Compatible Solution‐Processed Interface Materials for Improved Efficiency of Polymer Solar Cells

The electron transport layer (ETL) in an organic solar cell is one of the main components that play a crucial role in the extraction of charges, improving efficiency, and increasing the lifetime of the solar cells. Herein, solution‐processed PBDTTT‐C‐T:PC71BM‐based organic solar cells are fabricated using conjugated PDINO molecules, sol‐gel derived under stoichiometric titanium oxide (TiOx), and a mixture of the same as an ETL. For PBDTTT‐C‐T:PC71BM‐based organic solar cells, a blend of organic‐inorganic ETLs demonstrates reduced bimolecular recombination and trap‐assisted recombination than a single ETL of either two materials. Furthermore, in both, fullerene and nonfullerene systems, the efficiency of the devices employing the blend ETL as compared to the single ETLs show some performance improvement. The strategy of integrating compatible organic and inorganic interface materials to improve device efficiency and lifetime simultaneously, and demonstrate the universality of different systems, has potential significance for the commercial development of organic solar cells.


Introduction
As one of the most attractive candidates for third-generation thin-film photovoltaics, polymer-based organic solar cells (OSCs) have attracted considerable attention due to their production advantagesuse of "green" halogen-free solvents [1] and low-cost roll-to-roll processing, resulting in lightweight products. [2] With the introduction of novel nonfullerene acceptors, power conversion efficiencies (PCEs) exceeding 18% have been achieved in recent years. [3] In the visible and infrared spectral range, efficient material systems usually have a broad and strong absorption spectrum, which is mostly employed for harvesting photons to create free charges. One key to the innovation and improvement of organic solar cells lies in the charge transport layer, including electron (ETL) and hole transport layers (HTL), which play a vital role in the development of OSCs. [4] The HTL and ETL are naturally located on opposite sides of the active layer and are responsible for extracting and transporting holes and electrons from the active layer, respectively. To meet the goal of enhancing device performance, an effective HTL or ETL should modify the electrode, respectively, the active layer-charge extraction interface to lower the energy barrier and surface defects, increase the extraction and transmission efficiency of excitons by the charge transport layer, and reduce interfacial charge recombination. [5] Generally, interlayer materials can be distinguished into two kinds: either organic or inorganic materials. The main representatives of organic ETLs like poly[ (9,9-bis(3′-(N,Ndimethylamino)propyl)-2,7-fluorene)-alt-2,7-(9,9-dioctyfluorene)] (PFN), poly(9,9-bis(3′-(N,N-dimethyl)-N-ethylammoiniumpropyl-2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene))dibromide (PFN-Br), N,N″-bis(N,N-dimethylpropan-1-amine oxide)perylene-3,4,9,10-tetracarboxylic diimide (PDINO), and N,N″-bis{3-[3-(dimethylamino)propylamino]propyl}perylene-3,4,9,10-tetracarboxylic diimide (PDINN) exhibit a high electron-transport ability, efficiently lower the work function of cathode materials, and provide a smooth and homogeneous film morphology that can be universally applied in conventional OSCs. [6] The inorganic ETLs such as Zinc Oxide (ZnO), Titanium Oxide (TiO x ), and Tin Oxide (SnO 2 ) are the most extensively utilized materials in inverted OSCs resulting in good air stability, ease of The electron transport layer (ETL) in an organic solar cell is one of the main components that play a crucial role in the extraction of charges, improving efficiency, and increasing the lifetime of the solar cells. Herein, solution-processed PBDTTT-C-T:PC 71 BM-based organic solar cells are fabricated using conjugated PDINO molecules, sol-gel derived under stoichiometric titanium oxide (TiO x ), and a mixture of the same as an ETL. For PBDTTT-C-T:PC 71 BM-based organic solar cells, a blend of organic-inorganic ETLs demonstrates reduced bimolecular recombination and trap-assisted recombination than a single ETL of either two materials. Furthermore, in both, fullerene and nonfullerene systems, the efficiency of the devices employing the blend ETL as compared to the single ETLs show some performance improvement. The strategy of integrating compatible organic and inorganic interface materials to improve device efficiency and lifetime simultaneously, and demonstrate the universality of different systems, has potential significance for the commercial development of organic solar cells. manufacture, and capacity to process alcohols. [7] However, inorganic-based ETLs frequently require an annealing temperature of above 100 °C, and there are rough surfaces and UV doping effects, which cause contact difficulties with the active layer and charge trapping in the surface, reducing the stability of the devices. An approach of an organic-inorganic combined preparation of ETL is utilized to increase the charge extraction capability of ETL based on inorganic materials and even the stability of associated devices.
Already several successful examples of modifying an inorganic electron extraction layer with an additional organic layer have been demonstrated. Han et al. employed top-casting of PFN to modify the ZnO ETL in inverted OSCs and those devices exhibited an enhanced performance and stability. The device of PTB7:PC 71 BM:ICBA-based OSC achieved a PCE of 8.7%, which is more than an 8% improvement compared to the ZnO-or PFN-only devices. The stability of the device with ZnO/PFN bilayer preserved 60% of the initial PCE after 90 days in a drying cupboard, while the device of pristine PFN dropped to 30% after 90 days in the same storage environment. [8] Zhu et al. used glycine (Gly) modified ZnO as the ETL, leading to a decreased work function from 4.11 eV for ZnO to 4.02 eV for ZnO/Gly, which increased the built-in electric field and improved the electron extraction. The device of PM6:IT-4F with ZnO/Gly obtained a PCE of 14.0% which indicated an ≈9% increase of the PCE with the Gly modification of the ETL, as well as presenting lower bimolecular and trap-assisted recombination. [9] Kong et al. modified the SnO 2 ETL by using PDINO in inverted OSCs, which overcome the drawbacks of SnO 2 and restrained defect-caused molecular recombination in the device. The OSCs of PM6:Y6 with PDINO/SnO 2 showed an enhanced performance from 12.7% to 14.9%, which improved by 17% from pristine SnO 2 . [10] The bilayer ETL-based OSCs above demonstrate a considerable improvement in PCE, but the two-step deposition procedure adds manufacturing complexity. Lee et al. devised a simple method for combining ZnO and PFN to fabricate a single organic-inorganic hybrid ETL by simply mixing both solutions. Thereby the interfacial contact between the active and electron extraction layer was improved and the series resistance of the PSC device was decreased. A PTB7:PC 71 BMbased OSC with ZnO-PFN gave a PCE of 9.2%, compared to a ZnO-based device with only 7.2%. The device with ZnO-PFN gained a 92% retention of the initial PCE, while the ZnO maintained 86% after being stored in N 2 filled glovebox for 1200 h. [11] Liu et al. doped PEI into ZnO and obtained a thermal annealing-free ETL, with decreased work function as compared to the ZnO layer and consequently a reduced interfacial barrier. Both of which were revealed to be beneficial for an improved electron extraction and a declined charge recombination. The PTB7:PC 71 BM based device with ZnO-PEI exhibited an enhanced charge extraction ability with a PCE of 9.31% as compared to pristine ZnO-ETL with 7.96%. [12] Meanwhile, this ease of processing suggests a use for mass production of commercial solar cells. However, while there is no challenge in processing ETLs on ITO, as done in inverted OSCs, using a conventional device layer stack requires the ETL to be processed on top of the photoactive layer without extreme thermal annealing steps. [13] When it comes to inorganic-organic materials, it typically plays the purpose of optimizing the surface morphology and regulating the energy level, which is beneficial for boosting device efficiency. However, in order to reduce the production cost, the preparation process of the interface layer is also particularly important. Simple processing such as annealing-free can effectively reduce the energy consumption and the fabrication cost during the preparation process, and also helps to shorten the energy payback time.
In this study, the ETL is under focus. Understanding the impact of single and blended ETL on device efficiency and longterm stability is critical. The goal of this study is to establish a link between photovoltaic properties and the different ETL stacks for device optimization. Furthermore, the universality of the blend method was examined by demonstrating it for both, fullerene and nonfullerene systems. Photovoltaic performance, charge recombination, and dissociation, as well as device stability and energy payback time, [14] were considered for evaluation. Our approach is based on TiO x as the inorganic portion, which has already been effectively employed in both OPVs and perovskite solar cells and have a lower price. We combine that with the popular organic material PDINO, which usually exhibit good film-forming properties, and both, PBDTTT-C-T:PC 71 BM [15] and PM6:Y6 systems were selected for this study. The blending strategy helps reducing the number of processing steps and thus production costs as well as energy consumption, while solving the problem of perforation caused by inorganic materials and form high-quality and uniform thin films, obtain better ohmic contact with the aluminum electrode. It is thought to alter ETL's physical features, including work function and electron mobility. We hypothesized that due to an inevitable difference in the surface energy, a layering would also occur for the blend-ETL, which would affect the device's photovoltaic performance and stability positively.

Results and Discussion
The OSCs were fabricated with a conventional structure of ITO/PEDOT:PSS/PBDTTT-C-T:PC 71 BM/Interface layer/Al as shown in Figure 1a. The molecular structures of polymer donor PBDTTT-C-T, fullerene acceptor PC 71 BM, nonhalogenated solvent m-xylene, nontoxic additive o-vanillin, and the alcohol processed organic material PDINO are shown in Figure 1b PDINO was mixed with TiO x in different ratios v/v and spin-coated onto the active layer as the PDINO:TiO x layer. The energy level alignment of the layer stack as shown in Figure 1c between the TiO x of −4.35 eV and PDINO −4.10 eV, the highest occupied molecular orbital energy level of PDINO:TiO x was −4.20 eV. The matching energy level can be beneficial to achieving higher short circuit current density (J sc ) and fill factor (FF) in these devices.
For TiO x -only devices, we first investigated the effects of thermal annealing (at 110 °C in air, 10 min) versus without thermal annealing (keep it in the air at 22 °C for 10 min, humidity ≈35%), J-V curves, and photovoltaic characteristics are shown in Figure S2 (Supporting Information). The devices without thermal treatment performed better than those with thermal treatment at 110 °C for 10 min. Figure S3 and Table S1 (Supporting Information) showed an insensitive treatment time ranging from 7 to 30 min. Those interesting results suggest the easy approach without annealing is simpler and more www.advancedsciencenews.com www.advmatinterfaces.de suited for fabricating TiO x layer and yielding beneficial properties. For inverted device structures, TiO x is often annealed at high temperatures. In this system, 6% PCE was achieved following light soaking treatment, which is less than 7.47% of the conventional device, as shown in Figure S4 and Table S2 (Supporting Information). It suggests that TiO x in the PBDTTT-C-T:PC 71 BM system performs more efficiently in the conventional device structure. Then, we identified the optimal mixing of PDINO:TiO x v/v ratios, and compared the different treatments in air and N 2 atmospheres, as shown in Figure S5 and Table S3 (Supporting Information). The performance was improved as TiO x concentration was increased, yielding the best PCE of PDINO:TiO x , with 7.47% in air, and 7.94% in N 2 at the 60:40 blend ratio . At higher TiO x concentrations the performance dropped. The J-V curve for PDINO:TiO x blend ETL devices treated in the glove box acquires an S-shape when the concentration of TiO x reaches 100%, resulting in a drop in FF and PCE. This S-shape may be caused by unfilled trap states, which results in interfacial recombination and thus in a lower overall charge carrier concentration. [17] In fact, the best results were obtained when the PDINO:TiO x mixing ratio is 60:40 processed in an N 2 -filled glove box. We manufactured the bilayer ETLs of PDINO/TiO x and TiO x /PDINO at various spin coating frequencies in order to understand the results better, as shown in Figure S7 and Table S4 (Supporting Information). It turned out to be beneficial to use a thinner layer of TiO x on top of a somewhat thicker layer of PDINO, leading to slightly improved opencircuit voltage (V oc ). While bilayer ETLs of PDINO (bottom)/ TiO x (top) and blended ETLs produced similar results, we used blended ETLs in the device due to the reduced processing efforts of single-layer ETLs. The possible erosion between TiO x and PDINO during sequential depositions, as seen in Figure  S12 (Supporting Information), illustrates the risk of predeposited ETLs being washed off during postdeposition. Additional advantages of the blend ETL have been identified, such as lower cost than pristine PDINO and a possible thicker ETL (for reliable large area coating) due to the reduced optical absorption of the same, as shown in Figure S11 (Supporting Information). Figure 2 shows the J-V curves under simulated solar light, in the dark, external quantum efficiency (EQE), and PCE histogram of three different interface layer materials-based devices. The related photovoltaic parameters including V oc , J sc , FF, serial resistance (R s ), parallel resistance (R p ), and PCE were summarized in Table 1. The PCE of 7.94% is one of the highest reported efficiencies for halogen-free solvent processed PBDTTT-C-T:PC 71 BM-based organic solar cells. For PDINO:TiO x -based devices, the values of V oc , FF, and R s are all comparable to TiO x -only and PDINO-only devices. However, there is a distinct increase in J sc , which might originate from an optical spacer effect due to the blend ETL, since TiO x is not showing any parasitic absorption but contributes to the ETL layer thickness. [7b,18] On the one hand, in conventional devices, the optical spacer layer spatially redistributes the optical-electric field intensity within the device, allowing the active layer to cover the ideal intensity region and, as a result, boosting the short-circuit current and enhancing efficiency. On the other hand, the blended interfacial layer has a better overall low absorption intensity between 300 and 800 nm than the other two, shown in Figure S11 (Supporting Information). It means that when sunlight passes through the active layer, the interface layer of PDINO:TiO x , will absorb relatively little light, and the remaining sunlight is reflected through the top aluminum electrode for secondary absorption in the active layer, resulting in a higher J sc . The computed EQE values for TiO x , PDINO, and PDINO:TiO x -based device is 14.99, 15.04, and 15.56 mA cm −2 , respectively, corresponding to corrected PCEs of 6.73%, 7.07%, and 7.16%. The disparity between computed J sc from EQE and measured J sc from IV may be due to an incorrect distance between the simulator and the test bench, or due to a spectral mismatch. [11] The dark J-V curves shown in Figure 2c exhibited that the PDINO:TiO x -based device had a reduced leakage current and higher short circuit current density, which is the fundamental cause for enhanced OSCs performance. As a result of this method, PDINO:TiO x exhibited a considerable improvement in device performance when compared to single-component ETL materials.
To attain the surface properties of ETLs (TiO x , PDINO, and on PDINO:TiO x blends) processed on the photoactive layer of PBDTTT-C-T:PC 71 BM, we performed atomic force micro scopy (AFM) on the same. In Figure 3, the AFM image of the PDINO  [16] layer shows orderly gullies with uniform depressions of around 40 nm in width. The AFM height image of TiO x reveals uniformly distributed pore-like features with a dia meter distribution around 50 nm. Due to the abundance of voids and low-lying areas in both TiO x and PDINO, resulted a larger roughness (root mean square) value, the charge extraction and transport might be hampered by the large hole size and consequent discontinuities. The film with a pore size that is somewhat smaller than pristine TiO x was produced with the addition of some PDINO. Complementarity with large trenches and pore widths occurs after combining a low concentration TiO x with PDINO, leading to more homogeneous films with smaller pore sizes. This surface structure possibly enhances the effectiveness of electron transfer and extraction ability. In addition, some bright spots/areas are observed in all cases, which may originate from the underlying active layer and not from impurities on top of all layers (such as dust particles), as shown in Figure S8 (Supporting Information). This can be concluded for two reasons: i) the bright spots are visible in the film of the active layer and ii) typical PDINO phase-signals are found at those locations for the active layer overcast with PDINO, showing that these higher level regions originate from below. The scanning electron microscopes (SEM) images are shown in Figure S9 (Supporting Information), but were obtained for ETLs processed on glass and not on the photoactive layer. The PDINO:TiO x exhibited smaller hole-like and more homogenous surface structures than either TiO x or PDINO. It is suggesting that the PDINO:TiO x might provide a smoother film structure, which could be beneficial for improving ohmic contact with the top electrode. Thus SEM and AFM results consistently show the PDINO:TiO x blend ETL to exhibit a more flat and more homogeneous film morphology, possibly leading to improved charge collection. The average PCE is obtained from over ten independent devices; b) PDINO:TiO x = 60:40 v/v. The high-efficiency nonfullerene system PM6:Y6 was chosen for prove of principle, using the device structure reported in the literature to investigate the strategy's broad applicability. [19] In PM6:Y6-based devices with a conventional structure of ITO/ PEDOT:PSS/PM6:Y6/ETLs/Al, the optimized ETL materials TiO x , PDINO, and PDINO:TiO x have been used as well. Figure S10 (Supporting Information) displays the J-V curves and EQE spectra of the devices with TiO x , PDINO, and PDINO:TiO x , and all photovoltaic parameters are listed in These interesting results show that the strategy of simply mixing organic and inorganic material as an ETL can be effective in both fullerene and nonfullerene-based solar cells.
In parallel, we investigated the charge recombination characteristics of OSCs under short circuit current and open-circuit voltage conditions by measuring the light intensity (P light ) dependent J-V characteristics. The intensity of monochromatic light-emitting diode (LED) light (465 nm) varied between 0 and 2000 W m −2 . J sc is related to J sc ∝ (P light ) α , and α is computed from the fitting curve. In a nutshell, α equals 1 means that all free charge carriers in the device are swept away and extracted by electrodes before recombination. An α < 1 indicates that there exists weak recombination in the solar cell. [20] As present in Figure 4a, a higher α value of 0.945 was obtained for the PDINO:TiO x based device, compared to a slightly lower α value of 0.917 and 0.934 for TiO x and PDINO based www.advancedsciencenews.com www.advmatinterfaces.de device, respectively. This result suggested that the device with PDINO:TiO x suppresses bimolecular recombination more effectively than the TiO x or PDINO-based device, which may lead to the observed higher J sc and FF. Further, to analyze the trap-assisted recombination of OSCs under the open circuit voltage condition, V oc dependency on light intensity (P light ) was measured. The general functional dependence is described as V oc ∝ n(k B T/q)ln(P light ). [21] Here, k B , T, and q represent Boltzmann constant, temperature, and elementary charge, respectively. A slope of 2k B T/q (n = 2) indicates absolute trap-assisted recombination is the dominant recombination mode, while a slope of k B T/q (n = 1) suggests bimolecular recombination is the dominant model. As shown in Figure 4b, the device with PDINO:TiO x had a lower slope of 1.15 k B T/q than for TiO x or PDINO-based devices, which had a slope of 1.25 and 1.21 k B T/q, respectively. Thus, less trap-assisted recombination occurs in mixed ETL-based devices. In comparison to the single component ETL-based devices, the device with PDINO:TiO x ETL yielded a larger J sc , FF, and device performance. Overall, the light-intensity dependent J-V characterization shows that the mixed ETL applied in OSCs can lower both, bimolecular trapassisted recombination. To explore the impact of different ETL materials on PBDTTT-C-T:PC 71 BM-based devices, we determined the maximum exciton generation rate (G max ) and exciton dissociation probabilities [P(E, T)]. [22] The relationship of photocurrent density (J ph , defined as J L -J d , were obtained under illumination and dark condition, respectively) versus effective voltage (V eff , defined as V o -V appl , where V o was a voltage value when J L -J d = 0, and V as the applied voltage) was measured to further investigate the impact of charge generation and extraction properties on different ETL-based OSCs. In Figure 4c, the recorded J ph -V eff curves of OSCs with various ETLs are displayed. The saturated photocurrent densities (J sat ) for using TiO x , PDINO, and PDINO:TiO x as ETL are 17.80, 18.34, and 18.52 mA cm −2 at a high V eff of 3 V, which assumed that all photogenerated charge carriers were fully dissociated and removed by the electrode. The values of the G max have been calculated the values of G max using the equation J sat = qG max L, where q is the electronic charge and L is defined as the thickness of the photoactive layer. As shown in Table S6 (Supporting Information), the value of G max for TiO x -only, PDINO-only, and PDINO:TiO x -based devices are 1.11 × 10 22 , 1.15 × 10 22 , and 1.16 × 10 22 cm −3 s −1 , respectively. Therefore, after incorporating PDINO and TiO x into an ETL in the device, G max is slightly enhanced. Because the value of G max is a measure of the maximum number of photons absorbed, [23] this increased G max indicates the enhanced light intensity in the active layer of the PDINO:TiO x device. Furthermore, the P(E, T) of the OSCs were calculated by P(E, T) = J ph /J sat under short circuit conditions. [24] The results exhibit that TiO x , PDINO, and PDINO:TiO x -based devices have a value of P(E, T) is 91.6%, 92%, and 93.5%, respectively. The PDINO:TiO x -based device's higher P(E, T) value underlines an improved charge extraction ability over the single component ETL-based device. As a result, the increased exciton generation rate and the dissociation probability of PDINO:TiO x devices, resulted in the enhanced photocurrent of the OSCs.
The electron mobility properties of solar cells with PDINO, TiO x , and PDINO:TiO x as the ETL were evaluated with an electrononly device structure of ITO/ZnO/PBDTTT-C-T: PC 71 BM/ETL/Al,  Figure S13 (Supporting Information), the characterization of the J-V curves of single-carrier devices, and mobility was calculated by the space charge-limited current (SCLC) method. [25] As shown in Figure 4d and Table S7  To investigate the operational device stability independence of the three different ETLs, the solar cells were aged either at 45 °C in air under 1 sunlight intensity using white LED illumination or at room temperature in the glovebox (N 2 , dark). Electroluminescence imaging (ELI) was performed to analyze deficiencies in charge carrier injection from the electrodes, as well as for studying the general degradation of these solar cells. [18] Figure 5a shows the normalized photovoltaic parameters of the aged solar cells. Upon aging in a glovebox for 1000 h, all devices showed reasonably high stability and maintained over 95% of their initial PCE. Figure 5b shows how the electroluminescence emission of these devices scarcely changes after being placed in the glove box for 1000 h. This is another piece of proof that the solar cells remain stable in the glove box. Figure 6a shows normalized photovoltaic parameters obtained at 45 °C under white LED illumination in the air. As shown in Figure 6a, the V oc shows the most stable behavior among all parameters in OSCs. Even with high work function magnesium as the electron transport layer, the V oc of PBDTTT-C-T:PC 71 BMbased OSCs remains stable under continuous irradiation at 45 and 65 °C. [26] Because of this, all three ETL-interlayers did not show a different impact on the V oc stability, even though the material system largely differs. However, after 1000 h of continuous illumination at 45 °C, the PCE of all devices had fallen to less than 60% of the initial PCE, which is reflected by the decreases of J sc , FF, and R p already, as well as the increase in R s . As shown in Figure 6b, the three ETLs OSC's ELI show no discernible change in the first 300 h. The hue of the black dots on the TiO x -based device's interface began to darken at 500 h, and the surrounding dark areas based on the original black spots became more diffuse by 1000 h. Unlike TiO x , the black spots on PDINO-based devices became larger and were surrounded by a pale area after 500 h, and the pale area gradually expanded to the surrounding after 1000 h. For the PDINO:TiO x www.advancedsciencenews.com www.advmatinterfaces.de device more black spots in the center of a hole were surrounded by a corona of pale grey, which gradually increased over time.
It may be interesting to investigate the dark spot properties in future work. However, the variation in ETLs did not bring about remarkable differences in the stability, the PDINO:TiO x -based device had a greater lifetime energy yield within 0 -T 80 . As illustrated in Figure S14 and Table S8 (Supporting Information) The PM6:Y6 solar cells were aged under the same conditions as the ones mentioned above. As shown in Figure S15 (Supporting Information), all devices exhibited good stability upon aging in the glovebox for 1000 h, and they retained 90% of the initial PCE. In agreement, the ELI images shown in Figure S16 (Supporting Information) exhibit nearly no change upon 1000 h of aging. On the contrary, when the devices were aged at 45 °C under white LED illumination, they showed low stability, and short T 80 lifetimes, 15 h (TiO x ), 9 h (PDINO), and 44 h (PDINO:TiO x ), respectively, compare with Figures S17, S19, and Table S8 (Supporting Information). The device with PDINO:TiO x produced an improved energy output of 5.55 kWh m −2 within 0 -T 80 compared with TiO x (1.77 kWh m −2 ) and PDINO (1.03 kWh m −2 ). The PM6:Y6 aging under 1 sun simulator has been observed to have a limited device lifetime, which is perhaps due to oxidation of the photosensitive materials. [27] The decline in V oc and J sc primarily contributes to the rapid loss in PCE of all devices. After 500 h, the increased oxidation of the cell's edge due to air penetration may induce device failure, as shown in Figure S18 (Supporting Information). In Figure S18 (Supporting Information), the black blotches on the surface altered dramatically as the aging process progressed. The ELI image only indicated a rise in the number of black spots development in the first 500 h of use of the TiO x -based device. After 500 h, the device efficiency is close to zero, it should be triggered by water and oxygen eroding the section outside the cell region and diffusing inside the cell. In the early stages of aging, the ELI of PDINO-based device developed some black patches. As time passes, the black patches progressively spread to the surrounding region, forming a dark zone, posing a threat to the device's FF, J sc , and lowering the PCE. PDINO:TiO xbased ELI formed several little but thick patches without aging. As time passed, several of the black dots deepened in color but did not expand. At 1000 h, the ELI image became hazy, and the color of the black spots was very dark. The darker patches may have a greater influence on the device's photostability than places surrounded by light, according to the aging data.

Conclusion
We introduced a straightforward strategy of combining the organic PDINO and the mostly inorganic TiO x interface material into one blend ETL in conventional OSCs. The annealingfree processed blend ETL exhibited the most homogeneous film structure, which may yield favorable ohmic contacts with the aluminum electrode. Further, the mixed interface material resulted in enhanced device performance over the single ETLs. Compared with these, PDINO and TiO x , devices with PDINO:TiO x revealed better exciton generation and dissociation probability, less bimolecular recombination and trapassisted recombination, and more balanced charge mobility. These results in combination with a favorable optical spacer effect contributed to the increased short circuit current density and PCE of the device. For PDBTTT-C-T:PC 71 BM and PM6:Y6 based solar cells, the device with PDINO:TiO x boosted the PCE from 7.47% to 7.94% and from 14.07% to 15.10%, respectively. Finally, processing a single blend ETL (instead of a doublelayer) without annealing helps to shorten the energy payback time. This work provides a simple and adaptable method for merging two compatible organic and inorganic interfacial materials to improve the performance of conventional OSCs, and it has potential application in the manufacture of large-scale solar cells in the future. indole-2,10-diyl)bis(methanylylidene)) bis(5,6-difluoro-3-oxo-2,3-dihydro-1H-indene-2,1-diylidene)) dimalononitrile (Y6) were purchased from Brilliant Matters (Quebec City, Canada). The hole-transporting layer PEDOT:PSS (Clevios PH) was acquired from Heraeus (Hanau, Germany). The environmentally friendly additive o-vanilline was purchased from Sigma-Aldrich. The TiO x precursor was synthesized according to the reference, [7b] and was diluted to 0.15% v/v with isopropanol before casting.

Experimental Section
Device Fabrication: The polymer solar cells were fabricated on an indium tin oxide (ITO) based glass substrate with a conventional layer stack of ITO/PEDOT:PSS/PBDTTT-C-T:PC 71 BM/Cathode interlayer/Al. Glass/ITO substrate was cleaned with a cotton swab dipped in toluene and then transferred into an ultrasonic bath, sequentially immersing in toluene and isopropanol solution, each for 15 min, then dried by a nitrogen gun. After argon plasma treatment for 5 min, a 40 nm thick anode buffer layer of PEDOT:PSS was spin-casted on top of precleaned glass/ITO substrate at a spin coating frequency of 3500 rpm for 30 s, followed by annealing treatment of the thin-film at 178 °C for 15 min. After that, the samples were moved into the nitrogen-filled glove box (containing oxygen and moisture less than 5 ppm) and cooled to room temperature. The PBDTTT-C-T:PC 71 BM (1:1.5, wt/wt) were dissolved with the nonhalogenated solvent m-xylene with a total concentration of 25 mg mL −1 .
[15a] The solutions were stirred for at least 2 weeks at 50 °C in a nitrogen atmosphere. The nontoxic o-vanillin as an additive was added into the host blend solution before casting. The photosensitive layer was spun cast on top of the PEDOT:PSS film at 1200 rpm for 60 s. The PM6:Y6 (1:1.2, wt/wt) were dissolved in chloroform with a total concentration of 16 mg mL −1 , next process of the blend film was following the reported literature. [19] Afterward, the methanol solution of PDINO, or isopropanol solution of under-stoichiometric TiO x , or the PDINO:TiO x solution were deposited onto the active layer at 3000 rpm for 45 s. This resulted in an around 10 nm cathode interface layer. Thereafter 200 nm of Al were deposited onto the cathode interlayer by physical vapor deposition (PVD) under a vacuum of pressure < 2 ×10 −6 mbar. A shadow www.advancedsciencenews.com www.advmatinterfaces.de mask was applied to define the solar cell's active area of 0.42 cm 2 during evaporation. Finally, all samples were encapsulated by epoxy-based UV-curing glue under glass before testing. Polymer Solar Cell Characterization: Short circuit current densityvoltage (J-V) curves of devices under a class AM 1.5G solar simulator from Wavelab's (Sinus-70) and dark conditions were measured, and recorded by Keithley 2400 Source Meter Unit (SMU). External quantum efficiency (EQE) data were obtained using monochromatic light and additional halogen bias light provides about 1 sun excitation intensity. The wavelength of the system was controlled by the monochromator, and the standard single-crystal Si detector was selected to calibrate the EQE system. The reflection and transmission spectra were recorded on a spectrophotometer of Avantes AvaSpec-ULS3648-USB2-UA-25, and the absorption was calculated over a wavelength from 300 to 1200 nm. To investigate the stability of solar cells under the different aging conditions of ISOS-L1, which test temperature is 45 °C. The solar simulator uses a spectral light intensity close to the AM 1.5G spectrum of 100 mW cm −2 . J-V data are automatically collected by Keithley 2400 SMU and a Keithley 2700 multiplexer module every 30 min. ELI is used to study the inhomogeneity of organic solar cells. ELI was measured using an automatic computer program in dark conditions. Keithley 2400 SMU supports constant current, and the current density was set to 200 mA cm −2 . The Si CCD camera ANDOR iKon-M was applied to record emitted light from solar cells after cooling to −50 °C. The samples of SEM measurement were sputtercoated with 5 nm platinum (Pt) to enhance contrast and lessen the potential charging effect. The in-lens detector was used in conjunction with a field-emission scanning electron microscope (Sigma VP, Carl Zeiss AG, Jena, Germany) at an acceleration voltage of 6 kV to perform the SEM imaging. The sample's layer stack of AFM measurement is ITO/PEDOT:PSS/Active layer/ETLs. AFM images were produced by DriveAFM [29] (Nanosurf ) under dynamic mode with the probe of PP-NCLR (Nanosensors).The hole mobility with a hole-only layer structure of ITO/PEDOT: PSS (40 nm)/Active layer/MoO 3 (10 nm)/Al (200 nm). The electron mobility was prepared using the electron-only device structure of ITO/ZnO (40 nm)/Active Layer (≈100 nm)/ETL (10 nm)/Al (200 nm), and ZnO was prepared according to published literature. [28] Both hole-only and electron-only devices were measured by using the model of space charge limited current (SCLC), which is explained by the MOTT-Gurney Equation

9
/ 8 D 0 r 2 3 ε ε = J V L µ (1) where J D is the current density calculated under dark condition, ε 0 is defined as the vacuum permittivity (ε 0 = 8.85 C V −1 s −1 ), ε r is the relative permittivity of the material (for polymer system, assumed to be 3), µ is the mobility of hole or electron, V defined as V appl -V bi -V s , V appl is the applied voltage, V bi described as the built-in voltage (0 V), V s is the voltage drop result from the substrate's series resistance (for the electron-only device, the V s = 0, while the hole-only device, V s = IR), and the L is the thickness of the photosensitive layer.

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