Deterministic Generation of Arbitrary Fock States via Resonant Subspace Engineering

1. Problem Statement

The deterministic preparation of high-excitation Fock states (number states) and their arbitrary superpositions is a fundamental challenge in bosonic quantum information processing.

• Scalability Issues: Existing methods face severe scalability limitations. Measurement-based schemes suffer from exponentially decreasing success probabilities as photon numbers increase. Deterministic protocols based on full Hilbert space control suffer from "explosively growing" control complexity as the dimension of the Hilbert space expands.
• Complexity: While universal control tools exist (e.g., SNAP gates), preparing specific large-$n$ superpositions typically requires non-trivial numerical compilation or optimization. The experimental overhead (gate depth and time) grows rapidly with the mean photon number and the number of components in the superposition.
• Lack of Analytical Framework: There is no systematic mechanism that reduces quantum control complexity at the Hilbert space level while maintaining analytical solvability, high efficiency, and experimental feasibility for high-energy states.

2. Methodology: Resonant Subspace Engineering (RSE)

The authors propose Resonant Subspace Engineering (RSE), a protocol that analytically confines infinite-dimensional bosonic dynamics to a low-dimensional invariant subspace.

Core Concept

Instead of engineering complex transition paths across the entire Hilbert space (as in GRAPE or standard optimal control), RSE reduces the dynamics to a two-dimensional invariant subspace spanned by:

1. An initial reference state (typically a coherent state $|\alpha\rangle$).
2. The target state (a Fock state $|n\rangle$ or a superposition $|\psi\rangle$).

Theoretical Framework

• Generalized Oracle Unitaries (GOOs): The protocol utilizes two types of operations:
  • $O(|\alpha\rangle, \phi_0) = e^{-i\phi_0|\alpha\rangle\langle\alpha|}$: Imprints a phase on the reference coherent state.
  • $O(|\psi\rangle, \phi_1) = e^{-i\phi_1|\psi\rangle\langle\psi|}$: Imprints a phase on the target state.
• Effective Hamiltonian: By alternating these GOOs in small steps (Trotter-Suzuki decomposition), the system evolves under an effective Hamiltonian:
  $$H \propto \phi_1 |\psi\rangle\langle\psi| + \phi_0 |\alpha\rangle\langle\alpha|$$
  Crucially, this Hamiltonian acts non-trivially only within the 2D subspace $S = \text{span}\{|\psi\rangle, |\alpha\rangle\}$, leaving the orthogonal complement dark.
• Geometric Interpretation: The dynamics are mapped to a synthetic Bloch sphere. State transfer becomes a geodesic rotation from $|\alpha\rangle$ to $|\psi\rangle$.
• Conditions for Deterministic Transfer: To achieve perfect state transfer along the geodesic, two conditions must be met:
  1. Resonance Condition: The diagonal elements of the effective Hamiltonian in the subspace basis must be equal ($H_{11} = H_{22}$), eliminating detuning ($\sigma_z$ component).
  2. Phase-Matching Condition: The phase of the off-diagonal coupling ($H_{12}$) must match the relative phase between the initial and target states to ensure the rotation axis lies in the XY-plane.

Physical Implementation

The protocol is designed for circuit QED systems (bosonic cavity coupled to an ancillary qubit):

• Dispersive Regime: The qubit frequency shifts based on photon number ($n$), allowing frequency-selective driving to implement SNAP gates (selective number-dependent arbitrary phase).
• Operations: The continuous evolution is compiled into a discrete sequence of Displacement ($D(\alpha)$) and SNAP gates.
• Scalability: The method generalizes to $K$-component superpositions by expanding the invariant subspace to dimension $K+1$, enabling full $SU(K+1)$ controllability.

3. Key Contributions

• Dimensionality Reduction: A novel framework that reduces infinite-dimensional bosonic control to a 2D (or low-dimensional) problem, bypassing the exponential scaling of Hilbert space complexity.
• Analytical Solvability: Unlike numerical optimization methods (e.g., GRAPE), RSE provides closed-form analytical expressions for the required phases and evolution times.
• Sublinear Scaling: The protocol achieves a fundamental improvement in scaling. For preparing a single Fock state $|n\rangle$ from a coherent state $|\alpha\rangle$ (where $\alpha \approx \sqrt{n}$):
  • Evolution Time ($T$): Scales as $O(n^{1/4})$.
  • Gate Depth (Iterations): Also scales as $O(n^{1/4})$.
• Generalization: The method extends naturally to arbitrary superpositions of Fock states and logical operations on bosonic error-correcting codes (e.g., Cat codes) without re-optimization when parameters change.

4. Results

The authors validated the protocol through numerical simulations:

• Single Fock State Generation:
  • Successfully prepared $|n=100\rangle$ from $|\alpha=10\rangle$ with high fidelity.
  • Demonstrated that mismatched resonance conditions ($H_{11} \neq H_{22}$) lead to failed transfer and lower fidelity, confirming the necessity of the derived conditions.
  • Showed that for $n=100$, only 4 iterations of GOOs are required; for $n=380$, only 5 iterations are needed.
• Arbitrary Superpositions:
  • Simulated the preparation of complex superpositions (e.g., $|\phi_1\rangle = \sqrt{0.3}|70\rangle + \sqrt{0.7}|100\rangle$) from a coherent state.
  • Achieved unit fidelity ($F=1$) with very few iterations ($N=3$ to $5$) for states spanning photon numbers 70–100.
• Scaling Verification: The simulation results confirmed the sublinear $O(n^{1/4})$ scaling for both continuous evolution time and discrete iteration counts, significantly outperforming traditional methods.

5. Significance and Impact

• Scalable Bosonic Engineering: RSE provides a scalable, experimentally feasible pathway to generate large-scale non-Gaussian quantum resources, which are essential for quantum advantage in bosonic systems.
• Error Correction & Networking: The ability to deterministically prepare large Fock states and their superpositions is critical for:
  • Bosonic Quantum Error Correction: Generating logical states for Cat codes and other bosonic codes.
  • Quantum Metrology: Creating high-photon-number states for Heisenberg-limited sensing.
  • Multimode Entanglement: The framework extends to multimode architectures, enabling the deterministic generation of entangled states (e.g., NOON states, multimode Cat states) for distributed quantum sensing and networks.
• Experimental Feasibility: By relying only on standard displacement and SNAP gates (already demonstrated in superconducting circuits), the protocol avoids the need for exotic hardware, making it immediately applicable to current quantum computing platforms.

In summary, this work introduces a paradigm shift in quantum state engineering, moving from brute-force Hilbert space control to resonant subspace confinement, offering a mathematically transparent and highly efficient solution for generating complex bosonic quantum states.