Every ray of sunlight that strikes a surface carries heat, and for decades engineers have tried to convert that heat into useful electricity. A team of researchers in China has now reported a deceptively simple trick that dramatically improves one such conversion route: instead of only heating one side of a device with sunlight, they let the other side cool itself by radiating heat straight into the cold of outer space. The combination, described in npj Flexible Electronics, more than triples the electrical output of a flexible thermogalvanic cell and points toward self-powered sensors, smart windows, and wearable energy harvesters that need no batteries at all.

The device at the heart of the study is a thermogalvanic cell, or TGC, a class of electrochemical system that generates voltage from a temperature difference. Unlike solid-state thermoelectric generators, which rely on the Seebeck effect in semiconductors, thermogalvanic cells exploit the temperature dependence of a redox couple dissolved in an electrolyte. When one electrode is warmer than the other, the redox reaction equilibria shift unevenly, producing a measurable potential difference across the cell. Because the electrolyte can be gelled into a quasi-solid film, TGCs can be made thin, bendable, and lightweight, which makes them attractive for low-grade heat harvesting and for wearable or building-integrated applications where rigid modules are impractical.

The catch has always been the temperature gradient. The voltage a thermogalvanic cell produces scales with the difference in temperature, ΔT, between its two electrodes, and under direct sunlight that difference tends to collapse. Sunlight warms the whole device fairly uniformly, so both sides heat up together and the gradient that should drive the electrochemical potential all but disappears. Earlier solar-driven designs relied purely on photothermal heating, depositing a light-absorbing layer on one electrode to make it hot. That helps, but the heat inevitably spreads through the thin, flexible structure to the cold side, eroding the very gradient the cell needs. The result has been modest voltages and limited power, holding back the idea of harvesting sunlight with TGCs.

The Donghua University-led team, working with a colleague at Ruijin Hospital of Shanghai Jiao Tong University School of Medicine, attacked the problem from both ends at once. Their cell is built on a bacterial cellulose substrate, a flexible, porous biopolymer film that hosts the quasi-solid electrolyte. On one side they coated a layer of carbon nanotubes, which act as an efficient photothermal converter, absorbing sunlight and turning it into localized heat at the hot electrode. On the opposite side they coated titanium dioxide, a material chosen for its radiative cooling properties: it emits thermal radiation strongly in the mid-infrared atmospheric transparency window, allowing that side to shed heat passively to the sky even while sitting in full sunlight.

This asymmetric coating is what makes the design clever. Radiative cooling is a passive phenomenon with no moving parts and no energy input; a surface engineered to emit in the 8-to-13-micrometer wavelength band can dump heat through the atmosphere into the roughly three-kelvin cosmic background, dropping below ambient temperature on a clear day. By pairing a photothermal hot side with a radiatively cooled cold side, the researchers created a device that manages its own thermal asymmetry under illumination. The sunlight simultaneously acts as a heat source on one electrode and, through the cooling layer’s transparency to the sky, is prevented from spoiling the cold electrode on the other. The team describes this as using sunlight not only as a heat source but also as a trigger for asymmetric thermal management.

The measured results are striking. Under one sun of illumination, the standard intensity of noonday sunlight, the optimized device sustained a temperature difference of 14.11 kelvin between its electrodes, and it generated an open-circuit voltage of 18.07 millivolts, substantially exceeding an otherwise identical cell that had only the photothermal heating layer. The parasitic heating that normally bleeds warmth across the device was effectively suppressed, and the cold side stayed cold. The maximum power density reached 13.98 milliwatts per square meter, a 214 percent enhancement over the photothermal-only counterpart, meaning the output more than tripled simply by adding the radiative cooling layer and optimizing the structure.

Those numbers may sound small in absolute terms, and they are, but the context matters. Thermogalvanic cells are low-power devices by nature, and millivolts of thermopower from a thin flexible film is respectable for the field. The real significance of the 214 percent gain is architectural: it shows that thermal management, not electrochemistry, was the bottleneck for solar-driven TGCs, and that the bottleneck can be broken with a passive, coating-level solution. No lenses, trackers, pumps, or active cooling were required. The gradient arises spontaneously the moment sunlight hits the device, which is exactly what a self-powered, maintenance-free harvester needs.

The team also moved beyond bench-top measurements. In a practical outdoor demonstration, the flexible cell operated under real sky and real sunlight, supporting the claim that the photothermal-plus-radiative-cooling scheme works outside the controlled conditions of a solar simulator. As a proof of concept for functionality beyond raw power generation, the researchers coupled the device to a voltage amplifier and used it as an intelligent light-control switch for smart windows: the illumination level determines the voltage the cell produces, and that signal can drive a responsive switching action. In that configuration the same film serves simultaneously as an energy harvester, a light sensor, and a control element, hinting at a platform that integrates energy harvesting, environmental sensing, and responsive control in a single flexible laminate.

The materials choices also carry practical weight. Bacterial cellulose is inexpensive, biodegradable, and mechanically robust, and its use as the TGC substrate aligns with growing interest in sustainable flexible electronics. Carbon nanotubes are among the most broadband solar absorbers known, converting incident photons to heat with high efficiency, while titanium dioxide is cheap, stable, non-toxic, and widely used in coatings and pigments, which eases scaling. Because both functional layers are coatings on opposite faces of the same film, the fabrication route is compatible with roll-to-roll and textile-style processing, a natural fit for the dyeing and finishing expertise of the Donghua group and for applications such as smart fabrics, building skins, and conformable sensors.

There are, of course, caveats before such devices reach everyday use. Milliwatts per square meter will not power a phone, so near-term applications lie in ultra-low-power electronics: intermittent sensors, electronic skin, status indicators, and switching functions of the kind demonstrated. Performance will depend on sky conditions, since radiative cooling weakens under clouds and humidity, and on the long-term stability of the quasi-solid electrolyte and coatings under weathering. Still, the study establishes a clear design principle with broad reach: in any solar-thermal conversion device, the cold side deserves as much engineering attention as the hot side. By letting one face drink in the sun and the other face whisper its heat into space, the researchers turned a persistent weakness of thermogalvanic cells into their strongest asset, and showed that the sky itself can be recruited as a heat sink for flexible power generation.

Subject of Research: Solar-driven thermogalvanic cells enhanced by combined photothermal heating and passive radiative cooling

Article Title: Thermopower amplification in thermogalvanic cells via radiative cooling-mediated photothermal conversion

Article References: Geng, Y., Li, T., Fan, G., Zhong, R., Rong, L., Wang, B., Zhong, Y., Mao, Z., Xie, S., & Feng, X. (2026). Thermopower amplification in thermogalvanic cells via radiative cooling-mediated photothermal conversion. npj Flexible Electronics. https://doi.org/10.1038/s41528-026-00646-1

Image Credits: AI Generated

DOI: 10.1038/s41528-026-00646-1

Keywords: thermogalvanic cell, radiative cooling, photothermal conversion, flexible electronics, low-grade heat harvesting, carbon nanotubes, titanium dioxide, bacterial cellulose, thermoelectric conversion, smart windows, self-powered devices, temperature gradient