Nonlocal nonstabilizerness in free fermion models

This paper investigates nonlocal magic in free fermion systems by deriving a closed-form bound for Gaussian states, demonstrating its extensive nature in random ensembles, observing its suppression in gapped phases with peaks at criticality, and revealing a distinct separation between nonlocal magic and entanglement in XY chain quenches.

The Gist

Imagine you have a complex quantum system, like a collection of tiny magnets or particles, and you want to know how "truly quantum" it is. Scientists already have a ruler for one part of this: entanglement. Entanglement is like a super-strong glue that ties two parts of a system together so tightly that you can't describe one without the other.

However, the authors of this paper argue that entanglement isn't the whole story. You can have a lot of glue (entanglement) but still be able to simulate the system easily on a regular computer. To be truly powerful and "quantum," a system needs something else: Magic.

In the world of quantum computing, "Magic" (or non-stabilizerness) is the special ingredient that makes a system hard to simulate classically. It's the difference between a simple, predictable puzzle and a chaotic, unsolvable one.

Here is a breakdown of what the paper does, using simple analogies:

1. The Problem: Separating the "Local" from the "Global"

The authors are interested in Nonlocal Magic. Think of a quantum state as a giant, intricate tapestry woven by two people, Alice and Bob, sitting at opposite ends of a room.

• Local Magic: This is the complexity Alice or Bob could create just by rearranging their own thread (changing their local perspective).
• Nonlocal Magic: This is the complexity that remains even after Alice and Bob have done everything possible to simplify their own threads. It is the irreducible, "spooky" connection that exists between them. You can't get rid of it by just looking at your own side of the room.

Calculating this is usually incredibly hard, like trying to find the shortest path through a maze that changes shape every time you look at it.

2. The Solution: A Simple Formula for "Free Fermions"

The paper focuses on a specific type of quantum system called free fermions (particles that don't interact with each other in complex ways, like electrons in a simple metal).

• The Analogy: Imagine the system is a set of independent dancers. Even though they are dancing together, they aren't bumping into each other.
• The Breakthrough: The authors found a simple, closed-form formula (a neat mathematical recipe) to calculate the Nonlocal Magic for these systems. Instead of needing a supercomputer to solve a maze, they realized the answer depends entirely on the entanglement spectrum.
• The Metaphor: Think of the entanglement spectrum as a list of "dance pairs." Some pairs are dancing perfectly in sync (maximally entangled), some are dancing alone (not entangled), and some are in the middle. The authors found that the "Magic" only comes from the pairs that are in the middle—the ones that are entangled but not perfectly so. If the pairs are too simple or too complex, the Magic disappears.

3. Testing the Theory: The "Simulated Annealing" Check

To make sure their simple formula was actually the best possible answer, they ran a computer simulation called simulated annealing.

• The Analogy: Imagine trying to find the lowest point in a hilly landscape. You start at a random spot and take random steps. If you step downhill, you stay. If you step uphill, you might stay anyway (to avoid getting stuck in a small valley), but as time goes on, you become less likely to step uphill. This helps you find the absolute lowest valley.
• The Result: They ran this "search" over millions of possible local changes to the system. Every time, the lowest point they found matched their simple formula. This suggests their formula is indeed the "gold standard" for these systems.

4. What Happens in Random Systems?

They looked at what happens if you take a bunch of these systems and make them completely random (like shuffling a deck of cards).

• The Finding: The average amount of Nonlocal Magic grows steadily as the system gets bigger (it is "extensive"). However, it's still a relatively small amount compared to the total "quantumness" of the system. It's like finding a specific spice in a giant pot of soup; it's there, but it's a tiny fraction of the total volume.

5. The Kitaev Chain: A Quantum Phase Transition

The authors studied a famous model called the Kitaev chain, which can be in two different "phases":

• Trivial Phase: Like a calm, frozen lake.
• Topological Phase: Like a lake with a hidden, swirling current.
• The Critical Point: The exact moment the lake freezes or thaws.
• The Result: Deep inside the calm lake or the swirling current, the Nonlocal Magic is very low (suppressed). But right at the critical point (the phase transition), the Magic peaks.
• The Metaphor: It's like a crowd of people. When everyone is sitting still (trivial) or everyone is marching in perfect lockstep (topological), there is no "chaotic energy." But right when the crowd is deciding to stand up and start moving, there is a burst of chaotic, unpredictable energy. The Nonlocal Magic measures this burst.

6. Time and Dynamics: The XY Chain

Finally, they watched how this Magic changes over time when the system is shaken (a "quench").

• Random Circuits: When they used random gates to shake the system, the Magic grew like a drop of ink spreading in water (diffusively).
• The XY Chain (The Surprise): When they studied a specific version of the chain (the XX limit), they found something strange.
  • Entanglement (the glue) grew quickly and linearly, like a car speeding down a highway.
  • Nonlocal Magic (the complexity) grew very slowly, only logarithmically (like a snail).
• The Conclusion: This reveals a separation. In this specific case, the system becomes highly entangled (glued together) very fast, but it doesn't become "magical" (hard to simulate) at the same speed. The "glue" is there, but the "chaos" is missing. This happens because a specific symmetry (charge conservation) acts like a brake, preventing the Magic from building up, even though the entanglement is growing.

Summary

In short, this paper provides a simple, reliable way to measure the "irreducible quantum complexity" of a specific class of particles. They found that this complexity:

1. Is easy to calculate for these systems.
2. Peaks when the system is changing phases (critical points).
3. Can behave very differently from entanglement, sometimes growing much slower, revealing that a system can be "glued" together without necessarily being "complex" in a useful way.

Technical Summary

1. Problem Statement

The paper addresses the quantification of nonlocal magic (or nonlocal nonstabilizerness) in many-body quantum systems. While entanglement is a well-established resource for quantum advantage, it is insufficient on its own, as highly entangled "stabilizer states" (generated by Clifford circuits) can be efficiently simulated classically. Magic quantifies the non-Clifford resources required to prepare a state.

The central challenge is defining and computing nonlocal magic ($M^{NL}$), which represents the irreducible nonstabilizerness of a bipartite state after optimizing over local basis changes (local unitaries).

• The Difficulty: For generic many-body states, calculating $M^{NL}$ is computationally intractable because magic measures are nonlinear, and the local unitary orbit is exponentially large.
• The Gap: Explicit results for nonlocal magic in many-body settings are scarce. The authors aim to bridge this gap by focusing on pure fermionic Gaussian states, a class of states relevant to free-fermion systems, where analytical progress is possible.

2. Methodology

The authors employ a combination of analytical derivations, random matrix theory, and numerical simulations:

• Formalism: They utilize the $\alpha$-Stabilizer Rényi Entropy (SRE) as the measure of magic. For fermionic Gaussian states, the SRE can be expressed entirely in terms of the covariance matrix $\Gamma$ and its minors (via Wick's theorem and Pfaffians).
• Canonical Form Derivation: They derive a closed-form upper bound for nonlocal magic by transforming the bipartite covariance matrix $\Gamma$ into a local canonical form using local Gaussian unitary transformations ($O_A \oplus O_B$). This decomposition isolates independent entangled pairs characterized by singular values $\{\nu_k\}$ of the subsystem covariance matrix.
• Numerical Benchmarking: To test the tightness of their analytical bound, they perform simulated annealing over the local Gaussian unitary orbit. They minimize the SRE numerically for random Gaussian states and compare the results with their analytical bound.
• Random Matrix Theory (RMT): For the Gaussian Haar ensemble (random pure Gaussian states), they use the theory of the Jacobi ensemble (specifically the DIII circular ensemble) to compute the thermodynamic limit of the average nonlocal magic density.
• Model Applications: They apply their framework to:
  • The Kitaev chain (ground states).
  • Random Gaussian brick-wall circuits (dynamics).
  • The XY chain (Hamiltonian quenches).

3. Key Contributions

A. Analytical Bound for Gaussian States

The authors derive a simple, closed-form upper bound for nonlocal magic, denoted $M^{NL}_{\alpha,>}(\psi_\Gamma)$, which depends solely on the entanglement spectrum (singular values $\nu_k$ of the restricted covariance matrix):
$$M^{NL}_{\alpha,>}(\psi_\Gamma) = \frac{1}{1-\alpha} \sum_k \log \left( \frac{1 + \nu_k^{2\alpha} + \mu_k^{2\alpha}}{2} \right)$$
where $\mu_k = \sqrt{1-\nu_k^2}$.

• Significance: Unlike general bounds which are intricate, this expression is a sum of independent mode contributions. It vanishes for locally pure modes ($\nu_k=1$) and maximally entangled stabilizer-like pairs ($\nu_k=0$), isolating the "intermediate" modes responsible for non-Gaussian correlations.

B. Optimality Verification

• Local Optimality: They prove that the canonical form corresponds to a local minimum of the SRE along the local Gaussian orbit.
• Global Optimality (Numerical): Simulated annealing benchmarks on small systems ($L=8$) show that the analytical bound is never crossed by the numerical minimization, providing strong evidence that the bound is tight (i.e., it captures the global minimum over local Gaussian unitaries).

C. Thermodynamic Limit for Random States

For the Gaussian Haar ensemble, they show that the average nonlocal magic is extensive (scales linearly with system size $L$). They derive a closed-form expression for the magic density $J(\ell)$ in the thermodynamic limit ($L \to \infty$) as a function of the bipartition fraction $\ell = m/L$:
$$J(\ell) = \text{Re} \left( \dots \right)$$
The density peaks at the symmetric bipartition ($\ell=0.5$) but contributes only a small fraction to the total non-stabilizerness compared to Haar-random qubit states.

D. Phase Transitions and Criticality

In the Kitaev chain, they find that nonlocal magic is:

• Suppressed deep in both the trivial and topological gapped phases.
• Enhanced near the critical point ($\mu = 2$), where it exhibits a sharp peak.
• Logarithmically scaling with system size at criticality ($M^{NL} \sim \log L$), reflecting the broadening of the entanglement spectrum.

E. Dynamical Separation from Entanglement

The study of dynamics reveals a striking distinction between entanglement growth and nonlocal magic growth:

• Random Circuits: Nonlocal magic grows diffusively ($t^{1/2}$).
• XY Chain Quenches:
  • For generic parameters ($\gamma \neq 0$), magic grows linearly in time, similar to entanglement.
  • Crucial Finding: At the U(1)-symmetric XX point ($\gamma = 0$), entanglement grows linearly, but nonlocal magic grows only logarithmically.
  • Interpretation: The linear growth of entanglement at $\gamma=0$ is driven by a plateau of near-stabilizer modes ($\nu \approx 0$) in the spectrum. Since these modes are close to stabilizer states, they contribute negligibly to nonlocal magic. This demonstrates that nonlocal magic is sensitive only to the subleading crossover in the spectrum, not the leading ballistic quasiparticle contribution.

4. Results Summary

1. Closed-Form Bound: A simple, efficiently computable formula for nonlocal magic in fermionic Gaussian states based on the entanglement spectrum.
2. Extensivity: Nonlocal magic is extensive in random Gaussian ensembles, unlike in Haar-random qubit states where it scales differently.
3. Criticality: Nonlocal magic acts as a sensitive probe for quantum phase transitions, peaking at critical points in the Kitaev chain.
4. Dynamical Decoupling: In the XX limit of the XY chain, nonlocal magic and entanglement exhibit qualitatively different growth rates (logarithmic vs. linear), highlighting that charge conservation suppresses the buildup of irreducible nonstabilizerness.

5. Significance

• New Resource Metric: The paper establishes nonlocal magic as a viable and computable resource metric for free-fermion systems, a regime where magic was previously difficult to characterize.
• Beyond Entanglement: It provides concrete evidence that nonlocal magic captures physical phenomena distinct from entanglement, particularly in symmetry-protected dynamics (e.g., the XX limit).
• Computational Efficiency: The derived bound allows for the efficient characterization of non-Clifford resources in large many-body systems without the need for exponential classical simulation.
• Future Directions: The work opens avenues for studying symmetry-resolved magic, magic in topological phases of free fermions, and magic in monitored or driven dynamics.

In conclusion, this work provides a rigorous theoretical framework for quantifying the "quantumness" (beyond entanglement) of free-fermion states, revealing deep connections between magic, criticality, and dynamical symmetries.