EFRC scientists from Fred Schubert’s research group at RPI and at Sandia Labs have published a new paper titled “Genetic Algorithm for Innovative Device Designs in High-Efficiency III-V Nitride Light-Emitting Diodes” in the journal Applied Physics Express. The work is a starting point of applying artificial evolution to practical semiconductor devices, open new perspectives for complex semiconductor device optimization, and enable breakthroughs in high-performance LED design. According to the Applied Physics Express website, this publication was the their most downloaded article in January 2012.
Abstract: Light-emitting diodes are becoming the next-generation light source because of their prominent benefits in energy efficiency, versatility, and benign environmental impact. However, because of the unique polarization effects in III–V nitrides and the high complexity of light-emitting diodes, further breakthroughs towards truly optimized devices are required. Here we introduce the concept of artificial evolution into the device optimization process. Reproduction and selection are accomplished by means of an advanced genetic algorithm and device simulator, respectively. We demonstrate that this approach can lead to new device structures that go beyond conventional approaches. The innovative designs originating from the genetic algorithm and the demonstration of the predicted results by implementing structures suggested by the algorithm establish a new avenue for complex semiconductor device design and optimization.