Non-Symmorphic Symmetry

In the study of crystalline matter, symmetry is a foundational principle that dictates physical properties. While familiar symmetries like reflections and rotations are intuitive, a more subtle and powerful class exists: non-symmorphic symmetries. These operations, which indivisibly combine a rotation or reflection with a fractional translation through the crystal lattice, lead to consequences that are far from obvious. The central question this article addresses is how these 'symmetries with a twist' manifest in the quantum world, giving rise to exotic and potentially revolutionary material properties that simpler symmetries cannot explain. This article delves into the core of non-symmorphic symmetry across two chapters. In "Principles and Mechanisms," we will explore the fundamental nature of these symmetries, from the real-space concepts of glide planes and screw axes to their direct implications in momentum space, such as systematic absences and forced band degeneracies. Subsequently, "Applications and Interdisciplinary Connections" will reveal how these principles are observed experimentally and leveraged as a design tool to engineer novel topological materials, including Dirac semimetals and hourglass fermions, connecting the physics of crystals to the abstract beauty of group theory.

Symmetry is a concept we learn from a young age. We see it in the wings of a butterfly, the petals of a flower, the reflection in a mirror. In physics, and especially in the world of crystals, symmetry is not just a matter of aesthetics; it is a profound organizing principle that dictates the very laws governing the particles within. A crystal, at its heart, is a repeating pattern of atoms, and this repetition is its most fundamental symmetry—translation. If you shift the entire crystal by just the right amount, it looks exactly the same.

But nature, in its boundless ingenuity, doesn't stop with simple repetition. It employs symmetries that are more subtle, more intricate, and far more surprising in their consequences. Among the most fascinating of these are the ​​non-symmorphic symmetries​​. These are symmetries with a twist, a kind of built-in motion that interweaves rotation or reflection with a step through space.

Imagine walking up a spiral staircase. To get from one step to the one directly above it, you perform two actions at once: you rotate around the central axis and you translate upwards. You cannot separate these actions; the staircase only looks the same after you've done both. This is the essence of a ​​screw axis​​, a non-symmorphic symmetry that combines a rotation with a fractional translation along the axis of rotation.

Another type is a ​​glide plane​​. Picture yourself walking on a sandy beach, leaving a trail of footprints. If you imagine a line running between your left and right feet, each left footprint is a reflection of the right one before it, but also shifted forward. The pattern isn't symmetric under reflection alone, nor under translation alone, but only under the combined operation of reflection and translation. This is a glide plane: a reflection across a plane followed by a translation parallel to that plane.

The key word here is ​​fractional​​. The translation involved is not by a full lattice vector that would map the crystal onto itself anyway. It's a translation by a fraction of a lattice vector. The crystal as a whole is perfectly invariant under this composite operation, but it is not invariant under the rotation or reflection part alone, nor under the fractional translation part alone. This "indivisible" nature of non-symmorphic operations is what leads to their extraordinary consequences.

How can we be so sure that these peculiar symmetries exist? We can't peer into a crystal and see atoms performing this little dance of "reflect-and-shift". Instead, we see their fingerprints, left in the patterns of scattered waves. When we fire a beam of X-rays or neutrons at a crystal, the waves scatter off the atoms and interfere with each other, creating a diffraction pattern of bright spots. This pattern is essentially a map of the crystal's periodicities, a kind of structural Fourier transform.

Ordinarily, you might expect a spot for every possible repeating plane in the crystal. But crystals with non-symmorphic symmetries show a strange phenomenon: certain spots that should be there are systematically missing. These are called ​​systematic absences​​ or extinctions.

Imagine two atoms, A and B, that are related by a glide symmetry. A wave scattering off atom A and a wave scattering off atom B will travel to our detector. Because atom B is related to A by a reflection and a fractional shift, the wave from B will have a specific phase difference relative to the wave from A. For certain scattering directions, this phase shift from the fractional translation can be exactly half a wavelength, corresponding to a phase factor of $\exp(i\pi)=-1$. The two waves arrive perfectly out of phase and cancel each other out completely. The predicted bright spot vanishes.

One of the most famous examples is the diamond structure. Diamond is built on a face-centered cubic (FCC) lattice, which itself has certain systematic absences. But diamond has an additional non-symmorphic symmetry. As a result, certain reflections like the one indexed as $(2,0,0)$, which are allowed for a simple FCC structure, are mysteriously absent in diamond's diffraction pattern. This missing spot is a direct, unambiguous "fingerprint in the void", telling us that the underlying atomic arrangement possesses a non-symmorphic glide symmetry.

The most profound consequences of non-symmorphic symmetry are not on the static crystal structure, but on the behavior of electrons moving within it. In the quantum world of a crystal, an electron's energy is not arbitrary; it is confined to specific energy ranges called ​​bands​​. Normally, different bands corresponding to different quantum states can have different energies. But non-symmorphic symmetries can issue a command that cannot be disobeyed: at certain places in the Brillouin zone (the space of an electron's crystal momentum), different bands are forced to have the exact same energy. This is called a ​​forced degeneracy​​ or "band sticking".

How does this happen? The logic is as beautiful as it is inescapable. Let's call the operator for a non-symmorphic symmetry $G$. It could be a glide or a screw. If we apply this operation twice, we undo the rotation or reflection, but we add the fractional translation to itself. The net result of applying $G$ twice is simply a pure translation by a full lattice vector, say $\mathbf{R}$. So, we have the crucial algebraic relation: $G^2 = T_{\mathbf{R}}$.

Now consider an electron at a special location: the boundary of the Brillouin zone. A Bloch state $| \psi_{\mathbf{k}} \rangle$ at this boundary is defined by the condition that $\mathbf{k} \cdot \mathbf{R} = \pi$ (or an odd multiple of $\pi$). When the translation operator $T_{\mathbf{R}}$ acts on this state, it imparts a phase factor $\exp(-i\mathbf{k} \cdot \mathbf{R}) = \exp(-i\pi) = -1$. So, for an electron on the zone boundary, the action of $G^2$ is simply to multiply its state by $-1$. For these specific electrons, the symmetry operator has the property $G^2 = -1$!.

This simple equation, $G^2 = -1$, is the key. If an energy eigenstate $| \psi \rangle$ were an eigenstate of $G$ with eigenvalue $\lambda$, then $G^2|\psi\rangle = \lambda^2 |\psi\rangle$. But we know $G^2|\psi\rangle = -|\psi\rangle$, which means $\lambda^2=-1$. The eigenvalues of the symmetry operator must be $\pm i$. Now, if the system also has time-reversal symmetry $\mathcal{T}$, then the state $\mathcal{T}|\psi\rangle$ must have the same energy. If $G$ and $\mathcal{T}$ commute, $\mathcal{T}|\psi\rangle$ will be an eigenstate of $G$ with the complex-conjugated eigenvalue, $-i$. Thus, we have found two distinct states, $|\psi\rangle$ and $\mathcal{T}|\psi\rangle$, that are guaranteed to have the same energy. The bands must touch. We have a guaranteed two-fold degeneracy.

We can also see this from a more "hands-on" perspective by building a simple model of the crystal Hamiltonian. A one-dimensional crystal with a glide symmetry has two atoms per unit cell. The Hamiltonian can be written as a $2 \times 2$ matrix. The glide symmetry imposes strict constraints on the elements of this matrix. It turns out that at the Brillouin zone boundary ($k=\pi/a$), the symmetry forces the off-diagonal elements of the Hamiltonian to become exactly zero. The Hamiltonian becomes diagonal, and the two diagonal elements (representing the energies of the two atoms) are forced to be identical. The two energy bands become degenerate. This isn't an accident; it's a mathematical necessity dictated by the symmetry.

What if a crystal is endowed with more than one non-symmorphic symmetry? The results can be even more spectacular. At certain highly symmetric points in the Brillouin zone, we might find two distinct non-symmorphic operators, say $G_x$ and $G_y$, that both square to $-1$ at that point. But what's more, their fundamental algebra might require them to ​​anticommute​​: $G_x G_y = -G_y G_x$.

This trio of relations—$G_x^2 = -1$, $G_y^2 = -1$, and $G_x G_y = -G_y G_x$—is the defining algebra of the quaternions, famously discovered by William Rowan Hamilton. This algebra cannot be represented by simple numbers. The smallest matrices that can obey these rules are $2 \times 2$ matrices (like the Pauli matrices of quantum mechanics). This means that any set of electronic states at this point must form a basis for at least a two-dimensional representation. In other words, the electronic bands are forced into a two-fold degeneracy by the spatial symmetries alone!

Now, let's add the final ingredient: the electron's intrinsic spin. For a spin-1/2 particle, time-reversal symmetry is more powerful (Kramers' theorem) and by itself guarantees that every energy level is at least doubly degenerate. When this is combined with the two-fold degeneracy from the anticommuting non-symmorphic symmetries, the total minimum degeneracy becomes four. At these special points in momentum space, every single energy level is a meeting point for at least four distinct quantum states.

These forced degeneracies are not just mathematical curiosities. They are the seeds from which new and exotic forms of electronic matter can grow, and they have tangible consequences in the real space of the crystal.

Let's return to our pair of bands that are forced to touch. We can construct localized wave packets, or ​​Wannier functions​​, that describe the electrons in these bands. It turns out that the non-symmorphic symmetry that glues the bands together in momentum space also acts as a bridge between their corresponding Wannier functions in real space. The glide operator literally maps the Wannier function of one band onto the other. A startling consequence of this is that the "centers of charge" of these two electronic orbitals are forced to be separated by exactly the fractional translation vector of the glide symmetry, for instance, half a lattice constant, $a/2$. The abstract symmetry of the crystal lattice is imprinted directly onto the spatial configuration of the electron clouds themselves.

Perhaps the most exciting frontier is the role of non-symmorphic symmetries in birthing ​​topological matter​​. In some materials with both spin-orbit coupling and a glide symmetry, a remarkable "partner switching" occurs. At one point in the Brillouin zone (say, the center $\Gamma$), time-reversal symmetry pairs up states with different glide-symmetry eigenvalues. But at another point (like the boundary $X$), it pairs up states with the same eigenvalue. Since the bands must be continuous, and the eigenvalues are "sticky," the bands are forced to swap partners as they traverse the Brillouin zone from $\Gamma$ to $X$. The only way for this to happen is for them to cross. This isn't just a simple crossing; the band structure forms a shape reminiscent of an hourglass, with a guaranteed crossing point.

This protected crossing is not confined to a single line in momentum space. As we explore the entire plane defined by the glide symmetry, this crossing point traces out a continuous curve, forming a ​​nodal line​​ of degeneracies. This nodal line is not an accident that can be washed away by small imperfections. It is a robust topological feature protected by the non-symmorphic symmetry. Materials hosting these features, known as nodal-line semimetals, represent a new state of quantum matter with strange and potentially powerful electronic properties. Thus, a simple "symmetry with a step" gives rise to some of the most complex and fascinating phenomena in modern condensed matter physics, revealing once again the deep and beautiful unity between the geometry of space and the laws of nature.

In the previous chapter, we dissected the strange and beautiful nature of non-symmorphic symmetries—those peculiar crystallographic rules that combine rotations or reflections with "forbidden" fractional translations. We saw that they are not mere curiosities of crystal classification but fundamental rules that enforce surprising connections between wave-like states in a solid. Now, we ask the physicist's favorite question: So what? Where do we see these rules in action? What can we do with them?

This chapter is a journey through the consequences of taking a step that is not quite a full step. We will see how this simple geometric idea allows us to hunt for these symmetries in the laboratory, provides a toolkit for designing exotic new materials from the ground up, places profound constraints on the collective behavior of quantum matter, and ultimately reveals a stunning connection between the world of crystals and the abstract realm of pure mathematics.

How can we be sure that a crystal hidden deep inside a cryostat truly possesses a symmetry that involves a fractional shift of its lattice? We cannot see the atoms move, but we can see the consequences of their dance on the electrons that live among them. One of the most powerful tools for this is Angle-Resolved Photoemission Spectroscopy, or ARPES. In an ARPES experiment, we shine high-energy photons on a material, which knock out electrons. By measuring the energy and angle of these ejected electrons, we can reconstruct the electronic band structure—the "rules of the road" for electrons inside the crystal.

Now, imagine a crystal with a non-symmorphic glide plane. As we learned, this symmetry forces certain electronic bands to become degenerate—to touch—at the boundary of the Brillouin zone. But it does more than that. At such a touching point, the two degenerate electron states have opposite parity with respect to the mirror part of the glide operation. One state is "even," the other is "odd."

An ARPES experiment can often be set up to be sensitive to only one of these parities. What happens when we look at this degenerate pair of states? We see something remarkable: one of the bands is brightly visible, while its degenerate partner is completely absent. It’s as if one of a pair of identical twins is a ghost. This phenomenon, known as a ​​systematic extinction​​, is a direct and unambiguous fingerprint of the underlying non-symmorphic symmetry. We are not just inferring the symmetry; we are seeing its "selection rule" in action, a clear message from the crystal's quantum mechanical wavefunctions.

This principle is not limited to electrons. Any wave-like excitation that travels through the crystal must obey its symmetry rules. In a familiar material like silicon, which crystallizes in the diamond structure, the lattice vibrations themselves—the phonons—are governed by a non-symmorphic space group. At certain high-symmetry points in the Brillouin zone, phonon branches are forced to "stick together" in pairs, exhibiting degeneracies that would not be predicted by simpler, symmorphic symmetries alone. By scattering neutrons instead of photons off the crystal, physicists can map these phonon bands and again see the tell-tale sign of degeneracy, confirming the non-symmorphic character of one of the most important materials of our time.

Perhaps the most exciting application of non-symmorphic symmetries lies in the burgeoning field of topological materials. Here, these symmetries transform from a passive property to be discovered into an active design principle—a powerful tool for an electronics designer's toolkit to engineer materials with properties unimaginable a few decades ago. Non-symmorphic symmetries allow us to create and protect exotic electronic states of matter.

Pinning Down Reality: Dirac and Weyl Semimetals

In some materials, electrons behave not like the slow, hefty particles we learn about in introductory physics, but like massless, relativistic particles zipping along at a constant speed, obeying a version of Paul Dirac's famous equation. The places in the band structure where this happens are called Dirac points—fourfold degenerate crossings of the conduction and valence bands.

There are a few ways to create and protect such points. One way involves a combination of time-reversal and inversion symmetry. But this type of protection is somewhat flexible; the Dirac points can wander around along high-symmetry lines as the material's parameters are tweaked. Non-symmorphic symmetries offer a different, more rigid kind of protection. They have the remarkable ability to force Dirac points to appear, and to ​​pin them​​ to specific, high-symmetry locations on the very boundary of the Brillouin zone. It’s the difference between a rule that says "two roads must cross somewhere on this highway" and a rule that says "they must cross exactly at the state line."

The underlying reason is a beautiful piece of quantum algebra. A screw axis, for instance, is a rotation by an angle $2\pi/n$ followed by a fractional translation. When you apply this operation twice for a $180^\circ$ screw ($n=2$), you get a full lattice translation, but with a crucial minus sign from the electron's spin. This leads to an algebraic rule for the symmetry operator $\hat{S}_z$ that looks something like $\hat{S}_z^2 = -1$ precisely at the Brillouin zone boundary. This is the same algebra that the time-reversal operator obeys, and it guarantees that every energy level must be at least doubly degenerate. When combined with other symmetries like inversion, this can enforce a robust, four-fold degeneracy—a Dirac point—that is unmovable from its high-symmetry perch. The crystal structure itself conspires to create an emergent, protected relativistic universe for its electrons.

From Points to Lines: Nodal-Line Semimetals

Nature does not have to stop at creating degenerate points. What if the bands touched not just at a single point, but along an entire line, forming a "nodal line" or ring? Electrons in such a material would have a whole continuum of states where they behave as if they are massless. Non-symmorphic symmetries are a prime mechanism for creating such structures.

One common recipe involves two ingredients: band inversion and differing symmetry labels. Imagine two bands that, due to the chemistry of the crystal, want to swap places as we move from the center of the Brillouin zone to its edge. Now, if these bands are "tagged" with different eigenvalues of a glide symmetry, they cannot simply mix and push each other away (a process called hybridization). Like two trains on different tracks, they are destined to cross. The glide symmetry protects this crossing. Because this protection exists over a continuous region of momentum space, the single-point crossing extends into a line of crossings—a nodal line. In some simple models, the non-symmorphic translation forces the part of the Hamiltonian that couples the bands to have a peculiar momentum dependence, like $\cos(k_y/2)$. Such a term is mathematically guaranteed to be zero when $k_y = \pi$—that is, right at the Brillouin zone boundary—carving out the nodal line exactly where the symmetry dictates.

The Crystal Hourglass

The most visually stunning consequence of non-symmorphic symmetry is arguably the "hourglass fermion." This phenomenon arises from a delicate interplay between the non-symmorphic symmetry and the time-reversal symmetry that all spin-$1/2$ electrons obey.

Let's use an analogy. Think of the electronic states at a high-symmetry point as dancers in a pair, linked by time-reversal symmetry (a Kramers pair). A glide symmetry assigns each dancer a name tag (an eigenvalue). Now here's the trick: because of the fractional translation, the nature of these name tags changes as we move across the Brillouin zone.

At the center of the zone ($\Gamma$ point), the rules of the glide symmetry require that the two dancers in a Kramers pair have opposite name tags, say $+1$ and $-1$. But at the edge of the zone ($X$ point), the same rules require that they have identical name tags, say $+i$ and $+i$. Now, suppose a band starting at $\Gamma$ with a $+1$ name tag has to connect to a band at $X$. It must connect to a band with a $+i$ name tag. Its Kramers partner at $\Gamma$, with a $-1$ tag, must connect to a band at $X$ with a $-i$ tag. The partners are forced to separate and connect to different states! For the bands to maintain their continuity, they must cross paths somewhere in the middle. This enforced partner-switching and subsequent band crossing creates a unique dispersion shape that looks exactly like an hourglass. This is not an accident; it is an unavoidable consequence of the non-symmorphic symmetry. A crystal with only a simple mirror symmetry, lacking the fractional translation, has no such rule change for its name tags, and therefore no enforced hourglass crossing. The hourglass fermion is perhaps the most elegant demonstration of the profound connectivity that a "step-and-turn" symmetry imposes on the quantum world.

So far, we have focused on the behavior of individual electrons (or phonons) as described by band structures. But non-symmorphic symmetries also place powerful constraints on the collective, many-body behavior of quantum systems.

Consider a magnet made of atoms with half-integer spin. According to a powerful theorem known as the Lieb-Schultz-Mattis (LSM) theorem, such a system, under certain general conditions, cannot settle into a simple, boring, gapped ground state. It is constitutionally forbidden from being trivial. When a non-symmorphic glide symmetry is added to the mix, this theorem becomes even more prescriptive. Not only must the system be exotic and gapless, but the gapless excitations are pinned to a specific corner of the Brillouin zone, such as the point $(\pi/a, \pi/a)$. The crystal geometry reaches into the complex world of quantum many-body interactions and dictates where the interesting action must happen. This provides a powerful guiding principle in the search for exotic phases of matter like quantum spin liquids, which could be the foundation for future quantum computers.

At its deepest level, the distinction between symmorphic and non-symmorphic space groups is not just a physical one, but a profound mathematical one, connecting crystallography to the abstract field of group theory.

A space group $G$ can be thought of as being built from two smaller groups: the translation group $T$ (the infinite lattice of steps) and the point group $P$ (the set of rotations and reflections about a point). A symmorphic group is the simplest way to combine them, known as a semidirect product. It's like having a set of rotational symmetries that all share a common origin.

A non-symmorphic group, however, represents a more "twisted" or "entangled" way of combining translations and point group operations. There is no single point that is left unchanged by all the rotations and reflections. The mathematical tool used to classify all the possible ways of twisting two groups together is called ​​group cohomology​​. The set of distinct space groups for a given lattice and point group is classified by the "second cohomology group," denoted $H^2(P, T)$.

In this language, the symmorphic space group corresponds to the boring, trivial element—the "zero"—of this cohomology group. All the non-symmorphic groups correspond to the non-zero, non-trivial elements. For example, for the point group $D_6$ (the symmetries of a hexagon), the relevant cohomology group is the cyclic group $\mathbb{Z}_6$, which has six elements: $\{0, 1, 2, 3, 4, 5\}$. This means there are six fundamentally different ways to build a space group with this symmetry. The '0' element corresponds to the symmorphic group $P622$. The five non-zero elements correspond to the non-symmorphic groups containing screw axes, such as $P6_122$, $P6_222$, and so on. The space group $P6_k22$ corresponds directly to the element $k$ in $\mathbb{Z}_6$. This provides an incredibly deep and elegant classification, revealing the hidden mathematical skeleton that underpins the diversity of crystal structures.

What began as a simple observation about crystal patterns concludes with a connection to some of the beautiful structures of modern mathematics. From a missing spot in an ARPES measurement to the classification of group extensions, the journey of non-symmorphic symmetry reveals, time and again, the remarkable and often unexpected unity of scientific thought.