Negativity Cores in Gaussian Entanglement

• Negativity cores are localized entanglement carriers in mixed bipartite Gaussian states, defined using the partial transpose of the covariance matrix.
• They isolate paired modes between disjoint subsystems through local symplectic transformations, quantifying entanglement via contributions to logarithmic negativity.
• The core–halo decomposition, revealed in both lattice and continuum settings, provides actionable insights for entanglement extraction and simulation in quantum field theories.

Negativity cores are localized entanglement carriers defined with respect to the partial-transpose structure of a mixed bipartite state. In the Gaussian quantum-field and many-body settings where the term is used explicitly, a negativity core is not merely a region-level scalar such as the logarithmic negativity, but a mode-resolved decomposition in which the entanglement between two disjoint subsystems is concentrated into paired modes, one on each side of the bipartition. In lattice scalar-field vacuum states this concentration is achieved by local symplectic transformations that isolate a tensor-product “core” of entangled two-mode Gaussian states from a residual “halo” (Klco et al., 2021). In the continuum vacuum of a $1+1$-dimensional massless real scalar field, each negativity core is a pair of localized interval modes associated with a partially transposed symplectic eigenvalue $\tilde{\nu}_n<1$, contributing exactly $-\log_2 \tilde{\nu}_n$ to the total logarithmic negativity (Pye et al., 29 Jun 2026). The notion therefore turns entanglement negativity into a structural description of where mixed-state entanglement resides in field space, rather than only how much of it is present (Pye et al., 22 May 2026).

1. Definition and terminological scope

In the lattice formulation developed for non-interacting scalar field theory in one spatial dimension, the vacuum entanglement between two finite disjoint regions $A$ and $B$, initially a $(d_A\times d_B)_{\rm mixed}$ Gaussian continuous-variable state, is locally transformed into a tensor-product core of $(1_A\times 1_B)_{\rm mixed}$ entangled pairs (Klco et al., 2021). The number of such pairs is $n_c=n_-$, where $n_-$ is the number of partially transposed symplectic eigenvalues below unity, and each pair is associated with one negativity-contributing mode. The remainder of the degrees of freedom is termed the “halo.”

The continuum definition is sharper. For two compact spacelike-separated intervals $A$ and $\tilde{\nu}_n<1$0 in the vacuum of a $\tilde{\nu}_n<1$1D massless real scalar field, the term “negativity cores” denotes the pairwise local mode decomposition of the entangled content of the bipartite Gaussian state: modes with $\tilde{\nu}_n<1$2 form the core, while the orthogonal complement forms a halo that does not contribute to the logarithmic negativity (Pye et al., 29 Jun 2026). A complementary formulation describes a negativity core as the pair of localized field modes, one in $\tilde{\nu}_n<1$3 and one in $\tilde{\nu}_n<1$4, that together capture a single “modewise” contribution to the mixed-state entanglement between the two regions (Pye et al., 22 May 2026).

This usage is therefore specific and technical. It does not refer generically to regions of negative spectrum, nor to a coarse spatial subregion, but to a canonical mode pair singled out by the spectrum of the partially transposed Gaussian structure.

2. Gaussian-state and partial-transpose framework

The negativity-core construction is formulated in phase space. In the lattice Gaussian setting, the vacuum covariance matrix is written as

$\tilde{\nu}_n<1$5

and partial transpose for a bipartition $\tilde{\nu}_n<1$6 is implemented by the continuous-variable rule $\tilde{\nu}_n<1$7 for $\tilde{\nu}_n<1$8, or equivalently $\tilde{\nu}_n<1$9 at the covariance-matrix level (Klco et al., 2021). The partially transposed symplectic spectrum

$-\log_2 \tilde{\nu}_n$0

determines the logarithmic negativity,

$-\log_2 \tilde{\nu}_n$1

The negativity core is obtained by constructing local symplectic transformations

$-\log_2 \tilde{\nu}_n$2

from the partial-transpose eigenvectors and organizing the state as

$-\log_2 \tilde{\nu}_n$3

where $-\log_2 \tilde{\nu}_n$4 is the core, $-\log_2 \tilde{\nu}_n$5 the halo, and $-\log_2 \tilde{\nu}_n$6 a positive semidefinite classical-noise matrix (Klco et al., 2021).

The continuum formulation recasts the same structure in Kähler language. For a Gaussian state with covariance matrix $-\log_2 \tilde{\nu}_n$7, the linear complex structure is

$-\log_2 \tilde{\nu}_n$8

and partial transpose gives

$-\log_2 \tilde{\nu}_n$9

The partially transposed symplectic eigenvalues $A$0 are then the spectral data of $A$1, and

$A$2

follows directly (Pye et al., 29 Jun 2026). A notable refinement of this framework is a basis-independent definition of transpose: a linear map $A$3 is a transpose iff

$A$4

This isolates the essential operation behind partial transpose without tying it to a preferred coordinate basis (Pye et al., 29 Jun 2026).

3. Exact continuum solution in the $A$5D massless scalar vacuum

The most complete realization of negativity cores is presently available for the vacuum of a free massless real scalar field in $A$6 dimensions. The field Hamiltonian is

$A$7

with two disjoint bounded intervals

$A$8

and with observables restricted to derivative degrees of freedom via smearing functions of vanishing integral on each interval, which removes the zero-mode divergence (Pye et al., 22 May 2026). In the right/left-moving basis,

$A$9

the two chiral sectors decouple and contribute equally, producing a twofold degeneracy in the negativity spectrum (Pye et al., 29 Jun 2026).

For this problem, all interval-geometry dependence is encoded in the conformal cross ratio

$B$0

The diagonalization of the partially transposed restricted complex structure reduces to a boundary-value problem in the complex plane. The analytic function $B$1, defined off the cuts $B$2, has the unique solution relevant for $B$3,

$B$4

with

$B$5

Imposing the vanishing-integral constraints discretizes the spectrum through

$B$6

so that the admissible $B$7 are zeros of a conical/Legendre function and

$B$8

for the full theory (Pye et al., 29 Jun 2026).

The resulting negativity-core modes are given in closed form. Their smearing functions on $B$9 and $(d_A\times d_B)_{\rm mixed}$0 have endpoint-controlled square-root denominators and oscillatory phases $(d_A\times d_B)_{\rm mixed}$1, where

$(d_A\times d_B)_{\rm mixed}$2

These modes are canonical, interval-localized, and pairwise entangled; as $(d_A\times d_B)_{\rm mixed}$3 increases, the oscillations become more rapid. The continuum solution thereby identifies not only the total negativity but also the explicit local mode pairs that carry it (Pye et al., 22 May 2026).

4. Core–halo decomposition, hierarchy, and asymptotics

The core–halo split is central to the meaning of negativity cores. In the lattice construction, the transformed state can be resolved as

$(d_A\times d_B)_{\rm mixed}$4

with each two-mode block $(d_A\times d_B)_{\rm mixed}$5 carrying one contribution

$(d_A\times d_B)_{\rm mixed}$6

to the total negativity (Klco et al., 2021). Accessible entanglement is exactly the sum of these core-pair contributions and matches the full region-region logarithmic negativity. The ordering is hierarchical: the strongest partial-transpose mode becomes the strongest physical core pair, the next strongest becomes the next pair, and so forth. Numerically this hierarchy is typically exponential, so that at moderate separations

$(d_A\times d_B)_{\rm mixed}$7

and only a few pairs dominate the long-range entanglement (Klco et al., 2021).

The halo is more subtle than a simple separable remainder. The lattice analysis shows that the halo can be $(d_A\times d_B)_{\rm mixed}$8-separable in isolation and also separable from the core with respect to another partition, yet these two separability statements are not simultaneously compatible in the full mixed state because positivity of the combined classical-noise structure fails. This motivates the interpretation of halo correlations as inaccessible or bound entanglement: they are obscured by classical correlations and may be required by the preparation protocol even though the distillable entanglement is concentrated in the core (Klco et al., 2021).

The continuum solution adds exact asymptotics. For equal-length intervals with $(d_A\times d_B)_{\rm mixed}$9, the $(1_A\times 1_B)_{\rm mixed}$0-th negativity contribution obeys

$(1_A\times 1_B)_{\rm mixed}$1

where $(1_A\times 1_B)_{\rm mixed}$2 is the $(1_A\times 1_B)_{\rm mixed}$3-th zero of $(1_A\times 1_B)_{\rm mixed}$4, and the dominant decay rate is $(1_A\times 1_B)_{\rm mixed}$5 (Pye et al., 22 May 2026). As the intervals approach each other, $(1_A\times 1_B)_{\rm mixed}$6, the total negativity diverges as

$(1_A\times 1_B)_{\rm mixed}$7

which combines the universal adjacent-interval logarithm with an additional double-logarithmic term attributed to the infrared structure of the massless scalar (Pye et al., 22 May 2026). In physical terms, increased separation forces the core modes to become more oscillatory, exhibiting an explicit ultraviolet–infrared connection.

5. Conceptual interpretations in the wider negativity literature

Negativity cores are a mode-localized notion, but they resonate with broader interpretations of negativity. One such line of thought is the dimension-counting interpretation of negativity. For a bipartite state $(1_A\times 1_B)_{\rm mixed}$8, the quantity

$(1_A\times 1_B)_{\rm mixed}$9

was introduced as an estimator of how many local degrees of freedom are entangled; in particular, $n_c=n_-$0 is a lower bound on the Schmidt number for arbitrary states, and for axisymmetric states one has

$n_c=n_-$1

exactly (Eltschka et al., 2013). This suggests a natural interpretation of negativity cores: the core need not be viewed only as a set of localized modes, but also as the minimal entangled subspace certified by the partial transpose.

A second line concerns the distinction between bipartite and multipartite correlations. In monitored stabilizer circuits, mutual negativity obeys

$n_c=n_-$2

whereas

$n_c=n_-$3

Negativity therefore counts only the EPR-pair content $n_c=n_-$4, while mutual information also counts GHZ-type content $n_c=n_-$5 (Sang et al., 2020). Although that work does not introduce negativity cores as a technical construction, it implies that a negativity-based decomposition isolates a genuinely bipartite entanglement backbone rather than total or multipartite correlation. In that sense, the core language aligns with a broader operational reading of negativity as the part of a mixed-state correlation structure that remains directly distillable and pairwise.

These parallels are interpretive rather than definitional. The explicit notion of a negativity core remains tied to Gaussian mode decompositions, but its conceptual significance is strengthened by the wider literature on negativity as a witness of entangled dimensionality and of strictly bipartite quantum correlations.

6. Applications, extensions, and open problems

The principal application of negativity cores is entanglement localization. In the lattice scalar-field setting, the core pairs identify the dominant region modes from which vacuum entanglement could be extracted into a spatially separated pair of quantum detectors, while the partial-transpose eigenmodes provide a guide for constructing effective detector couplings when the local preprocessing cannot be performed directly in the continuum (Klco et al., 2021). The same work argues that entanglement consolidation should persist in higher dimensions and may aid classical and quantum simulations of asymptotically free gauge field theories such as quantum chromodynamics (Klco et al., 2021).

The continuum analysis extends this program by giving a complete mode-resolved characterization for arbitrary disjoint intervals in a massless scalar vacuum and by exhibiting an exact complex-analytic route to the negativity spectrum (Pye et al., 29 Jun 2026). Several extensions are stated explicitly: higher dimensions, where related boundary-value reformulations of Gaussian entanglement problems already exist; fermionic fields, where the Kähler framework persists but the partial transpose is subtler; and alternative modal decompositions together with the energetic cost of extracting entanglement from the field (Pye et al., 22 May 2026). The same continuum work also remarks that all negativity cores are mutually commuting across different $n_c=n_-$6, so the entire family of core modes can in principle be extracted simultaneously (Pye et al., 22 May 2026).

Limitations remain equally clear. The most complete results are for free Gaussian theories. The halo structure is only partly understood, especially when inaccessible or bound entanglement is implicated in state preparation. Even within the exactly solved massless scalar problem, high-$n_c=n_-$7 cores are increasingly oscillatory and become harder to resolve numerically on a lattice regularization (Pye et al., 22 May 2026). A plausible implication is that negativity cores furnish a canonical language for mixed-state entanglement geometry in free theories, while interacting and non-Gaussian systems still require an analogous structural theory.