Melting point depression of charge density wave in 1T-TiSe$_2$ due to size effects

Using in-situ cryogenic electron microscopy on 1T-TiSe$_2$ nanoflakes, this study demonstrates that charge density wave melting points depress as flake size decreases below 100 nm due to finite-size effects cutting off correlation length divergence, thereby confirming that electronic phase transitions in correlated states follow classical nucleation theory.

The Gist

Imagine a crowded dance floor where everyone is trying to move in perfect, synchronized steps. In the world of materials science, this synchronized movement is called a Charge Density Wave (CDW). It's a special state where electrons in a material (specifically a crystal called 1T-TiSe2) lock into a rhythmic pattern, creating a wave-like structure that changes how the material conducts electricity.

Usually, this dance happens naturally when the material gets cold. But what happens if you shrink the dance floor down to the size of a tiny speck? That is exactly what this paper investigates.

Here is the story of their discovery, broken down into simple concepts:

1. The "Too Small to Dance" Problem

In the big, bulk world (a large chunk of the material), the electrons can easily find their rhythm and form this wave when cooled below about 210–230 Kelvin (roughly -60°C).

However, the researchers took this material and chopped it into tiny, flat flakes, some smaller than a human hair is wide. They found a surprising rule: The smaller the flake, the harder it is for the electrons to dance.

• The Analogy: Imagine a massive stadium full of people doing "The Wave." It's easy for the wave to travel across the whole crowd. But if you only have a tiny group of 10 people in a small room, it's very hard to get them all to coordinate a wave. If the room gets too small, the wave simply can't form at all.

2. The Melting Point Drop

In physics, when a material changes from one state to another (like ice melting into water), we call it a "phase transition." For this material, the "melting" is when the electron dance stops and the material goes back to being chaotic.

• The Finding: In large chunks, the dance stops (melts) at about -60°C. But in their tiny flakes (smaller than 100 nanometers), the dance started falling apart at much warmer temperatures.
• The Result: For the tiniest flakes (around 50 nanometers), the electrons refused to dance entirely, even when the researchers cooled them down to nearly absolute zero (-273°C). The "dance floor" was just too small for the wave to exist.

3. Why Does This Happen? (The "Bouncer" Theory)

The researchers wanted to know why the dance failed in small spaces. They looked at the material under a super-powerful microscope (an electron microscope) and found the culprit: Defects.

• The Metaphor: Think of the electrons as dancers who need a "bouncer" or a "captain" to tell them where to stand and start the wave. In this material, those captains are tiny clusters of extra Titanium atoms (defects) that naturally get stuck inside the crystal during its growth.
• The Discovery: These "captains" are spaced about 10 to 50 nanometers apart.
  • If your flake is huge, it has plenty of captains to organize the dancers.
  • If your flake is tiny (smaller than the distance between captains), it might not have any captains at all. Without a captain to start the rhythm, the electrons can't organize themselves, and the Charge Density Wave never forms.

4. The "Freezing" of the Wave

The paper also explains that as the flake gets smaller, the "wave" tries to grow, but the edges of the flake cut it off. It's like trying to grow a giant tree in a tiny pot; the roots hit the sides before they can spread out.

The researchers used a mathematical model (called the Ginzburg-Landau model) to predict this. Their model perfectly matched what they saw in the lab:

• Big Flakes: The wave forms easily.
• Medium Flakes: The wave forms, but it melts (breaks down) at a warmer temperature than usual.
• Tiny Flakes: The wave cannot form at all because the "pot" is too small to hold the necessary pattern.

Summary

This paper proves that for certain electronic states, size matters immensely. Just as a small room can't hold a large crowd's synchronized dance, a tiny nanoflake can't support the complex electron wave found in bulk materials.

The researchers showed that the "melting point" of this electronic state isn't fixed; it depends on how big your sample is. If you make the sample too small, the electronic state disappears completely because there isn't enough room for the pattern to establish itself, and there aren't enough "captains" (defects) to start the process.

This is a fundamental observation about how nature behaves when you shrink things down to the nanoscale, showing that the rules of the "big world" don't always apply to the "tiny world."

Technical Summary

Technical Summary: Melting Point Depression of Charge Density Wave in 1T-TiSe2 due to Size Effects

Problem Statement
Classical nucleation theory predicts that nucleation and melting processes become size-dependent at the nanoscale due to surface and confinement effects. However, observing these effects in correlated electronic states, such as charge density waves (CDWs), has been rare. This is largely attributed to the extremely small length scales required to observe such phenomena for electronic states. While superconducting nanowires and other quantum systems have shown size-dependent phase instability when sample dimensions approach the governing correlation length, $\xi(T)$, systematic experimental studies on lateral confinement effects in 2D CDW systems have been inconclusive. Specifically, for 1T-TiSe2, a prototypical 2D CDW material, the stability of the CDW order in nanoflakes with lateral dimensions approaching the intrinsic correlation length or domain size had not been fully explored.

Methodology
The authors investigated the lateral size dependence of CDW stability in 1T-TiSe2 nanoflakes using in-situ cryogenic electron microscopy (cryo-TEM).

• Sample Synthesis: 1T-TiSe2 bulk powder was synthesized via solid-state reaction and subsequently exfoliated into nanoflakes using liquid-phase exfoliation in methanol. The resulting flakes had effective radii ranging from <250 nm down to the tens of nanometers.
• Characterization: Structural integrity was confirmed via High-Angle Annular Dark-Field Scanning Transmission Electron Microscopy (HAADF-STEM) and Electron Energy Loss Spectroscopy (EELS) to verify the 1T phase and identify surface oxidation (amorphous TiOx edges).
• Thermal Measurements: In-situ experiments were conducted using liquid helium (down to 20 K) and liquid nitrogen (down to 100 K) cryo-TEM holders. Nanobeam electron diffraction patterns were collected along the <001> zone axis while gradually heating the samples from base temperature to room temperature in 10–20 K steps.
• Data Analysis: The melting of the CDW phase was determined by the disappearance of the 2x2 superlattice peaks in the diffraction patterns. The melting temperature ($T_{melt}$) was defined as the point where the peak-to-background intensity ratio fell below 5.
• Modeling: Experimental data were fitted to a Ginzburg-Landau (G-L) model for a finite-size system to estimate the zero-temperature correlation length ($\xi_0$). The model assumes a scalar order parameter with vanishing boundary conditions.

Key Results

1. Size-Dependent Melting Point Depression: A systematic decrease in the CDW melting temperature ($T_{melt}$) was observed as the lateral size of the nanoflakes decreased. For flakes with effective radii larger than 100 nm, $T_{melt}$ remained consistent with the bulk value (210–230 K). However, for flakes with effective radii < 100 nm, $T_{melt}$ began to depress. For flakes with radii < 60 nm, the depression exceeded 50 K compared to bulk.
2. Complete Suppression of CDW: In the smallest nanoflakes (effective radii ~50 nm), the CDW transition was completely suppressed; no superlattice peaks were observed even at the base temperature of 20 K.
3. Correlation Length Estimation: Fitting the experimental $T_{melt}$ vs. size data to the G-L model yielded a zero-temperature correlation length ($\xi_0$) in the range of 10–50 nm. When restricting the analysis to flakes with high circularity (shape factor near 1), the estimated $\xi_0$ narrowed to 10–20 nm.
4. Microscopic Origin (Defect Starvation): Atomic-resolution imaging revealed the presence of Ti self-intercalant clusters within the van der Waals gaps, which act as nucleation sites for CDW formation. The spacing between these distorted regions (nucleation sites) was measured to be between 28 nm and 58 nm. This spacing aligns with the $\xi_0$ derived from the G-L model. The authors attribute the complete suppression of CDW in the smallest flakes to "defect starvation," where the flake dimensions are smaller than the mean distance between nucleation sites, preventing the formation of a CDW domain.

Significance and Claims
The paper demonstrates that the electronic phase transition of a CDW in 1T-TiSe2 follows classical nucleation theory, exhibiting size-dependent melting and stability limits dictated by the correlation length.

• Validation of Theory: The observation confirms that when the lateral size of a 2D CDW system approaches the correlation length, the divergence of the CDW correlation length near the transition is cut off, limiting long-range order and lowering the transition temperature.
• Defect Engineering: The study links the macroscopic melting behavior to microscopic defect spacing, suggesting that the distance between Ti intercalants sets a fundamental size limit for CDW stability.
• Technological Implication: The findings indicate a minimum device size required to exploit CDW properties in microelectronic applications, as devices smaller than the correlation length (or the mean distance between nucleation sites) may fail to sustain the CDW state.
• Modest Scope: The authors explicitly note that their temperature readings are based on holder sensors and may have offsets (up to ~15–20 K), but these do not alter the central observation of melting point depression. They also clarify that while thickness variations exist, the CDW transition in 1T-TiSe2 is robust against thickness changes down to monolayers, isolating the observed effects to lateral confinement.