Vitamin C modified cathode interlayer for efficient opaque and semitransparent organic photovoltaics

Abstract

Minimizing interfacial electrical losses is a key pathway for both opaque and semitransparent organic photovoltaics (STOPVs). Here, we propose a new method of using vitamin C (VC) to modulate the performance of cathode interlayers (CILs) and thus optimize the electrical losses of devices. Molecular dynamics simulation reveals that VC can improve the long-range packing order of CIL molecules, which is beneficial to charge transport. In addition, the CIL modulation strategy facilitates the establishment of a better contact between the active layer and the metal cathode and thus suppresses the charge carrier recombination. Comprehensive optimizations, including electrode thickness, donor : acceptor ratio and optical engineering, are further carried out to optimize the energy conversion and visible transmission for high-performance STOPVs. As a result, the opaque device exhibits a champion power conversion efficiency of 19.8%, and the best STOPV device shows a high light utilization efficiency of 4.53%, where the efficiency and the average visible transmission are 12.1% and 37.4%, respectively.

---

Jiayu Wang

Jiayu Wang received his B.S. degree in Chemistry from Beijing Institute of Technology in 2014. Then he joined Prof. Xiaowei Zhan’s group at Peking University and obtained his Ph.D. in Advanced Materials and Mechanics in 2019. After working as a research assistant in Prof. Xiaowei Zhan’s group at Peking University and a postdoctoral researcher in Prof. Xinhui Lu’s group at The Chinese University of Hong Kong, Dr Wang joined College of Polymer Science and Engineering, Sichuan University as an associate professor in 2022. His research interests are focused on the design and synthesis of organic functional materials and their applications in optoelectronics.

1 Introduction

In recent years, organic photovoltaics (OPVs) have attracted much attention in the renewable energy fields due to their lightweight, flexible, and solution-processable properties.^1–10 Through material design and device engineering, OPVs can absorb ultraviolet or near-infrared light for power generation and allow visible light to pass through, which makes semitransparent organic photovoltaics (STOPVs) show promising applications in the field of building-integrated photovoltaics.^11–13 High average visible transmittance (AVT) is a prerequisite for STOPV applications in power generating windows;^14–17 however, this tends to reduce the light absorption capacity of the device, which inevitably reduces its power conversion efficiency (PCE).^18–20 Light utilization efficiency (LUE), defined as the product of the PCE and the AVT (LUE = PCE × AVT), is introduced to characterize the balance between photovoltaic performance and transparency.^21–23

Reducing the electrical losses as well as balancing absorption and visible transmission are two key pathways to achieving high-performance STOPVs.^24–26 As a key component of devices, the cathode interlayer (CIL) plays a crucial role in minimizing electrical losses.^27–31 Metal oxides such as ZnO and amine/imine-functionalized polymers,³² perylene diimides³³ and naphthalene diimides³⁴ are common CILs materials. Besides the design of new materials, modifying CILs with other compounds is another effective strategy to improve device efficiency.^35–38 For example, an in situ-derived SiO_xN_y passivation layer was deposited on a ZnO CIL, achieving PCEs over 18% in 100 mm² devices and an estimated operational lifetime over 16 years.³⁹ Trihydroxybenzoic acid was used to optimize the morphological and electrical properties of the PDINN CIL, and thereby improved PCEs from 18.2% to 19.3%.⁴⁰ A suitable CIL can form an effective ohmic contact between the active layer and the metal cathode, which is conducive to reducing the interface trap density and mitigating carrier recombination, thus enhancing the PCE of the device.^41,42 In addition, light field modulation of STOPVs is conducive to optimize absorption and visible transmission of the device, which enables to achieve high LUE.^43–46

Based on this principle, we propose a new method of using ascorbic acid, also known as vitamin C (VC), which is a simple but essential nutrient for humans, to modulate the performance of the CIL and thus minimize the electrical loss of the device. VC can form hydrogen bonds with PDINN via its hydroxyl groups interacting with the carbonyl or secondary amine groups of PDINN, thereby enhancing the long-range packing order as well as the charge transport properties of the CIL. In addition, the interfacial contact between the CIL and the active layer can be improved by adding an appropriate amount of VC to PDINN, which promotes the formation of a smoother CIL film morphology. The CIL modulation strategy facilitates the establishment of a better contact between the active layer and the metal cathode and suppresses the carrier recombination, thus reducing the electrical losses of the device. Compared with the PDINN-based device (PCE of 19.0%), the PDINN:VC-based device exhibits a PCE of up to 19.8%. STOPVs are further prepared using the above optimized CIL. And the absorption and visible transmission of the devices are optimized by adjusting the silver (Ag) electrode thickness, modulating the donor : acceptor (D : A) ratio, and using an optical engineering strategy. Ultimately, the optimized STOPV device achieves a LUE of up to 4.53% with a PCE of 12.1% and an AVT of 37.4%.

2 Results and discussion

2.1 Preparation and characterization of the VC-based cathode interlayer

The chemical structures of VC and PDINN are shown in Fig. 1a. We first explored the effect of VC incorporation on the morphology evolution of the CIL using scanning electron microscopy (SEM). As shown in Fig. 1b, the pristine PDINN film exhibits an uneven surface morphology with some cracks. After the addition of VC, the PDINN:VC film exhibits a smoother and flatter surface morphology, meanwhile the cracks disappear (Fig. 1c). The smoother and flatter surface morphology of PDINN:VC facilitates its electrical contact with the metal cathode.^47,48 The surface properties of the CILs are further investigated by measuring the contact angles of water (H₂O) and diiodomethane (DIM) on PDINN and PDINN:VC films. The contact angles of H₂O and DIM on PDINN films are 21.6° and 44.8°, respectively, while those on PDINN:VC films are 22.8° and 47.6°, respectively (Fig. S1†). The surface energies of PDINN and PDINN:VC films are calculated as 71.3 mN m⁻¹ and 70.0 mN m⁻¹ using the Owens–Wendt–Rabel–Kaelble (OWRK) formula (Table S1†). Compared to PDINN, the PDINN:VC CIL shows a closer surface energy with the PM6:BTP-eC9 active layer (18.9 mN m⁻¹),⁴⁰ which could contribute to better surface contact.

Fig. 1 (a) The chemical structures of VC and PDINN. SEM morphology of (b) PDINN and (c) PDINN:VC films.

Molecular modelling was conducted to reveal the effects of VC on PDINN at the molecular level. The geometries of PDINN and VC were optimized first with density functional theory using ORCA 5.0.4.^49,50 Molecular dynamics (MD) simulations of PDINN and PDINN:VC films were conducted using Gromacs-2025.0 (ref. 51) and the results were visualized using VMD.⁵² PDINN molecules tend to form packing clusters through π–π stacking of the molecular backbones in both PDINN and PDINN:VC films (Fig. 2a and b). For clearer demonstration, a packing cluster consisting of 16 PDINN molecules with neighbouring VC molecules is extracted (Fig. 2c). The electrostatic potential map indicates that the oxygen atoms of carbonyls and nitrogen atoms of amines in PDINN are negatively charged, while the hydrogen atoms of hydroxyls in VC are positively charged (Fig. S2†); thus multiple hydrogen bonds are formed in the cluster. Specifically, the hydrogen bonds are formed between the hydroxyl of VC and carbonyl or secondary amine of PDINN (Fig. 2d), through which VC can influence the packing behaviour of PDINN.

Fig. 2 The MD-simulated (a) PDINN and (b) PDINN:VC films. The molecular backbones of PDINN molecules are in green, the side chains of PDINN molecules are in grey and the VC molecules are in cyan. (c) A packing cluster of PDINN molecules with neighbouring VC molecules in the PDINN:VC film and (d) a zoomed-in section of a packing dimer in the cluster, where orange dashes refer to hydrogen bonds. (e) The radial distribution function (RDF, g(r)) of the molecular backbones of PDINN in the corresponding films. (f) N 1s and (g) O 1s XPS spectra of PDINN and PDINN:VC films.

The radial distribution function (RDF) was calculated for the molecular backbones of PDINN in the simulated films (Fig. 2e), which describes the average distribution of backbone atoms around a reference backbone atom at a given distance (r). The positions of RDF peaks in both films are almost identical, with the first peak located at r = 0.36 nm, which is within the typical π–π stacking distance range. The r values of other peaks increase by ∼0.35 nm one by one, referring to the stacking of more molecules. Relative to the PDINN film, the PDINN:VC film shows more peaks at larger r, indicating that VC can induce an increased long-range packing ordering of PDINN molecules, which is beneficial to charge transport.

X-ray photoelectron spectroscopy (XPS) was used to experimentally investigate the interactions between VC and PDINN. The N 1s and O 1s XPS spectra of PDINN and PDINN:VC films are shown in Fig. 2f and g, respectively. The N 1s peak of the PDINN film is located at 399.3 eV, which is shifted toward a higher binding energy of 399.6 eV due to the addition of VC (Fig. 2f). Meanwhile, compared with the O 1s peak of PDINN (located at 531.2 eV), the O 1s peak in the PDINN:VC film shifts toward a lower binding energy of 531.0 eV (Fig. 2g). The decrease in the electron cloud density of nitrogen atoms and the increase in the electron cloud density of oxygen atoms indicate the formation of hydrogen bonding interactions between PDINN and VC.

The effect of VC incorporation on the optical and electrical properties of the CIL is further characterized. Fig. S3a† shows the absorption spectra of PDINN and PDINN:VC films. The maximum absorption peak of both films is located at 473 nm and the absorption edge is located at 631 nm, indicating that the addition of VC has basically no effect on the absorption of PDINN. In addition, the conductivity of the CIL is measured using ITO/PDINN or PDINN:VC/Ag structures. The current–voltage (I–V) characteristic curves of the films are shown in Fig. S3b.† The PDINN film exhibits a conductivity of 2.82 × 10⁻⁴ S cm⁻¹ and the introduction of VC increases the conductivity of the PDINN:VC film to 4.36 × 10⁻⁴ S cm⁻¹.

2.2 Photovoltaic performance of opaque devices

To reveal the effects of VC on photovoltaic performance, conventional opaque devices (Fig. 3a) were fabricated. Fig. 3b and c show the chemical structures and absorption spectra of PM6 and BTP-eC9, respectively. The current density–voltage (J–V) curves of the devices based on ITO/PEDOT:PSS/PM6:BTP-eC9/CIL/Ag structures at different VC contents are shown in Fig. 3d. The detailed photovoltaic performance parameters are shown in Table S2.† The PDINN-based device shows a PCE of 17.2% with an open circuit voltage (V_OC) of 0.849 V, a short circuit current density (J_SC) of 26.6 mA cm⁻² and a fill factor (FF) of 76.3%. As the VC content increases, the PCE of the devices increases. When the mass fraction of VC is increased to 10%, the PDINN:VC-based device exhibits the highest PCE of 17.8% (V_OC of 0.849 V, J_SC of 27.3 mA cm⁻² and FF of 76.7%). Upon further increasing the VC content to 60%, the PCE of the PDINN:VC-based device reduces to 15.1%, which is mainly due to the significant decrease in V_OC (0.821 V) and FF (70.2%). SEM reveals that a lot of particles appear on the film surface as the VC content increased to 60% (Fig. S4†). This rough surface morphology of the CIL can lead to poor electrical contact with the metal cathode, which induces more surface traps and thus severely affects the PCE. The external quantum efficiency (EQE) spectra of the best devices based on PDINN and PDINN:VC (10%) are shown in Fig. 3e. The integrated photocurrent density (J^cal_SC) of the PDINN-based device is 24.9 mA cm⁻², which is lower than that of the PDINN:VC-based device (25.5 mA cm⁻²), resembling the same trend in J–V measurements. To assess the impact of VC on device stability, maximum power point (MPP) tracking tests were conducted under continuous illumination of a white LED sunlight simulator. The PDINN:VC based devices demonstrated good stability, with a 2% attenuation after 4500 min, which is similar to the trend of changes in PDINN based devices (Fig. S5†).

Fig. 3 (a) Schematic structure of the opaque device. (b) Chemical structures and (c) absorption spectra of PM6 and BTP-eC9. (d) J–V curves of the devices with different VC contents. (e) EQE spectra of optimal devices based on the PDINN and PDINN:VC (10%). (f) J_ph–V_eff curves, (g) V_OC–P_light curves, and (h) J_SC–P_light curves.

In addition, the PDINN:VC CIL exhibits excellent universality in multiple active layer systems, including PM6:L8-BO and D18:L8-BO (Fig. S6†). Devices with the structure of ITO/2PACz (Fig. S6†)/active layer/PDINN or PDINN:VC (10%)/Ag are prepared. The J–V curves and EQE curves of the optimal devices based on PM6:L8-BO are shown in Fig. S7.† The PDINN-based device shows a PCE of 18.2% (V_OC of 0.872 V, J_SC of 26.7 mA cm⁻² and FF of 78.2%). In comparison, the PDINN:VC-based device exhibits a higher PCE of 18.9% with a V_OC of 0.874 V, a J_SC of 27.4 mA cm⁻² and an FF of 79.0%. More importantly, the PDINN:VC-based device achieves a higher PCE of 19.8% in the D18:L8-BO system than in the PDINN-based counterpart (19.0%), with an improved J_SC of 27.2 mA cm⁻² and an FF of 80.1% (Fig. S8 and Table S3†).

2.3 Charge carrier dynamics

Insights into the reasons for the improved photovoltaic performance of the PDINN:VC-based device were investigated by characterizing the charge extraction and recombination of the devices. In order to investigate the effect of the introduction of VC on the charge extraction, the photocurrent density (J_ph) under different effective voltages (V_eff) was measured (Fig. 3f). Assuming the same exciton separation and charge transport in the same active layer under the same V_eff, the ratio of J_ph/J_sat indicates the charge extraction efficiency. Compared with the PDINN-based device, the PDINN:VC-based device exhibits higher J_ph/J_sat values, which suggests that the incorporation of VC can promote the charge extraction. The PDINN:VC-based device exhibits better charge extraction, which is conducive for reducing the charge accumulation at the cathode interface and suppressing charge recombination.

The charge carrier recombination behaviour is evaluated by analysing the relationships between V_OC or J_SC and light intensity (P_light).^53,54 The relationship between V_OC and P_light follows the equation of V_OC ∝ nkT/q ln(P_light). As shown in Fig. 3g, the fitted n values of devices based on PDINN or PDINN:VC are 1.18 or 1.10, respectively, suggesting VC can hinder the trap-assisted recombination. The relationship between J_SC and P_light can be expressed using the J_SC ∝ P_light^α equation, where the exponential factor α is used to characterize the degree of bimolecular recombination. Compared with the PDINN-based device (α = 0.961), the PDINN:VC-based device shows a slightly higher α of 0.964 (Fig. 3h), suggesting that VC can suppress bimolecular recombination to some extent.

To gain a deeper understanding of the charge recombination, we used spectroscopy to quantitatively analyse the recombination process in devices, following the approach reported in our previous studies.^24,41 The capacitance, built-in potential (V_bi) and charge carrier densities of devices are shown in Fig. S9,† based on which the experimental recombination current densities (J_rec,_exp.) are fitted (Fig. 4a–d). The Langevin prefactor (ξ), bulk trap density (N_t,bulk), and surface trap density (N_t,surface) are used as fitting parameters to quantify bimolecular recombination, bulk trap-assisted recombination, and surface trap-assisted recombination,^55,56 respectively. The N_t,bulk values of the PDINN and PDINN:VC-based devices are 2.10 × 10¹⁶ cm⁻³ and 1.85 × 10¹⁶ cm⁻³, respectively (Fig. 4c), indicating the similar bulk trap assisted recombination owing to the use of the same active layer. The N_t,surface value of the PDINN-based device is 1.27 × 10¹⁰ cm⁻², which is higher than that of the PDINN:VC-based device (3.88 × 10⁹ cm⁻²) (Fig. 4d), suggesting that the incorporation of VC reduces the surface trap assisted recombination, which is conducive to obtain higher FF and J_SC of the devices.^41,55

Fig. 4 Fitting results of recombination current density for the (a) PDINN-based device and (b) PDINN:VC-based device. (c) Bulk trap density and (d) surface trap density for the corresponding devices.

2.4 Semitransparent devices

Based on the above optimized PDINN:VC-based devices, STOPVs with the structure of ITO/PEDOT:PSS/PM6:BTP-eC9 (10 : 12, w/w)/PDINN:VC/Ag are prepared. The J–V curves and transmittance spectra of STOPVs based on different Ag electrode thicknesses are shown in Fig. S10.† The detailed photovoltaic performance parameters of STOPVs are shown in Table S4.† When the Ag electrode thickness is 25 nm, the STOPV device exhibits excellent photovoltaic performance with a PCE of 15.8% and a low AVT of 10.3%, thus resulting in a low LUE of 1.63%. When the thickness of the Ag electrode is 8 nm, the STOPV device achieves an AVT of 26.8%. However, the high sheet resistance, arising from the Ag electrode thickness below its permeation threshold, results in a significant decrease in the photovoltaic parameters, showing an FF of only 71.7% and a PCE of 11.4%. In contrast, STOPVs with Ag electrode thicknesses between 10 and 20 nm exhibit superior comprehensive performance.

Adjustment of the D : A ratio is another useful approach for achieving high-performance STOPVs, via balancing the invisible light harvesting and visible light transmission. The absorption range of the donor PM6 between 500 and 700 nm (Fig. 3c) overlaps with the photon response range of the human eye, resulting in a sharp decrease in AVT. Considering that the absorption in the visible region is mainly caused by the donor, it would be beneficial for the STOPVs to achieve a higher LUE if the content of the donor in the active layer can be reduced while maintaining efficient photon to electron conversion. Fig. S11a† shows the normalized absorption spectra of PM6:BTP-eC9 films with different D : A ratios. The absorption of the active layer in visible light decreases significantly as the donor content decreases. And the AVT of the PM6:BTP-eC9 films increases from 40.3% to 70.0% when the D : A ratio decreases from 10 : 12 to 4 : 12 (Fig. S11b†). The effects of the D : A ratio on the LUE are further studied by comparing the device performance of STOPVs with Ag electrode thicknesses between 10 and 20 nm (Fig. S12†). The PCE and AVT variations of STOPVs with different D : A ratios are presented in Fig. S13a.† The detailed photovoltaic parameters are shown in Table S5.† The STOPV device with D : A = 6 : 12 exhibits the best LUE of 3.98%, which outcompetes the others (LUE of 3.28%, 3.84% and 3.36% at the D : A ratio of 10 : 12, 8 : 12 and 4 : 12, respectively) at the Ag electrode thickness of 10 nm (Fig. S13b†). Similarly, STOPVs with D : A = 6 : 12 exhibit the best LUE at other Ag electrode thicknesses. Therefore, D : A = 6 : 12 is selected for further optical optimization.

Optical dissipation loss at the Ag electrode interface is an important reason for the low AVT of the device, which can be solved by evaporating a simple optical layer with a high refractive index. Molybdenum trioxide (MoO₃) has a high refractive index value and is deposited on top of the ultrathin Ag electrode as an transmittance-enhancing optical layer, which suppresses the light dissipation at the metal surface and contributes to the improvement of AVT.⁵⁷ Based on the above optimized D : A ratio, STOPVs with the structure of ITO/PEDOT:PSS/PM6:BTP-eC9 (6 : 12, w/w)/PDINN:VC/Ag/MoO₃ are prepared (Fig. 5a). The thickness of transparent ultrathin Ag for STOPVs is optimized with a fixed MoO₃ thickness of 35 nm, by considering the double effects on light absorption and transmission. The J–V curves and EQE and transmission spectra of the devices with varied Ag thicknesses are presented in Fig. 5b–d. The detailed photovoltaic parameters are summarized in Table 1. With the Ag thicknesses increasing from 10, 15 to 20 nm, the J_SC of STOPVs increases from 15.5, 17.7, to 19.6 mA cm⁻², respectively, while the AVT decreases from 45.3%, 40.6%, to 37.4%, respectively. This result suggests that the thickness of Ag is an important factor to balance the light absorption and visible transmission. As a result, the STOPV device with an electrode thickness of 20 nm shows the best LUE of 4.53%. The comparison between our results and STOPVs reported in the literature is summarized in Table S6.†Fig. 5e–h show the photos of plants and STOPVs with different Ag electrode thicknesses.

Fig. 5 (a) Schematic structure of STOPVs. (b) J–V curves and (c) EQE and (d) transmittance spectra of STOPVs with different Ag electrode thicknesses. (e–h) Photos of plants and STOPVs with different Ag electrode thicknesses.

Table 1 Detailed photovoltaic performance parameters of STOPVs with PM6:BTP-eC9 (6 : 12, w/w) as the active layer and 35 nm MoO₃ as the optical structure

Electrode thickness	V OC (V)	J SC (mA cm^−2)	FF (%)	PCE (%)	AVT^a (%)	LUE (%)
a AVT is obtained by calculating the arithmetic mean of the transmittance between 400 and 700 nm.
20 nm	0.833	19.6	74.3	12.1	37.4	4.53
15 nm	0.831	17.7	74.1	10.9	40.6	4.43
10 nm	0.826	15.5	72.7	9.3	45.3	4.21

---

3 Conclusions

In summary, a novel CIL modification method utilizing vitamin C is proposed to suppress the electrical losses of devices. The addition of VC to PDINN not only improves the conductivity, but also results in flatter film morphology of the CIL. VC modulation enables the establishment of a better contact between the active layer and the metal cathode, which can suppress charge carrier recombination and thus improve the PCE of the device. The optimal PDINN:VC-based device exhibits a PCE of up to 19.8%, which is significantly higher than that of the PDINN-based device (19.0%). STOPVs are prepared based on the above optimized CIL, accompanied by the optimization of the electrode thickness, D : A ratio, and optical structure. Ultimately, the optimized STOPV device achieves a high LUE of up to 4.53% with a PCE of 12.1% and an AVT of 37.4%.

Data availability

The data supporting this article have been included in the manuscript and ESI.†

Author contributions

H. Y., J. W. and P. C. conceived the idea and designed the experiments. H. Y. and Y. H. fabricated the devices and conducted related measurements. J. W. conducted molecular modelling. H. Y. and J. W. wrote the paper. C. Y. and Q. R. contributed to data analysis. J. W. and P. C. supervised the project. All authors discussed the results and commented on the manuscript.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgements

The authors are thankful for the financial support from the National Natural Science Foundation of China (Grant No. 52403331 and 52403239) and National Key Laboratory of Advanced Polymer Materials (Grant No. sklpme2024-2-15). The Hefei Advanced Computing Center is acknowledged for molecular modelling.

Notes and references

1. X. Li, X. Kong, G. Sun and Y. Li, eScience, 2023, 3, 100171.
2. J. Wang, P. Xue, Y. Jiang, Y. Huo and X. Zhan, Nat. Rev. Chem., 2022, 6, 614–634.
3. L. Zhu, M. Zhang, G. Zhou, Z. Wang, W. Zhong, J. Zhuang, Z. Zhou, X. Gao, L. Kan, B. Hao, F. Han, R. Zeng, X. Xue, S. Xu, H. Jing, B. Xiao, H. Zhu, Y. Zhang and F. Liu, Joule, 2024, 8, 3153–3168.
4. H. Zhang, Y. Liu, G. Ran, H. Li, W. Zhang, P. Cheng and Z. Bo, Adv. Mater., 2024, 36, 2400521.
5. S. Yang, Z. Chen, J. Zhu, D. Yang, H. Wang, P. Ding, J. Wu, P. Yan, L. Xie, F. Chen, Y. Wang, J. Zhang and Z. Ge, Adv. Mater., 2024, 36, 2401789.
6. J. Wang, Y. Xie, K. Chen, H. Wu, J. M. Hodgkiss and X. Zhan, Nat. Rev. Phys., 2024, 6, 365–381.
7. Y. Zhang, J. Wang, Y. Mo, H. Li, F. Zhao, H. Sun, W. Kong, C. Yan, H. Wang, Z. Ni and P. Cheng, Chem. Commun., 2025, 61, 3343–3346.
8. Q. Chen, Y. H. Han, L. R. Franco, C. F. N. Marchiori, Z. Genene, C. M. Araujo, J.-W. Lee, T. N.-L. Phan, J. Wu, D. Yu, D. J. Kim, T.-S. Kim, L. Hou, B. J. Kim and E. Wang, Nano-Micro Lett., 2022, 14, 164.
9. C. Lin, R. Peng, J. Shi and Z. Ge, Exploration, 2024, 4, 20230122.
10. Y. Wang, C. Gao, W. Lei, T. Yang, Z. Liang, K. Sun, C. Zhao, L. Chen, L. Zhu, H. Zeng, X. Sun, B. He, H. Hu, Z. Tang, M. Qiu, S. Li, P. Han and G. Zhang, Nano-Micro Lett., 2025, 17, 206.
11. N. Gao, H. Bin, S. Wu, R. A. J. Janssen and Y. Li, Sci. China:Chem., 2024, 67, 2248–2256.
12. Y. Li, X. Huang, H. K. M. Sheriff and S. R. Forrest, Nat. Rev. Mater., 2023, 8, 186–201.
13. W. Kong, J. Wang, Y. Hu, N. Cui, C. Yan, X. Cai and P. Cheng, Angew. Chem., Int. Ed., 2023, 62, e202307622.
14. C. J. Traverse, R. Pandey, M. C. Barr and R. R. Lunt, Nat. Energy, 2017, 2, 849–860.
15. D. Wang, Y. Li, G. Zhou, E. Gu, R. Xia, B. Yan, J. Yao, H. Zhu, X. Lu, H.-L. Yip, H. Chen and C.-Z. Li, Energy Environ. Sci., 2022, 15, 2629–2637.
16. Z. Hu, J. Wang, X. Ma, J. Gao, C. Xu, K. Yang, Z. Wang, J. Zhang and F. Zhang, Nano Energy, 2020, 78, 105376.
17. S. Liu, H. Li, X. Wu, D. Chen, L. Zhang, X. Meng, L. Tan, X. Hu and Y. Chen, Adv. Mater., 2022, 34, 2201604.
18. C. Xu, Y. Chen, Z. Zhao, B. Yang, J. Suo, K. Ba, A. Tarasov, Y. Wu, X. Lian, M. Luo, Y. Zhan, Y. Chen, J. Gao, X. Mo, W. C. H. Choy, J. Wang, H. Zhang and J. Chu, Energy Environ. Sci., 2025, 18, 2095–2135.
19. H. W. Cheng, Y. Zhao and Y. Yang, Adv. Energy Mater., 2021, 12, 2102908.
20. X. Yan, J. Wang, W. He, T. A. Dela Peña, C. Zhu, H. Yu, Y. Hu, C. Yan, S. Ren, X. Chen, Z. Wang, J. Wu, M. Li, J. Xia, L. Meng, S. Lu, D. Zhao, M. Artemyev, Y. Li and P. Cheng, J. Energy Chem., 2024, 96, 351–358.
21. S. Sun, W. Zha, C. Tian, Z. Wei, Q. Luo, C. Q. Ma, W. Liu and X. Zhu, Adv. Mater., 2023, 35, 2305092.
22. X. Duan, Y. Yang, J. Yu, C. Liu, X. Li, M. H. Jee, J. Gao, L. Chen, Z. Tang, H. Y. Woo, G. Lu and Y. Sun, Adv. Mater., 2024, 36, 2308750.
23. X. Xu, Q. Wei, Z. Zhou, H. He, J. Tian, H. L. Yip, Y. Fu, X. Lu, Y. Zhou, Y. Li and Y. Zou, Adv. Funct. Mater., 2023, 34, 2305017.
24. Y. Li, J. Wang, C. Yan, S. Zhang, N. Cui, Y. Liu, G. Li and P. Cheng, Joule, 2024, 8, 527–541.
25. H. Yu, J. Wang, Q. Zhou, J. Qin, Y. Wang, X. Lu and P. Cheng, Chem. Soc. Rev., 2023, 52, 4132–4148.
26. J. Ding, H. Mou, H. Chen, J. Xu, W. Sun, J. Zhu, Y. Wang, Y. Huang, Y. Li and Y. Li, Adv. Mater., 2025, 37, 2420439.
27. H. Tang, Y. Bai, H. Zhao, X. Qin, Z. Hu, C. Zhou, F. Huang and Y. Cao, Adv. Mater., 2023, 36, 2212236.
28. L. Xiao, H. Zhang, S. Li, Q. Cheng, Y. Zhang, Y. Li, Y. Zhang and H. Zhou, Sol. RRL, 2024, 8, 2400126.
29. Z. You, J. Wen, W. Liu, Z. Fink, X. Wu, H.-G. Seong, Y. Wang, L. Zhang, X. Wang, T. P. Russell and Y. Liu, Adv. Mater., 2025, 37, 2500450.
30. C. Zhu, X. Wang, W. Liu, Y. Liu and X. Zhan, Trends Chem., 2024, 6, 37–50.
31. Y. Yu, J. Wang, Y. Cui, Z. Chen, T. Zhang, Y. Xiao, W. Wang, J. Wang, X.-T. Hao and J. Hou, J. Am. Chem. Soc., 2024, 146, 8697–8705.
32. Z. He, C. Zhong, S. Su, M. Xu, H. Wu and Y. Cao, Nat. Photonics, 2012, 6, 591–595.
33. J. Yao, B. Qiu, Z.-G. Zhang, L. Xue, R. Wang, C. Zhang, S. Chen, Q. Zhou, C. Sun, C. Yang, M. Xiao, L. Meng and Y. Li, Nat. Commun., 2020, 11, 2726.
34. Y. Wang, J. Wen, Z. Shang, Y. Zhong, H. Zhang, W. Liu, W. Han, H. Yang, J. Liu, J. Zhang, H. Li and Y. Liu, Angew. Chem., Int. Ed., 2025, e202506252.
35. Y. Zheng, J. Zhao, S. Yang and H. Yu, Adv. Opt. Mater., 2025, 2500569.
36. W. Zhang, X. Ding, J. Lv, X. Sun, D. Hu, G. Zhang, C. Gao, Y. Xue, Y. Zhong, G. Li and H. Hu, Adv. Funct. Mater., 2025, 2420588.
37. H. Yu, J. Wang, Y. Gong, C. Yan, Q. Ran, L. Meng, Y. Li and P. Cheng, Adv. Funct. Mater., 2025, 2504816.
38. Q. Zhang, T. Liu, S. Wilken, S. Xiong, H. Zhang, I. Ribca, M. Liao, X. Liu, R. Kroon, S. Fabiano, F. Gao, M. Lawoko, Q. Bao, R. Österbacka, M. Johansson and M. Fahlman, Adv. Mater., 2024, 36, 2307646.
39. B. Liu, O. J. Sandberg, J. Qin, Y. Liu, S. Wilken, N. Wu, X. Yu, J. Fang, Z. Li, R. Huang, W. Zha, Q. Luo, H. Tan, R. Österbacka and C.-Q. Ma, Nat. Photonics, 2025, 19, 195–203.
40. X. Ding, J. Lv, Z. Liang, X. Sun, J. Zhao, M. Lu, F. Wang, C. Zhang, G. Zhang, T. Xu, D. Hu, Z. Kan, C. Ruan, Y. Shi, H. Lin, W. Zhang, G. Li and H. Hu, Adv. Energy Mater., 2024, 14, 2401741.
41. H. Yu, J. Wang, Y. Li, T. Liu, Y. Gong, C. Yan, Q. Ran, Q. Luo, Z. Mao, X. Wei, L. Meng, C. Ma, Y. Li and P. Cheng, Chem. Eng. J., 2024, 499, 156475.
42. Z. Wang, J. Guo, Y. Pan, J. Fang, C. Gong, L. Mo, Q. Luo, J. Lin and C. Ma, Energy Environ. Mater., 2024, 7, e12592.
43. S. Guan, Y. Li, K. Yan, W. Fu, L. Zuo and H. Chen, Adv. Mater., 2022, 34, 2205844.
44. Y. F. Zhang, W. S. Chen, J. D. Chen, H. Ren, H. Y. Hou, S. Tian, H. R. Ge, H. H. Ling, J. L. Zhang, H. Mao, J. X. Tang and Y. Q. Li, Adv. Energy Mater., 2024, 14, 2400970.
45. T. Xu, B. Deng, K. Zheng, H. Li, Z. Wang, Y. Zhong, C. Zhang, G. Lévêque, B. Grandidier, R. Bachelot, M. Treguer-Delapierre, Y. Qi and S. Wang, Adv. Mater., 2024, 36, 2311305.
46. C. Yang, C. Zhao, Y. Sun, Q. Li, M. R. Islam, K. Liu, Z. Wang, S. Qu and Z. Wang, Carbon Energy, 2021, 3, 4–23.
47. J. Wen, R. Lin, Y. Wu, H. Zhang, H. Zhou, Z. Liu, Y. Xie, D. Ye and X. Ouyang, EcoMat, 2023, 5, e12318.
48. L. Li, J. Xu, W. Luo, K. Zhong, X. Zhao, Y. Hu, Z. Yuan and Y. Chen, J. Mater. Chem. A, 2023, 11, 24136–24145.
49. F. Neese, Wiley Interdiscip. Rev.:Comput. Mol. Sci., 2022, 12, e1606.
50. F. Neese, Wiley Interdiscip. Rev.:Comput. Mol. Sci., 2012, 2, 73–78.
51. M. J. Abraham, T. Murtola, R. Schulz, S. Páll, J. C. Smith, B. Hess and E. Lindahl, SoftwareX, 2015, 1–2, 19–25.
52. W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics, 1996, 14, 33–38.
53. L. J. A. Koster, V. D. Mihailetchi, R. Ramaker and P. W. M. Blom, Appl. Phys. Lett., 2005, 86, 123509.
54. C. M. Proctor and T.-Q. Nguyen, Appl. Phys. Lett., 2015, 106, 083301.
55. J. Vollbrecht, V. V. Brus, S. J. Ko, J. Lee, A. Karki, D. X. Cao, K. Cho, G. C. Bazan and T. Q. Nguyen, Adv. Energy Mater., 2019, 9, 1901438.
56. V. V. Brus, C. M. Proctor, N. A. Ran and T. Q. Nguyen, Adv. Energy Mater., 2016, 6, 1502250.
57. X. Huang, X. Ren, Y. Cheng, Y. Zhang, Z. Sun, S. Yang, S. Kim, C. Yang, F. Wu and L. Chen, Energy Environ. Sci., 2024, 17, 2825–2836.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03837e

---