PORTLAND, Ore. Rice University researchers have demonstrated that molecule-sized electronic devices can be used for both logic and memory, despite being randomly wired, error-prone and inaccurately formed at the nanoscale.
"The dream of making memory and logic by programming disordered nanosized arrays via micron-sized addresses is now a reality," said Rice professor James Tour.
Tour said his work demonstrates that today's chip makers can achieve increases of two to three orders of magnitude in chip density by leveraging the lithographic tools they already have to form random-access addresses into arrays of nanoscale molecular memories.
Tour recently showed the Defense Advanced Research Projects Agency (Darpa), which is helping to fund his research, that logic gates not just memory can be utilized at the nanoscale.
"We have made nonvolatile memory from these disordered arrays, addressing the nano via the micro," said Tour. "We are more recently showing [to Darpa] that indeed we can program logic into these disordered systems as well."
Tour and his team performed the research along with Pennsylvania State University professor Thomas Mallouk and North Carolina State University electrical engineering professor Paul Franzon.
Tour de force
The problem with nanoscale devices from carbon nanotubes to molecular switches comes in using them after they are fabricated. Tour has fabricated a variety of molecule-sized electronic devices from switches to memories to diodes to resistors but like everyone else, his team had been unable to form working circuits from them. Individual components could be isolated and tested, but connecting them into circuits had remained illusive until now.
By using a search algorithm to find interconnection patterns that perform desired functions, Tour's group was able to demonstrate that his method produced nanoscale functions. His memory nanocells were shown to retain their data for more than a week without refreshing.
Tour cautioned that his work is still research, but pointed out that if only a few percent of the interconnections result in working devices, they might still lead to denser arrays than conventional micron-sized circuits.
"State-of-the-art silicon is so sophisticated that there would need to be a world of improvement and testing [of our devices] before anything commercially viable could be manufactured," Tour said.