Steady Voltage Instead of Pumping: Multiphoton Dissipation in a Superconducting Circuit ✛
Context
Quantum systems inevitably interact with their environment and lose coherence. However, dissipation can be turned into a resource: if you couple the desired mode to a rapidly decaying buffer mode, the losses start to remove entropy and stabilize the desired state. In superconducting circuits, this is done with Josephson junctions under parametric pumping, but such schemes suffer from spurious nonlinearities: self-Kerr, cross-Kerr between modes, and other unwanted terms. These effects limit the precision of quantum control, especially at high pump powers. This work fits into the research on dissipation engineering for bosonic qubits, alongside experiments on mechanical Schrödinger cats and multilayer bridges for superconducting electronics, but it differs by fundamentally forgoing microwave pumping.
Methods
The authors fabricated a superconducting circuit from tantalum on a sapphire substrate with aluminum Josephson junctions. A high-Q memory resonator (frequency 4.0429 GHz) and a buffer resonator (7.56 GHz) are galvanically connected to a SQUID, which is biased with a DC voltage through a bias tee. Tunneling of a Cooper pair across the junction releases energy 2eV, which can be tuned to match the energy difference between n memory photons and one buffer photon. To suppress leakage of memory photons into the line, a four-stage stepped filter with a stopband of 2 GHz around the memory frequency was used. The memory state is read out via a transmon qubit through a dispersive shift of 1.51 MHz and Wigner tomography.
Results
At zero voltage, the SQUID acts as a tunable inductance, but when a DC voltage is applied, spurious frequency shifts average to zero. Exchanges of 1-to-1, 2-to-1, and 4-to-1 photons between memory and buffer were experimentally confirmed. The maximum two-photon exchange rate g2 was 18.5±0.5 MHz. In the adiabatic elimination regime of the buffer, an effective two-photon dissipation rate κ2/2π = 0.51 MHz was obtained, consistent with the prediction 4g2²/κb. The self-Kerr effect of the memory does not exceed 10 kHz, which is three orders of magnitude lower than typical values for pumped circuits. Wigner tomography showed collapse of a coherent state to the manifold |0⟩, |1⟩ within 50 ns, but without negativity due to residual dephasing from spurious modes.
Implications
This work demonstrates for the first time that a DC voltage across a Josephson junction can create pure multiphoton dissipation without the spurious nonlinearities that limit parametrically pumped circuits. Unlike neighboring works that used quasiparticle tunneling or complex multilayer structures, here Cooper-pair tunneling with a DC bias provides high process selectivity. This opens the way to stabilizing cat qubits with exponential suppression of bit errors, as well as more complex dissipative error correction schemes.
Future development
Further improvement of voltage filtering and the use of Josephson DC voltage sources could eliminate residual dephasing and spurious losses. Adding two-mode pumping or a dedicated nonlinear contact would allow stabilizing the memory state and turning it into a full-fledged cat qubit. Scaling to higher orders of photon exchange, such as 6-to-1 or 8-to-1, is possible for four-legged cat codes.
Impact
The results will impact the development of superconducting quantum processors with bosonic error correction codes, as well as microwave quantum optics.
Next steps
Immediate next steps include improving the bias circuit design to reduce voltage noise and testing cat-state stabilization in a continuous regime.
Key open problems
In the field of superconducting quantum computing, the problem of achieving qubit lifetimes sufficient for useful computation remains unsolved. This work directly addresses the challenge of suppressing spurious nonlinearities in dissipative schemes for stabilizing cat qubits, which is a key obstacle to scalable quantum error correction.
The knowledge machine's view
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The authors replaced parametric pumping of a Josephson junction with a DC voltage across a SQUID, which enabled multiphoton dissipation (1-to-1, 2-to-1, and 4-to-1 exchanges) without spurious nonlinearities. Wigner tomography measurements confirmed pure two-photon dissipation at a rate of 0.51 MHz and suppression of the intrinsic Kerr effect below 10 kHz.
Where its strength is
This work is the first to demonstrate that a DC voltage across a Josephson junction can generate pure multiphoton dissipation without the spurious nonlinearities inherent in pumped schemes.
Where it could go next
- It might be interesting to see how the DC-biased approach combines with ideas from the quantum switch paper, where discrete graphene modes control superconductivity, to further suppress spurious modes.based on: Quantum Switch: How Discrete Graphene Modes Control Superconductivity arXiv:2608.16711v1
- A related avenue is the work on losses that create order: there, dissipation is used to stabilize quantum states, which resonates with your goal of stabilizing cat qubits.based on: Loss Creates Order: The Quantum Dance of Photons arXiv:2604.05422
- One might also look at the paper on a superconducting web in graphene: defects there create quantum order, and similar mechanisms could help control dissipative processes.based on: Superconducting Web: How Defects in Graphene Weave Quantum Order arXiv:2607.02267
Worth trying
- Test whether residual dephasing can be further suppressed using techniques from the paper on a quantum switch for discrete modes.
- Try stabilizing cat states with your scheme and compare with results from the paper on losses that create order.
- Consider integrating ideas from the paper on a superconducting web to improve mode isolation.
Why this may matter
The proposed method could pave the way for more robust quantum systems where dissipation is used as a resource rather than a nuisance, with the potential for exponential error suppression in cat qubits.
Nearby works in our archive
- Crackling in quantum processor silenced for secondssimilarity 0.52
- A Superconducting Bridge Between Worldssimilarity 0.51
- Quantum Batteries: Instant Charging with Squeezed Lightsimilarity 0.51
- How Rotation Warps Superconducting Currentssimilarity 0.51
- Superconducting Web: How Defects in Graphene Weave Quantum Ordersimilarity 0.50
- Loss Creates Order: The Quantum Dance of Photonssimilarity 0.50
- Quantum Switch: How Discrete Graphene Modes Control Superconductivitysimilarity 0.50
The archive is sparse around this work: the closest review is only 0.52, against a usual 0.64. It looks like few have drilled nearby.
A two-photon dissipation rate of 0.51 MHz means that a pair of photons leaves the resonator in 2 microseconds, comparable to the lifetime of a single photon in the same resonator (3.1 μs).
The idea of using dissipation to stabilize quantum states evokes the concept of a 'quantum Maxwell's demon': instead of fighting losses, you channel them, as in Greg Egan's novel 'Permutation City', where quantum states are maintained by actively controlling the environment.
Key numbers
- two-photon dissipation rate: 0.51 MHz
- maximum two-photon exchange rate: 18.5 MHz
- upper bound on self-Kerr effect: 10 kHz
- collapse time of coherent state: 50 ns
- memory frequency: 4.0429 GHz