🔬 materials science

Revealing the Atomic Structure of NiO/Ga $_{2}$ O $_{3}$ Interfaces

This study utilizes aberration-corrected scanning transmission electron microscopy and modeling to reveal the atomistic structures of NiO/Ga $_{2}$ O $_{3}$ interfaces across various substrate orientations, identifying (100)-oriented Ga $_{2}$ O $_{3}$ as the optimal candidate for fabricating high-quality, low-defect power electronic devices while clarifying the impact of imaging artifacts on structural interpretation.

Original authors: Michelle A. Smeaton, Krishna Acharya, Anna Sacchi, Renae N. Gannon, M. Brooks Tellekamp, Andriy Zakutayev, Vladan Stevanovic, Steven R. Spurgeon

Published 2026-08-12

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Michelle A. Smeaton, Krishna Acharya, Anna Sacchi, Renae N. Gannon, M. Brooks Tellekamp, Andriy Zakutayev, Vladan Stevanovic, Steven R. Spurgeon

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the world of electronics as a bustling city where electricity is the traffic. For decades, this city has relied on silicon roads to manage the flow, but as our devices get faster and hotter, those roads are hitting a traffic jam. Enter a new, super-highway material called Gallium Oxide (Ga2O3). Think of it as a futuristic, ultra-wide bridge that can handle massive amounts of electrical power without melting down. However, to make this bridge work in a real device, you need to connect it to a "traffic controller" that can handle the opposite flow of electricity. That's where Nickel Oxide (NiO) comes in, acting as the perfect partner. The big question for scientists has been: when you build a wall between these two materials, does it fit together like a perfect Lego set, or is it a messy pile of mismatched bricks? If the connection is messy, the electricity gets stuck, heat builds up, and the device fails. Understanding exactly how these atoms line up at the microscopic level is the key to building the next generation of super-efficient power electronics for everything from electric cars to the power grid.

In this study, a team of researchers decided to take a microscopic peek at the "seam" where these two materials meet. They grew thin films of Nickel Oxide on three different types of Gallium Oxide crystals, each oriented in a slightly different direction—like laying a carpet on a floor that is tilted at three different angles. Using a super-powerful microscope called a Scanning Transmission Electron Microscope (STEM), which acts like a super-sharp flashlight to see individual atoms, they took high-resolution pictures of the interface. They didn't just look; they also built digital 3D models and ran computer simulations to see if the pictures matched their theories.

The results were a tale of three very different interfaces. When they grew the Nickel Oxide on the (100) oriented Gallium Oxide, the connection was a dream come true. The atoms lined up in a sharp, orderly fashion, like soldiers marching in perfect formation. The researchers found this interface was "abrupt," meaning the transition from one material to the other happened instantly without a messy middle layer. This suggests that the (100) orientation is the best candidate for building high-quality, reliable devices.

However, the other two orientations were much more chaotic. On the (001) and (¯201) surfaces, the atoms didn't line up as neatly. Instead of a sharp wall, there were "reconstructed" layers where the atoms seemed to scramble and rearrange themselves, creating a bumpy, uneven transition. The researchers noticed some strange patterns in the images that looked like they might be a new, unwanted material forming between the layers—specifically a spinel phase called NiGa2O4, which some previous studies had suggested might appear during high-temperature operation.

Here is where the story gets tricky and the researchers had to be very careful. Because the microscope takes a 3D object and squishes it into a 2D picture, sometimes atoms that are actually on top of each other in the third dimension can look like they are forming a new pattern. The team used their computer simulations to test this. They found that those "hexagon" patterns that looked like the mysterious spinel material were actually just optical illusions caused by the way the sample was sliced and viewed. By simulating the images, they proved that these patterns were likely just step edges or overlapping layers, not a new chemical phase forming. This is a crucial finding because it means that in these freshly made samples, the messy interfaces are due to structural mismatch, not necessarily a chemical reaction creating a new layer.

Ultimately, the paper suggests that if we want to build the best power electronics, we should focus our efforts on the (100) oriented Gallium Oxide. It offers the cleanest, most stable connection with the fewest defects. While the other orientations are more complex and prone to strain and disorder, the (100) version provides a clear, sharp path for electricity. The researchers also highlighted a vital lesson for everyone in the field: when looking at these tiny atomic pictures, you have to be careful not to mistake a 3D projection trick for a real new material. By combining sharp eyes, careful slicing of the samples, and smart computer modeling, they've cleared up some confusion and pointed the way toward building better, more robust power devices.

Technical Summary: Revealing the Atomic Structure of NiO/Ga2O3 Interfaces

Problem Statement
NiO/Ga2O3 heterojunctions are critical candidates for next-generation power electronics, leveraging the ultrawide bandgap and wafer-scale availability of $\beta$ -Ga2O3 alongside the controllable p-type doping of NiO. Despite significant progress in understanding carrier transport and crystallinity, the atomic structure of the NiO/Ga2O3 interface remains underexplored. This lack of understanding stems from the complexity of joining dissimilar crystal structures: cubic NiO and monoclinic $\beta$ -Ga2O3. Precise atomic positions and bonding environments at the interface are vital for determining band alignment, electron/phonon transport, and device stability. Furthermore, the potential formation of interlayer phases, such as the NiGa2O4 spinel observed in high-temperature cycling, requires a detailed understanding of the initial interface structure to be effectively utilized or mitigated in device design. While most research focuses on (001)-oriented Ga2O3 substrates due to bulk availability, systematic comparisons of interface structures across different Ga2O3 orientations ((100), ( $\bar{2}$ 01), and (001)) are lacking.

Methodology
The authors investigated the atomistic structure of NiO/Ga2O3 interfaces on (100), ( $\bar{2}$ 01), and (001) oriented $\beta$ -Ga2O3 substrates. NiO films (approx. 90 nm thick) were grown via pulsed laser deposition (PLD). Structural characterization was performed using aberration-corrected scanning transmission electron microscopy (HAADF-STEM) at 200 kV. To ensure data integrity, specimens were prepared using a plasma focused ion beam (PFIB) with low-energy polishing to minimize damage and artifacts.

Crucially, the study employed a rigorous comparative approach combining experimental imaging with computational modeling:

Interface Modeling: Structure matching algorithms based on Lennard-Jones energy minimization were used to generate candidate interface structures, followed by density functional theory (DFT) relaxation.
Image Simulation: Multislice algorithm simulations were performed on the relaxed models to generate theoretical STEM images.
Validation: Experimental HAADF-STEM images were directly overlaid with both the atomic models and the simulated images to validate structural interpretations.
Artifact Differentiation: The authors explicitly analyzed the effects of specimen thickness and 3D-to-2D projection to distinguish real crystal variations from imaging artifacts, a critical step given the complexity of the monoclinic-to-cubic interface.

Key Results

NiO/Ga2O3(100): This interface was found to be nearly atomically abrupt with a well-defined crystallographic registry. The NiO grows in the (001) orientation with an in-plane relationship of NiO [110] || Ga2O3 [010]. The experimental images matched the calculated interface models and simulations with excellent agreement, indicating a sharp, low-defect interface.
NiO/Ga2O3( $\bar{2}$ 01): This interface is more complex. While the NiO grows in the ( $\bar{1}$ 11) orientation with the same in-plane relationship as the (100) case, the interface exhibits reconstructed lattice layers and additional atomic columns not fully captured by the initial models. These features are attributed to the "corrugated" nature of the Ga2O3 ( $\bar{2}$ 01) surface, which influences nucleation.
NiO/Ga2O3(001): This orientation presents the most complex interface. The NiO growth orientation is tilted, estimated to be approximately the (10 10 3) plane rather than the previously suggested (331). The study identified an epitaxial relationship between the tilted Ga2O3 (101) plane and the NiO (1 $\bar{1}$ 1) plane that likely drives initial nucleation. The interface shows significant disorder and structural variation compared to the (100) case.
Projection Effects and Interlayer Phases: The study demonstrated that certain contrast features in STEM images, such as weak atomic columns or hexagonal motifs, can arise from projection effects (e.g., step edges or overlapping lattices) rather than the formation of new phases. Specifically, the authors concluded that the hexagonal motifs observed in the ( $\bar{2}$ 01) and (001) samples do not match the $\gamma$ -Ga2O3 or NiGa2O4 spinel structures, indicating that these interlayer phases have not formed in the as-deposited samples.

Significance and Claims
The paper claims that its primary contribution is the direct visualization and validation of the atomic structure of NiO/Ga2O3 interfaces across three different substrate orientations. By correlating experimental STEM data with DFT-based models and image simulations, the authors provide a pathway to distinguish real structural variations from imaging artifacts in complex heterointerfaces.

The authors assert that their comparative analysis supports the increased focus on (100)-oriented Ga2O3 as the superior candidate for fabricating high-quality, low-defect-density NiO/Ga2O3 heterojunctions. The sharp crystallographic registry and limited structural variation observed in the (100) orientation suggest better interface stability and fewer nucleation sites for deleterious interlayer phase formation (like NiGa2O4) during high-temperature device operation.

Finally, the work highlights the critical importance of considering specimen thickness and projection effects when interpreting STEM data for these interfaces. This caution is deemed essential for future studies regarding interface stability and interlayer phase evolution under actual device operating conditions. While the (100) orientation shows promise, the authors note that low homoepitaxial growth rates on this orientation remain a challenge for efficient device fabrication, though recent advances suggest this is a viable route.

More like this