EPFL, CSEM and their international partners take triple-junction perovskite-silicon cells beyond a useful threshold for industry, space and networks

Next-generation photovoltaics doesn't advance simply through the accumulation of small, incremental improvements. In some stages of its evolution, it proceeds through architectures—that is, new ways of stacking materials, distributing light, managing electrical losses, and making industrially credible solutions that for years have remained confined to the laboratory. It is in this perspective that the results achieved by EPFL e CSEM: A perovskite-silicon triple-junction solar cell has achieved independently certified efficiency of 30,02 percent, exceeding the previous certified benchmark of 27,1 percent.

The data, published in Nature and announced on March 18, 2026, concerns a device composed of a lower silicon cell, on which two thin perovskite cells are deposited. The logic is that of multi-junction cells: not requiring a single material to convert the entire solar spectrum, but assigning a different portion of the radiation to each layer. This allows for more selective use of light and reduced energy losses. The novelty lies not only in having surpassed a symbolic threshold, but in having done so with materials and processes that aim for more accessible manufacturing compared to the III-V cells used primarily in space.

The group of Photovoltaics and Thin-Film Electronics Laboratory of EPFL's School of Engineering, together with CSEM, has been working on a key aspect of the industry: achieving very high performance without inheriting the costs of more advanced technologies. Multi-junction III-V cells can achieve up to 37% efficiency, but they are constructed with multiple, expensive semiconductor layers and, according to the researchers, can cost about a thousand times more per watt than conventional terrestrial cells. The industrial challenge is therefore twofold: increasing efficiency while maintaining a cost trajectory compatible with large-scale terrestrial applications.

We demonstrate that, with intelligent design and manufacturing, we can approach performance levels traditionally reserved for the more expensive III-V multi-junction solar cells used in space, which are composed of multiple semiconductor layers. These can achieve up to 37 percent efficiency and cost about a thousand times more per watt than terrestrial cells. Our approach paves the way for a new generation of high-efficiency, industrially viable multi-junction photovoltaics.

said Kerem Artuk, first author of the study, former PhD student at EPFL and now at CSEM.

Three junctions to capture more solar spectrum

For decades, traditional silicon photovoltaic cells have been the cornerstone of the solar industry. Their success stems from a difficult-to-replicate combination: material availability, reliability, mature production chains, and reduced costs thanks to global scale. But silicon, like any semiconductor material, has physical limitations in converting light. Tandem and multi-junction architectures were developed precisely to overcome these limitations by pairing materials with different bandgaps and thus capable of absorbing different portions of the spectrum more effectively.

In the device developed in Switzerland, the lower cell in silicon absorbs the part of the light most suited to its properties, while the two upper cells in perovskite They work as thin films deposited on top of the substrate. Perovskite, in this context, is not a single material, but a family of semiconductors with optical and electronic properties tunable through their chemical composition. This flexibility is one of the reasons for its industrial appeal: it allows for the design of layers with complementary absorption characteristics, without completely abandoning the silicon production platform.

The 30,02 percent result is also significant because it marks a leap from the group's first demonstration, which in 2018 achieved 13 percent. In less than a decade, the experimental trajectory has moved from an initial test to a configuration certified to exceed 30 percent. This doesn't mean the technology is ready for a rooftop or utility-scale installation, but it does indicate that the architecture has real room for development. Christophe Ballif, head of the PV-Lab, emphasized that triple-junction cells have higher efficiency potential than single-junction cells and even tandem cells, with theoretical and design values ​​well above 40 percent.

“Our first demonstration in 2018 was only 13 percent efficient, so reaching over 30 percent in a triple-junction device today is a remarkable achievement. Triple-junction solar cells have even higher efficiency potential than single-junction and tandem cells, well above 40 percent.”

has explained Christophe Ballif, head of the PV-Lab.

The key was to reduce defects, losses and wasted light

The research addressed two typical limitations of perovskite-silicon triple-junction cells: the low voltage in the upper perovskite cell and the limited current generation in the intermediate cell. These are technical problems, but they have very concrete industrial significance. Any defect in crystalline growth can result in charge loss; any incorrectly absorbed portion of the spectrum reduces the current; any optical misalignment between layers compromises the architectural advantage.

The team introduced three main modifications. The first involves a molecule capable of guiding the formation of perovskite crystals and eliminating defects. This intervention allowed the upper cell to generate a higher voltage, equal to 1,4 volt under solar illumination. The second innovation is a three-step fabrication method for the intermediate cell, designed to enhance light absorption in the near-infrared region. The third involves inserting nanoparticles between the lower silicon cell and the intermediate perovskite cell, reflecting additional light toward the central layer and increasing the current produced.

These details show why contemporary photovoltaic innovation is no longer just a matter of "new materials." Interface chemistry matters, but so does optical engineering. Thin-film deposition matters, but so does managing the photons that pass through layers with different thicknesses, refractive indices, and functions. In a triple junction, the final efficiency is the result of a delicate balance: if one layer produces less current than the others, the entire device suffers.

The presence of cells measuring 1, 4, and 54 square centimeters in the images released by the lab also indicates a shift in experimental scale. This isn't yet the size of a commercial module, but it's an important signal: a credible photovoltaic technology must be able to transcend the micro-laboratory area and demonstrate process continuity across progressively larger surfaces. This is where the role of the CSEM comes in, geared toward technology transfer and industrial scalability.

From space to terrestrial networks, the expected cost changes

The comparison with III-V cells is one of the most interesting aspects of the story. In space, where every gram, every watt, and every square centimeter counts, the cost per watt can be much higher than in terrestrial applications. Satellites require high efficiency, endurance, and reliability in extreme conditions. On Earth, however, the competition is based on cost, durability, manufacturability, and integration with modules, inverters, storage, and power grids. For this reason, a technology that approaches the performance of space cells using cheaper materials could change the scope of high-efficiency applications.

Le cells III-V They will continue to play a role in the most demanding segments, but the EPFL-CSEM result suggests another path: using perovskites to bring some of that performance to more affordable devices. This prospect concerns both terrestrial and space applications. In the first case, interest lies in the possibility of generating more energy from the same surface area, a key factor for urban rooftops, systems in restricted areas, integrated infrastructure, and systems with limited space. In the second case, the focus is on reducing the cost of certain photovoltaic platforms without sacrificing high efficiency.

However, a cell record shouldn't be confused with a commercial product. The photovoltaic market evaluates a technology based on more stringent parameters: long-term stability, humidity resistance, thermal behavior, compatibility with encapsulants and glass, material availability, process repeatability, and large-area yield. Perovskites have made significant progress, but durability remains one of the most scrutinized issues by manufacturers. Therefore, the next work indicated by the group concerns the manufacturing scale-up, stability testing, and integration into future products. According to industry analysts,

"The industrial value of a triple-junction cell is measured not only by peak efficiency, but also by its ability to fit into production lines that are controllable, repeatable, and compatible with the silicon supply chain. A technology of this type becomes attractive when the increased efficiency offsets additional complexity, process costs, degradation risks, and the need for new quality controls throughout the module's supply chain."

The Swiss supply chain focuses on technology transfer

The project is also an example of collaborative innovation. In addition to EPFL and CSEM, the research involved contributions from Fraunhofer ISE and the University of Freiburg in Germany. Empa in Switzerland, Northwestern University in the United States, Helmholtz-Zentrum Berlin, the University of Queensland in Australia, the University of Potsdam, Arizona State University, the ALBA Synchrotron in Spain, the University of Groningen in the Netherlands, and EPFL Valais-Wallis. It's a broad network, consistent with the nature of the problem: multi-junction cells require expertise in semiconductor physics, materials chemistry, microscopy, optical characterization, process engineering, and reliability.

The funding involved European programmes such as Horizon TRIUMPH and VIPERLAB, the Green Energy Fund of the Geneva Industrial Services, the Swiss State Secretariat for Education, Research and Innovation, the Swiss Federal Office of Energy, and the Swiss National Science Foundation. This aspect is also significant: advanced photovoltaics is an energy technology, but it is also an industrial policy. Whoever controls the materials, processes, and patents for future solar architectures can influence value chains that are currently highly concentrated and subject to geopolitical pressure.

Switzerland doesn't compete with major global manufacturers in terms of the volume of standard modules, but it can play a significant role in applied research, instrumentation, metrology, transfer to industry, and specialization in high-value solutions. In this sense, the collaboration between an academic laboratory like EPFL and an industrialization-oriented center like CSEM is a recurring model of Swiss innovation: fundamental science, precision engineering, and experimental validation all work together.

"This project illustrates the power of combining fundamental science with Swiss engineering expertise. By demonstrating that low-cost perovskite materials can approach the performance of the most advanced space photovoltaics, this research sets a new benchmark for multi-junction photovoltaics."

said christian wolff, team leader of the EPFL.

The technical curiosity, ultimately, is that exceeding the 30 percent threshold doesn't depend on a single "trick," but on a combination of targeted adjustments. A molecule that better arranges the crystals. A three-step process that helps the intermediate cell absorb more light. Nanoparticles that redirect photons where they're needed. This is a useful lesson beyond photovoltaics: mature technologies evolve when materials, interfaces, and architectures are designed as a single system, not as separate components.

For the solar sector, the message is cautious but clear. Silicon remains the dominant manufacturing base, while perovskites continue to advance as a complementary platform for increasing efficiency. The EPFL-CSEM triple-junction doesn't end the commercial game, but reinforces a direction: the future of solar energy may not be a complete replacement of existing photovoltaics, but rather a progressive integration of new functional layers over the existing industrial infrastructure. It is in this space, between laboratory records and factory floors, that the next phase of solar innovation will unfold.

Here are three insights that might interest you:

How albedo drives bifacial photovoltaic efficiency
New Swiss practice for the "double face" capture of solar energy
A super photovoltaic thanks to "hybrid" perovskite paints