Mitigating the Efficiency Loss of Organic Photovoltaic Cells Using Phosphomolybdic Acid‐Doped PEDOT:PSS as Hole Transporting Layer

A phosphomolybdic acid (PMA)‐doped poly(3,4‐ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) hole transporting layer (HTL) is developed to effectively optimize the anode interface of inverted organic photovoltaic (OPV) devices. Since a deep‐lying work function of HTL is in favor of reducing the interfacial barrier for hole transporting, superior power conversion efficiencies (PCE) of 9.22% and 10.8% are achieved in fullerene‐based and nonfullerene‐based devices in inverted architecture. Thus‐prepared OPV devices also exhibit excellent photostability and retain 85% of the initial PCE after 1000 h of irradiation under 100 mW cm−2 irradiation. The overall result suggests the practicability and mitigation of efficiency loss when replacing the conventional PEDOT:PSS with PMA‐doped PEDOT:PSS as HTL to fabricate highly efficient, inverted OPV devices.


Introduction
Organic photovoltaic (OPV) is not just a solar energy technology. It is absolutely a groundbreaking energy harvesting technology with unique features, including mechanical flexibility, lightness, semitransparency, ease of large-area fabrication, and unlimited freedom of design. [1,2] Recently, the power conversion efficiency (PCE) of OPV cells over 18% has been reported. [3][4][5][6] Among these innovations, nonfullerene acceptors (NFAs) show a significant flexibility in molecular design and indubitably contribution in realizing a high PCE; such breakthrough also indicates the bright future of OPV technology. [7][8][9][10] However, the key factor for OPV commercialization is that the materials for production of OPV modules must be inexpensive and large-area processable; presently conjugated polymers and [6,6]-phenyl C 61 butyric acid methyl ester (PC 61 BM) bulk-heterojunction (BHJ) are still the mainstream systems for module manufacturing in industry due to their excellent solubility in halogenfree solvent and cost effectiveness. [11,12] Despite the fact that the overall PCE is lower in the system of fullerene-based BHJ layers, the PC 61 BM still shows the lowest synthetic complexity (SC) index compared to NFAs, and the fullerene-based BHJ system also displays a competitive figure of merit (SC/PCE ratio) than others. [13,14] OPVs composed of conjugated polymer:fullerene BHJ processed from green solvents have now achieved remarkable scalability, [2] values unthinkable only few years ago. These results also reflect the huge market value of OPV technology in the green energy landscape.
During the evolution of OPV technology from lab scale to industrial manufacturing, the scalability and processability of OPV for module production must be improved to mitigate the huge loss of PCE from a single cell to a large-area module. [12][13][14] Given that most of the inverted devices were fabricated with thermally evaporated molybdenum oxide (MoO 3 ) as the hole transporting layer (HTL) to reach high efficiency, [12,14,15] the loss of PCEs for large-area modules and the final performance of the device when using solutionprocessed HTL for the cost-effective coating process remain unanswered. [16,17] Among the numerous solution-processable HTLs, the polythiophene derivative, poly(3,4-ethylenedioxythiophene) (PEDOT), has a high carrier transporting ability and excellent environmental stability. Water-dispersible PEDOT colloid has been extensively used as a HTL in organic optoelectronic devices because of its ease of process and excellent carrier transporting ability. [18][19][20] PEDOT-based colloid can be composed of various PEDOT and water-soluble polymer segments, for example, polystyrenesulfonate (PSS) or perfluorinated ionomer (PFI), which results in different electronic characteristics of the resulting films. Careful adjustment of the colloid conformation may further tune the conductivity and energy level for different applications; [21] for example, PFI segments contain the fluorocarbon backbone and perfluorovinyl ether branch with sulfonic acid groups. PFI has been adopted to form a complex with PEDOT due to its highly electronegative fluorine, which results in a deep-lying highest occupied molecular orbital (HOMO) level. [18,22] An inverted, NFA-based OPV with solution-processed PEDOT:PFI HTL has been previously presented, with a PCE reaching 15% at the cell level, and a minimodule has also achieved a PCE of 10%. [14] DOI: 10.1002/aesr.202300006 A phosphomolybdic acid (PMA)-doped poly (3,4-ethylenedioxythiophene): poly (styrenesulfonate) (PEDOT:PSS) hole transporting layer (HTL) is developed to effectively optimize the anode interface of inverted organic photovoltaic (OPV) devices. Since a deep-lying work function of HTL is in favor of reducing the interfacial barrier for hole transporting, superior power conversion efficiencies (PCE) of 9.22% and 10.8% are achieved in fullerene-based and nonfullerenebased devices in inverted architecture. Thus-prepared OPV devices also exhibit excellent photostability and retain 85% of the initial PCE after 1000 h of irradiation under 100 mW cm À2 irradiation. The overall result suggests the practicability and mitigation of efficiency loss when replacing the conventional PEDOT: PSS with PMA-doped PEDOT:PSS as HTL to fabricate highly efficient, inverted OPV devices.
In addition, doping with transition metal oxide is also a common approach to modifying the energy level of the PEDOT-based HTL. Tungsten oxide or vanadium oxide-doped PEDOT:PSS not only shows a good ink dispersity but also modifies the energy level to align well with the HOMO level of BHJ layer, [23][24][25] which is also beneficial for improving the PCE of OPV cells. Among a variety of transition metal oxides that could be doped in PEDOT: PSS, polyoxometalate (e.g., phosphomolybdic acid, PMA) is recognized as one of the effective dopants to mix with PEDOT:PSS to achieve a deep-lying HOMO level with a BHJ layer in OPVs. PMA is a stable and soluble oxide cluster with desirable features, including excellent thermal stability and good solubility in many polar solvents. [26,27] Ji et al. reported a novel organic-inorganic composite of PMA and PEDOT:PSS that features good wettability with BHJ layer. Thus-prepared HTL can be easily deposited on a hydrophobic surface to form a thin film via solution coating to achieve a highly efficient OPV. [28] Herein, we report a simple method to optimize the energy level of the PEDOT:PSS without complicated synthesis. As shown in Figure 1a, by adding a tiny amount of PMA into a commercially available PEDOT:PSS solution, a well-dispersed PMA: PEDOT:PSS colloid solution could be obtained. Importantly, the energy level of the PMA:PEDOT:PSS thin film could be readily engineered by doping the PMA content. As a result, the PMA: PEDOT:PSS significantly outperformed the PEDOT:PSS in inverted OPV devices as a HTL, which would benefit the commercialization of highly efficient OPV technology with all solution processes.

Results and Discussion
In this work, the investigations were performed to reveal the effects of PMA-doped PEDOT:PSS interlayers in inverted OPVs. Figure 1b illustrates the device structure, with a BHJ system based on commercially available PV2000 and [6,6]-phenyl C 71 butyric acid methyl ester (PC 71 BM) blending, [12,[29][30][31] which is a representative material to verify the feasibility of novel HTLs. Herein, the performance of inverted OPV with three different HTLs, thermally evaporated MoO 3 , conventional PEDOT:PSS, and PMA:PEDOT:PSS, were evaluated. Figure 1c shows the plotted current density ( J) versus voltage (V ) curves for the devices measured under AM1.5 G solar irradiation at an intensity of 100 mW cm À2 . In the typically MoO 3 -based device, an opencircuit voltage (V OC ) of 0.81 V, a short-circuit current density ( J SC ) of 16.7 mA cm À2 , and a fill factor (FF) of 72.2% could be achieved, giving a PCE of 9.77%. For solution-processed HTL, here a commercially available, surfactant-added PEDOT:PSS, HTL solar, was used initially, and a lower PCE of 8.15% was obtained, with a V OC of 0.75 V, a J SC of 15.7 mA cm À2 , and FF of 69.2%. This result is consistent with the previous study, where poor device performance was attributed to the energy-level misalignment between PEDOT:PSS and PV2000. [18,23] Consequently, various concentrations (0.01, 0.05, 0.10, 0.20, 0.50, and 1.00 mg mL À1 ) of the PMA were added into PEDOT:PSS solution to obtain PMA:PEDOT:PSS complex. After doping the PMA, we found that the viscosity of PEDOT: PSS solution significantly increased from 3.2 to 13.2 cps as the PMA content increased (as shown in Table S1, Supporting Information). A minor gelation phenomenon was also observed within the PMA:PEDOT:PSS solution when the PMA concentration was higher than 1.00 mg mL À1 . Since PMA is a much stronger acid than PSS, it can donate protons to the PSS segment and reduce the Coulomb interaction between PSS and PEDOT, which results in gel-like liquid formation, which might lead to ink with a short shelf life. The appearance of gel-like solution is shown in Figure S1, Supporting Information. www.advancedsciencenews.com www.advenergysustres.com Some aggregations were observed when the PMA was added into PEDOT:PSS solution for 3 h. To avoid any issue caused by stability uncertainty, the mixed solution is used within 1 h after preparation. As for the film-forming quality, contact angle measurement shows that the surface energy of PEDOT:PSS solution changes with the addition of PMA. When the PMA and PEDOT:PSS solutions were tested on a hydrophobic PV2000:PC 71 BM surface, a significant increase in the contact angle was found ( Figure S2, Supporting Information). The data indicates that the contact angle increased from 9.3°to 22.8°as the PMA concentration increased, because PMA leads to a hydrophilic solution. Given that the surfactant already exists in PEDOT:PSS solution, the PMA and PEDOT:PSS solution can still be easily coated onto the hydrophobic PV2000:PC 71 BM surface from its water-based solution. No obvious dewetting issue was observed during coating, suggesting that the film-forming quality is good.
The surface topographies of the PEDOT:PSS-based HTLs were also examined by atomic force microscopy (AFM). As shown in Figure S3a, Supporting Information, the HTLs were deposited onto the PV2000:PC 71 BM layer by spin coating, and the AFM image suggests a pristine PEDOT:PSS surface with a root-mean-square roughness (R RMS ) of 1.14 nm ( Figure S3b, Supporting Information). As the pristine PEDOT:PSS was replaced by PMA-doped PEDOT:PSS, the R RMS values were found to be slightly increased to 1.16-1.27 nm ( Figure S3c-h, Supporting Information). The result shows that there is no obvious phase separation or aggregation in PMA:PEDOT:PSS complex, implying that PMA-doped PEDOT:PSS is in favor of forming a HTL for device fabrication.
By replacing the PEDOT:PSS with PMA:PEDOT:PSS, generally a higher PCE can be achieved, which mainly contributes to the enhancement of V OC and FF. As shown in Table 1, Figure 2, and S4, Supporting Information, (original J-V curves of devices), V OC monotonically increases from %0.76 V to %0.80 V as the PMA concentration increases from 0.01 to 1.00 mg mL À1 , with an improved FF from 68% to 72% as well. Notably, the PMA doping also influences the viscosity of PMA:PEDOT:PSS ink (Table S1, Supporting Information). With an identical coating parameter (3000 rpm for 60 s), the thickness of PMA:PEDOT: PSS layer increased from 65 to 152 nm when increasing the concentration of PMA. Given that series resistance is an important parameter in determining J SC of solar cell, J SC was also found to be 13.5 mA cm À2 as PMA concentration increased to 1.00 mg mL À1 , and the low J SC might be attributed to thicker HTL (152 nm thick, as shown in Table 1). The resistance and resistivity analyses of each HTL are shown in Figure S5 and Table S2, Supporting Information. Though the resistivity of PMA:PEDOT:PSS thin film reduced with increasing PMA, showing the acidic doping effect, the result obviously exhibits that the resistance of HTL increased from 6.68 to 10.8 Ω when increasing the thickness from 77 to 155 nm (as shown in Table S2, Supporting Information), which is the cause of low J SC for the device with 152 nm-thick HTL.
Eventually, the device composed of PMA:PEDOT:PSS with 0.2 mg mL À1 of PMA achieved a PCE of 9.01% (Table 1), with V OC , J SC , and FF of 0.80 V, 15.6 mA cm À2 , and 72.2%, respectively. A champion device with a V OC of 0.81 V, a J SC of 15.3 mA cm À2 , an FF of 74.4%, and a highest PCE of 9.22% was also realized after optimizing the thickness of PV2000: PC 71 BM BHJ layer from 250 to 220 nm and using PEDOT: PSS with 0.2 mg mL À1 of PMA. The device structure and J-V curve are shown in Figure 1b,c, and this champion device only gave a 5.6% difference in the PCE compared to that of the device composed of MoO 3 layer (PCE 9.77%), implying PMA doped-PEDOT:PSS is an efficiently solution processable HTL which is suitable for OPV production.
A photoelectron yield spectrometer was used to measure the ionization potential of the PEDOT:PSS and PMA:PEDOT:PSS films. Figure 3a plots the square root of the counting rate (CR) as a function of the photon energy and the photoemission threshold energy, which is also called the work function and can be determined from the crossing point of the background and the yield line. The work function of PEDOT:PSS was found to be À5.0 eV. When the content of PMA increased, the result indicated that an increased work function was achieved with the addition of PMA in the PEDOT:PSS solution, and the work functions also increased from 5.0 to 5.3 eV, which leads to barrier reduction for easier charge collection.
The HTL comprising PMA exhibits a deeper-lying energy level, which could explain the superior PCE than the devices using PEDOT:PSS-based HTL (Figure 3b). [27,28] The result suggests that the barrier height between the HTL and PV2000 strongly influences the device performance. PMA could mingle into PEDOT:PSS to effectively obtain a HTL with deeply lying HOMO level to use for preparing inverted OPV devices. [14] To evaluate the stability of PMA:PEDOT:PSS, the devices prepared with PEDOT:PSS or PMA:PEDOT:PSS were first  The light-soaking test was conducted under a metal halide lamp with an intensity of 100 mW cm À2 at 50°C. [32] As shown in Figure 4, V OC remained almost unchanged during light soak for 1000 h, and both devices showed slightly enhanced J SC at the beginning of the aging stage and yielded a higher PCE. Since the light-soaking process is expected to modify the energy level of electron-transporting materials, the enhancement of J SC can be attributed to the light activation effect of ZnO. [33] Subsequently, J SC and FF were mainly affected by light soaking, and the degradation might be attributed to the photodimerization of fullerene derivatives. [34] Eventually, the PCE of device composed of PMA:PEDOT:PSS  www.advancedsciencenews.com www.advenergysustres.com HTL can remain at 85% of the initial value after 1000 h of light soaking, which is slightly better than that of the device composed of PEDOT:PSS. Most of the NFA material system requires an extremely deep HOMO level of donor material that is more negative than À5.5 eV to maximize the built-in voltage and V OC of OPV devices. [8][9][10] However, this strategy hence forms a large energy barrier between the HOMO of the donor material and the work function of conventional PEDOT:PSS, which results in the poor electrical performance of inverted devices when using PEDOT:PSS as HTL. [18,23] The PMA:PEDOT:PSS was also used for testing the feasibility in NFA material system. In this work, a specific polymer, TPD-3 F, [23,35] and a typical NFA, IT-4 F, were combined as BHJ active layer to study the influence of barrier height between active layer and HTL in the inverted device architecture. Herein, TPD-3 F is an electron donor which has a very deep HOMO level, [23,35] and its detailed chemical structure is shown in Figure S7, Supporting Information. The detailed energy-level diagram of each material in the device is also shown in Figure 5a (Figure 5b), which boosts the PCE significantly from 6.30% (with PEDOT:PSS) to 10.8% (with PMA:PEDOT:PSS), suggesting that PMA-doped conducting polymer colloid is a promising HTL material for developing OPVs by solutioncoating process.

Conclusion
An approach for preparing solution-processed HTL with deeplying energy level was demonstrated. After careful optimization of the PMA concentration in PEDOT:PSS solution, a PMA-doped PEDOT:PSS layer can be easily spin coated onto the BHJ layer of inverted OPV, which enables optimizations of the energy-level alignment at the interface between BHJ layer and HTL. Devices based on PV2000:PC 71 BM and TPD-3 F:IT-4 F BHJ layers both achieved a boosted performance compared to the performance of devices with a conventional PEDOT:PSS, and the device also exhibits excellent photostability and retains 85% of the original PCE after 1000 h of irradiation under a 100 mW cm À2 metal halide lamp at 50°C. The result represents a breakthrough in HTL development for fabricating solutionprocessed inverted OPVs, indicating the feasibility of successful industrial production.

Experimental Section
Materials: The PV2000, TPD-3 F were synthesized by Raynergy Tek Incorporation. The PC 71 BM and PEDOT:PSS Clevios HTL Solar were purchased from SES Research and Heraeus Deutschland GmbH, respectively. HTL Solar is a commercially available PEDOT:PSS formulation with a surfactant to improve the wettability on hydrophobic surfaces, which is able to deposit directly on the photoactive layer. ITO-coated glass substrates (11 Ω sq À1 ) were purchased from AimCore Technology and patterned by Global Tech International. The other chemicals used in this article were purchased from the Aldrich Chemical Co. and used without further purification.
For PMA:PEDOT:PSS ink preparation, the PMA (0.1, 0.5, 1.0, 2.0, 5.0, and 10 mg) was dissolved in 1 mL of 2-propanol by stirring at room temperature in the dark for 60 min. 1 mL of the PMA solution was then added drop wise into a 20 mL vial containing 9 mL of the PEDOT:PSS solution (the ratio of the PMA solution to the PEDOT:PSS solution was 1:9) under a moderate stir speed, with continuous stirring at room temperature for an additional 5 min. Eventually, PMA:PEDOT:PSS solutions with different PMA concentrations (0.01, 0.05, 0.10, 0.20, 0.50, and 1.00 mg mL À1 ) were obtained, and the solutions were filtered with a 5 μm-sized filter before spin coating.
Device Fabrication: The OPV devices were fabricated by spin coating the ZnO precursor solution (consisting of diethylzinc/1.1 M in toluene dissolved in anhydrous tetrahydrofuran) on top of the precleaned ITO glass and then heated at 120°C for 10 min. The ZnO-coated ITO samples were transferred into a nitrogen-filled glove box. PV2000:PC 71 BM or TPD-3 F:IT-4 F with D:A ratio = 1:1.5 were dissolved in o-xylene, and the blend was spun coated on top of the ZnO to form a BHJ layer. Subsequently, the samples were annealed at 120°C in a glove box for 5 min. Finally, HTLs (MoO 3 , PEDOT:PSS, or PMA:PEDOT:PSS) and 100 nm-thick Ag electrode were deposited by thermal evaporation (for MoO 3 ) and spin coating (PEDOT:PSS or PMA:PEDOT:PSS). The active area of the cell was defined by the shadow mask with a area of 0.054 cm 2 . For solar cell measurement, an aperture mask with an area of 0.04 cm 2 was used to define the illumination area. For encapsulated devices prepared for photostability testing, the devices were encapsulated by cavity glass with UV-curable sealant (SAES, ZeoGlue-G) and the sealant was cured by a 400 W metal halide lamp for 30 s.
Characterization: J-V characteristics of OPV cell were measured using a Keithley 2400 source measurement unit under a simulated AM1.5 G spectrum with a 150 W solar simulator (SAN-EI XES-40S1), and the 100 mW cm À2 light intensity was calibrated using a KG5-covered Si-based reference cell. The photostability test was conducted in a customized oven with a metal halide lamp with a 100 mW cm À2 intensity at 50°C. The devices were light soaked under an open-circuit condition for a certain period. The device after light soaking was measured under a 100 mW cm À2 metal halide irradiation. The film thickness of each layer was determined by a KLA Tencor P-6 profilometer. All the AFM images were captured in tapping mode (Bruker Dimension Icon). The work functions of the materials were measured using an AC2 photoelectron spectrometer (Riken Keiki Co.), and the PEDOT:PSS and PMA:PEDOT:PSS films were prepared on the glass substrates for measurement. The resistance and resistivity of HTL were measured by ITO/HTL/Ag stacking. The current-voltage characteristics of samples were measured using a Keithley 2400 in the dark, and the resistivity (ρ) can be written as: where A is the area of active area (0.054 cm 2 ), l is the thickness of HTL, and R is the resistance of ITO/HTL/Ag stacking, which can be calculated from the current-voltage curves (1/R = slope). The viscosity of HTL solution was collected by Brookfield V1 voscosmeter, and the sameples were measured at 25°C with a spin speed of 100 rpm.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.