Quantum Scars

In the world of complex, interacting systems, chaos is often expected to be the great equalizer, driving any system toward thermal equilibrium and erasing all memory of its initial state. This principle, formalized for quantum systems as the Eigenstate Thermalization Hypothesis (ETH), suggests that at high energies, order gives way to a uniform, featureless blur. Yet, both simulations and experiments have revealed a startling exception to this rule: the existence of "quantum scars," special, non-thermal eigenstates that stubbornly persist within an otherwise chaotic environment. These scars pose a fundamental challenge to our understanding of quantum thermalization, representing a hidden layer of order and memory where none was thought possible. This article delves into the fascinating world of quantum scars. The first chapter, ​​Principles and Mechanisms​​, will uncover what these scars are, exploring the surprising connection to unstable classical orbits and the wave interference that gives them life, both in single-particle and complex many-body systems. Following this, the chapter on ​​Applications and Interdisciplinary Connections​​ will examine the real-world implications of this phenomenon, from its tangible signatures in quantum simulators to its potential for revolutionizing quantum computing and sensing.

Imagine a perfectly pristine snooker table, but shaped like a racetrack—a rectangle capped by two semicircles. This is the famous "stadium billiard," a classic playground for physicists studying chaos. If you were to hit a ball on this table, its path would be a textbook example of classical chaos. An infinitesimally small change in the ball's initial direction or speed would lead to a wildly different trajectory after just a few bounces. Over time, the ball would seem to visit every part of the table with equal probability. The long-term behavior is a featureless, uniform blur.

Now, let's switch from a classical ball to a quantum particle. The correspondence principle, a cornerstone of physics, suggests that at high energies, the quantum world should smoothly merge with the classical one. So, we'd expect a high-energy quantum particle in this stadium to also create a uniform probabilistic blur throughout the entire space. The particle's wavefunction, when squared to give its probability density, should be spread out evenly, just like the classical ball's long-term location.

But experiments and simulations reveal something astonishing. Amidst the expected quantum blur, certain high-energy states exhibit beautiful, intricate patterns. The probability of finding the particle is mysteriously enhanced along specific, well-defined lines. These ghostly patterns, a form of quantum memory in a system designed to forget, are called ​​quantum scars​​. Their existence is a profound wrinkle in the simple story of quantum-classical correspondence, showing that even in chaos, quantum mechanics can preserve a remarkable degree of order and structure.

What are these patterns tracing? In a shocking twist, they are echoes of ​​unstable periodic orbits​​ (UPOs) from the classical system. This is deeply counter-intuitive. A periodic orbit is a path that a classical particle could traverse over and over again. An unstable one is a path that is incredibly sensitive; like balancing a pencil on its tip, any tiny deviation sends the particle flying away. Why on Earth would a quantum particle, with its inherent fuzziness, "choose" to hang around a path that is classically so precarious?

The answer lies not in trapping the particle, but in the subtle magic of wave interference. Imagine launching a tiny quantum wave packet—a localized bundle of waves—along one of these unstable orbits. As it travels, the chaotic nature of the billiard causes the wave packet to spread out, just as a group of runners starting together would quickly disperse in a confusing maze. However, the very same geometry that causes this spreading can also, for a periodic orbit, have a re-focusing effect. After the wave packet completes one circuit, some part of it is guided back to the starting point. This process creates a series of "recurrences"—faint echoes of the initial wave packet returning home after one, two, three, or more trips around the orbit.

Each returning echo is weaker than the last, its amplitude diminished by the orbit's instability. Classically, this instability means the particle is long gone. But quantum mechanically, these faint echoes are all that matter. If the phases of these returning waves align just right, they undergo ​​constructive interference​​. They build upon one another, creating a surprisingly large wave amplitude precisely along the path of the unstable orbit. This is the heart of a scar: a quantum resonance effect, a conspiracy of wave phases that creates order out of chaos.

Of course, not all unstable orbits are created equal. The more unstable an orbit, the faster the returning echoes fade, and the harder it is for them to build up a significant interference pattern. Consequently, quantum scars are preferentially formed along the least unstable of the unstable periodic orbits. By analyzing the classical dynamics and identifying the UPOs with the smallest instability (characterized by a property called the Lyapunov exponent or, on a Poincaré section, a stability eigenvalue closest to 1), we can predict which classical paths are most likely to leave their imprint on the quantum world.

Now that we understand the mechanism, what does a scar actually look like? It's not a line of infinite density, but rather a "tube" of enhanced probability surrounding the classical path. In a simple model, we can imagine the probability density $|\psi(\mathbf{r})|^2$ inside this tube being a constant factor $\alpha$ times larger than the density outside. This factor $\alpha$ quantifies the scar's strength or "visibility."

To make this concrete, imagine you are given a picture of a probability density from a simulation. How would you objectively decide if it's scarred? You could define a tube of a certain width around a suspected classical orbit and calculate the average probability density within that tube. By comparing this "tube average" to the overall average density and its standard deviation across the entire system, you can compute a statistical score—like a Z-score—that tells you how exceptionally bright the tube is. A high score flags the presence of a scar.

This raises another beautiful question: what determines the physical width of the scar's tube? The answer provides a stunning link between the classical instability and the quantum uncertainty principle. The width of the scar in position, let's call it $\sigma_q$, is linked to its "fuzziness" in momentum, $\sigma_p$. The dynamics of the orbit are described by a stability matrix $M$, which tells us how small deviations in position and momentum evolve. The element $m_{12}$ of this matrix couples an initial uncertainty in momentum to a resulting spread in position after one period. A self-consistent picture emerges when we demand that the inherent spatial width of the scar, $\sigma_q$, must be equal to the spread it acquires in one trip due to its own momentum uncertainty. This simple physical argument, combined with Heisenberg's uncertainty principle, $\sigma_q \sigma_p \approx \hbar/2$, leads to a wonderfully simple formula for the scar's width:

Here, $\hbar$ is the reduced Planck constant. This equation is remarkable. It directly connects a feature of the classical orbit ($m_{12}$) to the size of a quantum object ($\sigma_q$). The more unstable the orbit (the larger $m_{12}$), the wider the resulting scar.

For those who appreciate the deep mathematical underpinnings, the formation of a scar is beautifully analogous to the ​​Stokes phenomenon​​ in the theory of asymptotic expansions. When approximating functions, one often finds that an exponentially small (subdominant) term, which is usually ignored, can suddenly and dramatically "switch on" and become important in certain regions of the complex plane. In the same way, the self-consistency of the wavefunction along the UPO requires us to account for subdominant wave components. These components, born from the chaos, switch on just right to interfere constructively and form the scar.

The idea of scars was born in the relatively simple world of a single particle in a billiard. But in recent years, the concept has been reborn in the far more complex and fascinating realm of quantum many-body systems—systems of countless interacting particles, like electrons in a solid or atoms in a quantum computer.

Here, the central question is no longer "Where is the particle?" but "Does the system thermalize?" The prevailing theory for chaotic many-body systems is the ​​Eigenstate Thermalization Hypothesis (ETH)​​. ETH is the modern, quantum version of the uniform blur we saw in the classical billiard. It states that in a complex, isolated quantum system, every individual high-energy eigenstate is, by itself, a thermal state. Any local measurement performed on such a state will yield the same result that a conventional thermometer would, as predicted by the laws of statistical mechanics. The system's complete loss of memory about its initial state is encoded in the thermal nature of each of its constituent eigenstates.

And yet, once again, quantum mechanics delivers a surprise. Certain non-integrable systems—systems that are expected to be fully chaotic and obey ETH—were found to possess a small number of very special, atypical eigenstates that stubbornly refuse to thermalize. These states are the modern incarnation of scars: ​​many-body quantum scars​​.

These scar states are rare exceptions. Their number grows only polynomially with the system size $L$ (e.g., like $L^2$), while the total number of eigenstates in the same energy range explodes exponentially (like $2^L$). This means that in the thermodynamic limit of a large system, the fraction of scar states becomes zero. They are a "measure-zero" subset of the full spectrum. This is why many-body scars are said to be a ​​weak violation of ETH​​. The "strong" version of ETH, which says every state is thermal, is broken. But a "weak" version, which says almost every state is thermal, still holds [@problem__id:3004298]. The system as a whole is still thermalizing, but it has these hidden, non-thermal skeletons in its closet.

Since a many-body scar state is not a simple picture of probability density, how do we identify one? We need a new set of tools, new diagnostics to measure its "non-thermalness."

One of the most powerful tools is ​​entanglement entropy​​. Thermal eigenstates are incredibly complex, containing a massive amount of entanglement between different parts of the system. For a subsystem of size $\ell$, the entanglement entropy should scale with its volume, a "volume law" ($S_A \propto \ell$). Many-body scar states, however, are far more structured and have anomalously low entanglement. Their entropy often scales much more slowly, for instance logarithmically ($S_A \propto \ln \ell$). This sub-volume-law entanglement is a smoking gun for a scar.

Another way is to look at expectation values of simple, local observables. According to ETH, these values should be nearly the same for all eigenstates at a given energy. For a scar state, they can be drastically different. We can even quantify this difference by calculating the ​​fidelity​​, a measure of overlap, between a scar state and a true thermal state at the same energy. For a perfect scar, this fidelity will be significantly less than one, providing a direct number that says, "This state is not thermal".

Perhaps the most dramatic consequence of many-body scars is revealed in the system's dynamics. If you prepare the system in a generic initial state, it will quickly thermalize, as ETH predicts. But if you are clever and prepare the system in a special initial state that has a large overlap with the scar states, the system's evolution will be anything but thermal. It will exhibit stunning, long-lived oscillations and periodic revivals, seeming to remember its initial configuration for an impossibly long time, just as a single wave packet's echoes returned in the stadium billiard.

From a simple billiard to the frontiers of quantum statistical mechanics, the story of quantum scars is a tale of unexpected order hiding within chaos. They represent the remarkable ability of quantum coherence to persist against the randomizing forces of a complex environment, revealing a deep and beautiful structure that lies just beneath the surface of our classical intuition.

Now that we have grappled with the peculiar mechanics of quantum scars, you might be left with a perfectly reasonable question: So what? Are these bizarre, non-thermal states merely a physicist's curiosity, a subtle footnote in the grand story of quantum statistical mechanics? Or do they represent something deeper, a principle that echoes across different fields of science and might even be harnessed for our own purposes?

The story of scars is a wonderful illustration of how nature works. It is not always a story of brute-force rules that apply everywhere, but often one of subtle exceptions, hidden pathways, and surprising connections. Scars are precisely such a pathway—a loophole in the otherwise iron-clad law of thermalization. As we follow this thread, we’ll find it weaving through the landscape of physics, from the untamed wilderness of classical chaos to the futuristic architecture of quantum computers.

Before "many-body scars" became a buzzword in quantum information, the original "quantum scar" was discovered in a very different context: the quantum mechanics of a single particle in a chaotic environment. Imagine a particle, like an electron, bouncing around inside a two-dimensional cavity, a "quantum billiard." If the shape of the billiard is irregular, like a stadium, the classical path of a ball inside it would be chaotic. A tiny change in its initial position or velocity leads to a wildly different trajectory. This is the famous "butterfly effect."

One might expect the quantum wavefunctions in such a system to be completely random, a turbulent, featureless sea of probability filling the entire space. And for most eigenstates, that is true—they look like random superpositions of plane waves, in perfect agreement with our expectations from statistical mechanics. But not all of them. Here and there, one finds eigenstates that are mysteriously, inexplicably "brighter" along the path of an unstable periodic orbit of the classical system. It is as if the ghost of a classical trajectory, one that a real classical particle could never stay on for long, is haunting the quantum world, leaving a "scar" on the fabric of the wavefunction.

This is not just a poetic metaphor; the connection is quantitative. The very instability of the classical orbit, characterized by its Lyapunov exponent $\lambda$, dictates the physical width of its quantum scar. Semiclassical analysis reveals that the transverse width of the scar, $w$, is tied directly to the classical chaos and a fundamental quantum constant, $\hbar$. For a scar that is as localized as quantum mechanics allows, it must obey the Heisenberg uncertainty principle. Combining this with the dynamics of the unstable orbit gives a beautifully simple result:

where $m$ is the mass of the particle. Think about what this means! The more unstable the classical orbit (larger $\lambda$), the more rapidly nearby classical paths diverge. Quantum mechanically, this forces the wavefunction to be "fatter" in position space to satisfy the uncertainty principle, resulting in a wider scar. It’s a delicate dance between classical instability and quantum uncertainty, a direct bridge between two worlds.

The modern excitement around scars ignited when a similar phenomenon was discovered not in a single particle, but in complex, interacting many-body systems. The perfect stage for this discovery turned out to be arrays of Rydberg atoms—atoms excited to very high energy levels. These systems, which are a leading platform for quantum simulation and computation, can be described by wonderfully effective but constrained models, like the "PXP model."

The most dramatic signature of many-body scars is their defiance of time. If you prepare a generic quantum system in a simple state, like a checkerboard pattern of spins, it will rapidly evolve into a complicated, high-entropy mess. Information about the initial state gets scrambled and appears lost to local measurements; the system thermalizes. But if you prepare a Rydberg atom chain in a specific simple state—the antiferromagnetic Néel state ($|1010\dots\rangle$)—something astonishing happens. The system evolves, scrambles, and then, against all odds, periodically returns to its initial state, like a magic trick. This phenomenon of perfect "fidelity revivals" is the calling card of quantum scars.

Why does this happen? It turns out that some special systems possess a "tower" of scar eigenstates whose energies are almost perfectly equally spaced, like the rungs of a ladder: $E_n = E_0 + n \hbar \omega$. When the initial state is a superposition of these scar states, each component evolves with a phase $e^{-iE_n t / \hbar}$. The regular energy spacing means that at a specific time, $t_{rev} = 2\pi/\omega$, all these phases realign perfectly, and the initial state is reborn. The system rings with a clear, single frequency $\omega$, instead of the discordant noise of a thermalizing system.

Of course, this magic trick only works if the initial state has a substantial overlap with these special scar states to begin with. And remarkably, for the PXP model, the simple, experimentally preparable Néel state does! It's not an eigenstate itself, but it carries a significant "imprint" of the scar states within it. This fortuitous overlap is what makes observing scarred dynamics experimentally feasible.

And this isn't just a story about Rydberg atoms. The same underlying models and scarred phenomena are predicted to appear in other quantum information platforms, such as chains of interacting nitrogen-vacancy centers in diamond. This universality suggests that scarring is not an accident of one particular system, but a more general principle of constrained quantum dynamics.

The ripples from this discovery have spread to the heart of condensed matter physics. One of the central challenges in this field is understanding the behavior of strongly correlated electrons in materials. The Hubbard model, a cornerstone for describing such systems, is notoriously difficult to solve. Yet, it turns out that in certain limits and on specific lattice geometries, this very model can host quantum many-body scars.

This implies that phenomena analogous to the revivals in cold atoms could, in principle, occur within the electronic structure of some materials, particularly those near a Mott metal-insulator transition. What makes these scar states so distinct in the language of many-body physics is their entanglement structure. A typical eigenstate at high energy in a chaotic system is expected to have "volume-law" entanglement—the entanglement entropy of a subsystem is proportional to its size, $L$. This signifies maximum scrambling. Scar states, however, defy this. Their entanglement grows much more slowly, often just logarithmically with system size ($S_A \propto \ln L$). This "sub-volume-law" entanglement is a smoking gun, a definitive fingerprint that these states are not thermal. They carry a form of hidden, non-local order that is invisible to simple probes.

Furthermore, scar states are not just "less entangled"; they have specific, structured properties defined by the symmetries of the system. For instance, certain scar states in the PXP model must have a staggered magnetization of exactly zero due to their definite parity under spatial inversion. They are not random at all, but are as specific and structured as the ground state, just living at a high energy density.

At this point, you might be thinking this all sounds too good to be true. Can we really cheat the second law of thermodynamics forever? The answer is no. Quantum scars are not a perpetual motion machine against thermalization. They are fragile.

Their beautiful coherent dynamics rely on the system being perfectly isolated from the outside world. In any real experiment, the system is inevitably coupled to an external environment, or "bath." This coupling acts as a source of noise, introducing randomness that breaks the delicate phase relationships responsible for the revivals. The fidelity oscillations do not last forever; they decay, and the system eventually thermalizes. Scars, therefore, don't prevent thermalization, but they can create a long-lived, intermediate regime known as ​​prethermalization​​, where the system remembers its origin for an anomalously long time before succumbing to amnesia.

Even in a hypothetically perfectly isolated system, a scar state will eventually decay. The "bath" it couples to is the sea of other thermal eigenstates that exist at the same energy within the system itself. The reason scars are long-lived is because they are, by their very nature, anomalously weakly coupled to this thermal sea. The rate of decay, $\Gamma$, can be understood through a competition: the sheer number of available thermal states to decay into, which is exponentially large in the system size $N$ and related to the entropy $s_S$, versus the exponentially weak matrix elements connecting the scar to these states, characterized by a parameter $\gamma$. Fermi's Golden Rule gives us a beautifully expressive formula for this process: