Surface Code

Definition

The surface code is a leading quantum error-correcting code that arranges physical qubits in a two-dimensional grid pattern, using only nearest-neighbor interactions — making it particularly practical to implement on hardware platforms like superconducting qubits that naturally support 2D grid layouts.

In depth

The surface code is the leading quantum error-correcting code for superconducting qubits. It arranges data qubits on a square grid with ancilla qubits between them, each ancilla measuring a four-body parity check of its neighbors, either for X-type or Z-type errors. Only nearest-neighbor interactions on a flat two-dimensional chip are needed, which matches the layout of fabricated devices. Its error threshold is about 1 percent for realistic noise, a level that hardware has now crossed. A distance-d version uses about 2 d squared minus 1 physical qubits to encode one logical qubit.

Kitaev proposed the toric code in 1997, and Bravyi and Kitaev adapted it to a planar surface in 1998. Fowler, Mariantoni, Martinis, and Cleland laid out a practical architecture in 2012. Experiments followed: Google showed a distance-3 and distance-5 code in 2021 and 2023, and in 2024 its Willow chip showed logical error falling as the distance grew from 3 to 5 to 7, with logical lifetime exceeding the best physical qubit. Logical gates are performed by lattice surgery, which merges and splits patches of surface code, and non-Clifford gates use magic states.

The drawback is overhead. A logical error rate of 10 to the minus 9 needs a distance around 17 to 25 for typical hardware, so roughly 600 to 1,300 physical qubits per logical qubit, and a useful algorithm needs thousands of logical qubits, leading to millions of physical qubits. Higher-rate codes, such as quantum LDPC codes, aim to reduce that, at the cost of nonlocal connections. The surface code remains the baseline against which other codes are compared, because of its high threshold, simple layout, and well-developed decoders.

Technical Definition

The surface code's key advantage is a relatively high fault-tolerance threshold and the requirement for only local (nearest-neighbor) qubit interactions, at the cost of needing a large number of physical qubits per logical qubit compared to some other theoretical codes.

Visual Explanation: An Analogy

Think of the surface code like a neighborhood watch system, where each house (qubit) only needs to communicate with its immediate neighbors to collectively detect and report when something has gone wrong, rather than requiring long-distance communication across the entire neighborhood.

Key facts

Where the idea comes from

Kitaev's toric code of 1997 and the planar surface code of 1998 led to Fowler's 2012 roadmap. Google's surface-code experiments of 2021 to 2024 and similar work by others made it the standard.

Real-World Use Cases

• The error-correcting code used in Google's Willow processor's below-threshold scaling demonstration
• The most commonly referenced error-correcting code in current quantum hardware roadmaps

How it differs from related ideas

• The surface code needs only local connections but many qubits, while qLDPC codes need fewer qubits but nonlocal connections.
• Compared with concatenated codes, it has a higher threshold in practice.
• Color codes allow some gates transversally but have lower thresholds.

Frequently asked questions

Related Terms

• Logical Qubit
• Fault Tolerance
• Code Distance
• Stabilizer Code
• Quantum LDPC Code
• Threshold Theorem

Keep exploring

• Algorithm: Quantum Error-Correcting Codes (Shor Code, Surface Code)
• Hardware: Google Willow

← Back to the dictionary