Thinking Small


You are here

Dr. Roee Ozeri, Prof. Adi Stern and Prof. Moty Heiblum. In the realm of quantum weirdness




Tomorrow's computer might be a quantum one based on the physics of particles smaller than atoms. No one is quite sure what a quantum computer should look like, or even whether it's possible to build a functional one; but scientists at the Weizmann Institute have been working on some of the basic questions that will need to be answered before we can begin to create this new kind of computer.

Is it a qubit?

At the level of atomic and subatomic particles, things work very differently from the macro world of everyday objects. For instance, there's wave-particle duality: The basic bits of matter and light behave sometimes as discrete particles and sometimes as waves, which can be in many locations at once. And there's quantum superposition – particles existing simultaneously in more than one state at a time – which could, theoretically, provide a dramatic increase in computing power. An electronic computer bit can be in only one of two states (0 or 1), whereas a quantum bit (called a qubit) can exist simultaneously in both 0 and 1, in an infinite number of different superpositions. The challenge for scientists is to connect these fragile quantum states to the larger world without destroying them.

Several years ago, Prof. Adi Stern of the Condensed Matter Physics Department came up with a way to test a system to see whether it could be used as a special kind of qubit – a topological quantum bit. The system in question involves, on the one hand, electrons moving in a very cold two-dimensional plane, with a strong magnetic field applied at right angles to the plane and, on the other, quasiparticles. These "imaginary" particles – which have electrical charges of one-third, one-quarter or one-fifth of an electron – don't exist in nature, but they have been created and measured in the lab of Prof. Moty Heiblum of the Condensed Matter Physics Department.

Such a system must meet several criteria before it can be considered a possible qubit. The particles must be able to exchange places, and this exchange must leave a sort of trail that can be traced – that is, it must preserve information. In Stern's theoretical experiment, two parallel lines of current run through such a system with a separation "wall" containing quasiparticles between them. An odd number of quasiparticles should cause the electrons in the current to behave as particles, flowing in line through the material. But if they are separated by an even number, the electrons in the system should act as waves, producing interference patterns at the end of the current pathways.

In addition to the number of fractionally charged particles in a system, the fraction itself is relevant. The quasiparticles Heiblum measured in the 1990s had odd denominators, and these don't leave traces when they exchange places in the plane, making them unfit for storing information. Even-denominator fractions might be better for the purpose, but they're harder to produce. This past year, Heiblum and Stern, together with research student Merav Dolev and Drs. Vladimir Umansky and Diana Mahalu, all of the Condensed Matter Physics Department, succeeded in creating a nanoscopic device in which quasiparticles with one-quarter the charge of an electron were measured for the first time. They are now continuing their experiments to find out if quarter-charge quasiparticles are truly suitable candidates for quantum bits.

Can quantum errors be corrected?



Quantum "weirdness" – the strange reality that rules the world of ultra-tiny particles – presents some unique challenges. For instance, how can one perform calculations in a system in which the very act of measurement changes the basic configuration of that system?

Quantum superposition has been demonstrated in particles such as electrons, but it has never been observed in larger objects composed of many particles. The reason, scientists believe, is that in larger groups the particles interact with one another and with their environment, forcing the quantum superposition of the system into a single classical state. (Measurement is one form of interaction.) This transition is called decoherence. One could conceivably build a very simple quantum computer with only a few qubits, but how to create one that has millions?

Since joining the Weizmann Institute in 2007, Dr. Roee Ozeri and his students Nitzan Akerman, Yinnon Glickman, Shlomi Kotler, Yoni Dallal and Anna Keselman have been setting up a lab in the Physics of Complex Systems Department, and they have recently begun to conduct experiments that may one day help overcome the limitations. Ozeri is especially interested in error correction in quantum computing. Today's electronic computers compensate for possible errors by building in redundancy and using error-correction protocols. In analog quantum protocols, different kinds of error correction may help overcome decoherence and keep superpositions of many particle states "alive." Ozeri is also investigating ways of creating complex quantum logic gates – the basic operations of quantum computing – in which actions performed on one qubit can, under the right conditions, change the state of a second. Because quantum systems can't be measured directly without affecting the result, Ozeri must use roundabout methods that ascertain whether there are errors in the qubits' final state.

His experimental quantum system is based on ions – specifically, atoms of the element strontium that have undergone "laser surgery" to remove some of their electrons. Several of these ions are fired into a vacuum chamber, where they're trapped in an array of electrical fields, while another laser cools them to within a few millionths of a degree of absolute zero. Although Ozeri's experiments trap just a few ions at a time, he can examine the effects of decoherence by applying an electromagnetic field to create noise in the ions' environment. For "communicating" with the ions, he uses yet more lasers, which are precisely tuned to interact with various transitions between strontium ion states.

While the challenge of creating the basis of future computers is compelling, it is ultimately the questions of basic physics that Ozeri finds most fascinating: "We've been exploring the physics of the quantum world for around 100 years, and those of macro systems for much longer, but we still don't know much about the point at which one takes over from the other, how the transition happens or whether it's possible to push the limits and extend the quantum superposition principle into many-particle systems. This research might help provide answers to some of these very basic mysteries."


Do quantum codes communicate better?

Assuming quantum computers become a reality one day, what will they be used for? Will they be more efficient for every type of operation? For example, factoring large numbers – a process that could be used to break some encryption codes – is believed to take an impossibly long time on today's computers, but it could be done quite efficiently on a quantum computer. Prof. Ran Raz of the Computer Science and Applied Mathematics Department investigates whether communication between computers would be better with quantum methods. Quantum computers may be far in the future, but quantum communication has already been successfully demonstrated in experiments.
An example of a problem involving communication is a program for setting up a two-person meeting. The minimum number of bits needed to be communicated today to find a common free hour in each participant's network calendar equals the number of calendar slots that must be checked (n). But a quantum communication protocol could perform the same task using just the square root of (n) bits. Raz found that the difference for some other communication problems could be even greater: The improvement would be logarithmic. In other words, as the value of n rises, quantum communication protocols should quickly leave classic ones behind in the dust.

Prof. Moty Heiblum's research is supported by the Joseph H. and Belle R. Braun Center for Submicron Research. Prof. Heiblum is the incumbent of the Alex and Ida Sussman Professorial Chair of Submicron Electronics.

Dr. Roee Ozeri's research is supported by the Chais Family Fellows Program for New Scientists.