Thomas Edison once said, “I have not failed. I've just found 10,000 ways that won't work.” All invention is a combination of science and sweat; theory provides the framework, but trial-and-error gives the final approval. The problem is that those 10,000 experiments are costly, time consuming, and usually lead to dead ends. Scientists at the Center for Inverse Design, a division of the National Renewable Energy Laboratory (NREL), are streamlining that process by running millions of virtual experiments in order to create new “designer” materials, some of which may become the next generation of solar cells.
Image courtesy of NREL
The process, dubbed “Inverse Design” and illustrated above, is a high-tech variation of the standard engineering design methodology. In any design, engineers begin with the desired outputs: what task should the product perform? In the case of a solar cell, the goal is to produce electricity from sunlight. Next is the brainstorming stage, where engineers put a bunch of possible solutions on the table. Then they start weeding out ideas that are less feasible and focus on a few of the best solutions. Designers will look at various materials with photovoltaic properties and determine which are likely to perform the best under a given set of conditions. They may combine materials to get the best properties of each. Some of those become prototypes that are tested in the lab. Finally, they settle on one solution and implement it.
The Inverse Design process does the brainstorming, weeding, and simulating in a computer using a sophisticated algorithm developed by NREL scientists. This allows the designers to evaluate millions of possibilities in a short period of time. Imagine trying to do that manually! But the “inverse” part is what makes it unique. Instead of asking “I wonder what happens if we put these materials together?” and then testing the combination, the inverse design process looks at the desired outputs and determines the atomic properties necessary to make it happen. Run that process for a million combinations, perform simulations of each, and eventually a handful of winners will emerge. Those are sent to the lab for testing. Targeted experimentation takes most of the guess work out of the process.
"We can cover in a week what used to take a researcher's entire post-doctoral career," says NREL researcher David Ginley. "Being able to do material discovery as fast as the theorists can do the theory creates the ability to really carry out inverse design."
If Edison were alive today, he might say, “Genius is one percent inspiration, twenty percent perspiration, and seventy-nine percent computation.” But Edison was quite egotistical; he might actually say, “Get your damn computers out of my lab - I’m the brains of this operation!”
Tesla, on the other hand...