ACAP/ANU researchers' radical design transforms stability of perovskite–silicon tandem solar cell

Creative thinking by ANU’s Dr Daniel Walter and Kunlin Chen led to a counterintuitive approach of deliberately not optimising for peak efficiency to achieve better real-world performance in perovskite-silicon tandem solar cells.

Perovskite solar cells represent a quantum leap in renewable energy technology. When layered on top of conventional silicon cells in a tandem configuration, they can capture far more of the sun's energy than either material alone – pushing efficiencies to record-breaking levels that could dramatically reduce the cost of solar power. Companies are racing to commercialise this technology, with some already achieving efficiencies above 28%.

However, Perovskite–silicon tandem solar cells can look stable in the lab, but struggle in the real world. In real operation, a solar module never sits at a perfectly fixed voltage. Grid interactions, inverters, shading, clouds, temperature swings, and MPPT tracking all cause the voltage across the cell to constantly change.

For perovskite tandems, these voltage swings are especially problematic.

Perovskites suffer from hysteresis, an instability where the cell's performance fluctuates unpredictably depending on recent voltage changes. Charged particles within the perovskite material migrate in response to voltage fluctuations that are common in real-world conditions. This migration not only makes reliable performance measurement nearly impossible, but also accelerates long-term degradation, threatening the technology's commercial viability.

Rather than fighting to eliminate this instability – which is the approach most researchers have pursued – ACAP-supported scientists at ANU took a radically different path.

Most tandem researchers will design the stacked sub-cells so they generate the same electrical current, allowing the device to operate efficiently without one layer limiting the overall power output. This called 'current matching'.

Led by Dr. Daniel Walter, the team discovered that by deliberately running tandem cells in a slightly mismatched mode, where the silicon layer limits the overall current rather than seeking the ideal balance between the two materials, they could dramatically suppress the problematic voltage fluctuations in the perovskite layer.

"Our work shows that if we configure the tandem in a specific way, we can shift some critical stresses to the silicon cell, away from the perovskite cell," explains Dr. Walter.

"Because state-of-the-art silicon solar cells are highly robust and well-adapted to handle the stress, this is a way we can extend the lifetime and performance of silicon-perovskite tandem solar cells by maximally exploiting the robustness of silicon solar cells."

The approach requires accepting a modest efficiency reduction – by as little as 5% – but delivers transformative improvements in stability, measurement reliability, and total energy production over the device's lifetime. By letting the robust silicon layer shoulder more operational stress, the fragile perovskite remains stable far longer.

Real-world voltage fluctuations put heavy stress on perovskite–silicon tandem solar cells, causing hysteresis and faster degradation when the perovskite is current-limiting (blue), or current-matched (green). This helps explain why lab tests often overestimate stability. When Walter and Chen engineered the tandem so that the robust silicon cell limits the current (red) and shoulders more operational stress, voltage cycling is much less damaging to the perovskite, and the tandem’s stability closely matches fixed-bias laboratory conditions (dark grey). 

This stability-first design philosophy represents a fundamental shift in thinking. Rather than chasing maximum peak performance, the research demonstrates how strategic engineering can unlock the commercial potential of ultra-high-efficiency tandem cells, accelerating ACAP's mission to deliver sustainable, affordable solar technology for net-zero economies worldwide.

ANU’s Dr Daniel Walter and Kunlin Chen.