In-Process CT Scanning for Void and Fiber Misalignment Detection in Thick CFRP Laminates for Semiconductor Robot End-Effectors

Thick carbon fiber reinforced polymer (CFRP) laminates used in semiconductor robot end-effectors must meet stringent quality requirements to ensure precision, stiffness, and cleanliness. Traditional destructive testing and post-process inspection are insufficient for detecting internal voids and fiber misalignment that can compromise performance. In-process CT scanning—integrated directly into the manufacturing cell—offers a non-destructive, volumetric method to detect these defects at the point of production. This guide explains the physics, standards, and practical implementation of in-process CT scanning for thick CFRP laminates, with a worked example and key parameters.

Why Thick CFRP Laminates Demand In-Process CT Scanning

Thick CFRP laminates (typically >6 mm [0.24 in]) are prone to process-induced defects such as porosity, resin-rich pockets, and fiber waviness. In semiconductor robot end-effectors, these defects can cause dimensional instability, reduced stiffness, and particle generation—critical in cleanroom environments. Traditional ultrasonic testing (UT) struggles with thick, multi-layered composites due to signal attenuation and complex geometries. In-process CT scanning overcomes these limitations by providing 3D volumetric data with micron-level resolution.

Key drivers for in-process CT:

• Real-time process control: Detect defects immediately after layup or cure, enabling corrective action before additional value is added.
• Complex geometry: CT handles intricate end-effector shapes with internal channels and varying thickness.
• Quantitative analysis: Measure void volume fraction (VVF) and fiber orientation angle (FOA) per ASTM E1441 and ASTM E1695.
• Cleanroom compatibility: Non-contact, no coupling media required.

Fundamentals of CT Scanning for Void and Fiber Misalignment Detection

Industrial CT uses X-ray attenuation to reconstruct cross-sectional images. For CFRP, the contrast between carbon fibers (density ~1.8 g/cm³) and epoxy resin (density ~1.2 g/cm³) allows void detection down to 10 µm. Fiber misalignment is detected via local variations in X-ray attenuation and edge detection algorithms.

Void detection: Voids appear as low-density regions. The void volume fraction (VVF) is calculated as:

VVF = (Σ V_void / V_total) × 100%

Per MIL-HDBK-17, VVF >2% is typically unacceptable for structural composites. In-process CT can achieve a resolution of 5–10 µm, enabling detection of voids as small as 20 µm in diameter.

Fiber misalignment: Fiber orientation is measured by analyzing the gradient of CT values. Misalignment angle (Δθ) is defined as the deviation from the designed orientation. For end-effectors, Δθ >3° can reduce tensile modulus by up to 15% (based on Toray T700S data: 230 GPa [33.4 Msi] at 0° vs. 195 GPa [28.3 Msi] at 3° misalignment).

Worked Numerical Example: Void Detection in a Thick CFRP End-Effector Link

Consider a 10 mm [0.39 in] thick CFRP link made from Toray T700S/Hexcel 8552, with a target VVF <1%. A CT scan with 8 µm voxel resolution reveals three voids: diameters 30 µm, 45 µm, and 60 µm, all spherical.

Step 1: Calculate void volumes.

V_void = (4/3)πr³

• 30 µm: r=15 µm → V=1.41×10^-14 m³
• 45 µm: r=22.5 µm → V=4.77×10^-14 m³
• 60 µm: r=30 µm → V=1.13×10^-13 m³

Total void volume = 1.75×10^-13 m³.

Step 2: Calculate total volume of the link.

Assume link dimensions: 100 mm × 20 mm × 10 mm = 2.0×10^-5 m³.

Step 3: Compute VVF.

VVF = (1.75×10^-13 / 2.0×10^-5) × 100% = 8.75×10^-7% ≈ 0.000000875%.

This is far below the 1% threshold, so the part is acceptable. However, if voids were larger or more numerous, VVF could exceed limits. For instance, a single 1 mm void would yield VVF = (5.24×10^-10 / 2.0×10^-5) × 100% = 0.0026%, still acceptable, but a cluster of 100 such voids would reach 0.26%.

Step 4: Fiber misalignment check.

CT data shows a local misalignment of 2.5° in a critical load-bearing region. Using the composite lamina transformation equations (ASTM D3039), the longitudinal modulus E₁ = 230 GPa. The transverse modulus E₂ = 15 GPa, and shear modulus G₁₂ = 5 GPa. The off-axis modulus E_x at angle θ is:

1/E_x = cos⁴θ/E₁ + (1/G₁₂ - 2ν₁₂/E₁)sin²θcos²θ + sin⁴θ/E₂

With ν₁₂ = 0.3, at θ=2.5°:

E_x ≈ 228 GPa, a reduction of ~0.9%—acceptable. At θ=5°, E_x ≈ 220 GPa, a 4.3% reduction—potentially unacceptable for stiffness-critical applications.

Key Parameters for In-Process CT Scanning of Thick CFRP

For thick CFRP (>10 mm), a 225 kV microfocus tube with 10 µm voxel is recommended. Scanning time can be reduced by using region-of-interest (ROI) scanning.

Integrating In-Process CT into Production: Best Practices

To maximize ROI, integrate CT scanning after cure but before final machining. This allows early defect detection and rework. Key steps:

1. Automated part handling: Use a robot to load/unload parts from the CT system.
2. Reference standards: Scan a calibrated phantom daily to ensure system accuracy per ASTM E1695.
3. Software automation: Implement AI-based defect recognition to flag VVF >1% or Δθ >3°.
4. Data traceability: Store CT data with part serial numbers for quality records.

At Flex Precision Composites, we use a Zeiss Contura CMM with CT capability to inspect end-effectors up to 500 mm [19.7 in] in length, achieving ±0.05 mm tolerance.

Standards and Specifications

Relevant standards for CT scanning of composites include:

• ASTM E1441: Standard Guide for Computed Tomography (CT) Imaging.
• ASTM E1695: Standard Test Method for Measurement of Computed Tomography (CT) System Performance.
• ASTM D3039: Standard Test Method for Tensile Properties of Polymer Matrix Composite Materials.
• MIL-HDBK-17: Composite Materials Handbook, Vol. 3—provides acceptance criteria for voids.
• ISO 527: Plastics—Determination of tensile properties.

These standards ensure consistent, reproducible CT results and facilitate customer acceptance.