The Signal
PRX Quantum Paper Automates Boundaries for Planar Quantum Codes
A paper published in PRX Quantum on October 23 presents algorithms for constructing gapped boundaries and defects in two-dimensional topological stabilizer codes. Its bivariate bicycle code examples include a proposed planar family encoding
Quantum.world Team · Written with AI · fact-checked ·
PRX Quantum published a new algorithmic framework for designing boundaries in planar quantum error-correcting codes on October 23. The work by Zijian Liang, Bowen Yang, Joseph T. Iosue and Yu-An Chen tackles a practical snag for quantum low-density parity-check (qLDPC) codes: many promising constructions use periodic lattices, while hardware layouts have edges.
Simply cutting a periodic code down to a finite patch can leave boundary degrees of freedom uncontrolled, create local logical operators or change the code’s distance and the number of encoded qubits. The authors set out a systematic way to derive the boundary operators and Hamiltonians needed to avoid those failures, rather than relying on case-by-case design.
Why cutting a lattice can change the code
In a periodic code, stabilizer checks constrain the system throughout the lattice. At an open edge, some checks are truncated or removed, and new operators may commute with all the remaining bulk checks without being products of them. Those operators form a boundary gauge algebra; if left unaddressed, they can support unwanted local logical information.
“A systematic microscopic theory of gapped boundaries is therefore a key missing ingredient for converting periodic topological qLDPC codes into useful low-overhead planar codes,” the authors write. Their framework starts from a truncated bulk code, identifies boundary gauge operators and the excitations and strings they support, then constructs boundary Hamiltonians that remove local logical operators at the edge.
Turning boundary design into matrix calculations
For codes that repeat across a lattice, the researchers represent stabilizers with Laurent polynomials. Hermite and Smith normal-form calculations then reduce the search for boundary operators, anyons and string operators to finite matrix operations. The method covers qudits with prime and composite dimensions, including systems described by Zd.
Once the boundary anyons are identified, the construction selects which groups can condense at an edge and adds compatible local terms to complete the boundary Hamiltonian. The resulting gapped boundary lets the selected excitations terminate without an energy penalty. The same algebra also recovers fusion rules, topological spins and braiding statistics from boundary operators, and the paper extends the construction to domain walls and defects.
Bivariate bicycle codes expose long strings
The team tests the framework on toric and color codes, double-semion and six-semion codes, and an anomalous three-fermion code. Its bivariate bicycle (BB) code examples show why an automated procedure may matter: the bulk stabilizers can be local even when the anyon strings have unusually long translation periods. In the (3,3)-BB example, the period is 12; in the (2,−3)-BB example, it reaches 1,023 lattice steps.
For the (3,3)-BB family, the paper gives an expected planar construction on a square lattice of side 12L, with parameters ⟦288L²; 8; d = 12L − O(1)⟧. That notation describes 288L² physical qubits, eight encoded logical qubits and a distance expected to grow roughly with 12L. The distance statement depends on a bulk-string weight assumption, and the paper does not present a hardware demonstration.
A construction method, not a hardware result
The work’s contribution is a way to derive microscopic boundary rules from a code’s bulk topological data. It addresses a mathematical and architectural step between periodic qLDPC constructions and planar layouts; it does not show that the proposed code has been implemented on a quantum processor or benchmarked under device noise.
The authors also identify a limit in the current computation: they want to optimize it for stabilizers with larger weights, which they say is important for studying BB codes. For now, their concrete planar example is a proposed family with eight logical qubits and an expected distance scaling—not an experimental result.
Zijian Liang, Bowen Yang, Joseph T. Iosue and Yu-An Chen, PRX Quantum