Hierarchical Interdiffusion Kinetics in Nanoscale Ni/Al Multilayers

By combining fast differential scanning calorimetry with correlative STEM across a wide range of heating rates, this study reveals that interdiffusion in nanoscale Ni/Al multilayers proceeds hierarchically, transitioning from grain boundary-dominated transport at low temperatures to lattice diffusion at higher temperatures, thereby establishing grain boundaries as the primary control on reaction onset and microstructural design.

The Gist

Imagine you have a microscopic sandwich made of alternating, ultra-thin slices of nickel and aluminum. These aren't just any slices; they are stacked so tightly that the whole thing is only a few hundred atoms thick. Scientists call these "reactive multilayers." When you heat them up, they are supposed to snap together and react violently, releasing a burst of energy. This is useful for things like igniting tiny rockets or welding parts together without a torch.

But here's the mystery: What happens in the very first split second before that big explosion?

For a long time, scientists knew the sandwich would eventually react, but they didn't understand the "pre-game" warm-up. It's like knowing a car will eventually speed down the highway, but not understanding how the engine turns over or how the gears shift before it hits top speed.

This paper solves that mystery by looking at the nickel and aluminum layers as they start to mix, using a clever combination of super-fast heating and high-tech electron microscopes.

The "Super-Fast Oven" and the "Freeze-Frame" Camera

To see what's happening, the researchers needed to heat the sandwich incredibly fast—up to 10,000 times hotter per second than a normal oven. They used a special chip-based device (a "Fast Differential Scanning Calorimeter") that acts like a super-fast oven.

But heating it up isn't enough; you need to see the result. So, they used a trick: they heated the sandwich to a specific point, then instantly "froze" it (quenched it) so fast that the atoms couldn't move anymore. It's like taking a high-speed photo of a hummingbird's wings. They did this at different stages of the heating process to create a "stop-motion" movie of the reaction.

The Two-Stage Mixing Dance

When they looked at the heat data and the frozen snapshots, they discovered that the mixing doesn't happen all at once. It happens in two distinct steps, like a dance with two different partners:

Step 1: The "Hallway" Run (Low Temperature)
At the beginning, the nickel atoms are shy. They don't want to walk through the middle of the aluminum blocks. Instead, they run along the "hallways" or "corridors" between the aluminum blocks. In scientific terms, these are called grain boundaries.

• The Analogy: Imagine a crowded party in a large room. At first, people (nickel atoms) only move along the edges of the room or the aisles between groups of people (aluminum grains). They haven't entered the crowd yet.
• The Result: The nickel spreads quickly along these edges, but the middle of the aluminum blocks stays mostly empty. This stage releases a little bit of heat.

Step 2: The "Room Invasion" (Higher Temperature)
As the heating continues, the nickel atoms get bolder. They stop just sticking to the edges and start pushing into the middle of the aluminum blocks.

• The Analogy: Now, the people from the aisles start walking into the center of the room, mixing with everyone else. They are invading the "grain interiors."
• The Result: This takes more energy to start, but once it happens, the mixing speeds up dramatically, releasing a lot more heat.

Why This Matters (According to the Paper)

The researchers found that the "hallway" running (grain boundary diffusion) is the main trigger that starts the whole reaction. If you want to control when the sandwich reacts, you need to control the size of the "rooms" (the aluminum grains).

• Small rooms (small grains): More hallways (grain boundaries). The nickel can run everywhere easily, and the reaction starts sooner.
• Big rooms (large grains): Fewer hallways. The nickel has a harder time getting started.

The Big Picture

Before this study, scientists thought the mixing was just one smooth process. This paper shows it's actually a hierarchical process:

1. First, the atoms run along the edges (fast, low energy).
2. Then, they flood into the centers (slower to start, higher energy).

By using their "super-fast oven" and "freeze-frame" camera, the team proved that the "hallways" between the aluminum grains are the most important highways for the reaction to begin. This gives engineers a new way to design these materials: if they want a reaction to start quickly, they should make the aluminum grains smaller to create more "hallways" for the nickel to travel through.

In short: The paper reveals that before the big explosion, the atoms take a two-step dance: first running along the edges of the aluminum grains, and then diving into the middle. Understanding this dance allows us to predict and control exactly when the reaction starts.

Technical Summary

Technical Summary: Hierarchical Interdiffusion Kinetics in Nanoscale Ni/Al Multilayers

Problem Statement
Reactive metallic multilayers (RMLs) store significant chemical energy within their nanostructured architecture, releasing it via self-propagating high-temperature synthesis (SHS) reactions upon thermal, electrical, or mechanical stimulation. While these materials are utilized in applications ranging from reactive joining to micro-propulsion, a comprehensive understanding of the reaction kinetics—specifically the initial stages preceding ignition—remains incomplete. Two fundamental experimental challenges hinder this understanding:

1. Kinetic Regime Limitations: SHS ignition occurs at heating rates of $10^3$–$10^4$ K/s, while propagation exceeds $10^6$ K/s. Conventional calorimetry is typically restricted to rates below a few K/s, failing to access the relevant kinetic regimes where reaction onset is determined.
2. Microstructural Correlation: Reaction pathways are highly sensitive to microstructural features (grain size, defects, interfaces). However, most studies lack near-atomic-scale characterization of transient states, making it difficult to directly correlate calorimetric heat-flow signatures with structural evolution. Furthermore, custom nanocalorimetry setups often deposit films directly onto sensors, limiting throughput and preventing systematic exploration of kinetic regimes.

Methodology
The authors developed an experimental workflow combining fast differential scanning calorimetry (FDSC) with correlative advanced electron microscopy on freestanding Ni/Al multilayers.

• Sample Preparation: Equimolar Ni/Al multilayers with a 20 nm bilayer periodicity (approx. 8 nm Ni / 12 nm Al) were deposited via DC magnetron sputtering. Freestanding films (1 µm and 5 µm thick) were mechanically delaminated from Si substrates to enable high-throughput measurements and ex-situ analysis.
• Calorimetry: FDSC (Mettler-Toledo FDSC 2+) was employed to vary heating rates over five orders of magnitude (0.1 to $10^4$ K/s). This allowed the detection of weak exothermic signals associated with early-stage interdiffusion and the preservation of intermediate states via rapid quenching ($10^4$ K/s).
• Kinetic Analysis: An isoconversional Kissinger–Akahira–Sunose (KAS) analysis was applied to the heat-flow data to extract conversion-dependent activation energies ($E_A$) without assuming a specific reaction model.
• Microstructural Characterization: Samples were arrested at specific reaction stages (as-deposited, ID-01, ID-02, and Peak A) and analyzed using 4D-STEM, High-Angle Annular Dark-Field (HAADF) STEM, and Energy-Dispersive X-ray Spectroscopy (EDX) to map compositional and structural evolution.

Key Results

1. As-Deposited State: The multilayers are nanocrystalline with columnar grains (Ni ~5 nm, Al ~10–several tens of nm) and a strong {111} out-of-plane texture. Significant "premixing" is observed, characterized by pronounced Ni enrichment throughout the Al layers (centers >30 at.% Ni) and ~7 nm wide intermixing zones (IMZs), attributed to a combination of thermodynamic driving forces and ballistic mixing during sputtering.
2. Calorimetric Signatures: Across all heating rates, the heat-flow curve exhibits a broad exothermic shoulder preceding the sharp peaks of intermetallic formation (Al$_9$Ni$_2$, Al$_3$Ni, etc.). This shoulder resolves into two distinct interdiffusion intervals: ID-01 and ID-02.
3. Kinetic Evolution: KAS analysis reveals a non-constant activation energy ($E_A$) that evolves with conversion ($\alpha$).
   • ID-01: $E_A$ reaches a local minimum of 81 ± 24 kJ/mol near $\alpha \approx 0.02$.
   • ID-02: $E_A$ rises linearly to a plateau of 168 ± 17 kJ/mol near $\alpha \approx 0.2$.
   • These values correspond to the empirical relation $E_{lattice} \approx 2E_{GB}$ for fcc metals.
4. Microstructural Evolution:
   • ID-01: SAED confirms no intermetallic phase formation. STEM-HAADF and EDX show only minor compositional changes.
   • ID-02: Pronounced redistribution of Ni occurs within the Al layers. EDX maps reveal localized Ni enrichment with a characteristic spacing of ~5–10 nm, matching the Al grain size. This suggests Ni transport is confined to Al grain boundaries.
   • Peak A: Intermetallic phases (Al$_2$Ni$_9$ and Al$_3$Ni) begin to form, marked by new diffraction reflections.

Key Contributions

• Resolution of Hierarchical Diffusion: The study identifies a hierarchical interdiffusion mechanism where Ni transport proceeds sequentially: first along Al grain boundaries (ID-01, low $E_A$) and subsequently into the Al grain interiors via lattice diffusion (ID-02, high $E_A$).
• Direct Correlation: By combining high-throughput FDSC with state-resolved microscopy, the authors directly link specific heat-flow intervals to distinct microstructural transport pathways, resolving a long-standing ambiguity regarding the relative contributions of grain boundary versus lattice diffusion in RMLs.
• Methodological Framework: The use of freestanding films on chip-based calorimeters enables the systematic mapping of kinetic regimes and the preservation of transient states for ex-situ analysis, overcoming the "single-shot" limitation of sensor-integrated RMLs.

Significance and Claims
The paper claims that grain boundaries act as the dominant transport pathways controlling the onset of reaction in nanoscale Ni/Al multilayers. Consequently, grain size and grain-boundary density are identified as key microstructural design parameters for controlling reaction pathways and heat release. The authors posit that previous literature reporting intermediate activation energies likely reflects the concurrent, unresolved activation of both grain boundary and lattice diffusion.

The study concludes that the combined FDSC–microscopy approach provides a powerful framework for studying defect-mediated transport and non-equilibrium phase transformations in materials driven far from equilibrium, where diffusion, clustering, and phase transformations overlap in time and space. The findings offer a mechanistic basis for understanding reaction initiation in reactive multilayers beyond the specific Ni/Al system.