Subtleties of UV-crosslinking in microfluidic particle fabrication: UV dosage and intensity matter

TL;DR

This work reveals that UV-crosslinking of PEGDA in microfluidic droplets is governed more by UV intensity than total dosage, with higher concentrations requiring stronger illumination to achieve solid-like restoration. By integrating droplet generation, inline curing, and multiple mechanical readouts, the authors show that gelation proceeds from the inside out and that many droplets retain an outer uncured water shell likely due to oxygen quenching. Cross-validation across in-flow deformation, capillary micromechanics, and pendant-droplet experiments demonstrates the presence and tunability of this shell, which impacts measured mechanics and morphology. The findings provide practical guidelines for designing inline curing protocols and highlight how controlled partial curing and shell manipulation can enable bespoke microgel architectures for drug delivery and cell therapy applications.

Abstract

Curable hydrogels have tunable properties that make them well-suited for applications in drug delivery, cell therapies, and 3D bioprinting. Advances in microfluidic droplet generation enable rapid fabrication of polymer-filled droplets. UV-curable polymers offer a clear path toward using fluidic generation to produce monodisperse microgels with uniform properties. In flow, polymer concentration and UV exposure both control the degree of crosslinking. High UV intensity is often used to ensure complete gelation and avoid complications that may arise from partial curing. Optical microscopy can assess droplet and particle sizes in flow. However, optimizing formulations for mechanical properties usually requires removal of generated material and external measurement outside of flow. In this study, we couple droplet generation, microgel fabrication, and mechanics assessment within a single fluidic device. We make and measure soft polyethylene glycol diacrylate (PEGDA) microgels by curing polymer-filled water drops in mineral oil. Crosslinking is tuned by varying UV dosage, allowing us to study how gelation degree influences microgel properties. Within the device, we use shape deformation in flow to measure the restoring stress of both droplets and particles. Our results suggest that PEGDA droplets gel from the inside out. If gelation is incomplete, a particle resides within a fluid drop. Independent measurements outside of flow corroborate this observation. Crosslinking PEGDA-filled droplets in a pendant drop geometry, with dye, suggests the persistence of an aqueous shell around the gel. Similarly, microparticles in PEGDA-filled drops undergoing gelation exhibit diffusive arrest near the drop center, while maintaining mobility in an outer region. Together, these results suggest the importance of considering the extent of gelation when fabricating microgels using fluidics.

Figures (16)

• Figure 1: Schematics of experimental setups. a and b show the two fluidic device designs. Both devices feature an upstream droplet maker and a long residence channel for UV exposure. The device in Fig. \ref{['UVschematic3']}a has a channel width which sequentially decreases from $200 \mu$m to $40 \mu$m, in intervals of $40 \mu$m. The device in Fig. \ref{['UVschematic3']}b has a channel with width $500 \mu$m that is interrupted by a series of constrictions ranging from 100 $\mu$m to 40 $\mu$m. Panel c shows a schematic of the UV light setup: UV light shines from the tip of the spot curing light guide. The UV light shines through the PDMS slab of the device to the droplet flow path on the glass slide. The inset to Fig. \ref{['UVschematic3']}c shows a calibration of equation \ref{['eq: UV intensity final']} to experimental values of $I_{UV}$ taken using a UV intensity meter.
• Figure 2: Fig. \ref{['fig: restoring stress']}a: An example of the measured deformation in each constriction plotted as a function of viscous shear stress. Restoring stress, $\tau$ is the inverse slope of the linear fit. The [id=R2]gray dashed error bars on each data point represent the standard deviation for all [id=R2]$\sim1000$ droplets or particles measured in each constriction. [id=R2]The standard error of the mean is also plotted, in black, and is smaller than the size of each data point. The error bars on the measurement of $\tau$ are given by the [id=R2]standard error of the [id=R2]linear fit [id=R2]corresponding to a 68% [id=R2]95% confidence interval [id=R2]of the linear regression of the entire data set. Fig. \ref{['fig: restoring stress']}b: Restoring stress as a function of UV dosage, $\Omega$, for PEGDA concentrations $9$ to $39\%$ as shown in the legend. Error bars represent error propagation from the linear fit in Fig. \ref{['fig: restoring stress']}a. Fig. \ref{['fig: restoring stress']}c shows the same data as in Fig. \ref{['fig: restoring stress']}b, plotted as a function of UV intensity at $x=10$ mm from the beginning of light exposure ($I_{UV_{10}}$). Fig. \ref{['fig: restoring stress']}d: A four-dimensional representation of $\tau$, indicated by the colorbar, as a function of $I_{UV_{10}}$, $\Omega$ and $t_r$ for $39$% PEGDA. The green line on the colorbar represents $\tau=230$ Pa.
• Figure 3: Fig \ref{['fig: kinetics compilation_1']}a: Gelation time in a bulk gel decreases with illumination intensity $I_{UV}$ for bulk gels made of $c=9$ and $39$% PEGDA. Gelation time is measured as the time between illumination $t_i$ and the arrest of tracer particles, $t_a$. Error bars represent the standard deviation between three samples. Fig \ref{['fig: kinetics compilation_1']}b: Gelation time plotted on a log-log axis as a function of $c\Omega$ nearly collapses onto a power law with slope $-1$, as shown by the blue dashed line. Fig \ref{['fig: kinetics compilation_1']}c: Conditions explored in the fluidic measurements are plotted on log-log axis like those used in b. Time on the $y$ axis refers to both residence time in the fluidics tests and gelation time in the bulk gels, as indicated by the dashed line. The color bar indicates measurements of $\tau$ as in Fig. \ref{['fig: restoring stress']}d.
• Figure 4: Fig.s \ref{['fig: capillary micromechanics']}a-e show examples of capillary micromechanics measurements. The examples correspond to: Fig. \ref{['fig: capillary micromechanics']}a, $c=19\%$ homogeneous gel particle after exposure to 159 J/cm$^2$ with labels for $R_{band}$ and $L_{band}$, Fig. \ref{['fig: capillary micromechanics']}b, 29% PEGDA after exposure to 159 J/cm$^2$, Fig. \ref{['fig: capillary micromechanics']}c-e, $c=29\%$ PEGDA after exposure to $\Omega=6$ J/cm$^2$ with pressure increasing from left to right, from $100$ to $500$ to $1000$ mbar. In Fig.s \ref{['fig: capillary micromechanics']}a-e, blue notations refer to gels while orange notations refer to uncured liquid; all scale bars are $50\mu$m. Fig. \ref{['fig: capillary micromechanics']}f shows a compilation of measurements of $G_m$ of gels in the capillary micromechanics compared to measurements of $\tau$ from fluidics and $G_b$ from bulk rheology on PEGDA gels of the same chemistry. Anselmo2015
• Figure 5: Fig \ref{['fig: pendant drop images']}a: Controls of pendant droplet set-up exposed to 27.4 mW/cm$^2$ with images of exposure at 1 s and 70 s. The first row depicts a sample of water with no PEGDA. The second row depicts $c=100\%$ PEGDA with no photoinitiator. The last row displays $c=99\%$ PEGDA with 1% photoinitiator. Red error bars represent 0.5 mm. Fig. \ref{['fig: pendant drop images']}b: Normalized grayscale image intensity versus exposure time, $t$, for $c=39\%$ PEGDA and five UV intensities. Shaded regions represent the deviation of three samples per UV exposure. The inset is the time for the normalized image intensity to plateau ($t_p$) as a function of $I_{UV}$. Fig. \ref{['fig: pendant drop images']}c: Pendant drop images of $c=39\%$ PEGDA at $t=1$ s and $t\geq t_p$ for three different UV intensities, intensity increasing from the top row to the bottom row. The red scale Red error bars in Fig \ref{['fig: pendant drop images']}a and Fig \ref{['fig: pendant drop images']}c represent 0.5 mm.