Building Supramolecular Chirality in Organic Solar Cells Enables High-Performing Circularly Polarized Light Detection

High-performing direct circularly polarized light (CPL) detectors are urgently needed for the development of chiral optoelectronics. We herein report the direct CPL detectors based on chiral organic small-molecule donor–fullerene acceptor bulk heterojunction organic solar cells (OSCs). By building supramolecular chirality in active layer of OSCs, the chiral OSC demonstrates its highest short-circuit current dissymmetry factor of 0.17 among state-of-the-art direct CPL detectors possessing intrinsic chirality. It demonstrates a signicant difference in responsivity of 40 mA W − 1 upon opposite CPL illumination, four times higher than that of the best reported chiral perovskite CPL detectors to date. The association between the device metrics crucially relating to CPL detection and the anisotropic factor (g-factor) of the active layers revealed that supramolecular chirality is vital to the CPL detectability of OSCs. Given unique but ubiquitous observed supramolecular chirality induction and transfer in organic conjugated systems, highly sensitive CPL detectors based on chiral OSCs show great potential for promoting the development of practical direct CPL detection.


Introduction
Polarization is one of the most important properties of light. When the electric eld vector of a light beam propagates along the alternative clockwise or counterclockwise helical trajectory, it can be de ned as right-or left-handed circularly polarized light (r-or l-CPL), accordingly. Although the human visual system is insensitive to the handedness of CPL, some creatures, such as beetles, crabs, cuttle sh and mantis shrimp take advantage of unique interactions between their intrinsic chiral tissues or structures and the polarization states of CPL for camou age 1 , visual contrast, communication, bioinformation transfer and navigation [2][3][4] . The unique polarization state enables CPL to be applied to some cutting-edge science and technology breakthroughs, such as drug screening 5 , quantum optics 6,7 , optical communications 7 and imaging [8][9][10] . An intriguing topic closely associated with these issues is the sensitive discrimination of opposite handedness between l-and r-CPL by electronic devices [11][12][13][14][15][16][17][18] . However, the realization of CPL detection in inorganic photodetectors excessively relies on packaged lters consisting of quarter-wave plates and linear polarizers because of the inherent achiral property of inorganic semiconductors. This feature consequently impedes the simpli cation and miniaturization of photodetectors for further integration in conventional electronic equipment 19,20 .
As such, direct CPL detection has greatly progressed by using electronic devices based on active materials with intrinsic chirality. These materials include organic conjugated small molecules and polymers 11,12,16,21 , hybrid organic-inorganic perovskite materials 14,15,18 and plasmonic metamaterials 22,23 . Among these materials, the potential of organic conjugated molecules has attracted great attentions in promoting the advancement of chiral optoelectronics, including e cient CPL detection and beyond. Single-handed intermolecular aggregations from organic conjugated molecules result in the remarkable enhancement in anisotropic factor (g-factor) under unique characters in supramolecular chirality induction, transfer and ampli cation 24,25 . Circular dichroism (CD) represents the difference between the absorptions of l-and r-CPL resulted from the molecular and supramolecular chirality of samples, which originate from the electronic transitions of the chromophore and provide detailed information on chiral interaction 26,27 . Whereas the g-factor is derived from but quanti es the CD intensity thus it is in favor of the lateral comparison of CD intensity among the samples 27 . The enhanced g-factor in organic semiconductors indicates an intensi ed predominant absorption in r-or l-CPL at certain wavelengths, which are vital to the distinct disparity of photoinduced charge carrier densities in photodetectors. This feature immensely caters for the demands for sensitive CPL detection.
Organic solar cells (OSCs) acquire photocurrent via charge transportation and collection after exciton separation upon photoexcitation in active blends consisting of an electron-donating material and an electron-accepting material. Bulk heterojunctions (BHJs) formed in the blends via phase separation between the donor and acceptor from their mixed solutions. As bicontinuous donor and acceptor phases are in favor of the corresponding hole and electron transport and a uent interface is bene cial to the photo-induced exciton separation with a high e ciency, BHJs are recognized as kernel for an e cient photoelectric conversion 28 . Massive organic conjugated semiconductors have been constantly developed for OSCs considered as one of the most promising next-generation energy conversion devices for sustainable clean energy. The structure-function relationship between the aggregation morphologies of the active blends and photovoltaic performance of OSCs has been well established. OSCs are also powerful platforms for optoelectronic detection, particularly as CPL detection diodes 16,21 ; however, this feature is a promising area that is still under development [29][30][31][32] . OSCs with small molecule as a donor and fullerene as an acceptor were extensively studied with well-de ned charge carrier transport and photogeneration on account of appropriate phase separation. Donors possess conjugated chromophores, so they are the predominant counterpart responsible for light absorption in such systems. The appropriate phase separation with a high purity in the active blends enables minimal optical property perturbation, which in turn, elects OSCs to act as one of the most potential candidates for CPL detection.
Supramolecular chirality was built in OSCs via the single-handed assemblies of donors. The chiral OSCs exhibit a speci c detection scenario analogous to the Cotton effect of their chiral donor phases. In the quanti ed study of the g-factor and CPL detection sensitivity of OSCs, an increase in chiral centers of the donor endowed active blends with drastically enhanced g-factor, which in turn produced chiral OSCs to detect CPL accurately. The explicit selective CPL absorption with expanded wavelengths resulted from progressive chirality induction and transfer can be generally observed in organic semiconductors. Our strategy is likely to manifest the potential of chiral OSCs in promoting the development of direct CPL detection.
CD and ultraviolet-visible (UV-vis) spectrums were combined to study the molecular and supramolecular chirality in solutions and spin-coating lms of neat DPP6T donors and their corresponding blends with PC 61 BM acceptors (Fig. 1c, d, e and f). The absorptions of all these chiral DPP6T donors in the solution were almost identical in the UV-vis regions (dash lines in Fig. 1c and Supplementary Fig. 2). The absorption peaks at ca. 640 and 600 nm could be ascribed to the 0-0 and 0-1 vibration bands of the intramolecular charge transfer (ICT) between the DPP-accepting unit and thiophene donor moieties; the peaks at ca. 380 nm were due to intramolecular local π-π* transitions 35 .
These chiral DPP6T donors were CD silent in solutions in near UV-vis (NUV-vis) regions ( Supplementary  Fig. 3). However, supramolecular chirality was easily built in the DPP6T lms via intermolecular π-π stacking, which resulted in the su cient and expanded CD absorption bands of the lms in almost all of the NUV-vis regions ( Fig. 1d and Supplementary Fig. 4). In (s,s,s,s)-and ( , , , )-DPP6T lms, the bisignate CD bands at ca. 390 and 430 nm corresponded to the absorption peak at ca. 420 nm in the UVvis spectrum, which stemmed from the distinct exciton coupling of the intramolecular π-π* transitions. The bisignate CD bands that appeared at 540 and 650 nm corresponded to the main absorption peak at ca. 600 nm in the UV-vis spectrum, which was derived from intramolecular chiral charge transfer. The CD bands that appeared at ca. 740 nm corresponded to the absorption peak at ca. 740 nm in the UV-vis spectrum, which was derived from the intermolecular chiral π-π transitions. The CD absorption was much weaker in the (s,s)-and ( , )-DPP6T lms than in the (s,s,s,s)-and ( , , , )-DPP6T lms as determined by measuring lms with comparative thickness (Fig. 1d and corresponding g-factor see Supplementary Fig. 4). The lms from DPP6T attaching s and pendants showed mirrored CD spectra but identical UV-vis spectra. This result revealed that the enantiomeric stereocenters of the peripheral alkyl chains induced the supramolecular chirality in the DPP6T aggregations.
In comparison with the neat DPP6T lms, the DPP6T donors added with PC 61 BM acceptors showed no de ned effect on the light absorption of the DPP6T counterpart (Fig. 1e). However, the CD intensity remarkably reduced in the (s,s)-and ( , )-DPP6T based blends comparing with their neat lms, but no distinct CD decline was observed in the (s,s,s,s)-and ( , , , )-DPP6T based blends ( Fig. 1d and f, an enlarged view is presented in Supplementary Fig. 5). In addition, no speci c induced chirality belonging to the PC 61 BM acceptor was observed in the blends (Fig. 1f), indicating that no distinctive interaction existed between the DPP6T donors and PC 61 BM acceptors at a supramolecular level. The changes in the CD and UV-vis spectra indicated that explicit phase separation might occur between the donor and the acceptor in the blends. Furthermore, increasing the amount of enantiomeric side chains intensi ed such phase separation and resulted in CD intensity maintenance in the (s,s,s,s)-and ( , , , )-DPP6T counterparts in the blends.
Interestedly, no distinct helical chirality was observed in neither neat nor blended DPP6T lms, though right-handed (s,s,s,s)-DPP6T nano bers and left-handed ( , , , )-DPP6T nano bers were observed from their precipitations via a fast self-assembly process in solution ( Fig. 1g and h, detail see Supplementary Method 2). The molecular chirality from these enantiopure DPP6T transferred to the nanoscale helical chirality of the assemblies via helical intermolecular π-π stacking and these helical nano bers possess accordant CD and UV features with their lm counterparts ( Supplementary Fig. 6). As CD originates from the electronic transitions of the chromophore and provides detailed information on chiral interaction 26,27 . These results indicated that though the expression of such nanoscale helical chirality was restricted in spin-coating lms, they share consistence intermolecular distance and twisting orientation with their helical bril counterparts thus their chiroptically generated excitons were retained. The similar phenomenon was observed in a blended lm of achiral poly uorene polymer with a chiral helicene derivative by Yang et al 36 . The blended lms exhibited strong chiroptical properties despite that no helical structure was observed in the isotropic granular nanoscale morphology. This result would be further veri ed through the structural and morphological analysis of the neat and blended lms.
Molecular stacking and morphological characteristics of the neat and blended chiral lms. Grazing incidence wide-angle X-ray scattering (GIWAXS) measurements were conducted to characterize the molecular stacking and crystallinity of the neat lms and blended active layers. The neat ( , , , )-and (s,s,s,s)-DPP6T lms presented lamellar peaks in both in-plane (IP) and out-of-plane (OOP) directions at q = 0.21 Å −1 ( Fig. 2a and b; corresponding integration in Supplementary Fig. 7a and b). The multiple highorder diffraction peaks in the OOP direction at q = 0.21, 0.44, 0.66, 0.86 Å −1 could be assigned to (100), (200), (300) and (400) re ections, respectively, which were attributed to the ordered arrangement along the chiral alkyl-stacking direction. The neat ( , )-and (s,s)-DPP6T lms showed distinct diffraction peaks at q = 0.56 Å −1 (d = 11.22 Å ) and q = 1.14 Å −1 (d = 5.51 Å ) in the OOP direction corresponding to (200) and (400) re ections, respectively (  The chiral OSCs were used to conduct CPL detection under ambient conditions. The excitation wavelengths were selected on the basis of the CD maximums (valleys and peaks) and minimums (crossovers) of the chiral blends. The I-T switching characteristics were explored by turning the l-or r-CPL on and off with a cycle period of 40 s (Fig. 3a, b, c and d; Supplementary Figs. 9, 10, 11 and 12). The self-powered photovoltaic short-circuit current (I sc ) was coupled with the handedness of CPL. For example, in the chiral OSCs based on (s,s,s,s)-DPP6T, the CPL irradiation at CD maximums (around 426, 540, 606, and 640 nm) with either correct or opposite handedness caused an increase in I sc of chiral OSCs, but the CPL irradiation with the correct handedness caused a more pronounced increase (Fig. 3a and Supplementary Fig. 9a, c and f). This was owing to the fact that CPL irradiations with the correct handedness induced a more e cient photon absorption than the opposite ones, resulting in a relatively higher density of exciton generation in the chiral active layers. The photo-induced excitons subsequently dissociated into free charge carriers in the BHJ active layers and the predominant carrier density induced by the correct CPL brought about a more pronounced increase in I sc than that by the opposite CPL.
Whereas, the CPL irradiation at CD minimums (around 475 and 570 nm) with both the correct and opposite handedness caused a strong but indiscriminate increase in I sc because the CD minimums coincided with the absorption maximums with an e cient light absorption ( Supplementary Fig. 9b and d). The detection scenario was similar but was mirrored by using the ( , , , )-DPP6T-based chiral OSC because of the mirrored chiroptical properties (Fig. 3b and Supplementary Fig. 10a-f). The OSCs based on the (s,s)-and ( , )-DPP6T presented obscure detection sensitivity on CPL owing to their weak chiroptical absorption (Fig. 3c and d; Supplementary Figs. 11 and 12).
To further quantitatively assess the relationship between the CPL detection performance and g-factor of OSCs, we introduced the device metrics, including the dissymmetry factor of short-circuit current (g sc ), responsivity (R), and speci c detectivity (D*) to quantify the relative difference in I sc , the strength of the photoresponse, as well as the noise equivalent power to the device area and electrical bandwidth of the noise measurement upon alternating irradiation of l-and r-CPL 19,21 . We introduced the g-factor as it is in favor of the lateral comparison of CD intensity among the samples 27 Fig. 3f and j). The error bar in the gures was obtained from 10 repeated measurements. The signs of g sc by using alternating OSCs with enantiomeric chirality were opposite. With the g-factor of the OSCs enhanced in the order of magnitudes from 10 −4 to 10 −2 , their g sc was clearly increased in the order of magnitudes from 10 −2 to 10 −1 .
We also calculated R and D* at the selected wavelengths of CD maximums and minimums. The  Fig. 13). These values were one order of magnitude higher than those of the best reported quasi-2D perovskite CPL detectors (1.1 ×10 12 Jones) 15 recently as well as those of the commercialized Si photodiode (10 12 Jones) 37 . The detailed comparisons are presented in Supplementary  Table 4. Additionally, while the g-factor of the OSCs was enhanced in the order of magnitudes, the ΔR and ΔD* (the relative differences in R and D* upon the alternating irradiation of l-and r-CPL, respectively) ampli ed in the order of magnitudes (Fig. 3g, h, k, and i). The results suggested that the response to the CPL of chiral OSCs was consistent with the CD spectrum in the position and intensity of CD absorption bands. The enhanced g-factor enabled the enhanced distinguishability of chiral OSCs between l-and r-CPL; for example, the better selectivity was obtained in the (s,s,s,s)-and ( , , , )-DPP6T:PC 61 BM-based OSCs than that in the (s,s)-and ( , )-DPP6T:PC 61 BM-based OSCs. The highest ΔR (ca. 40 mA W −1 ) was obtained in the (s,s,s,s)-DPP6T-based OSCs at 606 nm, which was four times higher than that of the best reported 1D perovskite CPL detectors (ca. 10 mA W −1 , Supplementary Table 4) 14 .
We extensively explored how the intensity of the incident CPL and the thickness of chiral active layers in uence the CPL detection sensitivity of OSCs because of their great importance in practical CPL detection. The intensity of photocurrent under CPL irradiation was positively dependent on the incident light intensity at given wavelengths, such as 540, 606 and 640 nm, by using ( , , , )-DPP6T:PC 61 BM OSCs (Fig. 4a, and b; Supplementary Fig. 14a); the enhanced irradiation intensity generates increased densities of photon-excited electron-hole pairs and consequently leads to enhanced I sc 38 . Signi cantly, both the J sc and the relative difference in J sc upon alternating irradiation of l-and r-CPL exhibited positively linear increase as the CPL intensity increased (in a double logarithm scale) (Fig. 4c, d, and Supplementary Fig.   14b). The broad detection in the intensity range of CPL is highly important in practical CPL detection 32 , though we could not measure the linear dynamic range (LDR) of these chiral OSCs at the present stage because of the intensity limitation of our CPL source. R and D* were negatively correlated with the intensity of CPL, indicating the detection potential of the chiral OSCs upon weak CPL irradiation. Using the formula derivation, we deduced that the g sc , g R , and g D* , which are the dissymmetry factors of J sc , R, and D*, respectively, were independent of the CPL intensity ( Fig. 4c-f; Supplementary Fig. 14b and 15a, b; Supplementary Note 1). These three parameters are crucial in the real-word applications, which can be used to accurately perform CPL detection without considering the CPL intensity.
To interpret how the thickness of the chiral active layer in uences chiroptical responses, we fabricated OSCs with a series of active layer thicknesses. For example, in ( , , , )-DPP6T:PC 61 BM-based OSCs with chiral active layer thicknesses of 53, 72, 80, 89 and 124 nm, as the thickness increased, the g-factor unexpectedly increased until the thickness reached ca. 100 nm (Fig. 4g, and h). The g-factor should be independent of sample concentration and optical pathlength in a proper range 27 ; however, a thickness of <100 nm was too thin to t in such a range, which led to a decrease in the g-factor. The g sc also decreased because of the decrease in g-factor of the chiral active layer ( Fig. 4i; Supplementary Fig. 16). This result further revealed that the g-factor of the active layers played a key role in the CPL detection sensitivity of chiral OSCs. However, the saturation of g sc lagged behind that of the g-factor as the thickness of the chiral active layers increased (Fig. 4i). Very thin active layers (e.g., <100 nm) were de cient in light absorption and could result in CPL re ection from metal electrodes in OSCs. The re ected light would present a reversed handedness from the incident CPL 11 . As a consequence, the absorption of the re ected CPL with the reversed handedness by the active layer leading to such lag in g sc . Chiral OSCs with active layers thicker than 130 nm may suffer from charge recombination due to long exciton diffusion length; as such, 100-130 nm should be the optimal thickness for effective CPL detection applied to the present case. This nding further revealed that the thickness of chiral active layers is highly important in chiral optoelectronic devices [39][40][41] .

Discussion
In summary, we built supramolecular chirality in OSCs by using enantiomeric DPP6T donors and achiral PC 61 BM acceptors. The supramolecular chirality of the OSCs was explicitly coupled to the polarization state of CPL with a positive CD absorption responding to l-CPL and a negative CD absorption to r-CPL. The detection scenario was analogous to Cotton effect from the supramolecular chirality of donor aggregations, though their molecular chirality in relevant wavelengths was silent. As the g-factor in the active layers increased, the crucial device metrics for CPL direct detection, including g sc , relative differences in responsivity (ΔR) and speci c detectivity (ΔD*) upon the alternating irradiation of l-and r-CPL, were drastically enhanced. By further exploring the performance of these chiral OSCs by adjusting both intrinsic and external parameters, mainly the thickness of chiral active layers and the intensity of incident CPL, we found that g sc, g R , and g D* were independent of CPL intensity, revealing their great feasibility in the practical CPL direct detection under ambient conditions. Given that the enhanced gfactor and expanded CD absorption in organic semiconductors via the principle of supramolecular chirality are unique with vast spaces and ample approaches for further technical evolution, the strategy we demonstrated here is likely to act as one of the most potent ones to constantly promote the development of direct CPL detection. Subsequently, the mixture solution was spin-coated at a rate of 1500 rpm min − 1 to form the active layer.    i, j, k and l The g-factor, gsc (gR, gD*), ΔR and ΔD* (the relative differences in R and D* upon alternating irradiation of l-and r-CPL ) at CD maximums and minimums for the (s,s)-and ( , )-DPP6T:PC61BM-based OSCs.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. LiuLXYangYCPLdetectionbyDPP6TOSCsSI20200620Plain.docx