High-precision digital light processing (DLP) printing of microstructures for microfluidics applications based on a machine learning approach

Xinhui Wang,Jinghang Liu,Ruihai Dong,Michael D. Gilchrist,Nan Zhang,Xinhui WangJinghang LiuRuihai DongMichael D. GilchristNan Zhanga Centre of Micro/Nano Manufacturing Technology (MNMT-Dublin),School of Mechanical & Materials Engineering,University College Dublin,Dublin 4,Irelandb School of Mechanical Engineering,Technological University Dublin,Dublin 1,Irelandc School of Computer Science,University College Dublin,Dublin 4,Ireland
DOI: https://doi.org/10.1080/17452759.2024.2318774
IF: 10.6
2024-02-24
Virtual and Physical Prototyping
Abstract:Digital light processing (DLP) is renowned for its precision, but the challenge lies in the identification of optimal print parameters to minimise print errors and enhance overall print accuracy. This study introduces a groundbreaking approach by integrating 'Random Forest' (RF) models with DLP printing to construct a predictive model for printing errors to achieve unparalleled precision. We conducted experiments using common commercial resins to print, resulting in a comprehensive dataset of 690 experimental datasets. Consequently, we validated this approach by fabricating Y-type microfluidic structures with a minimum feature size printing of 2 μm and herringbone mixer structures with feature sizes of 21.2 μm. We successfully controlled the average print error below 2.3 μm. The 'RF' model was also extended from the printing of microfluidics (2.5D) to complex spatial three-dimensional (3D) Triple-Periodic Minimal Surface (TPMS) structures. We have successfully achieved the printing of TPMS-Gyroid and TPMS-Diamond structures with a minimum feature size of 20 μm, while the printing errors were all kept within 1.5 μm. This study underscores the generalizability and immense potential of integrating machine learning techniques with high-precision DLP printing, offering valuable insights for future research in the realm of high-precision and expeditious 3D printing.
engineering, manufacturing,materials science, multidisciplinary
What problem does this paper attempt to address?
### Problems the Paper Aims to Solve This paper aims to address the challenges faced in high-precision digital light processing (DLP) printing of microstructures, particularly how to optimize printing parameters to minimize printing errors and improve overall printing accuracy. Specifically, the researchers introduced a Random Forest (RF) machine learning model to build a model that predicts printing errors, achieving unprecedented printing accuracy. ### Main Background Issues 1. **Importance of Microfluidic Technology**: - Microfluidic technology has garnered widespread attention since the 1990s. It precisely manipulates fluids and biological samples through tiny channels and chambers, enabling experiments on a microscopic scale, increasing reaction rates, and overall efficiency. - The initial manufacturing of microfluidic devices relied on expensive and time-consuming traditional methods such as photolithography, etching, and deposition. 2. **Advantages and Challenges of DLP Printing**: - DLP printing offers advantages such as rapid printing, high precision, and excellent surface finish in the production of microfluidic molds. - However, achieving the printing of small feature sizes while ensuring printing accuracy remains a challenge, especially since the minimum feature size of current microfluidic molds is still greater than 20 micrometers. 3. **Shortcomings of Existing Research**: - Current research mostly focuses on the minimum printable feature size, neglecting a comprehensive assessment of the final printing error. - Traditional DLP printing methods typically rely on trial and error to determine suitable printing parameters, significantly limiting the efficiency of parameter optimization. ### Solution 1. **Application of Random Forest Model**: - By integrating the Random Forest model with DLP printing, the researchers constructed a model to predict printing errors, achieving high-precision printing. - The experiment collected 690 sets of experimental data using common commercial resins, validating the effectiveness of this method. 2. **Experimental Results**: - Successfully printed Y-shaped microfluidic structures with a minimum feature size of 2 micrometers and fishbone mixer structures with a feature size of 21.2 micrometers, with an average printing error controlled within 2.3 micrometers. - Extended the model to complex three-dimensional (3D) triply periodic minimal surface (TPMS) structures, successfully printing TPMS-Gyroid and TPMS-Diamond structures with a minimum feature size of 20 micrometers, maintaining printing errors within 1.5 micrometers. ### Conclusion This study demonstrates the significant potential of combining machine learning technology with high-precision DLP printing, providing valuable insights for future research in high-precision and rapid 3D printing. This approach not only improves printing accuracy but also significantly reduces the time for parameter optimization, providing strong support for the development of microfluidic chip devices and the advancement of printing technology.