A joint research team from Chiba University's Research Center for Advanced Science and Technology (led by Project Professor Hirohiko Fukagawa), NHK Science & Technology Research Laboratories, Kyoto University, and others has identified the energy condition that lets organic thin-film solar cells (OPVs) maintain photocurrent while suppressing voltage loss. The team fabricated about 40 devices with different donor-acceptor combinations, measured their energy levels in the solid state, and analyzed the results. The paper was published online in the journal Advanced Materials on October 7, 2026. In OPVs, a larger energy offset between materials makes it easier to generate current but lowers the voltage that can be extracted. According to Chiba University's announcement, the condition under which the donor's optical gap roughly matches the energy of the free-charge state, in which electrons and holes are fully separated, serves as a "practical thermodynamic boundary" for maintaining photocurrent. The result does not set a new conversion-efficiency record. Rather, it offers a design guideline for reducing voltage loss without sacrificing current, and its applicability to practical blended devices has not yet been verified.
Three Steps from Light to Current, and Where Voltage Loss Arises
According to Chiba University, the power conversion efficiency of OPVs is already approaching 20%. However, "voltage loss," in which part of the light's energy cannot be extracted as voltage, remains a major obstacle to higher efficiency. To understand why it occurs, it helps to follow the process by which absorbed light becomes current.
When an OPV absorbs light, it first generates an "exciton," an electron and a hole (the vacancy left by an electron) bound together by mutual attraction. When the exciton reaches the interface between the electron-donating donor material and the electron-accepting acceptor material, the electron moves to the acceptor side, forming a "charge-transfer (CT) state." At this stage the electron and hole sit in separate materials but are still bound by electrostatic attraction. When they separate far enough to become "free charges," each moves to an electrode and is extracted as current.
Traditionally, the key factor thought to govern charge generation was the energy difference in going from the exciton to the CT state. This difference is called the "driving force." The larger it is, the more favorable charge generation becomes, but the lower the voltage that can be extracted. Conversely, a smaller offset reduces voltage loss but risks decreasing photocurrent. Views have differed, depending on the material combination, on how much energy difference efficient charge generation actually requires.
The team examined this problem by looking not only at CT-state formation but also at the generation of free charges beyond it. In Chiba University's English-language announcement, Fukagawa explained that the team focused on whether electrons and holes can reach a fully separated free-charge state, not just whether a CT state forms. In other words, instead of the CT-state energy that had been emphasized, they used the energy of the ultimately generated free charges as the reference for understanding the relationship between photocurrent and voltage loss.
About 40 Planar-Junction Devices Reveal Where Photocurrent Drops
High-efficiency OPVs commonly use a "bulk heterojunction" structure in which donor and acceptor materials are mixed together. In this structure, however, the two materials are intricately interwoven, and factors such as molecular arrangement and interface geometry also affect power generation, making it difficult to isolate the effect of energy levels alone.
The team therefore adopted planar-junction model devices in which donor and acceptor layers are stacked separately. The structure was designed so that visible light is absorbed mainly by the donor layer, and both layers were fixed at a thickness of 30 nm to minimize the effects of differences in light absorption and film morphology. Because the interface is well defined, this structure is suitable for investigating the relationship between energy levels and charge generation. Planar-junction organic solar cells date back to the bilayer device reported by C. W. Tang in Applied Physics Letters in 1986, which achieved a conversion efficiency of about 1% under simulated AM2 sunlight.
For the experiments, the team prepared five donor materials with differing molecular-backbone rigidity and 15 acceptor materials with electron affinities in the range of 2.0 to 3.1 eV. Electron affinity is the energy change when a material accepts an electron, and serves as an indicator of how readily it takes one up. Combining these materials, the team fabricated and compared about 40 OPV devices.
For evaluating energy levels, the team emphasized measurements in the solid state, the same as in actual devices. The ionization energy of the donor materials (the energy needed to remove one electron) was measured by ultraviolet photoelectron spectroscopy (UPS), and the electron affinity of the acceptors was determined by low-energy inverse photoelectron spectroscopy (LEIPS). The energy of the free-charge state was then calculated from the two measured values. Because typical electrochemical measurements in solution cannot adequately reflect polarization and interfacial environments inside a solid, directly measuring solid-state levels allowed comparison under conditions closer to real devices.
The results showed a clear relationship between photocurrent and energy difference. When the donor's optical gap (a measure of the minimum energy needed for light absorption) was larger than the free-charge energy, relatively large photocurrents were obtained. When the free-charge energy exceeded the optical gap, photocurrent tended to decline. The boundary lay where the difference between the two was roughly 0 eV, and this change was especially clear for four donors with rigid molecular backbones (ν-DABNA, PXZ-BN, DABNA-1, and DABNA-2).
Chiba University positions this condition as a "practical thermodynamic boundary" for maintaining photocurrent. It does not presuppose a specific pathway for CT-state generation; rather, it is a criterion for whether the energy of absorbed light can carry electrons and holes all the way to a fully separated free-charge state. Power conversion efficiency also tended to be higher near this boundary, but the university's announcement only describes the trend and gives no specific efficiency figures. What matters is that the team identified, from measured energy levels, a condition under which photocurrent can be maintained without a large energy offset.
Same 2.60 eV Free-Charge Energy, Different Voltage Depending on Molecular Rigidity
The team also examined how molecular-backbone rigidity affects the CT-state energy. In Chiba University's announcement, devices combining the rigid donor "ν-DABNA" and the flexible donor "HN-D2," which have the same ionization energy, with the same acceptor "A-6" are compared. Both have a free-charge energy of 2.60 eV, but their CT-state energies differ.
The measure used to evaluate this difference is the "effective CT binding energy." It is the free-charge energy minus the effective CT-state energy determined from the emission spectrum. It indicates how far the CT state is stabilized below the free-charge state; a smaller value means the CT-state energy is closer to that of the free-charge state.
Comparing the measurement results for the two devices shown in Chiba University's release materials (Figure 3) gives the following:
| Item | ν-DABNA (rigid) + A-6 | HN-D2 (flexible) + A-6 |
|---|---|---|
| Free-charge energy | 2.60 eV | 2.60 eV |
| Effective CT-state energy | 2.52 eV | 2.42 eV |
| Effective CT binding energy | 0.08 eV | 0.18 eV |
| Open-circuit voltage | 1.99 V | 1.94 V |
Although HN-D2 has the larger optical gap, its open-circuit voltage was 0.05 V lower than that of the ν-DABNA device. With the flexible HN-D2, the CT state was stabilized to an energy 0.18 eV below the free-charge state, whereas with the rigid ν-DABNA the difference was only 0.08 eV. In this comparison, then, ν-DABNA, whose CT state is not excessively stabilized, extracted a higher voltage from the same free-charge energy. Note, however, that the 1.99 V figure is the open-circuit voltage of a device using a donor with a large optical gap and does not imply high conversion efficiency as a solar cell.
Across the whole set of materials, rigid donors were associated with less relaxation of the CT state and less broadening of its energy distribution. Indeed, several combinations using rigid donors yielded small effective CT binding energies of 0.1 eV or less. Chiba University says these values approach those reported for metal halide perovskites. However, the effective CT binding energy is not determined by backbone rigidity alone; charge delocalization, molecular packing, and the electrostatic environment at the donor-acceptor interface also play a role. These results alone cannot establish that making a molecule rigid will always lower the binding energy.
The DABNA-type molecules used here were originally developed as light-emitting materials for organic EL displays. In 2016, Professor Takuji Hatakeyama of Kyoto University and colleagues reported blue thermally activated delayed fluorescence (TADF) materials that exploit the multiple-resonance effect in Advanced Materials. At the time, Hatakeyama was at Kwansei Gakuin University; he moved to Kyoto University in 2022. In 2019, they further reported ν-DABNA, with a narrower emission spectrum, in Nature Photonics. Molecules developed to improve display luminous efficiency and color purity have thus been put to use in this study as materials for clarifying the energy conditions of solar cells.
How Does "0.2–0.3 eV" in 2019 Differ from "Near 0 eV" Now?
To understand this result, it is important to see how it differs from research announced in 2019 by RIKEN, Chiba University, and others. In a study published in Nature Communications on June 7, 2019, an international research group including Kyohei Nakano and Keisuke Tajima of RIKEN and Professor Hiroyuki Yoshida of Chiba University examined 16 planar heterojunction devices combining four donors and four acceptors. They estimated that light is converted to current most efficiently when the energy difference between the excited state and the CT state is 0.2–0.3 eV. Chiba University's Japanese announcement also cites this study as a reference for the conventional view.
Comparing the 2019 "0.2–0.3 eV" and the current "near 0 eV" directly makes it look as if the energy difference required for charge generation has shrunk dramatically. However, the reference energy levels differ. The 2019 study evaluated the difference between the optical gap and the CT-state energy, whereas the present study evaluates the difference between the optical gap and the free-charge energy. The conditions and main results of the two studies are summarized below:
| Item | 2019 study (RIKEN, Chiba University, et al.) | 2026 study (Chiba University, NHK, Kyoto University, et al.) |
|---|---|---|
| Device structure | Planar heterojunction | Planar junction (both layers 30 nm thick) |
| Materials and devices | 4 donors, 4 acceptors, 16 devices | 5 donors, 15 acceptors, about 40 devices |
| Energy-level measurement | Measured by LEIPS and other methods | Measured in the solid state by UPS and LEIPS |
| Energy difference compared | Optical gap minus CT state (Eg,opt − ECT) | Optical gap minus free-charge energy (Eg,opt − ECS) |
| Guideline obtained | Efficient charge generation at 0.2–0.3 eV difference | Boundary for maintaining photocurrent at a difference near 0 eV |
| Result on CT binding energy | No clear correlation between ECS − ECT and charge-generation efficiency | Multiple combinations with rigid donors showed effective values of 0.1 eV or less |
| Main quantity evaluated | Charge-generation efficiency | Photocurrent and voltage loss |
Both studies used planar-junction devices with separated donor and acceptor layers and evaluated them on the basis of measured energy levels. LEIPS is a technique developed by Professor Hiroyuki Yoshida of Chiba University, and in the 2019 study too it was used to evaluate the free-charge-state energy with high accuracy. However, the conditions the two studies sought to clarify differ. The 2019 study focused on the energy difference in going from the exciton to the CT state, while the present study examines the energy condition spanning the entire path from exciton to free charge. This is why Fukagawa stresses reaching free charges rather than only forming a CT state.
It would also be inappropriate to interpret the present result as overturning the 2019 conclusion. In the 2019 study, the energy difference between the CT state and the free-charge state (ECS − ECT) showed no clear correlation with charge-generation efficiency. The effective CT binding energy highlighted in the present study is calculated using the effective CT-state energy determined from emission spectra, so it is defined differently from the quantity evaluated in 2019. Moreover, the 2019 study mainly examined charge-generation efficiency, whereas the present study emphasizes the relationship between photocurrent and voltage loss. Because the measured quantities and evaluation metrics differ, the two results cannot simply be treated as contradictory.
Domestic OPV Research Since 2019, and Challenges Remaining for Practical Blended Devices
The paper's references include, besides the 2019 RIKEN–Chiba University study, recent results from research institutions in Japan. One is a Nature Communications paper involving Taku Oono of NHK Science & Technology Research Laboratories and researchers at Tokyo University of Science and elsewhere. Published online on December 7, 2025, and in its final version on January 20, 2026, it reported a method of achieving both light emission and power generation in a single device using organic semiconductors. Chiba University also announced it in January 2026 as an "organic EL device that can generate power."
On March 23, 2026, a joint team from Hiroshima University, Kyoto University, RIKEN, the University of Tsukuba, Toray Research Center, and others published in Communications Materials results on PTNT1-F, a polymer semiconductor with a rigid molecular backbone. The study held voltage loss from "non-radiative recombination," in which electrons and holes recombine without emitting light, to 0.182 V while achieving high charge-generation efficiency. Combined with the acceptor Y12, the comparison materials D18 and PM6 gave 0.214 V and 0.235 V, respectively. A rigid molecular backbone is thought to make charges spread more readily, which is believed to be related to improved charge-separation efficiency.
In April 2026, a team including Associate Professor Seiichiro Izawa and Qingjun Shui of Institute of Science Tokyo reported in Advanced Materials a highly luminescent OPV using multiple-resonance molecules. The study reduced the rate of non-radiative recombination by more than five orders of magnitude compared with conventional state-of-the-art OPVs and obtained a high voltage close to the theoretical limit. According to Institute of Science Tokyo's announcement, a key was combining different light-emitting molecules and adopting an energy structure that suppresses transitions to non-emissive triplet excited states.
These studies share the common challenge of improving the trade-off between charge generation and voltage loss through energy-level tuning and molecular-backbone design. However, because the device structures, materials, and evaluation metrics differ among them, the reported voltages and efficiencies cannot be compared directly to judge which is superior. Fukagawa, who led the present study, moved to Chiba University in 2024 after more than ten years of research at NHK Science & Technology Research Laboratories, and he has remained involved in these lines of work. The study was supported by JST CREST, JSPS KAKENHI grants, and the JKA's Keirin-racing-funded subsidy program.
That said, the thermodynamic boundary revealed here may not apply to all OPVs. The framework presented in the paper targets materials with large optical gaps, and it cannot be determined from this study alone whether the same condition holds for materials with small optical gaps. Moreover, the boundary was identified using model devices with separate donor and acceptor layers and has not been verified in bulk-heterojunction devices, which are the norm in high-efficiency OPVs. Although the study showed that conversion efficiency tends to be higher near the boundary and that molecular rigidity is associated with a lower effective CT binding energy, whether these findings hold across a wide range of materials remains a task for the future.
The team plans to apply this design guideline to practical blended OPVs and test its generality. Fukagawa explains that if the principle clarified here can be incorporated into bulk-heterojunction materials, it could lead to solar cells that combine efficient charge generation with low voltage loss. Finding, from measured energy levels, a boundary at which photocurrent can be maintained offers a clue for making OPV materials design more efficient, but confirming its effectiveness in practical devices will be the next important step.