Lieb–Robinson Bound

The Lieb–Robinson theorem gives a state-independent upper bound on how quickly a bounded local observable can develop a large commutator with a distant bounded observable under sufficiently short-range lattice dynamics.

Helpful background. The full Lieb–Robinson Bounds treatment defines the interaction norm, derives the constants, proves the path-sum mechanism, and separates finite- from infinite-volume dynamics. This card is for lookup, not a substitute for that hypothesis ledger.

The exponential locality estimate

Let $(\Gamma,d)$ be a countable metric site set. For finite $X,Y\subset\Gamma$, let $\mathcal A_X$ and $\mathcal A_Y$ be the bounded local observable algebras. Suppose a bounded interaction $\Phi(Z)$ has finite exponentially weighted $F$ norm $\|\Phi\|_{F_a}$ for some $a>0$, where $F$ is uniformly summable and satisfies the convolution condition, $F_a(r)=e^{-ar}F(r)$, and $C_{F_a}$ is the convolution constant of $F_a$.

For disjoint $X,Y\subset\Lambda$, $A\in\mathcal A_X$, and $B\in\mathcal A_Y$, the finite-volume Heisenberg dynamics obeys

$$\begin{aligned} \|[\tau_t^\Lambda(A),B]\| &\leq \frac{2\|A\|\|B\|}{C_{F_a}} \left( e^{2C_{F_a}\|\Phi\|_{F_a}|t|/\hbar}-1 \right)\\ &\qquad\times \sum_{x\in X}\sum_{y\in Y} F_a(d(x,y)). \end{aligned}$$

The constants do not depend on the finite volume $\Lambda$. The full theorem has a sharper interaction-boundary prefactor; the displayed form is a clean corollary.

With

$$d(X,Y) = \min_{x\in X,\,y\in Y}d(x,y), \qquad v_a = \frac{2C_{F_a}\|\Phi\|_{F_a}}{a\hbar},$$

one obtains

$$\begin{aligned} \|[\tau_t^\Lambda(A),B]\| &\leq \frac{2\|F\|}{C_{F_a}} \|A\|\|B\| \min\{|X|,|Y|\}\\ &\qquad\times e^{-a[d(X,Y)-v_a|t|]}. \end{aligned}$$

Use the smaller of this expression and the trivial bound $2\|A\|\|B\|$.

Hypothesis checklist

• Geometry: a countable metric site set with a uniformly summable decay function and finite convolution constant.
• Kinematics: bounded local observables supported in finite, initially disjoint regions.
• Interaction: bounded self-adjoint intersite terms with finite weighted $F$ norm. Finite-range bounded interactions on bounded-geometry lattices are standard examples.
• Norm: the operator norm; the result is uniform over states.
• Volume: the displayed constants are independent of $\Lambda$, enabling a thermodynamic-limit argument.
• Units: $\|\Phi\||t|$ appears as $\|\Phi\||t|/\hbar$ when $\hbar$ is not set to one.

Finite-dimensional on-site Hilbert spaces are convenient but not necessary. Arbitrary self-adjoint, possibly unbounded, on-site Hamiltonians can be handled in the interaction picture when the intersite terms and tested observables remain bounded. Generic unbounded intersite interactions require a different theorem.

Interpretation

When $d(X,Y)>v_a|t|$, the commutator is exponentially suppressed. Evolution is therefore quasi-local: a local observable generally develops small tails, not compact support bounded by a strict cone.

The velocity $v_a$ depends on the decay parameter, interaction representation, geometry, and norm estimates. It is normally an upper-bound parameter, often far above a measured quasiparticle, butterfly, entanglement, transport, or signal-front velocity.

The estimate directly controls dynamical influence. Conclusions about the generation of correlations or entanglement require assumptions about the initial state. Exponential ground-state clustering additionally requires a spectral gap and the hypotheses of a clustering theorem.

Principal limitations

• The tail is generically nonzero, so this is not relativistic microcausality.
• The theorem supplies no lower bound on propagation and does not imply ballistic transport.
• Power-law interactions require decay- and dimension-dependent bounds; the exponential formula cannot be reused unchanged.
• Odd fermionic observables require graded locality or an even-observable restriction.
• Unbounded observables and unbounded intersite interactions require specialized estimates.
• A locality bound alone does not imply a gap, an area law, phase stability, or efficient simulation.

Quick checks

Initial time

Why does the exact $F$-function expression contain the term “$-1$”?

Solution

For disjoint supports, $[A,B]=0$ exactly at $t=0$. The factor

$$e^{2C_{F_a}\|\Phi\|_{F_a}|t|/\hbar}-1$$

also vanishes at $t=0$, so the bound records that exact initial condition. The simplified cone form uses $e^u-1\leq e^u$ and sacrifices this sharpness.

A tolerance contour

If $\|[\tau_t(A),B]\|\leq K e^{-a(r-v|t|)}$, find a sufficient separation for the bound to be at most $\varepsilon<K$.

Solution

Solve $K e^{-a(r-v|t|)}\leq\varepsilon$. Taking logarithms gives

$$r \geq v|t| + \frac{1}{a}\log\frac{K}{\varepsilon}.$$

The second term is the accuracy-dependent buffer outside the nominal line $r=v|t|$.

Correlation versus influence

Can two far-separated sites have a nonzero connected correlation while their Lieb–Robinson commutator is zero?

Solution

Yes. Prepare the sites in a Bell state and set $H=0$. Suitable Pauli operators have a nonzero connected correlation, but all Heisenberg operators are time independent and disjoint local operators commute. The theorem constrains dynamically generated influence, not pre-existing state correlations.

References

• S. Bravyi, M. B. Hastings, and F. Verstraete, “Lieb–Robinson Bounds and the Generation of Correlations and Topological Quantum Order,” Physical Review Letters 97, 050401, 2006, doi:10.1103/PhysRevLett.97.050401.
• M. B. Hastings and T. Koma, “Spectral Gap and Exponential Decay of Correlations,” Communications in Mathematical Physics 265, 781–804, 2006, doi:10.1007/s00220-006-0030-4.
• E. H. Lieb and D. W. Robinson, “The Finite Group Velocity of Quantum Spin Systems,” Communications in Mathematical Physics 28, 251–257, 1972, doi:10.1007/BF01645779.
• B. Nachtergaele and R. Sims, “Lieb–Robinson Bounds in Quantum Many-Body Physics,” Contemporary Mathematics 529, 141–176, 2010, doi:10.1090/conm/529/10429.
• B. Nachtergaele, R. Sims, and A. Young, “Quasi-Locality Bounds for Quantum Lattice Systems. Part I. Lieb–Robinson Bounds, Quasi-Local Maps, and Spectral Flow Automorphisms,” Journal of Mathematical Physics 60, 061101, 2019, doi:10.1063/1.5095769.