Dislocations in (011)-oriented vertical Bridgman $\beta$-Ga$_2$O$_3$ substrates

This study utilizes X-ray topography and reticulography to characterize dislocation arrays and domain boundaries in (011)-oriented vertical Bridgman $\beta$-Ga$_2$O$_3$ substrates, revealing their specific crystallographic orientations and providing critical insights into defect formation relevant to epitaxial growth and device performance.

The Gist

Imagine you are trying to build a skyscraper, but instead of concrete and steel, you are building it out of a special, ultra-hard crystal called beta-gallium oxide (β-Ga2O3). This crystal is like a super-hero material for future electronics because it can handle massive amounts of electricity without breaking, making it perfect for high-power devices like electric car chargers or smart grid systems.

To build a good skyscraper, you need a perfect foundation. In the world of electronics, this foundation is a substrate (a slice of the crystal). Scientists have been trying to figure out the best way to slice this crystal. For a long time, they sliced it one way, but it was full of tiny cracks and pits that ruined the building. Recently, they started slicing it a different way (the (011) orientation), and it seemed much smoother and stronger.

However, even with this "better" slice, there were still some invisible problems hiding inside. This paper is like a detective story where the researchers used special "X-ray glasses" to see these hidden flaws in the (011) crystal slices.

Here is what they found, explained simply:

1. The "X-Ray Glasses" (The Tools)

The researchers didn't just look at the crystal with a regular microscope. They used X-ray topography, which is like taking a 3D X-ray movie of the crystal.

• Transmission Mode: They shot X-rays through the crystal (like looking through a window) to see defects deep inside.
• Reflection Mode: They bounced X-rays off the surface (like a mirror) to see what was happening right at the top.
• Reticulography: This was their "grid test." They projected a mesh pattern onto the crystal. If the crystal was perfect, the grid would look straight. If the crystal had twisted sections, the grid would warp. This helped them find invisible boundaries between different crystal sections.

2. The "Traffic Jams" (Dislocation Arrays)

Inside the crystal, the atoms are supposed to line up in perfect rows, like soldiers in a parade. Sometimes, a row gets messed up, creating a "dislocation" (a defect).

• The Findings: The researchers found that many of these defects weren't just random scattered soldiers. They were lined up in long, straight arrays (like a traffic jam on a highway).
• The Location: These traffic jams were sitting on a specific flat plane inside the crystal called the (001) plane.
• The Direction: The defects were stretching out along the [010] direction (think of this as the crystal's "spine" or main axis).
• The Cause: These arrays were actually marking the borders between different "neighborhoods" in the crystal called domains. Imagine a city where one neighborhood is built slightly tilted compared to the next. The line where they meet is where these defect traffic jams form. The researchers measured this tilt to be incredibly small (about 0.00001 radians), but enough to cause issues.

3. The "Ghost Defects" (The (011) Plane)

There was a specific type of defect that scientists were worried about. In the old way of slicing the crystal (the (001) orientation), these defects would poke out of the surface and create long, ugly scratches (line-shaped pits) that ruined the electronics.

• The Good News: When they looked at the new (011) slices, they found that most of these "scratch-makers" were lying flat, parallel to the surface, so they didn't poke out. This explains why the (011) surface is so smooth.
• The Twist: However, the researchers did find some defects lying on the (011) plane, stretching along the [100] direction. But here is the catch: these were different from the "scratch-makers" found in the old crystals. They didn't look the same.
• The Mystery: The paper notes that the "scratch-makers" found in previous studies were grown using a different method (called EFG), while these new crystals were grown using a method called Vertical Bridgman (VB). This suggests that how you grow the crystal matters just as much as which way you slice it.

4. The Big Picture

The main takeaway is that the (011) crystal isn't just a "perfect" version of the old one. It has its own unique personality.

• It has fewer surface scratches (which is great).
• But it has these hidden "traffic jams" of defects along the domain boundaries.
• The type of defects you find depends heavily on the growth method (VB vs. EFG).

In summary: The researchers used advanced X-ray techniques to map the hidden "fault lines" inside a new type of super-crystal. They discovered that while this new crystal orientation avoids the surface scratches of the past, it still has internal structural boundaries where defects gather. Understanding exactly where these defects live and how they behave is crucial for engineers who want to build the next generation of powerful, efficient electronics.

Technical Summary

Technical Summary: Dislocations in (011)-Oriented Vertical Bridgman $\beta$-Ga$_2$O$_3$ Substrates

Problem Statement
While $\beta$-Ga$_2$O$_3$ is a promising candidate for next-generation power devices due to its ultra-wide bandgap and high breakdown field, achieving high-quality homoepitaxy remains challenging. The choice of substrate orientation is critical; while (100) and ($\bar{2}$01) orientations suffer from twin defects, and (001) and (010) orientations often exhibit line-shaped pits and surface roughness requiring post-growth polishing, the (011) orientation has emerged as a promising alternative. Previous studies suggest that (011) substrates yield smooth, pit-free surfaces because dislocations responsible for line-shaped pits on (001) are nearly parallel to the (011) plane and do not intersect the surface. However, fundamental questions remain regarding the detailed nature of dislocations in (011) substrates, particularly those grown by the Vertical Bridgman (VB) method versus the Edge-defined Film-fed Growth (EFG) method. Specifically, the behavior of dislocations propagating along the b-axis ([010]) and the existence of other dislocation types in VB-grown (011) crystals are not fully understood.

Methodology
The authors investigated dislocation behavior in (011)-oriented $\beta$-Ga$_2$O$_3$ substrates grown via the Vertical Bridgman method in an N$_2$+O$_2$ ambient using a Pt–Rh crucible. The substrates, approximately 460 $\mu$m thick, were mechanically polished and chemically mechanically polished (CMP) on both sides to facilitate clear observation.

The study employed a multi-modal X-ray characterization approach:

1. Transmission X-ray Topography (XRT): Performed at SPring-8 (BL24XU) using the anomalous transmission (Borrmann effect) with a diffraction vector $g = 7\bar{1}2$ and $\lambda = 0.124$ nm. This allowed for the visualization of dislocations penetrating the entire substrate thickness.
2. Reflection XRT: Conducted at KEK-PF (BL-3C) using asymmetric ($g = 73\bar{2}\bar{2}$) and symmetric ($g = 02\bar{2}\bar{2}$) reflections to probe the near-surface region (penetration depth of several micrometers).
3. X-ray Reticulography: Performed based on reflection XRT with a metallic mesh to detect domain boundaries and measure misorientation angles with high sensitivity.

Key Results

• Dislocation Arrays and Domain Boundaries: Transmission XRT revealed that most dislocations lie on the (001) plane and extend along the [010] direction (b-axis). These dislocations form arrays (designated DA type-I) that terminate at the surface along the [100] and [001] directions. Reticulography confirmed that these arrays are associated with domain boundaries exhibiting misorientations on the order of $10^{-5}$ rad.
• Dislocations on the (011) Plane: The study identified dislocations lying on the (011) plane extending along the [100] direction. While some showed deflection onto other planes, these were distinct from the dislocations responsible for line-shaped pits on (001) epilayers.
• Growth Method Dependence: The observed dislocations on the (011) plane in VB-grown crystals differ from those reported in EFG-grown crystals (which were linked to line-shaped pits). This suggests that the specific dislocation population and their impact on surface morphology may depend on the bulk crystal growth method (VB vs. EFG).
• Reflection vs. Transmission Correlation: Reflection XRT images showed good agreement with transmission results. The analysis of spot contrast and size in reflection mode allowed for the classification of dislocation types, distinguishing between those in arrays (uniform contrast) and isolated dislocations (higher intensity, larger size, and head-tail features).

Significance and Claims
The paper claims that the dislocation behavior in (011)-oriented $\beta$-Ga$_2$O$_3$ substrates is more complex than previously assumed. While the (011) orientation offers advantages such as reduced chlorine incorporation and smooth surfaces, the presence of domain boundaries and specific dislocation arrays (DA type-I) aligned along (001) planes indicates that these defects can propagate from the substrate into the epilayer.

The authors emphasize that optimizing substrates for epitaxial growth requires considering not only the crystallographic orientation but also the specific growth conditions of the bulk crystal. The findings provide insight into defect formation mechanisms relevant to epitaxial growth and device performance, highlighting that the reduction of line-shaped pits in (011) substrates does not necessarily imply the absence of all critical extended defects. The study underscores the necessity of evaluating domain boundaries and dislocation types when comparing (001) and (011) substrates for high-breakdown-voltage device applications.