Cyclic Quantum Codes Overview

• Cyclic quantum codes are defined by imposing cyclic structures on classical codes, stabilizer matrices, or preparation circuits to enable efficient error correction.
• They leverage algebraic techniques, such as defining-set methods and dual-containment criteria, to achieve robust CSS constructions and burst-error correction.
• Applications include entanglement-assisted schemes, mixed-alphabet and skew generalizations, and innovative stabilizer-cyclic and preparation-centric frameworks.

Cyclic quantum codes are a heterogeneous class of quantum error-correcting codes in which cyclic structure is imposed either on the classical ingredient codes, on the stabilizer generators, or on the preparation circuit. In the literature surveyed here, cyclicity appears through ideals in $\mathbb{F}_q[x]/\langle x^n-1\rangle$, through circulant or bipartite-circulant stabilizer matrices, and through translation-invariant cluster-state constructions. This algebraic regularity supports explicit dual-containment criteria, defining-set descriptions, Gray-map reductions, synchronizability, burst-error correction, locality constraints, entanglement assistance, and several ring- and Ore-theoretic generalizations (Guardia, 2017, Pereira, 2019, Hastings, 8 Sep 2025).

1. Algebraic foundations and meanings of cyclicity

A classical linear code $C\subseteq \mathbb{F}_q^n$ is cyclic when it is invariant under cyclic shift. Under the standard identification of vectors with residue classes in $\mathbb{F}_q[x]/(x^n-1)$, cyclic codes correspond to ideals and hence are generated by a monic divisor $g(x)\mid (x^n-1)$. The defining-set language is central: if $\beta$ is a primitive $n$-th root of unity, then

$$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$

and for cyclic codes $\dim C=n-|Z(C)|$. The defining set is a union of $q$-cyclotomic cosets, which also underlies BCH constructions and one-coset quantum-code constructions (Pereira, 2019, Guardia, 2017).

Duality is especially transparent in defining-set form. For Euclidean duals,

$$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$

while for cyclic codes over $C\subseteq \mathbb{F}_q^n$0,

$C\subseteq \mathbb{F}_q^n$1

These formulas make dual containment, LCD conditions, and intersection dimensions computable directly from root data (Pereira, 2019).

A recurring terminological distinction is that several papers explicitly construct quantum codes from cyclic classical ingredients without claiming that the resulting quantum code is cyclic in a strict stabilizer-theoretic sense. Other works use an intrinsically cyclic stabilizer description, such as circulant check matrices or translation-invariant preparation graphs. This distinction is explicit in the literature and resolves a common source of ambiguity (Rajpurohit et al., 8 Jun 2026, Suprijanto et al., 2021, Liang et al., 28 Jan 2025).

Sense of cyclicity	Representative description	Representative papers
Classical cyclic ingredients	Ideals in $C\subseteq \mathbb{F}_q^n$2, defining sets, cyclotomic cosets	(Guardia, 2017, Pereira, 2019)
Stabilizer-circulant cyclicity	Stabilizers generated by cyclic shifts; circulant $C\subseteq \mathbb{F}_q^n$3	(Ball et al., 2023, Liang et al., 28 Jan 2025)
Preparation-graph cyclicity	Translation-invariant bipartite CZ graph on a cyclic ring	(Hastings, 8 Sep 2025)

2. CSS and stabilizer constructions from cyclic classical codes

The standard bridge from classical cyclic codes to quantum codes is the CSS theorem. If $C\subseteq \mathbb{F}_q^n$4, then there exists a quantum code with parameters

$C\subseteq \mathbb{F}_q^n$5

and in the one-code case $C\subseteq \mathbb{F}_q^n$6, one obtains

$C\subseteq \mathbb{F}_q^n$7

This framework appears repeatedly, including in cyclic $C\subseteq \mathbb{F}_q^n$8-qLRC constructions and in asymmetric quantum cyclic codes (Rajpurohit et al., 8 Jun 2026, Aly et al., 2010).

One influential line constructs nonbinary CSS codes from a single $C\subseteq \mathbb{F}_q^n$9-cyclotomic coset containing consecutive integers. If a defining set consists of one coset $\mathbb{F}_q[x]/(x^n-1)$0 of size $\mathbb{F}_q[x]/(x^n-1)$1, and that coset contains $\mathbb{F}_q[x]/(x^n-1)$2 consecutive integers, then the BCH bound gives $\mathbb{F}_q[x]/(x^n-1)$3, while the dual-containing condition $\mathbb{F}_q[x]/(x^n-1)$4 yields a quantum code

$\mathbb{F}_q[x]/(x^n-1)$5

The significance is that BCH-designed distance is achieved with a very small defining set, preserving dimension (Guardia, 2017).

A second classical route uses maximal affine-invariant extended cyclic codes and duadic cyclic codes. Maximal affine-invariant extended cyclic codes of length $\mathbb{F}_q[x]/(x^n-1)$6 produce pure quantum codes with parameters

$\mathbb{F}_q[x]/(x^n-1)$7

and, in the prime-field case, pure codes

$\mathbb{F}_q[x]/(x^n-1)$8

For duadic codes, Hermitian dual containment yields $\mathbb{F}_q[x]/(x^n-1)$9 codes with

$g(x)\mid (x^n-1)$0

and under additional number-theoretic conditions one obtains degenerate families answering the open problem cited there (0711.2050).

The asymmetric extension keeps separate distances for phase and qubit-flip protection. If $g(x)\mid (x^n-1)$1, the paper on asymmetric quantum cyclic codes defines

$g(x)\mid (x^n-1)$2

$g(x)\mid (x^n-1)$3

and constructs AQECs $g(x)\mid (x^n-1)$4, together with asymmetric subsystem codes. Two generic cyclic methods are given: one via generator-polynomial multiplication $g(x)\mid (x^n-1)$5, and one via enlarging defining sets by cyclotomic cosets (Aly et al., 2010).

3. Entanglement assistance and locality-preserving cyclic families

Entanglement-assisted quantum codes remove the dual-containing restriction. In the cyclic setting, the key innovation is that the number of consumed ebits can be expressed directly through defining-set intersections. For Euclidean constructions, cyclic codes $g(x)\mid (x^n-1)$6 yield a QUENTA code whose quantum dimension and ebit count depend on $g(x)\mid (x^n-1)$7; in the Hermitian case, the corresponding quantity is $g(x)\mid (x^n-1)$8. If a cyclic code is LCD, the resulting EAQECC is maximal-entanglement, and MDS LCD cyclic codes yield MDS QUENTA codes (Pereira, 2019).

The paper develops explicit families from Reed–Solomon, BCH, and general cyclic codes. In particular, it derives MDS EAQECC families from Euclidean and Hermitian Reed–Solomon inputs, a BCH-based family with parameter sets stated to be new, and Hermitian cyclic families that are almost MDS or almost near-MDS. A long-length maximal-entanglement family is obtained from Hermitian LCD cyclic codes of length $g(x)\mid (x^n-1)$9, motivated there by possible relevance to hashing-bound applications (Pereira, 2019).

Locality adds another structural layer. A classical cyclic $\beta$0-LRC with $\beta$1 and $\beta$2 yields, via CSS, a quantum $\beta$3-LRC

$\beta$4

The decisive defining-set criterion is

$\beta$5

which is necessary and sufficient for dual containment of a cyclic code in that setting. Three explicit qLRC families are constructed: a length-$\beta$6 family, a $\beta$7 unbounded-length family at $\beta$8, and a more general unbounded-length family under $\beta$9. Two of these are optimal with respect to the quantum Singleton-like bound whenever the resulting qLRCs are pure, and Constructions 2 and 3 impose no bound on code length relative to field size (Rajpurohit et al., 8 Jun 2026).

4. Synchronization, burst correction, and correlated-error cyclic codes

Quantum synchronizable codes exploit cyclic structure to recover both Pauli errors and block misalignment. The general mechanism uses nested cyclic codes $n$0 with generator polynomials related by $n$1; synchronization is correctable whenever

$n$2

Quadratic-residue cyclic codes and their supercodes give families

$n$3

for Mersenne primes $n$4, with $n$5, hence maximal synchronization tolerance $n$6 (Xie et al., 2014).

Finite-geometry cyclic codes extend synchronizable constructions beyond prime and Mersenne lengths. Projective-geometry cyclic codes $n$7 and Euclidean-geometry cyclic codes $n$8 are shown to be cyclic, dual-containing in the relevant parameter ranges, and nested. Their quotient polynomial has full order equal to the code length, so the resulting QSCs attain the best possible synchronization recovery allowed by the cyclic-polynomial framework. The projective family is notable because lengths

$n$9

are generally neither prime nor Mersenne (Fujiwara et al., 2013).

Sextic cyclotomy provides another synchronization-optimal source. For $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$0 prime and $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$1, cyclic codes from sextic cyclotomic classes satisfy explicit dual relations $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$2 and $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$3, hence are dual-containing. Factor removal from the sextic-class polynomials produces nested augmented supercodes, yielding two QSC families

$$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$4

and

$$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$5

with maximal synchronization tolerance $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$6 (Wang et al., 2021).

Burst-error correction uses cyclicity differently. Quantum cyclic redundancy check codes start from a binary CRC check matrix $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$7 and form a stabilizer matrix

$$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$8

The resulting QCRC codes detect burst errors of length $$Z(C)=\{\,i\in \mathbb{Z}_n: c(\beta^i)=0\text{ for all }c(x)\in C\,\},$$9 and, under the c-property, correct burst errors of length $\dim C=n-|Z(C)|$0, thereby attaining the quantum Reiger bound $\dim C=n-|Z(C)|$1. A special family

$\dim C=n-|Z(C)|$2

admits a linear-time decoding algorithm based on quasi-cyclic interleaving (Ball et al., 2023).

5. Ring, mixed-alphabet, and skew/Ore generalizations

A substantial branch of cyclic quantum coding passes through non-field alphabets and Gray maps. Over the non-chain ring

$\dim C=n-|Z(C)|$3

cyclic codes decompose into five cyclic $\dim C=n-|Z(C)|$4-constituents $\dim C=n-|Z(C)|$5, and dual containment is equivalent to

$\dim C=n-|Z(C)|$6

The Gray image preserves duality and distance, yielding quantum codes $\dim C=n-|Z(C)|$7 (Suprijanto et al., 2021).

The ring

$\dim C=n-|Z(C)|$8

leads to a three-component theory. Every cyclic code can be written as

$\dim C=n-|Z(C)|$9

with each $q$0 binary cyclic, and the Gray map $q$1 converts cyclic ring codes into binary quasi-cyclic codes of index $q$2. Dual containment reduces to

$q$3

which yields quantum codes

$q$4

(Dertli et al., 2014).

Mixed-alphabet constructions over $q$5 provide exact self-duality criteria for cyclic codes. If

$q$6

then self-duality is characterized by

$q$7

Via the Gray map, self-orthogonal examples yield binary stabilizer codes including $q$8, $q$9, and $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$0, with the latter two reported as optimal (Aydogdu et al., 2017).

Skew and Ore generalizations broaden cyclicity from commutative $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$1 ideals to left modules over skew polynomial rings. For mixed alphabets $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$2, $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$3-skew cyclic codes admit a separable form $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$4, an exact dual-containing criterion in terms of right divisibility of $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$5 and $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$6, and Gray-image CSS quantum codes $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$7 with improved parameters (Prakash et al., 2023). More generally, $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$8-cyclic codes over the non-chain rings $$Z(C^\perp)=\mathbb{Z}_n\setminus \{-i\mid i\in Z(C)\},$$9 decompose into $C\subseteq \mathbb{F}_q^n$00-cyclic field constituents, and the Gray-image CSS construction gives $C\subseteq \mathbb{F}_q^n$01 codes, several of which improve prior tables (Prakash et al., 2024). The product-ring framework $C\subseteq \mathbb{F}_q^n$02-cyclic over $C\subseteq \mathbb{F}_q^n$03 adds Euclidean dual-containing and annihilator dual-containing CSS constructions, together with orthogonality-preserving Gray maps and MDS or almost MDS quantum codes (Akanksha et al., 3 Jan 2025).

6. Stabilizer-cyclic and preparation-centric families

Recent work includes intrinsically cyclic stabilizer families tailored to nonstandard noise models. Quantum XYZ cyclic codes are defined by cyclic permutations of the seed stabilizer

$C\subseteq \mathbb{F}_q^n$04

with code length

$C\subseteq \mathbb{F}_q^n$05

Their binary check matrices are circulant, $C\subseteq \mathbb{F}_q^n$06, and for the subfamily

$C\subseteq \mathbb{F}_q^n$07

the minimum-weight logical $C\subseteq \mathbb{F}_q^n$08 operator has weight $C\subseteq \mathbb{F}_q^n$09. The paper states, to the authors’ knowledge, that this is the only known family of quantum cyclic codes whose code distance increases with code length. Numerically, these codes exhibit code-capacity thresholds around $C\subseteq \mathbb{F}_q^n$10 for pure $C\subseteq \mathbb{F}_q^n$11, $C\subseteq \mathbb{F}_q^n$12, and $C\subseteq \mathbb{F}_q^n$13 noise, around $C\subseteq \mathbb{F}_q^n$14 for depolarizing noise, and around $C\subseteq \mathbb{F}_q^n$15 for $C\subseteq \mathbb{F}_q^n$16-biased noise with $C\subseteq \mathbb{F}_q^n$17; the paper also states that, for the same code distance, the physical-qubit overhead is much less than that of the XZZX surface code (Liang et al., 28 Jan 2025).

A different notion of cyclicity is preparation-centric. Bipartite cyclic cluster codes arrange qubits as $C\subseteq \mathbb{F}_q^n$18 with $C\subseteq \mathbb{F}_q^n$19 taken modulo $C\subseteq \mathbb{F}_q^n$20, require the CZ graph to be bipartite and invariant under simultaneous translation $C\subseteq \mathbb{F}_q^n$21, and then convert the resulting code to CSS form by applying Hadamards on one bipartition. In this framework, rotated toric codes with odd $C\subseteq \mathbb{F}_q^n$22 appear as BCC codes with parameters

$C\subseteq \mathbb{F}_q^n$23

and computer search yields further examples such as $C\subseteq \mathbb{F}_q^n$24 and $C\subseteq \mathbb{F}_q^n$25. The central claim is not stabilizer sparsity but simplicity of noiseless preparation, followed by explicit discussions of ancilla-assisted fault-tolerant preparation (Hastings, 8 Sep 2025).

Taken together, these developments show that cyclic quantum coding is not a single construction paradigm but an algebraic design space. In one direction, cyclicity organizes defining sets, duals, and nested classical codes for CSS, EAQECC, synchronizable, asymmetric, and subsystem constructions. In another, it appears directly in circulant stabilizers or translation-invariant preparation graphs. A plausible implication is that the most productive future generalizations will continue to move between these viewpoints: commutative cyclic theory, skew/Ore extensions, entanglement-assisted relaxations, and intrinsically cyclic stabilizer architectures (Pereira, 2019, Liang et al., 28 Jan 2025, Hastings, 8 Sep 2025).