🔬 materials science

Excitons and Optical Response in Excitonic Insulator Candidate TiSe $_2$

Through fully ab-initio Bethe-Salpeter equation calculations, this study demonstrates that while excitonic fluctuations emerge near the transition temperature, the charge density wave phase in TiSe $_2$ is primarily driven by structural distortions rather than a purely electronic excitonic insulator mechanism.

Original authors: Dino Novko

Published 2026-02-25

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Dino Novko

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a crowded dance floor inside a crystal called Titanium Diselenide (TiSe₂). This isn't just any dance floor; it's a place where electrons (the dancers) and the atoms they dance on (the floorboards) are constantly interacting.

For decades, scientists have been arguing about what causes a specific "dance move" in this crystal, known as the Charge Density Wave (CDW). This is a moment where the dancers suddenly line up in a perfect, repeating pattern, changing the rhythm of the whole room.

There are two main theories about who leads this dance:

The "Floorboard" Theory: The atoms themselves wiggle and shift, forcing the electrons to line up. (This is the electron-phonon mechanism).
The "Dancer" Theory: The electrons get so attracted to each other that they pair up and spontaneously organize, dragging the floorboards along with them. (This is the Excitonic Insulator or EI theory).

This paper is like a high-speed, super-accurate camera recording the dance floor to see who is actually leading.

The High-Temperature Dance (The Normal Phase)

First, the researchers looked at the crystal when it's hot (above 200 Kelvin). In this state, the dance floor is a bit chaotic.

What they found: They saw a very strong, energetic "duet" between an electron and a hole (a missing electron) happening at a high energy level (1.6 eV). Think of this as a couple doing a fast, high-energy spin near the center of the room.
The Missing Clue: The "Dancer" theory predicted that if the electrons were about to take over and form a new order, there should be a "soft," lazy, low-energy wobble happening near the edges of the room (the Brillouin zone).
The Verdict: The camera didn't see this lazy wobble. The "Dancer" theory (Excitonic Insulator) didn't seem to be the main force driving the dance. The electrons weren't spontaneously organizing on their own.

The Low-Temperature Dance (The CDW Phase)

When the crystal cools down below 200 K, the dance floor changes. The atoms physically shift their positions (a Periodic Lattice Distortion), creating a new, tighter pattern.

What happened: Because the floorboards moved, the energy gap between the dancers opened up. Suddenly, two new, slow, low-energy dances appeared (at 0.4 eV and 80 meV).
The Twist: As the temperature gets closer to the transition point (just before the dance floor snaps back to normal), these two slow dances start to "soften." They get slower and slower, almost stopping completely, right before the pattern breaks.

The Big Conclusion

The researchers realized that the "Dancer" theory (Excitonic Insulator) isn't the main director of this show. The atoms moving (the floorboards) are the ones who started the party.

However, there is a small, fascinating moment near the end of the show. Just as the dance floor is about to snap back to normal, those two slow dances get so soft they almost disappear. This suggests that while the electrons didn't start the dance, they do get involved right at the edge of the transition, creating a brief moment of "excitonic fluctuations."

The Takeaway Analogy

Think of it like a marching band:

Old Theory: The musicians (electrons) decided to spontaneously march in a perfect line, and the ground (atoms) just followed along.
This Paper's Finding: The ground (atoms) actually started to vibrate and shift first, forcing the musicians to march in a line.
The Nuance: But, right as the ground stops vibrating, the musicians get so synchronized that they almost form a new, ghostly pattern of their own before the music stops.

Why does this matter?
Understanding who leads the dance helps scientists design better materials for solar cells, superconductors, and ultra-fast computers. It tells us that in TiSe₂, the atoms are the bosses, but the electrons are the clever assistants who show up right at the critical moment to add a little extra magic.

1. Problem Statement

The origin of the Charge Density Wave (CDW) phase in Titanium Diselenide (TiSe $_2$ ) remains a subject of intense debate. The central controversy lies in identifying the primary driving mechanism for the phase transition below $T_{CDW} \approx 200$ K:

Electron-Phonon Coupling: Structural instabilities and soft phonon modes driving the transition.
Excitonic Insulator (EI) Mechanism: A purely electronic mechanism where a soft collective electronic mode (specifically, an exciton with finite momentum) drives the phase transition.

While experimental evidence (such as Electron Energy Loss Spectroscopy, EELS) has suggested the existence of a soft electronic mode supporting the EI scenario, other studies attribute these features to phonons or plasmons. Furthermore, theoretical models have yielded conflicting results, with some predicting a well-defined soft exciton and others finding only a weak continuum. The lack of consensus on the existence and nature of these soft electronic modes hinders the understanding of the CDW mechanism in TiSe $_2$ and other potential EI candidates.

2. Methodology

The authors performed fully ab-initio calculations to resolve these ambiguities, focusing on the electron excitation spectra across the CDW phase diagram. The computational framework included:

Electronic Structure: Calculations were based on Density Functional Theory (DFT) using both the semi-local PBE functional and the hybrid HSE functional. To ensure accuracy in band gaps and quasiparticle energies, the GW approximation was also employed.
Excitonic Effects: The electron-hole interaction was explicitly included using the Bethe-Salpeter Equation (BSE). This allowed for the calculation of optical absorption spectra and the identification of bound excitonic states.
Phases Studied:
- Normal Phase ( $T > T_{CDW}$ ): The high-temperature semimetallic phase ( $1 \times 1$ unit cell).
- CDW Phase ( $T < T_{CDW}$ ): The low-temperature phase with a $2 \times 2$ periodic lattice distortion (PLD). The structure was relaxed at various temperatures to simulate the temperature dependence of the lattice distortions.
Momentum Resolution: The study specifically investigated excitations at finite momentum ( $q \approx M$ , the edge of the Brillouin Zone), which is critical for testing the EI hypothesis, in addition to the zero-momentum ( $q=0$ ) optical response.

3. Key Contributions

First-Principles BSE Analysis: This work provides the first comprehensive ab-initio BSE study of TiSe $_2$ that simultaneously treats accurate electronic structures (via GW/HSE) and excitonic effects across both normal and CDW phases.
Resolution of the Soft Mode Controversy: The study directly addresses the conflicting reports regarding the existence of a soft electronic mode in the normal phase by rigorously calculating the finite-momentum excitation spectra.
Identification of CDW-Related Excitons: The authors identify and characterize two distinct low-energy excitonic modes that emerge specifically in the CDW phase due to band hybridization.
Phase Diagram Mapping: The work maps the temperature evolution of these excitonic modes, distinguishing between the region dominated by electron-phonon coupling and a narrow region near $T_{CDW}$ where EI fluctuations may be relevant.

4. Key Results

A. Normal High-Temperature Phase ( $T > T_{CDW}$ )

Dominant Exciton: The optical spectrum is dominated by a strong excitonic peak at 1.6 eV. This arises from vertical electron-hole transitions between Se- $p$ and Ti- $d$ states at the $M$ point of the Brillouin Zone. The calculated binding energy is $\sim 300$ meV, consistent with experimental optical absorption and EELS data.
Absence of Soft Electronic Modes: Crucially, at finite momentum ( $q \approx M$ ), no well-defined soft electronic mode was found. The excitation spectra show only a weak, broad interband electron-hole continuum.
Conclusion on EI: The lack of a temperature-dependent, soft exciton in the normal phase suggests that the purely electronic EI mechanism is not the primary driver of the CDW formation in TiSe $_2$ .

B. CDW Phase ( $T < T_{CDW}$ )

Band Gap Opening: Structural distortions (PLDs) induce a CDW gap ( $\sim 600-700$ meV) between hybridized Se- $p$ and backfolded Ti- $d$ states at the $\Gamma^*$ point.
Emergence of Low-Energy Excitons: Two new low-energy excitonic modes appear in the optical spectrum:
1. $\omega_{exc}^2$ (~0.4 eV): Arises from transitions across the CDW gap between hybridized valence ( $v_1$ ) and conduction ( $c_2$ ) states. This mode has a binding energy of $\sim 300$ meV and matches experimental infrared absorption peaks.
2. $\omega_{exc}^3$ (< 0.2 eV): A second low-energy mode formed by transitions between $v_1$ and $c_1$ states.
Temperature Dependence: As temperature increases toward $T_{CDW}$ $T_{C D W}$ :
- The CDW gap closes.
- The $\omega_{exc}^2$ mode redshifts and is quenched (disappears) as the gap closes.
- The $\omega_{exc}^3$ mode softens and approaches zero energy even before the system fully transitions to the normal state.

C. Near $T_{CDW}$ Fluctuations

While the EI mechanism is not the main driver, the softening of the $\omega_{exc}^3$ mode and the persistence of finite absorption intensity above $T_{CDW}$ (observed in experiments but not fully captured by static calculations) suggest a region near the transition temperature where EI fluctuations and soft CDW phonon modes may hybridize and influence each other.

5. Significance

Mechanism Clarification: The study strongly argues against the EI mechanism being the sole or dominant cause of the CDW in TiSe $_2$ , favoring a scenario where electron-phonon coupling initiates the transition, with excitonic correlations playing a secondary or corrective role.
Guidance for EI Candidates: The findings provide a benchmark for identifying true excitonic insulators. The absence of a soft mode in the normal phase serves as a critical test for other candidate materials.
Light-Matter Applications: The identification of strong, low-energy excitonic modes with large binding energies in the CDW phase suggests potential for enhancing light-matter interactions in TMDs, similar to observations in TaSe $_2$ and kagome metals.
Theoretical Framework: The work demonstrates the necessity of combining GW corrections with BSE to accurately describe the complex interplay of lattice distortions and electron-hole correlations in correlated materials.