Sulfonate-Conjugated Polyelectrolytes as Anode Interfacial Layers in Inverted Organic Solar Cells

Conjugated polymers with ionic pendant groups (CPEs) are receiving increasing attention as solution-processed interfacial materials for organic solar cells (OSCs). Various anionic CPEs have been successfully used, on top of ITO (Indium Tin Oxide) electrodes, as solution-processed anode interlayers (AILs) for conventional devices with direct geometry. However, the development of CPE AILs for OSC devices with inverted geometry is an important topic that still needs to be addressed. Here, we have designed three anionic CPEs bearing alkyl-potassium-sulfonate side chains. Their functional behavior as anode interlayers has been investigated in P3HT:PC61BM (poly(3-hexylthiophene): [6,6]-phenyl C61 butyric acid methyl ester) devices with an inverted geometry, using a hole collecting silver electrode evaporated on top. Our results reveal that to obtain effective anode modification, the CPEs’ conjugated backbone has to be tailored to grant self-doping and to have a good energy-level match with the photoactive layer. Furthermore, the sulfonate moieties not only ensure the solubility in polar orthogonal solvents, induce self-doping via a right choice of the conjugated backbone, but also play a role in the gaining of hole selectivity of the top silver electrode.


Introduction
The use of interfacial layer materials to improve the charge selectivity and minimize the energy barrier of electrodes plays a central role in promoting their performance and stability in organic electronic devices, such as organic solar cells (OSCs) [1][2][3]. In view of the technological need to attain fully solution-processed devices, most of the attention has been addressed in the search of efficient interfacial materials with good solubility in polar solvents, included water, which allows deposition from orthogonal solvents to the active layers [4][5][6]. Conjugated polyelectrolytes (CPEs), composed of a conjugated backbone with side-chains bearing ionic functional groups, have emerged as a promising class of interfacial materials with their proven ability to improving photovoltaic (PV) performances through solution processing [7][8][9][10]. CPEs combine several advantages including: solubility in aqueous/alcoholic solvents, orthogonality to organic solvents used for the deposition of the active layers, robust film formation, and chemical flexibility in tailoring both the conjugated backbone as well as the polar/ionic lateral functionalities. Most of the CPEs developed so far are effective in reducing the electrode work function [11][12][13] and might behave as electron transport layers [14,15]. Therefore there is a huge library of CPEs in the groups as P1 and P2 but with a different conjugated backbone based on a fluorenic unit, known to be un-fitted for self-doping [34,[38][39][40]. In the P3 copolymer, the side chains have been tailored to favor the interaction with the active layer by alternating alkyl-potassium sulfonate and non-polar alkyl side chains. By comparing the results obtained with the three CPEs we tried to elucidate the role that pendant sulfonate groups play in the AIL functionality of this class of polymers, demonstrating that sulfonate anionic CPEs could be a valid approach to obtain solution-processable anode interlayers for inverted OSCs devices.

Synthetic Methods
General information for synthesis: All glassware was oven-dried. Unless specifically mentioned, all chemicals are commercially available and were used as received. The dialysis membrane (MWCO: 3500-5000 Da) was purchased from Membrane Filtration Products Inc. 1 H NMR spectra were recorded at 600 MHz in D2O.

Synthetic Methods
General information for synthesis: All glassware was oven-dried. Unless specifically mentioned, all chemicals are commercially available and were used as received. The dialysis membrane (MWCO: 3500-5000 Da) was purchased from Membrane Filtration Products Inc. 1 H-NMR spectra were recorded at 600 MHz in 3.76 mmol, 1.0 equivalent), and tetrabutylammonium bromide (60.6 mg, 0.188 mmol, 0.05 equivalent), were dissolved in anhydrous DMSO (18.4 mL), and the solution was degassed by bubbling Ar for 5 min. 50% KOH in H 2 O (4.2 g) was added via syringe, followed by the addition of 1,4-butanesultone (924 µL, 9.02 mmol, 2.4 equivalent). After stirring at room temperature for 3 h, the reaction mixture was poured into acetone (100 mL) and the yellowish precipitate was collected by filtration and washed with acetone. The crude was used in the next step without further purification. 1 [34] (see Scheme S1 in Supplementary Materials).

Synthesis of 4-Potassium butanylsulfonate
]-dithiophene (CPDT, 300 mg, 1.68 mmol, 1.0 equivalent) and tetrabutylammonium bromide (27 mg, 0.084 mmol, 0.05 equivalent) were dissolved in anhydrous DMSO (8.2 mL), and the solution was degassed by bubbling with Ar for 5 min. 50% KOH in H 2 O (1.8 g) was added via syringe, followed by the addition of 1,4butanesultone (924 µL, 9.02 mmol, 1.2 equivalent). After stirring at room temperature for 3 h, the reaction mixture was poured into acetone (50 mL) and the yellowish precipitate was collected by filtration and washed with acetone. The crude was used in the next step without further purification. 1

Device Fabrication and Photovoltaic Characterization
Direct geometry devices fabrication: Solar cells were assembled with the conventional structure Glass/ITO/PEDOT:PSS or AIL/P3HT:PC 61 BM/Ag. Glass ITO (Kintec, Hong Kong) 15 Ω/sq substrates were mechanically cleaned with peeling tape and paper with acetone and then were washed in a sonic bath at 50 • C for 10 min sequentially with water, acetone, and isopropanol. After drying with compressed nitrogen flow, 10 min plasma treatment in the air was used to enhance the ITO wettability for the next deposition. PEDOT:PSS (Al VP 8030 from Heraus, Hanau, Germany) was filtered on a 0.45 µm nylon filter and spin-coated in the air at 2500 rpm for 50 s. P1 (5 mg/mL in H 2 O:MeOH 1:1), P2, and P3 (1 mg/mL in EtOH) were spin-coated at 2000 rpm and 4000 rpm for 60 s. Finally, the substrates were stored in a glovebox and annealed at 110 • C for 10 min. The device assembly was then performed in the glovebox. The active layer was composed of a blend dissolved at 1:0.8 wt/wt of P3HT:PC61BM solution in 1,2-dichlorobenzene at a total concentration of 27 mg/mL. The P3HT was purchased from Plextronics (Pittsburgh, PA, USA, Plexcore 0S2100, Mn: 62602; Mw: 119010, 99% regioregularity) and PC 61 BM (99.5 % purity) was purchased from Solenne BV. The solution was stirred for 12 h on a hotplate in a glovebox at 60 • C. The active layer was spin-coated from the warm solution at 1000 rpm for 60 s, which resulted in a thickness of 130 nm; then, the active layer film was slow dried under a glass petri dish for 1 h. Finally, a 100 nm-thick aluminum electrode was evaporated on the top of the device through a shadow mask under a pressure of 2 × 10 −6 mbar. The deposition rate was 0.5 nm/s. There were six devices on a single substrate, each with an active area of 6.1 mm 2 .
Inverted geometry devices fabrication: Solar cells were assembled with the conventional structure Glass/ITO/PEIE/active layer/AIL/Ag. Glass ITO (Kintec, Hong Kong) 15 Ω/sq substrates were mechanically cleaned with peeling tape and paper with acetone and then were washed in a sonic bath at 50 • C for 10 min sequentially with water, acetone, and isopropanol. After drying with compressed nitrogen flow, ITO was treated under UV light from a solar simulator for 20 s to remove O 2 species adsorbed on its surface. Polyethylenimine ethoxylated (PEIE) 0.4% w/w in 2-Metoxyethanole solution was spincoated in N 2 at 5000 rpm for 60 s. PEIE film was then washed with H 2 O to remove excess polymers (200 µL were deposited for 10 s on substrate and then removed at 4000 rpm for 60 s). The P3HT:PC 61 BM active layer was prepared upon blending the two components at a 1:0.8 weight ratio in 1,2-dichlorobenzene, at a total concentration of 27 mg/mL. The solution was stirred for 12 h on a hotplate in glovebox at 60 • C; the active layer was spin-coated from the warm solution at 1000 rpm for 60 s, which results in a thickness of 130 nm; the films were then covered with a glass petri dish to perform slow drying for an hour and then treated at 120 • C for 10 min. The PTB7-Th was purchased from Cal-Os (batch number 6) and PC 71 BM from Solenne BV (99.5% purity). The PTB7-Th:PC71BM active layer was prepared by dissolving the two components at a 1:1.5 weight ratio in chlorobenzene, with a solute concentration of 25 mg/mL. This solution was stirred overnight at 65 • C. Next, 10 min after adding 2% v/v of 4-methoxybenzaldehyde to the solution, the blend was spin coated at 1200 rpm, permitting the obtaining of films with thickness of 100 nm. The samples were left at room temperature for 10 min and then annealed for 20 min at 60 • C. The AILs were deposited on top of the active layers by dissolving P2 or P3 in ethanol at a concentration of 1 mg/mL, and 20 µL of this solution was dropped on a device rotating at 4000 rpm for 60 s. P1 was dissolved in H 2 O:Methanol (1:1) at a concentration of 5 mg/mL, and 20µL of this solution was dropped on a device rotating at 4000 rpm for 60 s. Finally, a 100 nm-thick silver electrode was evaporated on the top of the device through a shadow mask under a pressure of 2 × 10 −6 mbar. The deposition rate was 0.3 nm/s. A device without any AIL and a device with an evaporate MoOx AIL were prepared for comparison. The thickness of the MoOx layer was 10 nm, achieved with a deposition rate of 0.1 nm/s. There were six devices on a single substrate, each with an active area of 6.1 mm 2 .
Device characterization and measurements: The devices were characterized through current density-voltage and external quantum efficiency characterization. Current densityvoltage measurements were performed directly in the glovebox where the solar cells were assembled, with a Keithley 2602 source meter, under an AM 1.5G solar simulator (ABET 2000). The incident power, measured with a calibrated photodiode (Si cell + KG5 filter), was 100 mW/cm 2 . The EQE spectral responses were recorded by dispersing an Xe lamp through a monochromator, using an Si solar cell with a calibrated spectral response to measure the incident light power intensity at each wavelength. The devices were taken outside the glovebox for the EQE measurements, after mounting them on a sealed cell to avoid moisture and oxygen exposure. For the aging measurement, the devices were stored in air and measured in glovebox.

Results and Discussion
The chemical structures of the three copolymers P1, P2, and P3, are shown in Figure 1. The polymers were synthesized via Suzuki cross-coupling and their purifications were performed according to the literature. The selected experimental procedures were chosen according to the literature [41,42]; P1 was prepared according to the procedure described by Mai et al. [34] starting from the commercially available 2,1,3-Benzothiadiazole-4,7bis(boronic acid pinacol ester) and 2,6-dibromo-4,4-bis-potassium butanylsulfonate-4Hcyclopenta-[2,1-b;3,4-b']-dithiophene. The same procedure was applied to P2, starting from the newly synthesized 2,6-dibromo-4-potassium butanylsulfonate-4H-cyclopenta-[2,1-b;3,4-b']-dithiophene, while P3 was synthetized according to the literature [39,40] (synthesis details and characterizations are provided in the Supplementary Materials). GPC measurements confirmed the structure of the synthetized polymers, however they only provided an indicative estimation of the real value of the molecular weight due to solubility problems and structural difference with the polystyrene used as standard [43].
Cyclic voltammetry and optical spectroscopy were used to assess the electronic properties of the copolymers and to estimate their HOMO and LUMO energy levels. See Table 1 and the Supplementary Materials.  The HOMO and LUMO levels of P1, P2, and P3 were obtained from the onsets of the oxidation and reduction potential [44], respectively, and are reported in Table 1 together with UV−vis spectroscopic data. The deviation between the optical and electrochemical results are in quite satisfactory agreement, as previously described in literature [45].
The energy levels of P2 are similar to P1, consistently with the same conjugated backbone and lateral substituents. P3 shows a variation of the HOMO and LUMO which is coherent to its polyfluorene backbone [39,40]. Importantly, as previously reported [46][47][48][49] the functionalization with the polar side chains does not interfere with the conjugated π-orbitals of the polyfluorene backbone and thus leads to no modification in the position of the HOMO and LUMO levels.
The different concentration in the copolymers of the sulfonated groups and alkyl chains is a tool to obtain different hydrophilic/hydrophobic characteristics: P1 is the most hydrophilic owing to its two sulfonate groups per monomeric unit. The presence of only one polar side chain in P2 and the presence of two long alkyl chains in P3 are two approaches to reduce the CPEs hydrophilicity/polarity. This trend is confirmed by contact angle measurements on top of the P1, P2, and P3 films using glycerol as a polar solvent. As shown in Figure S10 in the Supplementary Materials, the contact angles were found to be respectively 36 • for P1, 49 • for P2, and 68 • for P3. Such reduction in the wettability to polar solvent going from P1 to P3 suggests that thanks to our design we were able to finely modulate the different hydrophilic/hydrophobic characteristics of the polymers and consequently the interaction with the a-polar active layer.
As a consequence, P1 has a good solubility in water but it is not soluble in pristine methanol or ethanol. P2 shows an overall poor solubility in water/alcohol solvents but its dispersion in ethanol at low concentrations, 1 mg/mL, leads to stable suspensions which are suitable for film deposition by spin casting. P3 can be also dissolved and deposited from ethanol at low concentrations (1 mg/mL). A good solubility in alcohol is particularly important with a view to preparing devices in an inverted configuration where a thin CPE film should be deposited on top of the hydrophobic active layer surface.

Self-Doping Behaviour
The self-doping behavior of P1, P2, and P3 was investigated by electron paramagnetic resonance (EPR) measurements in aqueous solutions, see Figure 2b. A strong EPR symmetric signal consistent with the presence of unpaired electrons is observed for both P1 and P2, with peak to peak linewidths and calculated g values respectively of 0.28 mT and 1.9955 for P1 and 0.25 mT and 1.99314 for P2, which are the typical signatures of polaron formation in conjugated polymers [34,36]. It is important to underline that the values of the integrals are comparable and therefore the signal of the P2 polymer is more intense than that of the P1 polymer. Moreover, as it is well known, EPR is an intrinsically quantitative technique, since the EPR signals are due to the number of the excited spins [50]. As a consequence, at the same temperature and experimental conditions, EPR signal heights and, even more so double integrals, are comparable. P1 and P2, with peak to peak linewidths and calculated g values respectively of 0.28 mT and 1.9955 for P1 and 0.25 mT and 1.99314 for P2, which are the typical signatures of polaron formation in conjugated polymers [34,36]. It is important to underline that the values of the integrals are comparable and therefore the signal of the P2 polymer is more intense than that of the P1 polymer. Moreover, as it is well known, EPR is an intrinsically quantitative technique, since the EPR signals are due to the number of the excited spins [50]. As a consequence, at the same temperature and experimental conditions, EPR signal heights and, even more so double integrals, are comparable. With regard to P3, no evidence of polaronic features is detected in the EPR spectrum. Polarons formation was previously reported for a P1-type CPE in aqueous solution and it was explained by a self-doping mechanism [33,34] in presence of a proton source like water [51]. Similarly, the polaronic features in the P2 EPR spectrum indicate that self-doping in solution is occurring also when the PCPDTBT backbone is bearing only one alkylsulfonate lateral group per monomeric unit. On the other side, the absence of polarons in the P3 EPR spectrum is consistent with the fact that self-doping can only take place in CPEs with a relatively low ionization potential, as is the case for P1 and P2, but not for polyfluorene-based CPEs such as P3 which have a high ionization potential/deep HOMO energy level. The UV/Vis-NIR absorption spectra of P1, P2, and P3 films are reported in Figure 3a. The spectrum of P1 exhibits two main bands centered at 405 nm and 660 nm that are peculiar for polymers with conjugated ciclopentadithiophene-benzothiadiazole (PCPDTBT) backbones and a weak and broad band peaked at 1140 nm which could be ascribed to the formation of polarons (radical cations) which are stabilized by the pendant sulfonate groups [34]. As shown in Figure 3a, a similar band is observed upon doping a pristine PCPDTBT film through exposure to I2 vapors. This confirms a polaronic transition assignment for the NIR absorption feature of the P1 film With regard to P3, no evidence of polaronic features is detected in the EPR spectrum. Polarons formation was previously reported for a P1-type CPE in aqueous solution and it was explained by a self-doping mechanism [33,34] in presence of a proton source like water [51]. Similarly, the polaronic features in the P2 EPR spectrum indicate that selfdoping in solution is occurring also when the PCPDTBT backbone is bearing only one alkylsulfonate lateral group per monomeric unit. On the other side, the absence of polarons in the P3 EPR spectrum is consistent with the fact that self-doping can only take place in CPEs with a relatively low ionization potential, as is the case for P1 and P2, but not for polyfluorene-based CPEs such as P3 which have a high ionization potential/deep HOMO energy level. The UV/Vis-NIR absorption spectra of P1, P2, and P3 films are reported in Figure 3a. The spectrum of P1 exhibits two main bands centered at 405 nm and 660 nm that are peculiar for polymers with conjugated ciclopentadithiophene-benzothiadiazole (PCPDTBT) backbones and a weak and broad band peaked at 1140 nm which could be ascribed to the formation of polarons (radical cations) which are stabilized by the pendant sulfonate groups [34]. As shown in Figure 3a, a similar band is observed upon doping a pristine PCPDTBT film through exposure to I 2 vapors. This confirms a polaronic transition assignment for the NIR absorption feature of the P1 film absorption spectrum. The P2 spectrum exhibits two main absorption bands, peaking respectively at 410 and 715 nm, which are similarly assigned to the PCPDTBT backbone transitions. The P2 D-A (donoracceptor) band at 715 nm has a broad tail in the NIR region which could arise from light scattering effects. A scattering tail is also observed in the absorption spectrum of a P2 dilute solution, indicating that P2 is partially aggregated even in diluted solution (0.1 mg/mL in ethanol or water, see Figure S2 in Supplementary Materials). The granular morphology in the solid state, evident from the atomic force microscopy (AFM) images reported in Figure S3, confirms the presence of aggregation and it is consistent to significant light scattering from P2 films. As a matter of fact, the P2 polarons spectral features, already evidenced by EPR spectra, are probably covered by the mentioned scattering, in the P2 film absorption spectrum. Consequently, we can reasonably suppose that, similar to the well-known P1, self-doping is maintained also in P2 films. The P3 film shows a π-π* transition peak at 365 nm as expected for a polyfluorene backbone. No polaronic features are detected, which is consistent to the previously discussed absence of self-doping in solution for this fluorene-based CPE.
polarons spectral features, already evidenced by EPR spectra, are probably covered by the mentioned scattering, in the P2 film absorption spectrum. Consequently, we can reasonably suppose that, similar to the well-known P1, self-doping is maintained also in P2 films. The P3 film shows a π-π* transition peak at 365 nm as expected for a polyfluorene backbone. No polaronic features are detected, which is consistent to the previously discussed absence of self-doping in solution for this fluorene-based CPE. The self-doping of CPEs films can be monitored by IR (Infra-Red) absorption spectroscopy. The absorption spectral signatures of polaronic charged states in conjugated polymers consist of both electronic transitions and IR vibrational modes, the so-called IRAV bands [52,53]. Figure 3b displays the IR absorption spectral pattern of P1 and P2, with a pristine and I2-doped PCPDTBT for comparison. The P1 and P2 IR spectra show two intense and broad peaks at 1035 and 1165 cm −1 . These bands are attributed to the symmetric and antisymmetric stretching modes of the SO3 − groups [54]. In P3, which has the same alkylsulfonate side chains, similar intense SO3 − stretching bands are observed ( Figure S13, Supplementary Materials). Both P1 and P2 polymers show a weak band at 1285 cm −1 , that is not observed in the P3 spectrum of Figure S3. This suggests that an alkylsulfonate vibrational mode cannot account for this feature. Interestingly, this weak band does not match the pristine PCPDTBT IR spectrum but instead has a close similarity to one of the IRAV bands growing up upon p-doping PCPDTBT with I2 vapors. On the basis of these considerations, we suggest that the band observed at 1285 cm −1 in the IR The self-doping of CPEs films can be monitored by IR (Infra-Red) absorption spectroscopy. The absorption spectral signatures of polaronic charged states in conjugated polymers consist of both electronic transitions and IR vibrational modes, the so-called IRAV bands [52,53]. Figure 3b displays the IR absorption spectral pattern of P1 and P2, with a pristine and I 2 -doped PCPDTBT for comparison. The P1 and P2 IR spectra show two intense and broad peaks at 1035 and 1165 cm −1 . These bands are attributed to the symmetric and antisymmetric stretching modes of the SO 3 − groups [54]. In P3, which has the same alkylsulfonate side chains, similar intense SO 3 − stretching bands are observed ( Figure S13, Supplementary Materials). Both P1 and P2 polymers show a weak band at 1285 cm −1 , that is not observed in the P3 spectrum of Figure S3. This suggests that an alkylsulfonate vibrational mode cannot account for this feature. Interestingly, this weak band does not match the pristine PCPDTBT IR spectrum but instead has a close similarity to one of the IRAV bands growing up upon p-doping PCPDTBT with I 2 vapors. On the basis of these considerations, we suggest that the band observed at 1285 cm −1 in the IR spectra of P1 and P2 is an IRAV band of the conjugated backbone arising from polaronic charged defects. These features support our previous conjecture that self-doping is maintained also in the P2 films. To sum up, both P1 and P2 display self-doping behavior, whereas P3 does not. We will show in the following paragraphs that self-doping is one of the important characteristics of anionic CPEs for attaining effective interfacial anode modification in OSCs devices with inverted geometry.

OSCs Devices Characterization
Direct geometry device: We have characterized the copolymers' interlayers in devices with direct geometry, made with a P3HT:PC 61 BM active layer, taking a standard PEDOT:PSS anode modifier as a reference. The devices architecture and PV characteristics are reported in the Supplementary Materials, Figure S14 and Table S2, while the devices' assembly procedures are given in the experimental section. The P1 devices exhibited the best PV performances with a power conversion efficiency (PCE) of 2.17, with comparable PV characteristics to the PEDOT:PSS reference device (see Table S1). A similar trend was already reported by the Bazan group using a P1-type CPE interlayer in P3HT:PC 61 BM devices [27]. Interestingly, the relatively high fill factor obtained using P1 (0.58), brings sup-Molecules 2021, 26, 763 9 of 17 port for an effective hole extraction when using P1. This can be ascribed to a combination of factors, including the good matching among the HOMO energy levels of P1 and P3HT, which is preventing a barrier formation to hole extraction; and the self-doping of P1, which is known to induce beneficial effects for the hole transport and for the engineering of the ITO electrode work function [22]. With regard to P2, a reduction of the PV performances were observed (PCE 1.39%). P2 has the same backbone/electronic features than P1 but lower processability, owing to the previously mentioned low solubility and aggregates formation in solution, that affects its film forming properties as confirmed by the presence of thick aggregates with granular morphology in the AFM images (see Figure S3). Such features can account for the lower performances obtained with P2 instead of P1. Using P3 as anode interlayer, there is a drastic drop of the PV performances, similarly to the previously reported pristine anionic CPE AILs, with a high energy gap and no self-doping, as P3 [23]. Kelvin probe measurements were used to investigate how the copolymers' interlayers modify the ITO electrode work function (see Figure S15 in the Supplementary Materials). While the effective work function of the ITO electrode decreased from 4.8 eV to 4.7 eV with the P3 interlayer, an increase to 5.02 eV and 4.98 eV was observed respectively with P1 and P2, as expected for self-doped anionic CPEs [23]. Hence, the hole selectivity of the ITO electrode increases when using P1 and P2, facilitating charge extraction, but it is reduced with P3.
Inverted geometry device: In order to investigate the key characteristics that anionic CPEs should have to obtain solution-processable anode interlayers for inverted OSC devices, we have tested P1, P2, and P3 interlayers in P3HT:PC 61 BM-based devices, with a device architecture displayed in Figure 4. PEIE thin film was used as cathode interlayer on top of the ITO electrode. This non-conjugated amino-containing polymer interlayer affords good stability and efficient performance when applied in fullerene-based polymeric solar cells [55,56]. For this study, PEIE offers the advantage of being scarcely affected by the air soaking treatments herein reported. In inverted geometry the interlayers are deposited on top of the hydrophobic active-layer surface and to increase the wettability we used ethanol for the AILs processing; this was possible for P2 and P3, but not for P1 which needs a water/ethanol mixture to be dissolved and processed. The details of the devices' assembly procedures are given in the experimental section. The PV characteristics are reported in Figure 4, Figure S16 in the Supplementary Materials, and Table 2.  The typical current density-voltage (J-V) curves of the devices featuring the polymeric interlayers P1, P2, and P3 are depicted in Figure 4b; we have taken for comparison a device with a state-of-the-art evaporated MoOx interlayer (MoOx/Ag) and a device prepared without any AIL (Ag). The PV characteristics are summarized in Table 2. The devices with the copolymer interlayers showed poor performances, with PCEs around 0.3-0.25%, similar to the Ag devices and about one order of magnitude lower than the reference MoOx/Ag devices, with a PCE around 2.5%. In one of our previous works we have highlighted that the air exposure of devices with polar polymers interfacial layers improved their performances [57]. Interestingly, after a short air soaking treatment of 15 min, a significant enhancement of the PV performances was observed in the P2 and P3 devices, see Figure 4c and Table 2. Namely, the devices made with P2 reached the best performances, comparable to the reference MoOx/Ag device, with a PCE of 2.62%; in P3 The typical current density-voltage (J-V) curves of the devices featuring the polymeric interlayers P1, P2, and P3 are depicted in Figure 4b; we have taken for comparison a device with a state-of-the-art evaporated MoOx interlayer (MoOx/Ag) and a device prepared without any AIL (Ag). The PV characteristics are summarized in Table 2. The devices with the copolymer interlayers showed poor performances, with PCEs around 0.3-0.25%, similar to the Ag devices and about one order of magnitude lower than the reference MoOx/Ag devices, with a PCE around 2.5%. In one of our previous works we have highlighted that the air exposure of devices with polar polymers interfacial layers improved their performances [57]. Interestingly, after a short air soaking treatment of 15 min, a significant enhancement of the PV performances was observed in the P2 and P3 devices, see Figure 4c and Table 2. Namely, the devices made with P2 reached the best performances, comparable to the reference MoOx/Ag device, with a PCE of 2.62%; in P3 devices, the PCE was 1.9%. In contrast to the P2 and P3 devices and irrespective to air exposure, P1-based devices showed the same poor performances of the Ag devices, prepared without any AIL. Such poor P1 functionality in inverted devices might arise from the strong hydrophilic character of P1, leading to scarce adhesion and film formation on top of the hydrophobic P3HT:PC 61 BM active layer. The active layer (AL) coverage by the copolymer interlayers was analyzed by AFM. As shown in Figure 5, the surface morphology after P1 deposition is similar to the AL substrate, thus suggesting a bad AL coverage by the P1 interlayer. On the other hand, after P2 and P3 deposition, the morphology and roughness are significantly different. RMS values are reported in Figure 5 and are consistent with the presence of P2 and P3 on top of the active layer.  To confirm the adhesion of our polar CPEs to the active layer, we measured the water contact angles prior to and after the interlayers deposition. As reported in the Supplementary Materials, Figure S12, the AL water contact angle reduces respectively from 107° to 96° and 102° after the P2 and P3 interlayers deposition. Such increase of hydrophilicity arises from the adhesion of the P2 and P3 polar polymers on top of the AL hydrophobic surface. Interestingly, by spin casting P1 the water contact angle remained the same (at 107° as for the active layer) revealing a scarce/absent adhesion of P1 on the active layer. This explains the identical PV characteristics observed for the P1 device and the control Ag device.
The above observations highlight an important requirement that an AIL CPE for To confirm the adhesion of our polar CPEs to the active layer, we measured the water contact angles prior to and after the interlayers deposition. As reported in the Supplementary Materials, Figure S12, the AL water contact angle reduces respectively from 107 • to 96 • and 102 • after the P2 and P3 interlayers deposition. Such increase of hydrophilicity arises from the adhesion of the P2 and P3 polar polymers on top of the AL hydrophobic surface. Interestingly, by spin casting P1 the water contact angle remained the same (at 107 • as for the active layer) revealing a scarce/absent adhesion of P1 on the active layer. This explains the identical PV characteristics observed for the P1 device and the control Ag device.
The above observations highlight an important requirement that an AIL CPE for solution-processed inverted solar cells has to satisfy: a correct balance between the hydrophilic and hydrophobic parts. This ensures a good adhesion to the active layers while maintaining solution processing from orthogonal solvents, which is a key for devices multi-stacking fabrication.
As depicted in Table 2, the pristine Ag devices exhibit quite low V OC , poor FF, relatively high R s , and low R p . Such features are the typical fingerprints of the bad hole selectivity of a silver electrode. As a matter of fact, the Ag work function was reported at -4.3 eV, while the P3HT HOMO energy level was around −5 eV. As such, there is a bad electrode-organic band alignment, leading to a barrier to hole extraction.
The insertion of the P2 or P3 interlayers in the devices, combined with an air soaking treatment (15 min), induces a significant enhancement of the V OC and FF parameters as compared to the pristine Ag. This testifies the ability of the P2 and P3 interlayers to improve the hole collecting character of the top Ag electrode. As depicted, the V OC and FF parameters of the P2 and P3 devices are reaching values close to the ones obtained with the state-of-the-art MoOx interlayer. This behavior evidences an overall effective modification of the silver electrode-active layer band alignment by the insertion of these copolymers interlayers. However, P2 appears to be a more effective anode modifier than P3, as self-doping occurs in P2 but not in P3. Self-doping should favor the hole transport within the P2 interlayer, leading to beneficial effects for hole extraction in the devices. Besides self-doping, P2 exhibits a better energy level alignment to the P3HT AL component than P3, see Figure 2. This reduces the barrier to hole collection at the AL/CPE interface when using P2 rather than P3 [25].
The above comparison of P2 and P3 AILs clearly highlights that the choice of the conjugated backbone is a key in the development of effective AILs for inverted solar cells. To favor hole collection at the top silver electrode, the conjugated backbone should be designed to grant self-doping and provide a good band energy alignment at the AL-AIL interface.
Interestingly, in direct geometry, P3 did not function as anode interlayer. In inverted geometry, even if P3 is a less effective AIL than P2, due to the mentioned absence of selfdoping and poor band alignment the P3 interlayer is able to induce a good hole collecting character to the top Ag electrode. Since both P3 and P2 have alkylsulfonate side groups, but different conjugated backbones, we deduce that the sulfonate side groups are playing a role in the anode engineering of the inverted solar cells.
It should be noticed that the ability of the sulfonate groups to impart a hole collection character to the top electrode could be not necessarily relate to this peculiar side group, but just to its inherent hydrophilicity. A hydrophilic material close to an Ag electrode upon air exposure may induce a shift from the vacuum of the Ag electrode work function via an oxidation mechanism [58].
In an attempt to clarify this issue, we have monitored, within a time scale of few days, the evolution of the J-V curves and PV characteristics of our devices versus their time of storage under air atmosphere. As depicted in Figure 6, with the P2 and P3 interlayers, most of the gain in hole selectivity of the top electrode is obtained rather quickly, at the very beginning of air exposure ("day 0" here corresponds to the previous 15 min air treatment). The Ag control device exhibits a completely different behavior upon aging under ambient atmosphere. At the very beginning of air exposure, the PV characteristics are not substantially affected. For longer times, a steady and continuous gain in hole selectivity of the top Ag electrode is observed. Such effect is ascribed to the formation of a thin oxide layer at the inner Ag surface, shifting the electrode work function to −5 eV [58]. It is known that this is a gradual process, driven by the slow diffusion of oxygen from the edges of the electrode [59]. For this reason, the gain in hole selectivity in the reference Ag devices is not completed in a few days [58,59]. By inserting the P2 and P3 interlayers, a faster process could be envisaged, owing to the hydrophilic nature of the CPEs which is attracting moisture at the buried silver. Moisture, in fact, provides a medium for adsorption of gases and the subsequent formation of an oxide layer [60,61]. However, a gradual oxidation process would be anyhow expected [18], which is by far different from the P2 and P3 trend depicted in Figure 6. For this reason, the formation of an Ag 2 O layer is not enough to explain the functional behavior of the P2 and P3 anode interlayers. This was indeed confirmed by monitoring the P2 device PV performances after 15 min storage under different ambient conditions, see the J-V curves in Figure 7 and corresponding PV parameters in the Supplementary Materials. We have found that a N 2 atmosphere, without oxygen but with a moisture content as in ambient air (50% RH), is identically effective for the PV performances as a standard air soaking treatment. Therefore, it is moisture rather than oxygen which plays a role in the P2 and P3 functional behavior. As shown in Figure 7b, by exposing the devices back to dry N 2 atmosphere and/or by drying them up by vacuum treatments, the anode modification is very stable and almost no reversibility is observed.     To summarize, the above results indicate that the mechanism that explains the common ability of P2 and P3 in modifying the top anode electrode is related to the combined effect of the sulfonate side groups and water molecules. Note that in fact water is a proton source able to dope this class of polymers, as recently reported by Bazan and coworkers [51].
The importance of sulfonate groups in the engineering of the top Ag electrode was also confirmed by using a PTB7-Th-based active layer (see the Supplementary Materials). Here, the non-optimal alignment of the levels between the PTB7-Th and AIL led to PV poor performance, but similarly to P3HT-based devices, a quite relevant gain in hole selectivity was also observed.
According to the above discussion, it has been identified that the choice of the polar group is another extremely important factor in the design of the AILs. We infer that the sulfonate moieties not only assure the solubility in polar orthogonal solvents, inducing self-doping via a right choice of the conjugated backbone, but also play a role for the anode engineering in inverted solar cells.
However, a simple picture that may possibly explain this mechanism is proposed in the following Figure 8. In the presence of moisture/water, SO 3 − anions get partially solvated and therefore the ionic pair with the alkaline metal K + is weakened, gaining the freedom to better interact with the silver surface. Moisture may also facilitate the orientation of the sulfonate groups towards the metal interface by sweeping the voltage (electric field) in the device. As a result, dipoles oriented towards the inner silver electrode should be formed, shifting from vacuum the silver electrode work function. As a result, the hole collecting character of the top electrode is improved. ated and therefore the ionic pair with the alkaline metal K + is weakened, gaining the freedom to better interact with the silver surface. Moisture may also facilitate the orientation of the sulfonate groups towards the metal interface by sweeping the voltage (electric field) in the device. As a result, dipoles oriented towards the inner silver electrode should be formed, shifting from vacuum the silver electrode work function. As a result, the hole collecting character of the top electrode is improved.

Conclusions
In this work we have conducted a study about the unexplored application of sulfonate anionic CPE solution-processable anode interlayer materials in inverted organic photovoltaics. Based on the results obtained in our study it is possible to establish a good understanding of AIL structure-property-PV performance relationships for inverted devices applications. In fact, by designing and investigating the functional behavior of three different polymers P1, P2, and P3 bearing a different number of sulfonate groups and modifications of their conjugated backbone, we could assess the important material features that should be taken in account for the development of effective AIL CPE materials for inverted OSC devices. First, it is mandatory to develop anionic CPEs with good

Conclusions
In this work we have conducted a study about the unexplored application of sulfonate anionic CPE solution-processable anode interlayer materials in inverted organic photovoltaics. Based on the results obtained in our study it is possible to establish a good understanding of AIL structure-property-PV performance relationships for inverted devices applications. In fact, by designing and investigating the functional behavior of three different polymers P1, P2, and P3 bearing a different number of sulfonate groups and modifications of their conjugated backbone, we could assess the important material features that should be taken in account for the development of effective AIL CPE materials for inverted OSC devices. First, it is mandatory to develop anionic CPEs with good wettability to the active layer and this can be achieved by the correct balancing of hydrophilic and hydrophobic substituents. Second, similarly to conventional direct geometry devices, the conjugate backbone should be suitably designed to ensure the self-doping of the CPE materials and grant a good energy level match with the photoactive layer. Finally, the choice of the polar groups is another important factor in the design of the AILs. The sulfonate moieties not only assure the solubility in polar orthogonal solvents, induce self-doping via a right choice of the conjugated backbone, but also play, combined with moisture exposure, a role for the anode engineering in inverted solar cells. Moisture annealing is a simple, easily accessible, and low cost procedure and could be a valuable alternative to electrical or thermal annealing. However a more comprehensive understanding of the importance of sulfonate for the self-doping mechanism of CPEs and its effect on charge transport and mobility is required to improve the use and design of this class of polymers. We believe that our insight could give a valuable contribution to the advancement in the development of engineering all polymeric solution-processable inverted solar cells.