Introduction

The development of flexible circularly polarized light (CPL) photodetectors has garnered significant attention, not only for their potential applications in areas such as security enhanced or encrypted communications1,2,3,4,5, 3D imaging6,7,8,9, and bionic synapse10,11,12,13,14,15,16,17, but also for facilitating applications in flexible electronics and wearable technology18,19,20. This integration of CPL detection capabilities with flexible and wearable technology includes the potential to expand the human cognition and perception resembling mantis shrimp and jewel scarap beetle, thereby opening the new application fields21,22,23. Despite the potential benefits, a significant challenge remains in designing optoelectronic devices that can simultaneously achieve both high chiroptical activity and mechanical flexibility. Specifically, maintaining an efficient balance between optimizing CPL detectivity and ensuring effective charge transport in flexible materials is difficult, as the structural flexibility often compromises charge carrier mobility24. Therefore, addressing these challenges is critical for advancing the practical implementation of CPL photodetectors in flexible and wearable applications.

Extensive research has been conducted on developing various types of chiral semiconductor materials to fabricate high-performance flexible devices4,25,26,27. Among them, chiral π-conjugated polymers have emerged as promising materials offering several distinct advantages28. Their unique ability to modulate light absorption and optical rotation arises from their inherently tunable molecular structures, which can be controlled by introducing structural asymmetry at the molecular level29,30. This structural flexibility allows researchers to precisely control the chiroptical properties, thereby enhancing their interaction with CPL. Moreover, these chiral π-conjugated polymers offer significant advantages in device fabrication, as they can be readily processed into thin film structures using solution-based methods, such as spin-coating, drop-casting, and inkjet printing31,32,33. These methods are not only cost-effective but also compatible with large-area and flexible substrates, making them ideal for the development of next-generation optoelectronic devices. Consequently, chiral π-conjugated polymers represent a versatile and highly adaptable platform for advancing chiral optoelectronic technologies, particularly in applications requiring both high chiroptical performance and mechanical flexibility.

One strategy for constructing polymer-based CPL photodetectors involves using chiral materials as electron donors and acceptors or as additives to form the bulk heterojunction (BHJ) structures. One representative example is the work by Meskers and colleagues, who reported the first CPL detection using polymer photovoltaic cells with a chiral π-conjugated polymer, poly(fluorene-alt-dithienylbenzothiadiazole) (PFDTBT), as the electron donor, and C60 as the electron acceptor in the active layer34. Lim and coworkers successfully constructed the BHJ devices by blending an achiral carboxylic acid-functionalized poly(3-alkylthiophene) with the chiral small molecule 1,1′-binaphthyl-2,2′-diamine to achieve efficient CPL detection35. Our group recently introduced a novel approach to fabricate helical polymers by using chiral small molecules as the template. Through intermolecular interactions with the chiral small molecules, spiral ordering of polymer chains was induced upon thermal removal of the chiral small molecules to produce chiral flexible layers36. However, chiral devices developed using these blend-based approaches often exhibit a strong dependence on the composition, leading to a trade-off between chiroptical properties and photodetection performance. Additionally, the fabrication of such blend systems typically requires careful self-assembly processing towards high performance.

In contrast, developing intrinsic chiral π-conjugated polymers offers a significant advantage, as they inherently possess strong interactions with CPL while maintaining the excellent charge transport properties37. These single-component polymeric active layers not only simplify the fabrication process but also offer a more stable and efficient platform towards high-performance flexible CPL detection. Furthermore, among organic materials, electron transporting (n-type) organic semiconductors are susceptible to charge trapping induced by water and oxygen under ambient conditions, causing them to lag behind their p-type counterparts. Therefore, the development of novel high-performance chiral n-type π-conjugated polymers is crucial for bridging this performance gap, paving the way for advanced chiral optoelectronic devices with robust and flexible operations.

In this study, we synthesized a modified N2200 copolymer38, a new class of n-type chiral π-conjugated polymer based on naphthalenediimide (NDI) with (S/R)-1-methylheptyl (MH) groups on their side chains and bithiophene (T2) units. Notably, the distinct enhancement of chiroptical activity of (S/R)-P(NDI2MH-T2) thin films is observed after annealing at 250 °C39. A respectable absorbance dissymmetry factor ( | gabs | ) of 2.34 × 10−2 at 382 nm and 1.38 × 10−2 at 670 nm after annealing is observed, which is higher than that of 6.87 × 10−3 at 382 nm and 8.25 × 10−3 at 670 nm before annealing. Theoretical calculations reveal that after annealing, the polymer chains become more extended, facilitating their ordered stacking and promoting the exciton coupling between chiral chromophores, leading to the chiral amplification. In addition, organic phototransistors (OPTs) based on P(NDI2MH-T2) annealed thin films exhibit high electron mobility reaching 7.9 × 10−2 cm2 V−1 s−1. These OPTs show the maximum photoresponsivity (R), external quantum efficiency (EQE), and specific detectivity (D*) of 92 A W−1, 17,600%, and 1.1 × 1012 Jones, respectively. Finally, the first flexible OPTs based on the n-type chiral polymers have been fabricated, which exhibit the maximum gph value of 0.043, as well as good operational/ambient stability and bending resistance. Although the current maximum |gph| value still falls short of the threshold generally regarded as sufficient for optics-free polarization discrimination, our findings demonstrate great potential of chiral π-conjugated polymers in achieving efficient flexible CPL detection, which paves the way for their applications in advanced optoelectronic devices.

Results

Chiroptical, morphological, and structural characterization of chiral polymers

A new class of n-type chiral NDI-based polymers [poly{[N,N′-bis((S/R)-2-methylheptyl)]-1,4,5,8-naphthalene-dicarboximide-2,6-diyl]-alt-5,5′-(2,2′-bithiophene)}, (S/R)-P(NDI2MH-T2) were synthesized via a Pd-catalyzed Stille polymerization (Fig. 1a, Supplementary Figs. 1–6, synthetic details of (S/R)-P(NDI2MH-T2) were reported in the Supplementary Information). The number average molecular weights (Mn) and dispersity (Ð) were determined to be 10.2 kg mol−1 (Ð = 3.30) for (S)-P(NDI2MH-T2) and 12.6 kg mol−1 (Ð = 3.54) for (R)-P(NDI2MH-T2) (Supplementary Tables 1, 2, Supplementary Figs. 7, 8). Given that the properties of the (S)- and (R)-enantiomers, including optical properties, thermal stability, thermal transitions, and surface morphology are essentially identical, we only focused on one of the enantiomers

Fig. 1: Molecular structures and optical/chiroptical properties of (S)- and (R)-P(NDI2MH-T2).
figure 1

a Chemical structures of (S)- and (R)-P(NDI2MH-T2). b UV-vis absorption spectra of annealed (R)-P(NDI2MH-T2) thin films. c Circular dichorism (CD) spectra and d gabs of annealed (S)- and (R)-P(NDI2MH-T2) thin films.

Full size image

for further investigation. The optical absorption spectra of polymer thin films exhibited two main absorptions, one at the UV region (π–π* band) and one at the red region (charge transfer band), located at 380 and 650 nm for (R)-P(NDI2MH-T2) (Fig. 1b).Then, we examined the thermal stability and crystalline properties of (R)-P(NDI2MH-T2). All polymers were thermally stable below 300 °C based on thermogravimetric analysis (Supplementary Fig. 9). The glass transition, crystallization transition and melting temperature were 148.2 ℃, 231.8 ℃, and 246.1 ℃, respectively, observed in differential scanning calorimetry (Supplementary Fig. 10). As shown in Supplementary Fig. 11, the cyclic voltammetric results showed that the LUMO is at −3.78 eV and the HOMO is at −5.75 eV for (R)-P(NDI2MH-T2), which were consistent with the theoretical calculations: LUMO = −3.39 eV and HOMO = −5.42 eV (Supplementary Fig. 12). To investigate the chiroptical properties of the polymer at different annealing temperatures, the (S)-P(NDI2MH-T2) polymer films were subsequently annealed at 150 °C, 200 °C and 250 °C. As shown in Supplementary Fig. 13, the chiroptical activity of the polymer films gradually increased with increasing annealing temperature. Notably, a remarkable |gabs| of 2.34 × 10−2 at 382 nm and 1.38 × 10−2 at 670 nm were obtained after annealing, which was higher than the |gabs| of 6.87 × 10−3 at 382 nm and 8.25 × 10−3 at 670 nm before annealing. (Fig. 1c, d).

Atomic force microscopy (AFM) was conducted to investigate the morphological changes of chiral π-conjugated polymer thin films upon thermal annealing. Both the pristine and annealed films exhibited smooth, uniform morphologies with low surface roughness, as shown in Fig. 2a, b. Interestingly, the root-mean-square roughness (Rrms) of the chiral polymer thin films increased from 0.540 nm to 0.875 nm after annealing at 250 °C, accompanied by the formation of distinct aggregates. This increase in surface roughness, along with the appearance of aggregates, suggested enhanced intermolecular interactions among polymer molecules and an increase in crystallinity. These morphological changes upon annealing indicated that thermal treatment effectively promotes the molecular ordering in polymer films, which increased their chiroptical properties40.

Fig. 2: Morphological and structural characterization of pristine and annealed P(NDI2MH-T2) thin films.
figure 2

AFM height images of a pristine and b annealed (R)-P(NDI2MH-T2) thin films. 2D GIWAXS patterns of c pristine and d annealed (S)-P(NDI2MH-T2) thin films. GIWAXS-scattering profiles of pristine and annealed (S)-P(NDI2MH-T2) thin films, along e qxy and f qz.

Full size image

In order to characterize the evolution of intermolecular packings of chiral π-conjugated polymer thin film upon annealing at 250 °C, the grazing incidence wide angle X-ray scattering (GIWAXS) was employed. The pristine (S)-P(NDI2MH-T2) thin films only revealed a broad, weak GIWAXS peak along the vertical axis (nominally qz) at q = 1.51 Å−1, corresponding to loose π–π stacking (Fig. 2c). The annealed (S)-P(NDI2MH-T2) thin films, as displayed in Fig. 2d, exhibited more pronounced scattering features than the pristine thin films indicative of improved ordering. We also observed an analogous structural shift in the (R)-P(NDI2MH-T2) thin films upon thermal annealing, as shown in Supplementary Fig. 14. Interestingly, the π–π stacking distance of (S)-P(NDI2MH-T2) was decreased from 4.17 Å to 3.81 Å and the (100) lamellar packing crystalline domain size increased from 27 Å to 134 Å after thermal annealing, which further confirms that the polymer crystallinity was significantly enhanced after annealing (Fig. 2e, f).

Molecular dynamics simulation and conformation analysis of chiral polymers

To further investigate the conformational changes of the chiral polymer before and after annealing, molecular dynamics simulations were conducted. Each periodic boundary condition box contained 94 polymer chains, and each polymer comprising 12 repeating units, as shown in Fig. 3a, b. The radial distribution functions (RDFs, g(r)) were calculated for (R)-P(NDI2MH-T2) before and after annealing (Fig. 3c). The RDFs measure the probability of finding a particle at a certain distance from a reference particle, with higher g(r) peaks indicating greater packing density at specific distances. In the annealed chiral polymer system, significant increases in g(r) were observed around 5 Å, corresponding to the π–π stacking distances. This indicates that annealing causes the polymer chains to become more extended. This enhanced extension leads to the improved π-π stacking interactions, resulting in a more ordered structure. Thus, the annealing significantly enhances the structural organization of the chiral polymer system.

Fig. 3: Molecular dynamic simulations of the pristine and annealed (R)-P(NDI2MH-T2).
figure 3

a Snapshot of the pristine film structure. b Snapshot of the annealed film structure. c Radial distribution functions of the two systems. d Probability distribution of the mean square radius of gyration for the two systems. e Typical structure of the pristine film, with pink spheres representing spheres drawn with the mean square radius of gyration as the radius. f Typical structure of the annealed film, with pink spheres representing spheres drawn with the mean square radius of gyration as the radius.

Full size image

To verify whether the polymer exhibits better extensibility in the annealed films, we statistically analyzed the mean square radius of gyration (Rg²) of the polymer chains before and after annealing. A larger Rg² indicates a more extended molecular conformation. As shown in Fig. 3d–f, the Rg² of the pristine thin films is primarily centered around 15 Å, while the Rg² of the annealed thin films significantly increases to 20 Å. This increase confirms that the annealing process enhances the extensibility of the polymer chains, thereby improving the overall order of the system. As the polymer chains became more extended and ordered through annealing, the intermolecular and intramolecular packings were further optimized, thereby significantly improving the chiroptical properties and CPL detection.

Electrical and optoelectronic properties of chiral polymer-based OPTs

The solution-processable nature of chiral π-conjugated polymers facilitates the fabrication of thin film optoelectronic devices. Therefore, OPTs with a bottom-gate-top-contact configuration were fabricated through spin-coating of the chiral π-conjugated polymers on n-octadecyltrimethoxysilane (OTS)-treated Si/SiO2 substrates, then the films were annealed at 250 °C for 10 min. The n-type transfer and output characteristics of the OPTs based on annealed (R)-P(NDI2MH-T2) thin films were measured and shown in Fig. 4a, b. Interestingly, they exhibited weak ambipolar charge transport behavior with typical V-shaped transfer characteristics owing to the narrow bandgap of the chiral polymer and its energy level alignment with the gold electrodes41. From the transfer curves, averaged field-effect electron mobility of 7.9 × 10−2 cm2 V−1 s−1 and hole mobility of 1.3 × 10−3 cm2 V−1 s−1 were estimated, respectively (Supplementary Fig. 15, Table 3). The derived electron mobilities were similar to those reported for OFETs based on commercial P(NDI2OD-T2) (commonly known as N2200), which has the same backbone as the chiral polymers42,43,44,45. These results suggest that the chiral arrangement of polymer side chains did not significantly disrupt electron transport pathways. In addition, the short alkyl chains might boost the charge transport properties by promoting short lamellar spacing and π-π stacking distances46,47. As a result, high-performance n-type operation was achieved from the chiral polymer-based thin films, without the need for complicated self-assembly or doping processes48,49. However, non-ideal current behavior was observed in the high-voltage regime, possibly originating from various factors including bias stress, charge trapping, contact resistance, and carrier-carrier interactions50,51. This implies that there remains considerable room for improvement in both material and device engineering.

Fig. 4: Electrical and optoelectronic performance of (R)-P(NDI2MH-T2)-based OPTs.
figure 4

a Transfer and b output characteristics of (R)-P(NDI2MH-T2)-based OPTs in dark condition. c IDS–VGS characteristics of (R)-P(NDI2MH-T2)-based OPTs under monochromatic light illumination (λ = 650 nm) varying in intensity (VDS = 100 V). Optoelectronic performances of (R)-P(NDI2MH-T2)-based OPTs. d EQE, D*, e R, and P values were plotted as a function of VGS under 650 nm monochromatic light irradiation at VDS = 100 V. f Real-time photoswitching graph with biexponential fitting curves for rise and decay time measurement.

Full size image

Under illumination of the monochromatic light (λ = 650 nm), (R)-P(NDI2MH-T2)-based OPTs clearly exhibited a negatively shifted threshold voltage and increased drain current due to the contribution of photogenerated charge carriers (Fig. 4c). To further quantify the photodetection properties of OPTs, we evaluated R, EQE, photosensitivity (P), and D* under various light intensities (Fig. 4d, e). The maximum R and D* values could reach 92 A W−1 and 1.1 × 1012 Jones, under the illumination intensity of 50 µW cm−2. The transistor configuration allowed for signal amplification through the application of an additional bias, thereby enhancing the sensitivity of photodetectors. The maximum values of these photodetection properties at different illumination intensities are summarized in Supplementary Table 4, Fig. 16. The overall photodetection properties of OPTs decreased as the incident light intensity increased, which originated from the shorter-lived photogenerated charge carriers due to saturated charge trap states under higher intensities52,53. In addition, real-time photoswitching tests were further conducted. Under pulsed illumination, (R)-P(NDI2MH-T2)-based OPTs also exhibited stable and reversible operation (Supplementary Fig. 17). The experimental results of photoswitching and the corresponding biexponential fitting curves were shown in Fig. 4f. The rise and decay times were extracted as the time required for the current to rise from 10 to 90% of its peak value after illumination was turned on, and from 90% to 10% of its maximum value for the current to decay after illumination was turned off, respectively. The extracted rise and decay times were around 6 ms and 20 ms, respectively, which are quite fast compared to the previously reported chiral organic semiconductor based OPTs54, implying their high potential for real-time chiral optoelectronic devices.

Flexible CPL photodetector applications

We have successfully fabricated the first flexible CPL photodetector based on the chiral π-conjugated polymer, by using polydimethylsiloxane (PDMS) and polyethylene naphthalate (PEN) as the flexible dielectric and substrate, respectively (Fig. 5a). The detailed device fabrication procedures can be found in Supplementary Fig. 18. The direct film transfer of chiral polymer films from OTS-treated Si/SiO2 to PEN/ITO/PDMS layers allowed high-temperature thermal annealing of the chiral polymer films prior to transfer. This process resulted in improved polymer chain packings, enhancing both charge transport and chiroptical properties. The n-type transfer characteristics of the (R)-P(NDI2MH-T2)-based flexible devices were shown in Fig. 5b. Then, the mechanical flexibility and stability of the fabricated devices were tested. As shown in Fig. 5c, electron mobility retained 87% of its initial value, and the threshold voltage remained relatively unchanged after 100 bending cycles with a bending radius of 25 mm.

Fig. 5: Flexible OPT applications of P(NDI2MH-T2).
figure 5

a Schematic illustration (left) and photograph (right) of P(NDI2MH-T2)-based flexible OPTs. b IDS-VGS characteristics of (R)-P(NDI2MH-T2)-based flexible OPTs in the dark condition. c The change of mobility and threshold voltage depending on the bending cycles at 25 mm bending radius. d Real-time photoresponse of (R)-P(NDI2MH-T2)-based flexible OPTs under CPL irradiation (λ = 670 nm). e gph of (R)-P(NDI2MH-T2)-based flexible OPTs after bending. Bending tests were conducted on the fabricated device by moving the bending stage at a speed of 0.1 cm/s. Regarding gph values were acquired from more than 5 individual OPTs.

Full size image

To investigate the CPL detection capability of the (R)-P(NDI2MH-T2)-based flexible OPTs, we measured the real-time photoresponse of the flexible OPTs under alternating left-handed (LH) and right-handed (RH) CPL (λ = 670 nm) (Fig. 5d). The gph was derived by following equation:

$${g}_{\text{ph}}=\frac{2({I}_{\text{L}}-{I}_{\text{R}})}{{I}_{\text{L}}+{I}_{\text{R}}}$$
(1)

where IL and IR describe the photocurrent of the device under LH- and RH-CPL irradiation, respectively. The calculated average |gph| was approximately 0.043, demonstrating efficient CPL detection. Additionally, the fabricated OPTs also exhibited CPL detection capability in p-type operation (Supplementary Fig. 19). Interestingly, the chiral P(NDI2MH-T2)-based flexible OPTs exhibited a reversal in CPL handedness preference in photocurrent compared to the absorption preference observed in chiral polymer films. The major absorption of the preferred CPL handedness did not contribute to the photocurrent in channel at semiconductor-dielectric interface positioned below55. Therefore, less absorbed CPL that reached in the below channel generated a larger photocurrent, resulting in the reversal of CPL handedness preference. This interpretation is further supported by thickness-dependent CPL detection measurements, which show that thicker films yield larger |gph | (Supplementary Fig. 20). (R)-P(NDI2MH-T2)-based flexible OPTs exhibited stable CPL detection after bending at radii of 25 mm and 20 mm, with a slight decrease of gph from 0.043 to 0.027 and 0.024, respectively (Fig. 5e, Supplementary Fig. 21). These findings indicate that the chiral flexible device maintains reliable and consistent CPL detection capabilities during flexible operation.

Discussion

In summary, we have successfully synthesized and demonstrated a novel n-type chiral π-conjugated polymer, (S/R)-P(NDI2MH-T2), which exhibits great promise towards flexible CPL detections. Following annealing at 250 °C, the (S/R)-P(NDI2MH-T2) thin films showed significant chiroptical enhancement, with |gabs| reaching 2.34 × 10−2 at 382 nm and 1.38 × 10−2 at 670 nm. This improvement is attributed to increased polymer chain extension and enhanced exciton coupling between chiral chromophores, as confirmed by AFM, GIWAXS, and molecular dynamics simulations. This chiral amplification led to the fabrication of the first chiral π-conjugated polymer (S/R)-P(NDI2MH-T2) based flexible CPL photodetectors, which exhibited a maximum gph of 0.043. Additionally, the fabricated organic thin film transistors demonstrated high electron mobility, photoresponsivity, and detectivity. Meanwhile the flexible device maintained excellent operational stability, ambient resistance, and bending durability, after repeated bending cycles. These results confirm the feasibility and robustness of chiral π-conjugated polymers for flexible CPL detection and underscore their potential in advancing high-performance optoelectronic applications. However, scaling the synthesis and fabrication for industrial use presents challenges, including maintaining material consistency, optimizing large-scale production, and reducing costs. Future efforts should focus on scalable synthesis techniques and cost-effective fabrication methods to support commercial applications. Our work demonstrates the potential of integrating single component chiral π-conjugated polymer into future wearable and flexible electronic technologies, enabling the development of next-generation, high-performance CPL photodetectors.

Methods

Synthesis of chiral polymers

All reagents and deuterated solvents were used as purchased without further purification. Compounds 2,6-dibromonaphthalene-1,4,5,8-tetracarboxylic dianhydride (NDA-Br2), (S/R)-2-aminooctane and 5,5′-bis(trimethylstannyl)-2,2′-bithiophene purchased from commercial suppliers.

For the synthesis of chiral N,N′-bis(1-methylheptyl)-2,6-dibromonaphthalene-1,4,5,8-bis(dicarboximide) [(S/R)-NDI2MH-Br2] compounds: A mixture of NDA-Br2 (2.50 g, 5.87 mmol), (S/R)-2-aminooctane (2.18 g, 17.10 mmol), propionic acid (15 mL), and o-xylene (45 mL) was stirred under argon at 140 °C for 24 h. Upon cooling to room temperature, solvents were removed in vacuo and the residue was purified by column separation on silica gel, with eluent CH2Cl2 : hexane (1:1, v/v) to afford (S)-NDI2MH-Br2 as slightly yellow solid (yield 7.09%) and (R)-NDI2MH-Br2 as slightly yellow solid (yield 6.57%).

For the synthesis of chiral polymer poly{[N,N’-bis(1-methylheptyl)-1,4,5,8-naphthalene diimide-2,6-diyl]-alt-5,5’-(2,2’-bithiophene)} [(S/R)-P(NDI2MH-T2)]: A mixture of (S/R)-NDI2MH-Br2 (0.30 g, 0.46 mmol), 5,5’-bis(trimethylstannyl)-2,2’-bithiophene (0.23 mg, 0.46 mmol) and Pd(PPh3)2Cl2 (16 mg, 0.02 mmol) in anhydrous toluene (16 mL) were stirred under argon at 90 °C for 4 days. Bromobenzene (0.20 mL) was then added and the reaction mixture was maintained at 90 °C for an additional 12 h. Upon cooling to room temperature, a solution of potassium fluoride (1.00 g) in water (2.00 mL) was added. This mixture was stirred at room temperature for 2 h before it was extracted with chloroform (60 mL × 3). Organic layers were combined, washed with brine (50 mL × 3), then water (50 mL × 3), dried over anhydrous sodium sulfate, and concentrated on a rotary evaporator. The residue was redissolved with a small amount of chloroform and precipitated in methanol three times. The obtained dark blue solid product was purified by Soxhlet extraction with acetone for 72 h. The remaining solid residue was redissolved in chloroform (20 mL) and the resulting mixture was heated to boil. Upon cooling to room temperature, the chloroform solution was filtered through a 5 μm filter, and the filtrate was added slowly to methanol (250 mL). The precipitates were collected by filtration, washed with methanol, and dried in vacuum, leading to afford (S)-P(NDI2MH-T2) as dark blue solid (yield 52%) and (R)-P(NDI2MH-T2) as dark blue solid (yield 43%).

Computational methods

All molecular dynamics simulations were conducted utilizing the GROMACS (2023.3) software package56, employing the GROMOS force field. The restrained electrostatic potential fitting technique was applied to derive partial charges for each molecule.

For the (R)-P(NDI2MH-T2) thin film, 94 (R)-P(NDI2MH-T2) molecules, each containing 12 repeat units, were randomly placed in a 50 × 50 × 50 nm3 simulation box. For the non-annealed film, the system underwent energy minimization, followed by a 200 ns simulation at 300 K and 20 bar to achieve a reasonable density. This was succeeded by a 50 ns equilibration phase and a 50 ns production run at 300 K and 1 bar. For the annealed film, the system underwent energy minimization, followed by a 100 ns simulation at 523 K. Subsequently, a gradual annealing process from 523 K down to 300 K at 1 bar was conducted over 100 ns, followed by a 50 ns equilibration phase and a 50 ns production run at 300 K and 1 bar.

The LINCS algorithm was utilized to constrain covalent bonds involving hydrogen atoms57. A simulation time step of 1.0 femtoseconds was employed. Both the pressure and temperature were regulated using the Parrinello–Rahman barostat and the Nose–Hoover thermostat58,59, respectively. All analysis was performed on the frames extracted from the production run. The graphics were processed by the Visual Molecular Dynamics (VMD) program60. Analyses were conducted using the MD Analysis61 Python library.

Device fabrication

OPTs were fabricated using heavily doped silicon wafers (n-type, < 0.004 Ω·cm) with 300-nm-thick SiO2 (Ci = 10 nF cm−2). The SiO2/Si wafers were cleaned with piranha solution for 30 min (a mixture of 70 vol% H2SO4 and 30 vol% H2O2) and treated with a UV-ozone cleaner. And then, the wafer surface was treated with OTS solution to form a self-assembled monolayer (SAM). The OTS solution (3 mM in trichloroethylene) was spin-coated at 500 rpm for 5 s, at 1500 rpm for 30 s, and at 500 rpm for 5 s onto the wafers. Afterwards, the wafers were kept in a glass desiccator under NH4OH vapor overnight. The wafers were sequentially washed with toluene, acetone and isopropyl alcohol, and dried with nitrogen blowing. Synthesized P(NDI2MH-T2) (10 mg mL–1 in chloroform) was spin-coated at 600 rpm for 60 s and at 1200 rpm for 10 s onto the wafers after 30 s wetting time. After the substrates were thermally annealed at 250 °C in a glove box, gold source/drain electrodes (40 nm thickness) were thermally evaporated in a rate of 0.1 Å s–1 under high vacuum condition (< 5 × 10–6 torr). The source/drain patterns were formed using shadow masks with a channel length (L) of 50 µm and a channel width (W) of 1000 µm (W/L = 20).

To enable the annealing process of polymer films in flexible devices, P(NDI2MH-T2) (10 mg mL–1 in chloroform) was spin-coated onto an OTS-treated wafer and thermally annealed using the same process as in OPT fabrication. PEN/ITO was used as the substrate, and PDMS was used as the dielectric layer for the flexible device. A dilute PDMS solution (Sylgard 184, 1.8 g base with 0.18 g cross-linker in 10 mL of hexane) was spin-coated onto an polymer-coated wafer at 5000 rpm for 60 s. In order to prevent deformation of the flexible substrates, PDMS in contact with the PEN/ITO substrate was cured at 120 °C overnight under pressure applied by weights, with the polymer-coated wafer placed on top in an inverted position. After the dielectric layer on the PEN/ITO substrate was fully cured, the wafer was carefully removed to transfer the polymer thin film onto the cured PDMS on the PEN/ITO substrate. Finally, gold source/drain electrodes (40 nm thickness) were thermally evaporated onto the polymer layer.

Optoelectrical measurements

The current-voltage characteristics of the OFETs were measured inside a vacuum chamber, using a Keithley 4200-SCS parametric analyzer. In the OPT measurements, a 650 nm laser (19.6 mW cm–2) was used to generate linearly polarized light. CPL illumination was generated through a linear polarizer and a quarter-wave plate (Thorlab) between the light source and samples.

Estimation of optoelectrical properties

The field−effect mobility (μ) and VT were estimated in the saturation regime with the following equation:

$${I}_{{DS}}=\frac{W}{2L}{{\mu }C}_{{\rm{i}}}{({{V}_{{\rm{GS}}}-V}_{\text{T}})}^{2}$$
(2)

where Ci is the capacitance per unit area of the gate dielectric.

In order to investigate the photodetection properties for CPL photodetectors, R and P were calculated from I-V characteristics coupled with light irradiation. The R and P values were determined using the following equations:

$$R=\frac{{I}_{\text{ph}}}{{P}_{\text{inc}}}=\frac{{I}_{\text{light}}-{I}_{\text{dark}}}{{P}_{\text{inc}}}$$
(3)
$$P=\frac{({I}_{\text{light}}-{I}_{\text{dark}})}{{I}_{\text{dark}}}$$
(4)

where Iph is the photocurrent, Pinc the incident illumination power on the channel of the device, Ilight the drain current under illumination, and Idark the drain current in the dark, respectively. In addition, EQE of devices was calculated which can be defined as the ratio of number of photogenerated carriers that practically enhances the drain current to the number of photons incident onto channel area, using the following equation:

$${\text{EQE}}=\frac{({I}_{\text{light}}-{I}_{\text{dark}}){hc}}{{e{P}_{\text{inc}}A\lambda }_{\text{peak}}}$$
(5)

where h is the plank constant, c the speed of light, e the fundamental unit of charge, A the area of the transistor channel, and λpeak the peak wavelength of the incident light, respectively.

Detectivity usually describes the smallest detectable signal, which allows comparisons of phototransistor devices with different configurations and areas. D* was evaluated within this study using the following equations:

$${D}^{* }=\frac{\sqrt{A}}{{NEP}}$$
(6)
$${NEP}=\frac{\sqrt{{I}_{n}^{2}}}{R\sqrt{\varDelta f}}$$
(7)

In these equations, NEP is the noise equivalent power, \({I}_{n}^{2}\) the measured noise current, and Δf the bandwidth. The specific detectivity can be understood as the signal-to-noise ratio produced by a photodetector with an active area of 1 cm2, operating under an incident power of 1 W and an electrical bandwidth is 1 Hz. If the major limit to detectivity is shot noise from the drain current under dark conditions, D* can be simplified as equation:

$${D}^{* }=\frac{R}{\sqrt{(2e\cdot {I}_{{dark}}/A)}}$$
(8)

Materials analysis

The 1H NMR and 13C NMR spectra were measured on Bruker Avance 400 MHz or 600 MHz spectrometer. The absorption spectra were measured using a Cary 5000 UV-vis-NIR spectrophotometer. The CD results were obtained using a J-1700 Spectropolarimeter (JASCO). Differential scanning calorimetry was conducted under a nitrogen atmosphere at a ramping rate of 10 °C min−1 using a Discovery DSC instrument (TA Instruments). Thermogravimetric analysis was conducted using a Discovery TGA instrument (TA Instruments). Film morphology, thickness and roughness were analysed via AFM (NX-10, Park Systems, Korea) in True Non-Contact mode. GIWAXD measurements were conducted on the PLS-II 9 A U-SAXS beamline at the Pohang Accelerator Laboratory in Korea. X-rays emitted from the in-vacuum undulator were monochromated at 11.08 keV using a double-crystal monochromator and the incidence angle of the X-ray beam was set to a range of 0.12°. Polarized optical microscopy images were obtained using an Olympus BX53F microscope equipped with polarizers oriented in a perpendicular (crossed) arrangement.