GIWAXS 2D diffraction patterns were collected to compare the molecular orientations of the SC and VESD P3HT:PCBM films. Figure 2 shows the GIWAXS results for both the SC and VESD P3HT:PCBM films before and after annealing. P3HT typically exhibits two primary orientations: face-on and edge-on26. In Fig. 2(a), the unannealed SC film displays (h00) diffraction peaks attributed to the lamellar structure with a spacing of ~ 1.6 nm in the out-of-plane direction. These peaks are located at qz ~ 0.38 and 0.75 Å−1, corresponding to the primary (100) and secondary (200) reflections of P3HT27,28. This finding suggests that P3HT adopts a thermodynamically favorable edge-on orientation early in the SC process29,30. Additionally, a broad, diffuse symmetric ring was observed at q ~ 1.4 Å−1, indicating the presence of randomly oriented nanoscale PCBM domains31,32. Ring patterns at q ~ 1.5, 2.2, and 2.5 Å−1 are attributed to the ITO substrate33. After annealing [Figure 2(b)], the diffraction patterns sharpened, reflecting improved crystallinity. A weak (010) π-π stacking peak in the out-of-plane, with a spacing of ~ 3.7 Å and located at qz ~ 1.7 Å−1, also appears. Consequently, the annealed SC film exhibits a predominantly edge-on orientation in the mixed phase.
In contrast, the unannealed VESD film showed no significant diffraction patterns in either the out-of-plane or in-plane directions [Figure 2(c)]. However, after annealing [Figure 2(d)], there was a remarkable improvement in both P3HT crystallinity and the formation of nanoscale PCBM domains. Notably, distinct (010) π-π stacking in the out-of-plane and (h00) lamellar stacking in the in-plane diffraction patterns emerge. The prominent (010) π-π stacking peak indicates a substantial increase in face-on orientation, which is likely to enhance vertical charge transport due to stronger π-π interactions. These results clearly demonstrate that VESD, when combined with annealing, yields better crystallinity and a more pronounced face-on orientation compared to SC. The rapid solvent drying characteristics of VESD may have contributed to this difference.
Figure 3 shows the UV-vis absorption spectra of the P3HT:PCBM films deposited via SC (blue) and VESD (red) before (dashed line) and after (solid line) annealing. All four spectra display absorption features originating from PCBM at approximately 345 nm and P3HT at approximately 500 and 600 nm. For both the SC and VESD films, annealing enhanced the absorption intensity of P3HT. The shoulder around 600 nm, known as interchain absorption and attributed to ordered P3HT domains34,35, is accompanied by a red shift in the absorption feature near 500 nm. The intensity of the absorption at 600 nm served as an indicator of P3HT ordering and crystallization. This feature behaves differently depending on the film preparation method (SC or VESD). Before annealing, the SC film exhibited a small but distinct shoulder at 600 nm, suggesting the crystallization of P3HT during the SC process. In contrast, the VESD film exhibited no such signal, indicating that it was amorphous. After annealing, the VESD film demonstrated a higher overall absorption, including the shoulder at 600 nm, than the SC film. These results are in good agreement with the GIWAXS results, suggesting a higher crystallinity and molecular ordering in the VESD film.
Based on the GIWAXS and UV-vis results, the initial state of the VESD film before annealing appeared to be almost amorphous, with highly mobile P3HT chains. These P3HT chains likely exhibit stronger intermolecular interactions than interactions with the substrate, adopting a face-on orientation after annealing. Further studies are required to understand the fundamental mechanisms underlying the molecular orientation of organic films prepared using VESD.
As mentioned previously, the VESD method effectively eliminates the solvent from the injected solution droplets through differential pumping before they reach the substrate, thereby helping to preserve the interface. To evaluate the effect of the residual solvent, cross-sectional TEM images were obtained for both the anode/light-absorbing layer and the light-absorbing layer/cathode interfaces. Figure 4 shows TEM images of the ITO/MoO3/P3HT:PCBM (anode) and P3HT:PCBM/LiF/Al (cathode) interfaces. In the SC sample [Figure 4(a)], interfacial mixing between the MoO3 and P3HT:PCBM layers is evident as a darker region, which is probably caused by the interaction of the residual solvent with the underlying MoO3 layer. Similarly, at the cathode interface, the evaporation of the residual solvent through the Al layer results in damage, which is also manifested as a darker region. In contrast, the VESD method deposits a P3HT:PCBM layer onto MoO3 without residual solvent, thereby maintaining an intact interface, as shown in Fig. 4(b). The same is true for the cathode interface.
To determine the elemental composition of the interfacial region, EDS measurements were conducted in two areas: near the anode (labeled 1) and near the cathode (labeled 2). As shown in Fig. 4(c), the Mo and Al contents were significantly higher in the SC sample than in the VESD sample. This increase can be attributed to the residual solvent, which facilitates the diffusion of materials from the hole extraction and cathode layers into the light-absorbing layer. Such diffusion hinders charge transport and reduces the operational lifetime because these mixed-element regions act as charge trap sites, accelerating device degradation.
To investigate the effects of face-on orientation and interface preservation of the VESD film on the OPV performance, devices with an ITO/MoO3/P3HT:PCBM/LiF/Al structure were fabricated by incorporating either the VESD or SC light-absorbing layer. The J–V characteristics under AM 1.5G, 1-sun illumination are presented in Fig. 5, and the corresponding photovoltaic parameters are summarized in Table 1. The VESD OPV exhibited a higher JSC of 8.51 mA cm− 2 and a FF of 60.0% compared to the SC OPV, which had a JSC of 7.94 mA cm− 2 and a FF of 57.9%. Both devices exhibited the same open-circuit voltage (VOC) of 0.61 V. Consequently, the VESD OPV achieved a higher PCE of 3.11% compared to 2.80% for the SC OPV. Based on the measured J–V characteristics, the series resistance (RS) of the VESD OPV was calculated to be 8.84 Ω cm2, which is lower than that of the SC OPV (10.31 Ω cm2). This reduction in RS is primarily attributed to the improved charge transport facilitated by the face-on stacking geometry and the intact P3HT:PCBM/electrode interfaces with minimal trap sites. These advantages of the VESD film contributed to the higher JSC and FF values observed for the VESD OPV compared to the SC OPV.
The device lifetimes of SC and VESD OPVs were also compared. The OPVs were stored under ambient conditions without encapsulation, and their PCEs were measured after 3, 6, 9, and 12 days. As shown in Fig. 5(b), the PCE of the VESD OPV degraded much slower than that of the SC OPV. This difference in device lifetime can be attributed to the interfacial diffusion of Mo and Al, as observed in the TEM images. The diffusion of Mo and Al atoms into the SC BHJ film creates charge-trap sites, accelerating the degradation of the light-absorbing layer. In contrast, the VESD-BHJ film maintains an intact interface with fewer trap sites, resulting in a significantly longer device lifetime.
Table 1
Photovoltaic parameters of SC and VESD OPVs.
| JSC (mA cm− 2) | VOC (V) | FF (%) | PCE (%) | RS (Ω cm2) |
SC | 7.94 | 0.61 | 57.9 | 2.80 | 10.31 |
VESD | 8.51 | 0.61 | 60.0 | 3.11 | 8.84 |
The impact of traps on the charge dynamics was further investigated using impedance spectroscopy to assess the electron lifetime, which is a key factor in determining the current loss in photovoltaic devices. Figures 6(a) and 6(b) show the Nyquist plots of the SC and VESD OPVs under forward bias in the dark. The electron lifetime values were extracted by modeling an equivalent circuit that incorporated the diffusion-recombination mechanism, as shown in Fig. 6(c)36. The equivalent circuit comprised several components: contact and wiring resistance (Rc), electron transport resistance (rt), electron recombination resistance (rrec), chemical capacitance due to excess electrons under bias (cn), and geometrical capacitance (Cg). The fitting lines generated by the equivalent circuit closely match the experimental data plotted in Figs. 6(a) and 6(b).
From the impedance spectroscopy results, the electron diffusion time was calculated as the product of rt and cn, whereas the electron lifetime was derived from the product of rrec and cn. Using the Nernst-Einstein relationship, the electron mobility of the SC OPV was determined to be 2.1 × 10− 3 cm2 V− 1 s− 1, aligning with values reported in the literature36,37. In contrast, the VESD OPV exhibited significantly higher electron mobility of 6.2 × 10− 3 cm2 V− 1 s− 1, approximately three times greater than that of the SC OPV. Furthermore, as depicted in Fig. 6(d), the electron lifetime of the SC OPV was calculated to be 0.20 ms at 0.50 V, in agreement with previous reports36,37. This value decreases with increasing bias voltage. In comparison, the electron lifetime of the VESD OPV was 0.66 ms at 0.50 V, more than three times higher than that of the SC OPV and consistently higher across all measured bias regions. These results can be attributed to the reduced charge traps, as observed in the TEM images (Fig. 4), and they correlate well with the PCE and device lifetime results, as shown in Fig. 5. Collectively, all the data clearly demonstrate the superior performance of the VESD OPV compared with that of the SC OPV.