Subsystem Symmetries

Symmetry is one of the most powerful and elegant principles in physics, dictating everything from fundamental conservation laws to the classification of elementary particles. Typically, we think of symmetries as global rules—the total charge or energy in a closed system is conserved. But what if nature could be more parochial? What if conservation laws were enforced not on the whole universe, but independently on countless lower-dimensional slices within it, like on every line of latitude on a globe? This question opens the door to the strange and fascinating world of subsystem symmetries. This article addresses the knowledge gap that arises when we move beyond conventional global symmetries, exploring the novel phenomena that emerge from these constrained rules.

This exploration is divided into two parts. In the first chapter, ​​Principles and Mechanisms​​, we will delve into the fundamental workings of subsystem symmetries, uncovering the mathematical machinery that enforces them and see how they give rise to bizarre, immobile particles called fractons in celebrated theoretical frameworks like the X-cube model. In the second chapter, ​​Applications and Interdisciplinary Connections​​, we will survey the profound impact of these ideas, from their potential to revolutionize quantum computing with error-resistant codes to their role in defining entirely new, robust phases of quantum matter. To begin our journey, we must first understand the core rules of this new game and how they differ from the symmetries we thought we knew.

So, we've been introduced to a curious new idea in the world of symmetries. But what are they, really? How do they work? To get a feel for them, let us first remember what we usually mean by a symmetry in physics. Think about electric charge. The total charge in the universe is conserved. If you take a sealed box, the total charge inside that box can only change if a current of charges flows through its walls. This is a ​​global conservation law​​. The rule applies to the entire system. Nature's bookkeeping is perfect, but it only cares about the grand total.

Now, let's imagine a world with a much stricter, almost bizarre, set of rules. What if, on our planet, charge was not only conserved globally, but was conserved on every single line of latitude? A charge created in New York could move east or west along the 40th parallel, perhaps to Madrid or Beijing, but it could never travel north to Montreal or south to Rio de Janeiro. It is eternally confined to its line. This is the essence of a ​​subsystem symmetry​​. The symmetry transformation—and the corresponding conservation law—doesn't apply to the whole system at once, but to independent, lower-dimensional slices of it. It’s a local edict, enforced line by line, plane by plane.

How could Nature possibly enforce such a peculiar rule? A simple law like "total charge is constant" won't do. You need a more sophisticated mechanism, a clever piece of mathematical machinery. The trick, it turns out, is to define the charge itself in a more intricate way.

In our familiar world, we have a charge density $\rho$ and a current density vector $\mathbf{j}$, related by the continuity equation $\partial_t \rho + \nabla \cdot \mathbf{j} = 0$. This ensures global conservation. To enforce conservation on, say, every horizontal line in a 2D plane, we need something stronger. Imagine if the charge density $\rho$ were itself the derivative of some other quantity along the conserved direction. For example, what if the charge density at a point $(x, y)$ was given by an expression like $\rho(x,y) = \partial_x F(x,y)$ for some function $F$?

Then the total charge along a line at a specific height $y_0$ would be $Q(y_0) = \int_{-\infty}^{\infty} \rho(x, y_0) \, dx = \int_{-\infty}^{\infty} \partial_x F(x, y_0) \, dx$. By the fundamental theorem of calculus, this integral is just the value of $F$ at the endpoints: $F(\infty, y_0) - F(-\infty, y_0)$. If our fields are well-behaved and vanish far away, this total charge is automatically zero! Or, more generally, it's conserved. A charge dipole can exist on the line, but a net charge cannot be created out of thin air.

This is precisely the kind of mechanism at play in theories with subsystem symmetries. These theories are often described not by the familiar vector potentials ($A_\mu$) of electromagnetism, but by more complex objects called ​​symmetric tensor gauge fields​​ ($A_{\mu\nu}$). These fields naturally give rise to charge densities that are spatial derivatives of field components. For instance, in a (2+1)-dimensional theory, the charge density might take the form $\rho = \mathcal{N} \cdot \partial_x ( \partial_x A_{0y} - \partial_y A_{0x} )$. This structure mathematically guarantees that the charge is conserved along individual lines, just as we imagined. The rules of the game are encoded in the very structure of the fields themselves.

This might sound like a theorist's fantasy. Can we actually construct a physical system that obeys these rules? The answer is a resounding "yes," and the most celebrated example is a toy model of quantum spins called the ​​X-cube model​​.

Imagine a vast three-dimensional cubic lattice, and on every edge of every cube, there sits a tiny quantum bit, or ​​qubit​​. The interactions between these qubits are governed by a very specific set of local rules (the Hamiltonian). One rule involves the 12 qubits on the edges of any elementary cube, and another rule involves the qubits that meet at a vertex. A system in its lowest energy state, the ground state, satisfies all these rules perfectly.

Now, if we violate one of the vertex rules, we create an excitation, a tiny point of energy. These excitations are the "charges" of this system, and they have a spectacular property: they are completely, utterly immobile. We call them ​​fractons​​. A single fracton, once created, is stuck. It cannot hop to a neighboring vertex. It is pinned in place by the very laws that define its existence.

This bizarre immobility is a direct consequence of a set of planar subsystem symmetries. In the X-cube model, the number of fractons (modulo 2) is conserved independently within each $xy$-plane, each $yz$-plane, and each $xz$-plane of the lattice. To move a fracton from one point to another, you would have to violate this conservation law. The only way to move a fracton at all is to create an anti-fracton nearby, forming a dipole. This neutral pair can move, but only in a limited way—for example, a dipole of fractons created on a z-axis can only separate along that axis. The individual charges are immobile, but certain composites of them can move, restricted to lines or planes.

The operators that generate these symmetries are themselves fascinating. For each plane in the lattice, one can construct an operator—a product of simple Pauli spin operators—that acts on all the qubits within that plane. These are the generators of the subsystem symmetry. Interestingly, these generators are not always fully independent. For a stack of $N \times N$ planes, while you can construct $N^2$ such generators, there is a single global constraint among them, leaving $N^2-1$ truly independent symmetry operations. This reveals a subtle interconnectedness even among these independent subsystems.

The name "fracton" is no accident; it hints at a deep and beautiful connection to ​​fractals​​. The operators that create and move these constrained particles, or the symmetry operators themselves, can have a fractal structure.

Consider a thought experiment. Let's try to build the operator that generates a subsystem symmetry through a recursive process. We start with a single point. In the next step, we replace that point with a small cluster of four points. Then, in the step after that, we replace each of those four points with its own identical cluster, scaled down. We repeat this process ad infinitum. The dusty, lacy object we create is a fractal. If we measure how the number of points in this object scales with its size, we find it has a ​​fractal dimension​​ of $D_f=2$. It's a ghostly plane-like object living in three dimensions. The symmetry itself is a fractal!

This connection becomes even more dramatic when we consider how fractons behave in a fractal environment. Imagine taking a system that hosts fractons and placing it in a potential landscape shaped like a ​​Menger sponge​​—that famous fractal cube full of holes. We'll see later that this has profound consequences for the system's energy, but for now, it's enough to appreciate how the physics of constrained mobility naturally intertwines with the geometry of fractals.

Symmetries in the real world are often not perfect. Sometimes, a system in its ground state must "choose" a configuration that doesn't respect the full symmetry of the laws governing it. This is called ​​spontaneous symmetry breaking​​. A profound result known as ​​Goldstone's theorem​​ tells us what happens next: the system gains the ability to support massless, wave-like excitations called Goldstone bosons. They are the ripples of the system exploring the other equivalent ground states it could have chosen. For a normal symmetry, these waves move at a constant speed, and their frequency $\omega$ is proportional to their momentum $k$.

But what happens when you spontaneously break a subsystem symmetry? You get a twisted, constrained version of Goldstone's theorem. The resulting Goldstone modes are not free to travel in all directions. Their movement is restricted, reflecting the nature of the broken symmetry. In a system with a broken planar subsystem symmetry, for instance, the Goldstone mode might have a bizarre dispersion relation like $\omega(\mathbf{k}) = A |k_y k_z|$. This means the wave has zero energy unless it has momentum components in both the $y$ and $z$ directions. It's a mode that can only propagate diagonally across a plane, a ghost of a normal particle.

Things get even stranger if the symmetry is also explicitly broken by a small external field. The "would-be" Goldstone bosons are no longer massless; they become ​​pseudo-Goldstone bosons​​ with a small mass gap. For fractonic systems, the consequences are even more exotic. If we take our system with its broken subsystem symmetry and place it in that Menger sponge potential, the lowest-energy excitation—the pseudo-Goldstone fracton—gets trapped in the largest holes of the fractal. Its energy gap doesn't settle on a constant value but scales with the overall size of the system $L$ as $\Delta \propto L^{-2}$. This is completely unlike any ordinary particle in a box; it is another signature of the strange, constrained world of fractons.

The story does not end there. Perhaps the most beautiful aspect of this physics is how different kinds of symmetries—subsystem symmetries, spatial symmetries of the lattice, global symmetries—can talk to each other, forming a rich and intricate structure.

For example, performing a simple spatial rotation on the X-cube lattice can transform one subsystem symmetry into another. A $90^\circ$ rotation around the $z$-axis will map the symmetry that acts on the $x=k$ plane into the symmetry that acts on the $y=k$ plane. This isn't just a relabeling; it shows a fundamental entanglement between the geometry of space itself and the internal rules of the system. Symmetries can act on other symmetries.

The deepest manifestation of this interplay is the concept of a ​​'t Hooft anomaly​​. You can think of an anomaly as a "quantum pact" between two symmetries that makes them inseparable. In the X-cube model, there is a global symmetry (flipping all spins everywhere) and the planar subsystem symmetries we've discussed. It turns out these two sets of symmetries have a mixed anomaly. This anomaly is a profound, unremovable feature of the quantum ground state. It means the two symmetries are linked in such a way that if you were to treat one as a gauge symmetry, the other would be dragged along with it.

We can even put a number on the strength of this anomaly. It corresponds to the number of fundamental mobile particle types (in this case, lineon dipoles) that are charged under the global symmetry. For the X-cube model, this number is $\nu=3$, one for each of the three planar orientations ($xy, yz, zx$) in which lineons can move. This integer is a robust, topological property of the phase. It tells us we have stumbled upon something truly special: a ​​symmetry-enriched topological phase​​, where the exotic topological order of fractons is fundamentally intertwined with its symmetries in a way that cannot be undone. This is where the journey takes us: from a simple, curious rule about conservation on a line, to a deep and unified new picture of the phases of matter.

In our last discussion, we journeyed into the strange world of subsystem symmetries. We saw that nature, or at least our theoretical imagination of it, doesn't always play by the democratic rule of treating every point in space the same. Sometimes, symmetries pick favorites, applying their rigid laws only to lower-dimensional slices of reality—planes, lines, or even intricate fractals. This might seem like a peculiar, abstract game. A rule that says "you can do this, but only if you stay on the yellow brick road... or perhaps within this single sheet of paper."

But what's the point? Why should we care about such constrained symmetries? The answer, as is so often the case in physics, is that when you impose a new kind of rule, you get a new kind of universe. The consequences of subsystem symmetries are not just minor curiosities; they are profound. They force us to rethink our most basic notions of particles, of phases of matter, and of information itself. They open a door to a veritable zoo of exotic phenomena, a landscape teeming with new theoretical creatures and tantalizing technological possibilities. Let us now take a walk through this zoo.

Perhaps the most startling consequence of subsystem symmetries is the birth of a completely new kind of particle: the ​​fracton​​. Imagine you create an excitation in a material—a little blip of energy, a quasiparticle. Ordinarily, you'd expect to be able to nudge it, to move it from one place to another with some local operation. Not so in a world governed by subsystem symmetries.

Consider a model where we enforce a symmetry on every horizontal plane in a 3D crystal. The rules of this world are such that these planar symmetries must be preserved at all times. If you now create a single, isolated charge excitation, you'll find yourself in a bizarre predicament. Any local operator you might apply to try and move this charge inevitably violates the symmetry of at least one of the planes it has to cross. The system simply won't allow it. The result? The excitation is stuck. It's completely immobile, pinned to its location as if by an invisible force. This is a fracton. Other excitations might find themselves only able to move along a specific line or within a specific plane—these are called lineons and planons, respectively.

This isn't just a lattice trick; it's a deep statement about the mobility of information and energy. When we look at the low-energy effective field theory that describes such systems, we find that the "photons" of this world behave very strangely. Instead of the linear dispersion relation $\omega = ck$ of a familiar light particle, these emergent gauge bosons can have a bizarre quadratic relation, like $\omega \propto k^2$. This means low-energy (long-wavelength) modes move incredibly slowly. It's a universe where "light" crawls rather than sprints, a direct consequence of the underlying constraints on movement imposed by the subsystem symmetries.

What good is a particle you can't move? It turns out this immobility is not a bug, but a feature—a spectacular one for building robust quantum memories. The challenge of quantum computing is decoherence; quantum information is fragile, easily destroyed by local noise. But what if your quantum information is stored in a way that is inherently resilient to local errors?

This is the promise of fracton codes, a new class of topological stabilizer codes built upon the foundation of subsystem symmetries. Models like the X-cube model or the even more arcane Haah's cubic code use these principles to create a protected ground state subspace for encoding logical qubits. The restricted mobility of excitations is key here: a local error creates fractons, but because these fractons are stuck, the error is easily pinpointed and cannot easily spread to corrupt the stored information.

Furthermore, the subsystem symmetries give us an unprecedented level of control. In a system like the X-cube model on a torus, the number of logical qubits it can store—its ground state degeneracy—is directly tied to the system's geometry and symmetries. By "gauging" a subsystem symmetry, a procedure that essentially promotes the symmetry to a fundamental constraint of the system, we can selectively remove logical qubits from the code. Gauging the symmetry associated with a stack of planes, for instance, can precisely reduce the number of logical qubits by one. This is like having a quantum hard drive with physically rewritable sectors, all managed by flipping symmetry switches. The same principle applies to more complex systems like Haah's code, where gauging its beautiful fractal-shaped symmetries also tunes its information capacity.

So far, we've considered systems where the symmetries apply everywhere. But what happens if we're not so democratic? What if we "paint" a subsystem symmetry onto just one part of a larger system? The results are fantastic.

Imagine a standard 3D topological material, like a $\mathbb{Z}_N$ toric code, where point-like charges and loop-like fluxes exist. These particles, called anyons, have fascinating braiding statistics—when you loop one around another, the wavefunction picks up a specific topological phase. This phase is the basis for topological quantum computation. Now, suppose we introduce a planar defect into this system by gauging a $\mathbb{Z}_k$ subsystem symmetry just on that one plane (where $k$ is a divisor of $N$). Suddenly, the rules of the game change on this 2D surface. The flux becomes quantized in different units, and as a result, the braiding phase between a charge and a flux is altered—from $2\pi/N$ in the bulk to $2\pi/k$ on the plane. We've essentially created a custom circuit board for anyons, where their interactions can be programmed by the choice of the painted symmetry.

This idea of locally gauging symmetries leads to an even more profound concept: emergence. If you take a model like the X-cube model and gauge the subsystem symmetry on a 2D region, that region effectively transforms into a 2D toric code—a completely different topological state. Now, this new 2D world has its own set of symmetries, so-called one-form symmetries, which are tied to its topology. If we choose our gauged region to have a complex structure, say with many holes like a Sierpinski carpet, we can generate a vast number of these new, emergent symmetries. It's a stunning hierarchy: we start with one set of rules (subsystem symmetries), use them to build a new world (the gauged region), and find that this new world has its own, richer set of rules.

Symmetries don't just constrain dynamics; they can also protect entire phases of matter. Symmetry-Protected Topological (SPT) phases are gapped states that are quantum-mechanically distinct from a trivial insulator, but only so long as a particular global symmetry is preserved. It should come as no surprise that subsystem symmetries can protect their own unique, and arguably more bizarre, SPT phases.

These "fracton SPTs" are characterized by astonishing physical responses. Consider an SPT phase protected by both $U(1)$ charge conservation and the conservation of charge dipoles within planes. While the bulk is a featureless insulator, its boundaries are anything but. If you take a single fracton charge from deep in the bulk and bring it up to the surface, it induces a cloud of bound charge on that 2D surface. This is strange enough, but the true magic is that this induced charge cloud is itself a topological state—a 2D integer quantum Hall droplet with a non-zero Chern number, whose value is quadratically proportional to the fracton's charge. The secret fractonic order of the bulk manifests as a tangible topological phenomenon on the boundary!

The bulk of such a material also has a "memory" of events that ordinary materials lack. If you drag a lineon excitation through the vacuum from one end of the sample to the other, it leaves behind a permanent polarization in its wake. This induced charge isn't uniform; it arranges itself in a distinctive quadrupolar pattern, a direct fingerprint of the conserved dipole moment that protects the phase. These kinds of topological response phenomena, including analogues of the famous Witten effect from high-energy physics where magnetic-like objects acquire electric-like charges, show that fracton SPTs represent a genuinely new, deeply ordered state of quantum matter.

At this point, you might be thinking that this is all a wonderful theoretical playground, but surely nothing so strange could exist in a real laboratory. You would be underestimating the ingenuity of experimental physicists. The concepts of subsystem symmetry are now guiding the search for new materials and inspiring novel experimental designs.

One of the most promising platforms is ultracold atoms trapped in optical lattices. By precisely tuning lasers, physicists can create artificial crystals of light to trap atoms. More importantly, they can engineer the very Hamiltonian that governs the atoms' behavior. It is conceivable to create a system where atoms don't just hop to their nearest neighbors, but can hop coherently between any two sites along an entire row or column. This would be a direct realization of a Bose gas with subsystem symmetries. Theoretical calculations show that such a system, if it entered a superfluid phase, would exhibit a highly anisotropic superfluid stiffness with an unusual scaling with the system size—a measurable signature of its underlying fractonic nature.

The journey that began with a simple, constraining rule has led us to immobile particles, programmable quantum memories, designer anyon circuits, and new topological states of matter with exotic responses. Subsystem symmetries show us that the world of quantum mechanics is even richer and stranger than we imagined. They are not just a curiosity; they are a new organizing principle, a new paintbrush with which we can color the canvas of quantum matter, connecting the esoteric world of topological field theory with the concrete possibilities of quantum technologies and condensed matter experiments. And, as always in science, we have likely only scratched the surface.