Investigation of various commercial PEDOT:PSS (poly(3,4-ethylenedioxythiophene)polystyrene sulfonate) as a hole transport layer in lead iodide-based inverted planar perovskite solar cells

Inverted-type perovskite solar cells have drawn remarkable attention due to solution-processable, straightforward configuration, low-cost processing, and manufacturing at very high throughput, even on top of flexible materials. The hole transport material (HTM) plays a vital role to achieve high performance in inverted type of perovskite solar cells. Herein, we report on the effect of different commercial PEDOT:PSS such as PH 1000, PH 500, P VP AI, and P T2, on the performance of CH3NH3PbI3-based planar perovskite solar cells.


Introduction
As a result of expanding population and industrial growth, worldwide energy utilization has escalated, and then demands for renewable and viable energy sources became inevitable. Alternative resources should keep a balancing act between efficiency, expenditure, technology issues, and environmental effects. Using photovoltaic devices for gathering energy from sunlight offers a fresh perspective for producing energy globally. Due to high cost, hightechnology requirements such as high vacuum and temperature for the fabrication, and additionally the existence of toxic components, the use of first and second generation of solar cells is limited. Hence, the new third-generation of PVs with low-cost, high flexibility, low molecular weight, and high efficiency are demanded. For a decade, perovskite-based solar cells have attracted a great deal of attention because of its superb characteristics such as low-cost, long carrier diffusion length, high absorption coefficient, low-temperature processing, low recombination rate [1][2][3], high electron and hole mobility, tunable bandgap [4], and a rapid increase in the efficiency from 3.9 [5] to more than 25% [6] which is not common for a photovoltaic device to show extremely rapid development in such a short period. Perovskite solar cell fabrication can be classified into two: evaporation-based methods [7,8], and solution-based methods. Solution-based methods have three different approaches including the one-step deposition method [9,10], spray coating [11], and two-step deposition method [12]. Evaporation-based techniques guarantee a pin-hole-free perovskite growth with a moderate uniformity of grains, but these processes require high vacuum technology. One-step deposition includes two forms: one with solvent annealing [13,14] and one with antisolvent washing treatment [15]. The latter method is usually carried out, where when the substrate is spinning a selected antisolvent such as toluene or chlorobenzene is dropped on the perovskite precursor layer and end with thermal annealing. This pratic method produces a conformal and dense perovskite layer, but the grain size is relatively small (around 200 nm). On the other hand, in terms of device architecture, perovskite solar cells can be classified as mesoporous and planar types. In the mesoporous type, the device requires high temperature (more than 450°C) through manufacturing progress which leads to increasing fabrication cost [16], while in planar structure the processing temperature is low. Due to the analogous efficiency of planar to mesoporous, employing planar device structure in PSCs is feasible. Different perovskite solar device architectures are shown in Fig. 1 [17]. In n-i-p (normal) structure, TiO 2 layer is used as an electron transport layer (ETL) and spiro-OMeTAD (2,2,7,7-tetrakis(N,N-di-p-methoxyphenylamine)-9,9spirobifluorene) as HTL, while in the p-i-n (inverted) structure commonly PEDOT:PSS is used as a hole transport layer (HTL) and [6,6]-Phenyl-C61-Butyric Acid Methyl Ester (PCBM) as ETL which need a buffer layer for reduction of interfacial charge recombination such as BCP [18][19][20][21] and LiF [22,23]. Applying high temperature for the sintering TiO 2 layer is inevitable as that increases the manufacturing costs and trapped charges at the interface between the perovskite and TiO 2 layer [16]. Moreover, spiro-OMeTAD has low stability and high cost [24]. Alternatively, by using PEDOT:PSS in p-i-n device structure, these problems can be eliminated. Both organic and inorganic materials can be employed as ETL and HTL by evolving from sensitized mesostructure to planar structure [25].
Generally, PEDOT:PSS is considered as an HTL in inverted-type perovskite solar cells due to high conductivity and transparency, low-temperature solution-processable HTL, and suitable for the fabrication of flexible devices [26,27]. Despite these advantages, PEDOT:PSS has some disadvantages such as hygroscopic and acidic nature, mismatching between work function energy level of PEDOT:PSS and HOMO energy level of perovskite which leads to less V oc [28][29][30], and consequently, unfavorable effects on the performance of PSCs. Various approaches are applied for modifying these properties such as solvent additive methods [31,32], UV treatment [33,34], doped PEDOT:PSS [35,36], and interfacial engineering [37,38]. P VP AI has been commonly utilized as a hole transport layer in inverted perovskite solar cells [39][40][41]. Shin et al. used P VP AI as a hole transport layer and by doping HTL with Triton X-100 (TX), the electronic structure of the doped HTL-based devices were improved [42]. Huang et al. employed DMSOtreated P VP AI and boosted PCE of the device about Fig. 1 Schematic illustration of different perovskite solar device architectures. a n-i-p mesoscopic, b n-i-p planar, c p-i-n planar, and d p-in mesoscopic [17] 37% [31]. Although the gel particles in PH 1000 solution is larger than the P VP AI [43], PH 1000 provides a better film-forming potential and is appropriate for large-area film production [44]. Double hole transport layers (PH 500, P VP AI) are used to minimize interfacial recombination in perovskite/polymer monolithic hybrid tandem solar cells by Chen et al. [45]. Despite all, there is no information related to P T2 PEDOT:PSS-employed perovskite solar cells and there is not a direct performance comparison between all these commercially available PEDOT:PSS for their use in inverted-type perovskite solar cells in the literature. In this work, the performance of inverted perovskite solar cells, with different commercial PEDOT:PSS hole transport layers, was studied in a configuration of ITO/PED-OT:PSS/CH 3 NH 3 PbI 3 /PCBM/BCP/Ag, where all the experiments were performed in ambient air under humidity 40-50% (Fig. 2).
The fabrication of the device with the structure ITO/PEDOT:PSS/ CH 3 NH 3 PbI 3 /PCBM/BCP/Ag was as follows: Indium tin oxide (ITO)-coated glasses were cut into small slides of 1.5 9 1.5 cm 2 . The etching process was used to pattern the ITO-coated glasses by applying a mixture of HCl:HNO 3 :H 2 O  In order to get high and fast crystallinity, toluene (60 lL) was dropped onto spinning substrates for antisolvent washing when the rotation speed reached 4000 rpm. Perovskite growth was completed in ambient air with heat treatment on a hot plate at 100°C for 20 min. Subsequently, PCBM film was formed by coating it at 1500 rpm for 30 s and dried at 90°C for 2 min; afterward, BCP layer was spincoated at 4000 rpm for 40 s [47][48][49]. Ag is used as the top electrode and is deposited by the thermal evaporation with a thickness of 110 nm (Fig. 3).

Characterization
UV-Vis spectroscopy was used to investigate the bandgap of the perovskite layer with a PG Instruments T80 spectrophotometer. Atomic force microscope (AFM) was performed using Hitachi 5100 N Incident photon-to-current efficiency (IPCE) measurements were conducted with New Port measurement system, including an optical system with a xenon lamp, a filter wheel, a mechanical chopper, and a monochromator (Acton SP150), and recorded by using a lock-in amplifier to extract theoretical maximum short-circuit current densities [50].  4 Results and discussion Figure 4 depicts the AFM images for PH 1000, PH 500, P VP AI, and P T2-based PEDOT:PSS films deposited onto ITO-coated glass substrates in two different areas of 2 9 2 lm 2 and 10 9 10 lm 2 . The mean roughness (S a ) values of PH 1000, PH 500, P T2, and P VP AI were obtained as 1.122 nm, 1.101 nm, 1.170 nm, and 1.407 nm, respectively, for 2 9 2 lm 2 . While for 10 9 10 lm 2 area the mean roughness (S a ) values of PH 1000, PH 500, P T2, and P VP AI were obtained as 2.504 nm, 2.300 nm, 2.644 nm, and 1.122 nm, respectively. This shows P VP AI layer is more uniform and homogeneous at large area. Although the difference between the roughness of different PEDOT:PSS layers is not noticeably significant, the surface of P VP AI film is smoother as compared to others. Since a more uniform and smoother film is favorable for charge transport enhancement, P VP AI-employed devices are expected to lead to a better photovoltaic performance. Furthermore, the effect of the surface morphology of PEDOT:PSS on the perovskite growth quality was investigated by scanning electron microscopy (Fig. 5). SEM images reveal that perovskite layer growth on the P VP AI was successfully achieved with larger grain sizes and homogeneous grain distribution without any pin holes. Increasing grain size and reducing grain boundaries in the perovskite film is known to be advantageous for high short-circuit current and fill factor values. On the other hand, in the case of PH 500, the perovskite film suffers from small grain size which consequently leads to a reduction in the performance of the device. The perovskite layer on PH 1000 presents a uniform and homogeneous film as P T2 layer. However, the grain sizes of the respective perovskite films on Ph1000 PEDOT:PSS layers are obviously smaller as compared to that of P VP AI PEDOT:PSS layers, which is directly linked to the lower J SC values of the PH 1000-employed PSCs. Insulating PSS is used as a dopant in conducting PEDOT for balancing counter ion [51]. Due to different electrical properties of PEDOT and PSS (Table 1.), different commercial PEDOT:PSS electrically differ from each other which leads to a variation in conductivity and wettability of the PEDOT:PSS films.    [56,57]. Perovskite film grown on the P VP AI sample illustrates higher and sharper (110) diffraction peak than the other types, demonstrating a considerable rise in crystallinity, which is consistent with SEM observation (Fig. 5). The full width at half maximum (FWHM) values of the (110) diffraction peak are extracted as 0.27, 0.29, 0.37, and 0.38 for P VP AI, P T2, PH100, and PH500, respectively. The narrower FWHM of the P VP AI could be due to the comparatively large perovskite crystalline size, as evidenced by the SEM images in Fig. 5.
The bandgap achieved from the absorption spectra of perovskite films on different HTL as depicted in Fig. 7a is around 1.60 eV which is in a good agreement with expected value for CH 3 NH 3 PbI 3 [58]. The Urbach energies (E u ) of different PEDOT: PSS layers are determined in order to analyze the electronic quality in the films. Figure 7b shows Urbach energy plots for various PEDOT:PSS HTLs. By using the following formula [59] log a ¼ log a 0 þ ht E U the Urbach energies are acquired as 229.41 meV, 183.39 meV, 182.28 meV, and 159.15 meV for PH 500, P T2, PH 1000, and P VP AI respectively. Diffusion length reduces as Urbach energy increases, which is expected since higher Urbach energy means higher recombination rates, which results in shorter diffusion length [60]. Due to lower Urbach energy of P VP AI, utilizing P VP AI as a HTL in PSCs helps to reduce the recombination losses. The smaller Urbach energy, the smaller degree of energy disorder, which leads to a better device performance [61]. Figure 8a  indicates the UV-Vis transmittance spectra (between 300 and 1100 nm) of different types of PEDOT: PSS, coated on ITO-coated glasses as substrates. Transmittance spectrum shows the percentage of the incident light passing through the sample without absorbing, reflecting, or scattering at different wavelengths.
As can be seen, the transmittance of PH 1000 is lower, while the PH 500, P T2, and P VP AI are similar. The higher transmittance means the more photons are reaching to the photoactive layer, which can increase the probability of charge generation. There is a clear transmittance increase around 400 nm for PH 500 and P T2 and a slight rise with a different pattern, for the P VP AI. Additionally, transmittances slightly increased after 800 nm for P VP AI.
A surface with lower wettability minimizes the number of sites for perovskite nucleation which spans to the larger crystal growth [62]. The resistivity of the PEDOT:PSS precursors is known to be different, and also following the PEDOT:PSS film formation by spinning and annealing, the resulted layers may exhibit different resistances. To determine the electrical conductivity of different types of PED-OT:PSS films, current-voltage (I-V) was characterized across the ITO/PEDOT:PSS/Ag structure (Fig. 8b). The current-voltage characteristics for the PEDOT:PSS layers display different slopes for each indicating various conductivities [63], which were calculated to be 13.6 9 10 -5 S cm -1 , 19.2 9 10 -5 S cm -1 , 20 9 10 -5 S cm -1 , and 24 9 10 -5 S cm -1 , for P VP AI, P T2, PH 500, and PH 1000, respectively, which is consistent with the literature [64]. Even though the high electrical conductivity of HTL is favorable for charge transport, higher J SC in P VP AIemployed PSCs can be attributed to the increased perovskite grain sizes and enhanced interface between perovskite layer and HTL which will provide an improvement in both FF and J SC of the devices [65][66][67]. On the other hand, the lower J SC values of PH 500-and P T2-employed PSCs might originate from the pin-hole-dominated perovskite growth. Figure 8c shows current density-voltage (J-V) curves of devices consisting of different PED-OT:PSS layers. The devices employing P VP AI PEDOT:PSS as a HTL demonstrated the best performance, while the device that PH 500 employed exhibited the lowest performance. The photovoltaic parameters are summarized in Table 2.
The P VP AI PEDOT:PSS-based PSC have an opencircuit voltage V oc of 925 mV, a short-circuit current density J sc of 19 mA cm -2 , and a fill factor FF of 0.64, resulting in a PCE of 11.2%. The highest V OC and FF values of the devices employing P VP AI can be explained with higher work function of P VP AI and enhanced interface properties concluded from SEM and AFM measurements. For PH 500-based PSC, J sc value is 13 mA cm -2 and V oc is found to be 900 mV, while the fill factor was obtained as 0.44, which resulted in a PCE of 5.1%. In the case of P T2-based PSCs, J sc was 16 mA cm -2 , and the FF was 0.55. PH 1000 PEDOT:PSS-based PSCs exhibited a PCE of 8.8% with the highest fill factor of 0.70, while Jsc was14 mA cm -2 which is lower than that of P VP AI and P T2 PEDOT:PSS-employed devices. For PH 1000 PEDOT:PSS-employed PSCs, V oc is found to be similar to that of PH 500 PEDOT:PSS-employed devices. Additionally, the series and shunt resistances are calculated and given in Table 2.
Since the series resistance is associated to charge carrier transport and contact resistance, the minimum R series value for P VP AI-employed device is understandable. On the other hand, R series increased to 25.0 X cm 2 for PH 500-employed devices and it has the highest value in as compared to the others, which leads to low photovoltaic performance decreasing the J sc s.
The lower J SC values of the PH 500-based devices may be attributed to the low shunt resistance with higher I 0 current values compared to the P VP AIbased devices, rather than the resistive nature of the PEDOT:PSS layers. Figure 9 depicts dark and illuminated J-V curves of the PSCs. Figure 10 describes the box charts graphics for the photovoltaic parameters of perovskite solar cells employing different types of PEDOT:PSS layers over 15 devices, including J SC , V OC , FF, and PCE. The steady photovoltaic performance indicates a good reproducibility. Figure 11a shows the incident photon-to-current conversion efficiency (IPCE) spectra and integrated current densities. The photocurrent onset at 800 nm was in a good agreement with the perovskite (CH 3-NH 3 PbI 3 ) bandgap. The devices employing P VP AIbased PEDOT:PSS layers as HTL presented higher quantum yields compared to the others, which is consistent with the PV results. The integrated photocurrent for P VP AI-based PSC was 18.5 mA cm -2 , while the values for PH 500, PH 1000, and P T2 were observed as 12.7 mA cm -2 , 13.8 mA cm -2 , and 15.8 mA cm -2 , respectively.
The results acquired from IPCE spectra were comparable with the ones achieved from current density-voltage measurements. Figure 11b presents the normalized power conversion energy (PCE) of different types of PEDOT:PSS-employed un-encapsulated devices, under ambient air (humidity 40-50%, temperature * 25°C). It can be seen that devices employing P VP AI and PH 1000 showed better stability over a period of time as compared with others.
The charge extraction efficiencies of the different type of PEDOT:PSS thin films were examined by steady-state photoluminescence (PL) experiment with perovskite films on top of them. Photocarriers are created in the perovskite absorbing layer during photoexcitation and disperse toward the interfaces. The photocarriers created near the interfaces are extracted prior to becoming a part of the PL signal resulting in PL quenching. As shown in Fig. 12, P VP AI demonstrated stronger PL quenching ability compared to the others, inferring that more effective charge extraction and transport probably occurrs at P VP AI/perovskite interface. The PL curves of the samples were symmetric, implying that no significant lattice distortion existed [68].

Conclusions
We investigated the photovoltaic performance of the methylammonium lead iodide-based inverted perovskite solar cells based on different types of PED-OT:PSS HTLs with commercial names of PH 1000, Fig. 11 a IPCE spectra and integrated photocurrent obtained from the corresponding IPCE spectra. b normalized PCE of different type of PEDOT:PSS layer employed solar cell devices under ambient condition Fig. 12 Steady-state PL curves with a ITO/HTL/perovskite/ PCBM/BCP configuration PH 500, P T2, and P VP AI precursors. The device with P VP AI PEDOT:PSS HTL exhibited the best photovoltaic performance of all, while the device using PH 500 presented the lowest photovoltaic performance as compared to the others. The main reason refers to large grains-dominated pin-hole-free growth of the perovskite layer on the top of the P VP AI layer which was confirmed by XRD analysis. Further, Urbach energies extracted using absorption spectra of the perovskite layers also indicated lowered disorder on the P VP AI-based HTLs. Moreover, photoluminescence measurements confirmed that the best charge transport was achieved in the case of P VP AI-type PEDOT: PSS utilization. Also, it was observed that the devices employing P VP AI and PH 1000 exhibit more resistance against degradation in the ambient environment over a short period of time.