Photoactive layer formation in the dark for high performance of air-processable organic photovoltaics

Recent progress in organic photovoltaics (OPVs) has led to an increased importance of laboratory-scale fabrication in ambient air using solution processes. However, the effect of the existence of both ambient air and light during the formation of a photoactive layer on the performance of fabricated devices has not been elucidated thus far in detail. Here, we show that photoactive layer formation in completely dark conditions enables air-processable OPVs with a high power conversion efficiency. The degradation in OPV performance caused by the coexistence of air and room light was confirmed by systematically examining atmospheric and room-light irradiation conditions during the formation and drying of the photoactive layer. Moreover, the degradation rate was much faster than that in the case of dried solid photoactive layers exposed to room light in ambient air. The photoactive layer with non-fullerene acceptors showed a much slower degradation rate, owing to room light, than that with fullerene acceptors. Based on these findings, we demonstrate that by eliminating light during formation, slot-die-coated OPVs in ambient air show comparable performance to that of spin-coated OPVs in an inert glovebox.


Introduction
The remarkable progress of organic photovoltaics (OPVs) indicates that the field is anticipating a transition from laboratory-scale research and development to industrial-scale manufacturing [1,2]. The power conversion efficiency (PCE) of single-junction cells is approaching 18% with the development of electron donor and acceptor materials for photoactive layers [3]. Even in the all-coating fabrication process, a PCE of 11.9% has been achieved by fully solution-processed OPVs with improved charge transport layers and coated electrodes [4]. Researchers have attempted to replace the solution process in laboratory-scale spin coating with roll-to-roll compatible coating processes, such as doctor-blade coating [3,5,6] and slot-die coating [1,7]. Recently, a benchmark PCE value of 9.5% was achieved for large-area OPVs with an effective area of 216 cm 2 , using doctor-blade coated layers in a glovebox [6].
In addition to the device stability under ambient conditions, air processability is important for the industrialization of OPVs. To maximize the potential of OPVs, roll-to-roll fabrication under ambient air conditions is an important aspect for lowering costs. Although most of the existing photoactive materials require an inert glovebox during formation because of the instability in air, recent material advancements enable good performance through ambient formation processes [5,7,14,28,29]. Further, studies on the humidity dependency of cell performance show that photoactive layers with certain polymeric acceptors are relatively insensitive to water levels in the air [15,29]. Additionally, solvent vapor annealing is a good approach for improving the efficiency of the ambient air fabrication process [30].
However, little attention has been paid to the coexistence of ambient air and light during the formation process of photoactive layers. When considering industrial production, the influence of ambient light, rather than strong illumination (such as 1 sun), should be carefully evaluated.

Fabrication of organic photovoltaics
To investigate how light and ambient air influence the formation and drying of photoactive layers, we performed the spin coating under three conditions: (a) in the nitrogen-filled glovebox with both oxygen and H 2 O concentrations below 1 ppm and with the glovebox light off but room light on (illuminance of 210-230 lx inside the glovebox), (b) ambient air (21 • C-26 • C, 35%-45% RH) with LED room-light illumination (880-980 lx), and (c) ambient air with no illumination (approximately 0 lx). Following the layer formation, we stored the samples under three different conditions to dry the film: (a) in the glovebox, (b) uncontrolled ambient air (21 • C-26 • C, 35%-45% RH) with room-light illumination, and (c) in an environmental test chamber (SH-242 bench-top-type temperature and humidity chamber; Espec) with controlled temperature and humidity (25 • C, 30% RH) in ambient air and dark conditions. Unless otherwise stated, the drying time was set to 30 min.
The dried samples were installed in a vacuum evaporator, and a hole-transporting layer of molybdenum oxide (MoO X , 7.5 nm) and an Ag anode (100 nm) were sequentially deposited by thermal evaporation at <3 × 10 −4 Pa. Finally, a 1 µm-thick parylene layer was deposited by chemical vapor deposition to form a passivation layer. The effective area of the fabricated cell was 0.04 cm 2 .

Photoactive layer formation with slot-die coating
To confirm the roll-to-roll compatibility, we formed a photoactive layer using a slot-die coater (Mini-40, Die-Gate Co., Ltd) under ambient air conditions. The fullerene photoactive layer solution was coated with a 30 µm-width slot at 10 mm s -1 and a target wet thickness of 6 µm in ambient air under completely dark conditions. The coated sample was immediately installed in a vacuum chamber, and then MoO X (7.5 nm) and Ag (100 nm) were sequentially deposited by thermal evaporation. The effective area of the fabricated cell was 0.04 cm 2 .

Device characterization
The current density-voltage (J-V) characteristics of the OPVs were recorded under AM 1.5 G (100 mW cm −2 , with the intensity calibrated using a silicon reference solar cell) using a SourceMeter (Series 2400; Keithley) under ambient laboratory conditions. External quantum efficiency (EQE) measurements were performed with monochromatic light (SM-250 F; Bunkoukeiki) calibrated with a silicon reference diode. The absorbance of the films was characterized using an ultraviolet-visible-near-infrared (UV-vis-NIR) spectrophotometer (V-780; JASCO Inc.). The surface morphology was imaged using atomic force microscopy (AFM, SPM-9700HT; Shimadzu) in the tapping mode. Ultraviolet photoelectron spectroscopy (UPS) and x-ray photoelectron spectroscopy (XPS) measurements were performed using a photoelectron spectroscopy system (PHI5000 VersaProbe II; ULVAC-PHI Inc.). The light source for UPS was He I excitation light (21.2 eV). For all UPS measurements, a −5.0 V bias was applied to the samples. The x-ray source for XPS was monochromated Al Kα (1486.6 eV) radiation with an operating power of 50 W (15 kV voltage). The diameter of the analyzed area was 200 µm. The take-off angles were 90 • and 45 • to the sample substrate for UPS and XPS, respectively. Photoelectron yield spectroscopy (PYS) was conducted using a photoelectron spectrometer (AC2; Riken Keiki Co. Ltd.).

Results and discussion
Figures 2(a)-(c) present the J-V characteristics measured from the representative devices based on fullerene and non-fullerene acceptors under different photoactive layer formation and drying conditions, and table 1 summarizes the corresponding average photovoltaic parameters. The photovoltaic parameters of reference fullerene cells, which were formed and dried in a nitrogen-filled glovebox, exhibit a current density (J SC ) of 16.1 ± 0.3 mA cm -2 , an open-circuit voltage (V OC ) of 0.749 ± 0.004 V, and a fill factor (FF) of 0.740 ± 0.006, resulting in a PCE of 8.92 ± 0.07%. All parameters were significantly lower for cells with a photoactive layer formed and dried in air under room-light illumination. In particular, the average J SC was 2.1 ± 0.1 mA cm -2 . The average PCE was 0.49 ± 0.05%, which corresponds to a maximum power output normalized with the reference cell of that of the reference devices (P max /P max0 ) = 0.055. A similar sharp degradation was observed when spin coating was conducted in the glovebox, followed by drying in air under room-light illumination (1.46%, figure S1 and table S1 (available online at stacks.iop.org/JPMATER/4/044016/mmedia)).
Surprisingly, drying in air under completely dark conditions led to a dramatic improvement in the PCE. The average J SC of the devices whose photoactive layer was formed in air with room-light illumination and dried in air in the dark was 15.2 ± 0.2 mA cm -2 , resulting in a PCE of 6.34%, which corresponds to P max /P max0 = 0.711. Both formation and drying under dark conditions further improved the PCE, with the photovoltaic parameters J SC of 15.7 ± 0.3 mA cm -2 , V OC of 0.705 ± 0.005 V, and FF of 0.636 ± 0.007, resulting in a PCE of 7.02 ± 0.19%, which corresponds to P max /P max0 = 0.787. In sharp contrast to cells having a photoactive layer formed and dried in air with room-light illumination, the devices having a photoactive layer dried in vacuum and then exposed to air with room light exhibit a minor PCE decrease (7.26%, figure S1 and table S1). This shows that the existence of both air and room-light illumination during drying of the photoactive layer causes severely degraded performance.
The degradation, induced by air and room-light irradiation, was significantly slower in the case of the non-fullerene acceptors than that of the fullerene acceptor (figures 2(b) and (c)). The photovoltaic parameters of reference PBDTTT-OFT:IEICO-4F cells, which were formed and dried in the nitrogen-filled glovebox, exhibit a J SC of 22.9 ± 0.2 mA cm -2 , V OC of 0.674 ± 0.005 V, and FF of 0.708 ± 0.007, resulting in a PCE of 10.91 ± 0.23%. When the photoactive layer was formed and dried in ambient air with room-light illumination, both J SC and FF decreased by approximately 10% compared to the reference cells, resulting in a PCE of 8.61 ± 0.42% (P max /P max0 = 0.789). By blocking the room-light illumination during the drying of the photoactive layer, all parameters maintained approximately the same values as those of the reference cells, resulting in an average PCE of 10.87 ± 0.27% (P max /P max0 = 0.996). The spin-coating process in the dark in air led to a small improvement in the PCE. The photovoltaic parameters of non-fullerene OPVs having the photoactive layer formed and dried in air under dark conditions exhibit a J SC of 22.6 ± 0.2 mA cm -2 , V OC of 0.672 ± 0.006 V, and FF of 0.716 ± 0.008, resulting in a PCE of 10.88 ± 0.29%, which is comparable with that of the reference cells (P max /P max0 = 0.997). The photovoltaic parameters of the reference PM6:Y6 cells, which were formed and dried in the nitrogen-filled glovebox, exhibit a J SC of 24.4 ± 0.2 mA cm -2 , V OC of 0.805 ± 0.003 V, and FF of 0.704 ± 0.022, resulting in a PCE of 13.84 ± 0.46%. When the photoactive layer was formed and dried in ambient air with room-light illumination, J SC and FF decreased by approximately 6% and 13%, respectively, compared with the reference cells, resulting in a PCE of 11.07 ± 0.72% (P max /P max0 = 0.800). By blocking room-light illumination during the formation and drying of the photoactive layer, the cells exhibit a J SC of 23.5 ± 0.3 mA cm -2 , V OC of 0.803 ± 0.004 V, and FF of 0.713 ± 0.014, resulting in a PCE of 13.45 ± 0.46%, which is comparable to that of the reference cells (P max /P max0 = 0.972).
The influence of the room-light illumination during the drying process on the OPV performance was further evaluated (figure S2). The photoactive layers were dried in ambient air with and without room-light illumination for different durations, followed by vacuum drying and anode evaporation. Figure 2(c) shows the PCE obtained as a function of drying time. For both drying with room light and in the dark, a linear approximation with a negative slope to the logarithm of drying time is applicable. The absolute slope of the drying process in the dark was much smaller than that of the drying with room light. This shows that the drying of the photoactive layer in air with room light can be an important degradation factor, and this can be minimized by blocking the light. We also evaluated the continuous operational stability of PBDTTT-OFT:IEICO-4F OPVs formed in the glovebox and ambient air under dark conditions ( figure 2(d)).  The PCE change obtained from the maximum power point (MPP) tracking shows that the drying condition has a negligible effect on the operation stability. The EQE spectra of these devices were compared to understand the mechanism of degradation caused by room-light illumination during the drying of the photoactive layer film. The fullerene acceptor cells show a largely decreased EQE within the measured wavelengths ranging from 300 to 1000 nm ( figure 3(a)). Based on the EQE curves normalized by the maximum value of each condition, it can be observed that the decrease in EQE is more noticeable for wavelengths ranging from 400 to 600 nm ( figure S3(a)). According to the absorption spectra of the active layers, PC 71 BM mainly covers the absorption spectra ranging from 400 to 600 nm ( figure S3(c)) [31]. This implies that the exciton diffusion efficiency in the PC 71 BM domain becomes poor for the films dried in air with room light. It can be interpreted that the domain size of PC 71 BM becomes larger when the photoactive layer films were dried under room-light illumination; therefore, the excitons cannot reach the donor/acceptor interface. In the case of the PBDTTT-OFT:IEICO-4F acceptor, a minor decrease in EQE for all wavelengths was observed for the cells with the photoactive layer dried in air with room light (figures 3(b) and S3(b)). Conversely, the EQE was almost unchanged when the film was dried under light shielding in air. In addition to the EQE, the absorbance of the active layer was measured. For both PBDTTT-OFT:PC 71 BM and PBDTTT-OFT:IEICO-4F photoactive layers, the drying conditions had a limited influence on the absorption (figures 3(c) and (d)). This implies that the reduction in the EQE is not attributed to the absorption changes of the active layer film.
For both PBDTTT-OFT:PC 71 BM and PBDTTT-OFT:IEICO-4F photoactive layers, the drying conditions have limited influence on the absorption. For fullerene photoactive layers, the possible degradation of the materials was investigated using XPS and UPS ( figure S4). The C 1s, O 1s, F 1s, and S 2p spectra were identical between the samples of glovebox/glovebox and air (room light)/air (room light), indicating that light irradiation did not induce oxidation or partial decomposition of molecules. We evaluated the density of states (DOS) of the polymer in the BHJ film using UPS and PYS. The results are shown in figure S5. The secondary electron cut-off energies of these samples were almost identical ( figure S5(a)); no surface dipoles were induced by the degradation. The DOS near the highest occupied molecular orbital (HOMO) of these samples was also the same ( figure S5(b)), indicating that degradation did not cause unexpected charge doping of the polymer. The gap states were evaluated using PYS following a previous method [32]. The HOMO edges of these samples had a Gaussian shape, and no obvious gap states were observed ( figure S5(c)). These photoelectron spectroscopies strongly suggest that there was no material degradation by light irradiation.
We also compared the performance of OPVs obtained from photoactive layer solutions prepared under different environmental conditions. The presence of air and room light during the preparation of the photoactive layer solution had a negligible effect on the photovoltaic performance ( figure S6). The influence of humidity in the drying process on the device performance was also studied, in which the photoactive layer was dried in air under dark conditions for 30 min at a constant temperature of 25 • C and relative humidities of 30%, 60%, and 90% RH. The OPV performance was almost unchanged for the three humidity conditions (figure S7). However, this insensitivity to humidity is inconsistent with the results from a previous study [15], in which both FF and V OC degraded when the photoactive layer with PCBM was formed under high humidity conditions. Our results differ probably because the previous study did not consider the effect of room light.
Surface imaging using AFM for both the height and phase modes revealed that fullerene cells show topological differences under different formation/drying conditions. The air formation and drying led to a small increase in the dark region in the height image, and the grain boundaries became blurred (figures 4(a) and (b)). More notably, the phase images show that the width of the mesh-like network covering each domain increased when the active layer was formed in ambient air with room light. The brighter region in the phase mode exhibits softer and/or higher adsorption surfaces; therefore, the mesh-like network implies an aggregated polymer region. It is reasonable to consider that the donor or acceptor aggregates during the drying process in ambient air with room light. By contrast, the PBDTTT-OFT:IEICO-4F photoactive layers did not exhibit clear differences in the AFM images (figures 4(c) and (d)).
A series of experimental results suggested that the coexistence of air and room light during the transformation of the photoactive layer from a liquid to a solid film leads to the phase separation of the donor polymer and acceptor. Aggregation of fullerene acceptors is known to occur especially in photoactive layers with additives to enhance the solubility [33]. Furthermore, the degradation of the fabricated fullerene OPVs was accelerated by ultraviolet light illumination [27,34]. Conversely, IEICO-4F is a highly hypo-miscible system with poly[4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b;4,5-b']dithiophene-2,6-diyl-alt-(4-(2ethylhexyl)-3-fluorothieno [3,4-b] thiophene-)-2-carboxylate-2-6-diyl)] (PTB7-Th or PBDTTT-EFT). Such a system can suppress both de-mixing and crystallization, leading to the long-term stability of OPVs. The donor polymer PBDTTT-OFT used in this study has a similar chemical structure and a superior thermal stability to PTB7-Th [35], insensitivity to the drying conditions can be achieved. According to the comparison parameters summarized in table 1, the V OC of fullerene OPVs shows a larger change under different active layer formation conditions than non-fullerene cells. The lowest unoccupied molecular orbital of fullerene acceptors is sensitive to their aggregation conditions in solid states [36]. The change in V OC under different drying conditions also supports the morphological changes in the active layers.
Finally, the slot-die coating process was applied to the fullerene photoactive layer films (figure 5). The coating process was conducted in ambient air under dark conditions, and the obtained film was immediately dried in a vacuum chamber. The obtained photovoltaic parameters are J SC of 16.3 ± 0.2 mA cm -2 , V OC of 0.745 V, and FF of 0.722, resulting in a PCE of 8.75 ± 0.3%, which is almost comparable to that of reference cells having a spin-coated photoactive layer formed in the glovebox (table S4). This confirms that the roll-to-roll compatible fabrication process in ambient air can be fully utilized to achieve comparable performance to that of the controlled formation and drying of photoactive layer films.

Conclusion
In this study, we systematically investigated how the coexistence of air and room light during the formation of the photoactive layer affects the OPV performance. Compared to the fabricated devices that have completely dried photoactive layer films, performance deterioration in response to weak room light is much more rapid during the liquid-solid phase change. Weak room-light illumination, which has received less attention in performance studies, is an important degradation parameter during the drying of photoactive layer films. Based on these findings, we demonstrated that air-processed OPVs achieve comparable performance to that of OPVs formed in a nitrogen-filled glovebox by eliminating light during the process. The strong effects of both ambient air and weak light illumination during formation/drying are confirmed for both fullerene and non-fullerene BHJ films, and it was also confirmed that the air stability of the BHJ film is another important parameter for decreasing the degradation speed during the formation/drying process of photoactive layers. This suggests that other BHJ films can also be formed in ambient air under dark conditions if they are not extremely unstable in ambient air. By combining this technology with an air-stable photoactive layer and other materials, roll-to-roll manufacturing in ambient air can be fully applied for the fabrication of large-area and high-efficiency OPVs.