⚛️ quantum physics

Fast Bosonic Control via Multiphoton Qubit-Oscillator Interactions

This paper proposes and validates an $n$ -photon Law-Eberly protocol and numerical optimal control strategies that utilize multiphoton qubit-oscillator interactions to significantly accelerate the preparation of rotationally symmetric bosonic states, thereby enhancing the performance and fault tolerance of bosonic quantum error correction codes on superconducting hardware.

Original authors: Noah Gorgichuk, Mohammad Ayyash, Matteo Mariantoni, Sahel Ashhab

Published 2026-09-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Noah Gorgichuk, Mohammad Ayyash, Matteo Mariantoni, Sahel Ashhab

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Quantum computers promise to solve problems that are currently impossible for classical machines, but they are notoriously fragile. The information they store, known as quantum bits or qubits, can easily be scrambled by the slightest touch of heat or electromagnetic noise. To protect this delicate information, scientists have developed a strategy called bosonic quantum error correction. Instead of relying on a single, tiny particle, this approach uses a harmonic oscillator—a system that vibrates like a spring or a pendulum—to store data. Because these oscillators can vibrate with many different amounts of energy at once, they offer a vast, infinite space to hide information. The trick is to arrange this information in a specific, symmetrical pattern so that if a small error occurs, the system can detect it and fix itself without losing the data. However, creating these specific, symmetrical patterns of energy has been a slow and difficult process, especially when trying to build these systems on flat, chip-like surfaces rather than in large, three-dimensional metal boxes.

A team of researchers has now discovered a way to speed up this creation process dramatically. In their work, they proposed and tested a new method for preparing these special quantum states using a technique that involves moving bundles of energy packets, rather than adding them one by one. Imagine trying to fill a large bucket with water. The old method was like using a tiny spoon to scoop water from a source and drop it into the bucket, one spoonful at a time. It was precise, but it took a very long time to fill the bucket before the water could evaporate or leak away. The new method is like using a large ladle to scoop up a whole bucketful of water and pour it in at once. By moving these larger groups of energy simultaneously, the researchers found they could prepare the necessary quantum states in a fraction of the time it previously took.

The researchers focused on a specific type of interaction between a qubit and an oscillator. In standard setups, these two components talk to each other by exchanging single units of energy. The new approach uses a more complex interaction where they exchange multiple units of energy at the same time. This is not just a theoretical idea; the team showed that by using these multi-energy exchanges, they could generate important quantum states, such as "cat states" and "binomial codes," much faster. These states are the building blocks for error-correcting codes, which are essential for making quantum computers reliable. The team calculated that for certain complex states, this new method could reduce the preparation time by more than half compared to the best existing techniques. For example, preparing a specific type of cat state that required 110 nanoseconds with the old method took only 26 nanoseconds with the new multi-energy approach.

To ensure their findings were not just mathematical exercises, the researchers simulated the process using realistic parameters for superconducting circuits, which are the flat, chip-based hardware currently used to build quantum computers. They included the effects of noise and energy loss that happen in real-world devices. Even with these imperfections, the simulations showed that the new method remained robust and could produce high-quality states. The team also demonstrated that by combining different types of these multi-energy interactions, they could gain control over the entire range of possible states in the oscillator, not just the symmetrical ones. This flexibility is crucial for building more advanced quantum computers that can perform a wider variety of tasks.

The study also addressed the limitations of this approach. While moving large bundles of energy is faster, it becomes technically difficult to control when the bundles get too large or the process gets too complex. The researchers found that there is a practical limit to how many energy units can be moved at once before the control becomes too fast to manage with current technology. They determined that for the most common types of error-correcting codes, the new method works exceptionally well, but for the most extreme cases involving very high energy levels, the speed advantage might be offset by the difficulty of switching the interactions on and off quickly enough. Despite these constraints, the results suggest that this multi-energy control strategy is a powerful tool for the future.

By significantly shortening the time required to set up these protective quantum states, this work helps overcome a major bottleneck in building scalable quantum computers. The transition from large, three-dimensional cavities to flat, planar circuits is a necessary step for mass-producing quantum computers, but these flat circuits currently have shorter lifespans for their quantum states. The ability to prepare states faster means the system can finish its work before the information degrades. The researchers concluded that their protocol offers a viable path to improving the performance of bosonic codes on planar superconducting hardware, bringing the goal of a large-scale, fault-tolerant quantum computer closer to reality.

Technical Summary: Fast Bosonic Control via Multiphoton Qubit-Oscillator Interactions

Problem Statement
Bosonic quantum error correction (QEC) leverages the infinite-dimensional Hilbert space of harmonic oscillators to protect quantum information, utilizing codes such as cat, binomial, and Gottesman-Kitaev-Preskill (GKP) states. These codes often rely on rotational symmetries (invariance under specific rotations) to detect and correct photon loss. However, a significant bottleneck in scaling these codes on planar superconducting hardware is the long state preparation time required to populate the oscillator with the necessary number of photons. Current state-of-the-art demonstrations primarily rely on linear qubit-oscillator interactions (e.g., the Law-Eberly protocol) or conditional displacement schemes. Linear interactions populate the oscillator one photon at a time, leading to preparation times that scale poorly with the required Fock cutoff. While planar resonators offer scalability, their coherence times are currently lower than those of 3D cavities, necessitating control methods that significantly shorten state preparation and gate operation times to remain within the coherence window.

Methodology
The authors propose a multiphoton control framework that generalizes the Law-Eberly (LE) protocol. Instead of single-photon Jaynes-Cummings (JC) interactions, the protocol utilizes $n$ -photon JC ( $n$ JC) interactions between an auxiliary qubit and the oscillator, combined with qubit drives.

The $n$ -photon Law-Eberly ( $n$ LE) Protocol:
- This protocol generates arbitrary states with exclusive Fock support on the subspace $\mathcal{H}_{0,n}$ (states $|0\rangle, |n\rangle, |2n\rangle, \dots$ ).
- It operates by alternating between a qubit drive (rotating the qubit) and an $n$ JC interaction (swapping excitations between the qubit and the oscillator in bundles of $n$ photons).
- The authors derive an analytic inversion algorithm to determine the necessary coupling strengths and pulse durations for each step to synthesize a target state, ensuring that at the end of each step, the highest occupied Fock state is linked only to the qubit ground state.
- The protocol is extended to rotationally symmetric states (e.g., cat codes) which naturally reside in these subspaces.
Arbitrary State Preparation (Fine-Tune-Then-Populate - FTP):
- To synthesize arbitrary states that do not possess a single rotational symmetry (spanning multiple subspaces $\mathcal{H}_{k,n}$ ), the authors introduce the FTP protocol.
- This involves a "fine-tuning" phase to prepare a base state (superposition of the lowest $n$ Fock states) and a "populate" phase.
- The populate phase utilizes selective qubit rotations (conditional on specific Fock states via dispersive shifts) combined with $n$ JC interactions. This allows the system to climb specific "columns" in the Hilbert space (subspaces $\mathcal{H}_{k,n}$ ) independently, avoiding the incommensurate swap problems that arise when trying to swap multiple Fock states simultaneously with a single interaction order.
Optimization and Generalization:
- The authors perform numerical optimal control theory (OCT) calculations using the GRAPE algorithm to verify if continuous pulse optimization can further reduce times compared to the step-wise $n$ LE protocol.
- The protocols are generalized to multiple oscillators coupled to a single qubit, allowing for the preparation of rotationally symmetric multi-oscillator states.

Key Results

Preparation Time Reduction: The $n$ $n$ LE protocol substantially reduces state preparation times compared to linear ( $n=1$ $n = 1$ ) interactions. For example:
- Two-component cat states ( $\alpha=\sqrt{2}$ ): Preparation time reduced from 110 ns ( $n=1$ ) to 26 ns ( $n=2$ ), a 76% improvement.
- Four-component cat states: Reduced from 89 ns to 46 ns (48% improvement).
- GKP states (approximate logical zero): Reduced from 384 ns to 200 ns (48% improvement) for a specific parameter set.
Scaling Behavior: While higher-order interactions ( $n>1$ ) typically have weaker coupling strengths ( $g_n < g_{n-1}$ ), the reduction in the number of required steps and the scaling of the multiphoton Rabi swap frequency ( $\propto \sqrt{(jn)!/((j-1)n)!}$ ) result in faster overall preparation for states with significant Fock support.
Optimal Control: Numerical OCT calculations generally recover the $n$ LE protocol times or offer only marginal improvements (e.g., ~15% speedup for specific small states like $(|0\rangle+|2\rangle)/\sqrt{2}$ ), suggesting the analytic $n$ LE protocol is near-optimal for the alternating pulse structure.
Comparison to Conditional Displacement: The multiphoton sideband control schemes are significantly faster than conditional displacement (CD) schemes (e.g., Echoed-Conditional Displacement) for the states tested. CD times are largely independent of energy levels, whereas sideband swaps become faster at higher photon numbers.
Robustness: Simulations using realistic planar superconducting circuit parameters (including spurious terms, relaxation, and dephasing) show that the protocol maintains high fidelity (e.g., $F \approx 0.98$ for a four-component cat state) despite decoherence, though fidelity degrades as the cat size (circuit depth) increases.

Significance and Claims
The paper claims that the proposed multiphoton control protocols offer a viable pathway to implementing bosonic quantum error correction on planar superconducting hardware, which is currently limited by coherence times. By drastically reducing state preparation times, the protocols help mitigate the impact of decoherence, potentially enabling a "break-even" demonstration of bosonic QEC on planar circuits.

The authors emphasize that while the protocols offer substantial speedups, they are subject to practical limitations:

Gate Switching Speed: As the Fock cutoff increases, the required gate times become extremely short (e.g., ~1 ns), making adiabatic switching of interactions difficult and potentially introducing diabatic errors. This sets a rough cutoff (around 20 photons) for the current parameter regimes.
Rotating Wave Approximation (RWA): The validity of the effective Hamiltonian relies on the RWA, which degrades for large Fock numbers and long evolution times due to counter-rotating terms.
Selective Rotations: The dispersive regime required for selective qubit rotations limits the drive strength and becomes spectrally challenging for high Fock numbers.

The work concludes that these protocols are a critical tool for near-term improvements in bosonic code implementations, specifically facilitating the transition from 3D cavities to scalable planar architectures.

More like this