Molecular insights of exceptionally photostable electron acceptors for organic photovoltaics

Photo-degradation of organic semiconductors remains as an obstacle preventing their durable practice in optoelectronics. Herein, we disclose that volume-conserving photoisomerization of a unique series of acceptor-donor-acceptor (A-D-A) non-fullerene acceptors (NFAs) acts as a surrogate towards their subsequent photochemical reaction. Among A-D-A NFAs with fused, semi-fused and non-fused backbones, fully non-fused PTIC, representing one of rare existing samples, exhibits not only excellent photochemical tolerance in aerobic condition, but also efficient performance in solar cells. Along with a series of in-depth investigations, we identify that the structural confinement to inhibit photoisomerization of these unique A-D-A NFAs from molecular level to macroscopic condensed solid helps enhancing the photochemical stabilities of molecules, as well as the corresponding OSCs. Although other reasons associating with the photostabilities of molecules and devices should not excluded, we believe this work provides helpful structure-property information toward new design of stable and efficient photovoltaic molecules and solar cells.

The corresponding results and discussion were added in the manuscript and supporting information (Page 8), as shown in Fig S13 and  2. As the photoisomerization is revealed, it is better to test (or demonstrate) the point that the E-TFICH is less stable (or more reactive) than Z-TFICH. Response: We thank reviewer very much for his/her positive assessments.  (Chem. Rev., 2015, 115, 11718), and the bridged stilbenes (Angew. Chem. Int. Ed., 2020, 59, 10566).
(2 The corresponding results and discussion were added in the manuscript and supporting The authors have addressed the issues in the previous round of review, and the paper can be accepted now. Reviewer #2 (Remarks to the Author): The manuscript reports the photoisomerization of exocyclic vinyl groups as a critical step towards the subsequent photodegradation of NFAs. The work reported here is interesting. The authors have revised the manuscript responding to the questions raised previously. However, the main questions are still not properly addressed and some of them are listed below. It still requires further explanation and clarification. I do not recommend publishing this manuscript as it is in this journal. The authors should revise the manuscript based on the issues pointed out below before considering resubmission even to another journal. 2. Photoisomerization: The authors stated that the distorted C=C bonds after photoisomerization appears to be more reactive towards photooxidation that is likely a critical step opening up the photo degradation of NFAs. However, they have not explained why this is the case for their materials. Why does photoisomerization increase the propensity for photooxidation? I feel this is a key observation in the paper but it hasn't been explained.
3. Molecular structure: The authors stated that the planar non-fused skeleton and outward hexyl chains lead to exceptional photostability of A-D-A NFAs. I am not convinced by the fused vs unfused ring explanation for better photostability, except that the unfused structure has less "pointy out" side groups, allowing better packing. The authors cited that the fused ring acceptors Y6, fits well with their explanation. However, the stability of Y6 is reported to be due to the restricted rotation induced by the side chains.
4. Raman spectra: The authors showed Raman spectra of different films and indicated the vinyl group (assigned to 1561 cm-1 for IT-4F and 1623 cm-1 for HF-PCIC) along with other characteristic signals, disappeared during photodegradation. However, the IT-4F Raman spectra are not clear at all, as there seems to be no clear changes in relative peak intensities, instead all peaks decrease their intensities.
5. BHJ morphology: The authors provided more optical data to show photo-stabilities of NFAs. However, it still does not provide the information about other critical factors such as BHJ morphology which is known to strongly influence device stability. The detailed studies of structural and stacking properties of NFAs in BHJ blends are needed to exclude them as main contributor for the device stability.

NFAs is linked to the structural factors of molecules, and come up solutions: 1) to suppress the photoisomerization of vinyl groups by installing outward-chain in A-D-A NFAs (Scheme 1d in the main text); 2) to promote dense packing of molecules with planar sp 3 carbon-free backbones (Scheme 1e in the main text).
As for reviewer question, why does long-lived excited state in film not lead to more degradation as found in solution? PL lifetime has correlation, but not the sole factor to influence the photochemical reaction of molecules. Non-fused PTIC appears with both good photochemical stability and relatively long PL lifetime in solid. They are not contradictory factors. It is because non-fused acceptors, as discussed in main text, are mediated by non-covalent interaction that adapt a rotatable conformation in solution and transit into planar conformation with dense stacking in solid. Therefore, there appears difference for the PL lifetime of non-fused acceptors in solution and in solid. (

1) In solution, molecules are dissolved, and surrounded by solvent. For semi-fused HF-PCIC and non-fused PTIC, there have rotatable single-bond in backbone, which are incline to quench the excited state via the non-radiative decay channel, for instance molecular vibration and motion. Whereas, fused molecule itself with rigid and chemical-bond locked backbone appears with long-excited life time in solution. Meanwhile, IT-4F has lack of protection to vinyl groups, which gives high probability towards photoisomerization and photooxidation, as revealed by NMR and TRPL measurements. (2) In condensed solid, non-fused PTICs were tightly stacked with planar conformation in solid (as revealed by crystal, XRD and TEM), wherein the vinyl groups are well protected by outward-chain and packed molecules. The molecular motion is also constrained to mitigate non-radiative decay channel. Therefore, PL lifetimes of neat-films for PTIC is largely elongated to that in solutions. These combined factors embed PTIC with the elongated and photooxidation-resistive excited state that is beneficial for the stable and efficient photo-to-electron conversion of solar cells.
Overall, non-fused acceptors, representing one of rare existing samples, exhibit appearing characteristics of both good photochemical tolerance and long exited lifetime in solid. In addition, they can be prepared through the extremely simple chemistry. These pleasant properties are not contradictory factors, just needs more understandings and explorations. We believe the insights reported herein are beneficial for community to further develop stable and efficient organic photovoltaic materials.

Regarding to reviewer comments, description has been added into main text:
It is consistent with time-resolved photoluminescence (TRPL) spectroscopy (Fig. 2b)  appears to be more reactive towards photooxidation that is likely a critical step opening up the photo degradation of NFAs. However, they have not explained why this is the case for their materials. Why does photoisomerization increase the propensity for photooxidation? I feel this is a key observation in the paper but it hasn't been explained. The outcomes of this work are far more than the above description.
Response: Thanks for bringing this again. This comment is also asked by Reviewer 1 that was previously well addressed. From UV-vis and NMR measurements, it clearly shows that photo-isomerism of TFICH from ZS to ES occurs prior to the photooxidation (after 220 hours illumination). This experimental evidence serves as a solid foundation to support that Z-form is more stable than E-from. Besides, among all the reported single crystal data, A-D-A non-fullerene acceptors all adapt Z-form of the terminal double bonds, suggesting Z-form is thermal dynamically more stable conformation than that of E-from. The isomerization of exocyclic vinyl groups occurs as volume-conserving photoisomerization in solid (Supplementary Scheme), which is similar to its presence in the green fluorescent protein (GFP) chromophore (J. Am. Chem. Soc. 2019, 141, 15504).

Fig. 5b:
The photo-isomerism of TFICH in solution state after 10 hours illumination (from ZS to the co-mixture ZS + ES) that are reversible process upon heat treatment (Recovered ZS), and further decomposition of TFICH after 220 hours illumination (After Oxidation).
4. Raman spectra: The authors showed Raman spectra of different films and indicated the vinyl group (assigned to 1561 cm -1 for IT-4F and 1623 cm -1 for HF-PCIC) along with other characteristic signals, disappeared during photodegradation. However, the IT-4F Raman spectra are not clear at all, as there seems to be no clear changes in relative peak intensities, instead all peaks decrease their intensities.
Response: Thanks for the comments. The Raman signals are sensitive to film properties. IT-4F spectra appears with good signal-to-noise ratio, as can be assigned to 1561 cm -1 for vinyl bond at initial stage, which is disappeared due to fast degradation under illumination in ambient. We have added a supplementary figure including Infrared spectrum and Raman spectrum. These complementary signals are clear enough to assign molecular structures. Description has been added into main text: The degradation of IT-4F and HF-PCIC based OSCs appears to be the decay of V oc and J sc parameters, ascribing to the generation of trap states due to photochemical reaction of photoactive materials with low degree of oxygen permeation (Fig. S30).
Supplementary Figure: UV-vis absorption spectra and photo images of encapsulated films (100 o C annealing for 10 minutes and then encapsulated in glovebox with epoxy sealant) under one-sun equivalent illumination in ambient.
6. Devices: The authors discussed about photooxidation but then they looked at encapsulated devices (figure 6). Description has been added into main text: Slow oxygen permeation is involved for these devices tested in ambient under harsh illumination without UV-filter (Fig. S30).

Response
Updated Figure 6  7. There are many English errors throughout the manuscript. The authors should check carefully.
Response: Thanks for comments. We have carefully checked manuscript and made updates.
In the past few years, the field of organic solar cells achieved much progress due to the development of non-fullerene acceptors. As the efficiencies are quite high, dealing with the stability problems becomes more urgent. This study discussed the chemical stability of the nonfullerene acceptors and provides some insightful results. I carefully checked the comments from reviewer #2 and the corresponding response of the authors. I think the characterizations were well performed and the manuscript was well prepared. I support the publication of this manuscript as I think the results are meaningful and timely although there are debates on some details. The followings list some small questions. 1) About the photostability test, more details about the conditions should be provided, such as the spectra range of the lamp, the film thickness. Does the humidity affect the stability? 2) I noted that the polymer PBDB-TF also shows quick photodegradation. The authors should put more discussion on why? What are the differences when compared to the non-fullerenes, fused, or non-fused? 3) After the illumination treatment, do the degraded fragments volatilized? The optical density is nearly zero. 4) About the conclusion, I think it's better to point out there may be other reasons that affecting the chemical stability of these materials as the studied systems are quite limited when compared to the boom of OSC materials. 5) I think the authors should update the references as the field is growing very fast. It's better to conclude the latest progress reporting record efficiencies, such as 17-18% efficiencies obtained using non-fullerene acceptors like bo-4cl, eC9.