From the Abacus to Supercomputers to Quantum Computers

Estimated reading time: 2 minutes

If using quantum mechanics to compute problems that are unsolvable with today’s fastest supercomputers sounds outrageously ambitious, that’s because it is. There are many experts who say that it can’t be done.

But that’s not stopping Jungsang Kim, professor of electrical and computer engineering at Duke University, from pursuing the impossible. A pioneer in translating theoretical quantum physics into physical hardware, Kim has been engineering the components for a quantum computer at Duke for more than a decade.

And he’s starting to sniff the finish line.

“We’ve put together and demonstrated all of the individual components needed to build a large, scalable quantum computer,” said Kim. “We are convinced that within the next few years we could turn this technology into much more sophisticated quantum computers with the potential to solve problems considered impossible today.”

Imagine a computer trying to put together a jigsaw puzzle. Because computer code is binary, either a piece fits or it doesn’t, the most efficient method would be to pick a piece at random and attempt to attach every other available piece until one fits. Today’s computers would then take that two-piece unit, and repeat the entire process over and over until the puzzle is completed.

Even with today’s supercomputers, this process would take a long time because it must be done sequentially. Quantum computers, however, have the advantage of occupying many different states at the same time.

Now imagine a quantum computer with enough “qubits”—individual pieces of memory analogous to today’s transistors—to assign one to each puzzle piece. Thanks to quantum mechanics, all possible configurations are stored into a quantum memory, which is manipulated in a very careful way so that all the non-answers fade away very quickly and all the real answers emerge in a systematic way. This allows the quantum computer to converge on a solution much more efficiently than a classical computer.

“Nobel Laureate Bill Phillips said that using quantum principles to compute is as different from classical computing as a classical supercomputer is from an abacus,” said Kim. “There are, however, several different ways that one might achieve this. Our group has focused on approaches using individually trapped ions.”

The qubits in Kim’s quantum computer are individually trapped ions—atoms with electrons stripped away to give it a positive electric charge. That charge allows researchers to suspend the atoms using an electromagnetic field in an ultra-high vacuum. Kim and his colleagues then use precise lasers to manipulate their quantum states.

The method is promising. Kim and colleague Christopher Monroe at the University of Maryland have secured more than $60 million in grants to transition these ideas into large, scalable quantum computers. And they’re not alone—many other big companies like Google, IBM, Microsoft and Intel are starting to make big investments as well.

With the potential to revolutionize industries such as materials design, pharmaceutical discovery and security encryption, the race is on. And Kim and his colleagues are the only ones betting on trapped ions, having started a company called IonQ to pursue commercialization of the technology.

“Our collaboration actually has a small qubit quantum computer that's very generally programmable,” said Kim. “We think we know how to take this system and turn it into a much bigger system that is reliable, stable and much more scalable. We've come to a point where we believe that even commercially viable systems can be put together.”

Related video: