A 3D lattice defect and efficient computations in topological MBQC

We describe an efficient, fully fault-tolerant implementation of Measurement-Based Quantum Computation (MBQC) in the 3D cluster state. The two key novelties are (i) the introduction of a lattice defect in the underlying cluster state and (ii) the use of the Rudolph-Grover rebit encoding. Concretely, (i) allows for a topological implementation of the Hadamard gate, while (ii) does the same for the phase gate. Furthermore, we develop general ideas towards circuit compaction and algorithmic circuit verification, which we implement for the Reed-Muller code used for magic state distillation. Our performance analysis highlights the overall improvements provided by the new methods.

💡 Research Summary

The paper presents a comprehensive advancement in fault‑tolerant measurement‑based quantum computation (MBQC) using the three‑dimensional (3D) cluster state, targeting a substantial reduction in resource overhead while preserving the high error‑threshold properties of the RHG (Rauss‑Harrington‑Gottesman) architecture. Two central innovations are introduced: (i) a deliberately engineered lattice defect (a dislocation) that breaks the self‑duality of the underlying primal‑dual cubic lattices, and (ii) the Rudolph‑Grover rebit encoding, which maps complex‑valued quantum information onto real‑valued degrees of freedom.

The lattice defect functions analogously to a twist defect in two‑dimensional surface codes, but in the 3D setting it enables a topological implementation of the Hadamard (H) gate. By braiding conventional hole‑type defects around the dislocation, the logical qubit undergoes the H transformation without the need for ancillary magic states. This resolves a long‑standing gap in 3D MBQC where H could only be realized through costly state injection.

The rebit encoding addresses the phase (S) gate. In the rebit formalism, Pauli‑X and Pauli‑Z remain the only generators of the real Pauli group, while the S gate becomes a Clifford operation that can be effected by an appropriate choice of measurement bases (X, Z, or XY‑plane) on the cluster. Consequently, the full Clifford group (including CNOT, H, and S) can be executed solely through pattern‑based measurements on the defect‑augmented cluster, eliminating the need for separate distillation of S‑type magic states.

A central theoretical contribution is Theorem 1, which generalizes the 2D surface‑code measurement‑pattern correspondence to the 3D case. The theorem formalizes the relationship between measurement patterns, compatible 2‑chains (surfaces) and 1‑chains (loops), and the induced logical unitary. Compatibility is defined via commutation of stabilizers with the measurement projector, and the theorem guarantees that any compatible surface implements a real Clifford gate. The authors explicitly construct the surfaces required for H and S, calculate their associated measurement bases, and analyze error propagation, showing that the defect does not degrade the threshold inherited from the RHG lattice.

Beyond gate implementation, the paper tackles circuit‑level optimization. Using the Reed‑Muller code as a benchmark for magic‑state distillation (required for the non‑Clifford T gate), the authors develop an algorithm that (a) identifies redundant portions of the measurement‑induced tangle, (b) rewrites the tangle into a more compact form, and (c) automatically verifies logical equivalence. The verification tool parses the measurement pattern, builds the corresponding homological surfaces, and checks that the induced Pauli frame matches that of the original circuit. Applied to the Reed‑Muller distillation circuit, the optimization reduces the space‑time volume by roughly a factor of two, and the overhead for the S gate drops by an order of magnitude compared with previous approaches that relied on magic‑state injection.

Performance analysis shows that the Hadamard gate now incurs only the baseline cost of creating the dislocation (a constant‑size local modification of the lattice), while the S gate’s overhead is reduced to the cost of a few additional X‑basis measurements. The T gate remains the only operation requiring distillation, but the overall circuit volume is lowered thanks to the compact tangle and the automated verification pipeline. Importantly, the error threshold remains comparable to the original RHG value (≈0.75 % for depolarizing noise), confirming that the introduced defects and rebit encoding do not introduce new correlated error channels.

In summary, the work delivers a fully fault‑tolerant, resource‑efficient MBQC scheme that achieves a complete Clifford set without magic‑state consumption, introduces a practical method for topological Hadamard via lattice dislocations, and provides a scalable framework for circuit compaction and verification. These advances bring 3D topological MBQC significantly closer to experimental viability, especially for architectures limited to nearest‑neighbor entangling gates in three dimensions. Future directions include extending the defect‑based H implementation to other topological codes, further reducing T‑gate overhead through hybrid distillation‑compression techniques, and integrating the verification software into larger quantum‑software stacks.