Witnessing genuine multipartite entanglement in phase space with controlled Gaussian unitaries

TL;DR

The paper develops phase-space-based witnesses for genuine multipartite entanglement in continuous-variable systems, enabling certification with far fewer measurements than full tomography. By leveraging controlled Gaussian unitaries and measurements of the Wigner function or its Fourier transform, the authors present five concrete schemes that integrate parity, displacement, and beamsplitter operations to detect GME in paradigmatic states such as W, Dicke, N00N, and entangled cat states. They prove two main theorems linking Wigner negativity and smoothed-Wigner criteria to GME, and assess robustness to noise, finite resolution, and finite-region sampling. The methods are designed for state-of-the-art platforms, including cQED, cQAD, trapped ions, and atoms, where dichotomic phase-space readouts are standard; they promise exponential reductions in measurement overhead and practical pathways to benchmark GME resources in quantum metrology and distributed quantum information tasks.

Abstract

Many existing genuine multipartite entanglement (GME) witnesses for continuous-variable (CV) quantum systems typically rely on quadrature measurements, which is challenging to implement in platforms where the CV degrees of freedom can be indirectly accessed only through qubit readouts. In this work, we propose methods to implement GME witnesses through phase-space measurements in state-of-the-art experimental platforms, leveraging controlled Gaussian unitaries readily available in qubit-CV architectures. Based on two theoretical results showing that sufficient Wigner negativity can certify GME, we present five concrete implementation schemes using controlled parity, displacement, and beamsplitter operations. Our witnesses can detect paradigmatic GME states like the Dicke and multipartite $N00N$ states, which include the W states as a special case, and GHZ-type entangled cat states. We analyze the performance of these witnesses under realistic noise conditions and finite measurement resolution, showing their robustness to experimental imperfections. Crucially, our implementations require exponentially fewer measurement settings than full tomography, with one scheme requiring only a single measurement on auxiliary modes. The methods are readily applicable to circuit/cavity quantum electrodynamics, circuit quantum acoustodynamics, as well as trapped ions and atoms systems, where such dichotomic phase-space measurements are already routinely performed as native readouts.

Figures (17)

• Figure 1: Consider that like in (a), we wish to measure ${\tr}(\rho\prod_{n=N}^1 U_{n})$ with respect to the CV state $\rho$ using only qubit readouts. This can be implemented using (b), the controlled unitary $cU = \ketbra{\uparrow}\otimes\mathbbm{1} + \ketbra{\downarrow}\otimes U$. Then, ${\tr}(\rho \prod_{n=N}^1 U_{n})$ can be measured by first preparing all the readout qubits as $\ket{+}$, then either (c) performing all controlled unitaries on one readout qubit and measuring $\ev{\sigma_+}$, (d) performing them each on their own readout qubit and measuring $\ev{\sigma_{+}\otimes\sigma_{+}\otimes\cdots\otimes\sigma_{+}}$, or (e) some combination of both.
• Figure 2: Witness \ref{['wit:P']}: Implementation of \ref{['col:GME-Wigner-witness']} through local parity measurements on each mode. The measurements are performed for values of $\alpha$ that cover a finite region $\alpha \in \omega\subseteq\mathbb{C}$.
• Figure 3: Witness \ref{['wit:D-BS']}: Implementation of \ref{['col:GME-Wigner-witness']} using a controlled beamsplitter and local displacement measurements on each mode. The measurements are performed at a finite number of points $\xi \in \{\xi_n-\xi_{n'}\}_{n \leq n'}$.
• Figure 4: Witness \ref{['wit:BS-A']}: Implementation with a controlled beamsplitter and auxiliary modes. Only a single measurement setting is needed.
• Figure 5: Witness \ref{['wit:BS-Rand']}: Implementation with a controlled beamsplitter and random displacements. The output from the control qubit is averaged over many random displacements $\hat{\beta}$, which is sampled from a normal distribution.

Theorems & Definitions (13)

• Corollary 1: GME criterion with Wigner function measurements over a finite region
• Theorem 2: From Ref. CV-GME-letter: Negativity of the smoothed Wigner function of the center-of-mass implies GME
• Corollary 2: From Ref. CV-GME-letter: Enough nonclassicality depth of the center-of-mass implies GME
• Proposition 1: Bochner's theorem and bounds of absolute volume of displaced parity
• Proposition 2: Parity operations on collective modes as multiport beam splitters
• Proposition 3
• Proposition 4
• Proposition 5
• proof : Proof of Proposition \ref{['prop:bochner-extended']}
• proof : Proof of Proposition \ref{['prop:centre-of-mass-parity-is-beamsplitter']}