When a team of German physicists set their simulations running on one of Europe’s most powerful supercomputers, they were hunting for flaws — atomic-level imperfections at solar cell interfaces that could be smoothed away to boost performance. What they found instead stopped them short.
Certain defects, it turned out, weren’t hurting energy transfer. They were helping it. The finding arrives at an uncomfortable moment for the field: despite decades of progress, the average solar panel still converts only about 22 percent of incoming sunlight into electricity, and researchers have been racing to close that gap by pursuing ever-more-perfect materials.
The efficiency ceiling that frustrates solar energy
Germany’s solar story is one of the most cited success cases in renewable energy. The country generated less than one percent of its electricity from solar in 2000; by 2022, that share had climbed to roughly 11 percent. Subsidies, falling hardware costs, and political will all played a role. But beneath that headline growth sits a stubborn technical limit.
The average solar panel today converts only about 22 percent of incoming sunlight into electricity — most of the rest is lost before it can do any useful work. High-energy photons, violet light for instance, carry around three electron volts of energy, but silicon can only convert about 1.1 eV into electricity. The remainder escapes as heat.
That heat loss isn’t just inefficiency on paper. It degrades cell performance and durability over time, compounding the problem across a panel’s entire working life.
Tetracene, singlet fission, and the quest for a perfect interface
One promising approach to recovering that wasted energy involves layering a molecule-thin film of tetracene — an organic semiconductor — on top of a silicon cell. Tetracene has a useful property: when it absorbs a high-energy photon, it can split the resulting excitation into two lower-energy excitations through a process called singlet fission. Those lower-energy excitons can then be passed into the silicon layer, where most of their energy gets converted into electricity rather than heat.
The challenge lies at the boundary between the two materials. Researchers have spent considerable effort trying to engineer a flawless tetracene-silicon interface, working from the reasonable assumption that any atomic irregularity would disrupt the delicate transfer of excitons. The field was moving toward cleaner, more ordered structures.
A supercomputer reveals the surprise in the imperfection
Prof. Wolf Gero Schmidt’s team at the University of Paderborn approached the problem using ab initio molecular dynamics simulations — computationally intensive calculations that track how hundreds of atoms and their electrons interact over time, advancing in femtosecond steps. To run them, the team relied on HLRS’s Hawk supercomputer in Stuttgart, one of Europe’s most capable high-performance computing systems.
What the simulations showed contradicted the prevailing assumption. Silicon dangling bonds — atoms at the interface that aren’t fully bonded to their neighbors — are typically treated as liabilities in electronic systems. Here, they were doing something unexpected: actively fostering exciton transfer across the tetracene-silicon interface rather than blocking it. The team published the findings in Physical Review Letters.
Why ‘defect’ doesn’t have to mean ‘flaw’
The word “defect” carries obvious negative connotations, but Prof. Uwe Gerstmann, a collaborator on the project, pushes back on that framing. Semiconductor physics already relies on strategically introduced impurities — donors and acceptors — to build the diodes and transistors underpinning modern electronics. In that context, a defect is a design tool, not a mistake.
Dr. Marvin Krenz, the paper’s lead author, put the finding in sharper relief. The field had been moving toward removing defects at all costs. This result points elsewhere. “Our paper might be interesting for the larger research community because it points out a different way to go when it comes to designing these systems,” he said.
That reframing — from “perfectly clean” to “perfectly imperfect” — is more than semantic. It opens a genuinely different set of engineering possibilities for hybrid solar cell design.
Designing the next generation of ‘perfectly imperfect’ solar cells
The Paderborn team isn’t claiming a finished solution. What they have is a new direction, and they intend to follow it systematically. Using future computing allocations, the group plans to deploy AIMD simulations to map out interface configurations with optimized dangling-bond patterns — essentially designing imperfection with precision.
The potential payoff is meaningful. Schmidt estimates that consistently applying singlet fission could, in principle, boost efficiency by a factor of 1.4. Solar cell efficiency has improved at roughly one percent per year across various architectures over recent decades, a slow but steady trend. Work like this, Schmidt suggests, supports the case that the trend can continue.
Whether dangling bonds eventually become a standard feature of commercial solar cell design depends on how well simulations translate into physical prototypes. But the more immediate consequence is already here: the assumption that perfection is always the goal just got considerably harder to defend.