Resource state generation for a multispin register in a hybrid matter-photon quantum information processor

Hybrid quantum architectures that integrate matter and photonic degrees of freedom present a promising pathway toward scalable, fault-tolerant quantum computing. This approach needs to combine well-established entangling operations between distant registers using photonic degrees of freedom with direct interactions between matter qubits within a solid-state register. The high-fidelity control of such a register, however, poses significant challenges. In this work, we address these challenges with pulsed control sequences which modulate all inter-spin interactions to preserve the nearest-neighbor couplings while eliminating unwanted long-range interactions. We derive pulse sequences, including broadband and selective gates, using composite pulse and shaped pulse techniques as well as optimal control methods. This ensures a general pulse sequence in the presence of spin-position bias, and robustness against static offset detunings, and Rabi frequency fluctuations of the control fields. The control techniques developed here apply well beyond the present setting to a broad range of physical platforms. We demonstrate the efficacy of our methods for the resource state generation for fusion-based quantum computing in four- and six-spin systems encoded in the electronic ground states of nitrogen-vacancy centers or other molecular solid-state qubits. We also outline other elements of the proposed architecture, highlighting its potential for advancing quantum computing technology.

💡 Research Summary

The paper addresses a central challenge in hybrid matter‑photon quantum information processors: the deterministic generation of high‑fidelity multipartite entangled resource states within a solid‑state spin register while preserving the desired nearest‑neighbour (NN) couplings and suppressing unwanted long‑range interactions. The authors propose a pulsed dynamical decoupling (DD) scheme that works with a global driving field, eliminating the need for individual spin addressability, which is often impractical in dense solid‑state environments such as nitrogen‑vacancy (NV) centers or molecular spin qubits.

The core idea is to interleave segments of free evolution under the natural zz‑type Hamiltonian Hzz = Σ_{i<j} g_{ij} σ_z^i σ_z^j with pairs of selective multi‑spin π‑pulses about the x‑axis, denoted R_x^π(i(k)) and its inverse R_{‑x}^π(i(k)). In each segment k, the set i(k) specifies which spins are flipped. Because flipping a single spin in a pair (i,j) changes the sign of the corresponding coupling term, the effective coupling during that segment becomes f_{ij}(k) g_{ij} where f_{ij}(k)=(-1)^{N_{ij}(k)} and N_{ij}(k) counts how many of the two spins are flipped. By choosing a sequence of m segments with appropriate i(k) and durations τ_k, the accumulated phase θ_{ij}=g_{ij} Σ_k f_{ij}(k) τ_k can be engineered to equal π for all NN pairs and zero for all non‑NN (long‑range) pairs.

Mathematically, the problem reduces to solving a linear system M·τ = α, where M encodes the modulation factors f_{ij}(k) for each segment, τ is the vector of segment durations, and α contains the target phases (π/g_{ij} for NN, 0 for others) together with the total evolution time T_c. The matrix M is constructed directly from the chosen flip patterns; its invertibility for small rings (four‑ and six‑spin) guarantees a unique non‑negative solution τ ≥ 0, ensuring physical feasibility. Importantly, the formalism does not rely on a perfect lattice; once the actual couplings g_{ij} are measured (including variations due to implantation errors), the same linear framework yields the appropriate pulse schedule, automatically compensating for inhomogeneities.

To implement the selective π‑flips, the authors develop composite pulse constructions that achieve frequency selectivity using only the global field. A selective π‑pulse is built from two shaped (π/2)_x pulses (e.g., Gaussian or DRAG‑shaped) sandwiching two composite π‑pulses with different phases, effectively flipping only the resonant spin while leaving others untouched. These pulses are further refined with optimal control techniques (GRAPE) to render them robust against static detuning offsets and Rabi‑frequency fluctuations, which are typical in experimental setups.

The methodology is demonstrated on two resource‑state generators required for fusion‑based quantum computing (FBQC): a four‑spin square ring and a six‑spin hexagonal ring. For each, the authors calculate the exact coupling matrix g_{ij} for realistic NV‑center spacings, construct the M matrix, solve for τ, and simulate the full sequence including realistic noise. The resulting state fidelities exceed 99.8 % with total gate times on the order of 10 µs. Sensitivity analysis shows that even with up to ±5 % positional disorder (affecting g_{ij} by similar amounts), the fidelity drops by less than 0.2 %, confirming the robustness of the scheme.

Beyond these specific examples, the paper discusses scalability to larger registers (≈20 spins) by exploiting spatial symmetries to keep the number of independent pulse patterns manageable. The authors argue that the same principles apply to other solid‑state platforms—silicon‑based donor spins, superconducting qubits with tunable couplers, or molecular magnets—provided the dominant interaction is of Ising‑type and a global drive can be applied.

In summary, the work delivers a versatile, experimentally realistic toolbox for engineering interaction graphs in dense spin ensembles. By combining composite‑pulse selectivity, optimal‑control robustness, and a linear‑algebraic pulse‑schedule design, it overcomes the twin obstacles of long‑range coupling suppression and spin‑position disorder. This enables deterministic preparation of the four‑ and six‑spin graph states that serve as the building blocks of FBQC, thereby advancing the feasibility of hybrid matter‑photon quantum processors toward fault‑tolerant operation.