Effect of Droplet Surface on the Wetting of Gelatin-Rich Phase.

The PEG and gelatin blend spontaneously separated in the aqueous solution. Fig. 1A shows the complete separation of a PEG/gelatin solution after incubation for 60 min at 30 °C, with the PEG-rich L-phase on top and the gelatin-rich G-phase on the bottom. This phase separation has been well characterized (18, 19). Here, we used 1.7 wt% PEG20000 to set the phase separation point, T_p, slightly above the gelation point, T_g. Under this condition, these transition points (T_p and T_g) are plotted against the gelatin concentration in Fig. 1B (19). With a decrease in the temperature, the PEG/gelatin blend first separates into two liquid phases at T_p (L-L phase), and then the gelatin-rich phase turns into a gel below T_g (< T_p) (L-G phase). The double-quench conditions enable us to regulate the coupling between the phase separation and gelation. Thereafter, the temperature was shifted from 60 °C (above T_p, one phase) to 30 °C (just below T_p, L-L phase), retained for a period t at 30 °C, and then quenched to 20 °C (below T_g, L-G phase). A fast quench with a smaller t delays the phase separation by the gelation and produces spherical microgels or a porous gel according to their volume fractions, as previously reported (Fig. S1) (19, 20).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

(A) An example of phase-separated solution of PEG 1.7 wt% and gelatin 5.0 wt% in tube. (B) Schematic phase diagram of PEG/gelatin blends containing the PEG 1.7 wt% and various concentrations of gelatin. The black solid and dashed lines indicate the phase separation point, T_p, and the gelation point, T_g, respectively. (C) Phase separation of PEG and gelatin 5.0 wt% blend in PE and PC droplets after the incubation time t = 60 min. Droplets are shown in (Left) DIC images, (Center) fluorescent images of the gelatin-rich gel phase, and (Right) schematic illustrations.

To break the spherical symmetry of the microgel shapes by the interactions between the gelatin and interface, we confined the PEG/gelatin solution in microdroplets coated by a lipid layer. First, we tested the symmetry-breaking effect of the droplet confinement with a slow quench. Two types of zwitterionic lipids were examined to prepare the droplets: 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (PE) (negatively charged in water) and 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine (PC) (electrically neutral in water) (21, 22). For the PE, all droplets spontaneously formed a hollow microcapsule of the gelatin-rich G-phase (t = 60 min), trapping an isolated domain of the PEG-rich L-phase in the center; i.e., the gelatin-rich phase completely wetted to the PE membrane (Fig. 1C). In contrast, the gelatin-rich phase partially wetted to the PC membrane. We found that the partial wetting of the gelatin to the droplet surface yielded microgels with spherically asymmetric shapes.

Effect of Droplet Size on Stable Patterns.

To determine the effect of the droplet size on the microgel shapes, we investigated the stable gel patterns under a slow quench (t = 60 min). Under this condition, all PE droplets reached the macrophase separation and formed a microcapsule of gelatin gel, regardless of the droplet size. We analyzed the relation between the average thickness of the gel wall, l, and the droplet radius, R, for the PE droplets containing PEG and gelatin 5.0 wt%. According to volume conservation, the volume fraction of the gelatin-rich phase, v_g, in a droplet with radius R is expressed as follows:43πR3vg=43πR3−43π(R−l)3,[1]which leads to a linear relation between l and R: l = R(1 − (1 − v_g)^1/3). In fact, the experimental results indicated that the thickness l increased proportionally to the droplet size R (Fig. 2A). This means that the hollow capsules in all PE droplets have similar shapes. Thus, the stable patterns of the PE droplets can be determined according to the droplet size R, i.e., by adjusting the thickness of the spherical microcapsule. From a linear fit, v_g was found to be 56 vol%, higher than that for the bulk system, which was ∼45 vol%. This suggests that the interactions between the gelatin and lipid membranes slightly vary the compositions of the two phases.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Droplet size dependence of the stable patterns and domain coarsening in droplets containing PEG 1.7 wt% and gelatin 5.0 wt%. (A) Average thickness of microcapsules in PE droplets after t = 60 min is plotted against the droplet radius, R. (B) Frequency distribution histogram of PC droplets having a monodomain (n_d = 1; black) and multidomains (n_d > 1; gray). (C) (Left) Time development of domain size, d, in PC droplets with a different R. The data are fitted with a power low (d ∼ t^α). The dashed and solid lines represent α = 1/3 and 1, respectively. (Right) The relation between α and R is analyzed from 30 droplets.

In contrast, the PC droplets exhibited a strong size dependence on the number of domains, n_d (Fig. 2B). Although the smaller PC droplets with R < 40 μm reached the macrophase separation (i.e., n_d = 1) as the PE droplets did, the larger PC droplets with R > 40 μm had some wetting domains (n_d > 1). To elucidate the effect of the PC droplet size on n_d, we analyzed the domain coarsening at the center of the PC droplets with a different R. Small isolated domains of the gelatin-rich L-phase grew in size to ∼15 μm with time as t^α by the collision and coalescence (Fig. 2C, Left). By fitting the data at the early stage of phase separation, we determined the relation between α and R: α is constant at 1/3 for R > 40 μm, and when R < 40 μm, it increases from 1/3 to ∼1 as R decreases (Fig. 2C, Right). This indicates that the phase separation is accelerated in smaller droplets having a large area-to-volume ratio. This accelerated behavior in the smaller droplets resembles a polymer blend confined in a 2D space, wherein the thickness of the wetting layer increases as t^α with α = 1 or 2/3 at the early stage (23⇓–25). In the late stage of the phase separation, the migration of the large wetting domains stopped. Therefore, the stable state is obtained only for the smaller PC droplets, unlike the case for the PE droplets.

Shape Deformation of Smaller PC Droplets upon Dewetting Transition.

Although all PC droplets in the L-L phase had a spherical shape, the smaller PC droplets began to deform along the PEG/gelatin interface by the second quench below T_g and then had a stable shape. Fig. 3A shows examples of the deformed PC droplets. To quantify the R dependence of the shape, we analyzed the aspect ratio of 50 droplets with various R, i.e., the ratio of the short axis along the domain boundary, a, to the long axis along the deformed direction, b. The aspect ratio, a/b, is plotted against the initial droplet radius R above T_g (Fig. 3A, Right). a/b suddenly decreased from 1 to ∼0.6 when R was smaller than the critical size, R*. R* was ∼25 and ∼10 μm for the PEG and gelatin 5.0 and 7.0 wt% systems, respectively. The deformed droplets returned to their initial spherical shape as the temperature increased above T_g immediately after the second quench.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

(A) Exampled of deformed PC droplets along the domain boundary. The aspect ratio, a/b, is plotted against the initial radius of spherical PC droplets. (B) Temperature dependence of contact angle, θ, between an air bubble and gelatin–gel surface (square), and of the reduced interfacial tension between PEG–sol and gelatin–gel phases, Γ (circle). The arrow indicates the phase separation point, T_p (i.e., Γ = 0).

The aforementioned results indicate that the interfacial tension between the PEG and gelatin phases, Γ, in the L-G phase is larger than that in the L-L phase. We estimated Γ according to the balance of the interfacial tensions at the triple point of the gelatin–gel surface, the PEG solution, and an air bubble (Fig. 3B, Left). Γ should satisfy the Young equation: γ_ag = Γ + γ_apcosθ, where γ_ag and γ_ap are the interfacial tensions between the air/gelatin and air/PEG phases, respectively, and the θ is the contact angle between the air bubble and the gelatin–gel surface. We measured θ by changing the temperature and obtained a value of Γ by assuming the previously reported values of γ_ag ∼ 70 mN/m (26, 27) and γ_ap ∼ 45 mN/m (28). The obtained Γ in the L-G phase (>20 mN/m) was far larger than the estimated value in the L-L phase [<0.1 mN/m (29)] (Fig. 3B, Right). The temperature dependence of Γ suggests a sudden increase near the gelation point, as Γ should be zero above T_p (indicated by an arrow). We concluded that this increasing interfacial tension below T_g triggered the shape deformation (dewetting transition).

Trapped States upon Dynamical Coupling Between Phase Separation and Gelation.

The effect of the incubation time, t, on the pattern formation of the microgels in the PE droplets was evaluated (Fig. 4A). Under a fast-quench condition (t = 0 min), the PE droplets containing the PEG and gelatin 7.0 wt% blend (volume fraction of gel phase: ψ_g ∼ 0.55; right lines in Fig. 4A) formed porous microgels entrapping a domain of the PEG-rich phase. Reducing t caused the number of PEG-rich domains (porous) to increase and the porous size to decrease. In the case of the PEG and gelatin 5.0 wt% blend (ψ_g ∼ 0.45; left lines in Fig. 4A), reducing t decreased the volume of the gelatin-rich wetting layer that forms the capsule and increased the number of isolated microgels inside the capsule. In contrast to the linear relation between the thickness l and droplet size R after t = 60 min (Fig. 2B), l after t = 0 min was fixed as 1.9 ± 0.9 μm, regardless of R (Fig. 4B). This strongly suggests the appearance of the wetting layer of the gelatin at the early stage of the phase separation (23).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

(A) Various hollow microcapsules in PE droplets in response to changes in volume fraction of gelatin-rich phase (x axis) and number of isolated domains regulated by an incubation time, t (z axis). PE droplets are shown in (Left) fluorescent images of gelatin-rich phase and (Right) schematic images. (Scale bar: 50 μm.) Red lines indicate the thickness of the capsule wall. (B) The thickness of capsules under a fast quench condition (t = 0 min) is plotted against the droplet size (gelatin 5.0 wt%).

The PC droplets exhibited variously shaped microgels as t was changed. Fig. 5 shows some examples, with respect to the volume fraction of the gel phase ψ_g (x axis), droplet radius R (y axis), and number of isolated domains n_d (controlled by t; z axis). The droplets with R ∼ 50 μm in the x–z plane show the effects of the gel patterns on n_d and ψ_g. When the droplets had over 50 domains (n_d >> 50), they maintained their spherical shapes irrespective of R and formed isolated smaller microgels (ψ_g ∼ 0.45) or a porous gel (ψ_g ∼ 0.55). When n_d was smaller than ∼10, spherically asymmetric microgels were observed, depending on the ψ_g: hemisphere-, disk-, and star-shaped. In addition, when the size of the droplets having a few domains was smaller than ∼25 μm, those droplets deformed their shapes upon dewetting. Consequently, the dynamical coupling between the phase separation and gelation yielded a great variety of microgel shapes for the smaller PC droplets, accompanied by the partial wetting and shape deformation.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Multiple trapped patterns of microgels in PC droplets in response to changes in volume fraction of gelatin-rich phase, droplet size, and number of isolated domains. In the x–z plane, large droplets with R ∼ 50 μm are shown in fluorescent images and schematic illustrations, where gelatin-rich phase is shown in white and yellow, respectively. In the y–z plane, small droplets with R ∼ 20 μm are shown as DIC images.

Collection of Asymmetrical Microgels from Droplets to Aqueous Solutions.

Finally, we collected the patterned microgels from the droplets. To maintain the gel shape, we added a PEG 2.0 wt% solution to the PC droplets having a hemisphere microgel using a microinjection system. This procedure disrupted the PC membrane and exposed the hemisphere microgel (Fig. S2). By comparing the gel shape between the droplets (yellow lines) and PEG 2.0 wt% solution (red lines), the microgel was found to be swollen. In contrast, the addition of 50 vol% ethanol having a low affinity for the gelatin solution, shrank the gel. The volume was again increased by the addition of the PEG 2.0 wt% solution. The collected microgel maintained the hemisphere shape after the volume phase transition. In addition, the change in the curvature of the wetting side (indicated by an arrow in Fig. S2) was slightly smaller than that of the nonwetting side, indicating that the wetting of the gelatin affects the gel structure.