A [[30,6,5]] code may allow quantum experiments to scale up faster

A Compact Encoder for the [[30,6,5]] Code Could Accelerate Quantum Experiments

A team from RIKEN’s Center for Quantum Computing, Toshiba, and lead author Hayato Goto has unveiled an efficient encoder for the [[30,6,5]] symplectic double code. The code packs six logical qubits into 30 physical qubits—a 20 % encoding rate—while preserving a distance of five, meaning it can correct up to two errors. By redesigning the circuit that prepares encoded states, the researchers reduced the required two‑qubit gates from 108 to 67, a 38 % improvement that directly translates into lower error accumulation on noisy hardware.

---

Why High‑Rate Codes Matter

Quantum error correction (QEC) is the bottleneck for scaling quantum processors. Traditional surface‑code approaches demand dozens of physical qubits for each logical qubit, quickly exhausting the modest qubit counts of today’s devices. High‑rate codes like the [[30,6,5]] code aim to flip that ratio, delivering more logical qubits per hardware qubit. The trade‑off is often higher circuit complexity or reduced fault‑tolerance. Goto’s group tackles both issues by streamlining the encoder, the first step in any QEC workflow.

---

Encoder Design and Gate Savings

The new encoder builds on a “plus‑state” preparation strategy. Starting from a stabilizer matrix \(HZ'\), the authors apply Gaussian elimination to isolate an identity block, allowing the first 12 qubits to be initialized in the \(|0\rangle\) eigenstate of \(Z\) and the remaining 18 in superposition via Hadamard gates. The crucial innovation lies in identifying overlapping “1” entries across column vectors of \(HZ'\). Where two rows share multiple ones, a single CNOT can replace several, collapsing the original 108‑gate network to just 67 CNOTs without compromising the encoded state’s stabilizer eigenvalues.

A complementary “zero‑state” encoder was also constructed using the \(H_X'\) matrix, but simulations showed the plus‑state version performed better under a realistic circuit‑level noise model. This suggests that the plus‑state encoder is more robust to the dominant error channels in current superconducting and trapped‑ion platforms.

---

Fault‑Tolerant Operations and Universality

Beyond state preparation, the paper outlines how transversal physical Hadamard and SWAP gates can implement logical Hadamard operations on the encoded qubits. Transversal gates are inherently fault‑tolerant because they act independently on each physical qubit, preventing error propagation across the code block. The authors also describe an arbitrary‑state encoder that can inject any quantum state into the code space, a prerequisite for universal quantum computation.

---

Numerical Validation

Using a circuit‑level noise model that captures depolarizing and measurement errors, the authors simulated both encoders. The plus‑state encoder maintained logical error rates comparable to the original 108‑gate design, confirming that the gate reduction does not sacrifice reliability. These results provide a concrete pathway for experimental groups to test the [[30,6,5]] code on near‑term hardware, where gate depth and qubit coherence are at a premium.

---

Outlook

If realized experimentally, the compact encoder could enable early demonstrations of fault‑tolerant logical operations with far fewer physical resources than surface‑code prototypes. The work also illustrates a broader design principle: systematic gate‑level optimization, guided by stabilizer matrix structure, can yield substantial overhead reductions for a wide class of quantum low‑density parity‑check (qLDPC) codes. As quantum processors grow beyond a few hundred qubits, such high‑rate, low‑overhead encodings will become essential for practical quantum advantage.

---

The research appears in Science Advances (doi:10.1126/sciadv.aec4142).