Atomic-Scale Insights into Slip Deformation and Nanometric Machinability of 6H‑SiC

Introduction

Silicon carbide (SiC) is a cornerstone of third‑generation semiconductor technology, prized for its wide bandgap, high breakdown field, superior radiation tolerance, high carrier‑velocity saturation, exceptional thermal conductivity, low dielectric constant, and robust chemical stability. These attributes make SiC indispensable in high‑temperature, high‑frequency, high‑power, anti‑radiation, and short‑wavelength optoelectronic applications, including Schottky rectifiers, thyratrons, and MOSFETs.

Among SiC polytypes, 3C, 4H, and 6H are most widely used. Conventional machining—grinding, lapping, and polishing—remains the primary fabrication route for single‑crystal SiC. However, the hardness ratio between diamond and SiC (~2:1 for processing depths <50 nm) falls short of the recommended 5:1, leading to accelerated tool wear and subsurface damage that compromise wafer quality. Extensive research on 3C‑SiC has elucidated plastic deformation mechanisms, tool wear, friction behavior, anisotropy, and temperature effects during cutting. In contrast, 6H‑SiC, with its complex ABCACB stacking and a bandgap of 3 eV, has received comparatively less attention, despite its superior technological maturity and industrial prevalence.

To improve the surface and subsurface integrity of 6H‑SiC, it is crucial to identify the optimal crystal plane (machining surface) and orientation (machining direction) that minimize damage while maximizing removal efficiency. Scratch experiments and simulations are standard tools for probing material removal at the nanoscale, offering insights that guide abrasive machining. Here, we employ molecular dynamics scratch simulations to dissect the influence of crystallographic anisotropy on 6H‑SiC’s nanometric machinability.

Methodology

All cutting simulations were performed with the Large‑Scale Atomic/Molecular Massively Parallel Simulator (LAMMPS). Model visualization and defect identification utilized OIVTO and a diamond‑structure detection algorithm. The workpiece and tool were modeled as fully deformable bodies, allowing us to capture wear evolution during the scratch. Each model comprised a boundary layer, a thermostatic layer, and a Newtonian region. Atoms in the boundary layer at the bottom and right of the workpiece were fixed; the remaining atoms obeyed Newton’s second law. Periodic boundaries were applied along the y direction. Prior to scratching, a 50 ps NVE relaxation with a Berendsen thermostat ensured a steady energy state.

The abrasive tip was a spherical crown triangular pyramid with a 90° edge‑to‑edge angle. We selected the three most common 6H‑SiC planes—basal (a‑plane), prismatic (m‑plane), and c‑plane—as machining surfaces. Considering crystal symmetry, we examined six surface/orientation combinations: (0001)/[2‑1‑1‑0], (0001)/[10‑1‑0], (01‑1‑0)/[2‑1‑1‑0], (01‑1‑0)/[0001], (11‑2‑0)/[1‑1‑00], and (11‑2‑0)/[0001]. Simulation parameters are listed in Table 1. Before the scratch, the abrasive was positioned 50 Å below the top surface, 30 Å from the workpiece, well outside the cutoff radius of the interatomic potential. The abrasive moved along +x to complete the scratch.

a MD model of nanoscratching simulation. b The morphology of the tool. c Structure of model. d Axis direction

Schematic diagram of scratching process, where a–f correspond to (0001)/[2‑1‑1‑0], (0001)/[10‑1‑0], (01‑1‑0)/[2‑1‑1‑0], (01‑1‑0)/[0001], (11‑2‑0)/[1‑1‑00], and (11‑2‑0)/[0001] respectively

The analytical bond‑order potential (ABOP) developed by Erhart and Albe was chosen to describe Si–Si, C–C, and Si–C interactions due to its proven accuracy for SiC mechanical properties and removal mechanisms. Parameters are given in Table 2.

Results and Discussion

Nanometric Machinability Analysis

Figures 3 and 4 demonstrate that 6H‑SiC’s anisotropy strongly affects nanometric machinability—machined depth, removal mode, material removal, and subsurface damage depth (SSD). By relating machining surface and direction to the basal plane/c‑axis, we classify processing modes into three groups: (i) basal plane as machining surface; (ii) basal plane perpendicular to the surface with the c‑axis perpendicular to the machining direction; and (iii) c‑axis parallel to the machining direction.

Topography of machined surface under different crystal plane/orientation conditions, where a–f correspond to (0001)/[2‑1‑1‑0], (0001)/[10‑1‑0], (01‑1‑0)/[2‑1‑1‑0], (01‑1‑0)/[0001], (11‑2‑0)/[1‑1‑00], and (11‑2‑0)/[0001]

Nanometric machinability under different crystal plane/orientation conditions. a shows the influence of anisotropy on machined depth and damage depth; b illustrates removal amount, wear, and grinding ratio. Numbers 1–6 correspond to the process methods listed above. The damage layer depth represents the maximum depth of internal crystal defects caused by scratching. The theoretical depth is the preset removal depth, while the machined depth is the residual depth after scratching. Removal amount is the number of atoms removed from the workpiece, and wear amount is the difference in abrasive atom count before and after scratching.

Key observations:

1. Modes (0001)/[2‑1‑1‑0] and (0001)/[10‑1‑0] produced similar machinability, but the latter favored a brittle removal mode.
2. Modes (01‑1‑0)/[2‑1‑1‑0] and (11‑2‑0)/[1‑1‑00] achieved higher machined depth and removal rates. For a theoretical depth of 5 nm, (11‑2‑0)/[1‑1‑00] removed 3.4 × more material than (01‑1‑0)/[0001], with a material‑removal‑to‑wear ratio of 10.1, though SSD was 2.3 × greater than (0001)/[10‑1‑0]. This mode is suited for high‑throughput nanogroove processing where subsurface damage is acceptable. In contrast, (01‑1‑0)/[0001] offers a comparable removal rate but only 50 % of the SSD, yielding a smoother, more mechanically robust groove—ideal for precision machining.
3. When the machining direction aligns with the c‑axis, the abrasive tip wears heavily early on, leading to lower removal rates (≈1.0) and higher SSD. Although these modes exhibit excellent wear resistance, they are not recommended for micro‑nano groove fabrication but could be valuable for designing cutting tool rake faces.

Analysis of Slip Motion and Subsurface Damage Distribution

Schmid Factor Distribution in the Scratching Process

Hexagonal SiC features basal, prismatic, and pyramidal slip systems. Slip propensity depends on generalized stacking‑fault energy and the Schmid factor. Basal slip typically occurs on the <11‑20> shuffle set, while prismatic slip activates along and directions. Figures 5 and 6 illustrate the relevant slip systems and Schmid factors for each loading direction.

Slip systems of the hexagonal crystal

The resolved shear stress τ_ss on a glide plane is τ_ss=σ_cont·m, where σ_cont is the contact stress and m is the Schmid factor. The global coordinate system is fixed while the local system follows the crystal orientation.

During scratching, the rake face dominates contact until severe tip wear shifts the primary contact to a composite of rake and tip top, altering the loading direction. The loading directions are f₁=(1,0,−√2) for the rake face and f₂=(0,0,−1) for the tip top.

Using the crystallographic parameters (c/a≈4.901), basal slip requires lower critical shear stress than prismatic slip. Thus, when machining the basal plane (0001)/[2‑1‑1‑0] or (0001)/[10‑1‑0], basal slip initiates first. For orientations (01‑1‑0)/[2‑1‑1‑0] and (11‑2‑0)/[1‑1‑00], basal slip is suppressed, and prismatic slip dominates. In c‑axis‑parallel modes, severe tip wear elevates the influence of f₂, leading to symmetric prismatic slip across the YZ plane.

Surface/Subsurface Damage Distribution

Figures 6a and 6b show that, for (0001)/[2‑1‑1‑0] and (0001)/[01‑1‑0], slip primarily occurs on (0001)/<11‑20>, producing nanocrystalline grains and lattice distortions that extend to the subsurface. The amorphous phase covers the entire machined surface, with the dislocation depth matching the lattice distortion layer.

The cross‑section of machined area: D dislocation, A amorphous phase, SCF single‑crystal form, O other defects. Labels a–f correspond to the six process modes.

For (01‑1‑0)/[2‑1‑1‑0], despite a Schmid factor of zero for basal slip, the 5.3° misalignment between loading and <11‑20> slip direction activates basal slip on both sides of the V‑groove before prismatic slip. In (11‑2‑0)/[1‑1‑00], larger misalignments (24.7° and 35.3°) suppress basal slip, and prismatic slip on <1‑2‑10> dominates, resulting in deeper subsurface damage and a pronounced sp³→sp² transition.

When the machining direction aligns with the c‑axis, the tip’s severe wear drives prismatic slip during stable scratching. In (01‑1‑0)/[0001], symmetric cross‑slip on <1‑2‑10> and <−1‑1‑20> (60° apart) pinches slip, limiting SSD to t·tanθ/2·cotα/2. In (11‑2‑0)/[0001], irregular tip wear reduces slip to a single system.

Overall, subsurface damage consists mainly of dislocations, lattice distortion (torsion/sliding), and amorphous phase. Slip deformation, non‑crystallization, and irregular lattice distortion govern the mechanical response of 6H‑SiC during nanomachining.

Processed surface/subsurface damage distribution. a xy cross‑section. b xz cross‑section

Concluding Remarks

We have delineated the deformation mechanisms and nanometric machinability of 6H‑SiC across various crystal surface/direction combinations. Our principal findings are:

1. The nanoscratch response of 6H‑SiC is governed by a combination of amorphous phase transition, lattice distortion, and dislocation slip. The depth of the dislocation line determines subsurface damage.
2. Basal and prismatic slip dominate the slip deformation during scratching. The Schmid factor provides a reliable predictor of slip activation.
3. Processing mode (01‑1‑0)/[2‑1‑1‑0] delivers high removal rates with minimal abrasive wear, making it ideal for surface machining. Conversely, the basal plane and c‑axis represent the most challenging faces and directions, offering guidance for cutting‑tool design.

Availability of Data and Materials

All data generated or analyzed during this study are included in the article.

Abbreviations

A:

Amorphous phase

ABOP:

Analytical bond order potential

D:

Dislocation

d:

Undeformed chip thickness

GSF:

Generalized stacking fault energy

LAMMPS:

Large‑scale atomic/molecular massively parallel simulator

MD:

Molecular dynamics

MOSFET:

Metal‑oxide‑semiconductor field‑effect transistor

NVE:

Number, volume, and energy

O:

Other type of defect

SCF:

Single‑crystal form

SPDT:

Single‑point diamond turning

SSD:

Subsurface damage depth

SSD_max:

Maximum subsurface damage depth