Reviews on Machine Learning Approaches for Process Optimization in Noncontact Direct Ink Writing

Area of Science:

• Materials Science
• Additive Manufacturing
• Machine Learning

Background:

• Noncontact direct ink writing (NDIW) is an emerging 3D printing technology for low-cost, customized microelectronic devices.
• Current NDIW methods, including inkjet printing (IJP) and aerosol jet printing (AJP), face limitations in printing quality, impacting electrical performance.
• Machine learning (ML) offers novel process modeling and optimization but often focuses on specific aspects of NDIW.

Purpose of the Study:

• To systematically review NDIW printing principles, influencing factors, and limitations.
• To compare traditional and ML-based optimization strategies for IJP and AJP.
• To identify challenges and future directions for a comprehensive ML approach to NDIW process optimization.

Main Methods:

• Systematic review of noncontact direct ink writing principles and technologies (IJP and AJP).
• Classification of process optimization requirements into four main aspects.
• Comparative analysis of traditional and state-of-the-art machine learning strategies for NDIW process optimization.

Main Results:

• NDIW technologies like IJP and AJP have inherent limitations affecting microelectronic performance.
• Existing ML strategies show potential but lack a holistic approach for NDIW process optimization.
• A comprehensive, systematic ML approach is required to address the multifaceted challenges in NDIW.

Conclusions:

• Significant advancements in NDIW are hindered by printing quality limitations.
• A unified ML framework is essential for optimizing diverse aspects of NDIW processes.
• Future research should focus on developing integrated ML solutions for enhanced NDIW performance and broader applications.