ABSTRACT Measurement‐based quantum computing relies on the preparation of large‐scale multipartite entanglement, typically realized through cluster states. Quantum frequency combs provide a naturally multimode platform for constructing such states, yet prior pulse‐shaping methods are inherently sequential, thereby hindering scalability and making the reconstructed state vulnerable to system drifts. Here, we address this limitation by developing a multipixel homodyne detection scheme that enables the parallel characterization of quantum noise across multiple frequency modes and the single‐step reconstruction of the full covariance matrix, ensuring high measurement stability independent of long‐term system drifts. In an eight‐mode quantum frequency comb system, we experimentally determine sixteen entangled bipartitions (out of 127 possible bipartitions) and extract three squeezed supermodes via eigenmode decomposition of the covariance matrix. Through constructing linear optical networks in classical post‐processing, these resources enable the deterministic construction of cluster states ranging from one‐dimensional chains based on supermodes to three‐dimensional topological graphs formed by bipartite entangled pairs. The synergy between the parallel measurement capability of multipixel homodyne detection and the scalability of quantum frequency combs provides a compact and versatile platform for large‐scale, reconfigurable quantum information processing.
Li et al. (Sun,) studied this question.