A Quantum Means Toward a Macro End: The Solid-State Processor

At the 1st Conference on Physics and Computation at MIT in 1981, Nobel Laureate Richard Feynman noted that no classical computer could simulate a large quantum system. At the same time, he proposed a truly stunning thought: perhaps only by using quantum mechanics could we simulate a quantum world.

Thus was born the idea of the quantum computer, a machine that uses the rules of quantum mechanics to do things impossible using a classical computer.

More recently, and closer to home, a group of Yale physicists took a major step forward in the quest to bring to life Feyn­man’s dream. The group, led by Robert Schoelkopf, Professor of Applied Physics & Physics, built the first solid-state quan­tum processor and used it to run a simple quantum search algorithm. Their results were published in Nature.

Classical vs. Quantum: What’s the Dif­ference?

The basis of all computing is the “bit,” which is how a computer stores informa­tion. A classical bit has two states: it can be either 0 or 1 – something or nothing. Information is stored in strings of bits known as a register.

The key difference between a quantum computer and a classical computer is that a quantum computer uses quantum bits, known in computer lingo as “q-bits.” Instead of being confined to 0 or 1, a q-bit can be in any combination of 0 and 1. Addi­tionally, two or more q-bits can become “entangled,” a strange property of a set of quantum mechanical particles that makes it impossible to describe the state of one of the particles without also describing the state of the other. The subsequent increase in computing power is enormous.

As a result, quantum computers are able to perform certain tasks, such as factoring large numbers, much more quickly than classical computers can.

Building the Q-bit

In theory, quantum computing is simple and elegant. But as any engineer can tell you, simple theory rarely translates into simple practice, and the engineering required to manipulate and use q-bits is especially daunting.

The possibility of manmade q-bits was first demonstrated in 1998 by a group of Japanese scientists, led by Yasunobu Nakamura, at NEC Corporation. At Yale, the new quantum processor is realized as a superconducting circuit that uses quantum electrodynamics. The processor consists of a cavity that resonates at a frequency of seven gigahertz (GHz) with two q-bits situated at opposite ends of the cavity. This cavity protects the q-bits from losing energy to the environment, and it also allows the q-bits to be read with and manipulated with a microwave tone generator. The q-bits are thus superconducting structures that act as effective two-level quantum systems.

Magnets are used to tune the transitions between energy levels of the individual q-bits. Any quantum algorithm involves what are called single q-bit operations, or operations that act only on one of the q-bits. The user must be sure that the opera­tions do not affect the other q-bit, and the group does this by tuning the two bits with significantly different frequencies.

However, a quantum algorithm also involves operations that require the q-bits to be coupled. In order to accomplish this, the team uses tuning frequencies that are very close.

The team needed to be able to switch between these two modes of tuning very quickly. As a result, they designed the pro­cessor to work using something called mag­netic-flux bias lines. Postdoctoral Associate Leo DiCarlo, lead author of the Nature paper, explains, “Having these flux bias lines that allow you to tune the q-bits on a very short time scale was an innovation. In a nanosecond, we can bring the q-bits from several gigahertz apart to being close enough that the interaction that entangles them can be turned on very suddenly.”

When a q-bit is bumped into its excited state, which must happen during quantum computing, it will naturally decay back into the ground state by losing energy to the environment. If this happens too quickly, the q-bit is useless, as the information stored in it is lost before it can be used. Therefore, the team needed the q-bits to remain in their excited state for approxi­mately one microsecond.

Indeed, their processor was able to achieve this. The result was a processor that essentially allows someone to choose the correct answer of four choices using only one guess instead of two or three.

Though many of the team’s techniques had been used before, their seamless com­bination of them into a working processor that was actually able to run an algorithm was groundbreaking. As DiCarlo says, “A lot of the value of this work is in the inte­gration of techniques that were previously already demonstrated.” The success of the new quantum processor may be best thought of as one important step forward in the long journey toward a fully functional quantum computer.

The 7-mm solid-state two q-bit processor built by Leo DiCarlo. It can perform simple computations.

The Long Journey

At Yale, that journey began just before the turn of the millennium, when Robert Schoelkopf joined the faculty in 1998 after working as a post-doc with Professor of Applied Physics & Physics Dan Prober. Right around that time, two results – one from Yasunobu Nakamura at NEC Corpo­ration and the other from Michel Devoret, who was then working in France but is now Yale’s Frederick W. Beinecke Profes­sor of Applied Physics and Professor of Physics – indicated that superconducting circuits could be used to create two-level quantum systems.

During his time in Prober’s lab, Schoel­kopf had experimented with building very sensitive electrometers, and the results of Nakamura and Devoret struck him as an interesting direction in which to apply that skill. “That was immediately obvious as a useful and novel way to measure whether you’d done some manipulation of one of these solid-state q-bits,” says Schoelkopf.

Schoelkopf began his own lab and worked on building and manipulating these solid-state q-bits. Devoret then came to Yale on sabbatical and returned several years later to join the faculty. At the same time, Steve Girvin, Eugene Higgins Professor of Physics and Professor of Applied Physics, also joined the Yale faculty. At that point, there was, in the words of Schoelkopf, “a meeting of the minds.”

Schoelkopf and Girvin had already been thinking about ways to couple q-bits using photons in the microwave range, and in 2003, they managed to measure a q-bit using photons instead of electrometers. “That was quite a breakthrough,” says Schoelkopf, who now has a printout of data from that day posted on the wall in his office. There are two question marks in the upper left hand corner of the printout, which, remarks Schoelkopf, “was my post-doc and student saying ‘maybe it’s real!’”

A few years later, the group redesigned their q-bits to make them much easier to work with and much longer lived than anyone had managed previously. Before this redesign, say Schoelkopf, “When we got a q-bit working it was a very finicky thing and you’d sort of spend all day at the bench top tweaking and measuring and tweaking and measuring until it finally worked nicely.” Yet after the redesign, “The q-bit was there immediately and we didn’t have to do any of this tweaking… the q-bit would sit there while we did some complicated experiment all day, instead of having to tweak all day and getting only two minutes of data.”

Schoelkopf recalls, “That was when we knew that our life was going to get much, much better. Up until that point, I hadn’t been too optimistic about being able to build two or three q-bit, but from that point on, I knew we’d be able to do a few things.”

The Future of the Quantum Computer? Ask Again Later

Those few things that Schoelkopf real­ized would be possible have now been achieved. Still, it is important to note that research into building a quantum computer is still in its infancy. Says Schoelkopf, “No one’s ever done an interesting calculation with a quantum computer that wouldn’t have been trivial with pencil and paper.” Indeed, Schoelkopf compares current quantum computer research to the early stages of computers in the 1940s. “We talk about our research as being on quantum computers,” he says, “but in a sense it’s more like working to develop the quantum transistor.”

So how far are we from the “interesting calculations?” No one knows for sure, but Schoelkopf says that such calculations would require around 100,000 q-bits. Considering that the processor that has sparked so much excitement contains only two q-bits, that could be a long way off.

In addition to the technical problem of building and coupling that many q-bits, Steve Girvin foresees an interesting theo­retical problem that may arise when com­puters reach 30 q-bits. Just as you need a modern classical computer to build another modern classical computer, it becomes dif­ficult to build a quatum computer with 30 q-bits without the aid of another quantum computer. With smaller quantum comput­ers, a classical computer can be used to model the system, but, says Girvin, “We don’t know how to simulate quantum systems with more than 20 or 30 q-bits. You have to diagonalize a one billion by one billion matrix, and we can’t do that on classical computers.”

How, or even if, this problem will be solved is currently unclear. What does seem clear is that a few more breakthroughs will be required for the quantum com­puter to become a reality. As Schoelkopf explains, “There are linear evolu­tions in the laboratory, and then there are revo­lutionary things where all of a sudden the project advances so much that everything you had done before is made obsolete. We need another one or two of those before you would really ever consider this as a way of doing information processing.”

Few in the field seem willing to predict the future of quantum computing with­out large amounts of hedging. And that is as it should be, for anyone claiming to know such a thing would be seen as either extremely arrogant or simply lying. How­ever, when forced to guess, Schoelkopf offers a clever stock answer that he credits to Nobel Laureate Bill Phillips: “The chance of having a technologically useful quantum computer is 50-50. The meaning of that is that there’s a fifty percent chance that it will happen sometime in the next fifty years.”

Whatever the future of quantum com­puting, there was an exciting pit stop along the way at Yale this summer. When and where next one comes, we don’t know, but you can be sure that Schoelkopf and his group hope to be there when it happens.

About the Author
CHRISTOPHER KACHULIS is a junior in Ezra Stiles College majoring in Physics.

The author would like to thank Professor Robert Schoelkopf, Professor Steven Girvin, and Leonardo DiCarlo for their assistance.