Here's a blog post that corresponds with what IBM presented at the 2019 March Meeting.

In this post, they claim they've measured qubit T1 relaxation times above 500 microseconds, and claim that they've seen evidence of genuinely entangled 18-qubit GHZ states.

This is far beyond anything that I've seen in published literature. The best results as far as I can find are 12 genuinely entangled qubits (up from 10) from the Pan group in China (1), and ~160 microsecond T1 times (2), also from IBM, but using a 3D cavity transmon qubit.


Assuming the claims are credible, what sort of advancements in fabrication (for the long coherence times) and microwave technology (for large entangled states) could allow such fast progress?


So, to begin, I would point out that the 500 micosec T1 time is for a single qubit in isolation, while the GHZ results are on a 20 qubit device. This device has an avg T1 time of around ~75 microsec.

The GHZ results were done by Ken Wei from IBM, and will be published shortly.

enter image description here

In short, the circuit is a standard GHZ building circuit, with a hadamard gate followed by a chain of CNOT gates (As pictured above). These gates are ordered in such a way as to minimize the construction time by running CNOT gates in parallel on the device(not pictured above). Then there is a layer of phase gates (U1 gates in the IBM notation), followed by the reverse sequence of gates to undo the GHZ state. The U1 layer in the middle gives the states a phase that is then kicked-back in the unrolling at the end, and measured by the top qubit.

To make everything happen, you need high-fidelity CNOT gates, and high-fidelity parallel CNOT gates. In addition, and equally important, you need to be able to correct for measurement errors in the final result. Think of this as measurement tomography if you like. This is a part of the IBM Qiskit Ignis package (qiskit.org). Further details of how all this works together will be in the forthcoming publication


Here is the paper: Wei, Ken X., et al. "Verifying Multipartite Entangled GHZ States via Multiple Quantum Coherences" arXiv preprint arXiv:1905.05720 (2019).

  • 2
    $\begingroup$ Hi Paul. You're clearly from the IBM group. It seems like careful calibration and the circuit above might be enough for the GHZ result. I'm quite impressed with the 500us coherence times too. Is there some technological breakthrough, maybe in the fab process, that caused just a jump in the coherence times? $\endgroup$ – psitae May 13 at 16:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.