Temperature Gradient Control with an Air-Pocket Design for Growth of High Quality SiC Crystal

Seung Jun Lee, Su Ho Kim, Jung‐Woo Choi, Jong Hwi Park, Jung Doo Seo, Myung Ok Kyun, Jung Gyu Kim, Kap Ryeol Ku, Yeon Suk Jang, Won Jae LeeDiffusion 2023 · DOI

TL;DR

Researchers have developed a novel "air-pocket" hot-zone design for growing high-quality 6-inch Silicon Carbide (SiC) crystals. This modification successfully creates a uniform temperature gradient during growth, significantly reducing crystal defects and improving overall quality compared to conventional methods. This advancement is crucial for producing better SiC substrates needed for high-performance electronic devices.

What Problem Does This Paper Address?

The demand for high-quality SiC crystals is rapidly increasing, driven by their use in advanced power electronic devices like MOSFETs and SBDs. The performance and yield of these devices are directly linked to the quality of the SiC substrate, specifically its low defect density and minimal warpage. Growing large-diameter SiC crystals, particularly 6-inch wafers, is challenging because maintaining a uniform temperature gradient across the growing crystal is difficult. In conventional methods, this non-uniformity leads to inhomogeneous growth fronts, multi-facet formation, and ultimately, an increased density of crystal defects such as polytype inclusions and dislocations. This paper addresses the critical need for a more controlled and uniform temperature gradient during the SiC crystal growth process to achieve higher quality, larger diameter crystals.

Methods

The core innovation presented is a modified "hot-zone" design for the Physical Vapor Transport (PVT) growth of SiC single crystals. The conventional hot-zone design was modified by incorporating an "air-pocket" behind the SiC seed holder. This air-pocket is designed to influence heat radiation and conduction, thereby controlling the temperature gradient in the horizontal direction.

The growth process itself followed established PVT methods, using a graphite crucible at 2400°C under an argon inert gas atmosphere (20 torr with 5% nitrogen) for approximately 6-inch SiC single crystal growth. The axial thermal gradient was maintained at 15-20°C/cm.

To evaluate the effectiveness of the modified design, the researchers employed several characterization techniques:

  • In-situ Growth Monitoring: The front shape of the SiC crystal ingot was measured at different growth stages (initial, growing, and finished) to assess the uniformity of the growth front.
  • Optical Microscopy and UV Fluorescence (UVF): These techniques were used to visually inspect the grown ingots for uniformity and to detect different polytypes.
  • X-ray Rocking Curve Measurement: This was used to quantify crystal quality by measuring the Full Width at Half Maximum (FWHM) of the rocking curve, which indicates crystal lattice perfection. Mapping measurements were performed across the wafers.
  • Chemical Etching (molten KOH): This method was used to reveal and quantify various types of crystal defects, such as Threading Screw Dislocations (TSD), Threading Edge Dislocations (TED), and Basal Plane Dislocations (BPD), by counting etch pits.

Datasets

Not discussed in the paper. The study focuses on experimental growth and characterization of SiC crystals rather than using pre-existing datasets for training or evaluation.

Key Insights and Ideas

  • The Air-Pocket's Role: The air-pocket acts as a thermal buffer, allowing for fine-tuning of the temperature gradient. By adjusting its shape, the researchers could achieve a more uniform temperature distribution across the seed crystal, especially in the critical initial growth phase.
  • Uniformity is Key: The study strongly emphasizes that a uniform temperature gradient is paramount for achieving high-quality SiC crystals. Inhomogeneous gradients lead to a cascade of problems, including uneven growth fronts, multi-facet formation, and increased defect generation.
  • Visualizing Quality: The difference in crystal quality was visually striking. Conventional growth resulted in ingots with varying colors under UV light, indicating different polytypes. The modified design produced ingots with a uniform color, suggesting a single, consistent polytype.
  • Defect Reduction is Significant: The quantitative data from X-ray rocking curves and etch pit density analysis clearly demonstrates the superiority of the modified design in reducing crystal defects.

Key Results

  • Improved Crystal Shape: The modified hot-zone design resulted in a consistently convex ingot shape throughout the growth process, indicating a stable and uniform temperature gradient. In contrast, the conventional design showed a bumpy initial growth front.
  • Reduced FWHM: The average FWHM value in X-ray rocking curve measurements was significantly reduced from 37.4 arcsec for the conventional design to 19.8 arcsec for the modified design, indicating much better crystal lattice perfection.
  • Lower Defect Densities:
    • TSD (Threading Screw Dislocation): Reduced from 4,500 ea/cm² to 200 ea/cm².
  • TED (Threading Edge Dislocation): Reduced from 19,300 ea/cm² to 400 ea/cm².
  • BPD (Basal Plane Dislocation): Reduced from 50,300 ea/cm² to 4,200 ea/cm².
  • Reduced Peak Shift: The shift degree of the X-ray peak position was reduced from 0.6° to 0.2°, which is considered advantageous for subsequent device fabrication steps.

Limitations and Future Work

  • Limitations: The paper focuses on 6-inch SiC crystal growth. While the principles are likely applicable to other sizes, specific optimization might be needed. The exact mechanism by which the air-pocket influences heat transfer is described qualitatively; a more detailed thermal modeling could provide deeper insights.
  • Future Work: The authors suggest that the improved crystal quality and reduced shift degree of the X-ray peak position from the modified hot-zone design are more advantageous for next-process steps in device fabrication. Further investigation into how these improvements translate directly to enhanced device performance and yield would be a logical next step. Exploring different air-pocket geometries and their impact on temperature gradients could also be a fruitful area for future research.

Summary updated

Abstract

The hot-zone design using an air-pocket was adopted to produce uniform temperature gradient in horizontal direction. In order to investigate the change of temperature gradient toward horizontal direction with growth time, the front shape of SiC growing crystal was measured with different growth stages such as initial, growing and finished stage. While SiC ingot grown in conventional hot-zone design exhibited inhomogeneous growth front in the initial stage of growth and multi facet formation in final stage, which could result in increased defect density, a homogeneous temperature gradient and improved crystal quality was obtained in the modified hot-zone design. Based on the mapping measurement of FWHM (Full width at half maximum) value in X-ray rocking curve, the crystal quality of SiC crystals grown with the modified hot-zone design was observed to be definitely better than conventional design.
BibTeX
@inproceedings{lee2023temperature,
  author = {Seung Jun Lee and Su Ho Kim and Jung‐Woo Choi and Jong Hwi Park and Jung Doo Seo and Myung Ok Kyun and Jung Gyu Kim and Kap Ryeol Ku and Yeon Suk Jang and Won Jae Lee},
  title = {{Temperature Gradient Control with an Air-Pocket Design for Growth of High Quality SiC Crystal}},
  booktitle = {Diffusion},
  year = {2023},
  doi = {10.4028/p-38rj7g},
}
Figure