Organic solar cells (OSCs) have attracted huge attention because of their unique merits^[1-3]. In last few years, thanks to the design of new materials and device engineering, the power conversion efficiencies (PCEs) of OSCs have surpassed 18%^[4-8]. The PM6:Y6 cells are efficient binary cells, offering high PCEs over 16%^[9-11]. The high performance originates from the efficient free charge generation and the ground state dipole field at the donor–acceptor interface that promotes the exciton dissociation^[12]. To further boost the performance of PM6:Y6 cells, ternary architectures were adopted. Significant improvements in short-circuit current density (J_sc) and open-circuit voltage (V_oc) were realized. However, most of the ternary cells still suffer from low fill factor (FF) (generally <78%) (Table S1). The FF is generally determined by the competition between recombination and extraction of charge carriers^[13-15]. The interfacial electronic structures have nonnegligible impacts on charge transport in OSCs, also influencing the FF^{[16, 17]}. Previously, Bao et al. demonstrated that the energy of positive integer charge transfer (ICT) states (E_ICT+) of PM6 is equal to the energy of negative ICT states (E_ICT–) of Y6, leading to no potential step at the PM6:Y6 interface and an ideal binary host system^[18]. To avoid the formation of potential step in the ternary system, the E_ICT– of the second acceptor should be lower than (or equal to) E_ICT+ of the donor, thus suppressing the ICT trap-assisted recombination^[1]. In this work, we carefully incorporate a second acceptor EH-IDTBR into the host PM6:Y6 blend (Fig. 1(a)). From the ultraviolet photoelectron spectroscopy (UPS)-derived energy levels (Fig. 1(b)), the negative ICT states of EH-IDTBR (E_ICT– = 4.25 eV) is smaller than the positive ICT states of PM6 (E_ICT+ = 4.5 eV), suggesting no ground state charge transfer at the PM6:EH-IDTBR interface, thus avoiding interfacial ICT trap-assisted recombination.






onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2021/9/PIC/21070011-1.jpg'"
class="figure_img" id="Figure1"/>



Download



Larger image


PowerPoint slide






Figure1.
(a) Chemical structures of PM6, Y6 and EH-IDTBR. (b) Relevant energy levels. (c) J?V curves for PM6:Y6, PM6:EH-IDTBR, and PM6:EH-IDTBR (5% w/w):Y6 OSCs. (d) EQE spectra. (e) FF vs PCE plots for PM6:Y6-based ternary cells. (f) Charge carrier lifetime (τ) as a function of charge density (n). (g) sEQE and EL spectra for the optimal ternary cell. The extended sEQE (orange line) is determined by EL and the blackbody emission (BB). (h) EQE_EL–current plots for binary and ternary cells.




The ternary and reference binary OSCs were fabricated with a structure of ITO/PEDOT:PSS/active layer/PFN-Br/Ag. A series of ternary OSCs with different EH-IDTBR content in acceptors were studied. The ternary cells with 5%(wt) EH-IDTBR gave a PCE of 17.59% (certified 16.9%, Fig. S6), with simultaneously improved V_oc of 0.853 V, J_sc of 26.03 mA/cm², and FF of 79.22% (Fig. 1(c), Fig. S3(a), Fig. S4 and Table S2). The integrated photocurrent densities from external quantum efficiency (EQE) spectra agree with the J_sc from J–V measurements (Fig. 1(d), Fig. S3(b) and Table S2). We analyzed FF and PCE distributions of total 60 devices in Fig. S5, and a good reliability is demonstrated. The higher efficiency of ternary cells than binary cells is mainly attributed to the increased FF. The 79.22% FF is among the highest values for OSCs containing PM6 and Y6 (Fig. 1(e) and Table S1).



To understand the working mechanism behind the impressive FF, we performed comprehensive characterization. A stronger photoluminescence (PL) quenching in ternary blend film was observed, suggesting more efficient exciton dissociation in the ternary blend film (Fig. S7)^[19]. The PL spectrum of EH-IDTBR overlaps with the absorption spectrum of Y6 (Fig. S8(a)). The enhanced PL at 906 nm (emission from Y6) in EH-IDTBR (5% w/w):Y6 blend film indicates that there is energy transfer from EH-IDTBR to Y6 (Fig. S8(b))^[20]. The J_sc of EH-IDTBR (5% w/w):Y6 cell is much larger than corresponding single-component cells, suggesting that there is charge transfer between EH-IDTBR and Y6 (Fig. S9)^[21]. Femtosecond (fs) transient absorption (TA) spectroscopy was employed to probe hole-transfer dynamics in the blend films (Fig. S10, Fig. S11 and Fig. S12). The PM6:EH-IDTBR (5% w/w):Y6 ternary blend film exhibits faster transfer rate with the time constants of τ_{1, h} = 88 fs and τ_{2, h} = 10.41 ps (τ_{1, h} = 119 fs and τ_{2, h} = 13.11 ps for PM6:Y6 binary blend film), respectively. The faster hole transfer might be due to the increased donor/acceptor HOMO offset caused by the introduction of EH-IDTBR into PM6:Y6 blend, which enhances the hole-transfer driving force. The exciton dissociation and charge extraction probabilities of the ternary cells are 97.05% and 91.05%, respectively, which are larger than that of binary cells (Fig. S13 and Table S3), consisting with the higher FF in ternary cells. The J–V characteristics of solar cells under various light intensities (P_light) were used to study charge recombination kinetics (Fig. S14). For J_sc vs P_light plots, the α values for binary PM6:Y6 and ternary PM6:EH-IDTBR:Y6 cells are both 1.00, suggesting negligible biomolecular recombination in both devices (Fig. S15(a)). For V_oc vs P_light plots, the ternary cells show a diminished slope of 1.11 as compared to PM6:Y6 cells (1.19), suggesting that EH-IDTBR can effectively suppress the trap-assisted recombination in the ternary cells (Fig. S15(b)). The space charge limited current (SCLC) method was applied to estimate the electron and hole mobilities (μ_e/h) in binary and ternary blend films (Fig. S16 and Fig. S17). Compared with PM6:Y6 film, the μ_e and μ_h for ternary blend film increase from 4.31 × 10^–4 and 3.74 × 10^–4 cm²/(V·s) to 4.54 × 10^–4 and 4.43 × 10^–4 cm²/(V·s), respectively, and the μ_e/μ_h ratio decreases from 1.15 to 1.03 (Table S4). The addition of EH-ITDBR could balance the charge transport, which is beneficial to FF and PCE^[18]. We further studied charge carrier lifetime (τ) and charge carrier density (n) for the ternary and reference binary cells by using transient photovoltage (TPV) and transient photocurrent (TPC) methods (Fig. S18 and Fig. S19). As shown in Fig. 1(f), the ternary cells show higher τ over the whole n range, indicating suppressed charge recombination. The corresponding recombination exponent (λ) for the ternary device is 0.99, indicating a nearly ideal (trap-free) bimolecular recombination process^[22].



The atomic force microcopy (AFM) and transmission electron microcopy (TEM) images show that EH-IDTBR is finely mixed with the host PM6:Y6 blend, and the PM6:EH-IDTBR (5% w/w):Y6 film features fibrillar structures (Fig. S20 and Fig. S21). Grazing incidence wide-angle X-ray scattering (GIWAXS) is conducted to investigate the effect of EH-IDTBR on molecular packing (Fig. S22 and Table S5)^[23]. In ternary blend film, the π–π stacking d-spacing (3.54 ?) is slightly shorter than that of PM6:Y6 film (3.55 ?) and the crystal coherence length (CCL) increases from 20.55 to 21.91 ?, indicating improved crystallinity by incorporating EH-IDTBR. Besides, the corresponding diffraction peak is more intense (Fig. S23), suggesting more ordered face-on orientation. The enhanced crystallinity and more ordered face-on orientation favor charge transport, thus leading to high FF^{[14, 24]}.



The energy loss was analyzed according to the framework of Marcus theory^{[25, 26]}. The total energy loss (E_loss) is described as: $ {E_{
m{g}}} - e{V_{{
m{oc}}}} = ({E_{
m{g}}} - {E_{{
m{CT}}}}) + Delta {E_{{
m{rad}}}} + Delta {E_{{
m{non-}}{
m{rad}}}} $ (E_g is optical bandgap; E_CT is the energy of charge transfer state; ΔE_rad and ΔE_non-rad represent the radiative and non-radiative recombination loss, respectively.). From binary to ternary cells, E_loss decreased from 0.579 to 0.563 eV. E_CT is determined by fitting the sub-gap sensitive EQE (sEQE) and EL (Eqs. (S11) and (S12), Fig. 1(g), Figs. S24 and S25, Tables S6 and S7)^[27]. The ternary device yields a smaller charge extraction loss (E_g – E_CT) of 0.082 eV as compared to the binary device (0.09 eV). ΔE_non-rad is quantified by EQE_EL measurements, expressed as –kTln(EQE_EL)^[28]. Fig. 1(h) indicates that the ternary device features a higher EQE_EL of 1.234 × 10^–4 than the binary device (6.344 × 10^–5). This led to smaller non-radiative recombination loss (ΔE_non-rad) of 0.233 eV in ternary cells. Therefore, the higher V_oc of the ternary cells can be attributed to the reduced charge extraction loss and non-radiative recombination loss.



In summary, efficient ternary OSCs were made by blending a polymer donor PM6 with two non-fullerene acceptors EH-IDTBR and Y6. A PCE of 17.59% and an impressive FF of 79.22% were obtained. Our results indicate that improving the FF via enhancing charge transport and extraction and suppressing recombination losses is a quite effective approach for boosting PCE.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (21875067, 11604099, 51811530011), the Fundamental Research Funds for the Central Universities, Shanghai Rising-Star (19QA1403100), and East China Normal University Multifunctional Platform for Innovation. M. Fahlman thanks the STINT grant (CH2017-7163). Q. Xue thanks the National Natural Science Foundation of China (51803060) and the Science and Technology Program of Guangdong Province (2018A030313045). L. Ding thanks the National Key Research and Development Program of China (2017YFA0206600) and the National Natural Science Foundation of China (51773045, 21772030, 51922032, 21961160720) for financial support. We thank Prof. Lin Sun for fruitful discussion. Q. Bao also thanks the open project of State Key Laboratory of Luminescent Materials and Devices (2021-skllmd-07).

Appendix A. Supplementary materials

Supplementary materials to this article can be found online at https://doi.org/1674-4926/42/9/090501.