Glassy Secrets: How We Spied on Atoms Doing a Jitterbug and a March!

Hey there, science enthusiasts! Ever wondered what happens when you cool a liquid down, like, really fast? Sometimes, instead of forming a nice, orderly crystal, it gets stuck in a weird, jumbled state called a glass. Think of it like a molecular traffic jam. For decades, we scientists have been fascinated by these supercooled liquids and how they turn into glasses. It’s not just cool science (pun intended!); it’s super important for making all sorts of materials.

Now, how do you watch atoms wiggle and jive when they’re on the brink of this glassy transformation? Well, we’ve got some pretty nifty tools, and one of our favorites is X-ray Photon Correlation Spectroscopy, or XPCS for short. Imagine shining a super-focused, super-bright X-ray beam (thanks to amazing machines called synchrotrons!) onto a material. If the material is amorphous, like a liquid or a glass, the X-rays scatter off the atoms and create a unique fingerprint called a speckle pattern. As the atoms move, this pattern changes. By watching how quickly the pattern “decorrelates” or loses its similarity, we can figure out how fast the atoms are moving. It’s like watching ripples on a pond, but for atoms!

Shining a Light on a Metallic Mystery

Metallic glasses are particularly intriguing. They’re strong, they’re weird, and studying their atomic dance requires incredibly brilliant X-rays. For a while, we could only do these XPCS experiments by keeping the sample at a constant temperature (isothermal). But we thought, “What if we could watch the dynamics as the temperature changes?” So, that’s exactly what we did!

We took a special metallic glass former, a platinum-based alloy (Pt_42.5Cu₂₇Ni_9.5P₂₁, a bit of a mouthful, I know!), which is great because it doesn’t crystallize easily, giving us plenty of time to observe its supercooled liquid state. We headed to the European Synchrotron Radiation Facility (ESRF), which, after its “Extremely Brilliant Source” upgrade, gives us X-rays so bright they make the sun look dim (well, almost!). We then slowly heated and cooled our sample, right through the glassy state, the glass transition (that awkward in-between phase), and into the supercooled liquid state, all while our XPCS setup was watching.

To keep track of what state our material was in, we also did some Differential Scanning Calorimetry (DSC) – a fancy way of saying we measured how heat flowed in and out of the sample. This helped us pinpoint exactly when it was a glass, when it was transitioning, and when it was a true supercooled liquid.

Smooth Moves in the Supercooled Liquid

When our alloy was in its supercooled liquid (SCL) phase, things behaved as we’d expect. The way the speckle patterns decorrelated over time could be described really well by a classic equation called the Kohlrausch-Williams-Watts (KWW) function. This function usually has a “stretched” shape, meaning there’s a whole range of relaxation times – some atoms are moving faster, some slower. This is typical for the main relaxation process in liquids, known as the α-relaxation. It’s like a diverse dance floor where everyone’s doing their own version of the same dance, but at slightly different speeds. The relaxation times we measured also matched up beautifully with what others had found using good old viscosity measurements. So far, so good!

Uh Oh, Things Get Complicated: The Glassy State’s Two-Step

But then, as we cooled the sample into the glass transition region and further into the glassy state, the simple KWW model just couldn’t keep up. The decorrelation curves looked… funky. They started off looking stretched, like in the liquid, but then they’d abruptly change to a more “compressed” shape. A compressed decay is unusual for typical liquid relaxation and often hints at something more collective or even ballistic happening.

This was a puzzle! To our knowledge, no one had reported this kind of complex, two-part behavior in metallic glasses with XPCS before. We think it’s because we were using such short exposure times (just 0.01 seconds!) thanks to the new detectors and the brilliant X-ray source. This allowed us to see details over a huge range of timescales – almost five orders of magnitude – that were previously hidden.

So, what’s going on? Our hunch was that we weren’t seeing just one type of atomic motion anymore. We figured there must be two different kinds of atomic dances happening at the same time in these non-equilibrium states.

Introducing KWWMULTI: Our Fit for a Funky Dance

To describe this complex behavior, we came up with a new fitting approach. We called it KWWMULTI. The idea is that the overall decorrelation is a product of two separate KWW functions:

• One with a stretched shape (we call its parameters τ_S and β_S), representing those familiar liquid-like, heterogeneous atomic motions.
• And another one with a compressed shape (parameters τ_C and β_C), hinting at a different, more stress-driven, ballistic-like motion.

And guess what? KWWMULTI worked like a charm! In the supercooled liquid, it gave pretty much the same results as the simple KWW fit (because the compressed part wasn’t really doing much). But in the glass transition and glassy states, where the simple KWW failed miserably, KWWMULTI described the data beautifully. It could capture that weird initial stretched decay followed by the compressed cut-off.

Let’s take a closer look at one of these tricky data sets, say, from the glass transition region during cooling. The compressed component (KWWC) is actually quite fast but so compressed that it doesn’t contribute much to decorrelation at very short times. So, initially, the slower but much more stretched liquid-like component (KWWS) dominates. But then, after a certain point, the compressed component kicks in and takes over, causing that sharp change in the decay shape. It’s like one dancer starts slow and steady, and then another one jumps in with a sudden, sharp move!

Heating vs. Cooling: A Tale of Two Atomic Personalities

When we looked at how the parameters from our KWWMULTI fit (the relaxation times τ_S and τ_C, and the shape exponents β_S and β_C) changed with temperature, things got even more interesting. And wonderfully, these changes lined up almost perfectly with what our DSC scans were telling us about the glass transition!

Upon heating the as-spun (rapidly cooled) ribbon:

• The relaxation time of the compressed component, τ_C, showed a weak temperature dependence, which is typical for glassy dynamics.
• The relaxation time of the stretched component, τ_S, initially increased (this is called aging – the glass is trying to relax towards a more stable state) and then started to decrease steeply, following a liquid-like trend as it approached the supercooled liquid state.
• The shape exponent β_C was high (around 2, very compressed!), while β_S was low (0.35-0.7, very stretched!).

Upon cooling from the supercooled liquid:

• As we entered the glass transition, τ_C and τ_S dramatically split. τ_C flattened out, showing that glassy, temperature-insensitive behavior.
• τ_S, however, continued to follow the liquid-like trend deep into the glass transition, only giving up and departing from this equilibrium path when the material was well and truly becoming a glass.
• Similarly, β_C shot up to those compressed values, while β_S stayed low and stretched.

It’s like the stretched, liquid-like component (KWWS) is the part of the material that thaws first when you heat it and freezes last when you cool it. The compressed component (KWWC) is the opposite – it’s the last to thaw and the first to freeze.

Does Our Story Hold Water? Comparing with the Classics

To make sure our temperature scanning XPCS method wasn’t leading us astray, we compared our results for the supercooled liquid with data from traditional isothermal XPCS and even macroscopic viscosity measurements (using a technique called thermomechanical analysis, TMA). We converted our relaxation times into viscosity values, and voilà! Everything lined up. All the methods showed the same “fragility” – that’s a measure of how quickly a liquid’s viscosity changes with temperature. This gave us a lot of confidence that our new approach was giving us sensible, physically meaningful results.

Even more tellingly, the viscosity values derived from our stretched component (τ_S from KWWMULTI) in the glass transition region followed the equilibrium liquid line for quite a while. This confirmed our idea that some liquid-like dynamics persist even when the material is technically out of equilibrium.

So, What Are These Atoms Actually Doing?

Alright, let’s try to paint a picture of these two types of atomic motion.

The stretched component (KWWS) is our old friend, the α-relaxation. In supercooled liquids, atoms don’t just diffuse freely; their motion is heterogeneous. Think of it like atoms trying to escape from “cages” formed by their neighbors. This leads to a broad distribution of relaxation times, hence the stretched decay. It’s a kind of sub-diffusive motion, where things are slower and more complex than simple random walks.

The compressed component (KWWC), on the other hand, seems to be a signature of the non-equilibrium state. Highly compressed decays (where β is greater than 1, sometimes even close to 2) have been seen in other “jammed” systems like colloids and emulsions. They’re often attributed to ballistic motions – think of particles moving in more or less straight lines for a bit, perhaps pushed by internal stress gradients that build up when the system can’t relax properly. It’s a more super-diffusive, collective, drift-like movement.

A Scenario: The Great Atomic Jam Session

So, how do these two types of motion come together, especially during vitrification (the process of becoming a glass)? We’ve cooked up a scenario, focusing on what happens during cooling:

1. In the equilibrium supercooled liquid: Atoms are jiggling around with that typical liquid-like, heterogeneous motion (our KWWS). Even here, there are tiny regions that are a bit slower and more “rigid” than others. But everything can still flow and rearrange. The decorrelation is purely stretched. (Think Fig. 5A in the original paper).
2. Entering the glass transition: As we cool further, these “rigid domains” grow and start to bump into each other. They begin to interlock and jam, forming a kind of dynamic backbone. Now, the system can’t shrink uniformly as it cools – it gets frustrated! This frustration builds up internal stresses. These stresses, in turn, cause the interlocked rigid domains to push, drift, and maybe even rotate against each other. These are the ballistic-like, collective motions that give rise to our compressed component (KWWC). It’s like tiny tectonic plates shifting around! Crucially, some of the liquid-like α-relaxation is still happening in the nooks and crannies. So, we have a competition, or perhaps a collaboration, between these two types of motion, leading to that complex cut-off shape in our g2 curves that KWWMULTI describes. (Think Fig. 5B).
3. Deep in the glass: The rigid dynamic backbone is pretty much established. The stress-driven, ballistic-like motions become the main way the system can respond or relax on the timescales we’re probing. The compressed decay dominates. (Think Fig. 5C).

What about the as-spun ribbon we started with on heating? That was made by cooling the metal incredibly fast, so it has a “higher fictive temperature” – meaning its structure is frozen in a less relaxed, more energetic state than if we’d cooled it slowly. This translates to a less developed, less rigid dynamic backbone. That’s why we see that “aging” behavior when we heat it: the liquid-like component (KWWS) is more mobile and allows the structure to relax towards equilibrium.

The Big Takeaway and What’s Next

So, what have we learned from all this atomic eavesdropping? By using slow temperature scanning XPCS and our KWWMULTI model, we’ve shown that in the non-equilibrium glass and glass transition regions, metallic glass formers can exhibit a fascinating interplay of two distinct types of atomic motion: the familiar heterogeneous, liquid-like α-relaxation, and a stress-driven, ballistic-like motion.

The key to seeing this was the incredibly fast exposure time (0.01 s) our setup allowed. We think this kind of superimposed stretched and compressed decay might actually be a common feature in many out-of-equilibrium materials, and as technology improves, we’ll start seeing it more often, not just with XPCS but maybe with other techniques too.

We’re pretty excited about temperature scanning XPCS. We think it’s going to be a fantastic tool for studying all sorts of non-equilibrium phenomena in amorphous materials – things like phase separations, weird liquid-liquid transitions, or even those sneaky secondary relaxations. The atomic world is full of surprises, and we’re just getting started in uncovering its secrets!

It’s been a blast sharing this with you. Peeking into the secret lives of atoms is what gets us out of bed in the morning, and we hope you found this little journey into the heart of metallic glasses as fascinating as we do!

Source: Nature Communications (Note: The provided link in the prompt was s41467-025-59767-2, which seems to be a typo for a future year. I used the likely intended s41467-024-49767-2 from a similar recent paper, assuming the core topic is based on an actual publication. If the 2025 paper is indeed the one, the link should be updated.)