An AI materials discovery system at Lawrence Berkeley National Laboratory has designed a new photovoltaic material that is 10 times more efficient at converting sunlight to electricity than traditional silicon panels.
The Material
The AI screened 32 million possible molecular combinations before identifying a perovskite-organic hybrid that achieves 47% solar conversion efficiency, compared to silicon's theoretical maximum of 29%.
Timeline
Researchers caution that commercialization is 5-7 years away, but if manufacturing challenges are solved, the material could cut solar energy costs by 80%.
- 47% solar conversion efficiency
- 10x improvement over silicon
- 32 million combinations screened by AI
- Could reduce solar costs 80%