Holstein-Primakoff Spin Codes

• Holstein-Primakoff spin codes are a framework that maps SU(2) spin operators to bosonic modes using polynomial expansions, ensuring accurate simulation in the physical subspace.
• They enable the import of bosonic error-correcting codes—like cat, binomial, and GKP codes—into symmetric spin ensembles, thus enhancing robustness against both collective and local noise.
• The approach facilitates efficient block-diagonalization of spin Hamiltonians and scalable implementations in circuit-QED, offering practical pathways for quantum simulation and fault-tolerant computing.

Holstein-Primakoff (HP) spin codes are a framework for mapping spin systems, especially those involving ensembles of spin-½ particles or larger SU(2) representations, to bosonic systems using the Holstein-Primakoff transformation. This mapping underpins advanced quantum simulation and error-correction protocols across circuit-QED, quantum computing hardware, and analytical diagonalisations. HP spin codes exploit the structure of the symmetric (or collective) spin subspace and facilitate the import of continuous-variable (CV) bosonic codes, such as cat, binomial, and GKP codes, into spin ensembles. Recent developments provide polynomial or even finite support (Hermitian) operator representations, noise robustness under both collective and local models, and efficient numerical implementations with drastically improved scaling.

1. The Holstein-Primakoff Transformation and Polynomial Expansions

The HP transformation encodes SU(2) spin operators in terms of bosonic creation and annihilation operators, $a^\dagger,a$, and number operators, $\hat n$, via: $$S^+ = a^\dagger \sqrt{2S - \hat n}, \qquad S^- = \sqrt{2S - \hat n}\, a, \qquad S^z = S - \hat n$$ The transformation preserves commutation relations $[S^+,S^-]=2S^z$, $[S^z,S^\pm]=\pm S^\pm$ on the physical subspace $\hat n = 0\ldots 2S$.

For finite spin S, particularly $S=½$, the square-root expressions can be polynomially expanded or exactly represented as finite polynomials on the physical Hilbert space. For $S=½$, the polynomially expanded HP mapping takes the form (Dudinets et al., 4 Jul 2025): $$S^+ \rightarrow a^\dagger (1 - n),\quad S^- \rightarrow (1 - n) a,\quad S^z \rightarrow n - \tfrac12$$ This representation is exact provided the bosonic dynamics remain restricted to $\ket{0}$ and $\ket{1}$ at each site, i.e., hard-core bosons. For arbitrary S, a differential equation method yields a unique polynomial of degree $2S$ exactly reproducing the square-root's matrix elements on the physical subspace, thus preserving the SU(2) algebra without leakage terms (Vogl et al., 2020).

The manifestly Hermitian finite polynomial HP mapping overcomes the lack of Hermiticity in mappings like Dyson-Maleev, ensuring compatibility with variational or mean-field approaches. Polynomial truncations also localize support and facilitate error detection for population leakage out of the physical spin code space (Vogl et al., 2020).

2. HP Spin Codes: Construction and Code Properties

HP spin codes leverage the HP mapping to transfer the structure of bosonic codes to the symmetric subspace of a spin ensemble. For a permutation-symmetric ensemble of $N$ spin-½ (the Dicke subspace), the mapping

$$\ket{n}\mapsto\ket{J_{\max}, M=J_{\max} - n},\qquad J_{\max}=N/2$$

establishes an isomorphism between bosonic Fock codewords and collective spin states (Omanakuttan et al., 24 Jan 2026).

This correspondence enables:

• Importing binomial codes, where logical Fock basis codewords map onto Dicke states of definite $M$.
• Mapping cat and GKP codes, where bosonic displacement- and grid-based superpositions correspond to finite-angle SU(2) rotations and superpositions of spin-coherent states on the Bloch sphere.

The theoretical justification relies on the HP approximation, which is accurate for large $N$ and $\langle \hat n \rangle \ll N$, i.e., small-angle excursions away from the "north pole." In this regime, $J_x \approx \sqrt{J}\,q$, $J_y \approx \sqrt{J}\,p$ and quadratic noise processes map faithfully.

A central property is that HP spin codes, constructed from bosonic KL-correctable codes, are robust against both collective (permutation-invariant) and local spin noise, up to $O(1/N)$ corrections (Omanakuttan et al., 24 Jan 2026). The code structure enforces self-similarity of codeword features (grid, lobe, fringe) across the irreducible representations populated by noise-induced transitions.

3. HP-Based Hamiltonians, Simulation, and Circuit-QED Implementation

The HP transformation maps spin chain Hamiltonians, such as the Heisenberg model,

$$H_\text{spin} = -\sum_{j<k} J_{jk} \left[\tfrac12(S_j^+ S_k^- + S_j^- S_k^+) + S_j^z S_k^z\right] + \sum_j h_j S_j^z,$$

to an extended Bose-Hubbard (EBH) Hamiltonian via the HP mapping: $$\begin{aligned} H_\text{EBH} &= -\sum_{j<k} \frac{J_{jk}}{2} [a_j^\dagger a_k - a_j^\dagger (n_j + n_k) a_k + h.c.] \ &\quad -\sum_{j<k} J_{jk} (n_j - ½)(n_k - ½) + \sum_j h_j(n_j - ½) \end{aligned}$$ This mapping, and its hard-core bosonic character, enables analog realization in circuit-QED platforms, where each spin is encoded in a nonlinear microwave oscillator (Josephson junction plus capacitor), with nearest-neighbor couplings engineered via inductive and capacitive links (Dudinets et al., 4 Jul 2025).

The experimental design directly matches HP boson parameters (hopping, cross-Kerr, occupation-dependent terms) to the Josephson-junction-array Hamiltonian in the rotating-wave approximation, under the conditions that ensure the isolation of $|0\rangle$ and $|1\rangle$ Fock states per mode. The resulting simulator can access spin dynamics, entanglement measures, and can be scaled up by increasing the number of coupled resonators.

4. Noise Robustness and Measurement-Free Local Error Recovery

HP spin codes are robust to both collective (permutation-invariant) and local spin noise channels. For collective errors generated by $\{\hat J_x, \hat J_y, \hat J_z\}$, the HP-mapped code inherits the same Knill-Laflamme protection as the original bosonic code (Omanakuttan et al., 24 Jan 2026). For local errors (e.g., single-spin depolarizing noise), the action on the symmetric subspace can be expressed via projection into collective observables, with errors mapping into lower $J$ irreducible representations of the collective spin algebra.

A key feature is the availability of measurement-free local-error-recovery (MFLER). Since local errors predominantly move the population among symmetric and neighboring $J$-subspaces without damaging the internal code structure, a two-stage collective SWAP protocol using entangling gates can transfer the erroneous subspaces back into a fresh symmetric ancilla, thereby converting local noise into collective noise that can be corrected using standard HP code recovery routines. The logical error rate following MFLER scales as $O(\gamma t/N)$ for local depolarizing noise at rate $\gamma$—a substantial suppression in large $N$ ensembles (Omanakuttan et al., 24 Jan 2026).

5. Computational Methods and Block-Diagonalization

HP spin codes also enable efficient numerical diagonalization of isotropic multispin Hamiltonians. By mapping $N$ spins (possibly of various lengths $s_i$) to HP boson modes and introducing an index compression map,

$$I = \sum_{i=1}^N \nu_i \prod_{j=i+1}^N (2s_j + 1),\quad \nu_i = s_i - m_i,$$

the many-body basis is classified by total boson number. Since the mapped Hamiltonian commutes with total boson number, it is block-diagonal in the occupation-number basis, and each block corresponds to a fixed value of $n=\sum_i \nu_i$. The computational cost is dominated by the largest block, which can be polynomial in $N$ and $s$ for higher spins or constraints, substantially mitigating the exponential scaling of full Hilbert space diagonalization (Gyamfi et al., 2018).

The method generalizes to arbitrary spin graphs with pairwise interactions and can be expressed as efficient pseudocode grouping basis vectors by compressed index and occupation number, constructing block matrices, and diagonalizing each block independently.

6. Practical Constraints, Limitations, and Error Mechanisms

HP mappings and spin codes are exact only within the physical subspace (occupation up to $2S$ per site). Leakage to unphysical Fock states or errors populating $n_j>1$ in spin-½ systems introduce systematic deviations in observables (Dudinets et al., 4 Jul 2025). Ensuring strict population constraints requires hardware (finite anharmonicity, low photon loss) and initialization (preparation in $|0\rangle,|1\rangle$ only).

Truncation errors, non-RWA terms, and population outside HP regime (e.g., for GHZ-type or widely delocalized states) further limit code performance. MFLER is most efficient when leakage among $J$-subspaces is $O(1)$ and the codewords are localized near the north pole; otherwise, the collective gates fail to act uniformly on all populated irreps (Omanakuttan et al., 24 Jan 2026).

7. Significance and Applications

HP spin codes unify continuous-variable and spin-based quantum error correction, providing a route for importing GKP, cat, and binomial code constructions into qubit ensembles with collective interactions. They yield protocols and physical implementations compatible with both classical simulation and near-term hardware (circuit-QED, spin ensembles) (Dudinets et al., 4 Jul 2025, Omanakuttan et al., 24 Jan 2026). Their mathematical framework supports scalable simulation, analytical block-diagonalization, and operator synthesis with manifest Hermiticity and finite support, facilitating both theoretical work and experimental realization. These advances contribute substantially to the program of scalable, fault-tolerant quantum computation in architectures governed by collective or bosonic dynamics.