Exciton Binding Energy of Non‐Fullerene Electron Acceptors

Abstract

The past three years have witnessed the power conversion efficiency (PCE) of organic solar cells (OSCs) rocketing to over 18%, due to outstanding advantages of non‐fullerene acceptors (NFAs). However, large exciton binding energy (E _b) caused by strong Coulombic force is still one of the main limiting factors for high‐performance OSCs. Thus, it is critical to reduce the E _b for further enhancement of device performance. Many strategies have been developed to reduce the E _b of organic materials previously. In this perspective, the calculation methods for E _b and the relationship between E _b and voltage loss (V _loss) are discussed. Then, the effects of the properties of small‐molecule acceptors on E _b from the perspectives of fused‐ring donor cores, end groups, side chains, and molecular packing are discussed. Finally, the potential directions for reducing E _b and pointing out the trade‐off between E _b and bandgaps/miscibility are put forward. It is hoped that this perspective could provide a new thinking of a molecular design for the breakthrough of OSCs.

Introduction

Bulk heterojunction (BHJ) organic solar cells (OSCs), consisting of mixed donor (D) and acceptor (A) in the active layer, have attracted considerable attention, due to their potential advantages, such as portability, mechanical flexibility, and roll-to-roll machinability.^{[ 1–4 ]} Upon excitation by incident light, an exciton, rather than free charge carriers, is generated in the active layer,^{[ 5 ]} which is ascribed to the low relative dielectric constant (ε _r < 3–4) of organic semiconductors. Then, the exciton, bound by strong Coulombic force, can dissociate into free charge carriers under the energy level difference at D/A interfaces.^{[ 6 ]} This energy, required to dissociate exciton into free charge carriers, is called the exciton binding energy (E _b).^{[ 7,8 ]} E _b for organic materials is usually in the range of 0.1–1.0 eV,^{[ 9 ]} which is much higher than those for perovskite and inorganic semiconductors.^{[ 10 ]} The high E _b leads to large energy loss and severe geminate recombination for OSCs.^{[ 9 ]} Furthermore, because of the large E _b and short exciton lifetime, it is necessary to employ a BHJ structure with bicontinuous interpenetrating networks with nanoscale phase separation to facilitate exciton dissociation with the energy level difference, leading to morphology control issues. Thus, reducing E _b is one of the key issues to achieve highly efficient OSCs.

The past 3 years have witnessed the power conversion efficiency (PCE) of OSCs exceeding 18%^{[ 11,12 ]} due to outstanding advantages of non-fullerene acceptors (NFAs) as follows: 1) strong absorption in the visible and near-infrared (NIR) area, benefiting for high short-circuit current density (J _SC); 2) adjustable energy levels to achieve reduced energy loss and high open-circuit voltage (V _OC); and 3) various and simple chemical modifications. These advantages are believed to be closely related to its unique structure, which consists of a fused ring donor unit with side chains attached to a bridged atom (C or N), and two strongly electron-withdrawing units at both ends. With considerable efforts devoted to highly efficient NFAs, a large pool of NFAs have been developed, which offers a chance to investigate the E _b from the viewpoint of the structures of NFAs.

Several strategies have been put forward to successfully adjust E _b via modifying the structure of NFAs. For example, Li et al.^{[ 13 ]} reported that using [1,2,5]oxadiazolo[3,4-d]pyridazine to replace benzo[c][1,2,5]thiadiazole (BT) in NFAs results in the reduction of E _b deduced from density functional theory (DFT) calculation. However, the inner relationship between E _b and the structure of NFAs is complicated and has not been systematically investigated yet. In this perspective, we discuss the methods to determine the E _b and the relationship between E _b and open-circuit voltage loss (V _loss) of OSCs. Then, we focus on the relationship between E _b and the structure of NFAs from the aspects of fused-ring donor cores, end groups, and side chains, as well as molecular packing. At last, we figure out the potential directions to reduce E _b for high-efficiency OSCs. We hope that this perspective could provide a new view of molecule design for the breakthrough of OSCs.

The Calculation Methods of E _b

The E _b for a material M (homogeneous or heterogeneous) is defined as the energy required for the process. 1 $M^{*}=M^{+}+M^{-}$ where M ^* is the lowest excited state of the material in the solid material and M ⁺ and M ⁻ represent noninteracting holes and electrons in the material. Considering energy state change of a system absorbing light, E _b is calculated by the energy difference between the fundamental gap (E _fund) and optical gap (E _g ^opt).^{[ 14 ]} 2 $E_{\mathrm{b}}=E_{\text{fund}}-E_{\mathrm{g}}^{\text{opt}}$

As shown in Figure  1 , $E_{\mathrm{g}}^{\text{opt}}$ is the energy from the ground state (S₀) to the first optically allowed excited state (S₁), which is often determined experimentally from the absorption band edge. E _fund is the energy difference between ionization potential (IP) and electron affinity (EA). 3 $E_{\text{fund}}=\text{IP}-\text{EA}$

[IMAGE OMITTED. SEE PDF]

The $E_{\mathrm{g}}^{\text{opt}}$ is accessible by absorption measurement, whereas E _fund is relatively difficult to acquire accurately.^{[ 15 ]} One direct way to obtain E _fund is by measuring the IP and EA through ultraviolet photoelectron spectroscopy (UPS) and inverse photoelectron emission spectroscopy (IPES), respectively.^{[ 16 ]} It should be noted that using UPS and IPES must exclude nuclear relaxation processes and yield electron and hole transport levels that are shifted of 100–150 meV with respect to the fully relaxed polaron level relevant to carrier transport.^{[ 16 ]} However, few studies have used these methods to obtain E _fund in organic semiconducting materials due to the expensive equipment and complicated measurement.

Another method to obtain E _fund is by measuring the oxidation and reduction potential through cyclic voltammetry (CV).^{[ 17 ]} According to Hartree Fook's calculation and Koopmans’ theorem, IP can be considered as highest occupied molecular orbital (HOMO) energy level and EA can be considered as lowest unoccupied molecular orbital (LUMO) energy level in considering molecular orbits theory.^{[ 18 ]} Therefore, in many cases, E _b is normally expressed as 4 $E_{\mathrm{b}}=\text{LUMO}-\text{HOMO}-E_{\mathrm{g}}^{\text{opt}}=E_{\mathrm{g}}-E_{\mathrm{g}}^{\text{opt}}$ where E _g is the energy difference between HOMO and LUMO energy levels. However, for CV measurement, an equally and likely more important issue is that the effective (low-frequency) ε _r in the electrolyte solution is generally quite a bit larger than in an organic solid. This will lead to the deviation that the electrochemical gap (E _ox − E _red, E _ox is the oxidation potential and E _red is the reduction potential) is generally quite a bit smaller than the E _fund, and thus E _b will be underestimated. However, due to the universality of this method, there are still many studies to measure E _g by CV so far.

In addition, there is another classical way to calculate E _b with some basic parameters, which can be easily measured. 5 $E_{\mathrm{b}}\approx e^2/(4\pi {\epsilon }_0{\epsilon }_{\mathrm{r}}R)$ where e is the elementary charge, ε ₀ is the vacuum dielectric constant, and R is the average electron–hole distance.^{[ 19,20 ]} This method is limited as R is usually the distance between electrons and holes in a single molecule obtained by quantum-chemical calculations. ε _r is acquired by measuring the relationship between frequency (a common frequency for the comparison of ε _r values is 1 MHz this frequency is chosen because its time scale is comparable to that of nongeminate recombination (∼1 μs), which is the main loss mechanism to reduce with higher dielectric values.)^{[ 21 ]} and capacitance in the film. This method of calculating E _b is generally used in inorganic semiconductors. Due to the disorder of organic materials in thin films, it is not suitable for most organic semiconductors.

In addition to this experiment method, E _b can be calculated theoretically. For instance, Nayak et al. studied the shapes and radius of some small organic conjugated molecules with the B3LYP hybrid DFT method.^{[ 15 ]} Kraner et al. pointed out that E _b was the difference between the Coulomb interactions of the electron–hole pair and the kinetic energy of these quasi-particles and assumed that the kinetic energy is half of the Coulomb interactions.^{[ 22 ]} Furthermore, they designed a triad molecule and predicted that it had a very small E _b (about 39 meV) by utilizing the long-range, corrected hybrid density functional CAM-B3LYP.^{[ 23 ]} However, this research only focused on a single molecule. It is far from reality, as intermolecular interaction in film, related to molecular packing and arrangements, cannot be ignored.

The Relationship between E _b and V _loss in OSCs

V _loss in OSCs

Due to the large binding energy for Frenkel exciton, the V _loss of OSCs is considered to be a large and unavoidable loss.^{[ 24 ]} The V _loss in OSCs is defined as 6 $V_{\text{loss}}=\frac{E_g}{q}-V_{\text{OC}}$ where E _g is the lowest optical bandgap among the donor and acceptor components and q is the elementary charge.^{[ 24 ]} In the framework of Shockley–Queisser detailed balance model,^{[ 25,26 ]} the overall V _loss can be divided into three parts. 7 $V_{\text{loss}}=\Delta V_{\text{oc}}^{\text{SQ}}+\Delta V_{\text{oc}}^{\text{rad}}+\Delta V_{\text{oc}}^{\text{non}-\text{rad}}$ where $\Delta V_{\text{oc}}^{\text{SQ}}$ is the minimum V _loss while assuming an ideal semiconductor, $\Delta V_{\text{oc}}^{\text{rad}}$ comes from the absorption below the optical gap in a real device, and $\Delta V_{\text{oc}}^{\text{non}-\text{rad}}$ is the V _loss due to nonradiative recombination.^{[ 25 ]} The first term, $\Delta V_{\text{oc}}^{\text{SQ}}$, is the detailed balance V _loss limit under the given solar irradiation near the earth's surface and omnipresent blackbody radiation at room temperature. It is initially proposed to be 0.25–0.30 eV for an ideal semiconductor, solely depending on its optical gap under standard working conditions. For $\Delta V_{\text{oc}}^{\text{rad}},$sub-band gap states, such as the charge transfer (CT) states and the band-tail states, will provide additional transitions below the optical gap, and thereby extend the optical response to a lower energy region and reduce the maximum achievable voltage. The $\Delta V_{\text{oc}}^{\text{non}-\text{rad}}$ can be approximately quantified as follows. 8 $\Delta V_{\text{oc}}^{\text{non}-\text{rad}}=-\frac{kT}{q}\text{ln}({\text{EQE}}_{\text{EL}})$ where q is the elementary charge, k is the Boltzmann constant, T is the temperature, and EQE_EL is the external quantum efficiency of electroluminescence. Regardless of the methods used to categorize the V _loss, the $\Delta V_{\text{oc}}^{\text{non}-\text{rad}}$, stemming from nonradiative recombination, dominates the V _loss in most OSCs and has spurred tremendous efforts to mitigate this loss.

Binding Energy of CT States

For organic semiconductors, the low ε _r contributes to their high E _b, which results in a large Coulombic interaction between tightly bound excitons.^{[ 21 ]} These excitons are mainly separated by CT at organic donor/acceptor interfaces, driven by energetic offsets between donor and acceptor energy levels. A BHJ blend morphology is typically employed to facilitate photogenerated excitons to access this interface. Following exciton separation, charges may either dissociate into free charges or remain at donor/acceptor interfaces as electron–hole pairs. Such interfacial electron–hole pairs are generally referred to as CT states, which can be observed by the determination of the transient absorption. It needs extra energy to overcome the Coulomb force between the hole in the donor and the electron in the acceptor and separate the CT states in OSCs, which is the binding energy of CT states.^{[ 27 ]} The energy level offset, formed between the donor and the acceptor, would contribute to the separation of CT states. Therefore, the binding energy of CT states is generally less than E _b in OSCs.

Relationship between E _b, CT States, and V _loss

In BHJ OSCs, CT states are the active centers of charge separation and recombination. In fullerene-based OSCs, charge separation occurs at donor/acceptor interfaces and is considered to be driven by HOMO(D)–HOMO(A) or LUMO(D)–LUMO(A) energy level offsets (ΔE _HOMO or ΔE _LUMO). This energy consumption requirement is generally regarded as a basic limitation of OSC efficiency. A large energy level offset is usually related to suppressing the CT states recombination loss, but it also reduces the electronic bandgap of the blend (IP_donor−EA_acceptor), limiting the V _OC of devices. Conversely, a small energy level offset can increase V _OC, but it will limit the charge separation efficiency. Therefore, the trade-off between V _OC and J _SC is often observed in OSCs.^{[ 24 ]} Recently, it has been reported that CT states can hybridize with singlet exciton states under low energy level offset in high-efficiency NFA-based systems.^{[ 28,29 ]} The hybridization will make the CT states electronically couple with the singlet states, which has high oscillator strength for radiative transition, and thus improves the radiative efficiency of the blend to reduce the nonradiative recombination loss. Although the energy of CT states is closely related to E _b and V _loss, there is no necessary connection between E _b and V _loss. Though it has been reported that, in some NFA systems, lower E _b is accompanied by a reduction in energy loss, the inner relationship is unclear yet.

Effect of the Structure of NFAs on E _b of Excitons

E _b of excitons is closely related to the molecular structure of materials in the active layer. We discussed the relationship between E _b and the structure of NFAs benefiting from the large pool of the reported NFAs in this section. To obtain more reliable trends in E _b, the molecule and its E _b for comparison need to be acquired from the same article. Moreover, the rest of the material remains unchanged to study the effect of a certain structural unit on E _b. We extracted the required data from a number of studies and use Equation (3) to calculate E _b to investigate the relationship between E _b and the structure of NFAs. The data of E _b for relevant molecules are listed in Table  1 .

Table 1 E _g , E _g ^opt, and E _b for NFAs (E _b was calculated by Equation (3))

Acceptor	E g	E g opt	E b	Ref.
IDIC	1.79a)	1.62b)	0.17	[32]
IEIC	1.60a)	1.57b)	0.03	[33]
ITIC	1.75a)	1.59b)	0.16	[34]
INIC	1.57a)	1.57b)	/	[35]
NFBDT	1.57a)	1.56b)	0.01	[36]
F6IC	1.64a)	1.36b)	0.28	[44]
F8IC1	1.44a)	1.32b)	0.12	[43]
F10IC1	1.39a)	1.29b)	0.10	[43]
Y6	1.55a)	1.33b)	0.22	[40]
BP5T-4 F	1.75a)	1.34b)	0.39	[45]
ABP4T-4 F	1.80a)	1.37b)	0.43	[45]
BP6T-4 F	1.76a)	1.39b)	0.37	[46]
ABP6T-4 F	1.73a)	1.36b)	0.37	[46]
IT-DM	1.76a)	1.63b)	0.13	[49]
IT-CF	1.64a)	1.57b)	0.07	[49]
IT-4 F	1.57a)	1.54b)	0.03	[49]
IT-IC-0 F	1.70a)	1.59b)	0.11	[50]
IT-IC-2 F	1.69a)	1.56b)	0.13	[50]
IT-IC-3 F	1.61a)	1.54b)	0.07	[50]
IT-IC-4 F	1.59a)	1.55b)	0.04	[50]
IT-IC-6 F	1.62a)	1.51b)	0.11	[50]
YA1	1.92c)	1.73c)	0.19	[51]
YA2	1.95c)	1.75c)	0.20	[51]
YA3	2.00c)	1.80c)	0.21	[51]
YA4	2.03c)	1.82c)	0.21	[51]
Y21	2.03c)	1.83c)	0.20	[51]
BTP-4Cl	1.49a)	1.35b)	0.14	[52]
C8-ITIC	1.72a)	1.55b)	0.17	[57]
ITIC	1.72a)	1.61b)	0.11	[57]
Y6-O	2.04a)	1.52b)	0.52	[59]
Y6	1.86a)	1.37b)	0.49	[59]
BTP-Ph	1.79a)	1.36b)	0.43	[58]
BTP-Th	1.82a)	1.34b)	0.48	[58]
BTP-C11	1.78a)	1.32b)	0.46	[58]
Y6-4O	1.63d)	1.39b)	0.24	[67]
Y6	1.65d)	1.35b)	0.30	[67]

Effect of Conjugated Backbones on E _b

Because of the great success of 4,9-dihydro-s-indaceno[1,2-b:5,6-b′]dithiophene (IDT) unit as an electron-donating core in DC-IDT2T and fluorene as a core in NFAs,^{[ 30 ]} the most common strategy to tune the optical physical properties of NFAs is increasing the number of conjugated rings in the fused core. Extending the donor core promises to enhance intramolecular charge transfer (ICT), thus enhancing HOMO energy level and expanding the absorption to NIR region.^{[ 31 ]} According to the calculation of E _b, extending conjugated rings has an impact on E _g and E _g ^opt, resulting in variation of E _b. Therefore, it is necessary to explore the influence of the conjugated fused rings on E _b. Figure  2 shows the chemical structures of the different conjugated units of the NFAs discussed in this section. Many research groups have reported various models to calculate the E _b of molecules to find the relationship between the number of fused rings and E _b. Zhu et al.^{[ 27 ]} used the systemically optimized time-dependent DFT and took into account the solid-state polarization effect to calculate E _b of six kinds of molecules, including IDIC,^{[ 32 ]} IEIC,^{[ 33 ]} ITIC,^{[ 34 ]} INIC,^{[ 35 ]} and NFBDT.^{[ 36 ]} Figure  3a shows E _b for molecules with different conjugated lengths. The results demonstrate that E _b decreases with the increased number of fused rings on the backbones when the side chains and end groups remain unchanged. For example, E _b of INIC, which has two more thiophene rings on the backbones, is smaller than that of NFBDT. The results from the calculation indicate that extending conjugation of the main chain leads to a decreased isotropic component of the molecule polarizability (α).^{[ 37 ]}As the ε _r increases when α decreases,^{[ 38 ]} extending fused rings in the donor core reduces E _b. These results are also verified in other systems. Zheng et al.^{[ 39 ]} simulated the calculation of Y6-series molecules and compared E _b of BTPTT-4F(Y6),^{[ 40 ]} BTP6T-4F, BTPTT-4F-2T, and BTPTT-4F-TT-TT6 on the basis of DFT with the tuned ωB97X/6-31G(d)/PCM theory level.^{[ 41,42 ]} The calculation results of E _b for these molecules are shown in Figure 3b. It showed that E _b decreased effectively by either increasing the number of fused rings of the main chain or increasing the π bridge linkage element.

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

We calculated E _b of the reported molecules according to Equation (3) and found the same pattern with the DFT calculation results mentioned earlier. Zhan et al.^{[ 43 ]} reported a series of fused-ring small molecule acceptors, namely F6IC,^{[ 44 ]} F8IC1, and F10IC1. According to the data reported in the literature, it was found that E _b gradually decreased with the increased number of fused rings under the condition of unchanged end groups and side chains. Reduced E _b facilitates exciton dissociation.^{[ 10 ]} As a result, exciton dissociation efficiency is increased with more fused rings in the conjugated structure. The J _SC /J _sat (saturated photocurrent) ratios for the F6IC-, F8IC1-, and F10IC1-based devices are calculated to be 93.7%, 95.6%, and 96.9%, respectively. As for asymmetric NFAs, BP5T-4F possesses one more fused ring than ABP4T-4 F, leading to smaller E _b.^{[ 45 ]} After that, another two NFAs, ABP6T-4F and BP6T-4F, based on nonacyclic electron acceptors were developed by the same group.^{[ 46 ]} Due to the extension of conjugated units, E _b of these two molecules are further decreased.

The asymmetric structure generally has a higher ε _r compared to the symmetric isomer with the same number of fused rings in the donor unit. According to Equation (4), the larger ε _r is, the smaller E _b will be obtained. As a result, better exciton dissociation can be expected. For instance, Jen et al.^{[ 46 ]} designed an asymmetric isomer of BT and developed the fused-ring electron acceptors, ABP6T-4F. The ε _r of ABP6T-4F is 2.22 times higher than that of BP6T-4F (1.78). As a result, ABP6T-4F presents a smaller E _b compared to BP6T-4F. It may be one of the reasons for more effective exciton dissociation in ABP6T-4F-based devices.

As summarized in Figure 3c, increasing the number of fused rings in the donor core, replacing different fused ring blocks, and introducing an asymmetric structure can decrease E _b effectively.

Effect of the End Groups on E _b

The end groups of NFAs not only influence the molecular energy levels (especially LUMO energy levels) and optical gaps through ICT,^{[ 47 ]} but also influence the dipole moment (μ) of NFAs. E _b was affected by the end groups of NFAs, because E _b is related to the energy levels and μ.^{[ 48 ]} It is impossible to distinguish the effect of energy levels and μ. The μ is the product of the distance between the center of a molecule's positive and negative charges and the amount of charge q in the center. 9 $\overrightarrow{\mu }=q\times \overrightarrow{r}$

Therefore, the dipole of the substituent bond to the carbon and the position of the substituent on the end groups determine the μ of the molecule. We summarized the tendency of E _b toward the different kinds of fluorine substitution on the end groups from the reported NFA series. For IDTT series, we take IT-DM, IT-CF, and IT-4F, for instance (Figure  4 ). E _b gradually decreased from the electron-donating group to the electron-withdrawing group on end groups.^{[ 49 ]} The OSC based on IT-4F with the smallest E _b exhibits the highest J _SC. For IT-IC-nF series (Figure 4),^{[ 50 ]} another class of IDTT-based fluorinated molecules, increasing the fluorine atoms of the IT-IC-nF end group could reduce E _g ^opt and deepen LUMO energy level, as shown in Figure  5a. It turns out that from IT-IC-2F to IT-IC-4F, E _b decreases. However, it should be noted that there is some exceptional case. E _b increases for IT-IC-6F, and E _b of the IT-IC-2F is larger than that of IT-IC-0F. Although E _b of IT-IC-6F is not the lowest, as shown in Figure 5b, IT-IC-6F-based devices afford the highest J _SC. In fact, the μ is a vector quantity. A large μ on the whole molecule can be achieved, only when the dipoles of the substituents on end groups are in the same direction. Therefore, it is necessary to study the effect of fluoride on dipole distance and E _b. Furthermore, IT-IC-4F affords the lowest E _b but not the highest J _SC, suggesting a balance between E _b and other photovoltaic parameters (e.g., V _OC).

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

These results are similar for the Y series NFAs. Muhammad et al.^{[ 51 ]} designed four novel Y series NFAs (YA1, YA2, YA3, and YA4 shown in Figure 4) through terminal modification of the recently reported Y21 molecule, and calculated E _b of them through the charge distribution on the molecules. The result from the calculation indicates that E _b decreases with the large dipole of the substituent on the end groups in Y series. For example, BTP-4Cl (Figure 4), with Cl at the end group,^{[ 52 ]} exhibits smaller E _b compared to the fluorinated counterpart Y6. As is well known, fluorine has a stronger electron-withdrawing ability than chlorine. The reason for the abnormal reduction of E _b may be that the μ of the chloro-carbon bond is larger than that of the fluorocarbon bond. The dipole direction in BTP-4Cl turns opposite compared to that in Y6. The reduced E _b suppresses recombination loss, which contributes to improved V _OC. In contrast, a large μ facilitates the charge separation and helps to achieve a high fill factor (FF) in the device.^{[ 53 ]}

In fact, as the μ is a vector, the dipole of the substituent must be as large as possible in the same direction to enhance the μ of the entire molecule. As shown in Figure 5c, the large μ design of the molecule will increase the molecular polarizability and reduce E _b.^{[ 54 ]}

Effect of the Side Chains on E _b

In NFAs, side chains are usually attached to the sp³ hybrid atoms C and Si or sp ² hybrid N atoms in the donor moiety to increase the solubility in the processing solvents, and to prevent excessive aggregation of NFAs for ideal morphology.^{[ 31 ]} The large steric hindrance and free rotation for the side chains will disturb the intermolecular π–π stacking and lead to a low crystallinity.^{[ 55 ]} Intermolecular aggregation often has a great influence on E _b. for example, Wei et al.^{[ 56 ]}reported that the 3D packing crystal of the OSCs can effectively reduce E _b with an extremely weak E _b of 0.04 eV calculated in 4TIC. We discussed the side chains in this part because the side chains can change the ε _r of the molecule and regulatory molecular aggregation and thus affect E _b. The chemical structures of the different side chains of the NFAs are shown in Figure  6 .

[IMAGE OMITTED. SEE PDF]

Heeney et al.^{[ 57 ]} synthesized C8-ITIC by replacing phenylalkyl side chains with alkyl side chains of ITIC. With the same E _g, C8-ITIC has a smaller E _g ^opt leading to a larger E _b. Similar results were found in the Y series NFAs. Yan et al.^{[ 58 ]} developed BTP-Th and BTP-Ph by introducing thiophene and benzene ring units at the β site of the thiophene group of Y6 (BTP-C11), respectively. It is found that the thiophene and benzene units on the side chains introduce more steric hindrance and thus slightly reduce the crystallinity of the molecule. Compared to Y6, phenyl side chain (BTP-Ph) decreases E _b to 0.43 eV, while thiophene side chain (BTP-Th) increases E _b. As a result, BTP-Ph-based devices offer a superior photovoltaic performance than BTP-Th and Y6, which should be partially ascribed to the reduced E _b of BTP-Ph. By introducing aromatic hydrocarbons on the alkyl chain, the molecular stacking method can be adjusted to affect E _b.^{[ 56 ]}

In contrast, alkoxy chains influence the E _b. Y6-O, an analogue of Y6, is obtained by substituting the alkyl chains with alkoxy chains at the outer thiophene β site.^{[ 59 ]} Alkoxy replacement obviously improves the LUMO energy levels and keeps HOMO energy levels unchanged. Compared to Y6, Y6-O suffers higher E _b. However, with the number of oxygen atoms in the side chain increasing, ε _r increases and E _b decreases. Lin et al.^{[ 60 ]} reported a new NFA named Y6-4O, which introduces an asymmetric highly polarizable oligo-(ethylene glycol) side chain onto the pyrrole unit of Y6. As shown in Figure  7a, at the frequency of 1 kHz, Y6-4O exhibits a ε _r of 5.13, obviously higher than that of Y6 (3.36). As a result, E _b decreases with the oligo-(ethylene glycol) side chain. The PCE of Y6-4O (15.2%) is the highest among NFAs with a ε _r greater than 5 (Figure 7b).

[IMAGE OMITTED. SEE PDF]

As summarized in Figure 7c, by introducing aryl groups on the side chains, the molecular aggregation can be adjusted to affect E _b. With the increase of the number of oxygen atoms in the side chain, ε _r of the molecule will increase, leading to the reduction of E _b.

Effect of the Molecular Packing on E _b

The optical and electrical properties of organic materials are deeply affected by molecular packing structures in the solid state. In general, organic materials will form two kinds of aggregates in the solid state: H- and J-aggregates. These aggregations cause molecules to form dimers and thus split energy levels. According to Kasha's molecular exciton model,^{[ 60 ]} when NFAs form H-aggregates, the singlet energy of the dimer state is higher than that of the single molecule. As for J-aggregates, when the slip angle is less than 54.7°, the singlet energy of the dimer state is lower than that of the single molecule.^{[ 60 ]} For most of the NFAs, it is usually easy to form J-aggregates in the film. Therefore, the film absorption of most NFAs redshifts, compared with solution absorption. Thus, the molecular packing will affect the E _b of organic materials.^{[ 61 ]} For example, Zheng et al.^{[ 62 ]} calculated the E _b values of the single molecule and dimer of subphthalocyanine chloride (SubPc) (Figure  8 ) and found that dimer s³ packing state of SubPc had the smallest E _b. Wei et al.^{[ 56 ]} reported that the 3D packing crystal of the OSCs could effectively reduce E _b with an extremely weak E _b of 0.04 eV calculated in 4TIC (Figure 8). Moreover, they further calculated the E _b values of IDIC in three different crystal phases and demonstrated that a higher backbone packing dimensionality was beneficial for obtaining a smaller E _b.^{[ 63 ]}

[IMAGE OMITTED. SEE PDF]

Summary and Outlook

Since 2015, OSCs have been greatly developed with the maximum PCE exceeding 18% and gradually approaching that of inorganic solar cells. Although great achievements in OSCs have been made recently, large E _b, the main factor limiting the photovoltaic performance of OSCs, calls for much attention. In this perspective, the relationship between molecular structure and E _b of NFAs is discussed as follows: 1) The donor moiety. Extending the donor core of NFAs is an effective way to enhance absorption and decrease E _b. In addition, the asymmetric structure can decrease E _b by improving ε _r. 2) The electron-withdrawing end groups. Increasing the electron-withdrawing capacity of end groups or halogen atomic substitution may enhance the μ of the molecule. The larger μ is, the smaller E _b will be achieved. 3) The side chains. Side chains affect the ε _r of NFAs, thus changing E _b. At present, an ethylene glycol chain instead of an alkyl chain can effectively decrease E _b by increasing ε _r. Introducing highly polarizable side chains seems an effective way to reduce E _b by enhancing ε _r of NFAs.

It should be noted that the methods to reduce E _b are more than the molecular design of NFAs. For instance, from the viewpoint of the donor, Janssen et al.^{[ 64 ]} reported that introducing 2,5-dihydropyrrolo-[3,4-c]pyrrole-1,4-dione (DPP) and dithieno[3,2-b:2′,3′-d]pyrrole (DTP) (Figure 8) into the polymer donor can effectively reduce E _b. In terms of interfacial modification, inorganic high-k nanostructured materials are introduced to decrease E _b by reducing Coulomb interaction.^{[ 65,66 ]}

E _b has a significant influence on the performance of OSCs. How to reduce E _b is one of the key issues for the next breakthrough. Therefore, more attention is suggested to be paid to the potential directions: 1) How to balance E _b with bandgaps of NFAs? As discussed in Section 3.2, fluorination of the end group reduces E _b, but the LUMO energy of NFAs declines at the same time. As a result, reduced E _b increases J _SC but decreases V _OC. Thus, the balance between E _b and bandgaps needs to be taken into consideration. 2) How to balance E _b with miscibility? Increasing ε _r by introducing highly polarizable side chains is an effective way to decrease E _b. However, the highly polarizable side chains may decrease the miscibility of NFAs with the donor. It may lead to inferior photovoltaic performance due to morphology issue. Therefore, how to balance E _b with miscibility needs attention. For example, the number, the length, and the position of the highly polarizable side chains are suggested to be carefully considered. 3) What is the relationship between E _b and the structure of donor molecules? The structure of donor molecules also affects their E _b. However, it is hard to summarize the regulation between E _b and donors, as the related literature are limited. Therefore, we appealed to investigate the effect of donor molecules on E _b as well.

Acknowledgements

Y.Z. and F.Z. contributed equally to this work. The authors thank the National Natural Science Foundation of China (No. 52173189) for financial support.

Conflict of Interest

The authors declare no conflict of interest.

References

L. Dou , J. You , Z. Hong , Z. Xu , G. Li , R. A. Street , Y. Yang , Adv. Mater. 2013, 25, 6642.

S. Guenes , H. Neugebauer , N. S. Sariciftci , Chem. Rev. 2007, 107, 1324.

A. L. Roes , E. A. Alsema , K. Blok , M. K. Patel , Prog. Photovolt: Res. Appl. 2009, 17, 372.

A. Heeger , Chem. Soc. Rev. 2010, 39, 2354.

C. M. Proctor , M. Kuik , T.-Q. Nguyen , Prog. Polym. Sci. 2013, 38, 1941.

R. Meng , Y. Li , C. Li , K. Gao , S. Yin , L. Wang , Phys. Chem. Chem. Phys. 2017, 19, 24971.

R. S. Bhatta , M. Tsige , Polymer 2014, 55, 2667.

K. Hummer , P. Puschning , S. Sagmeister , C. Ambrosch-draxl , Mod. Phys. Lett. B 2008, 20, 261.

M. Knupfer , Appl. Phys. A 2003, 77, 623.

F. Arabpour Roghabadi , N. Ahmadi , V. Ahmadi , A. Di Carlo , K. Oniy Aghmiuni , A. Shokrolahzadeh Tehrani , F. S. Ghoreishi , M. Payandeh , N. Mansour Rezaei Fumani , Sol. Energy 2018, 173, 407.

Q. Liu , Y. Jiang , K. Jin , J. Qin , J. Xu , W. Li , J. Xiong , J. Liu , Z. Xiao , K. Sun , S. Yang , X. Zhang , L. Ding , Sci. Bull. 2020, 65, 272.

Y. Qin , Y. Chang , X. Zhu , X. Gu , L. Guo , Y. Zhang , Q. Wang , J. Zhang , X. Zhang , X. Liu , K. Lu , E. Zhou , Z. Wei , X. Sun , Nano Today 2021, 41, 101289.

Y. L. Wang , Q. S. Li , Z. S. Li , Phys. Chem. Chem. Phys. 2018, 20, 14200.

H. Sun , Z. Hu , C. Zhong , S. Zhang , Z. Sun , J. Phys. Chem. C 2016, 120, 8048.

P. K. Nayak , N. Periasamy , Org. Electron. 2009, 10, 1396.

A. Kahn , Mater. Horiz. 2016, 3, 7.

N. Mercier , M. Giffard , G. Pilet , M. Allain , P. Hudhomme , G. Mabon , E. Levillain , A. Gorgues , A. Riou , Chem. Commun. 2001, 2722.

T. Tsuneda , J. W. Song , S. Suzuki , K. Hirao , J. Chem. Phys. 2010, 133, 174101.

P. K. Nayak , Synth. Met. 2013, 174, 42.

J.-L. Bredas , Mater. Horiz. 2014, 1, 17.

M. P. Hughes , K. D. Rosenthal , N. A. Ran , M. Seifrid , G. C. Bazan , T. Q. Nguyen , Adv. Funct. Mater. 2018, 28, 1801542.

Y. Zhang , A. Mascarenhas , Phys. Rev. B 1999, 59, 2040.

S. Kraner , G. Prampolini , G. Cuniberti , J. Phys. Chem. C 2017, 121, 17088.

J. Wu , H. Cha , T. Du , Y. Dong , W. Xu , C. T. Lin , J. R. Durrant , Adv. Mater. 2022, 34, 2101833.

W. Shockley , H. J. Queisser , J. Appl. Phys. 1961, 32, 510.

J. Yao , T. Kirchartz , M. S. Vezie , M. A. Faist , W. Gong , Z. He , H. Wu , J. Troughton , T. Watson , D. Bryant , J. Nelson , Phys. Rev. Appl. 2015, 4, 014020

L. Zhu , Y. Yi , Z. Wei , J. Phys. Chem. C 2018, 122, 22309.

F. D. Eisner , M. Azzouzi , Z. Fei , X. Hou , T. D. Anthopoulos , T. J. S. Dennis , M. Heeney , J. Nelson , J. Am. Chem. Soc. 2019, 141, 6362.

G. Han , Y. Yi , J. Phys. Chem. Lett. 2019, 10, 2911.

H. Bai , Y. Wang , P. Cheng , J. Wang , Y. Wu , J. Hou , X. Zhan , J. Mater. Chem. A 2015, 3, 1910.

S. Dey , Small 2019, 15, 1900134.

L. Yang , S. Zhang , C. He , J. Zhang , H. Yao , Y. Yang , Y. Zhang , W. Zhao , J. Hou , J. Am. Chem. Soc. 2017, 139, 1958.

Y. Lin , Z.-G. Zhang , H. Bai , J. Wang , Y. Yao , Y. Li , D. Zhu , X. Zhan , Energy Environ. Sci. 2015, 8, 610.

Y. Lin , J. Wang , Z. G. Zhang , H. Bai , Y. Li , D. Zhu , X. Zhan , Adv. Mater. 2015, 27, 1170.

S. Dai , F. Zhao , Q. Zhang , T. K. Lau , T. Li , K. Liu , Q. Ling , C. Wang , X. Lu , W. You , X. Zhan , J. Am. Chem. Soc. 2017, 139, 1336.

B. Kan , H. Feng , X. Wan , F. Liu , X. Ke , Y. Wang , Y. Wang , H. Zhang , C. Li , J. Hou , Y. Chen , J. Am. Chem. Soc. 2017, 139, 4929.

W. Volksen , R. D. Miller , G. Dubois , Chem. Rev. 2010, 110, 56.

J. P. Perdew , S. Kurth , Lect. Notes Phys. 2003, 500, 8.

W. Qiu , S. Zheng , Sol. RRL 2021, 5, 2100023.

J. Yuan , Y. Zhang , L. Zhou , G. Zhang , H.-L. Yip , T.-K. Lau , X. Lu , C. Zhu , H. Peng , P. A. Johnson , M. Leclerc , Y. Cao , J. Ulanski , Y. Li , Y. Zou , Joule 2019, 3, 1140.

R. O. Jones , O. Gunnarsson , Chem. Mater. 1989, 61, 689.

E. K. U. Gross , C. A. Ullrich , U. J. Gossmann , NATO ASI Ser., Ser. B 1995, 337, 149.

W. Wang , H. Lu , Z. Chen , B. Jia , K. Li , W. Ma , X. Zhan , J. Mater. Chem. A 2020, 8, 3011.

S. Dai , T. Li , W. Wang , Y. Xiao , T. K. Lau , Z. Li , K. Liu , X. Lu , X. Zhan , Adv. Mater. 2018, 30, 1706571.

W. Gao , H. Fu , Y. Li , F. Lin , R. Sun , Z. Wu , X. Wu , C. Zhong , J. Min , J. Luo , H. Y. Woo , Z. Zhu , A. K. Y. Jen , Adv. Energy Mater. 2020, 11, 2003177.

W. Gao , B. Fan , F. Qi , F. Lin , R. Sun , X. Xia , J. Gao , C. Zhong , X. Lu , J. Min , F. Zhang , Z. Zhu , J. Luo , A. K. Y. Jen , Adv. Funct. Mater. 2021, 31, 2104369.

J. Zhang , Y. Li , Z. Peng , F. Bai , L.-K. Ma , H. Ade , Z. Li , H. Yan , Mater. Chem. Front. 2020, 4, 1729.

K.-G. Lim , S. Ahn , T.-W. Lee , J. Mater. Chem. C 2018, 6, 2915.

M. Hao , T. Liu , Y. Xiao , L.-K. Ma , G. Zhang , C. Zhong , Z. Chen , Z. Luo , X. Lu , H. Yan , L. Wang , C. Yang , Chem. Mater. 2019, 31, 1752.

T. J. Aldrich , M. Matta , W. Zhu , S. M. Swick , C. L. Stern , G. C. Schatz , A. Facchetti , F. S. Melkonyan , T. J. Marks , J. Am. Chem. Soc. 2019, 141, 3274.

M. U. Khan , R. Hussain , M. Yasir Mehboob , M. Khalid , Z. Shafiq , M. Aslam , A. A. Al-Saadi , S. Jamil , M. Janjua , Mater. Horiz. 2020, 5, 24125.

Y. Cui , H. Yao , J. Zhang , T. Zhang , Y. Wang , L. Hong , K. Xian , B. Xu , S. Zhang , J. Peng , Z. Wei , F. Gao , J. Hou , Nat. Commun. 2019, 10, 2515.

B. Carsten , J. M. Szarko , H. J. Son , W. Wang , L. Lu , F. He , B. S. Rolczynski , S. J. Lou , L. X. Chen , L. Yu , J. Am. Chem. Soc. 2011, 133, 20468.

W. Gao , M. Zhang , T. Liu , R. Ming , Q. An , K. Wu , D. Xie , Z. Luo , C. Zhong , F. Liu , F. Zhang , H. Yan , C. Yang , Adv. Mater. 2018, 30, 1800052.

J. Ouyang , B. Zhang , G. Zeng , J. Zhang , X. Zhao , X. Yang , Sol. RRL 2020, 4, 2000359.

L. Zhu , Z. Tu , Y. Yi , Z. Wei , J. Phys. Chem. Lett. 2019, 10, 4888.

Z. Fei , F. D. Eisner , X. Jiao , M. Azzouzi , J. A. Rohr , Y. Han , M. Shahid , A. S. R. Chesman , C. D. Easton , C. R. McNeill , T. D. Anthopoulos , J. Nelson , M. Heeney , Adv. Mater. 2018, 30, 1705209.

Y. Chang , J. Zhang , Y. Chen , G. Chai , X. Xu , L. Yu , R. Ma , H. Yu , T. Liu , P. Liu , Q. Peng , H. Yan , Adv. Energy Mater. 2021, 11, 2100079.

Y. Chen , T. Liu , L.-K. Ma , W. Xue , R. Ma , J. Zhang , C. Ma , H. K. Kim , H. Yu , F. Bai , K. S. Wong , W. Ma , H. Yan , Y. Zou , J. Mater. Chem. A 2021, 9, 7481.

H. R. R. M. Kasha , M. A. El-Bayoumi , Pure Appl. Chem. 1965, 11, 371.

J. H. Kim , T. Schembri , D. Bialas , M. Stolte , F. Wurthner , Adv. Mater. 2021, 2104678.

X. Chen , S. Zheng , Int. J. Quantum Chem. 2020, 121, e26511.

L. Zhu , C. Yang , Y. Yi , Z. Wei , J. Phys. Chem. Lett. 2020, 11, 10227.

P. J. Leenaers , A. Maufort , M. M. Wienk , R. A. J. Janssen , J. Phys. Chem. C 2020, 124, 27403.

L. Zhu , J. Zhang , Y. Guo , C. Yang , Y. Yi , Z. Wei , Angew. Chem., Int. Ed. 2021, 60, 15348.

M. Engel , F. Kunze , D. C. Lupascu , N. Benson , R. Schmechel , Phys. Status Solidi RRL 2012, 6, 68.

T. Li , K. Wang , G. Cai , Y. Li , H. Liu , Y. Jia , Z. Zhang , X. Lu , Y. Yang , Y. Lin , JACS Au 2021, 1, 1733.