Entanglement Harvesting Protocol

• Entanglement harvesting protocol is a method in relativistic quantum field theory where localized detectors extract pre-existing vacuum correlations through controlled, weak local interactions.
• The approach quantifies genuine entanglement extraction by distinguishing nonlocal coherence from communication-induced correlations via the field’s Wightman function.
• Operational criteria ensure that only when detectors remain spacelike separated do the harvested correlations reflect the true underlying quantum field entanglement.

Entanglement harvesting is a protocol in relativistic quantum information theory wherein initially uncorrelated localized quantum systems (detectors) extract pre-existing quantum correlations from the vacuum state of a quantum field through local interactions. This operational approach enables physical observers to indirectly probe the entanglement structure of quantum fields by mapping it onto the entanglement of localized, accessible degrees of freedom. The dominant paradigm for entanglement harvesting employs two Unruh–DeWitt–type detectors with short, weak local coupling to a scalar quantum field. The protocol quantifies when and how this process constitutes genuine extraction of pre-existing field entanglement, as opposed to simply generating detector correlations via causal, field-mediated signaling.

1. Fundamental Principles and Detector–Field Model

The canonical entanglement-harvesting protocol utilizes two-level Unruh–DeWitt detectors, labeled $A$ and $B$, each with ground state $|g\rangle$, excited state $|e\rangle$, and energy gap $\Omega$. The interaction Hamiltonian in the interaction picture, neglecting backreaction, is

$$H_I(t) = \sum_{j=A,B} \lambda \, \chi_j(t) \, \mu_j(t) \, (F_j * \phi)(t, \mathbf{x}_j)$$

with:

• $\lambda$: dimensionless coupling strength,
• $\chi_j(t)$: real, compactly supported or smooth "switching" function specifying temporal localization,
• $\mu_j(t)$: monopole operator, $$\mu_j(t) = \sigma_j^+ e^{+i\Omega t} + \sigma_j^- e^{-i\Omega t}$$,
• $F_j(\mathbf{x})$: spatial smearing profile, typically $\delta^n(\mathbf{x})$ for pointlike detectors,
• $\phi(t,\mathbf{x})$: quantum field operator (e.g., real scalar field).

The detectors and field are initially in a factorized state $|g_Ag_B\rangle \otimes |0\rangle_{\text{field}}$. Time evolution to second order in $\lambda$ yields a two-qubit reduced density matrix for the detectors: $$\rho_{AB} = \text{Tr}_{\phi}[U \rho_0 U^\dagger] = \begin{pmatrix} 1-\mathcal{L}_{AA}-\mathcal{L}_{BB} & 0 & 0 & \mathcal{M}^* \ 0 & \mathcal{L}_{BB} & \mathcal{L}_{AB} & 0 \ 0 & \mathcal{L}_{BA} & \mathcal{L}_{AA} & 0 \ \mathcal{M} & 0 & 0 & 0 \end{pmatrix} + O(\lambda^4)$$ where $\mathcal{L}_{ij}$ are local excitation ("noise") terms and $\mathcal{M}$ is the nonlocal ("coherence") term.

2. Causal Structure, Wightman Function, and Correlation Splitting

The key to distinguishing genuine harvesting from communication-based entanglement transfer lies in the field's Wightman function $W(x,x') = \langle 0|\phi(x) \phi(x')|0\rangle$. This admits a decomposition: $$W(x,x') = \tfrac{1}{2}\big(C^+(x,x') + C^-(x,x') \big)$$ with:

• $C^+(x,x') = \langle\{\phi(x),\phi(x')\}\rangle$: field anti-commutator (Hadamard function), state-dependent, encodes pre-existing vacuum correlations,
• $C^-(x,x') = \langle[\phi(x),\phi(x')]\rangle$: field commutator, state-independent, encodes microcausal signaling.

The nonlocal term $\mathcal{M}$ naturally splits into two contributions: $\mathcal{M} = \mathcal{M}^+ + \mathcal{M}^-$

$$\mathcal{M}^+ = -\lambda^2 \int dt\,dt'\; \chi_A(t)\chi_B(t')\; e^{+i\Omega(t+t')} \tfrac{1}{2}C^+(x_A, x'_B)$$

$$\mathcal{M}^- = -\lambda^2 \int dt\,dt'\; \chi_A(t)\chi_B(t')\; e^{+i\Omega(t+t')} \tfrac{1}{2}C^-(x_A, x'_B)$$

• $\mathcal{M}^+$: contribution from harvesting of pre-existing field entanglement,
• $\mathcal{M}^-$: contribution due solely to causal signaling between detectors.

3. Operational Quantification and the Harvesting Criterion

Entanglement between the detectors is quantified, to leading order, via the two-qubit negativity: $$\mathcal{N} = \max\{0,\,|\mathcal{M}| - \mathcal{L}_{jj}\} + O(\lambda^4)$$ where $$\mathcal{L}_{AA} = \mathcal{L}_{BB} \equiv \mathcal{L}_{jj}$$ for identical detectors.

One further defines: $$\mathcal{N}^+ = \max\{0,\,|\mathcal{M}^+| - \mathcal{L}_{jj}\}$$

$$\mathcal{N}^- = \max\{0,\,|\mathcal{M}^-| - \mathcal{L}_{jj}\}$$

with

• $\mathcal{N}^+$: genuine harvested entanglement,
• $\mathcal{N}^-$: communication-induced entanglement.

A quantitative estimator for the communication fraction is: $$\mathcal{I} = \begin{cases} \mathcal{N}^- / \mathcal{N} & \mathcal{N} > 0 \ 0 & \mathcal{N} = 0 \end{cases}$$ Interpretation:

• $\mathcal{I}\ll1$: entanglement is genuinely harvested,
• $\mathcal{I}\approx1$: entanglement is entirely due to communication.

4. Dimensional and Field-Theoretic Dependence

The dependence on spacetime dimension and field properties is profound:

• Massless scalar field in (3+1)D flat spacetime: $C^-(x,x')$ is supported only on the lightcone. For purely spacelike separation ($|\Delta t| < L$), $C^-=0$, so $\mathcal{M}^- = 0$ and only genuine harvesting occurs ($\mathcal{N} = \mathcal{N}^+$). When the detectors become null- or timelike-related ($|\Delta t| \geq L$), $C^-(x,x')\ne 0$, $\mathcal{M}^-$ becomes nonzero, quickly dominating, and $\mathcal{N} \approx \mathcal{N}^-$: all entanglement comes from communication.
• Lower-dimensional massless fields: $C^-$ acquires support inside the lightcone (violation of the strong Huygens principle). For any causal contact (timelike or null separation), $\mathcal{M}^-$ dominates and genuine harvesting is suppressed.
• Massive scalar field, (3+1)D: The commutator $C^-$ has support everywhere inside the lightcone. For timelike-related detectors, both $\mathcal{M}^+$ and $\mathcal{M}^-$ contribute, typically comparably; harvesting and communication-based entanglement coexist over generic parameter regimes.

5. Causal Separation and the True Harvesting Regime

The defining criterion for genuine entanglement harvesting is the vanishing of the commutator between the support regions of the two detectors: $$C^-(x_A, x'_B) = 0 \;\Leftrightarrow\; \mathcal{M}^- = 0 \;\Leftrightarrow\; \mathcal{I}=0,\;\mathcal{N} = \mathcal{N}^+$$ In flat spacetime, this requires purely spacelike separation over the entire support (switching windows and spatial smearings) of both detectors.

Whenever $C^-(x_A, x'_B)\neq 0$ (i.e., any causal contact), communication-induced effects inevitably dominate to leading order, unless parameters are finetuned, which is generically impossible for massless fields in flat backgrounds.

6. Field Mass, Switching, and Separability Features

For massive fields in (3+1)D, $\mathcal{M}^+$ and $\mathcal{M}^-$ both survive for non-spacelike separations, and their relative magnitude and phase oscillate with time delay and field parameters. The average harvested negativity, for long enough detector interactions, is typically split evenly: $\mathcal{N}^+ \sim \mathcal{N}^-$.

In massless ($m=0$) flat space, the transition from harvested to communicative entanglement is sharp at the lightcone: numerical analysis shows $\mathcal{N}^-$ switches on with a discontinuity as soon as the detector windows touch the lightcone, while $\mathcal{N}^+$ rapidly decays. For lower dimensions, or backgrounds violating the strong Huygens principle, similar dominance of communication occurs for causal contact, but the decay of $\mathcal{N}^+$ is more gradual.

7. Implications and Protocol Limitations

Rigorous application of this splitting shows that standard entanglement-harvesting scenarios in massless flat-space QFT truly reflect extraction of pre-existing vacuum entanglement only when the detectors remain spacelike separated throughout their support. As soon as causal signaling is possible, measured detector entanglement is almost exclusively communication-induced. This constrains experimental and theoretical interpretations, as practical detector switching and spatial support must ensure spacelike separation for robust field-entanglement extraction.

For massive fields or curved backgrounds, both mechanisms may coexist, requiring detailed analysis of the commutator support and field correlators for proper attribution of harvested correlations.

Summary Table: Harvesting vs Communication Contributions

Field/Spacetime	Detector separation	Commutator $C^-$	$\mathcal{N}^+$	$\mathcal{N}^-$	Harvesting?
Massless, flat 3+1	Spacelike	$0$	$\neq0$	$0$	Yes
Massless, flat 3+1	Null/timelike	$\neq0$	$\simeq0$	$\neq0$	No
Massive, flat 3+1	Timelike	$\neq0$	$\sim$comm.	$\sim$harv.	Mixed
Massless, flat 1+1/2+1	Timelike/null	$\neq0$	$<$comm.	$\gg$harv.	No

• "When entanglement harvesting is not really harvesting" (Tjoa et al., 2021)