The Search For Leakage-free Entangling Fibonacci Braiding Gates

It is an open question if there are leakage-free entangling Fibonacci braiding gates. We provide evidence to the conjecture for the negative in this paper. We also found a much simpler protocol to generate approximately leakage-free entangling Fibonacci braiding gates than existing algorithms in the literature.

💡 Research Summary

This paper investigates a fundamental question in topological quantum computing with Fibonacci anyons: whether there exist exact, leakage-free entangling two-qubit gates that can be realized purely by braiding operations. The work provides strong evidence against the existence of such ideal gates while presenting a significantly simpler protocol for generating approximate versions.

The core of the problem lies in the encoding of quantum information. A two-qubit system is encoded using six Fibonacci τ anyons with a total topological charge of 1 (vacuum). The corresponding Hilbert space is five-dimensional. The intended computational subspace (V_C) is four-dimensional, encoding the two qubits, while the remaining one-dimensional state (|NC⟩) is non-computational. A braiding gate is “leakage-free” if it perfectly preserves this computational subspace (i.e., its matrix has a norm-1 entry in the |NC⟩⟨NC| component). It is “entangling” if its restriction to V_C cannot be decomposed as a tensor product of single-qubit operations (or a swap thereof). The existence of a braid that yields a unitary gate satisfying both properties exactly has been a long-standing open question.

The paper’s first major contribution is theoretical evidence suggesting such gates do not exist. The authors systematically construct a set of braids—including the standard generators σ1, σ2, σ4, σ5, a “half-twist” braid Δ, and a pure braid Σ—all of which produce leakage-free gates. They prove that the group generated by these specific braids consists solely of non-entangling gates on the computational subspace. For example, Δ produces a gate equivalent to the SWAP gate (up to a phase), and Σ produces I ⊗ R^2. This constructive proof demonstrates an entire infinite family of leakage-free braids that fail to be entangling.

To complement this analytical result, the authors performed an exhaustive computer search enumerating braid words in B6 up to length 7, checking each for the leakage-free and entangling properties. No braid satisfying both criteria was found. The convergence of the theoretical construction and the numerical search provides compelling, though not conclusive, evidence for the conjecture that exact leakage-free entangling Fibonacci braiding gates are impossible.

Given the likely impossibility of exact gates, the paper’s second major contribution is a novel and simpler protocol for generating approximately leakage-free entangling gates. The key insight is to focus on braids that preserve a specific two-dimensional subspace V = span{|NC⟩, |ττ⟩}, rather than the entire four-dimensional computational space. The authors identify braids (like σ2σ1σ1σ2, σ4σ5σ5σ4, and σ3) whose action on this subspace V reduces to single-qubit operations generated by ρ3(σ1^2) and ρ3(σ2), which are known to be universal for SU(2). One can then approximate a desired entangling gate (e.g., CNOT) on the full computational space by using a “magical iteration” technique. This method iteratively corrects the gate’s action on V and its orthogonal complement V⊥.

This approximation protocol is praised for its geometric intuition and simplicity compared to prior algorithms like the Solovay-Kitaev approach. However, the authors note a trade-off: while potentially easier to understand and implement, their method may produce longer braid words (sequences of elementary braiding operations) to achieve a given accuracy.

In conclusion, the paper makes a strong case that the ideal goal of exact, leakage-free, entangling gates in the pure Fibonacci braiding model may be unattainable, revealing a fundamental tension between universality and perfect subspace preservation in anyonic systems. It then offers a practical and conceptually clearer path forward via approximation, advancing the toolkit for potential experimental realization of Fibonacci anyon-based quantum computation.