Microfluidics technology, which manipulates fluids and various microspecies within micro-scaled channels[1, 2], boasts numerous appealing characteristics such as minimal sample consumption, ease of integration, high sensitivity, rapid response time, and low cost[3]. These advantages make microfluidics highly suitable for a broad range of applications, including biomedicine[4], drug screening[5], environmental detection[6], chemical synthesis[7], food security[8], etc. The quality of microfluidic chip is crucial for fulfilling the intended functionalities of these applications. Generally, researchers rely on photolithography technology to fabricate microfluidic master molds. This method, known for its precision and reliability, involves transferring geometric patterns onto a substrate, a critical step in producing the intricate channel networks essential for effective microfluidic devices. Soft photolithography, known for its submicron resolution, is highly effective for fabricating polydimethylsiloxane (PDMS) chips from silicon wafers, ideal for advanced microfluidic applications[9]. However, despite its prevalence in microfluidic laboratories, photolithography has notable limitations. It necessitates a cleanroom environment, involves complex and costly operational processes, and is typically limited to fabricating microstructures with rectangular cross-sections (planar, 2D structures). Producing microstructures with arbitrary cross-sections or varied heights (non-planar, 3D structures) requires multi-step processes involving the manual alignment of multiple masks. This not only complicates the fabrication process but also demands significant labor, especially for multiple exposures, highlighting the need for more streamlined and flexible fabrication methods.
Some candidate methods have been reported to fabricate microfluidic molds instead of photolithography such as CNC micromachining[10], ultrasonic etching[11], laser ablation[12], 3D printing[13], etc. While CNC micromachining offers effectiveness in fabricating microstructures with complex cross-sections, it suffers from limitations such as low surface roughness and precision. Additionally, high-resolution CNC machines tend to be expensive and bulky, making them impractical for laboratory settings. Ultrasonic etching and laser ablation, while viable alternatives, also encounter issues such as high cost and coarse features, limiting their applicability in microfluidics. These challenges underscore the need for more accessible and refined fabrication techniques to meet the demands of microfluidic research and development.
In recent years, 3D printing has been proposed increasingly for microfluidics[14], which mainly refers to hot-melt extrusion-based 3D printing[15], jetting 3D printing[16], and stereolithographic 3D printing[17]. Among the above 3D printing technologies, stereolithographic 3D printing has the appealing characteristics of good printing accuracy (as low as 10 µm) and surface quality (as low as 0.38–0.61 µm)[18], and thus it is more commonly utilized for microfluidic studies such as microfluidic valves with deformable membranes[19], microfluidic pressure sensors[20], miniature droplet generators[21], torque-driven micropumps[22], etc. As a promising candidate for low-cost and convenient fabrication, stereolithographic 3D printing mainly prints curable resin materials directly on a substrate. The printing process is automated, and almost any 3D structures with arbitrary cross-sectional shapes can be produced, thus some researchers use 3D printers to fabricate entire microfluidic devices directly[23]. Entire microfluidic devices are indeed attractive to researchers, however, evacuation of uncured resin from microchannels and cleaning of sacrificial support materials remains a challenge, which may result in defects in the form of microchannel distortion or clogging. In contrast to 3D printing of entire devices, 3D printing of master molds is an indirect way to fabricate microfluidic chips. As microfluidic master molds can be printed with good resolution and surface quality, high-accuracy PDMS chips can be also obtained through mold replication.
Among the stereolithographic 3D printing technologies, digital light processing (DLP) 3D printing is widely used in microfluidics due to its rapid and high-resolution capabilities[24]. DLP 3D printers have pixel sizes ranging from 2–40 µm, with the smallest reported microchannel measuring 18 × 20 µm[25, 26]. However, the high cost of research-grade DLP 3D printers, typically around 15–30 K USD, creates a significant entry barrier for many microfluidic laboratories. In contrast, liquid crystal display (LCD) 3D printers are a more affordable option for entry-level microfluidic applications. LCD 3D printers are priced between 150–600 USD and offer high pixel resolutions of 4K, 8K, and even 12K, with pixel sizes ranging from 18–50 µm[27]. They can achieve feature resolutions down to 35 µm horizontally and 10 µm vertically by adjusting printing parameters such as layer thickness and curing time. Despite these advantages, LCD 3D printing faces challenges such as overcuring of the light-sensitive resin due to light penetration, leading to dimensional inaccuracies in the printed molds. In addition, while many works have shown success in leveraging LCD 3D printing for the fabrication of microchannels with rectangular cross-sections[28], its potential for fabricating microchannels with complex cross-sections remains rarely explored.
In this work, we proposed a method for high-accuracy fabrication of microchannels with complex cross-sections by using a low-cost desktop LCD 3D printer. The optimal layer thickness was studied and employed to reduce the overcure effect of printed mold. To address the issue of uncured PDMS, commonly caused by resin mold poisoning, we introduced a simple post-treatment process to enhance PDMS curing. Following this, PDMS sheets were successfully replicated from the mold, and the microchannel dimensions were measured to analyze the printing accuracy. We also measured the bonding strengths of the PDMS chip in two directions by using two types of tensile testing device. To reduce the fabrication errors of microchannel with triangular cross-sections, we implemented a dimension compensation method to correct nominal dimensions during 3D printing. Finally, we demonstrated the application of the LCD 3D printing for developing an inertial microfluidic chip with a spiral microchannel of trapezoidal-like cross-section, which was effectively employed for separating particles of different diameters.