High Visibility In Superconducting Qubits

One of the more attractive approaches to building a quantum computer are the various proposals which utilize superconducting electrical circuits. One of the benefits of using superconducters to build a quantum computer is that it is expected that scaling up from a few to many qubits might be easier because of the advanced state of the fabrication for superconducing circuits (Although it must be said that “fitting everything together” will certainly still be a challenge. But it does seem like a slightly easier challenge, than, say loading 1000 ions into separate ion traps.)

One of the interesting problems with superconducting qubits has been that no one has been able to demonstrate high visibility of the qubits. One question which is particularly acute for solid state qubits is whether they can be sufficiently isolated from their environment to act as truly two level systems. Interestingly, some of the early experiments with superconducting circuits, while they demonstrated Rabi flopping of a single qubit, these experiments weren’t able to get high visibility of this flopping. What this means is that instead of observing the the qubit system flopping back and forth between 100% population in one of the qubit states to 100% population in the other qubit states, the experiments observed, say 100% in one state and then, say 30% in the other state. And in these experiments, while the flopping was not between 100% |0> and 100% |1>, after this initial reduction to, in our example, 30%, the qubits then Rabi flopped with a pretty slow decay. Thus it seemed that there was a “visibility” problem: the qubits were probably oscillating properly, but something in the scheme was causing the measurements to not see this full oscillation. Hence there was a “visibility” problem.

Now in todays Physical Review Letters, A. Wallraff et. al. from Yale, report on superconducting experiments in which they achieve nearly 95% visibility in Rabi flop measurements of superconducting qubits! The trick which these authors use is to couple their (charge) superconducting circuit to, basically, an electromagnetic cavity. Thus what these experimentalists are able to achieve is a superconducting qubit which can be strongly coupled to a quantum electrodynamic cavity mode. They then use this nice pure coupling to perform a measurement on the superconducting qubit with the beautiful result that they really high visibilities.

This is very exciting news. Ion trap proposals for quantum computers can still obtain higher visibility in their experiments, but this movement from visibilities less than 50 percent to 95 percent is an awesome jump. I can’t wait until they get this up to 99 percent and then to 99.9 percent. Then we will be rocking!

This entry was posted in Computer Science, Quantum. Bookmark the permalink.

2 Responses to High Visibility In Superconducting Qubits

  1. aram harrow says:

    If everything except hte measuremnet is reliable, then low visibility shouldn’t be that terrible. You can always do a measurement by making a cat state with a bunch of CNOTs and then doing majority voting.

    I might be wrong here, but I thought a bigger deal (or related deal) is the speed of the measurement.

  2. Dave Bacon says:

    Aram: this is partially true. I think you are assuming that the poor visibility comes from the measurement. The poor visibility just as easily could have come from the preparation. In the former case you can indeed use CNOTs and do majority voting. In the latter, you simply need to run the experiment multiple times and then take the majority vote of these experiments, or, better yet, you need to use a majority voting measurement as in the former case to prepare your quantum computer.

    It is true that having fast measurements is nice (because you really need to dump your entropy in quantum error correction), the method of multiple CNOTs effectively makes the measurement extremely slow!

    And, of course, I think anytime you can get high visiblity measurements you will be much happier!

Leave a Reply

Your email address will not be published. Required fields are marked *