The late physicist Erwin Schrodinger was probably relieved to know that flesh and blood cats are too big to behave according to the laws of quantum physics. His intellectual heirs, however, no longer have that luxury.
The line between the large and the small is not so clear cut as it was in Schrodinger's day. That's due as much to the work of physicist David Wineland as anyone else. Wineland received a Nobel in 2012 for his work trapping individual atoms. Today, to an audience of early career chemists at the opening session of the 63d annual Nobel Laureate Meeting in Lindau, Germany, he talked about the implications of the tremendous precision that scientists can bring to bear in the lab. When it's possible to manipulate single atoms, the dilemma of Schrodinger’s cat moves from thought experiment to the realm of practical problem-solving.
In Wineland's lab at the National Institute of Standards and Technology, he and his colleagues take a single ionized atom of, say, beryllium and put it into a magnetic trap. There it acts a bit like a marble rolling around in a bowl. It’s on the left side of the bowl, it's on the right, it's in both places at once. This particular bowl is 80 nanometers wide, and the marble is a “wave packet” that is all of seven nm. “You might argue that it isn't really Schrodinger’s cat because it's small,” Wineland said. But where do you draw the line? “We struggle with that, because at this point there's no way of finding that classical-quantum boundary.”
Wineland struggles with that boundary because, for one thing, he can, and that's the kind of thing physicists do. But there are practical reasons, such as the prospect of building a quantum computer. A quantum computer could have an absurd amount of processing power, if only someone could iron out the difficulties. Wineland is trying, and he's made some progress. (I say Wineland, but he spoke today for a team of more than 100 scientists at NIST's labs, and seemed uncomfortable in not having time to list every single one of them.)
The idea of a quantum computer is to take advantage of the property of superposition to build a computer that is immensely more complicated and powerful than a classical computer [see our video explainer How Quantum ComputersWork]. (Classical computers are the kind used in real products, even Apple's.) In a classical computer, a byte—which consists of eight bits, each with two possible states, one or zero—can describe 256 different numbers. A byte consisting of 8 qubits (quantum bits, analogous to the ones and zeros of classical computers) can describe many more than that—how many depends on how many states each qubit can assume. The way Wineland described it, a classical computer can store a sentence in 300 bits or so. To match the information that a quantum computer could store in the same number of bits, a classical computer would require all the matter in the universe and then some. "That sounds pretty good," he said.
The catch (one of many catches) is that once you read a qubit, it “collapses”—it assumes a single, definite state, rather than all possible states at once. (Just as when you check up on Schrodinger’s cat, you it is either dead or alive, but not dead and alive.) What good would a super-powerful quantum computer be if, when you asked it where the nearest Starbucks was, it could hardly muster a sentence?
Wineland's team is trying to figure out how to read a few qubits without causing the entire quantum computer to lose its vast computational power. The idea is to build an array of qubits and read them selectively, a few at a time, without having to read them all at once. This requires a great deal of precision. Wineland described an experimental array of qubits consisting of 8 beryllium atoms magnetically suspended, or trapped, above a substrate of silicon. To read one, he shines a fine laser, which "reads" the ion.
One qubit, however, does not a computer make. What’s needed are logic gates—a way of taking an input or a couple of inputs and getting an output based on some pre-defined logic. For instance, a two-input OR gate will produce a 1 if one of its two inputs is also a 1, but it will produce 0 if both inputs are zero. A NOT gate will take a single input and produce the opposite: a one produces 0, a zero produces 1. Computers are made of layers upon layers of such logic gates.
What makes Wineland’s array a computer of a sort is that two qubits in close proximity affect one another in such a way as to act like a logic gate. In his array of 8 qubits, he’s been able to get two at a time to behave pretty close to how logic gates behave in a classical computer. His logic gates, he says, are “a bit of a cheat” because they don’t behave with complete consistency, but it works as a proof of concept. It’s just one way his lab is “playing with the idea of qubit logic gates," he says.
Given that this year’s Lindau meeting is devoted to chemistry, you might wonder why Wineland, a physicist, was asked to present his work. He figures that his linear qubit array is so small, it is “a kind of pseudo molecule."
“It’s the closest I'll get to chemistry.”