I have been checking out the parameters of the new ibm_washington processor and I have the following doubt about the calibration data provided by them. Checking out the relaxation and dephasing times I found out that some of their qubits are said to have $T_2>2T_1$. For example, see qubits Q2 or Q21. I understand that the dephasing times they provide are the ones obtained by Ramsey's experiments. However, relaxation and dephasing times are related by the expression \begin{equation} \frac{1}{T_2}=\frac{1}{2T_1}+\frac{1}{T_\phi}, \end{equation} where $T_\phi$ is the pure dephasing time. From this equation, it can be seen that qubits that have $T_2>2T_1$ make no physical sense since that would imply that the pure dephasing time is negative. Therefore, I am wondering what's going on with the decoherence time values that are being provided by IBM for the newest processor. I have thought about measurement error, but qubits Q2 and Q21 are not even close to the Ramsey limit $T_2\approx 2T_1$. Maybe that due to the novelty of the system the data is not still accurate? Or may I be missing something?
$\begingroup$
$\endgroup$
asked Nov 19, 2021 at 17:51
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Good catch! This is a result of the T1 and T2 properties of the qubits being estimated in separate measurement batches. What was happening is that a qubit fluctuation such as a TLS would cause a low T1 to be measured. Sometime later the fluctuation would disappear (and the qubit would recover its inherent T1) followed by the measurement of the T2s. The effect of separating these measurements is the appearance that some T2s might violate the T1 limit.
Going forward T1/T2s will be measured together to reduce the incidence of such events.
