Bacon-Shor Code
Understanding the Bacon-Shor Code
In quantum computing, physical qubits are highly susceptible to environmental noise and decoherence. To build fault-tolerant quantum computers, quantum error correction (QEC) is required to group multiple fragile physical qubits into a single, highly stable logical qubit. Standard stabilizer codes, such as the Shor 9-qubit code or topological Surface codes, require measuring multi-qubit stabilizer generators. In hardware, measuring operators of weight 4 or higher (involving four or more qubits simultaneously) is technically challenging and introduces additional channels for error propagation.
The Bacon-Shor code resolves this challenge by operating as a subsystem code. Rather than directly measuring stabilizer generators, it divides the stabilizer group into a gauge group. Error information is obtained by measuring weight-2 gauge operators: XX operators between horizontally adjacent qubits and ZZ operators between vertically adjacent qubits. The products of these gauge measurements are then used to reconstruct the stabilizer generators (which are weight-2n operators spanning entire rows or columns).
Because only 2-qubit measurements are required, the control routing and RF readout circuitry on the quantum processor are simplified. In superconducting quantum systems, where microwave pulses are routed through coaxial lines to read qubit states, lowering the weight of measured operators reduces RF cross-talk, thermal dissipation at cryogenic temperatures, and gate complexity. However, this simplicity comes at a cost. The Bacon-Shor code has a lower threshold compared to the Surface code and requires n² physical qubits to achieve a distance of n, resulting in a less efficient physical-to-logical qubit overhead scaling.
Key Equations
n² physical qubits → 1 logical qubit (with distance d = n)
Weight-2 Gauge Operators:
X-type (horizontal pairs): gi,jX = Xi,j Xi,j+1
Z-type (vertical pairs): gi,jZ = Zi,j Zi+1,j
Row/Column Stabilizer Construction:
Row Stabilizers: SiX = ∏j=1n Xi,j Xi+1,j
Column Stabilizers: SjZ = ∏i=1n Zi,j Zi,j+1
Comparison of QEC Codes (d=3)
| Code Type | Physical Qubits (d=3) | Stabilizer Weight | Readout Complexity | Error Threshold |
|---|---|---|---|---|
| Bacon-Shor Code | 9 (3×3 grid) | Weight 2 (Gauge operators measured) | Low (only 2-qubit interactions needed) | ~0.1% to 0.5% |
| Surface Code | 17 (Rotated Surface Code) | Weight 4 (stabilizers measured directly) | High (requires 4-qubit ancilla coupling) | ~1.0% |
| Shor 9-Qubit Code | 9 | Weight 6 (X phase flip) & Weight 2 (Z bit flip) | Medium (high weight stabilizers required) | Very low (non-fault-tolerant) |
Frequently Asked Questions
What is a quantum subsystem code like the Bacon-Shor code?
A quantum subsystem code is a generalization of stabilizer codes where the logical information is stored in a subsystem of the overall state space. In these codes, the physical qubits are partitioned into logical qubits, gauge qubits, and syndrome space. The presence of gauge qubits means that certain degrees of freedom do not affect the stored logical information. This allows engineers to measure lower-weight gauge operators (such as 2-qubit operators in the Bacon-Shor code) instead of higher-weight stabilizer generators. The stabilizers are subsequently computed by multiplying the gauge measurement results, simplifying physical implementation.
Why is the Bacon-Shor code attractive for superconducting quantum processors?
Superconducting quantum processors rely on microwave lines and resonators to control and read out qubit states. Measuring multi-qubit stabilizer generators (such as weight-4 stabilizers in Surface codes) requires complex coupling networks, extra ancilla qubits, and sophisticated RF routing, which increases cross-talk and cryogenic heat load. The Bacon-Shor code requires only weight-2 (2-qubit) gauge measurements. This significantly simplifies the RF routing on the quantum chip, minimizes the number of control channels, and reduces the likelihood of measurement-induced errors, making it highly compatible with planar superconducting qubit architectures.
What are the primary disadvantages of the Bacon-Shor code compared to Surface codes?
The main disadvantage of the Bacon-Shor code is its poor threshold scaling under local noise models. Unlike Surface codes, which exhibit an error threshold near 1% under circuit noise, the Bacon-Shor code threshold is typically lower (around 0.1% to 0.5%) and does not behave asymptotically as well in 2D architectures without active stabilization. Additionally, the physical-to-logical qubit overhead is relatively high, requiring n² physical qubits to achieve a code distance of n, whereas modern rotated Surface codes require fewer physical qubits for equivalent distances.