Translation symmetry-enforced long-range entanglement in mixed states

Ryan Thorngren, Lei Gioia, Carolyn Zhang · 2026

Read Translation symmetry-enforced long-range entanglement in mixed states by Ryan Thorngren, Lei Gioia, Carolyn Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-15200.txt.

Introduction

Ryan Thorngren, Lei Gioia, and Carolyn Zhang show by counting that translation symmetry enforces long-range entanglement in a mixed state. Symmetric short-range states exist, but too few span the zero-momentum sector. The fixed-point state of strong-to-weak spontaneous breaking of translation symmetry therefore cannot be written as a mixture of short-range states. This subtle long-range entanglement escapes long-range connected correlation functions, which see nothing unusual.

Catalog record

CREATORRyan Thorngren, Lei Gioia, Carolyn Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE35792
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.15200

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Translation symmetry-enforced long-range entanglement in mixed states Ryan Thorngren, Lei Gioia, Carolyn Zhang Abstract We show by a counting argument that even though translation symmetry admits symmetric short-range entangled (SRE) eigenstates, there are not enough such SRE eigenstates to span the zero momentum sector. This means that the fixed point strong-to-weak spontaneous symmetry breaking state of translation symmetry is long-range entangled: it cannot be written as a mixture of SRE states. This is a subtle form of long-range entanglement in mixed states that cannot be detected by long-range connected correlation functions. Introduction —Recent work on open quantum systems has highlighted the rich landscape of different patterns of entanglement unique to mixed states. We say a mixed state is long-range entangled (LRE) if it is not approximately a mixture of easy-to-prepare (i.e. “short-rangle entangled”) states Hastings 2011 ; Chen and Grover 2024a ; Chen and Grover 2024b 1 1 1 Note that under this definition, states like \frac{1}{2}\left(|\{Z_{r}=+1\}\rangle\langle\{Z_{r}=+1\}|+|\{Z_{r}=-1\}\rangle\langle\{Z_{r}=-1\}|+\right) is considered SRE, even though it has classical correlations precluding an SRE purification. This definition of SRE focuses on quantum entanglement. Also note that such states are sometimes called “separable”, even though in the quantum information literature “separable” is reserved for states that satisfy the stronger condition that they can be written as mixtures of product states . . It is natural to ask: what are general classes of LRE mixed states? What are the mechanisms that lead to this entanglement? And how can long range entanglement be detected? Symmetries have served as an important tool for exploring the above questions de Groot et al. 2022 ; Moharramipour et al. 2024 ; Sala et al. 2024 ; Lessa et al. 2025a ; Lessa et al. 2025b ; Li et al. 2025 ; Ziereis et al. 2025 ; Wang and Li 2025 . In particular, Ref. Lessa et al. 2025a showed that strong anomalous symmetries preclude a state from being multipartite separable. If a mixed state has a strong G symmetry, meaning U_{g}\rho\propto\rho,\,\forall g\in G , this implies that every state |\psi_{i}\rangle in its decomposition must be LRE. This led to the construction of mixed states that, unlike any pure state, are tripartite entangled but bipartite disentangled. On the other hand, on-site symmetries of the form U_{g}=\otimes_{r}U_{g,r} , where each U_{g,r} acts on a single site, clearly admit completely disentangled symmetric product states. In fact, for abelian G , their entire symmetry sectors can be spanned by symmetric product states: \rho\propto P_{\pi(g)}=\sum_{i}|\psi_{i}\rangle\langle\psi_{i}| where \{|\psi_{i}\rangle\} are product states and \pi(g) labels an irreducible representation of G . P_{\pi(g)} is also known as the maximally mixed state in a symmetry sector (MMIS), and gives the fixed point representative of a strong-to-weak spontaneous symmetry breaking (SWSSB) phase Ma et al. 2025 ; Lessa et al. 2025c ; Ziereis et al. 2025 ; Moharramipour et al. 2024 ; Weinstein 2025 . On the other hand, strong non-abelian finite and continuous symmetries enforce entanglement in the MMIS Moharramipour et al. 2024 ; Li et al. 2025 . Translation symmetry lies somewhere in between on-site symmetries and anomalous symmetries. Translation symmetry has the peculiar property that states in particular symmetry sectors, i.e. with nonzero momentum (transforming as T|\psi\rangle=e^{ik}|\psi\rangle with k\neq 0 mod 2\pi ), must be long range entangled Gioia and Wang 2022 . In the zero momentum sector, it is straightforward to write down translation invariant product states. However, unlike for on-site symmetries, it may not be possible to span the zero momentum sector with short range entangled zero momentum states. In the context of open quantum systems, the above question can be rephrased as – are states that SWSSB translation symmetry LRE? While the results of Ref. Lessa et al. 2025a rule out an SRE SWSSB state for anomalous symmetries, it does not rule this out for symmetries like translation and the recently introduced non-onsiteable anomaly-free finite internal symmetries Shirley et al. 2026 ; Tu et al. 2026 . In this work, we show that while translation symmetry is anomaly-free and admits symmetric product states, its zero momentum sector cannot be spanned by symmetric SRE states. Therefore, the projector onto the zero momentum sector, which demonstrates SWSSB of translation symmetry, is LRE. This long range entanglement is rather subtle because it is not detected by long range connected correlation functions sup . The mismatch between the space of zero momentum states and the space spanned by SRE zero momentum states gives a new mechanism for enforcing long-range entanglement. It is worth mentioning that from the perspective of parent channels for a mixed quantum state, which describe dynamics steering arbitrary states to the mixed state of interest, strong translation symmetry is very unusual. A quantum channel \mathcal{E}[\rho]=\sum_{i}K_{i}\rho K_{i}^{\dagger} is strongly symmetric if [U_{g},K_{i}]=0 for all g,i . This is impossible for translation symmetry unless K_{i} has support over the entire system. Indeed, we will give an example where translation SWSSB states show up naturally as steady states of quantum channels, and these quantum channels are nonlocal. Throughout, we work in 1d with periodic boundary conditions, on a chain of length n with local Hilbert space/physical dimensions q . However, our results generalize to higher dimensions (see Discussion ). Main Theorems —We study the maximally mixed state of translation invariant states, given by \displaystyle\rho_{\rm TI}=\frac{1}{D_{\rm TI}}\sum_{j}|\Psi_{j}\rangle\langle\Psi_{j}|\,, (1) where j sums over an orthonormal basis of translation-invariant (i.e. zero momentum) states |\Psi_{j}\rangle and D_{\rm TI} is the dimension of the space of these states. We will prove \rho_{\rm TI} is long-range entangled in the sense of the following theorem. Theorem 1 . The maximally mixed state of translation invariant states \rho_{\rm TI} is not well-approximated by any mixture of easily-preparable pure states. In particular, if \sigma_{\tau} is a mixture of pure states each preparable by time- \tau evolution from a product state by a range- r Hamiltonian (not necessarily all the same), and \sigma_{\tau} well-approximates \rho_{\rm TI} in the sense that \begin{gathered}D(\rho_{\rm TI},\sigma_{\tau})\leq 1-\eta\,,\end{gathered} (2) where D is trace distance and \eta>1/\text{poly}(n) , then \begin{gathered}\tau=\Omega(\frac{\sqrt{n}}{\text{polylog}(n)})\,,\end{gathered} (3) meaning \tau is lower bounded by a function which is asymptotic to \sqrt{n}/\text{polylog(n)} . The theorem means that to sample states from an ensemble with even very mild fidelity with \rho_{\text{TI}} requires a time growing quickly with the system size. The \sqrt{n} lower bound essentially comes from the volume of a light-cone for time- \tau evolution being size \tau^{2} . The first ingredient in the proof is a lemma which is generally useful when proving long-range-entanglement. It says that it is actually enough to show \rho_{\text{TI}} is not well approximated by “depth- d states” for large enough d . These are pure states created by applying a depth- d circuit to a product state. This lemma based on the fact that time- \tau evolution under a bounded geometrically local time-dependent Hamiltonian is well approximated by a depth- d circuit for d\leq d(\tau,n,\epsilon,r) , a universal function depending only on the number of sites n , the time \tau of evolution, and the range r of the Hamiltonian Haah et al. 2021 . See Theorem 2 in the Supplemental Material for the precise statement and the proof of the following. Lemma 1 . (Tails Lemma) There is a function d(\tau,n,\epsilon,r)=O(\tau\ \text{polylog}(n\tau/\epsilon)) such that if P_{d} is

Read the complete text