Parameter Optimization for High-Resolution Microfluidic Channel Fabrication Using a Commercial Low-Cost MSLA Printer

Abstract

Vat polymerization-based 3D printing has emerged as a promising approach for the rapid, low-cost, and scalable fabrication of microfluidic devices; however, achieving high-resolution and fully clog-free microchannels using commercial resins remains challenging. In this study, we systematically investigate key printing parameters—including channel orientation, length, layer thickness, and exposure time—to elucidate their effects on channel openness, dimensional fidelity, and surface morphology using a commercially available low-cost masked stereolithography (MSLA) printer and printing resin, thereby establishing quantitative fabrication boundaries that define the transition from fully open to blocked microchannels in practice. Under optimized printing conditions, microchannels with characteristic dimensions exceeding 200 µm were fabricated in a reliable and clog-free manner using standard commercial resins. In addition, by implementing a size-compensated design strategy, we achieved the fabrication of complex droplet generator arrays with a minimum central channel width of 400 µm, while maintaining an internal dimensional deviation below 2.5%. These investigations significantly expand the practical applicability of low-cost MSLA 3D printing for microfluidic device fabrication, providing a scalable and accessible pathway for producing high-fidelity microchannels without reliance on custom resins or post-processing-intensive workflows.

1. Introduction

Microfluidic systems, originally developed approximately 30 years ago with the goal of miniaturizing analytical technologies, have since become a core technology in a wide range of fields, including the fundamental studies of transport phenomena [1], diagnostic devices [2], and chemical analysis [3]. These systems enable experiments using only small volumes of samples and reagents, while offering high-throughput processing and ease of batch system implementation. As a result, they allow for precise and efficient analysis and control within complex miniaturized platforms [3,4]. Despite these advantages, most existing research platforms remain largely confined to two-dimensional design and fabrication approaches [5]. A representative example is soft lithography-based fabrication using polydimethylsiloxane (PDMS) [6,7,8,9]. This method typically involves microcontact printing, in which single-layer patterns are transferred onto substrates using stamps, or mold-based replication techniques [10,11]. Owing to its simple operation, excellent sealing performance, and broad material compatibility, soft lithography has made significant contributions to the development of various microfluidic platforms [12]. However, emerging microfluidic applications, such as massively parallel droplet generators [13], multilayer reaction networks, and organ-on-chip systems that require vertical perfusion and tissue-scale integration [14], are increasingly incompatible with strictly planar architectures, motivating the development of truly three-dimensional microfluidic platforms.

To fundamentally address these challenges, microfluidic chip fabrication using 3D manufacturing technologies has been actively explored in recent years [15]. For example, fused deposition modeling (FDM) employs extrusion of thermoplastic filaments through a nozzle, offering relatively low cost and a wide range of material options. However, its resolution is traditionally limited to several hundred micrometers, and issues such as surface roughness, difficulty in removing support materials, and insufficient chemical resistance restrict its suitability for microfluidic chips that require precise fluid control [16,17,18].

In contrast, Vat polymerization-based 3D printing enables the fabrication of microfluidic structures with tens of micrometers resolution and true three-dimensional channel architectures. In particular, MSLA offers a unique cost–effectiveness by replacing expensive laser scanners or projection optics with a simple LCD-based photomask, enabling full-field, pixel-level exposure while dramatically reducing system complexity and hardware cost. This architecture allows large build areas, high in-plane resolution, and high throughput to be simultaneously achieved on compact and affordable desktop platforms [19,20,21].

Nevertheless, the reliable fabrication of enclosed, clog-free microchannels using Vat polymerization-based 3D printing remains fundamentally challenging. First, excessive curing or residual uncured resin trapped within enclosed channels can lead to channel blockage and reduced mechanical stability between stacked layers, potentially affecting biochemical reactions [22]. Moreover, post-processing steps—including solvent washing, vacuum or pressure-driven flushing, and UV curing—are essential for resin removal. As channel width and height decrease, the required applied pressure for overcoming high fluidic resistance increases significantly, making resin removal increasingly difficult [23,24,25,26]. These failures are not governed by printer resolution alone, but arise from a coupled interplay between optical exposure, channel geometry, and printing orientation, which makes the printability of enclosed microchannels difficult to predict using single printing parameters [27,28].

To address these challenges, this study presents a quantitative and systematic investigation on parameters of Vat polymerization printing for fabricating microfluidic channels using a commercial desktop printer and a commercially available photocurable resin. We examine how printing orientation, channel geometry, layer thickness, and ultraviolet exposure collectively influence microchannel openness, dimensional fidelity, and structural reliability at the hundred-micrometer scale. By analyzing these coupled effects, we identify practical fabrication boundaries that distinguish clog-free from failure-prone enclosed microchannels under commercial printing conditions. By combining this parametric–geometric analysis with device-level validation using a three-dimensionally stacked droplet generator, we demonstrate how these fabrication boundaries can be applied to guide the design of functional three-dimensional microfluidic devices. These results offer practical, experimentally grounded guidance for designing enclosed microchannels using commercial photopolymer printers, helping to reduce reliance on purely trial-and-error parameter tuning.

2. Materials and Methods

2.1. MSLA Printing System, Materials, and Device Design

All microfluidic structures were fabricated using a commercial desktop MSLA 3D printer (Saturn 4 Ultra, Elegoo, Shenzhen, China) equipped with a 405 nm ultraviolet light source. A commercially available transparent standard photopolymer resin (eSun Standard Resin, Clear, Shenzhen, China) was used as the printing material without further modification. Three-dimensional microchannel geometries were designed using computer-aided design (CAD) software (SolidWorks 2025, Dassault Systèmes, Waltham, MA, USA). The designed models were imported into the ELEGOO SatelLite software (https://www.elegoo.com/pages/satellite-3d-slicer?srsltid=AfmBOooTK8cyCg3WA2SFDoN-0Wfqdabj3ua_P-IMSxyX0rAqqTDq6XJo: accessed on 2 March 2025) for slicing, layout arrangement, and printing parameter configuration before printing. In addition, the information on the approximate printing cost has been provided in the Supplementary Materials (Table S1).

2.2. Printing Parameters and Fabrication Procedure

To systematically investigate the effects of printing parameters on microchannel printability and dimensional fidelity, layer thickness, exposure time, and printing orientation were independently varied while all other conditions were held constant. All parameter conditions used for the experiments shown in the following figures are summarized in Table S2. Layer thicknesses of 25, 50, 75, and 100 µm and ultraviolet exposure times ranging from 0.8 to 3.0 s (0.2 s increments) were evaluated using the partitioned exposure function of the ELEGOO SatelLite software, enabling efficient parameter screening within a single print. Unless otherwise specified, fabrication was performed using a layer thickness of 50 µm and an exposure time of 1.2 s, which were identified as the optimized conditions through systematic analysis. Orientation-dependent effects were examined by fabricating microchannels either in the XY plane or along the Z-axis without internal support structures, while channel lengths and widths were strictly defined by the CADs and remained unchanged during parameter variation. All printing was conducted at room temperature under ambient laboratory conditions, without resin heating, surface modification, or custom optical calibration, ensuring that the results reflect the intrinsic performance of a commercial desktop MSLA printing system under standard operating conditions.

2.3. Post-Processing and Structural Characterization

After printing, the samples were detached from the build platform and rinsed in 99.9% isopropyl alcohol (IPA) to remove uncured resin. To enhance solvent penetration into the enclosed microchannels, the parts were subjected to ultrasonic cleaning (BFSF-15S, BNF Korea, Gimpo, South Korea; 40 kHz, 360 W) for 10–15 min. Compressed air generated by an oil-free air compressor (Z-900DF8, Zhejiang Aurbita Pneumatic Tools LLC, Taizhou, China) was delivered through a handheld air gun (ZINAP) to expel residual IPA and trapped resin from the internal channels, ensuring complete channel opening. The samples were then dried under ambient laboratory conditions.

Ultraviolet post-curing was performed using a multifunctional wash and commercial curing system (Anycubic Wash & Cure 3 Plus, Anycubic, Shenzhen, China) equipped with 405 nm UV LEDs and an adjustable auxiliary curing lamp. Post-curing was conducted for 3–5 min to complete polymer crosslinking and stabilize the printed structures.

Channel openness was further verified by perfusing aqueous solutions containing water-soluble food dyes through the printed microchannels. Representative dye perfusion results are shown in Figure S1, where continuous dye propagation and sharp color boundaries provided direct visual confirmation of full channel opening and flow confinement. Any channels exhibiting partial blockage or incomplete dye penetration was conservatively categorized as blocked and excluded from subsequent structural or functional analysis.

Channel morphology and dimensional fidelity were evaluated using an inverted fluorescence microscope (IX71, Olympus Corp., Tokyo, Japan), with at least five independent samples analyzed for each condition. Channel widths were measured using ImageJ (NIH, Bethesda, MD, USA) at multiple locations along the channel length and compared with the CAD values to determine mean dimensional deviations and standard deviations. Surface morphology and channel cross-sectional profiles were further characterized by scanning electron microscopy using a tabletop SEM (Cube-II, EmCrafts, Hanam-si, South Korea), operated at 5–10 kV with a working distance of approximately 8–10 mm.

2.4. Droplet Generation and Performance Evaluation

Droplet generation experiments were conducted using the printed 5 × 5 multilayer droplet generator array to evaluate device-level functionality and flow stability. The continuous phase consisted of mineral oil containing 1% (w/v) Span-80 surfactant, while deionized (DI) water was used as the dispersed phase; both phases were independently supplied to the device using syringe pumps through standard tubing connections. Prior to operation, the device was visually inspected to confirm complete channel opening without clogging and the absence of leakage or structural defects. Droplet formation at the flow-focusing junctions was monitored by optical microscopy under steady-state flow conditions, and continuous operation was used to assess flow stability across the entire array. Droplet size and uniformity were quantified by analyzing optical micrographs acquired during operation, with the diameters of more than 100 droplets measured for each condition using image analysis software. Droplet size distributions were used to calculate the coefficient of variation (CV), defined as the ratio of the standard deviation to the mean droplet diameter, as a quantitative metric of droplet mono-dispersity and generation uniformity.

3. Results and Discussion

3.1. Printing Principle and Preliminary Characterization

A MSLA-based fabrication and cleaning workflow was implemented to achieve clog-free microchannels and to characterize the fundamental printing performance at the microscale. Figure 1 illustrates the process of fabricating and characterizing microchannel structures using an MSLA resin printer. During printing, ultraviolet light passes through a perfluoro alkoxy film (PFA) and is projected onto the resin vat, where selective layer-by-layer curing solidifies the photocurable resin on the build plate to generate the microchannel structures. Upon completion of printing, the part is immersed in 99.9% IPA and subjected to ultrasonic cleaning to remove uncured resin from within the channels. Following sonication, syringe-assisted flushing is performed to expel residual cleaning solution from the channels using airflow (Figure 1A). This cleaning step is critical to ensure complete channel openness. Finally, the dried part is placed in a UV curing chamber for post-curing under specified conditions.

Figure 1B shows a representative microfluidic chip fabricated via the resin printer, demonstrating that the Saturn 4 Ultra can construct complex branched networks and achieve integrated 3D channel fabrication on resin substrates. To verify channel openness, we prepared a series of colored acrylic dye solutions and introduced them into the channels. The different colors were uniformly distributed through the branching network and reached the terminal chambers. The distinct interface between orange and yellow droplets indicates that the channels exhibit strong liquid confinement, effectively maintaining flow boundaries between two phases. Moreover, the color contrast facilitated direct visualization of fluid diffusion and mixing behavior within the channels (see Figure S1), providing clear experimental evidence for subsequent quantitative analysis. To further examine the printed surface morphology, scanning electron microscopy (SEM) characterization was performed on the microchannel chip (Figure 1C). The SEM images present microchannel arrays with two designed sizes, where the left column corresponds to square microchannels with a designed width of 500 µm, and the right column corresponds to square microchannels with a designed width of 400 µm. Local magnified views reveal the detailed cross-sectional morphology of the printed channels, including the channel corners and sidewall features, which primarily arise from volumetric shrinkage and light scattering during the photocuring process. Overall, these preliminary results confirm that the printer is capable of fabricating finely defined microchannel structures at the hundred-micron scale.

3.2. Microchannel Length and Opening Ratio

To evaluate the printability limits of microchannels at the hundred-micron scale using a commercial desktop MSLA printer, channel openness was systematically examined as a function of both channel width and length, with a direct comparison between horizontally printed (XY plane) and vertically printed (Z-axis) configurations. The number of successfully opened channels was used as a quantitative metric to assess fabrication reliability (Figure 2). For microchannels fabricated in the XY plane (Figure 2A), a strong dependence on both channel width and length was observed. Channels with widths larger than 400 µm remained nearly fully open over the entire tested length range up to 10 mm, indicating robust printability under lateral exposure conditions. Channels with a width of 300 µm exhibited a clear length-dependent failure behavior: partial closure became apparent when the channel length exceeded 5 mm, followed by a sharp reduction in openness at lengths of 9–10 mm. For the narrowest channels (200 µm), blockage occurred even at short lengths of 1–3 mm, and the number of open channels rapidly decreased with increasing length, reaching complete closure at approximately 8 mm.

This pronounced sensitivity to horizontal channel length and width originates from the coupled effects of increased fluidic resistance and cumulative photopolymerization of the resin confined within enclosed geometries during MSLA printing. First, uncured resin becomes trapped inside enclosed microchannels and is repeatedly exposed to laterally scattered and axially penetrating UV light from adjacent pixels and successive layer projections. As the channel length increases or the channel width decreases, the ratio of illuminated wall area to internal fluid volume increases, thereby enhancing the probability of unintended polymerization within the confined channel volume. Subsequently, once internal narrowing occurs, post-processing determines whether the residual uncured resin can be effectively removed. Because the hydrodynamic resistance of a microchannel scales linearly with channel length and approximately with the inverse fourth power of the channel width (R ∝ L/w⁴) when the cross-sectional aspect ratio is unity, long and narrow channels exhibit dramatically increased flow resistance. As a result, solvent penetration and flushing of residual or partially cured resin become particularly ineffective.

In contrast, vertically printed microchannels along the Z-axis demonstrated substantially improved openness (Figure 2B). Channels with widths larger than 350 µm remained almost fully open across the full 10 mm length range, indicating a higher tolerance to increased channel length. Intermediate-width channels (250–300 µm) shows a gradual decline in openness with length but still retained relatively high printability at lengths of 8–9 mm. However, channels with widths less than 200 µm exhibited a pronounced reduction in printable length, with 150 µm-wide channels becoming nearly fully closed at ~ 4 mm and 200 µm-wide channels reaching zero openness at approximately 7 mm. The enhanced performance observed in the Z-axis orientation can be attributed to reduced lateral light scattering and more localized polymerization, which mitigate cumulative overexposure effects along the channel walls. Figure 2C provides representative images of microchannels with lengths of 1–3 mm, further illustrating the formation quality at different widths. Based on these results, the practical channel length limits and corresponding printing conditions for achieving open microchannels are summarized in Supplementary Tables S3 and S4.

Collectively, these results demonstrate that both channel width and printing orientation critically determine the achievable channel length and openness in MSLA-printed microfluidic devices. From a design perspective, vertical printing is therefore preferable for long or narrow channels, while horizontal printing requires more conservative geometric constraints to ensure reliable channel formation.

3.3. Dimensional Deviation Analysis

To quantitatively evaluate printing accuracy and orientation-dependent dimensional fidelity, the designed dimensions of microchannels were systematically compared with experimentally measured values for structures fabricated in the XY plane and along the Z-axis. For each design size, at least five independent samples were analyzed by optical microscopy, and the mean values with standard deviations are summarized in Figure 3 and Figure 4.

Figure 3 compares the cross-sectional dimensional deviations of open-structured and fully enclosed microchannels. Compared with the open channels (Figure 3A), both W_XY and H_Z shift toward smaller values in the enclosed channels (Figure 3B). In open channels, the presence of side openings facilitates resin flow and dilution of partially UV-exposed resin into the surrounding uncured resin during printing, as well as efficient resin drainage during post-processing. As a result, partially cured channel boundaries become more susceptible to cleaning-induced edge erosion or boundary relaxation, leading to an enlarged effective channel width in the XY plane. In contrast, enclosed channels experience limited resin flow and cumulative photo-activation of resin confined within the enclosed geometry. Consequently, fully enclosed microchannels exhibit a distinct anisotropic dimensional deviation (Figure 3B), in which W_XY remains oversized while H_Z becomes consistently smaller than the designed value. These clear structural differences highlight the importance of careful dimensional characterization of 3D-printed microfluidic platforms and provide practical guidelines for adjusting W_XY and H_Z to achieve the originally intended dimensions.

To further assess vertical accuracy across a broader length scale, microchannels with designed dimensions in the hundred-micron range and lengths from 1 to 10 mm were analyzed (Figure 4). A strong linear correlation between the designed and measured dimensions is observed, indicating robust and reproducible vertical resolution. Notably, the measured values are consistently slightly smaller than the design, suggesting a systematic shrinkage effect associated with photopolymerization and post-curing. Within the tested range, this undersizing remains limited to approximately 5–10%, and the small error bars reflect good repeatability across samples. Taken together, these results demonstrate that dimensional deviations in resin-printed microchannels are anisotropic and geometry-dependent, and that geometric confinement (open vs. enclosed) plays a critical role in cross-sectional fidelity. Open designs primarily exhibit lateral widening, whereas enclosed geometries additionally suffer from Z-axis height loss, and this anisotropic behavior underscores the importance of selecting an appropriate printing orientation and, where necessary, applying compensatory design offsets to achieve high-fidelity target dimensions at the hundred- to micron-scale.

3.4. Layer Thickness Effect

The effect of printing layer thickness on microchannel printability and dimensional fidelity was systematically investigated by varying the layer thickness from 25 to 100 µm while maintaining identical design geometries and exposure time (fixed at 1.0 s). Channel openness, dimensional accuracy, and surface morphology were evaluated to elucidate the trade-offs associated with layer-wise resolution, as summarized in Figure 5.

As shown in Figure 5A, reducing the layer thickness markedly improves the ability to fabricate long and narrow microchannels. At a layer thickness of 25 µm, channels with a designed width of 200 µm could be reliably printed with high openness over lengths up to 7 mm, particularly in the short-to-intermediate length range (1–5 mm). A similar trend was observed at 50 µm, although the maximum printable length was slightly reduced. In contrast, increasing the layer thickness to 75 µm significantly constrained the achievable channel length and width, with a rapid decline in openness for narrow channels. When the layer thickness was further increased to 100 µm, severe delamination and structural failure occurred, resulting in complete channel collapse; consequently, openness data for 100 µm were excluded from the quantitative analysis.

This failure at large layer thickness reflects a coupled optical–mechanical limitation of bottom-up MSLA printing. As layer thickness increases, the effective UV penetration depth may become insufficient to fully crosslink each layer into the previously cured layer, weakening interlayer bonding. At the same time, thicker layers involve a larger cured volume and contact area with the vat window, substantially increasing the peeling force during layer separation. When this force exceeds the interlayer adhesion strength, interfacial cracking, layer shifting, or complete detachment occurs. The observed improvement in channel openness at thinner layers can be attributed to reduced voxel height and improved layer-by-layer definition, which facilitate more effective resin drainage and limit vertical overcuring within confined channel spaces. However, thinner layers also increase the total number of exposure cycles required to build a given structure, thereby amplifying cumulative exposure effects and increasing risk to partial blockage in extremely narrow features.

Dimensional fidelity analysis further reveals a strong dependence on layer thickness (Figure 5B). Consistent with the intrinsic behavior of resin-based MSLA printing, all measured channel widths exceeded their designed values. Nevertheless, the magnitude of oversizing was strongly thickness-dependent. Channels printed at 25 µm exhibited the smallest deviation, typically within 5–15%, whereas those printed at 75 µm shows deviations exceeding 50% for features smaller than 350 µm. Surface morphology characterization by SEM (Figure 5C) corroborates these trends. Structures printed at thinner layers (25–50 µm) exhibit smoother channel walls and more uniform surfaces, whereas thicker layers result in pronounced surface roughness, cracking, and interlayer delamination. These morphological defects not only compromise dimensional fidelity but also directly contribute to channel blockage and structural instability.

Overall, layer thickness governs a fundamental trade-off between interlayer mechanical stability, dimensional fidelity, and internal channel openness. Among the tested conditions, approximately 50 µm provides the optimal balance for reliable microchannel fabrication using the desktop MSLA systems.

3.5. Exposure Time Optimization

The influence of exposure time on microchannel printability and dimensional fidelity was systematically investigated using the partitioned exposure function of the printer. Exposure durations ranging from 0.8 to 3.0 s were evaluated at 0.2 s intervals, while the layer thickness was fixed at 50 µm and the channel length was maintained at 2 mm. Channel openness, dimensional deviation, and failure modes were analyzed to identify an optimal exposure window, as summarized in Figure 6.

As shown in Figure 6A, channel openness depends strongly on both feature size and exposure time. Narrow channels (200–300 µm) exhibit a limited tolerance: 300 µm-wide channels remain open only below ~1.4 s, while 200 µm-wide channels require exposure times below ~1.0 s to avoid blockage. Wider channels (400–500 µm) tolerate longer exposures, remaining open up to ~2.2 s for 400 µm and ~2.4 s for 500 µm, although partial blockage appears at higher exposures. These results show that smaller features require much tighter exposure control.

Dimensional fidelity analysis further reveals the dual role of exposure time in governing printing accuracy (Figure 6B). Across all exposure conditions, the measured channel dimensions exceed the corresponding design values, consistent with the intrinsic oversizing behavior of resin-based MSLA printing. However, the magnitude of oversizing decreases as the exposure time increases. Channels fabricated at longer exposure times (1.6–2.0 s) exhibit closer agreement with the design dimensions, whereas those printed at shorter exposure times (0.8–1.2 s) show larger deviations. This behavior suggests that insufficient exposure leads to incomplete boundary polymerization and poorly defined feature edges, thereby increasing apparent dimensional error despite improved channel openness.

Exposure time directly governs the degree of polymer conversion and the effective curing depth within each printed layer. Longer exposures enhance wall consolidation and boundary definition, which reduces dimensional scatter, but they also increase lateral and axial light penetration, thereby promoting polymer growth inside enclosed channels. Representative micrographs in Figure 6C illustrate the characteristic failure modes associated with exposure variation. At optimized exposure conditions, channels remain fully open with well-defined boundaries. Moderate overexposure results in partial blockage due to excessive polymer growth at the channel walls, while severe overexposure leads to complete channel occlusion.

Overall, exposure time emerges as a decisive parameter controlling both microchannel openness and dimensional accuracy. Shorter exposure times suppress internal overcuring and favor channel openness but increase the risk of incomplete wall formation and blurred feature edges, whereas longer exposure times improve mechanical integrity and boundary definition at the expense of increased clogging risk. Taken together, these results indicate that an exposure window of approximately 1.0–1.2 s provides an optimal balance between reliability and accuracy for channels with widths of 300–400 µm under the present printing conditions. For narrower channels (<200 µm), further optimization combining reduced layer thickness, refined exposure control, and compensatory design strategies is required.

3.6. Droplet Generator Characterization

To validate the applicability of the optimized printing parameters for complex and functionally integrated microfluidic systems, we designed and fabricated a stacked 5 × 5 multilayer droplet generator array as a representative test case. The overall device architecture and internal flow network are illustrated in Figure 7A,B. The array consists of twenty-five flow-focusing droplet generation units interconnected through a three-dimensional distribution network, imposing stringent requirements on dimensional fidelity, channel continuity, and alignment accuracy across multiple embedded layers. Using the optimized printing conditions identified above (50 µm layer thickness, 1.2 s exposure time, Z-axis orientation), all internal flow pathways were successfully fabricated, yielding continuous, fully enclosed microchannels throughout the multilayer structure without visible defects or delamination. A critical indicator of fabrication fidelity is the dimensional accuracy of deeply embedded microchannels.

The designed channel geometries shown in Figure 7B were quantitatively compared with the printed cross-sections obtained from microscopic imaging (Figure 7C). The measured deviation in channel width remained below 2.5%, demonstrating that the optimized parameter set enables precise geometric reproduction even in spatially confined, three-dimensional regions. The sharply defined flow-focusing junctions and uniform orifice geometries further indicate that overcuring, polymerization shrinkage, and lateral light-scattering effects—commonly encountered in volumetric MSLA printing—were effectively suppressed under the selected conditions. The functional performance of the printed device was evaluated using the experimental setup shown in Figure 7D. The continuous phase (mineral oil containing 1% Span-80) and the dispersed phase (DI water) were independently supplied by syringe pumps, and the fully assembled droplet generator module with integrated tubing connections is shown in Figure 7E. Stable operation was achieved without leakage or structural failure, confirming both the mechanical integrity and interconnect reliability of the printed device. As shown in Figure 7F, the droplet generator array produced highly uniform water-in-oil droplets across all channels. Statistical analysis of more than 100 droplets yielded a coefficient of variation below 10%, indicating monodisperse droplet production and consistent flow-focusing behavior throughout the array.

Collectively, these results demonstrate that the optimized MSLA printing strategy enables the fabrication of complex, multilayer microfluidic devices with high dimensional precision and robust functional performance. The successful realization and operation of a 5 × 5 droplet generator array confirm that parameter optimization at the microscale directly translates into predictable hydrodynamic behavior at the device level. This capability, achieved using a commercial desktop MSLA printer and standard resin, highlights the potential of low-cost additive manufacturing for rapid prototyping of advanced microfluidic platforms, including multiplexed droplet generators, emulsification systems, and integrated lab-on-a-chip devices.

4. Conclusions

This study establishes a systematic framework for understanding and extending the fabrication limits of fabricating enclosed microfluidic channels using commercial desktop MSLA 3D printers and standard photocurable resins. By quantitatively correlating printing orientation, channel geometry, layer thickness, and UV exposure time with microchannel openness and dimensional fidelity, we identify explicit fabrication boundaries that distinguish clog-free from failure-prone regimes over the hundred-micrometer to millimeter scale. We demonstrate that channel printability is not governed by nominal pixel resolution alone but by the coupled effects of optical overexposure, lateral light scattering, resin confinement, and curing-induced shrinkage, which produce fundamentally different and anisotropic error modes in the XY plane and along the Z axis. In particular, lateral channel openings in the XY plane exhibit systematic widening, while fully enclosed geometries additionally show a consistent reduction in effective Z-axis channel height, reflecting confinement-related curing and resin retention effects that directly reduce the functional flow cross-section. Channel viability is therefore jointly constrained by geometry, exposure conditions, and post-processing dynamics. Fully open channels are reliably attainable for widths ≥ 400 µm over centimeter-scale lengths, whereas narrower features remain limited by resin entrapment rather than exposure resolution alone. Guided by these insights, an optimized parameter set (50 µm layer thickness, ~1.0–1.2 s exposure time, and Z-axis channel orientation) enabled the high-fidelity fabrication of a 3D stacked 5 × 5 droplet generator array, exhibiting less than 2.5% deviation in dimension and stable production of monodisperse water-in-oil droplets (CV < 10%). It should be noted that the reported fabrication boundaries are derived from a specific commercial MSLA printer, resin formulation, and post-processing workflow. Variations in optical configuration, resin properties, or cleaning conditions may shift the exact printable limits. In addition, these boundaries were primarily determined for channel dimensions larger than 200 µm, and extension to smaller feature sizes or alternative material systems requires further investigation. Collectively, these results define practical, experimentally validated design guidelines that bridge consumer-grade additive manufacturing with reliable microfluidic functionality, significantly lowering the cost and technical barriers for rapid prototyping and deployment of complex lab-on-a-chip systems.