Fidelity bounds for adiabatic gates and other quantum operations with time-dependent dissipation

This paper generalizes analytical fidelity-reduction formulae to account for time-dependent dissipation in quantum operations, revealing that flux-dependent noise and qubit-coupler hybridization significantly degrade the fidelity of adiabatic controlled-Z gates in superconducting quantum computers.

The Gist

Imagine you are trying to perform a delicate dance routine with a partner. In the world of quantum computing, this "dance" is a quantum gate, a specific operation that manipulates tiny bits of information called qubits. The goal is to move from a starting position to a perfect ending position without stumbling.

However, the room you are dancing in isn't empty; it's full of invisible, jittery air currents (noise) that try to knock you off balance. This is decoherence, and it ruins the fidelity (accuracy) of your dance.

For a long time, scientists had a perfect rulebook for calculating how much these air currents would mess up a dance, but only if the currents were constant and predictable.

The Problem: The Wind is Changing

In modern quantum computers, specifically those using superconducting circuits, engineers don't just stand still. To make the dance happen, they have to dynamically tune the frequencies of the qubits, kind of like a musician sliding their finger along a guitar string to change the pitch while playing.

The problem is that as they slide that finger (modulate the frequency), the "wind" (noise) doesn't stay the same. It gets stronger or weaker depending on exactly where the finger is. The old rulebooks broke because they couldn't handle a wind that changes speed and direction in real-time.

The Solution: A New Rulebook for Changing Winds

The authors of this paper wrote a new, more flexible rulebook. They derived a mathematical formula that can calculate the accuracy of a quantum gate even when the noise is time-dependent—meaning it changes as the operation happens.

They treated the noise not as a static wall, but as a flowing river that changes its current strength every second. Their formula sums up the "damage" done by the river at every single moment of the dance to give a total score of how well the dance went.

The Case Study: The Adiabatic CZ Gate

To test their new rulebook, they looked at a very common and important dance move called the Adiabatic Controlled-Z (CZ) gate.

• The Setup: Imagine two dancers (qubits) who need to coordinate a move, but they are separated. A third dancer (a tunable coupler) acts as a bridge. To make the move, the bridge dancer slides closer to the other two, creating a temporary connection, and then slides back.
• The Catch: As the bridge slides, it mixes with the other dancers. This is called hybridization. It's like the bridge dancer briefly borrowing some of the other dancers' energy and vice versa.
• The Discovery: The authors found that this mixing is a double-edged sword.
  • The Good: Mixing makes the dance happen faster.
  • The Bad: Because the bridge dancer is now mixed with the others, if the bridge is noisy (which it often is), that noise gets "stolen" by the main dancers.
  • The Result: Their formula showed that while you can make the gate faster by increasing the mixing, the noise sensitivity also increases. However, because the gate finishes so much faster, the total error often goes down. It's a trade-off: a shorter, slightly bumpier ride is often better than a long, smooth one if the bumps are small enough.

The Takeaway

This paper provides the essential tools for engineers to predict exactly how much error a quantum gate will have when the noise is changing.

• For Designers: It tells them where to tune their "knobs" (like flux levels) to avoid the worst noise.
• For Error Budgets: It helps them calculate exactly how much "noise budget" they have left before the computer fails.

In short, they moved from saying, "The wind is bad," to saying, "The wind is bad, but if we dance this specific way at this specific speed, we can still win." This is crucial for building the large-scale quantum computers of the future, where every tiny bit of error counts.

Technical Summary

Technical Summary: Fidelity Bounds for Adiabatic Gates with Time-Dependent Dissipation

Problem Statement
The realization of large-scale quantum computers requires precise control over fragile quantum states, yet performance is fundamentally limited by decoherence arising from the coupling between the processor and its environment. While analytical formulae for average gate fidelity exist for systems subject to static Markovian noise, these expressions fail for increasingly common tunable quantum architectures. Specifically, operations such as baseband flux gates in superconducting circuits involve dynamically modulating qubit or coupler frequencies to implement two-qubit gates (e.g., adiabatic controlled-Z or CZ gates). This modulation causes noise sensitivity, decoherence rates, and jump operators to become inherently time-dependent. Existing heuristic estimates lack a complete analytical treatment from first principles, making it difficult to rigorously quantify error budgets and optimize control signals for these dynamic operations.

Methodology
The authors generalize the framework for average gate fidelity previously developed for static noise to accommodate time-dependent dissipation.

1. Generalized Fidelity Formula: Starting from the definition of average gate fidelity, the authors derive an analytical formula for the fidelity reduction of quantum operations under weak, time-dependent decoherence. They model the system using a master equation with $N_L$ dissipative processes, each characterized by a time-dependent rate $\Gamma_k(t)$ and a time-independent Lindblad jump operator $\hat{L}_k$. By expanding the solution to the master equation in the small parameter $\bar{\Gamma}_k \tau \ll 1$ (where $\tau$ is the gate duration), they show that each process contributes independently to the fidelity reduction, with the time-dependent rates acting as kernels in a time integral rather than constant prefactors.
2. Adiabatic Evolution Analysis: For adiabatic operations, the authors exploit the factorization of the time-evolution operator into a phase operator and a frame-transformation operator. They demonstrate that the fidelity reduction integrand is independent of the dynamical and geometric phases, allowing the problem to be simplified by working in the instantaneous eigenbasis.
3. Perturbative Expansion: To obtain analytical expressions in terms of Hamiltonian parameters (coupling strengths $g$ and detunings $\Delta$), the authors employ a Schrieffer–Wolff (SW) transformation. This maps the bare eigenstates of the uncoupled Hamiltonian to the instantaneous eigenstates of the full Hamiltonian, enabling a perturbative expansion of the jump operators.
4. Case Study: The framework is applied to a concrete model of an adiabatic CZ gate using two fixed-frequency qubits coupled via a tunable coupler. The system is modeled with a bosonic Hamiltonian where the coupler frequency is modulated by an external magnetic flux. The authors derive explicit analytical expressions for the fidelity reduction rate, accounting for both relaxation ($T_1$) and pure dephasing ($T_\phi$), and including the effects of mode hybridization.

Key Results

• Analytical Fidelity Bound: The paper provides a rigorous analytical formula (Eq. 3 and Eq. 5) for the average gate fidelity under time-dependent dissipation, generalizing previous static results.
• Adiabatic CZ Gate Analysis: For the specific case of an adiabatic CZ gate, the authors derive the fidelity reduction rate (Eq. 16). The results show that the reduction is a weighted sum of decay rates, where the weights depend on the hybridization between modes (scaling as $g^2/\Delta^2$).
• Trade-off Between Coupling and Gate Time: The analysis reveals a critical trade-off: while increasing the qubit-coupler coupling strength increases the instantaneous fidelity-reduction rate (due to stronger hybridization and leakage), it simultaneously reduces the gate time (as the effective interaction scales as $g^4/\Delta^3$). The total accumulated error decreases because the gate time reduction dominates the increase in the error rate.
• Leakage Mechanisms: The study identifies that accumulated leakage out of the computational subspace in adiabatic operations is governed exclusively by pure dephasing processes, not relaxation.
• Validation: The analytical results are validated against numerical simulations. The authors show excellent agreement between the analytical SW perturbation theory and numerical integration of the full master equation in the dispersive regime. They also demonstrate that the framework remains valid beyond the dispersive regime when using the exact frame-transformation operator, whereas simple time-averaged error estimates significantly underestimate the accumulated error.

Significance and Claims
The authors claim that this work provides essential theoretical tools for evaluating error budgets and optimizing the design of quantum operations in tunable quantum-computing architectures. By capturing the interplay between time-dependent dissipation and mode hybridization, the derived formulae offer physical insights necessary to refine control strategies for dynamically tuned hardware. The paper asserts that these tools are applicable not only to quantum computing but also to quantum sensing and communication protocols affected by time-dependent dissipation. Furthermore, the work highlights the limitations of standard time-averaged error estimates, which fail to capture the nuances of hybridization in tunable systems. The authors note that their analysis is based on a Markovian description and does not currently account for non-Markovian noise, identifying this as a direction for future work.