1. Introduction
Organic solar cells (OSCs), noted for their light weight, solution processability, and compatibility with flexible substrates, have become a focal point for researchers worldwide [
1,
2,
3,
4]. Currently, the state of the art single-junction OSCs have achieved a power conversion efficiency (PCE) exceeding 19% [
5,
6,
7]. High-efficiency OSCs primarily employ the bulk heterojunction (BHJ) structure, which is spontaneously formed via phase separation from a mixed solution of electron donors and acceptor during the film deposition process. To optimize the microstructure morphology, researchers have developed a range of strategies, such as thermal annealing (TA), solvent vapor annealing (SVA), and additive engineering [
8,
9,
10,
11,
12]. However, the photovoltaic performance of BHJ OSCs is highly dependent on molecular crystallinity, molecular orientation, phase separation, and vertical phase distribution within the active layer, which is often unpredictable and sensitive to material properties and processing conditions. This unpredictability fabrication challenges in scaling up for large-scale industrial production and limits their potential for commercial applications.
In recent years, the layer-by-layer structure, fabricated by depositing electron donor and acceptor in separate solutions, allows for controlled microstructure morphology and vertical component distribution, and also offers a reproducible and efficient technique for fabricating large-scale OSCs [
13,
14,
15]. During LBL deposition process, the interfacial diffusion of electron donor and acceptor can lead to the formation of a p-i-n type structure within the active layer [
16,
17,
18]. Such LBL-type devices are valued for their ease of reproducibility and reduced carrier recombination. By incorporating a wax additive to form nanoscale pores with the PM6 layer, a novel interdigitated heterojunction structure based on PM6/L8-BO bilayer heterojunction OSCs was successfully fabricated via LBL deposition method. This novel structure result to an ideal vertical phase distribution inside the active layer, leading to efficient exciton diffusion length and dissociation, and the reduced charge recombination [
19]. By adjusting the solution temperature and annealing processes, the D18 polymer's pre-aggregation behavior in solution can be controlled, leading to manipulation of the microstructure of the D18 bottom layer. This optimized bottom layer effectively facilitates the formation of suitable networks in the L8-BO upper layer for efficient charge transport and deliver an enhanced PCE of 18.02% [
20].
In the array of methods for regulating microstructure morphology of active layer, the use of additives stands out as a straightforward yet potent method to regulate the molecular packing, enhance the crystalline and improve the blend film morphology [
21,
22]. For the study on the additives, it is essential to delve deeper into the mechanisms how these additives influence on the film morphology evolution. For example, the introduction of a selective solvent as an additive can swell the donor domains and improve the donor-acceptor interfacial area, facilitating a more efficient charge separation and extraction processes. In addition to the chemical interactions, the physical properties of additives such as their boiling point and evaporation rate can also be critical. Solvent additive may remain in the film for a longer time during the film formation process, facilitating the molecules to self-assemble into a well-structured microstructure morphology [
23,
24]. The reported study have demonstrated the effectiveness of additive to fine-tune the morphology of the active layer, resulting to more ordered structure and higher charge mobility. For instance, a solvent additive of 1% 1,8-diiodoctane (DIO) was employed as the solvent additive to treat the J71:N2200 blends, resulting in more favorable phase separation and domain size. As a result, the corresponding OSCs device achieved an outstanding PCE of 9.34% with an ultrahigh fill factor (FF) of 77.86% [
25]. Furthermore, by replacing DIO with diiodomethane in PM6:L8-BO based OSCs, the energetic difference between the single excited state of L8-BO and the charge transfer state in blend film was effectively reduced, while the microstructure morphology and charge transport of the optimized blend film was not deteriorated, and thereby an increased open-circuit voltage (
VOC) [
26].
In this study, the polymers D18 and PYIT-OD were employed as donor and non-fullerene acceptor materials, respectively, for the fabrication of all-polymer solar cells (all-PSCs) using the LBL method. In addition, we systematically study how solvent additives influenced the photovoltaic performance of all-PSCs. The addition of 2% chloronaphthalene (CN) additive led to a PCE of 15.07%. This study also investigates the impact of different solvent additives on the exciton dynamics, molecular crystallinity and microstructure morphology within the D18/PYIT-OD based LBL all-PSCs devices. Our study demonstrated the great potential of additive engineering via LBL fabrication methods in regulating microstructure of active layers, suppressing carrier recombination and enhancing the photovoltaic performance of devices.
2. Result and Discussion
The chemical structure of D18 and PYIT-OD was illustrated in
Figure S1.
Figure 1a shows the extinction coefficient of the D18/PYIT-OD bilayer film fabricated from different processing conditions. The introduction of additives noticeably enhanced the extinction coefficient of the film, with values reaching 4.8×10
4 cm
-1, 6.9×10
4 cm
-1, and 6.1×10
4 cm
-1 under three different processing conditions. This indicates an improvement in the molecular packing of the film due to the presence of additives. Since 585 nm is the main absorption peak of D18, our main focus is on the changes at this position under different additive conditions. The absorption peak ratio Ι
0-0/Ι
0-1 of the D18 film under three different processing conditions rose from 0.99 in the controlled device to 1.06 and 1.02 for bilayer films with DIO and CN additives, suggesting that the D18 film exhibited a more uniform π-π stacking with the presence of additives. The increase in crystallinity and the optimization of molecular packing are likely to enhance charge transport and photovoltaic performance. To verify whether the DIO and CN additives could enhance the photovoltaic performance of OSCs, we fabricated devices using the LBL method with a structure of ITO/PEDOT:PSS/D18/PYIT-OD/PDIN/Ag. DIO and CN additives were added to the PYIT-OD solution in measured by volume percentage.
Figure 1b and
Figure S2 show the representative current density-voltage (
J-
V) characteristics of the devices at different concentrations of solvent additives, with specific photovoltaic parameters presented in
Table 1 and
Table S1. The PCE for the control device (without any additive) was 12.91% with a short-circuit current density (
JSC) of 20.44 mA cm
-2, an open-circuit voltage (
VOC) of 0.956 V, and a fill factor (FF) of 66.06%. When DIO or CN additive was added, the PCE of the devices first increased and then decreased, reaching maximum PCE of 15.07% and 13.72% at 2% CN and 1% DIO concentrations, respectively. The corresponding
JSC was 22.00 and 21.38 mA cm
-2,
VOC was 0.969 and 0.947 V, and FF was 70.71% and 67.73%, all of which were superior to the control devices. In order to validate the accuracy of the
J-V measurements, the external quantum efficiency (EQE) spectra for all devices were measured and illustrated in
Figure 1c and
Figure S3. The
JSC values derived from the EQE spectra aligned with those obtained from
J-V measurements. All devices exhibited a wide photo-response range spanning from 400 to 850 nm. In particular, when 2% CN was added into PYIT-OD solution, the EQE values between 450 to 575 nm and 725 to 825 nm were significantly higher than that of other devices.
To investigate the excitons dissociation and charge collection within devices, we conducted experiments to assess the relationship between photo-generated current (
Jph) and the effective voltage (
Veff), as shown in
Figure 1d. Here, the photo-generated current (
Jph) and the effective voltage (
Veff) are calculated by
Jph=J
L-J
D and
Veff=V
0-V, respectively, where J
L and J
D refer to the current density under illumination and dark conditions, V
0 is the voltage when
Jph is zero, and V is the applied voltage [
27,
28]. Under a large
Veff, the
Jph tends to be saturated (
Jsat), at which point nearly all excitons are dissociated into free carriers and collected by the electrodes [
29]. We also analyzed the probability of charge dissociation (P(E,T)) of devices under different processing conditions (
Table S2), which was calculated by the ratio of
Jph to
Jsat under short-circuit conditions. For the devices without additives, with the addition of 1% DIO additive and 2% CN additive, the P(E,T) values were 89.3%, 91.4%, and 94.9%, respectively. Based on the equation
Jsat=q
GmaxL, we calculated the corresponding maximum exciton generation rates (
Gmax) was to be 1.31×10
28, 1.32×10
28, and 1.38×10
28 m
-3 s
-1, respectively. Devices with 2% CN additive showed increased
Gmax, which correlated well with their higher
JSC, potentially contributing to the improved light absorption and more orderly molecular packing. These results demonstrate that devices with 2% CN additive exhibit superior performance in exciton and charge dynamics.
To delve into the charge recombination in different devices, we performed
J-V under various light intensities (
Plight) and plotted the correlation curves of
JSC and
VOC with
Plight. The relationship between
JSC and
Plight follows the formula
JSC∝(
Plight)
S, where S reflects the intensity of bimolecular recombination in the device. Generally, The S value being very close to the unit suggests that OSC devices experience minimal bimolecular recombination [
30]. As shown in
Figure 2a, for devices without additive, with 1% DIO additive, and with 2% CN additive, the corresponding S values are 0.96, 0.97, and 0.98, respectively. The results indicate that devices based on D18/PYIT-OD with 2% CN are most effective in suppressing bimolecular recombination. Meanwhile, the
VOC versus
Plight curves were shown in
Figure 2b, the controlled device exhibited the highest fitting slope, suggesting severe trap-assisted Shockley-Read-Hall (SRH) or bimolecular recombination. For the device with 1% DIO additives and 2% CN additives, the slopes were decreased to 1.60 kBT/q and 1.52 kBT/q, respectively. The smallest fitting slope for the device with CN additive indicated that SRH or bimolecular recombination is effectively suppressed, promoting effective charge transport and collection.
Single carrier devices based D18/PYIT-OD bilayer films were fabricated to investigate the charge transport behavior by the space charge-limited current (SCLC) method [
31], and summarized data was listed in
Table S3. The specific device structure for the hole-only device is ITO/PEDOT:PSS/D18/PYIT-OD/Ag, while the electron-only device is ITO/ZnO/D18/PYIT-OD/PFN-Br/Ag. As shown in
Figure 2c,d, the hole mobility (μ
h) and electron mobility (μ
e) for the device without additive are 3.71×10
-4 cm
2 V
-1 s
-1 and 4.63×10
-4 cm
2 V
-1 s
-1, respectively. After adding additives to PYIT-OD, the measured μ
h increased to 7.66×10
-4 cm
2 V
-1 s
-1 (1% DIO) and 9.34×10
-4 cm
2 V
-1 s
-1 (2% CN), while the μ
e increased to 6.93×10
-4 cm
2 V
-1 s
-1 (1% DIO) and 8.18×10
-4 cm
2 V
-1 s
-1 (2% CN). It is evident that the device treated with 2% CN additive not only has more efficient charge transport performance but is also more balanced charge transport, indicating an effective reduction in charge accumulation and recombination, thereby improving the
JSC of devices. In addition, exciton separation and charge transfer at the electron donor and acceptor interface are crucial for device photovoltaic performance. To investigate the exciton separation and charge transfer of bilayer films, we performed steady-state photoluminescence (PL) spectra of the pristine D18 film and the D18/PYIT-OD bilayer films, as shown in
Figure S4. The films were excited at 467 nm, and D18 exhibited typical emission peaks at 620 and 680 nm. The PL intensity at these emission peaks was significantly quenched in the D18/PYIT-OD bilayer films, indicating efficient charge transfer at the electron donor and acceptor interface. For the bilayer film with added 2% CN, this PL quenching is most pronounced, indicating most efficient charge transfer.
We used atomic force microscopy (AFM) to observe the influence of additives type and their concentrations on the surface morphology of bilayer films. As shown in
Figure 3, the root-mean-square (RMS) roughness value is 1.68 nm for the bilayer film without any additive. With increasing DIO concentration from 0.5% to 2%, RMS roughness values rose from 1.87 nm to 3.43 nm. When the CN concentration increased from 1% to 2%, the RMS roughness values slightly decreased from 1.70 nm to 1.54 nm, exhibiting a more refined structure. However, excessive additive (3% CN concentration) led to large-scale aggregation, resulting in a sharp increase in film RMS roughness, which severely damaged device efficiency. Additionally, by examination of the transmission electron microscopy (TEM) images for both the controlled film and those added with 1% DIO and 2% CN (
Figure 4), it becomes apparent that these additives precipitate morphological inhomogeneity within the film, likely attributable to the induced crystallization. Comparative analysis indicates that the addition of 2% CN results in a more uniform and fiber-like film structure, leading to higher FF in the corresponding devices. This also aids in effective separation of excitons and charge transport.
To gain a deeper understanding of the molecular packing and crystalline properties within the active layer, we utilized grazing-incidence wide-angle X-ray scattering (GIWAXS) measurements to examine the D18/PYIT-OD bilayer films under various processing conditions.
Figure 5a shows the 2D GIWAXS patterns of D18/PYIT-OD bilayer films without any additive (control), with 1% DIO, and with 2% CN. The corresponding line-cut plots in the in-plane (IP) and out-of-plane (OOP) directions were shown in
Figure 5b,c, respectively. From the GIWAXS 2D GIWAXS patterns under three different processing conditions, we observed a strong (010) π-π stacking peak in the OOP direction and a distinct (100) lamellar stacking peak in the IP direction, This indicates that D18/PYIT-OD bilayer films predominantly exhibited a face-on orientation. This result also implied that fabrication the upper layer of PYIT-OD does not significantly alter the molecular orientation of the D18 film [
32]. For the bilayer film with 1% DIO, the diffraction peak intensities are more pronounced for both the IP lamellar stacking and the OOP π-π stacking. In addition, It is evident that the bilayer film containing 2% CN exhibits a minimum full width at half maximum (FWHM) value of 0.292 Å at 1.69 Å
-1 in the OOP direction, indicating a strong π-π stacking. In comparison, the FWHM values for the controlled bilayer film and the film with 1% DIO are 0.352 Å and 0.317 Å, respectively. The reduced FWHM value in the film with 2% CN suggests a higher crystal coherence length (CCL). A larger CCL suggested an increase in the number of crystal repeating units within the active layer, which is more conducive to the charge transport. This is consistent with the results above where the OSC device containing 2% CN exhibited the highest hole and electron mobility. The bilayer film that incorporated the 1% DIO additive suffered from excessive aggregation, leading to compromised photovoltaic performance of the OSC device.