DIY 3D Microparticle Generation from Next Generation Optofluidic Fabrication

Abstract Complex‐shaped microparticles can enhance applications in drug delivery, tissue engineering, and structural materials, although techniques to fabricate these particles remain limited. A microfluidics‐based process called optofluidic fabrication that utilizes inertial flows and ultraviolet polymerization has shown great potential for creating highly 3D‐shaped particles in a high‐throughput manner, but the particle dimensions are mainly at the millimeter scale. Here, a next generation optofluidic fabrication process is presented that utilizes on‐the‐fly fabricated multiscale fluidic channels producing customized sub‐100 µm 3D‐shaped microparticles. This flexible design scheme offers a user‐friendly platform for rapid prototyping of new 3D particle shapes, providing greater potential for creating impactful engineered microparticles.

[Note S1] Pressure drop analysis: When scaling down a channel, higher hydraulic resistance increases the pressure drop across the channel necessary to match the Reynolds number. The necessary pressure increase can be calculated using a hydraulic circuit model and solving for the pressure drop ∆ across a channel as a function of the Reynolds number. The pressure drop in a channel can be calculated as follows: where Q is the volumetric flow rate and ℎ is the hydraulic resistance. The hydraulic resistance for a square channel can be approximated using Equation S2 [1] , where is the dynamic viscosity, is the channel length, and ℎ is the hydraulic diameter. The Reynolds number can then be solved in terms of the volumetric flow rate, as follows, where ρ is the density and avg is the average flow velocity. Finally, by solving Equation S3 for Q and plugging into Equation S1 in combination with Equation S2 allows the estimation of the pressure drop as follows, implies that when matching the Reynolds number in a channel that is 10 times smaller, the pressure drop across the channel increases by a factor of 10 3 .

[Note S2]
UV illumination conditions and throughput: UV light is patterned using a photomask, demagnified through an objective, and focused on the fluidic channel (Figure 1b, d). The necessary UV exposure time for polymerization depends on UV power, channel height, and objective. For example, a 2.5× objective is useful for polymerizing particles at the millimeter scale owing to a large field-of-view. However, when creating microparticles, the UV pattern can be demagnified to a smaller field-of-view, thus increasing the UV power density. Therefore, higher magnification objectives can polymerize smaller particles with shorter UV exposure times, although limited field-of-views and higher numerical apertures make higher magnified objectives less suitable for large and tall channels. Because of the larger S6 channel dimensions (6 mm × 1 mm), a 2.5× objective with a large field-of-view was used for polymerizing the S6 particles with 2 s UV exposure. For polymerizing the particles in the reduction section at location R, a 20× objective was used with 20 ms UV exposure due to the higher UV power delivery achieved from greater demagnification in the smaller channel section (~600 µm × 100 µm). After polymerizing a particle, the NG-OF process can be repeated with a cycle time of approximately 10 s. Particles were created with a throughput of 360 particles/hour in the presented work. Using photomasks with multiple patterns placed side-by-side, at least 10 particles could be polymerized at a time based on the current 1,200 µm wide field-of-view, thus leading to a potential throughput of 3,600 particles per hour.
To create a DIY pillar and overcome the oxygen inhibition layer, a 10× objective was used with a higher UV power density of >200 mW cm -2 to create an "anchor" pillar (see Movie S1 ). Subsequently, a 2.5× objective with a UV power density of ~30 mW cm -2 created the final pillar around the anchor pillar. This two-step process ensures that the anchor pillar is fixed in place using the more powerful UV exposure, while the less powerful 2.5× objective could still create the large millimeter scale pillars from the larger field-of-view. Longer UV exposures were observed to create larger pillar diameters ( Figure S10). However, little growth was observed after 8 s of UV exposure, and thus a UV exposure of 8 s was chosen in this study.

[Note S3] Experimental Section
Channel Fabrication: Polydimethylsiloxane (PDMS) channels from Sylgard 184 Silicone Elastomer (Dow Corning Corporation, MI, USA) were created using soft lithography on 3D printed channel molds. An Objet500 3D printer was used with the VeroWhitePlus material (Purple Porcupine, CA, USA) with a specified manufacturer x-y resolution of 42 µm, and a 16 µm layer thickness for printing channel molds. PDMS was prepared using a base-to-curing agent ratio of 5:1 to help stiffen the channels [2] for faster flow stop times (approximately 15% decrease in flow stoppage at Re = 5). The uncured PDMS was poured over channel molds, desiccated to remove air bubbles, and was allowed to cure at room temperatures for 24 hours to prevent heat

Figure S3. Channel geometries based on designed and fabricated results. a)
The tapered reduction section is designed to reduce the channel cross-section from 6 × 1 mm 2 to 600 × 100 µm 2 . b) The designed reduction cross-section shown at section AA' is rectangular and maintains the same 6:1 aspect ratio. c) A top view schematic shows the x-y taper angle of 10° leading to a 2 mm long reduction section. d) 3D printed channel mold with the reduction section labeled as DD'. e) The 3D printed channel mold has a rounded cross-section illustrated based on the section DD' owing to the intrinsic resolution limits of the photocurable inkjet 3D printer. f) Experimental cross-section of the reduction channel. A PDMS channel created from the 3D printed mold was filled with photocurable PEG-DA, and a thin slit of UV light was illuminated to polymerize a thin particle. The cross-section of the displayed particle represents the crosssectional shape of the channel mold. g) The simulated channel design was rectangular except for the reduction h) where the cross-sectional shape transitions towards the actual rounded crosssection from the 3D printed mold.