From Microscopic Damage Evolution to Atomic-Scale Removal Mechanisms

Figure 1. Manufacturing process flow of silicon carbide wafers.
With the rapid growth of new energy vehicles, photovoltaic energy storage, rail transportation, and high-frequency communication industries, silicon carbide (SiC) has emerged as one of the most important representatives of third-generation semiconductor materials. Compared with conventional silicon, SiC offers a wider bandgap, higher thermal conductivity, superior high-voltage tolerance, and excellent high-temperature stability, making it highly suitable for high-power and high-frequency electronic devices.
However, these exceptional properties also make SiC extremely difficult to machine. Due to its characteristic combination of high hardness and high brittleness, SiC is highly prone to crack formation, edge chipping, and subsurface damage during grinding, cutting, and polishing processes. These defects severely compromise device reliability and wafer yield.
As a result, achieving high-efficiency, low-damage, and ultra-precision machining of SiC wafers has become a major research focus in advanced manufacturing.
At present, researchers mainly investigate SiC material removal mechanisms and damage evolution behaviors through:
1. Indentation and scratching experiments
2. Practical grinding experiments
3. Molecular dynamics (MD) simulations
4. Stress-field theoretical models
These approaches aim to systematically reveal the fundamental transition mechanism from brittle fracture to ductile material removal.
 

1. Why Is SiC So Difficult to Machine?

SiC is a typical hard and brittle material with a hardness approaching that of diamond and relatively low fracture toughness. This means:

1. Cutting tools struggle to penetrate the material;
2. Plastic deformation is difficult to initiate;
3. Once local stress exceeds the critical limit, cracks propagate rapidly.

During practical machining, SiC wafers commonly exhibit:

• Microcracks
• Edge chipping
• Surface scratches
• Subsurface damage
• Crystal lattice defects
• Residual stress concentration

These issues not only degrade surface quality but may also cause device failure during subsequent operation.
Therefore, the core objective of SiC machining research is to achieve:

Ductile Mode Machining

That is, removing material through plastic deformation rather than brittle fracture.
 

2. Indentation and Scratch Experiments: Revealing the Microscopic Nature of SiC Removal

2.1 Single-Abrasive Scratching as a Fundamental Grinding Model

In actual grinding processes, the grinding wheel surface contains a large number of randomly distributed abrasive grains, each acting as a miniature cutting tool. Therefore, the complex grinding process can be simplified into:
The micro-cutting interaction between a single abrasive grain and the workpiece.
Researchers commonly employ:

• Nanoindentation
• Nanoscratch testing
• Single-grit scratching

to investigate:

• Plastic deformation behavior
• Crack initiation mechanisms
• Dislocation evolution
• Brittle-to-ductile transition behavior

This methodology effectively establishes the relationship between microscopic damage and macroscopic grinding performance.
 

2.2 SiC Is Not Entirely a “Brittle Material”

Figure 2. Anisotropy of friction coefficient and hardness in 3C-SiC.
Conventional understanding assumes that SiC removal is dominated by brittle fracture. However, extensive nanoscratch experiments have demonstrated that under nanoscale and low-load conditions, SiC can undergo localized plastic deformation.
Studies indicate that ductile removal in 3C-SiC primarily arises from:

• Dislocation slip
• Lattice distortion
• Amorphization transformation
• Local crystal reconstruction

Under low loading conditions, continuous pile-up chips are formed within the scratch region, exhibiting clear ductile cutting characteristics.
Meanwhile, strong crystallographic anisotropy has also been observed:

• The <110> orientation exhibits higher hardness;
• Higher friction coefficients are observed;
• Material removal rates become lower.

This indicates that SiC crystals possess pronounced machining anisotropy.
Therefore, crystal orientation selection is also a critical factor influencing machining quality in advanced wafer processing.
 

2.3 Cracks Often Initiate Beneath the Surface

Figure 3. Surface and cross-sectional TEM morphology of 6H-SiC after scratching.
Research has shown that as the applied load gradually increases, the material removal mode of SiC changes significantly.
For example:

• When the load is below approximately 7 mN, material removal is dominated by plastic flow;
• At around 10 mN, fragmented particles begin appearing in the chips;
• Subsurface microcracks rapidly propagate;
• Cracks eventually extend along crystal planes toward the surface.

Importantly, cracks do not first appear on the surface. Instead, they preferentially nucleate within the subsurface layer.
This occurs because:

• Maximum shear stress is typically located beneath the surface;
• Dislocation slip initially induces internal damage;
• Repeated abrasive interactions continuously propagate hidden cracks.

Once these subsurface cracks extend to the surface, they produce:

• Edge chipping
• Crack networks
• Surface spalling

This is one of the fundamental reasons why achieving ultra-smooth SiC surfaces is extremely challenging.
 

2.4 Critical Cutting Depth Governs the Brittle-to-Ductile Transition

Figure 4. Scratch depth and contact pressure during different material removal stages in 4H-SiC.
Researchers introduced a critical concept known as the:

Critical Depth of Cut

Its significance is:
When the cutting depth is below the critical value, material removal occurs through plastic deformation;
Once the cutting depth exceeds the critical threshold, cracks propagate rapidly and brittle fracture dominates.
For 4H-SiC, studies indicate that the critical brittle-to-ductile transition depth is approximately:
h_c \approx 90\ \mathrm{nm}
This implies that ultra-precision machining must strictly control the cutting depth of individual abrasive grains.
Otherwise, even if the surface appears smooth, severe subsurface damage may already exist.
 

3. Grinding Experiments: Closer to Real Machining Conditions

Although nanoscratch experiments reveal microscopic removal mechanisms, their strain rates are far lower than those encountered in actual grinding.
The strain rate in real grinding is typically:
10^3 \sim 10^6
times higher than in scratch experiments.
Therefore, scratch testing alone cannot fully represent practical machining behavior.
 

3.1 Damage in Real Grinding Is More Complex

 

Figure 5. Cross-sectional TEM image showing the subsurface region of 4H-SiC (0001).
TEM studies reveal that the subsurface layer of ground SiC contains:

• Median cracks
• Radial cracks
• Basal-plane dislocations
• Stacking faults
• Ripple-like damaged zones

In particular, microcrack sliding can induce basal-plane dislocation propagation, further aggravating internal damage.
Even if these defects do not immediately cause failure, they may gradually expand during:

• Thermal cycling
• High-voltage operation
• Device packaging processes

3.2 High Material Removal Rate Does Not Necessarily Mean Poor Quality

Some studies have shown that increasing the material removal rate (MRR) does not necessarily deteriorate surface quality.
This suggests that machining efficiency and surface quality are not inherently contradictory.
The key factors influencing machining quality include:

• Abrasive size
• Cutting depth
• Load distribution
• Cooling conditions
• Grinding wheel dressing condition

Only through coordinated multi-parameter optimization can the following be simultaneously achieved:

• High efficiency
• Low damage
• Superior surface quality

 

4. Molecular Dynamics (MD) Simulation: Understanding SiC Removal at the Atomic Scale

4.1 Why Is MD Simulation So Important?

Conventional experiments cannot directly observe:

• Atomic motion
• Dislocation formation
• Amorphization processes
• Phase transformation behavior

MD (Molecular Dynamics) simulation enables researchers to observe material deformation processes in real time at the atomic scale.
Therefore, it has become an essential tool for studying SiC removal mechanisms.
 

4.2 What Causes Plastic Deformation in SiC?

Figure 6. Formation mechanism of prismatic dislocation loops in 3C-SiC.
The prevailing view is that plastic deformation in SiC originates from the coupled effect of:

• Dislocation slip
• Amorphization transformation

MD simulations reveal that:

• Dislocation loops continuously expand;
• Multiple dislocations merge with each other;
• The crystal lattice gradually destabilizes;
• Local amorphous regions eventually form.

Furthermore, high shear stress can induce crystal phase transformation.
For example:
Approximately 9 GPa shear stress induces the formation of 3C-SiC layers;
Around 12 GPa leads to the formation of 3C-SiC grains.
This indicates that SiC undergoes complex crystal structure reconstruction under high-stress conditions.
 

4.3 Crack Formation Is a Progressive Process

Figure 7. Dislocation motion in 3C-SiC during nanometric cutting.
MD studies further show that crack formation is not instantaneous but evolves through:

• Amorphization
• Dislocation nucleation
• Slip band formation
• Local stress concentration
• Crack initiation and propagation

As cutting thickness increases, the material removal mode gradually transitions from ductile removal to brittle fracture, which is highly consistent with experimental observations.
 

5. Stress-Field Models: Predicting Where Cracks Begin

In addition to experiments and MD simulations, stress-field theoretical models are also important tools for investigating SiC removal mechanisms.
Their primary objective is to predict stress distributions and crack initiation regions during machining.
 

5.1 Tensile Stress Dominates Crack Formation

Figure 8. Scratch crack system and experimental validation.
Studies indicate that:

• Tensile stress tends to form on both sides of the abrasive grain;
• Cracks usually initiate and propagate from these regions.

Meanwhile:

• Compressive stress dominates in front of the abrasive grain;
• Plastic deformation is more likely to occur there.

Thus:

• Tensile stress promotes crack propagation;
• Compressive stress promotes plastic deformation.

 

5.2 Abrasive Geometry Influences Damage Modes

Figure 9. Stress distribution under abrasive scratching with different rake angles.
Research shows that:

• Abrasives with larger rake angles generate stronger tensile stress;
• Smaller rake angles are more likely to induce vertical cracks.

This means that abrasive geometry directly affects material damage behavior.
Therefore, grinding wheel dressing and abrasive control are critically important in ultra-precision grinding.
 

6. Major Challenges Remaining in Current Research

Although substantial progress has been achieved in understanding SiC removal mechanisms, many critical issues remain unresolved.
 

6.1 The Phase Transformation Mechanism Remains Controversial

There is still ongoing debate regarding whether SiC necessarily undergoes phase transformation during machining:

• Does amorphization always occur?
• Is graphitization formed?
• Does transformation into 3C-SiC occur?
• How do different polytypes transform into one another?

These questions require further investigation.
 

6.2 Crack Propagation Mechanisms Are Still Unclear

Existing models can predict: Where cracks initiate;
but cannot fully explain:

• How cracks propagate;
• How instability develops;
• How cracks penetrate subsurface layers.

This remains a major challenge in current SiC machining theory.
 

6.3 Lack of Unified Multiscale Models

Most current studies focus on:

• Nanoscale scratching;
• Atomic-scale MD simulation.

However, actual machining involves:

• Multi-abrasive interactions
• Thermo-mechanical coupling
• High-speed dynamic impact
• Macroscopic machine vibration

Therefore, a major future direction is the establishment of a unified multiscale machining model spanning:
Atomic scale → Microscopic scale → Macroscopic scale
 

7. Conclusion: SiC Ultra-Precision Machining Is Advancing Toward Controlled Ductile Removal

Overall, research on SiC wafer material removal mechanisms has gradually evolved from traditional empirical processing toward:

• Atomic-scale mechanistic understanding
• Multiscale damage analysis
• Intelligent process parameter optimization
• Controlled ductile-regime machining

An increasing number of studies demonstrate that SiC is not an entirely non-ductile material.
Through proper control of:

• Cutting depth
• Abrasive size
• Crystal orientation
• Stress distribution
• Machining parameters

it is entirely possible to achieve:
High-efficiency, low-damage, ultra-smooth surface machining.
In the future, with advances in in-situ characterization technologies, AI-driven optimization algorithms, and high-precision MD simulations, SiC wafer machining is expected to further overcome the limitations associated with hard and brittle materials, providing more reliable technological support for advanced power semiconductor manufacturing.

As the demand for high-performance SiC devices continues to grow, advanced wafer processing technologies are becoming increasingly critical to the semiconductor industry. From understanding microscopic damage evolution to achieving atomic-scale material removal, every breakthrough contributes to higher wafer quality, lower defect density, and improved device reliability.
JXT is committed to providing high-quality 2–8 inch silicon carbide (SiC) substrates for research, development, and industrial applications. With stable crystal quality and reliable supply capability, we support global customers in accelerating innovation in power electronics and next-generation semiconductor technologies.

 

Related products:

3inch 4H-N SiC Wafer

8inch 4H-N SiC Wafer

D159mm SiC Wafer