Random Local Stabilizer Codes in Three Dimensions without String or Self-Similar Fractal Logical Operators

Abstract: Quantum error-correcting codes (QECs) are essential components quantum computation and have deep connections to quantum phases of matter. A key obstruction to passive self-correcting QECs is the presence of string logical operators, which can generate logical errors through constant-energy-barrier processes. Haah's Codes (fracton codes) showed that three-dimensional stabilizer codes can forbid such string logical operators, but their translation-invariant structure supports self-similar fractal logical operators with a logarithmic energy barrier. We introduce the qutrit random cubic codes, a family of local qutrit Calderbank-Shor-Steane stabilizer Hamiltonians with similar cube-check structure as Haah's Code 1 but built from spatially varying stabilizers. We prove that these models retain the no-string property and numerically observe that they have properties distinct from translation-invariant fracton codes: the smallest ground-state degeneracy exponent is $k=2$ for odd $L$ and $k=4$ for even $L$; noncontractible plane-logical operators span the entire logical space; and charge-push diagnostics show that the self-similar fractal operators are absent. These results demonstrate that constrained randomness can fundamentally change the nature of stabilizer codes and improve their self-correction properties. They further point to broader families of quantum error-correcting codes and quantum phases beyond canonical topological and fracton orders.

• The paper demonstrates a novel family of 3D qutrit stabilizer codes that rigorously prohibit string and self-similar fractal logical operators.
• It employs constrained spatial randomness within a cubic lattice and adapts Haah’s deformation techniques to generalize the no-string theorem in the F3 setting.
• Numerical results reveal parity-dependent ground-state degeneracy and membrane-like operator support, suggesting enhanced partial self-correction for quantum memories.

Summary of "Random Local Stabilizer Codes in Three Dimensions without String or Self-Similar Fractal Logical Operators" (2606.19873)

Introduction and Context

This work addresses a central challenge in quantum error correction (QEC): the construction of three-dimensional (3D) local stabilizer codes that avoid low-energy logical operators—specifically, string-like or fractal operators—that permit logical errors via thermal processes with only logarithmic or even constant energy barriers. In translation-invariant fracton codes such as Haah’s Code 1 ("Haah's code"), although logical string operators are absent, self-similar fractal logical operators persist and limit self-correction to partial/no-go regimes. This paper introduces and analyzes a family of 3D qutrit (three-level system) stabilizer codes termed "Qutrit Random Cubic Codes" (QtRCC), which crucially break translation symmetry via constrained spatial randomness in the stabilizer coefficients while preserving strict locality and the basic cubic support structure.

Model Definition: Qutrit Random Cubic Codes

QtRCC generalizes Haah’s Code 1 to qutrits, allowing local stabilizer coefficients to vary spatially, subject to global and commutation constraints derived from the CSS (Calderbank-Shor-Steane) formalism. Each cube on the lattice carries a pair of local CSS stabilizer checks—Z-type and X-type—whose nonzero coefficient fields are elements of the multiplicative group $F_3^*=\{1,2\}$. Constraints from stabilizer commutativity and topology significantly restrict coefficient assignments, leading to a family of locally admissible random stabilizer Hamiltonians distinct from both translation-invariant fracton codes and previously studied engineered or welded codes.

Key technical ingredients include:

• Each lattice site hosts two qutrits.
• Each cube anchors Z and X checks on specified sets of corners.
• The commutation relations, boundary conditions, and requirement of nonzero coefficients define admissible configurations.
• The resulting code ensemble is random only within the constrained admissible class, not under an unconstrained distribution.

Rigorous Results: Absence of String Logical Operators

A principal achievement is the rigorous generalization of the no-string theorem: for all admissible random choices of stabilizer coefficients, QtRCC forbids logical operators supported on finite-width strings, even allowing for spatial disorder in coefficients. The proof closely adapts Haah's original geometric deformation/erasure argument and verifies that each commutative and algebraic move—corner erasure, good-edge erasure, exposed-edge “confusing” constraints—remains valid in the F$_3$ setting so long as all coefficients are nonzero. This ensures that any nontrivial logical operator cannot be supported on a narrow, linear segment and leads to a finite upper bound on the length of possible string segments at any fixed width.

Numerical Results: Ground-State Degeneracy and Logical Operator Geometry

Analytical and numerical studies on the periodic lattice ($L^3$) for $10 \leq L \leq 25$ reveal substantial differences from translation-invariant models:

• Ground-state degeneracy: The minimal observed ground-state degeneracy exponent exhibits only a simple parity dependence: $k=2$ for odd $L$, $k=4$ for even $L$. No arithmetic spikes or power-of-three structure appear, in contrast to the translation-invariant qutrit cubic model (QtHC), which shows strong system-size and arithmetic fluctuations.
• Logical operator support: The full logical operator space is spanned by plane-supported (membrane) logical operators; axis-aligned tube operators (of width $w < L$) are numerically absent even for the widest proper tubes, indicating that logical representatives are genuinely membrane-like.
• Minimum-weight plane logical operators: The minimal plane logical operator weight scales exactly as $L$ for odd $_3$0 and $_3$1 for even $_3$2. The exhaustive search confirms that every logical operator can be realized on a single plane without invoking more complex support.

Diagnostics: Fractal Charge Propagation

A critical diagnostic addresses the presence or absence of self-similar, fractal, power-of-three-supported logical operators. In uniform QtHC, a charge-push protocol generates such fractal operators that recur at layers $_3$3 with four charges. However, in sampled random-coefficient QtRCC instances, this recurrence is destroyed: the same charge-push protocol results in a distribution of 7–12 charges at layer $_3$4 (first nontrivial $_3$5 layer), and the fractal pattern fails to regenerate even at $_3$6, where arithmetic should force a recurrence in the uniform case. This implies the absence of regular self-similar fractal logical operators in QtRCC.

Implications and Theoretical Significance

Theoretical Implications

The results demonstrate that:

• Strictly local stabilizer codes can be constructed in 3D that are random yet, due to constraints, entirely avoid both string logicals and the canonical self-similar fractal logicals of Haah’s Code 1.
• Constrained spatial disorder can fundamentally alter the topological entanglement structure and error-correction thresholds, leading to random code families with potentially improved self-correction properties.
• The observed parity-dependent, but otherwise weak, system-size dependence of the ground-state degeneracy and the absence of self-similar fractal operators suggest a novel stabilizer code phase that is neither topological in the conventional sense nor fractonic in the strict translation-invariant sense.

Practical Implications and Future Outlook

Potential outcomes for memory properties depend on whether more irregular, non-membrane, or irregular fractal logical operators exist:

• If absent, QtRCC achieves all logical operators as membranes, with minimum area scaling $_3$7 and possibly a linear energy barrier, providing a candidate for partial self-correction in 3D local codes without fine-tuned translation symmetry.
• If some irregular sub-membrane logical operators exist, they are not captured by plane or tube searches and would require global code-distance or novel analytic methods to detect, possibly leading to a new regime of "disorder-induced" fractonicity.

Generalizing beyond [111]-line-symmetric subfamilies and studying the most generic admissible models remain pressing computational and analytic challenges. Other promising directions include extending to higher-dimensional systems, more exotic lattice topologies, and prime-qudit generalizations. Classification of the spatially nonuniform stabilizer phases invoked by QtRCC, which are inaccessible to polynomial/Laurent polynomial methods and translation-invariant order parameters, is an open theoretical frontier.

Conclusion

This work rigorously establishes and numerically characterizes a new family of 3D local qutrit stabilizer codes that break translation symmetry via controlled randomness, yet strictly forbid both string and canonical self-similar fractal logical operators. The results clarify which aspects of fracton order and topological encoding are robust to constrained randomness and identify a promising path toward local code designs with enhanced self-correcting capability. Future research should pursue the classification, memory lifetimes, and dynamical properties of such random local code families, as well as their potential for practical quantum memory and quantum phase engineering.