Direct nanoscale observation of melting and solute redistribution in a hypoeutectic Al-Cu alloy with in situ STEM

Using in situ STEM heating with MEMS technology, this study provides direct nanoscale observation of melting and solute redistribution in a hypoeutectic Al-Cu alloy, revealing that melting initiates at Cu-enriched grain boundaries, the Al$_2$Cu phase melts before the matrix, and liquid-state Cu redistribution extends over 258 micrometers, far exceeding solid-state diffusion limits.

The Gist

Imagine you have a tiny, ultra-thin sheet of metal, like a microscopic piece of aluminum foil mixed with a little bit of copper. This sheet is made of incredibly tiny grains, so small that they are invisible to the naked eye. Scientists wanted to watch what happens to this sheet when it gets hot, but not just "hot" like an oven—hot enough to melt, all while watching it through a super-powerful microscope called a STEM.

Here is the story of what they found, explained simply:

The Setup: A Tiny Hot Plate

The researchers put this tiny metal sheet on a special chip that acts like a miniature hot plate. This chip is so advanced that it can heat up the metal while the scientists watch it in real-time, frame by frame, like watching a high-speed movie. They could also measure how easily electricity flowed through the metal as it changed.

The Story of the Melting: A Crowd Moving Outward

When they started heating the metal, something interesting happened. It didn't melt all at once like an ice cube in a warm room. Instead, it started melting in the very center of the chip, which was the hottest spot.

Think of the metal grains like a crowded dance floor.

1. The Warm-Up: First, the dancers (the metal grains) got bigger and more organized. The tiny copper atoms, which were hiding between the aluminum dancers, started gathering at the edges of the dance floor (the grain boundaries).
2. The First to Melt: Because the copper gathered at the edges, those spots turned into liquid first. It's like the edges of the dance floor turned into a slippery, wet zone while the center was still solid.
3. The Wave: The melting didn't stop there. It started in the middle and spread outward, like a wave moving across a pond. The center of the metal sheet began to turn into a puddle.

The Great Escape: The Marangoni Effect

Once the metal turned into a liquid, it didn't just sit there. It started to move. The scientists saw the liquid metal flow away from the hot center and pile up at the cold edges of the chip.

Why did it do this? Imagine a drop of water on a hot pan. If one side of the drop is hotter than the other, the "skin" (surface tension) on the hot side gets weaker, and the skin on the cool side is stronger. The strong skin pulls the liquid toward the cool side.

In this experiment, the heat in the center made the liquid metal "slippery" (low surface tension), while the cooler edges were "sticky" (high surface tension). The sticky edges pulled the liquid metal away from the center, dragging the copper with it. This is called the Marangoni effect.

The Result: A Depleted Center and a Copper-Rich Edge

Because of this flow, the center of the metal sheet was left almost empty, like a stage after the actors have run off. The copper, which loves to move with the liquid, ended up piling up at the very outer edges of the chip.

The scientists measured this movement and found it was massive. In the time it took to melt, the copper traveled a distance that is thousands of times longer than it could ever travel if the metal were still solid. It was like watching a person run across a country in the time it usually takes them to walk across a room. This proved that the copper was moving through the liquid, not the solid.

The Electrical Clue

The scientists also watched the electricity. Before melting, as the grains got bigger, the electricity flowed easier (resistance went down). But the moment the metal started to melt and flow away, the electricity struggled to get through, and the resistance shot up until the connection was broken. This was like a bridge collapsing as the road washed away.

The Big Picture

This study is special because it's the first time anyone has watched these tiny processes happen in real-time with such detail. They saw exactly how the metal grains grew, how the copper gathered at the edges to start the melting, and how the liquid flowed away due to temperature differences.

This helps us understand what happens inside metals when they are heated quickly, which is important for things like 3D printing with metal, welding, or casting. But mostly, it showed us that when tiny metals melt, they don't just turn into a puddle; they dance, flow, and rearrange themselves in a very specific, predictable way driven by heat and surface tension.

Technical Summary

Technical Summary: Direct Nanoscale Observation of Melting and Solute Redistribution in a Hypoeutectic Al-Cu Alloy with In Situ STEM

Problem Statement
While the melting and solidification of eutectic systems are classical topics in physical metallurgy, the mechanisms governing these processes at the nanoscale remain poorly understood due to limitations in spatiotemporal resolution. Specifically, the behavior of nanocrystalline (NC) and ultrafine-grained (UFG) metallic materials differs from coarse-grained counterparts, with high grain-boundary volume fractions expected to cause boundary melting and spatially distributed onset temperatures. In multi-component NC/UFG systems, shortened diffusion distances should theoretically accelerate eutectic liquid formation and solute partitioning. However, previous in situ Scanning Transmission Electron Microscopy (S/TEM) studies have largely addressed grain growth and solid-state phase transformations, leaving the melting regime and its immediate solute-redistribution consequences essentially unobserved directly.

Methodology
The authors investigated a nanocrystalline hypoeutectic Al–Cu alloy (Al92Cu8 at.%) using a combined experimental approach centered on in situ electrothermal S/TEM.

• Sample Preparation: A 100 nm thick Al92Cu8 thin film was deposited via non-reactive magnetron sputtering directly onto an electrothermal MEMS chip (Protochips Fusion AX holder) featuring nine electron-transparent SiN windows.
• Experimental Setup: The study utilized a Thermo Fisher Scientific Talos F200X G2 microscope operated at 200 kV in STEM-HAADF mode with Energy Dispersive X-ray (EDX) spectroscopy.
• In Situ Conditions: The sample was heated using MEMS-based electrical biasing with temperature control via Keithley 2450 source meter and Protochips Fusion Clarity software (100 ms time step).
• Analysis Techniques: The study integrated real-time imaging, automated change-detection analysis (Python-based frame averaging and thresholding), Selected Area Electron Diffraction (SAED), and simultaneous four-point electrical resistance measurements to correlate microstructural evolution with electrical properties.

Key Results

1. Microstructural Evolution and Melting Onset:

   • Prior to melting, the film underwent grain coarsening, with grain sizes increasing from ~20 nm (as-deposited) to ~300 nm after annealing at 500 °C.
   • Melting did not occur instantaneously across the chip. Instead, it initiated in the hotter central region and propagated outward.
   • Grain boundaries acted as preferred sites for eutectic liquid formation due to Cu enrichment. The Al2Cu (θ-phase) melted prior to the complete melting of the Al-rich matrix.
   • Diffraction data confirmed the disappearance of θ-phase reflections above the eutectic temperature but before the complete melting of the Al matrix.

2. Lateral Solute Redistribution:

   • Upon melting, pronounced lateral redistribution of Al and Cu occurred over a distance of approximately 258 µm.
   • The central region of the observation window became depleted of both elements, while material accumulated at the outermost edges of the chip.
   • EDX mapping revealed that in the central depleted zone, Cu segregation was lost, whereas the outer rim showed significant Cu enrichment.
   • Post-quench aging experiments confirmed that Cu remained dissolved in the Al-rich matrix after melting and quenching, subsequently re-precipitating at grain boundaries during aging. The final Cu concentration (~2.16 at.%) was significantly lower than the initial film composition (8 at.%), confirming solute loss from the central region.

3. Transport Mechanisms:

   • The observed transport distance (258 µm) far exceeds the limits of solid-state diffusion (estimated at ~5 µm for the same duration) and aligns with liquid-state diffusion coefficients.
   • The authors attribute the material flow to the Marangoni effect, where a temperature gradient creates a surface tension gradient ($\partial\gamma/\partial T < 0$), driving liquid flow from the hot, low-surface-tension center to the cold, high-surface-tension edges.
   • Capillarity-driven retraction and breakup of the molten thin film further contributed to the formation of rim accumulations.

4. Electrical Response:

   • Resistance decreased by 35% during solid-state heating (up to 500 °C) due to grain growth and reduced grain boundary scattering.
   • Upon melting, resistance increased by several orders of magnitude until electrical contact was lost due to the large-scale redistribution of material.

Significance and Claims
The paper claims to provide, for the first time at nanoscale and millisecond spatiotemporal resolution, a direct observation of the complex melting and solute-partitioning evolution in a hypoeutectic system. The study demonstrates that in situ electrothermal S/TEM can directly probe:

• The spatial propagation of melting from hot to cold regions.
• The role of grain boundaries as facilitators for eutectic liquid formation.
• The transition from solid-state coarsening to liquid-mediated transport mechanisms (Marangoni flow).

The authors assert that this methodology offers a versatile platform for studying fundamental phase-transformation phenomena (including segregation, eutectic reactions, and local melting) in metastable nanostructured materials. These insights are directly relevant to understanding microstructural control in engineering processes such as additive manufacturing, welding, sintering, and casting, where transient partial melting and solute redistribution are critical. The work establishes a direct structure–chemistry–property relationship during dynamic phase transformations.