Many-Body Entanglement and Information

Entanglement describes correlations relative to a chosen division into subsystems. In a many-body system, that division may separate spatial regions, modes, orbitals, or operator degrees of freedom. The meaning of an entropy also depends on whether the complete state is pure or mixed.

This chapter explains how entropies, mutual information, Schmidt spectra, and tensor networks reveal different aspects of many-body states. Begin with Many-Body Entanglement Overview to choose the partition and quantity you need.

Required background. Reduced Density Operators supplies subsystem states and partial-trace logic. Thermodynamic Limit supplies controlled size sequences, held-fixed data, and order-of-limits discipline.

Helpful background. Finite-system entanglement entropy and Schmidt decomposition prepare the pure-state branches. Quantum Statistical Mechanics prepares thermal states; Phases prepares critical and topological claims; the Heisenberg picture prepares operator-growth questions.

The same subsystem entropy can mean different things

For a globally pure bipartite state, $S_A=-\operatorname{Tr}(\rho_A\ln\rho_A)$ measures entanglement between $A$ and its complement. We use natural logarithms here. The Bell state $(|00\rangle+|11\rangle)/\sqrt2$ has $\rho_A=I_2/2$ and $S_A=\ln2$.

Now consider the uncorrelated mixed state $I_2/2\otimes I_2/2$. Its subsystem state is the same, so $S_A=\ln2$ again, but the two parts are not correlated at all. The mixed state

$$\rho_{\rm cl}=\frac12\bigl(|00\rangle\langle00|+|11\rangle\langle11|\bigr)$$

also has $S_A=\ln2$. It contains classical correlation and is separable. These examples show why a subsystem entropy alone cannot measure entanglement of a mixed global state.

Mutual information, $I(A:B)=S_A+S_B-S_{AB}$, distinguishes their total correlations: its values are $2\ln2$, 0, and $\ln2$, respectively. It does not separate classical correlation from quantum entanglement. For that task, use a suitable mixed-state measure, such as the quantities introduced in Negativity and PPT.

In extended systems, also specify the shape and size of the region, boundary conditions, lattice spacing or other regulator, and total-system state. Area, logarithmic, and volume scaling are useful patterns, but none uniquely identifies a phase or dynamical mechanism on its own.

Choose an information quantity

Select the branch appropriate to the state and partition.

1. Establish the common language. Start with Many-Body Entanglement Overview, then use Entanglement Entropy in Many-Body Systems for spatial von Neumann and Rényi entropies, benchmark states, and scaling extraction.
2. Branch by state class. Thermal Entropy versus Entanglement Entropy separates global mixing from reduction. Mutual Information in Many-Body Systems measures total correlation between regions in pure or mixed states; it is not generally a mixed-state entanglement measure.
3. Branch by leading scaling. Area Laws treats boundary-dominated ground-state structure and its qualifications. Volume Laws compares thermal, eigenstate, random-state, and dynamical mechanisms that share an extensive leading term.
4. Resolve more than one entropy number. Entanglement Spectrum keeps Schmidt levels, symmetry sectors, and convention-dependent entanglement energies visible.
5. Choose a representation branch. Tensor products, singular-value decomposition, and Schmidt structure lead to Tensor Networks Preview, then Matrix Product States Preview for one-dimensional structure. Area laws and spectrum data explain capacity; they are not universal prerequisites for defining a tensor network.
6. Choose a critical branch. Entanglement Entropy plus quantum-transition and scaling preparation leads to Entanglement and Criticality. Add MPS only when finite-bond-dimension scaling is part of the question.
7. Choose a topological branch. Topological Order plus Entanglement Entropy and Area Laws leads to Topological Entanglement Entropy Preview. Entanglement Spectrum and mutual-information constructions are complementary diagnostics, not universal prerequisites.
8. Choose an operator-growth branch. Heisenberg evolution plus Schmidt and entropy language leads to Operator Entanglement and Scrambling Preview. Return to Nonequilibrium Many-Body Dynamics for quench, OTOC, thermalization, and chaos claims.

Example: a constant term in a spatial entropy

Suppose finite two-dimensional samples fit an entropy form with a boundary contribution and a negative constant intercept. The fit is compatible with a topological entanglement term only after region geometry, corners, ultraviolet convention, correlation length, boundary conditions, ground-state sector, and the size range relative to microscopic and correlation scales are controlled. One subtraction geometry or one linear fit does not establish topological order. The strongest defensible statement is consistency with the predicted subleading structure, strengthened by independent checks of the gap, sectors, quasiparticle content, entanglement spectrum, or other topological diagnostics.

Continue to algorithms or physical interpretation

Composite Systems and Entanglement supplies partial traces, Schmidt decomposition, and finite-system measures. Quantum Statistical Mechanics explains thermal mixing. For a phase interpretation, continue to Phases, Order, and Criticality.

Computational Many-Body QM connects Schmidt spectra to tensor-network contraction and optimization. Nonequilibrium Dynamics studies entanglement growth, operator spreading, and their relation to relaxation.

Common pitfalls

“Nonzero subsystem entropy proves entanglement.” This is direct for an appropriate globally pure bipartite setting. For a mixed global state, subsystem entropy also contains ordinary mixing.

“Mutual information is mixed-state entanglement.” Mutual information measures total correlation and can be nonzero in a separable state.

“An area law proves a gap and an efficient algorithm.” Neither implication holds without additional dimensional, state, Hamiltonian, contraction, and optimization hypotheses.

“A volume law proves thermalization or chaos.” Thermal states, chaotic eigenstates, random vectors, and quench states can share extensive leading entropy while differing physically.

“The entanglement spectrum is an energy spectrum.” Its levels derive from a reduced state and depend on the partition, normalization, and convention.

“A constant entropy intercept proves topological order.” Corners, finite size, regulator choices, fitting correlations, and sector dependence require control and independent checks.

“Bond dimension is an accuracy certificate.” It bounds representation capacity. Discarded weight, observables, energies, correlations, and size or bond-dimension convergence must still be checked.

“Logarithmic entropy proves a conformal field theory.” Geometry, boundary conditions, oscillatory and finite-size corrections, and competing gapless mechanisms must be separated.

“Operator entanglement growth is scrambling.” Operator support, operator entanglement, OTOCs, chaos, information loss, and thermalization are related but inequivalent diagnostics.

Exercises

Exercise 1: Select four information diagnostics

Route (a) a gapped-chain ground-state interval, (b) a regional reduction of a mixed Gibbs state, (c) a highly excited eigenstate with entropy proportional to subsystem size, and (d) a growing Heisenberg operator. Name the measure, scaling variable, and strongest initial claim for each.

Solution

For (a), use Entanglement Entropy → Area Laws and test saturation, boundaries, and finite correlation length. For (b), first use Thermal versus Entanglement; subsystem entropy alone is not an entanglement measure, so use Mutual Information for total regional correlation or a dedicated mixed-state measure for entanglement. For (c), use Volume Laws and compare eigenstate, thermal, random-state, and finite-size explanations before invoking thermalization. For (d), use Operator Entanglement and Scrambling, then Nonequilibrium Dynamics for support fronts, OTOCs, chaos, or thermalization.

Exercise 2: Test an entropy-based interpretation

A study fits one family of finite-region entropies to a boundary term plus a constant and observes low-lying levels in one entanglement spectrum. What additional checks are needed before claiming intrinsic topological order?

Solution

Specify the region and regulator, control corner and boundary corrections, demonstrate a size window large compared with microscopic and correlation scales, test multiple geometries or subtraction constructions, track fit covariance and finite-size uncertainty, fix the ground-state and symmetry sector, and exclude symmetry-breaking or critical alternatives. Then seek independent phase signatures such as a robust gap, characteristic sector structure, quasiparticle or anyon data, ground-state degeneracy under controlled topology, or other canonical topological diagnostics. The entropy and spectrum constrain the interpretation but do not establish it alone.

References

• L. Amico, R. Fazio, A. Osterloh, and V. Vedral, “Entanglement in Many-Body Systems,” Reviews of Modern Physics 80, 517–576 (2008).
• P. Calabrese and J. Cardy, “Entanglement Entropy and Quantum Field Theory,” Journal of Statistical Mechanics P06002 (2004).
• J. Eisert, M. Cramer, and M. B. Plenio, “Colloquium: Area Laws for the Entanglement Entropy,” Reviews of Modern Physics 82, 277–306 (2010).
• R. Orús, “A Practical Introduction to Tensor Networks: Matrix Product States and Projected Entangled Pair States,” Annals of Physics 349, 117–158 (2014).
• U. Schollwöck, “The Density-Matrix Renormalization Group in the Age of Matrix Product States,” Annals of Physics 326, 96–192 (2011).