Remote epitaxial frustration stabilizes a correlated interfacial state

In the world of materials science, building a new crystal often feels like trying to fit a complex puzzle piece onto a specific table. The goal is to make the piece sit perfectly flat, aligning its internal pattern with the table's surface so that the new structure grows in an orderly, predictable way. This process, known as epitaxy, usually relies on the new material "feeling" the pattern of the table beneath it. However, scientists have discovered a way to grow crystals on top of an ultra-thin sheet of carbon, called graphene, which acts as a transparent window. Through this window, the crystal can still sense the table below and align itself accordingly, a phenomenon called remote epitaxy. For years, researchers assumed this remote sensing always led to a single, perfect alignment, but the question remained: what happens if the forces pulling the crystal in different directions become equally strong?

A team of researchers at the University of Wisconsin-Madison, working with colleagues at Sandia National Laboratories and Pennsylvania State University, has now found that when these forces are balanced just right, the result is not a perfect crystal, but a state of "frustration." In this context, frustration does not mean the material is confused or broken; rather, it describes a situation where the competing pulls from the carbon sheet, the table below, and the interface between them are so evenly matched that the material cannot choose a single, long-range order. Instead, it settles into a unique, self-limited state where order exists only over a very short distance. This discovery is significant because it reveals a new way to engineer materials, creating stable, finite-range structures that exhibit unusual magnetic properties, distinct from both perfect crystals and random, amorphous solids.

To explore this, the team grew thin films of a material called GdAuGe, which contains the rare-earth element gadolinium, on top of silicon carbide crystals covered with varying numbers of graphene layers. They tested everything from no graphene at all to one, two, or more layers, and even a special case where the graphene was chemically separated from the silicon carbide. When they grew the films on bare silicon carbide or on the separated graphene, the result was exactly what traditional physics predicts: a clean, crystalline interface where the film aligned perfectly with the underlying table. However, when they used an intermediate number of graphene layers—specifically one or two layers—the outcome changed dramatically.

Using powerful electron microscopes to look at the atomic structure, the researchers observed that the interface between the film and the graphene did not form a long, continuous crystal lattice. Instead, it formed a thin, self-limited layer, only two or three atoms thick, that possessed local order but lacked long-range alignment. It was as if the atoms knew how to arrange themselves with their immediate neighbors but could not agree on a pattern that extended across the whole surface. This state was not a mistake or a result of the material failing to crystallize. The team proved this by heating the samples in stages. They watched the material start as a disordered, amorphous seed, transform into a perfect crystal, and then, upon further heating, evolve into this frustrated, short-range ordered state. This progression showed that the frustration was a stable, thermodynamic state that the material actively chose, rather than a kinetic trap where the atoms simply got stuck before they could finish ordering.

The researchers also noticed that the direction in which the crystal grew changed in a surprising, non-linear way. On bare silicon carbide, the film aligned in one specific direction. On the separated graphene, it aligned in that same direction. But on the intermediate layers of one or two graphene sheets, the film suddenly switched to a different orientation. This switch happened even within a single sample, where the film would change its alignment simply because the number of graphene layers beneath it changed by a single atomic sheet. This behavior confirmed that the material was responding to a complex interplay of forces transmitted through the graphene, rather than just copying the table below or bonding directly to the carbon.

To understand why this happened, the team turned to computer simulations to map the invisible forces acting on the film. They calculated the electrical potential landscape created by the silicon carbide, the graphene, and the reconstructed interface between them. Their calculations revealed that at intermediate graphene thicknesses, the forces coming from the substrate, the graphene sheet, and the interface reconstruction all had similar strengths but different patterns. Because these patterns did not line up, the film faced a landscape with multiple, nearly equal options for how to arrange itself. Unable to find a single, clear path to a perfect crystal, the system settled into the frustrated state. This microscopic competition provided the physical basis for the broken order observed in the experiments.

Perhaps the most striking consequence of this structural frustration was its effect on magnetism. The researchers measured the magnetic response of the films and found that the frustrated samples behaved very differently from their crystalline or amorphous counterparts. While the magnetic signal in normal crystals usually depends on the total volume of the material, the signal in the frustrated films scaled with the surface area of the interface. Furthermore, these films showed a strong, persistent magnetic irreversibility that lasted well above room temperature, up to 300 Kelvin. This means that the magnetic state of the material did not simply flip back and forth instantly; it held onto a memory of its magnetic history in a way that the other samples did not. This suggests that the short-range structural order created a unique magnetic environment, likely due to a mix of strain and atomic coordination that created a complex landscape for the magnetic spins.

The work demonstrates that by carefully tuning the thickness of a single atomic layer, scientists can move a material from a state of perfect order to a state of controlled frustration, and then to a state of disorder. This ability to stabilize finite-range order offers a new tool for designing materials with specific collective properties. The study shows that frustration is not just a barrier to be overcome in crystal growth, but a regime that can be harnessed to create new states of matter with distinct electronic and magnetic behaviors. By understanding how competing interactions can prevent a material from settling into a single pattern, researchers have opened a path to engineering interfaces where the whole is greater than the sum of its parts, creating stable, functional states that exist nowhere else in nature.