The power conversion efficiency (PCE) for donor–acceptor bulk-heterojunction organic solar cells (OSCs) has reached ~20%[1–3], approaching that for inorganic solar cells, due to the development of key photoactive materials[4–14]. The short-circuit current density (Jsc) and the fill factor (FF) for state-of-the-art OSCs are already close to the thermodynamic upper limit predicted by Shockley-Queisser theory[15], but the open-circuit voltage (Voc) is low, limiting the overall performance of OSCs.
LowVoc can be ascribed to the high radiative (Vr) and non-radiative (Vnr) voltage losses associated with the decay of photogenerated charge carriers[16]. Thus, understanding the mechanism of charge carrier decay dynamics in OSCs is the focus for device physicists. In 2009, Vandewalet al. analyzed the energetics of excited states in polymer-fullerene solar cells and revealed that the decay of charge carriers in OSCs occurredvia the charge transfer (CT) states formed at the polymer-fullerene interface[17]. Therefore, the high voltage loss was associated with the undesired electronic properties of CT state. Later, Vandewalet al. demonstrated that the key parameters determining the voltage loss included the energy of CT state (Ect), the reorganization energy (λ) of organic photovoltaic materials, and the absorption oscillator strength (fosc) of CT state[18].
Owing to the existence of CT states with energy lower than the singlet (S1) state of the pristine photovoltaic materials, the absorption tail of OSCs extends into the long wavelength region. This leads to a very high saturation current[19], yielding highVr. Reducing the energy difference between the CT state and the S1 state (ΔEct) is an effective approach to reduceVr[19]. However, the reduced ΔEct often led to reduced exciton dissociation rate and limited device quantum efficiency[20]. Reducing the density of CT state (Nctc) can also effectively reduceVr[21]. The reducedNctc resulted in deteriorated transport properties of charge carriers, thus limiting FF. HighVr was believed to be an intrinsic problem for OSCs[22].
In non-fullerene solar cells, high quantum efficiency could be obtained by using donor–acceptor blends with very low ΔEct or lowNctc[23]. Therefore, negligibleVr could be realized, and the major limit forVoc isVnr, which is associated with low external quantum efficiency of electroluminescence (EQEEL) for CT state[24], since
wherek is Boltzmann constant,q is elementary charge, andT is temperature. EQEEL is determined by following formula[25]:
wherekr andknr are the radiative and non-radiative decay rate constant, respectively. Thus, very highknr for CT state in organic blends leads to low EQEEL. Low EQEEL (<1 × 10–5) for organic solar cells yields highVnr (>0.3 V)[26]. Accordingly, more efforts have been spent on the manipulation of the decay dynamics of CT state in recent years to reduceknr and increase EQEEL. In 2017, Bunduhnet al. discovered that highknr of OSCs originated from strong vibrational coupling between CT state and the ground state[27]. The strong coupling resulted from high-frequency carbon vibration of organic molecules. Later, Ullbrichet al. reported that increasingEct could reduce vibrational coupling and reduceknr[28].Vnr was very low in OSCs with highEct. However, the use of high-bandgap material for highEct led to reduced spectral coverage of the solar cell, and thus reducingJsc and limiting overall device performance. Increasing the spacing between donor and acceptor molecules could also reduceknr and reduceVnr, thus improving overall device performance[29]. Furthermore, Azzouziet al. extended the model describing non-radiative decay rate of CT state[30], and demonstrated that reducingλ andfosc, or increasing static dipole moment of CT state, could reduceknr.
Qianet al. indicated that increasing the degree of hybridization between CT state and S1 state could effectively increase EQEEL[31]. The increase in EQEEL was ascribed to the intensity-borrowing mechanism of the excited states. Later, Eisneret al. built a model to describe the impact of hybridization on the dynamic process of excited states[32]. They found thatVnr could be reduced, while increasingVr. It is still unclear whether the hybridization could improve overall device performance.
Based on the improved understanding onVoc loss mechanism, many strategies, such as ternary strategy[33], mixed-solvent strategy[23], and thin-film deposition strategy[34], as well as material design strategies (e.g. side-chain engineering[35], double-cable structure[36]) have been developed to reduce voltage loss. Now the lowest voltage loss is below 0.4 V[37]. OSC performance will be further improved[38,39], thus paving the road to real commercialization.