was prevented by restraining the perovskite ions, metals and halides migration, emphasizing the pivotal role that can be played by the interface in the perovskite-additive-HTM stack.
Supplementary Materials: The following are available online at http://www.mdpi.com/1996-1073/13/8/2059/s1, Figure S1: Photovoltaic parameters of SnO 2 /Perovskite/CuSCN/Au based devices (with single and double spin coating of 25 mg mL −1 of CuSCN in dipropyl sulfide).The first two sets are related to single spin deposition while the set consider a double spin coating CuSCN deposition PSC with an active area of 0.09 cm 2 , Figure S2