High-throughput, Non-Destructive, Three-Dimensional Imaging of GaN Threading Dislocations with in-Plane Burgers Vector Component via Phase-Contrast Microscopy

This paper demonstrates a high-throughput, non-destructive phase-contrast microscopy (PCM) technique capable of three-dimensionally imaging threading dislocations and other defects in GaN by analyzing contrast shapes and focal plane shifts.

The Gist

The "X-Ray Vision" for Microscopic Flaws: Making Perfect Semiconductors

Imagine you are a master baker trying to make the world’s most perfect, delicate soufflé. To make it rise perfectly, the structure of the air bubbles inside must be flawless. If there is even a tiny, microscopic crack in the structure, the whole thing collapses.

In the world of high-tech electronics, engineers are trying to bake "digital soufflés" called GaN (Gallium Nitride) chips. These chips are the superstars of the future—they power everything from fast-charging your phone to the next generation of electric vehicles. But there is a problem: these chips often have tiny "cracks" called dislocations. These are microscopic imperfections in the crystal structure that act like leaks in a pipe, causing the electronics to fail or waste energy.

The Problem: The "Invisible Leak"

Finding these tiny cracks is incredibly hard.

• The old way (The Microscope approach): It’s like trying to find a single hair in a dark room using a tiny flashlight. It takes forever, and you can only see one tiny spot at a time.
• The "Destructive" way: It’s like having to smash the soufflé just to see if there’s a bubble in the middle. Once you check it, the chip is ruined.

The Solution: Phase-Contrast Microscopy (The "Shadow Puppet" Method)

The researchers in this paper have introduced a new way to "see" these flaws without breaking anything, using a technique called Phase-Contrast Microscopy (PCM).

Think of it like Shadow Puppets. If you hold your hand in front of a light, you don't need to touch your hand to know its shape; you just look at the shadow on the wall.

The researchers discovered that these microscopic cracks in the crystal create tiny "shadows" (distortions in light) that they can see clearly.

• Dots vs. Lines: If the crack is standing straight up (like a flagpole), it looks like a tiny dot. If the crack is leaning over (like a fallen tree), it looks like a line.
• 3D Vision: By shifting the focus of the lens, it’s like moving a flashlight up and down through a glass sculpture. They can actually trace the path of the crack from the top of the chip all the way to the bottom, seeing exactly how it "grows" through the material.

Why is this a big deal?

1. It’s Fast: Instead of scanning one tiny dot at a time (which is like reading a book through a straw), this method lets them see a wide area all at once. It’s like reading the whole page in one glance.
2. It’s Non-Destructive: They can inspect the chip and leave it perfectly intact, ready to be used in your next smartphone.
3. It’s a "Multi-Tool": Not only can they see the "cracks" (dislocations), but they can also spot scratches, tiny air bubbles (voids), and other "bruises" on the chip's surface.

The Bottom Line

This paper describes a new set of "super-vision" goggles for scientists. By using light to cast "shadows" of microscopic flaws, they can now inspect high-tech materials quickly and thoroughly, ensuring that the chips powering our future are strong, efficient, and leak-free.

Technical Summary

Technical Summary: High-throughput, Non-Destructive, 3D Imaging of GaN Threading Dislocations via Phase-Contrast Microscopy

1. Problem Statement

Gallium Nitride (GaN) is a critical wide-bandgap semiconductor for blue LEDs and power electronics. However, its performance is often degraded by threading dislocations (TDs), which act as leakage paths in devices. Existing non-destructive characterization methods face significant trade-offs:

• Cathodoluminescence (CL): Requires vacuum conditions and is limited by small observable areas.
• Multiphoton Excitation Photoluminescence (MPPL): Provides 3D capabilities but is extremely slow for wafer-scale inspection due to the need for laser scanning.
• X-ray Topography (XRT) & Polarization Microscopy (PM): Often lack the resolution required for high-density GaN wafers (exceeding $10^4 \text{ cm}^{-2}$).

There is a critical need for a high-resolution, high-throughput, and non-destructive method capable of wafer-scale 3D defect characterization.

2. Methodology

The researchers employed Phase-Contrast Microscopy (PCM) to inspect (0001) GaN chips grown via HVPE. The experimental setup and validation included:

• PCM Configuration: Utilized a 405 nm LED illumination source, a 20× objective lens (NA = 0.5), and a condenser lens (NA = 0.78) with a phase plate.
• 3D Imaging Technique: The focal plane was shifted from the top surface to the back surface in discrete steps (e.g., 12 $\mu$m) to visualize the propagation paths of defects.
• Validation via MPPL: To confirm that PCM was detecting actual dislocations, the researchers compared PCM images with Multiphoton Excitation Photoluminescence (MPPL) images. A one-to-one correspondence between PCM contrasts and MPPL dark spots was established.
• Structural Verification: Etch pits (KOH+NaOH) and Transmission Electron Microscopy (TEM) were used to definitively identify dislocation types (TEDs, TSDs, and TMDs).

3. Key Contributions

• Identification of Detection Limits: The study established that PCM detects dislocations through shear stress in the XY plane. This means PCM can detect Threading Edge Dislocations (TEDs) and Threading Mixed Dislocations (TMDs), but it cannot detect Threading Screw Dislocations (TSDs) because they lack the necessary in-plane Burgers vector component.
• Contrast Interpretation Model: The researchers proved that PCM contrast shapes correlate to dislocation inclination:
  • Dot contrasts $\rightarrow$ Vertical dislocations (perpendicular to the surface).
  • Line contrasts $\rightarrow$ Inclined dislocations.
• 3D Projection Capability: Demonstrated that a single PCM image represents a projection of all threading dislocations within a thickness of approximately 43 $\mu$m.

4. Results

• Resolution: PCM successfully resolved individual dislocations spaced as close as 1.3 $\mu$m apart.
• 3D Visualization: By shifting the focal plane, the researchers successfully mapped the 3D propagation paths of dislocations from the top surface to the back surface of the GaN chip.
• Multi-defect Detection: Beyond dislocations, PCM proved effective at detecting:
  • Scratches and subsurface scratches.
  • Facet boundaries.
  • Plate-like voids (sub-micrometer thickness).
• Efficiency: The PCM imaging was highly efficient, with an exposure time as short as 3 ms per image, enabling high-throughput inspection.

5. Significance

This research establishes Phase-Contrast Microscopy as a versatile, laboratory-accessible, and high-speed tool for the semiconductor industry. By providing a way to perform non-destructive, three-dimensional inspection of both dislocations and other structural defects (like voids and scratches) at a high resolution, PCM offers a practical solution for quality control in the production of wide-bandgap semiconductor wafers. The ability to combine PCM with MPPL also provides a strategic way to map the spatial distribution of TSDs, which are otherwise difficult to characterize.