At first glance, "quantum information" may seem like a contradiction in terms. After all, the quantum-mechanical properties of any object are unknown - and unknowable - because the object exists in a superposition of all possible properties at once, at least until a measurement is made.

Paradoxically, that situation offers great promise for certain kinds of computational tasks that are enormously difficult now. JQI scientists are investigating methods to produce and control quantum effects that can be exploited to process information in new ways, and results to date are highly encouraging. Their work epitomizes the stark differences between the quantum and classical worlds.

In conventional electronic computers, information is stored and processed in the form of strings of "bits" (binary digits). Each individual bit can have only one of two values: 0 or 1. That either-or digital condition is easy to represent with electrical charges in chips, tiny holes on a CD surface or microscopic magnetized spots on a disk. The longer the string of bits, the larger the number of values that can be represented. Thus, for example, it takes a string of three bits to represent any one of the eight numerical values from zero (000) to seven (111). So a total of 24 bits would be needed to represent all eight possibilities.

But in a quantum computer, information would be stored in qubits, each of which, thanks to the nature of superposition, can be 0, 1 or both at once . Thus only three qubits are needed to represent all eight digits simultaneously. That capability could make some mathematical operations exponentially faster.

One of them is the task of factoring the extremely large numbers that serve as the "public keys" in current encryption and data-protection schemes used to encode sensitive information in dozens of applications from bank transfers to military intelligence. The public-key numbers - which typically require a string of more than 1,000 bits to represent - are the product of a set of prime numbers known only to the sender and receiver. Decoding the information requires the user to know those factors. The encoded data are safe because it would take even the best modern computers years to run through all the possible combinations of prime numbers which, when multiplied together, produce the giant public keys. But a quantum computer could solve the problem in hours or even minutes.

Another potential use is in searching large, unstructured databases in which the entries are in no particular order. Locating the specific desired items could be done at vastly greater speed with quantum computation.

So far, JQI experimental research has generally focused on three kinds of objects with properties that can function as qubits: trapped neutral atoms, trapped ions and "artificial atoms." (The last category consists of assemblies of superconducting circuits that, for computational purposes, behave like atoms.)

But there are also other possibilities for storing and processing quantum information, including "topological quantum computing." This novel approach aims to avoid many sources of error by representing qubit data in the form of two-dimensional geometrical patterns of electrically charged "quasiparticles" that form under certain well-understood low-temperature conditions. Because the basic patterns are unchanged by minor variations at one point or another, they would be more robust and error-tolerant than other qubit alternatives.

JQI researchers are examining this option along with others on the road to an optimal design for quantum computing.