A 50°C change determines whether ultrathin magnetic films stay intact or disintegrate

Magnetic memory could be approaching a major materials breakthrough after researchers in Japan developed a two-step heating method that produces ultrathin iron–palladium films with an unusual combination of atomic order, magnetic stability, and surface control. The advance targets a long-standing problem in next-generation data storage: how to create magnetic films that are structurally perfect enough to retain information reliably while remaining smooth and continuous enough for practical devices.

The films are made from a 5-nanometer-thick alloy known as L1₀-FePd. In this crystal structure, iron and palladium atoms occupy alternating positions in a highly ordered tetragonal lattice. That arrangement generates strong magnetocrystalline anisotropy, a property that forces the material’s magnetization to favor a particular direction. In magnetic memory, high anisotropy is crucial because it helps prevent thermal fluctuations from randomly reversing the magnetization and erasing stored information.

L1₀-FePd is especially attractive for magnetic random-access memory because it also has relatively low magnetic damping. This means that its magnetization can potentially be switched with less energy than in many conventional magnetic materials. However, the properties of an ultrathin film depend not only on its chemical composition but also on how its atoms arrange themselves and how smoothly the film covers its substrate. Increasing the temperature usually improves atomic ordering, but it can simultaneously cause roughness, grain growth, and loss of film continuity.

A team led by Professor Hiroshi Naganuma at the University of Toyama investigated whether these competing effects could be separated through a carefully controlled thermal process. Using radio-frequency magnetron sputtering, the researchers deposited iron and palladium onto a strontium titanate substrate at selected relatively low temperatures. During this initial stage, the incoming atoms had limited mobility, allowing them to remain close to their landing sites and form a smooth film through layer-by-layer growth.

After deposition, the samples were cooled to room temperature and then annealed at 600°C. This second heating step supplied the energy needed for iron and palladium atoms to diffuse through the film and rearrange into the ordered L1₀ crystal structure. Instead of exposing the growing film to high temperatures throughout fabrication, the method separates smooth film formation from atomic ordering, creating a process that gives researchers independent control over two normally linked properties.

The results showed that a difference of only 50°C during the first stage dramatically changed the film’s later evolution during annealing. When the initial deposition temperature was 150°C, the film remained smooth and continuous. Its growth followed the Frank–van der Merwe mode, in which material spreads across the substrate in successive layers. After the 600°C anneal, this sample developed strong perpendicular magnetic anisotropy while preserving the flat morphology considered desirable for magnetic memory devices.

At an initial temperature of 200°C, the researchers observed a more surprising transformation. Atomic diffusion during the first stage appeared to create defects that later acted as starting points for solid-state dewetting. During dewetting, atoms migrate across an ultrathin film because the system can lower its total surface free energy by changing its shape. In this case, the continuous film developed square holes that extended into the underlying substrate, while the remaining material achieved nearly perfect L1₀ ordering.

When the initial temperature reached 300°C, diffusion became even more pronounced, and the film evolved into a rough, island-like structure. Although such a morphology is unsuitable for many conventional memory architectures, it may offer a route toward deliberately engineered nanoscale magnetic patterns. The researchers’ observations indicate that solid-state dewetting is not triggered solely by the surface-energy difference between FePd and the substrate. Defects produced during the initial heating stage appear to be necessary to initiate the instability.

First-principles calculations supported this interpretation by linking the driving force for dewetting to the relative surface free energies of the film and strontium titanate. The calculations also helped explain why the process did not begin identically in every sample. The initial thermal history determines the number and nature of defects, which then govern how atoms move during the high-temperature anneal. This finding transforms dewetting from a potentially destructive failure mode into a controllable tool for designing magnetic nanostructures.

The study suggests that films deposited at 150°C may be the most suitable for MRAM, where a flat, continuous surface and strong perpendicular anisotropy are essential for reliable device fabrication. Meanwhile, the controlled hole formation observed at 200°C could be useful for producing self-organized arrays and other nanoscale magnetic architectures for ultra-high-density storage. By tuning only the first-stage temperature, manufacturers may be able to select between smooth memory layers and intentionally patterned magnetic structures without introducing a separate lithography step.

The work, published online in the Journal of Alloys and Compounds, provides a practical strategy for engineering ultrathin magnetic alloys at the interface of thermodynamics, nanotechnology, and spintronics. It also offers a new way to think about imperfections: defects that would normally be viewed as damaging can serve as carefully positioned triggers for atomic rearrangement. As data centers, consumer electronics, and artificial-intelligence systems demand faster and more energy-efficient memory, this two-step approach could help advance magnetic storage technologies capable of holding denser data while consuming less power.

News Publication Date: 26-Jul-2026

Web References: https://doi.org/10.1016/j.jallcom.2026.189751

References: 10.1016/j.jallcom.2026.189751

Keywords

L1₀-FePd, magnetic films, MRAM, magnetic storage, spintronics, solid-state dewetting, epitaxial films, magnetocrystalline anisotropy, nanotechnology, thin-film materials science