Rapid non-destructive inspection of sub-surface defects in 3D printed alumina through 30 layers with 7 {\mu}m depth resolution

TL;DR

This paper introduces a rapid, non-destructive imaging technique using MIR OCT to inspect sub-surface defects in 3D printed alumina ceramics with high resolution, enabling defect detection throughout manufacturing stages.

Contribution

It demonstrates the use of MIR OCT for high-resolution, deep-penetration, non-destructive inspection of 3D printed ceramics, a significant improvement over existing methods.

Findings

Able to detect individual layers and defects in alumina parts

Effective defect tracking through all manufacturing stages

Potential for automated defect classification with AI

Abstract

The use of additive manufacturing (AM) processes for industrial fabrication has grown rapidly over the last ten years. The most well-known AM technologies are fused deposition modelling and stereolithography techniques. One particular industry where 3D printing is advantageous over traditional fabrication techniques is within ceramic components due to its flexibility. To establish a new and improved level of print quality and reduce resource consumption in the 3D printing ceramics industry, there is a need for fast integrated, sub-surface and non-destructive inspection (NDI) with high resolution. Several techniques have already been developed for high-resolution NDI, such as X-ray computed tomography (XCT), but none of them are both fast, integrable, and non-destructive while allowing deep penetration with high resolution. In this study, we demonstrate sub-surface monitoring of 3D printed alumina parts to a depth of $\sim$0.7 mm in images of 400$\times$2048 pixels with a lateral resolution of 30$~\mu$m and depth (or axial) resolution of 7$~\mu$m . The results were achieved using mid-infrared optical coherence tomography (MIR OCT) based on a MIR supercontinuum laser with a 4$~\mu$m center wavelength. We find that it is possible to detect individual printed ceramic layers and track predefined defects through all four processing steps: green, preconditioned, debinded, and sintered. Our results also demonstrate how a defect in the green phase could affect the final product. Based on the understanding of how defects develop in maturing printed parts, we pave the way for NDI integration in AM, which can be combined with artificial intelligence and machine learning algorithms for automatic defect classification in volume production of a new standard of high quality ceramic components.