New subcell current metrology for perovskite-silicon tandem solar cells

Researchers at Germany’s Fraunhofer Institute for Solar Energy Systems (Fraunhofer ISE) have developed a purely electrical method to determine the currents of both subcells in perovskite-silicon tandem solar cells within milliseconds. The approach exploits the silicon bottom cell’s inherent capacitive charge reservoir.

“Perovskite-silicon tandem cells are moving fast toward mass production, but manufacturers have lacked a way to check the current balance between the two subcells at production speed,” research project lead Christoph Messmer told pv magazine. “Existing techniques such as spectral response or spectrometric analysis take hours and require complex optical setups, making them unsuitable for inline quality control. Our method closes this gap: It determines both subcell currents within milliseconds, using only the standard current-voltage test that is already performed on every cell.”

“The key idea is to exploit the silicon bottom cell’s own capacitive charge reservoir,” Messmer said. “A short forward pre-bias charges this reservoir, and a subsequent voltage step triggers a current overshoot. Analyzing this transient reveals a perovskite-limited plateau followed by the silicon-limited steady state — giving us both subcell currents from a single, millisecond-scale measurement. Because the method requires no additional hardware, light sources, or spectral filters, it can be integrated directly into existing inline testers. This makes 100% inline sampling possible for the first time, allowing manufacturers to detect drift in perovskite deposition or optical-stack variations in real time, before they lead to yield loss.”

In the paper “Inline millisecond subcell current metrology for perovskite/silicon tandem solar cells,” published in Joule, Messmer and his colleagues explained that the silicon bottom cell typically limits the steady-state current when it generates less photocurrent than the perovskite top cell. Their method temporarily removes this limitation by first applying a forward bias, ideally close to the open-circuit voltage, to store excess charge in the silicon cell.

The device is then rapidly switched to a lower voltage, causing the stored charge to discharge. During this brief period, the additional discharge current prevents the silicon cell from limiting the tandem current, meaning the measured current corresponds directly to the perovskite subcell current.

Once the stored charge is depleted, the silicon cell becomes current-limiting again and the measured current drops to the silicon subcell current. The two current levels can therefore be used to separately determine the perovskite and silicon subcell currents in a two-terminal tandem device.

The scientists tested the methodology using simulations and experiments.

Their simulations showed that, after preconditioning at open circuit and rapidly switching to short circuit, the tandem current can display two distinct plateaus. The first corresponds to the perovskite current, while the second appears after the silicon charge reservoir is depleted and corresponds to the silicon current. The time between the two plateaus depends on the current mismatch between the subcells.

The researchers first experimentally demonstrated the method on a small in-house tandem cell and then tested it on a full-size industrial device. For the industrial cell, fast scans of about 7 ms to 12 ms produced two clear plateaus that closely matched the subcell currents obtained through conventional spectrometric characterization, which requires more than an hour. The measurements were also highly reproducible, with a relative standard deviation below 0.1%, according to the research team, indicating that repeated fast scans did not measurably degrade the device.

For practical measurements, the method can be integrated into a standard fast reverse j–V scan. When the scan is performed on a millisecond timescale, the discharge of the silicon reservoir produces the two current levels, allowing both subcell currents to be determined in a single measurement. However, the approach works when the silicon cell is current-limiting. If the perovskite cell already limits the tandem current, only one plateau appears and the silicon current remains hidden.

“From a manufacturing perspective, 100% inline subcell current sampling becomes feasible, enabling statistical process control charts that detect drift in perovskite deposition—thickness, composition, morphology—or optical-stack variations before they propagate into yield loss,” the academics concluded. “Because the measurement uses only bias conditions already present in standard j–V testing, integration into existing testers requires no hardware modification, only a firmware-level adjustment to the scan protocol.”

“The underlying principle of the new metrology is not limited to perovskite-silicon tandems,” Messmer added. “It applies to any multijunction architecture combining a direct-bandgap top cell with an indirect-bandgap bottom cell, including III–V-silicon tandems and triple-junction devices.”