EP4629142A1 - ARRANGEMENT FOR QUBIT RESET (IQM Finland)

EP4629142A1 - ARRANGEMENT FOR QUBIT RESET (IQM Finland)#

This seems a rather trivial patent application, but it is super-fun for me reading it due to my background in quantum non-linear optics. This application is basically art, like a thoroughly choreographed ballet that is part of a saga (even though there is no such thing as a “Nut Cracker 2”). It’s part of a saga because it is tied to some new developments and patents that are state-of-the-art. I learned many things.

TL;DR: The patent application uses a passive dissipation approach to the qubit reset problem. It navigates two main obstacles:

  1. It cannot be done via a single-photon, direct process due to parity preservation laws, but indirectly via a multi-photon process requiring fast, very precise pulse-shaping.

  2. The tasks of implementing coherent gates or measuring the qubit are diametrically opposed to the dissipative-reset task here.

This patent application is clear in its means and objectives, leveraging engineered anharmonicity (\(\alpha\)), a tunable Quantum Circuit Refrigerator (QCR) to “flush” excitations, and flux-tunable Purcell filters to control emission into the \(|g, 2\rangle\) state. It offers a novel path to a technical problem and implements a modern, choreographed approach to invent a technical solution.

Motivation and Challenge: The Dilution Fridge Bottleneck#

Controlling and measuring solid-state based quantum computers (superconducting ones such as transmon or fluxonium, or quantum dots) is a Herculean task. You have your qubits, each coupled to their read-out resonator—a cavity mode in the quantum optical sense.

The problem is they all sit in a fridge cooled down almost to absolute zero, and there is only one way in and out. You are sitting outside this fridge with your AWG and control electronics, sending signals inside through a single line in a sophisticated, multiplexed fashion.

Whether you are resetting qubits, implementing digital quantum gates, or measuring states, some goals are conflicting. When you apply gates, you want the read-out cavities to be as “empty” (free of photons) as possible to avoid decoherence. You apply Purcell filters that make sure qubit excitations do not spontaneously decay to the cavity. When you measure, you fill the cavity but must ensure the qubit’s state is not demolished.

Yet, when you need to reset your qubit state, you need that to happen ASAP. You have to “flush” the excitation out as swiftly and powerfully as possible, so you just externalize it. This is difficult on its own merit, but physical conservation laws, namely parity conservation laws, also act against you when you try to perform a reset.

The Parity Problem: Why We Can’t Just Drive It#

When we drive our qubit resonator, we pump in an electric field represented by an operator related to \((\hat{b} + \hat{b}^\dagger)\) [1]. This operator is “odd,” meaning it cannot implement single-photon transitions between states of the same parity.

We start from the state \(|e, 0\rangle\) (qubit in lowest excited state, resonator in zero-photon state), which has odd parity. The state we want to end up in, \(|g, 2\rangle\) (qubit in ground state, resonator in two-photon state), has even parity. Consequently, you do not have a direct single-photon transition route you can drive.

We can, however, use the progress of 2026 physics to our advantage: we can turn Purcell filters on and off to control the spontaneous decay, and we can use Quantum Circuit Refrigerators (QCR) to control how fast photons leak out of the resonator. To solve the parity issue, the inventors [2] propose two detours to \(|g, 2\rangle\). These paths would normally interfere destructively in a harmonic oscillator, but because transmons are “anharmonious” (\(\alpha\)), these paths work in conjunction to lead to decay.

The Reset “Sprint” Sequence#

  1. Frequency Alignment: The drive frequency is set to \(\omega_d = \omega_{g2} - \omega_{ge}\), where \(\omega_{g2}\) is the frequency difference between the two-photon state and the ground state. This ensures the \(|e, 0\rangle\) and \(|g, 2\rangle\) states are effectively degenerate in the rotating frame.

  2. Coherent Injection (\(\Omega\) Drive): The signal source provides a microwave drive (\(\Omega\)) to the qubit, activating the effective coupling \(\tilde{g}\) that bridges the parity-forbidden gap.

  3. Optimal Capture (Rabi Timing): The drive is applied until the Rabi oscillation of the population transfer from \(|e, 0\rangle\) to \(|g, 2\rangle\) is at its maximum. In simulations, this occurs at approximately 800 ns.

  4. Dissipation Trigger (QCR/Readout): Once the excitation is trapped in the resonator/Purcell filter, the Quantum Circuit Refrigerator (QCR) or a readout line dissipates the photon population. The QCR uses photon-assisted electron tunneling in a SINIS junction to dump the energy into a thermal bath.

  5. Isolation (Brake): The drive is stopped and the tunable environment is deactivated (or the Purcell filter is tuned away) to prevent the environment from decohering the qubit during subsequent computation.

The Two Interference Paths#

The effective coupling \(\tilde{g}\) that enables the \(|e, 0\rangle \to |g, 2\rangle\) transition is the result of two primary quantum interference paths:

  • Path A: Drive-First: (\(|e,0\rangle \to |f,0\rangle \to |e,1\rangle \to |g,2\rangle\)). The drive \(\Omega\) excites the qubit from \(|e\rangle\) to the second excited state \(|f\rangle\), followed by swap interactions \(g\) moving excitations to the resonator.

  • Path B: Swap-First: (\(|e,0\rangle \to |g,1\rangle \to |e,1\rangle \to |g,2\rangle\)). The coupling \(g\) swaps the qubit excitation to the resonator, then the drive \(\Omega\) excites the qubit back to \(|e\rangle\), before a final swap to \(|g, 2\rangle\).

Intuition and Equations#

The effective coupling strength \(\tilde{g}\) depends on the anharmonicity (\(\alpha\)) of the qubit. If the qubit were a perfectly harmonic oscillator (\(\alpha = 0\)), these two paths would interfere destructively and cancel out perfectly. The effective coupling is expressed as the following:

\[\tilde{g} = -\frac{\Omega \alpha g^2 e^{-i\phi}}{\sqrt{2} \Delta^2 (2\Delta + \alpha)}\]

Key takeaways from the math:

  • \(\tilde{g} \propto \alpha\): No anharmonicity means no reset path.

  • \(\tilde{g} \propto \Omega\): The reset speed is linearly tunable by the drive amplitude.

  • \(\tilde{g} \propto g^2\): The transition is a second-order process in terms of the qubit-resonator coupling.

Selection Rules: The drive \(\hat{H}_d\) is an odd operator. Since \(|e, 0\rangle\) is odd and \(|g, 2\rangle\) is even, the single-photon drive is permitted to bridge them.

Outro#

Ultimately, this arrangement is a masterclass in turning a physical constraint—the transmon’s anharmonicity—into a functional tool for error budget management.

In the race to scale superconducting processors, we can’t afford to wait for qubits to spontaneously decay at their own leisure. By engineering a “back door” through the two-photon state of a resonator and flushing it with a QCR, IQM has effectively designed a high-speed plumbing system for quantum information. It’s a pragmatic, highly technical solution to the “Herculean” task of keeping the dilution fridge clean of leftover excitations.

References:

[1] Zeytinoğlu, S., et al. (2015). “Microwave-induced amplitude- and phase-tunable qubit-resonator coupling in circuit quantum electrodynamics.” Phys. Rev. A 91, 043846.

[2] IQM Finland Oy. (2025). “Arrangement for Qubit Reset.” European Patent Application EP 4 629 142 A1.

Contents