Exciton Berryology

This paper establishes a gauge-invariant framework for defining unique electron and hole exciton Berry phases via a projected position operator, enabling the numerical calculation of exciton polarization and the characterization of topologically distinct exciton bands, including shift excitons, through crystalline inversion and $C_2\mathcal{T}$ symmetries.

The Gist

Here is an explanation of the paper "Exciton Berryology" using simple language and creative analogies.

The Big Picture: The Quantum Dance of a Couple

Imagine a semiconductor (like the silicon in your phone) as a crowded dance floor.

• The Ground State: Usually, all the dancers (electrons) are sitting in the front rows (the "valence band"), and the back rows (the "conduction band") are empty.
• The Exciton: If you shine a light, one dancer jumps up to the back row. This leaves a "hole" (an empty seat) in the front row. Because opposite charges attract, the dancer in the back and the empty seat in the front don't just wander off; they hold hands and dance together as a pair. This pair is called an exciton.

The paper is about understanding the topology (the shape and twistiness) of this dance. Specifically, it asks: Where exactly is this couple dancing, and how does their position change as they move across the floor?

The Problem: "Where is the Center?"

In physics, we often try to find the "center of mass" of a particle to know where it is. But an exciton is a two-body system (an electron and a hole).

• The Electron's View: If you ask the electron, "Where are you?" it might say, "I'm over here!"
• The Hole's View: If you ask the hole, "Where are you?" it might say, "I'm over there!"

In many materials, these two answers are different. The paper reveals that for a long time, physicists were using a "compromise" answer (the average of the two) to describe the exciton's position. The authors realized this was like trying to describe a couple's location by averaging their positions—it misses the nuance of who is actually leading the dance.

The Discovery: Two Different Maps

The authors discovered that there isn't just one way to define the "position" or "twist" (called the Berry phase) of an exciton. There are actually two unique, valid maps:

1. The Electron Map: This map tracks the position of the electron, assuming the hole is fixed.
2. The Hole Map: This map tracks the position of the hole, assuming the electron is fixed.

The Analogy: Imagine a couple walking through a city.

• If you track the husband, you might see him walking in a perfect circle.
• If you track the wife, you might see her walking in a slightly different circle.
• If you track their average, you get a third circle.

The paper shows that the "husband's circle" and the "wife's circle" are the fundamental truths. They tell us different physical things. The difference between these two circles is actually a measure of how far apart the couple is holding hands (their polarization).

The "Shift Exciton": The Magic Trick

The most exciting part of the paper is about Shift Excitons.

Imagine a building where every apartment (unit cell) has a kitchen in the exact center.

• Normal Electrons: If you put a single electron in the building, it naturally sits in the center of the kitchen.
• Shift Excitons: Now, imagine a couple (exciton) enters. Even though the building's rules say "kitchens are in the center," the couple decides to dance in the corner of the apartment.

This is a "Shift Exciton." The couple has shifted their position away from the center, not because the building changed, but because of how they interact with each other.

Why does this matter?
Usually, if the building (the material) is "boring" (topologically trivial), you don't expect any special edge states (like a door that only opens from the outside). But because the exciton couple has shifted to the corner, they create a special door at the edge of the building.

• Result: You can create light-emitting states at the very edge of a material that would otherwise be dead in the middle. This could lead to new types of super-efficient solar cells or LEDs.

The Symmetry Rules: When the Couple Must Agree

The paper also explains when the "Husband's Map" and "Wife's Map" must be identical.

• Inversion Symmetry (The Mirror): If the building has a perfect mirror down the middle, the couple must dance in a way that respects the mirror. In this case, the electron and hole end up agreeing on the same "center."
• Time-Reversal Symmetry (The Rewind): If the laws of physics allow you to play the dance backward perfectly, the couple is forced to dance in a specific, quantized way (either exactly in the center or exactly on the edge).

The authors showed that even if you break the mirror symmetry (making the building lopsided), as long as you keep the "rewind" symmetry, the couple is still forced to dance in these special, quantized spots. This means Shift Excitons can exist in materials that don't look symmetric at all!

The Takeaway

1. Excitons are complex: They aren't just single particles; they are couples with two distinct "personalities" (electron and hole).
2. Two views are better than one: To understand them, we need to look at the electron's position and the hole's position separately, not just the average.
3. Interaction creates topology: Even in "boring" materials, the way electrons and holes interact can create "Shift Excitons" that act like topological materials, creating special states at the edges.
4. New Tech Potential: This understanding helps us design better materials for solar panels and lights by manipulating how these electron-hole couples dance.

In short, the paper gives us a new pair of glasses to see how light and matter interact, revealing that even in simple materials, there is a hidden, complex dance that can be harnessed for future technology.

Technical Summary

Here is a detailed technical summary of the paper "Exciton Berryology" by Davenport, Knolle, and Schindler.

1. Problem Statement

The study of topological phases in condensed matter has largely focused on non-interacting electronic ground states. While the topological classification of interacting ground states is an active field, there is a growing need to understand the topology of interaction-induced excitations, specifically excitons (bound electron-hole pairs) in semiconductors and insulators.

Previous work on exciton topology has relied on specific definitions of the exciton Berry connection (often the average of electron and hole contributions, $\alpha=1/2$) without rigorously addressing the gauge freedom inherent in defining a many-body exciton wave function. The central problem addressed in this paper is:

• How many distinct, valid definitions of the exciton Berry connection exist?
• What is the physical meaning of these different connections?
• How do these definitions relate to observable quantities like the exciton Wannier center and the electron-hole separation?
• How do crystalline symmetries (Inversion $I$ and $C_2T$) constrain these quantities, and can "shift excitons" (excitons with displaced Wannier centers) be identified even without symmetry indicators?

2. Methodology

The authors employ a rigorous many-body formalism to derive exciton topological invariants without relying on smooth gauges, which are often difficult to maintain in numerical calculations.

• Exciton Wave Function Parametrization: They define a general exciton state $|p_{exc}\rangle$ at total momentum $p$ using a parameter $\alpha$ that determines how the total momentum is partitioned between the electron and hole momenta. This reveals an infinite family of valid exciton wave functions and corresponding Berry connections.
• Projected Position Operators: Inspired by the modern theory of polarization, the authors introduce genuine many-body projected position operators. Unlike standard approaches that take the thermodynamic limit immediately, they define periodic operators $\hat{Z}_c$ (projected onto the conduction band/electron) and $\hat{Z}_v$ (projected onto the valence band/hole).
  • $\hat{Z}_{c/v} = \hat{P}_{exc} \left( \sum_{R,i} e^{i\Delta R} \hat{n}_{R,i}^{c/v} \right) \hat{P}_{exc}$
  • Here, $\hat{P}_{exc}$ projects onto the exciton band, and $\hat{n}$ are band-projected number operators.
• Wilson Loops: By calculating the eigenvalues of these projected operators, the authors derive two distinct Exciton Wilson Loops ($W_{exc,c}$ and $W_{exc,v}$). These loops correspond to the Berry phases obtained when the electron or the hole, respectively, is maximally localized within the exciton Wannier state.
• Symmetry Analysis: The authors analyze how these Wilson loops transform under Inversion symmetry ($I$) and Spinless $C_2T$ symmetry (rotation followed by time-reversal). They utilize symmetry eigenvalues for $I$ and direct constraints on the electron-hole separation for $C_2T$.
• Numerical Models: They validate their theory using tight-binding models, including a $C_2T$-symmetric variant of the Su-Schrieffer-Heeger (SSH) model and a model with broken symmetries to demonstrate non-quantized shifts.

3. Key Contributions

A. Resolution of Berry Connection Ambiguity

The paper proves that there is not a unique exciton Berry connection but an infinite family of them, parameterized by $\alpha$.

• The connection $A_{exc}^\alpha(p)$ is a weighted sum of two fundamental connections: $A_{exc,c}(p)$ (electron-localized) and $A_{exc,v}(p)$ (hole-localized).
• $A_{exc}^\alpha(p) = \alpha A_{exc,v}(p) + (1-\alpha) A_{exc,c}(p)$.
• The authors show that previous works implicitly used $\alpha=1/2$ (center of mass), but this is just one specific choice among many.

B. Physical Interpretation: Two Distinct Wannier Centers

The most significant physical insight is that the two fundamental connections correspond to two physically distinct Wannier centers:

1. $\gamma_{exc,c}$: The shift of the Wannier center when the electron is maximally localized.
2. $\gamma_{exc,v}$: The shift of the Wannier center when the hole is maximally localized.

• In general, these two centers are not equal. Their difference is directly proportional to the electric polarization (mean electron-hole separation) of the exciton at momentum $p$.
• The difference $F(p) = A_{exc,c}(p) - A_{exc,v}(p)$ is a gauge-invariant quantity representing the exciton's internal dipole moment.

C. Gauge-Invariant Formulation

The authors derive discrete Wilson loop expressions (Eqs. 15 & 16) that allow for the numerical calculation of these Berry phases without requiring a smooth gauge for the electronic Bloch functions. This is crucial for practical application in numerical simulations where smooth gauges are hard to enforce.

D. Generalization of Shift Excitons

The concept of "shift excitons" (where the exciton Wannier center is shifted relative to the underlying electronic Wannier center) is generalized.

• Shift excitons can exist even if the underlying electronic bands are topologically trivial.
• The shift is determined by the interaction-induced topology, not just the single-particle band topology.

4. Key Results

In the Presence of Symmetries

• Inversion Symmetry ($I$):
  • $I$ symmetry forces the electron-hole separation to be antisymmetric: $F(p) = -F(-p)$.
  • Consequently, the integral of $F(p)$ over the Brillouin zone vanishes, implying $\gamma_{exc,c} = \gamma_{exc,v}$.
  • The unique exciton Berry phase is quantized to $0$or$\pi$and can be diagnosed by the product of the many-body exciton inversion eigenvalues at high-symmetry points ($p=0, \pi$).
• $C_2T$ Symmetry:
  • Although $C_2T$ is anti-unitary and lacks symmetry eigenvalues, it also forces $F(p) = 0$ for all $p$.
  • This implies $\gamma_{exc,c} = \gamma_{exc,v}$ and the Wannier centers are identical.
  • The Berry phase is quantized to $0$or$\pi$.
  • Crucial Finding: Shift excitons can exist in $C_2T$ symmetric systems even when the underlying electronic bands are trivial. These are topologically protected but cannot be diagnosed by symmetry indicators (since $C_2T$ has no eigenvalues), requiring the Wilson loop calculation instead.

In the Absence of Symmetries

• When both $I$ and $C_2T$ are broken, $\gamma_{exc,c} \neq \gamma_{exc,v}$.
• The two sets of Wannier states (electron-localized vs. hole-localized) have different centers of mass.
• The difference in these centers is determined by the momentum-dependent electron-hole separation $F(p)$.
• The Berry phases are no longer quantized but still provide physical information about the exciton's internal structure.

Chern Numbers in Higher Dimensions

• The authors show that while the Berry curvature distribution depends on $\alpha$, the Chern number is independent of $\alpha$.
• $C_{exc, \alpha} = \alpha C_{exc, v} + (1-\alpha) C_{exc, c}$. Since the Chern number must be an integer for all $\alpha$, it follows that $C_{exc, c} = C_{exc, v}$. Thus, the exciton Chern number is well-defined regardless of the choice of localization.

5. Significance and Impact

1. Foundational Clarity: The paper resolves the ambiguity in defining exciton topology, establishing that there are two fundamental, physically distinct topological invariants (electron and hole localized) rather than a single arbitrary one.
2. New Diagnostic Tools: The derivation of gauge-invariant Wilson loops provides a robust numerical method to calculate exciton topology in finite systems and without smooth gauges, making it applicable to real materials simulations.
3. Beyond Symmetry Indicators: It demonstrates that interaction-induced topology (shift excitons) can exist in systems with $C_2T$ symmetry where traditional symmetry indicators fail. This expands the known landscape of topological phases in interacting systems.
4. Physical Observables: The link between the difference in Berry phases and the exciton polarization (electron-hole separation) connects abstract topological invariants to measurable physical properties, such as shift currents in photovoltaics and the behavior of excitons in external fields.
5. Generalizability: The formalism is not limited to excitons; the authors suggest it applies to any two-body bound state (e.g., magnons, Cooper pairs) and can be extended to $N$-body systems, where $N$ distinct Wilson loops can be defined.

In summary, "Exciton Berryology" provides a comprehensive framework for understanding the topological nature of excitons, revealing that their topology is richer and more nuanced than previously thought, governed by the interplay between electron and hole localization and constrained by crystalline symmetries.