One-step synthesis of spherical/nonspherical polymeric microparticles using non-equilibrium microfluidic droplets

Abstract

A simple but versatile microfluidic process is presented for the production of monodisperse polymeric microparticles using non-equilibrium droplets. Oil-in-water (o/w) droplets were formed within microchannels by means of a water-soluble polar organic solvent containing polymer molecules as a dispersed phase. The droplets of solvent were rapidly dissolved into the continuous phase during flow through the microchannel, whereas water-insoluble polymers were precipitated to form monodisperse polymeric particles. In this way, we successfully synthesized particles with sizes significantly smaller than that of the initial droplets, using polystyrene and poly(methyl methacrylate) as the polymer molecules and typically ethyl acetate as the polar solvent. Particles obtained under several different conditions exhibited unique non-spherical morphologies, which were caused by the translocation and segregation of the precipitated polymer and diminishing solvent-rich droplet. We confirmed that various factors influence the particle morphology, including the polymer concentration, molecular weight of the polymer, type of the polar solvent, and the presence of additives in the dispersed phase.

---

Introduction

Monodisperse micrometer-sized polymeric particles, with controlled morphologies and compositions, are gaining increasing attention as a novel material for a wide range of industrial, clinical, and research applications. Currently, a variety of conventional methods are available to produce micrometer-sized polymeric particles, including bulk-scale dispersion polymerization,^1,2 membrane/column emulsification,^3,4 and aerosol polymerization.⁵ Compared to these conventional techniques, the use of microfluidic devices is highly advantageous because of their capability to produce monodisperse droplets using relatively simple experimental setups and operations.^6,7 Flow-focusing channels, T-shape junctions, and micronozzle devices have previously been utilized to produce monomer droplets, which are then further polymerized into solid particles.^8–11 In most of the conventional techniques, spherical particles are usually obtained because of the minimal interfacial energy of the spherical droplets. In contrast, microfluidic devices are capable of producing functional polymeric particles having non-spherical,^11–15 Janus,^16–21 hollow,²² or highly-complicated morphologies,^23–25 which is achieved by deforming droplets in narrow channels, generating multicomponent droplets in stable laminar flows, or polymerizing a specific region within the droplets.

Generally speaking, many microfluidic methods employ a two-step process for producing polymeric particles, including the formation of monomer droplets with initiators and their subsequent on- or off-chip polymerization via UV irradiation or heating. Although such a process is the most common way to produce particles, it does introduce several problems. Firstly, the particle size, composition, and degree of polymerization cannot be individually controlled at the same time, as the physicochemical characteristics of a monomer solution critically affect the generation behavior of droplets. Secondly, the polymerization process causes a volume change,⁸ and thus can potentially degrade the initial uniformity of the droplet size. Finally, it is not easy to produce monomer droplets with sizes much smaller than the microchannel dimensions unless specific schemes such as droplet jetting are employed.

As a method for producing polymeric particles without employing polymerization, researchers have developed new processes, often referred to as solvent diffusion, that rely on using previously-synthesized polymer molecules.^26–34 In such processes, droplets of a polymer-containing organic solvent are initially generated, from which the solvent is gradually removed by slow evaporation or dilution. This has allowed microparticles to be produced using various types of polymers including poly(lactic acid), polystyrene (PS), poly(methyl methacrylate) (PMMA), polysulfone, and polyfluorene. Moreover, polysaccharide gel beads have also been produced using water-in-oil (w/o) droplets in microchannels.^35,36 Most of these processes utilize the slow diffusion of the solvent from the o/w droplets through the continuous water phase. However, some studies have reported the use of rapidly dissolving, non-equilibrium droplets for particle synthesis; such droplets being temporarily formed in microchannels.^27,28,35,36 Being non-equilibrium in nature, such droplets cannot be generated in a controlled manner through conventional mixing processes, but rather require microfluidic devices. The rapid removal of solvents from non-equilibrium droplets shows great promise, as it is capable of producing particles with non-spherical morphologies.^27,36 In addition, particles with sizes much smaller than that of the initial droplets can be obtained. However, to the best of our knowledge, there has been only one report of particles of common polymers such as PMMA or PS having been fabricated from non-equilibrium droplets;²⁷ toroidal PMMA particles were produced using silicone oil and N,N-dimethylformamide as the continuous and dispersed phases, respectively, with the help of the axisymmetric laminar flow in microchannels.

In this study, we propose a simple but highly versatile approach utilizing non-equilibrium droplets for producing microparticles made of common polymers such as PS and PMMA. In this process (Fig. 1), non-equilibrium o/w droplets are first generated at the microchannel confluence using water and a water-soluble organic solvent. During flowing through the microchannel, the solvent in the droplets is rapidly dissolved into the continuous water phase. By adding a small amount of water-insoluble polymer molecules to the dispersed phase, the concentration of the polymer gradually increases. Polymeric particles are finally formed by completely removing the solvent via further dilution. By this means, we can control the size of the synthesized particles by adjusting either the polymer concentration or the droplet size; i.e., when the polymer concentration is sufficiently low, particles with a size much smaller than the initial droplet size are obtained. In a series of experiments, we prepared PS and PMMA particles by using several types of solvents, though mainly ethyl acetate (EA), and examined several factors (i.e., polymer concentration, molecular weight, and polar solvent) that affected the size and morphology of the particles obtained.

Fig. 1 Schematic illustration showing the microfluidic system used to produce polymeric microparticles from non-equilibrium droplets. A dispersed phase consisting of a polar organic solvent with polymer molecules and a continuous phase of distilled water with stabilizer were continuously introduced into the microchannel.

Experimental section

Materials

Polystyrene (PS, polymerization degree of ∼2000), poly(methyl methacrylate) (PMMA, M_w of 100 000), polyvinyl alcohol (PVA, polymerization degree of 500), methyl acetate (MA, purity of >98%), ethyl acetate (EA, purity of >99.9%), propyl acetate (PA, purity of >97%), and oleic acid (OA) were all obtained from Wako Pure Chemical Ind. Ltd., Osaka, Japan. Polystyrenes with different molecular weights (M_w/M_n = 13 200/12 400 and 983 400/950 400) were obtained from Sigma-Aldrich Corp., MO, USA. Pyrene was obtained from Kanto Chemical Corp., Tokyo, Japan. All other chemicals used were of analytical grade.

The average molecular weight and polydispersity of the PS with a polymerization degree of ∼2000 (from Wako) were determined by a Hitachi GPC modular system comprising a LaChrom L-7100 pump, a LaChrom L-7300 column oven, a TOSOH guard column α, and a JASCO RI-1530 intelligent RI detector. The molecular weight separation was performed by a column set of TOSOH TSKgel α3000 and α4000 using DMF as an eluent at a flow rate of 0.6 mL min⁻¹ at 40 °C. The calibration curve was prepared by commercial linear PS standards. The resulting data were analyzed using the Run Time Corporation software package Chromato-PRO-GPC Version 3.0.0, from which M_n and M_w were determined to be 59 600 and 114 700, respectively.

Preparation of polymeric particles

PDMS-glass microfluidic devices were fabricated using the standard soft lithography and replica molding techniques as described elsewhere.³⁷ A simple flow-focusing microchannel was employed to generate droplets. The microchannel width was a consistent 200 μm with the exception of the confluence point (50 μm). The channel depth was uniform at ∼55 μm. Hydrophilic microchannels were employed in order to stably generate o/w droplets; a PDMS plate with the microchannel structure and a flat glass slide being bonded via O₂ plasma treatment using a plasma reactor (PR500, Yamato Scientific Corp., Tokyo, Japan), and the as-prepared microdevices used immediately for experiments.

PS and PMMA were used as the polymer molecules, whereas EA, MA, and PA were used as the polar organic solvent. Polymer molecules were dissolved in the polar solvent at a concentration of 1–5% (w/v). Distilled water containing 2.5% (w/v) PVA was used as the continuous phase. These solutions were continuously introduced into the microchannel using syringe pumps (KDS200, KD Scientific Corp., MA, USA). The flow rates of the continuous and dispersed phases were 20 and 2 μL min⁻¹, respectively, unless noted otherwise. The generated particles were recovered through a Teflon tube (inner diameter 0.5 mm) into a glass vial containing distilled water. Particles were then washed trice with methanol via gentle centrifugation to remove any remaining polar solvent. Finally, the obtained particles were suspended in distilled water or dried, and their morphologies were observed using an inverted optical/fluorescence microscope (IX71, Olympus Corp., Japan) equipped with a CCD camera (DP72, Olympus) or by a scanning electron microscope (SEM; VE8800, Keyence Corp., Tokyo, Japan). For observation by SEM, particles were deposited onto a glass slide and their surfaces were coated with a thin Au layer by sputtering. The average diameters of the particles were calculated by measuring the longest axes for at least 30 particles for each condition. In the case of non-spherical particle having a concavity, the volume was estimated by regarding the particle shape as one of two parts of a sphere separated by a flat plane. To produce fluorescent PS particles, ethyl acetate with 5% (w/v) PS and 0.3% (w/v) pyrene was used as the dispersed phase.

Results and discussion

Generation of non-equilibrium droplets

There are several previous reports regarding the generation of non-equilibrium droplets in microfluidic devices using polar solvents as the dispersed phase. For example, o/w droplets of dimethyl carbonate (DMC) and ethyl acetate (EA) have been generated and applied to the production of polymeric^28,34 and silica particles,³⁸ and w/o droplets of EA and methyl acetate have been used to produce polysaccharide microbeads.³⁶ However, as far as we are aware, other types of polar solvents have not been used for generating non-equilibrium droplets. Furthermore, hydrophilic glass microchannels, not PDMS channels, have been used to produce non-equilibrium droplets, with only one exception employing a PDMS channel hydrophilized by using PVA.²⁸ Consequently, we first examined several types of water-soluble polar solvents to determine if they can form non-equilibrium droplets in PDMS microdevices hydrophilized by O₂ plasma treatment. The results are shown in Table 1 and Fig. 2. Some of the solvents including MA, ethyl formate, EA (Fig. 2(b)), and PA were found to form monodisperse o/w droplets that gradually dissolved into the continuous water phase, diminished, and then finally disappeared upon flowing through the microchannel. On the other hand, we did not observe the formation of o/w droplets when solvents such as acetone, acetonitrile, isopropanol, 2-butanol, and dimethyl sulfoxide (Fig. 2(a)) were used as the dispersed phase.

Table 1 Formation of non-equilibrium droplets in microfluidic devices

Organic solvent	Solubility in water at 25 °C [g/100 mL]	Formation of non-equilibrium droplets
a Miscible at any ratio.
Acetone	Miscible^a	No
Methanol	Miscible^a	No
Dimethyl sulfoxide	Miscible^a	No
2-Butanol	26	No
Methyl acetate	24.4	Yes
Ethyl formate	10.5	Yes
Ethyl acetate	8.3	Yes
Propyl acetate	1.9	Yes

---

Fig. 2 Formation of non-equilibrium droplets in the flow-focusing microchannel. (a) Dimethyl sulfoxide and (b) ethyl acetate were used as the dispersed phase.

From these results, it was presumed that the formation of droplets would be impossible in the case of polar solvents with a high solubility in water that do not form an interface, such as acetone and methanol. Moreover, the thin regions near the interface of the continuous phase and the non-equilibrium droplet would be in an equilibrium state. The formation of droplets would, therefore, be dependent on the interfacial tension between the two phases. We measured the interfacial tensions between polar solvents and water in equilibrium states; the values for the EA–water and 2-butanol–water interfaces measured as 4.4 and 0.7 mN m⁻¹, respectively, by using a tensiometer (CBVP-Z, Kyowa Interfacial Science). This difference is most likely the reason why 2-butanol did not form droplets in water.

Synthesis of PS particles

Next, we investigated the formation behavior of PS particles in the microchannel. Initially, PS with an average molecular weight of 114 700 was used, this being dissolved in EA at a concentration of 1% (w/v) and continuously introduced into the microchannel from Inlet 2 as the dispersed phase (Fig. 3(a)). As shown in Fig. 3(b), monodisperse droplets with a uniform diameter (∼55 μm) were generated at the confluence point. The droplet volume gradually decreased because EA, which has a solubility in water at 8%, dissolved into the continuous phase. This shrinkage caused solid polymers to start precipitating in the droplet when the PS concentration in the droplet reached saturation (Fig. 3(c)). At this point, the droplet volume was ∼50% of the initial value, indicating that the PS concentration in the dispersed phase was only ∼2%. However, the precipitation of solid PS particles, most likely due to the absorption of water molecules from the continuous phase by EA, resulted in a decrease in the solubility of PS in the dispersed phase. Several small PS particles were precipitated in a single droplet, which were gradually united into the one PS particle as the droplet dissolution proceeded (Fig. 3(d)). This precipitated polymer particle was accompanied by one droplet of the EA-rich phase (Fig. 3(e)), which eventually disappeared. After collecting the particles and washing with methanol, monodisperse polymer particles were obtained with an average diameter of ∼15 μm (CV value less than 5%), as shown in Fig. 4(a). From this result, we confirmed that the process presented is capable of producing polymeric particles of sizes much smaller than that of the initial droplets. Interestingly, the shape of the obtained particles was found to be not perfectly spherical, instead having a hemispherical concavity. This shape was clearly caused by the bispherical shape of the polymeric particle with an attached EA-rich droplet, which was temporarily formed in the continuous phase (Fig. 3(e)).

Fig. 3 Generation of PS droplets/particles in the microchannel. (a) Microchannel design used in this study. (b–e) Droplets flowing through the microchannel at points indicated in (a). The corresponding retention times were (b) 0.03 s, (c) 0.4 s, (d) 1.8 s, and (e) 3.0 s, respectively.

Fig. 4 (a–e) PS particles obtained by using (a) 1%, (b) 2%, (c) 3%, (d) 4%, and (e) 5% PS (M_w = 114 700) in EA as the dispersed phase. (f) PS particles obtained from 1% PS in EA, with the flow rates of the continuous and the dispersed phases of 70 and 2 μL min⁻¹, respectively. (g) Relationship between the initial droplet size and the particle's diameter with changing PS concentration. (h) Relationship between the initial PS concentration and the volume ratio of the particles to the initial droplets.

To produce single micrometer-sized particles from monomer droplets using microfluidic devices, single micrometer-sized microchannels should be employed;⁸ however, this introduces several problems concerning channel clogging and the high pressure required to introduce the requisite fluid flows. In contrast, the presented protocol is capable of producing particles as small as several micrometers (Fig. 4(f); average particle diameter of 4.4 μm) using 50 μm sized channels, and hence is highly advantageous by comparison. In addition, the polymerization process that is generally required in most conventional microfluidic techniques is dispensed with, thus greatly simplifying the experimental processes.

Effect of polymer concentration on particle size and morphology

Next we examined the effect of PS concentration on the size and the morphology of the particles. As shown in Fig. 4(b–g), the particle size was gradually increased with an increase in the PS concentration, although the initial droplet size remained almost constant. This relationship between the initial PS concentration and the volume ratio of the polymeric particles to the initial droplets is shown in Fig. 4(h). From this, we can see that the volume ratio of the particle was ∼1.8 times greater than the initial PS concentration under all these conditions. This result indicates that the polymer matrix of the particles would be porous, which is most likely caused by the removal of the solvent from the precipitated polymer after washing.

From the SEM observation, it is apparent that the polymer concentration critically affected the shape of the obtained particles (Fig. 4(b–e)); the size of the hemispherical concavity decreasing with an increase in the PS concentration. The most probable reason for this difference would be the stability of the polymer particle-encapsulating configuration of the EA-rich dispersed phase (Fig. 5). Note that ethyl acetate can dissolve water up to a concentration of 3%, and thus the non-equilibrium droplet would dissolve water within it from the continuous phase during the dissolution process. At lower PS concentration conditions, a relatively large amount of water could conceivably be dissolved in the droplets (Fig. 5(a-I)). As a result, the configuration of the droplet covering the inner PS particle would be unstable, and therefore easily turned into a twin morphology composed of a PS particle and an attached EA-rich droplet (Fig. 5(a-IV)). The shape of the particle–droplet complex at the time of this translocation and segregation determines the shape of the final particles, producing particles having a large crater-like concavity. On the other hand, under higher PS concentration conditions, the amount of water dissolved into the EA-rich phase would be reduced, possibly stabilizing the PS particle-covering configuration of the EA-rich phase (Fig. 5(b)). Consequently, the EA-rich phase is dissolved into the continuous phase covering the PS particle (Fig. 5(b-IV)), thus resulting in the formation of an almost completely spherical particle (Fig. 4(c and d)). In order to validate this presumption, we measured the water solubility in EA of different amounts of PS molecules. Distilled water was gradually added to PS solutions in EA, and the formation of precipitation was observed. From this, the solubilities of water in EA with 1% and 5% PS were found to be 2.3% (v/v) and 1.5% (v/v), respectively, which supports our hypothesis regarding the influence of PS concentration on particle morphology.

Fig. 5 Mechanism of forming particles with different morphologies under conditions of (a) low and (b) high PS concentration.

Effect of the polymer's molecular weight

The molecular weight of a polymer typically has a critical effect on its characteristics, including its solubility in solvents, viscosity of polymer solutions, and the physical properties of solid particles. To examine the effect of altering the molecular weight on the particle morphology, we employed PS molecules of different molecular weights (average M_w of 13 200 and 983 400) at a fixed PS concentration in EA of 1%. Other experimental conditions were the same as those described in the preceding sections (Fig. 3). The particles obtained are shown in Fig. 6(a and b). This shows that when the molecular weight of PS was 13 200, particles were obtained with a significantly large cavity in the core and multiple small craters on their surface. On the other hand, a higher molecular weight of PS (983 400) resulted in the formation of almost perfectly spherical particles. These results clearly indicate that the particle shape is also significantly affected by the molecular weight of the polymer.

Fig. 6 Effects of the polymer molecular weight and solvent on the particle morphology. (a and b) PS particles prepared using different molecular-weight polymers; (a) M_w of 13 200, and (b) M_w of 983 400. (c and d) PS particles prepared by using (c) methyl acetate (MA) and (d) propyl acetate (PA) as the polar solvent.

To elucidate the mechanism behind this difference in the particle morphology, we measured the solubility of water molecules in EA containing 1% PS. As a result, we found water solubilities of 2.7% and 1.2%, for PS with molecular weights of 13 200 and 983 400, respectively. This result indicates that the formation of particles with spherical or non-spherical morphologies can also be attributed to the stability of the three-phase state, as shown in Fig. 5. Similar morphologies in particles resulting from a three-phase system have been previously reported, wherein the interfacial tension and contact angle between liquid phases were controlled by changing the composition of the droplets.³⁹ Moreover, the solubility in EA and the diffusion speed of the low molecular weight PS are higher than those of the high molecular weight PS. This may have also affected the particle-covering or attaching configuration of the dispersed phase, and thus the segregation of the solid particle.

Effect of solvent

The solubility of the dispersed phase in water and speed of the droplet dissolution would be changed if different solvents were used. We therefore employed methyl acetate (MA) and propyl acetate (PA) as the polar solvent, which can be dissolved in water at 25% and 1.9%, respectively, and investigated the effect of each on the particle morphology. The concentration of PS (M_w: 114 700) was fixed at 1%. The flow rates of the continuous and dispersed phases were set at 40 and 4 μL min⁻¹ for MA, and 20 and 10 μL min⁻¹ for PA.

When MA was used, the dissolution speed of the droplets was found to be significantly faster than in the case of EA (data not shown). As a result, we obtained rock-like, non-spherical particles with a rugged surface (Fig. 6(c)). This shape would likely be generated by the coalescence of multiple small PS particles during droplet dissolution; a significantly fast dissolution of MA may result in insufficient recovery of the spherical shape of the aggregated small PS particles. In contrast, perfectly spherical particles with smooth surfaces were obtained when PA was used (Fig. 6(d)). The relatively-low dissolution speed of PA into the continuous phase, together with the low solubility of water molecules into the droplet, would contribute to the formation of this spherical shape. The volume ratio of the particles to the initial droplets was ∼1.4 times greater than the initial PS concentration for particles prepared using propyl acetate, indicating that the inner structure of the polymer matrix was affected by the polar solvent used. These results clearly show that the characteristics of the polar solvent, in particular the solubility in water, are the critical factors affecting the particle morphology.

Effect of additives

The formation mechanism of particles in the presented scheme is the sequential combination of droplet dissolution, polymer precipitation, coalescence of small polymer particles, and the translocation of the remaining droplet and the particle; thus resulting in the formation of spherical and non-spherical particles depending on the characteristics of the solvent and the polymer. We next investigated the effect of additives that are dissolvable in the polar solvent, but not in water, and which do not form a solid phase. Such additives are expected to influence the translocation of the solid/liquid phases and the coalescence of the small solid particles. We tested oleic acid (OA) as a water-insoluble additive, which was incorporated into the dispersed phase of EA.

The particles obtained are shown in Fig. 7. When the PS (M_w: 114 700) and OA concentrations were 1% (w/v) and 1% (v/v), respectively, PS particles were obtained with a highly-unique morphology, consisting of one smooth hemisphere, and one covered with multiple satellite particles (Fig. 7(a)). These satellites were stably attached to the main particles, and were not detached even after three washes with methanol via centrifugation. It was presumed that these satellites were generated after the translocation of the solid and the solvent phases, subsequently being attached onto the surface of the main particle. When the PS concentration was increased to 5% (w/v), the satellite particles were not formed, but the hemisphere of the main particle was wrinkled, as shown in Fig. 7(b).

Fig. 7 Complex PS particles prepared by the addition of oleic acid (OA) into the dispersed phase: (a) 1% PS and 1% OA, (b) 5% PS and 1% OA, (c) 1% PS and 0.1% OA, and (d) 1% PS and 0.2% OA.

The OA concentration was then reduced to 0.1% or 0.2% (v/v), whereas the PS concentration remained fixed at 1% (w/v). Under these conditions, satellite particles were formed on the surface of the main particles (Fig. 7(c and d)), much as in the case of 1% OA (Fig. 7(a)). However, these satellites were deposited on only a limited area of the main particle as a closely packed, crown-like aggregate. The number of satellite particles formed with an OA concentration of 0.1% was also observed to be smaller than that with an OA concentration of 0.2%. These results indicate that the relative volume of OA is closely related to the covering area of the main particle by an OA-rich droplet and the depositing area of the satellites.

Production of particles with different compositions

One of the remarkable advantages of the method presented is its ability to incorporate functional hydrophobic molecules into the polymer matrix by simply adding them into the dispersed phase. To demonstrate this concept, we prepared fluorescent particles by adding pyrene into the dispersed phase of EA with 5% PS. The particles obtained by this method are shown in Fig. 8(a). The surface of the fluorescence particles was slightly wrinkled probably because of the incorporated pyrene (Fig. 8(b)). It can clearly be seen that we successfully obtained monodisperse fluorescent particles with a strong blue-color emission, having an average size of ∼19 μm. Since the relative intensity of the first to third vibronic peak of pyrene fluorescence, (I₁/I₃), is known to reflect the microenvironmental polarity, the emission spectrum of the obtained beads was measured by a fluorescence spectrophotometer (F4010, Hitachi, Japan) with the excitation at 323 nm. The I₁/I₃ value was measured to be 0.89, which is closer to that observed in hydrophobic solvents like p-xylene (0.95) and toluene (1.02) than in polar solvents like EA (1.37) and water (1.87).⁴⁰ The pyrene is therefore considered to be completely encapsulated inside the PS matrix in view of both its low solubility in water and the aforementioned fluorescence result.

Fig. 8 (a and b) Fluorescence micrograph (a) and SEM image (b) of PS particles produced using EA with 5% (w/v) PS and 0.3% (w/v) pyrene as the dispersed phase. (c) Produced PMMA particles. (d) Relationship between the initial PMMA concentration and the volume ratio of the particles to the initial droplets.

Finally, we demonstrated the preparation of particles composed of a different polymer (PMMA). Fig. 8(c) shows particles made from a 5% PMMA solution in EA, resulting in an average diameter of ∼16 μm. These particles did not exhibited cavities on the surface; however, as in the case of PS particles, those made from a 1% PMMA solution did have a cavity on the surface that is likely to also be produced by the translocation of the generated particle and dissolving solvent (data not shown). The relationship between the PMMA concentration and the volume ratio of the particles to the initial droplets is shown in Fig. 8(d), indicating that the volume ratio of the particle was slightly (∼1.15) times greater than the initial PMMA concentration. These results clearly demonstrate that the presented mechanism can be applied to the formation of a variety of particles, composed of various different materials. Furthermore, particles composed of multiple polymer types could be obtained by using a polymer mixture as the precursor.

Conclusions

A highly versatile process to prepare monodisperse polymeric particles has been demonstrated using droplets in a non-equilibrium state, which are temporarily formed and rapidly shrinking in microchannels. This process is applicable to polymer molecules that are dissolvable in the polar solvent but not in water. Particles that are significantly smaller than the initial droplets are formed, owing to the dissolution of the droplets. In addition, a preparatory polymerization process is not required, thus enabling control over both the particle size and composition at the same time. We successfully obtained PS and PMMA particles 4–20 μm in size by using a 50 μm wide microchannel. Moreover, non-spherical particles could also be easily obtained, with this shape being associated with the rapid phase separation and autonomous translocation of the solid/liquid phases within the dispersed phase. Although the mechanism of forming particles with complex morphologies has not been fully elucidated, the presented technique is nonetheless highly unique and advantageous. Consequently, it is likely to become a promising method for the preparation of various forms and types of monodisperse micrometer-sized polymeric particles.

Acknowledgements

This study was supported in part by Grants-in-aid for Scientific Research (21800009 and 23655063) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. We wish to thank Mr Y. Sasaki at Chiba University for his technical assistance in the molecular weight measurement and spectrofluorometry.

Notes and references

1. R. Arshady, Colloid Polym. Sci., 1992, 270, 717.
2. T. Fujibayashi, Y. Komatsu, N. Konishi, H. Yamori and M. Okubo, Ind. Eng. Chem. Res., 2008, 47, 6445.
3. T. Nakashima, M. Shimizu and M. Kukizaki, Adv. Drug Delivery Rev., 2000, 45, 47.
4. M. Yasuda, T. Goda, H. Ogino, W. R. Glomm and H. Takayanagi, J. Colloid Interface Sci., 2010, 349, 392.
5. C. Esen, T. Kaiser, M. A. Borchers and G. Schweiger, Colloid Polym. Sci., 1997, 275, 131.
6. S. Y. Teh, R. Lin, L. H. Hung and A. P. Lee, Lab Chip, 2008, 8, 198.
7. J. T. Wang, J. Wang and J. J. Han, Small, 2011, 7, 1728.
8. S. Sugiura, M. Nakajima, H. Itou and M. Seki, Macromol. Rapid Commun., 2001, 22, 773.
9. T. Nisisako, T. Torii and T. Higuchi, Chem. Eng. J., 2004, 101, 23.
10. M. Yamada, S. Doi, H. Maenaka, M. Yasuda and M. Seki, J. Colloid Interface Sci., 2008, 321, 401.
11. M. Seo, Z. Nie, S. Xu, M. Mok, P. C. Lewis, R. Graham and E. Kumacheva, Langmuir, 2005, 21, 11614.
12. S. Xu, Z. Nie, M. Seo, P. Lewis, E. Kumacheva, H. A. Stone, P. Garstecki, D. B. Weibel, I. Gitlin and G. M. Whitesides, Angew. Chem., Int. Ed., 2005, 44, 724.
13. D. K. Hwang, D. Dendukuri and P. S. Doyle, Lab Chip, 2008, 8, 1640.
14. T. Nisisako and T. Hatsuzawa, Microfluid. Nanofluid., 2009, 9, 427.
15. S. Lone, S. H. Kim, S. W. Nam, S. Park, J. Joo and I. W. Cheong, Chem. Commun., 2011, 47, 2634.
16. T. Nisisako, T. Torii, T. Takahashi and Y. Takizawa, Adv. Mater., 2006, 18, 1152.
17. Z. Nie, W. Li, M. Seo, S. Xu and E. Kumacheva, J. Am. Chem. Soc., 2006, 128, 9408.
18. S. H. Kim, S. J. Jeon, W. C. Jeong, H. S. Park and S. M. Yang, Adv. Mater., 2008, 20, 4129.
19. S. Seiffert, M. B. Romanowsky and D. A. Weitz, Langmuir, 2010, 26, 14842.
20. S. N. Yin, C. F. Wang, Z. Y. Yu, J. Wang, S. S. Liu and S. Chen, Adv. Mater., 2011, 23, 2915.
21. J. Lee and J. Kim, Chem. Mater., 2012, 24, 2817.
22. Y. Pan, M. Ju, C. Wang, L. Zhang and N. Xu, Chem. Commun., 2010, 46, 3732.
23. D. C. Pregibon, M. Toner and P. S. Doyle, Science, 2007, 315, 1393.
24. K. W. Bong, S. C. Chapin and P. S. Doyle, Langmuir, 2010, 26, 8008.
25. Q. Zhang, B. Lin and J. Qin, Microfluid. Nanofluid., 2012, 12, 33.
26. J. Zhu and R. C. Hayward, Angew. Chem., Int. Ed., 2008, 47, 2113.
27. B. Wang, H. C. Shum and D. A. Weitz, ChemPhysChem, 2009, 10, 641.
28. L. H. Hung, S. Y. Teh, J. Jester and A. P. Lee, Lab Chip, 2010, 10, 1820.
29. A. J. C. Kuehne and D. A. Weitz, Chem. Commun., 2011, 47, 12379.
30. G. T. Vladisavljevic, W. J. Duncanson, H. C. Shum and D. A. Weitz, Langmuir, 2012, 28, 12948.
31. S. Liu, R. Deng, W. Li and J. Zhu, Adv. Funct. Mater., 2012, 22, 1692.
32. P. Zhang, Y. He, Z. Ruan, F. F. Chen and J. Yang, J. Colloid Interface Sci., 2012, 385, 8.
33. J. Jeong, E. Um, J. K. Park and M. W. Kim, RSC Adv., 2013, 3, 11801.
34. T. Watanabe, T. Ono and Y. Kimura, Soft Matter, 2011, 7, 9894.
35. E. Rondeau and J. J. Cooper-White, Langmuir, 2008, 24, 6937.
36. S. Sugaya, M. Yamada, A. Hori and M. Seki, Biomicrofluidics, 2013, 7, 054120.
37. D. C. Duffy, J. C. McDonald, O. J. A. Schueller and G. M. Whitesides, Anal. Chem., 1998, 70, 4974.
38. A. Fang, C. Gaillard and J. P. Douliez, Chem. Mater., 2011, 23, 4660.
39. W. Wang, M. J. Zhang, R. Xie, X. J. Ju, C. Yang, C. L. Mou, D. A. Weitz and L. Y. Chu, Angew. Chem., Int. Ed., 2013, 52, 8084.
40. D. C. Dong and M. A. Winnik, Can. J. Chem., 1984, 62, 2560.

---