Abstract
All-aqueous systems have attracted intensive attention as a promising platform for applications in cell separation, protein partitioning, and DNA extraction, due to their selective separation capability, rapid mass transfer, and good biocompatibility. Reliable generation of all-aqueous droplets with accurate control over their size and size distribution is vital to meet the increasingly growing demands in emulsion-based applications. However, the ultra-low interfacial tension and large effective interfacial thickness of the water–water interface pose challenges for the generation and stabilization of uniform all-aqueous droplets, respectively. Microfluidics technology has emerged as a versatile platform for the precision generation of all-aqueous droplets with improved stability. This review aims to systematize the controllable generation of all-aqueous droplets and summarize various strategies to improve their stability with microfluidics. We first provide a comprehensive review on the recent progress of all-aqueous droplets generation with microfluidics by detailing the properties of all-aqueous systems, mechanisms of droplet formation, active and passive methods for droplet generation, and the property of droplets. We then review the various strategies used to improve the stability of all-aqueous droplets and discuss the fabrication of biomaterials using all-aqueous droplets as liquid templates. We envision that this review will benefit the future development of all-aqueous droplet generation and its applications in developing biomaterials, which will be useful for researchers working in the field of all-aqueous systems and those who are new and interested in the field.
I. INTRODUCTION
Aqueous two-phase systems (ATPSs) are formed when two incompatible additives are mixed in water and undergo phase separation above a threshold concentration.1,2 ATPSs have attracted increasing attention and have been considered as a promising candidate for applications in cell separation,3 protein partitioning,4,5 and DNA extraction,6,7 due to their selective separation capability, rapid mass transfer, and good biocompatibility. All-aqueous emulsion droplets have been exploited for cell encapsulation,8 bioreactors,9 biomaterials fabrication,10,11 and delivery of biomolecules.12 Reliable generation of all-aqueous emulsion droplets with accurate control over their size and size distribution is of vital importance to meet the increasingly growing demands in emulsion-based applications.
Droplet microfluidics technology is capable of producing highly uniform droplets with tunable size, structure, and composition. It can effectively tailor the structure of the fabricated biomaterials (such as microparticles,13 micromotors,13–15 and microfibers16–18) and endow them with multiple functionalities.15 Therefore, researchers have recently begun to explore how ATPSs can be combined with microfluidic technologies. Scaling down ATPSs to microscale offers a new way for molecular purification with several advantages over macroscopic processes, such as rapid diffusion, increased controllability of individual flows and droplets, and online monitoring of the separation process. Microfluidic implementations of ATPSs for biomaterial separation and purification are usually based on all-aqueous emulsion droplets generated in microdevices.19–21 The all-aqueous emulsion droplets generated with the microfluidic technology have been exploited for cell encapsulation,22–27 enzyme encapsulation,28,29 protein purification,30 protein extraction,31 protein partitioning,32–35 protein encapsulation,36 recovery of biotechnological products,37 and DNA extraction.38 Therefore, the generation and stabilization of all-aqueous emulsion droplets are of utmost technological importance for practical applications. However, the ultra-low interfacial tension and large effective interfacial thickness of the water–water interface pose challenges for the generation and stabilization of uniform all-aqueous emulsion droplets, respectively. Various strategies have been proposed to facilitate the all-aqueous emulsion droplets generation and enhance their stability in recent years.
This review aims to introduce in detail the formation mechanism and controlling methods for all-aqueous emulsion droplets generation and summarize various strategies to improve the stability of all-aqueous emulsion droplets. In this review, we start by first introducing the properties of all-aqueous systems (including the bulk aqueous phases and the water–water interface). Then, we summarize in detail the recently developed techniques for producing all-aqueous emulsion droplets (single, core-shell, Janus, and other complex structures) with the microfluidics technology in terms of formation mechanism, microdevice configurations, and control methods. Finally, we summarize the various strategies for enhancing the stability of the all-aqueous emulsion droplets and the templated all-aqueous biomaterials.
II. THEORETICAL BACKGROUND
A. Background
1. History and development
ATPSs were first reported by Beijerinck in 1896 when he observed the separation of the turbid aqueous mixture of gelatin and agar (or soluble starch) into two layers: the top layer containing most of the gelatin and the bottom layer containing most of the agar (or starch).39 The ATPSs' application was presented during 1956–1958 by Albertsson for the partition of cell particles and proteins.40,41 Since then, ATPSs' partition has been applied to various materials, such as plant and animal cells, microorganisms, viruses, chloroplasts, mitochondria, membrane vesicles, proteins, and DNA, by exploiting the preferential distribution of materials between phases, a phenomenon known as partitioning (Fig. 1).19,42
FIG. 1.
The development of ATPS-based technologies.19
2. Types of aqueous two-phase system
ATPSs are formed by phase separation of two incompatible additives mixed in water when their concentration exceeds some threshold value.2 Many types of incompatible additives are capable of producing ATPSs, as summarized in Table I, of which the most widely explored are polymer–polymer and polymer–salt ATPSs.43–46 ATPSs consist of three regions: two bulk phases separated by density into a top phase and bottom phase, and the interface.
TABLE I.
Commonly used aqueous two-phase systems (ATPSs) formed with biocompatible additives.43–46 In each row, solute A and solute B represent a pair of incompatible solutes.
| Advantages and disadvantages | Type of ATPSs | Solute A | Solute B |
|---|---|---|---|
| Polymer–polymer ATPSs are preferably used for the separation, recovery, and purification solutes sensitive to the ionic environment as these systems pose low ionic strength. | Polymer–polymer ATPSs | Polyethylene glycol (PEG) | Dextran |
| Polyethylene glycol (PEG) | Poly (vinyl methyl ethyl ether) | ||
| Polyethylene glycol (PEG) | Polyvinyl alcohol (PVA) | ||
| Polyethylene glycol (PEG) | Polyvinylpyrrolidone (PVP) | ||
| Polyvinylpyrrolidone | Dextran | ||
| Ficoll | Dextran | ||
| Dextran | Poly (ethylene oxide) (PEO) | ||
| Dextran | Polyvinyl alcohol (PVA) | ||
| Dextran | Polyvinylpyrrolidone (PVP) | ||
| Dextran | Gelatin | ||
| Poly (ethylene oxide) (PEO) | Pullulan | ||
| Starch | Gelatin | ||
| Starch | Locust bean gum | ||
| Gelatin | Maltodextrin | ||
| Amylopectin | Xyloglucan | ||
| Sodium caseinate | Xanthan | ||
| Casein | Xanthan | ||
| High ionic strength is the only disadvantage of polymer–salt ATPSs. | Polymer–salt ATPSs | Polyethylene glycol (PEG) | (NH4)2SO4 |
| Polyethylene glycol (PEG) | Na2SO4 | ||
| Polyethylene glycol (PEG) | K2HPO4/KH2PO4 | ||
| Polyethylene glycol (PEG) | Na2CO3 citrate | ||
| Polyethylene glycol (PEG) | Tripotassium phosphate (K3PO4) | ||
| Polyethylene glycol (PEG) | Anhydrous dipotassium hydrogen phosphate (K2HPO4) | ||
| Polyethylene glycol (PEG) | Sodium citrate or sodium sulfate | ||
| Dextran | Methylcellulose | ||
| Polyvinyl alcohol (PVA) | Methylcellulose | ||
| Polyvinylpyrrolidone (PVP) | Methylcellulose | ||
| Methylcellulose | Na2HPO4 | ||
| Ammonium sulfate (AS) | Tetrabutylammonium bromide (TBAB) | ||
| Inexpensive as compared to polymers and co-polymers; low viscosity, easy constitute recovery, and reduced setting times. | Alcohol–salt ATPSs | Ethanol | K2HPO4/KH2PO4 |
| Ethanol | Na2SO4 | ||
| 1-Propanol | K2HPO4/KH2PO4 | ||
| 2-Methyl-2-propanol | K2HPO4/KH2PO4 | ||
| 2-propanol | K2HPO4 | ||
| A major drawback: many proteins are not compatible with alcohol-rich phase. | |||
| Ionic liquid-based ATPSs | 1-Butyl-3-methylimidazolium chloride | K2HPO4/KH2PO4 | |
| 1-Butyl-3-methylimidazolium chloride | K2CO3 | ||
| 1-Butyl-3-methylimidazolium BF4 | Na3C6H5O7/Na2C4H4O6 | ||
| 1-Butyl-3-methylimidazolium BF4 | K2CO3 | ||
| Useful for ionic environment sensitive solutes as nonionic surfactants can be used for the formation of these systems. | Micellar/reverse micellar ATPSs | Octylphenol ethoxylate | Nonionic/ionic surfactants |
| n-Decyl tetra(ethylene oxide) | Salts and/or low-molecular-weight solvents |
B. Thermodynamics of segregative phase separation
Phase separation can be explained by the thermodynamic properties of incompatible solutions, which leads to the formation of immiscible liquid phases. To further understand the thermodynamic process of ATPSs' phase separation and the final phase structure of ATPSs, the Gibbs free energy of the ATPSs should be considered
| (1) |
where △G, △H, and △S are the change in Gibbs free energy, enthalpy, and entropy between the mixed and phase-separated states, respectively. T represents the temperature. If, △G < 0, the mixing process is spontaneous, which means that ATPSs cannot form. Otherwise, when △G > 0, the contribution of entropy is less than the contribution of enthalpy, such that phase separation occurs and ATPSs form. Thermodynamically, the phase separation of ATPSs occurs when the entropic contribution that favors mixing becomes less than the contribution of enthalpy that opposes mixing.19,42
C. Phase diagram
The phase separation behavior of ATPSs can be described by a phase diagram, as shown in Fig. 2(a), which includes a binodal curve [TCB in Fig. 2(a)], tie lines [TB in Fig. 2(a)], volume ratio, and critical point C.19,20,42 Phase diagram provides a set of information like concentration of components for phase separation and their concentration in the top and bottom phases in equilibrium, with the concentration of the top-rich phase on the ordinate axis and the concentration of the bottom-rich phase on the abscissa axis. The binodal phase diagram can be determined by turbidimetric titration method, cloud-point titration method, node determination method,44 microfluidic platform,47 and 96-well microplate-based titration technique.48
FIG. 2.
ATPS properties. (a) Phase diagram of a typical ATPS. Concentrations above the binodal curve (TCB) form an aqueous two-phase system. (b) Images showing the three different states of polymer mixtures: supercritical (biphasic) state, critical state, and subcritical (monophasic) state. Reproduced with permission from Teixeira et al., Adv. Healthcare Mater. 7, 1701036 (2018). Copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.19 (c) A phase-separated ATPS from dextran and gelatin in a 1.2-cm-wide tube. The water–water interface is indicated by an arrow, and the spatial distribution of dextran and gelatin near the water–water interface is illustrated (inset). Reproduced with permission from Y. Chao and H. C. Shum, Chem. Soc. Rev. 49, 114 (2020). Copyright 2019 Royal Society of Chemistry.56
The curve separating the monophasic regime from the biphasic regime is known as the binodal curve. The binodal curve (TCB) represents the threshold concentrations for phase separation. When the total system composition lies below the binodal curve, the aqueous mixture remains as a single homogeneous phase as visualized on the macroscopic scale, while for compositions above the binodal curve, the aqueous mixture spontaneously separates into two immiscible phases [Figs. 2(a) and 2(b)]. On the binodal curve, the chemical potentials of the two polymers are the same in both phases.49
The straight lines (TB) in the diagram are the tie lines. The final constituents of the top-rich and the bottom-rich phase in equilibrium are given by the two end points (point T and B, respectively) of the straight tie lines. Point B represents the composition of the bottom phase, and point T represents the composition of the top phase of an ATPS. All ATPSs on the same tie-line [points 1–3 in Fig. 2(a)] have the same equilibrium compositions of “T” and “B” for the top phase and the bottom phase, respectively, while only the volume ratio between the two immiscible phases changes depending on the overall system composition.50 The longer the tie line is, the larger the difference between the two-phase compositions becomes, among many other physical properties, such as viscosity, surface tension, hydrophobicity, ionic strength, and overall charge.20
The volume ratio of the top and the bottom phases of the ATPS is: VR = , where VTP and VBP are the volume of the top and bottom phases, respectively.19
Point “C” on the binodal curve is the critical point, also called the plait point. When the system is closer to point “C,” the difference between the two phases gradually vanishes. At the critical point “C,” the length of the tie line equals zero, suggesting that compositions and volumes of the two phases theoretically become equal.42,50
D. Partitioning
Partitioning occurs when molecules or particles are mixed into the polymeric system, but does not contribute to phase separation. After equilibration, these materials distribute to the phase for which they have the greatest relative affinity.19,42 The partition of component A in ATPSs is generally described using the partition coefficient (K), defined as the ratio of the concentration of component A in the top and the bottom phase of the ATPS44,45
| (2) |
where are the concentration of component A in the top and bottom phases at equilibrium, respectively. The partition behavior in ATPS can be categorized into four groups: size dependent, electrochemical dependent, hydrophobicity dependent, and specific bioaffinity dependent.44,45
E. Physicochemical properties of the water–water interface
As water is the main component of both phases (in the range between 85% and 99%), ATPSs provide a gentle environment for biomolecules separation and purification, which has been used to avoid the deficiencies of the traditional water–organic system.42 Due to its all-aqueous nature, the water–water interface is featured with its ultralow interfacial tension and large effective thickness.50
The interfacial tension of ATPSs is typically in the range of 10−4–10−1 mN m−1,49,51,52 which is comparable to or slightly higher than that in pure lipid bilayer membranes (∼2 × 10−5–10−3 mN m−1),53 but several orders of magnitude lower than that in the water/oil systems (1–40 mN m−1)51 and that in lipid bilayer membranes in the presence of organic solvents (0.2–7.25 mN m−1).54 Therefore, we treated ATPSs as ultralow interfacial tension systems.
The thickness of the water–water (W/W) interface is theoretically calculated to be tens to hundreds of nanometers [Fig. 2(c)],55,56 much thicker than the oil–water interfaces. The length scale of the water–water interface is larger than that of the polymer. The thick water–water interface of ATPSs results in an ill-defined interface and non-barrier for small hydrophilic molecules when they move from one phase to another. Hence, small molecules could not adsorb on the water–water interface, affecting the stability of ATPS emulsion droplets. The poor stability is the main limitation for the practical application of ATPS emulsion droplets.49
III. MICROFLUIDIC GENERATION OF ALL-AQUEOUS SINGLE EMULSION DROPLETS
While droplet generation is driven by interfacial tension in conventional water–oil systems, the pinch-off of ultralow surface tension fluids is dominated by bulk diffusion.57 The ultralow interfacial tension of ATPSs poses severe challenges to the spontaneous generation of all-aqueous emulsion droplets in passive means, and an extended ATPS jet is usually produced in microchannels. The ATPS jet is either stable throughout the channel or breaks up erratically into non-uniform droplets. To prevent the ATPS jet formation and produce uniform ATPS droplets in a dripping manner, the capillary number should satisfy Ca = μu/γ < 1, and thus, the flow velocity u must be smaller than the visco-capillary velocity uv−i (uv−i ∼ γ/μ, γ, and μ are the interfacial tension and viscosity, respectively). The condition u < γ/μ (Ca < 1) can be easily met in water–oil systems. Nevertheless, the several-order-of-magnitude reduction in γ requires a corresponding decrease in the flow velocity u for ATPS. As such, the dripping mode of all-aqueous droplet generation is limited by a very low speed of fluid flow in passive ways of all-aqueous droplet generation.51
Therefore, an external force is introduced into the ATPS to facilitate the breakup of the stable ATPS jet for droplet generation. That is why most active methods were developed for all-aqueous emulsion droplets generation in earlier years (from 2007 to 2015). External actuation makes use of additional energy input in promoting interfacial instabilities for droplet generation. For most active methods, the integration of external components makes the fabrication of microfluidic chips complicated, and the droplet generation frequency is limited by the response time. With the development of microfluidic technology and improvements in the precision of driven components, researchers have managed to find strategies that could provide slow flow rates or driven pressures, making passive preparation of all-aqueous droplets possible. Therefore, since 2015 (especially since 2018), more and more passive methods have been proposed for ATPS droplet generation.
We summarize the historical timeline of the development in Fig. 3 and the classification of microfluidic all-aqueous single emulsion droplets generation in Fig. 4. In this review, the passive method is referred to those that have a constant flow rate or pressure source, with no additional energy input except the driving pressure of the fluid injection. The active method refers to those with additional energy input that is featured by an unstable or time-dependent flow rate (or pressure source).
FIG. 3.
Historical timeline of the development in microfluidic all-aqueous single emulsion droplets generation.
FIG. 4.
Summary of microfluidic all-aqueous single emulsion droplets generation by active and passive methods. Mechanical vibration. Reproduced from A. Sauret and H. C. Shum, Appl. Phys. Lett. 100, 154106 (2012), with the permission of AIP Publishing.59 Mechanical shaking. Reproduced from Shum et al., Biomicrofluidics 6, 026501 (2012), with the permission of AIP Publishing.61 Acoustic actuation (audio speaker). Reproduced with permission from De Lora et al., ACS Appl. Bio Mater. 2, 4097 (2019). Copyright 2019 American Chemical Society.62 Rupture of interfacial fingers. Reprinted with permission from Mak et al., Langmuir 34, 926 (2018). Copyright 2017 American Chemical Society.63 Piezo-electric bending disk. Reproduced with permission from Ziemecka et al., Lab Chip 11, 620 (2011). Copyright 2011 Royal Society of Chemistry.65 Pneumatic solenoid valve. Reproduced with permission from Liu et al., Small 14, 1801095 (2018). Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.25 Braille pin actuation. Reproduced with permission from Lai et al., Lab Chip 11, 3551 (2011). Copyright 2011 Royal Society of Chemistry.67 Piezo-electric Buzzer. Reproduced with permission from Tarameshlou et al., Int. J. Polym. Mater. Polym. Biomater. 63, 884 (2014). Copyright 2014 Taylor & Francis Group, LLC.68 Pulsating inlet pressure. Reproduced with permission from Moon et al., Lab Chip 15, 2437 (2015). Copyright 2015 Royal Society of Chemistry.69 Square-wave-changing injection. Reproduced with permission from Liu et al., Biomed. Microdevices 19, 55 (2017). Copyright 2017 Springer Science Business Media New York.70 Electrohydrodynamic. Reproduced with permission from Song et al., J. Chromatogr. A 1162, 180 (2007). Copyright 2007 Elsevier B.V.72 Electrospray. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75 Oil-mediated (active pico-injection). Reproduced with permission from Nan et al., Microsyst. Nanoeng. 6, 70 (2020). Copyright 2020 Springer Nature.76 Hollow-fiber membrane emulsification. Reproduced with permission from Breisig et al., J. Membr. Sci. 467, 109 (2014). Copyright 2014 Elsevier B.V.77 Shirasu porous glass (SPG) membrane emulsification. Reproduced with permission from Akamatsu et al., Langmuir 35, 9825 (2019). Copyright 2019 American Chemical Society.78 Weak gravity-driven hydrostatic pressure. Reprinted with permission from Moon et al., Anal. Chem. 88, 3982 (2016). Copyright 2016 American Chemical Society.79 Microneedle-assisted gravity-driven hydrostatic pressure. Reprinted with permission from Jeyhani et al., J. Colloid Interface Sci. 553, 382 (2019). Copyright 2019 Elsevier, Inc.81 Pressure/flow rate-controlled. Reproduced with permission from Y. Chao and H. C. Shum, Chem. Soc. Rev. 49, 114 (2020). Copyright 2019 Royal Society of Chemistry.56 Assembled evaporation pump. Reproduced with permission from Zhang et al., J. Mater. Sci. 54, 14905 (2019). Copyright 2019 Springer.84 Fracture-based variable-width microchannel. Reproduced with permission from Choi et al., Soft Matter 15, 4647 (2019). Copyright 2019 Royal Society of Chemistry.83 Oil droplet chopper. Reproduced with permission from Zhou et al., Lab Chip 17, 3310 (2017). Copyright 2017 Royal Society of Chemistry.51 Transient double emulsion. Reproduced with permission Zhou et al., from Lab Chip 21, 2684 (2021). Copyright 2021 Royal Society of Chemistry.85 Oil-mediated (passive flow focusing). Reproduced with permission from Nan et al., Microsyst. Nanoeng. 6, 70 (2020). Copyright 2020 Springer Nature.76 Water-head-driven microfluidic oscillator. Reprinted with permission from V. B. Dang and S. J. Kim, Lab Chip 17, 286 (2017). Copyright 2017 Royal Society of Chemistry.86 interface fingering instability method. Reproduced with permission from Chao et al., Langmuir 34, 3030 (2018). Copyright 2018 American Chemical Society.87
In addition, we contrast generation methods in Table II in terms of their characteristics in producing all-aqueous single emulsion droplets, including the radius and the coefficient of variation (CV) of droplets generated, and the generation frequency. In Table III, we summarize the properties of ATPSs (including their composition and interfacial tension), the microfluidic device geometry, the driven components, and external actuation components used in generating all-aqueous single emulsion droplets. Furthermore, we compare the advantages and disadvantages of different generation methods in Table IV. Microfluidic devices with passive methods have no external components, thus simple to fabricate but normally with a narrow range of flow conditions for generating droplets. Most active methods involve the integration of external components and complex chip fabrication. Moreover, for the electrical actuation methods, the electric field may impair the bioactivity of biomolecules and cells, imposing restrictions on biological applications. In generating all-aqueous single emulsion droplets, phase-separated ATPSs (phases in equilibrium) are normally used, although single-phased ATPSs (phase not in equilibrium) have also been used to manipulate the droplet size, such as in the membrane emulsification method.58
TABLE II.
Comparison of methods for generating all-aqueous single-emulsion droplets. All the droplet sizes are measured at thermodynamic equilibrium.
| Method | Radius (μm) | CV (%) | Maximum frequency (Hz) | Year | Reference | |||
|---|---|---|---|---|---|---|---|---|
| Active methods | Mechanical perturbation | Vibrating the microtubing | Mechanical vibration method | ∼65 to ∼100 | ⋯ | ∼9 | 2012 | 59 |
| ∼30 to ∼100 | < 3 | ∼20 | 2012 | 60 | ||||
| ∼20.5 to ∼92.5 | ⋯ | ∼35 | 2019 | 28 | ||||
| Mechanical shaking method | ⋯ | ⋯ | ∼8 | 2012 | 61 | |||
| Acoustic actuation (audio speaker) | 50–125 | 18.3 | ∼60 | 2019 | 62 | |||
| Acoustic actuation (rupture of interfacial fingers) | 1.5–8.5 | < 3 | 1000 | 2018 | 63 | |||
| Varying channel dimension | Piezo-electric bending disk method | 15–30 | < 10 | ∼50 | 2011 | 65 | ||
| 12.5–37.5 | ⋯ | ∼45 | 2011 | 66 | ||||
| ⋯ | ⋯ | ⋯ | 2012 | 64 | ||||
| Pneumatic solenoid valve method | ∼50 to ∼90 | 2.58–2.98 (for 1 channel) | ∼12.5 (one channel) | 2018 | 25 | |||
| 3.5 (for 8 channel) | ∼100 (8 channels) | |||||||
| Braille pin actuation method | ∼60 to ∼93 | ⋯ | ∼2.5 | 2011 | 67 | |||
| (0.9–3.4 nL) | ||||||||
| Piezo-electric buzzer method | 9.5–56 | < 10 | 40 | 2014 | 68 | |||
| ‘On-off’ Flow Control | Pulsating inlet pressure method | 22–177 | ⋯ | O(0.1) to O (1) | 2015 | 69 | ||
| Square-wave-changing injection method | 32.5–55.5 | ⋯ | ⋯ | 2017 | 70 | |||
| Electrical actuation | Electrohydrodynamic (EHD) method | ∼58.4 | ∼6.5 | ∼5 | 2010 | 71 | ||
| 2007 | 72 | |||||||
| 2008 | 73 | |||||||
| 2019 | 74 | |||||||
| Electrospray method | ∼22.5 to 1000 | 3.6 to 6.7 | ⋯ | 2015 | 75 | |||
| Oil-mediated method (active pico-injection) | ∼17.5 to ∼38.75 | ∼2 | ∼200 to 2400 | 2020 | 76 | |||
| Passive methods | Membrane emulsification | Hollow-fiber membrane emulsification method | ∼320 to ∼610 | 0.75–3 | ⋯ | 2015 | 58 | |
| Shirasu porous glass membrane emulsification method | ∼5 to ∼25 | 10.4–10.6 | ⋯ | 2019 | 78 | |||
| Precise control of liquid pressure | Weak gravity-driven hydrostatic pressure method | ∼5 to ∼60 | Dripping <1 | ∼12.5 per channel | 2016 | 79 | ||
| 2016 | 27 | |||||||
| Jetting ∼10 | ||||||||
| ∼20 to ∼65 | ⋯ | < 0.2 for dripping; >1.5 for jetting | 2017 | 80 | ||||
| Microneedle-assisted gravity-driven hydrostatic pressure method | ∼2.5 to ∼32.5 | ∼10 | 850 | 2019 | 81 | |||
| Pressure-/flow rate-controlled method | 7 to 14 | ⋯ | ∼300 | 2018 | 82 | |||
| Fracture-based variable-width microchannel method | ∼5 to ∼50 | 5 to 20 | 2 to 20 | 2019 | 83 | |||
| Assembled evaporation pump method | ∼22 to ∼46.5 | ∼3.2% | 7 – 8 (per unit) | 2019 | 84 | |||
| ∼60 (8 units) | ||||||||
| Transient oil-medium | Oil-droplet chopper method | ∼5 to ∼180 | 0.75 to 2.45 | 2137 | 2017 | 51 | ||
| Transient double emulsion method | ∼100 to ∼250 | 0.66 to 2.55 | ∼4 to 170 | 2021 | 85 | |||
| Oil-mediated method (passive flow focusing) | ∼21.5 to ∼36.5 | ∼2 | ∼270 to 1200 | 2020 | 76 | |||
| Others | Water-head-driven microfluidic oscillator method | ∼90 to ∼160 | ⋯ | ⋯ | 2017 | 86 | ||
| Interface fingering instability method | ∼250 to ∼750 | ⋯ | ⋯ | 2018 | 87 | |||
TABLE III.
Summary of the ATPSs, device, and external components of the all-aqueous single-emulsion droplets generation methods. % w/w: weight of polymer per weight of solution; % w/v: weight of polymer per volume of solution; wt. %: weight of polymer to the weight of the solution, DEX: Dextran; PEG: Polyethylene glycol; TBAB: Tetrabutylammonium bromide; and AS: Ammonium sulfate. PSS-1:poly(sodium 4-styrene sulfonate); PDADMAC: Poly (diallyldimethyl ammonium chloride); HEMA: 2-hydroxyethyl methacrylate.
| Method | Tension (mN m-1) | ATPS | Device | Driven component | External components | Reference | |||
|---|---|---|---|---|---|---|---|---|---|
| Active methods | Mechanical perturbation | Vibrating the microtubing | Mechanical vibration | ∼0.1 | Emulsion: 17 wt. % PEG (8 kDa) | Co-flow glass microcapillary device. | Syringe pumps | Mechanical vibrator | 59 |
| Continuous: 15 wt. % DEX (500 kDa) | |||||||||
| (PASCO SF-9324, U.K.) | |||||||||
| Sinusoidal-wave generator | |||||||||
| / | Emulsion: 5 wt. % DEX (500 kDa) +1 wt. %PEG (Mw 8 kDa) | Co-flow and flow-focusing glass microcapillary device. | Syringe pumps | Mechanical vibratorSinusoidal-wave generator | 60 | ||||
| Continuous: 8 wt. % PEG (8 kDa) + 20 wt. % glycerol | |||||||||
| Non-equilibrated ATPSs | Sinusoidal-wave generator | ||||||||
| / | Emulsion: 15 wt. % DEX (500 kDa) + 0.5 wt. % PDADMAC | Co-flow and flow-focusing glass microcapillary device. | Syringe pumps | Mechanical wave generator (Mechanical vibrator (PASCO SF-9324, U.K.) | 28 | ||||
| Middle: 17 wt. % PEG (8 kDa) | |||||||||
| ATPS 1(Mainly): Emulsion: 17wt.% PEG(8 kDa); | |||||||||
| Mechanical shaking | / | Co-flow glass microcapillary microfluidic device | Syringe pumps | Orbital shaker (VWR Symphony, USA) | 61 | ||||
| Continuous: 16 wt. % | |||||||||
| DEX(500 kDa) | |||||||||
| ATPS 2: PEGDA/DEX | |||||||||
| ATPS 3: PEG/K3PO4 | |||||||||
| ATPS 4: PEGDA/K3PO4 | |||||||||
| Acoustic actuation (Audio speaker) | 0.2 | Emulsion: 5% w/w DEX (550 kDa) | Co-flow glass micro-capillary device | Syringe pumps. | An audio speaker, an amplifier, a sinusoidal waveform source (Waveform generator or smartphone) | 62 | |||
| Continuous: 17% w/w PEG (8 kDa) | |||||||||
| Acoustic actuation (rupture of interfacial fingers) | O (10−3 –10−1) | Six types of ATPSs: | Co-flow glass micro-capillary device | Syringe pump | Loud speaker (SPA2210, Philips) | 63 | |||
| PEG, Na2CO3, C6H5Na3O7, K3PO4, Dextran | |||||||||
| Varying channel dimension | Piezo-electric bending disk | ∼0.10 | Emulsion: 20% w/w DEX (110 kDa) | Two-layer flow-focusing PDMS microfluidic device | Syringe pumps | A sinusoidal AC voltage, A functional generator, A linear amplifier, Piezo-electric bending disk actuator. | 65 | ||
| Continuous: 10% w/w PEG (35 kDa). | |||||||||
| ∼0.3 | Innermost: 7% w/w PEG (10 kDa) | PDMS-on-glass microdevice consists of a series of three flow focusing junctions with a 80 cm length channel. | 66 | ||||||
| Inner: 10% w/w DEX (500 kDa) | |||||||||
| Outer: 10% w/w DEX (500 kDa) | |||||||||
| Continuous: 7% w/w PEG (10 kDa) | |||||||||
| ∼0.3 | Emulsion: 10% w/w DEX (500 kDa) | Two-layer flow-focusing PDMS device with a piezoelectric bending disk embedded. | 64 | ||||||
| Continuous: 7% w/w PEG (10 kDa). | |||||||||
| Pneumatic solenoid valve | / | Continuous: PEG (5%, 10%, and 17%, w/w) | Three layer cross-flow PDMS microfluidic device | Syringe pump. | Solenoid valves controller; | 25 | |||
| Emulsion: DEX (5%, 10%, and 15%, w/w) | |||||||||
| Pneumatic single layer membrane (SLM) valve. | |||||||||
| DEX(70 kDa; 500 kDa); | |||||||||
| PEG (8 kDa; 20 kDa). | |||||||||
| Braille pin actuation | ∼0.01 | Emulsion: 3.2 wt. % DEX (500 kDa) | Hydrodynamic PDMS focusing device with multi-level channels and narrow orifice, fabricated by backside light lithography (BDLL). | Syringe pumps | Braille pin valving. (Controlled by computer) | 67 | |||
| Continuous: 2.5 wt. % PEG (35 kDa) | |||||||||
| Piezo-electric buzzer | / | Emulsion: 80 wt. % HEMA + 20 wt. % PEGDA | 3D PMMA flow-focusing microfluidic device | Syringe pumps | A Piezoelectric speaker (piezo buzzer); | 68 | |||
| Continuous: 40 wt. % Na2SO4 + 2 wt. % NaNO2 | |||||||||
| function generator. | |||||||||
| “On-off” flow control | Pulsating inlet pressure | / | Emulsion: 5% w/v DEX (100 kDa) | Flow-focusing PDMS-glass microfluidic device. | Emulsion: a pulsating applied pressure (pipette tip, pressure regulator, three-way solenoid valve); | A LabVIEW controlled pressure regulator; a tree-way pneumatic solenoid valve. | 69 | ||
| Continuous: 10% w/v PEG (8 kDa) | |||||||||
| Continuous:Syringe pump | |||||||||
| Square-wave-changing injection method | ∼0.01 to 0.1 | Emulsion: 10 wt. % DEX (70 kDa), alginate-Ph and HRP in PBS; | A PDMS-glass hybrid microfluidic device containing three inlets, followed by a narrow outlet opening and a subsequent open space. | The flow rates of all phases are controlled by a microflow controller. | A microflow controller (OBI, Elveflow, Paris, France) | 70 | |||
| Middle: 7.5 wt. %PEG (100 kDa) in PBS; | |||||||||
| Continuous: 7.5 wt. % PEG (100 kDa) and 1 mM H2O2 in PBS. | |||||||||
| Electrical actuation | Electrohydrodynamic (EHD) method | ∼4 to 5 | Emulsion: 30% w/w AS-rich | PDMS and glass microdevice with a T-junction structure | Syringe pumps. | DC power supply | 71 | ||
| 4.73 to 5.3 | 72 | ||||||||
| 73 | |||||||||
| Continuous: 15% w/w TBAB-rich | |||||||||
| / | |||||||||
| Simulation of the ATPS droplets generation by electrohydrodynamics (EHD) method | 74 | ||||||||
| Electrospray method | ∼0.03 | Emulsion: 15 wt. % DEX (500 kDa); | A tapered glass capillary | Syringe pumps. | DC power supply; | 75 | |||
| Continuous: 8 wt. % PEG (20 kDa) | |||||||||
| Metallic needle; | |||||||||
| Metallic ring. | |||||||||
| Oil-mediated method (active pico-injection) | ∼0.116 to 0.521 | Emulsion: DEX-rich (10 kDa) with PAH | PDMS device with flow-focusing and pico-injection channel. | / | Applied electric field. | 76 | |||
| Middle-1: Pure PEG-rich (8 kDa) | |||||||||
| Middle-2: PEG-rich (8 kDa with PSS-1 | |||||||||
| Continuous: HFE 7500 oil | |||||||||
| Passive methods | Membrane emulsification | Hollow-fiber membrane emulsification method | / | Emulsion: K2HPO4 | Porous polypropylene microfiltration hollow-fiber membrane device | Syringe pumps. | None | 58 | |
| Continuous: PEG (4 kDa) | |||||||||
| Shirasu porous glass membrane emulsification method | / | DEX (15 kDa–20 kDa): 4 wt. % | Shirasu porous glass (SPG) membrane | / | None | 78 | |||
| PEG (8 kDa): 2 wt. % | |||||||||
| Both PEG-rich and DEX-rich phase can be used as inner emulsion phase or outer continuous phase. | |||||||||
| Precise control of liquid pressure | Weak gravity-driven hydrostatic pressure method | 0.037 to 0.103 | Emulsion: 16% w/v DEX-rich (500 kDa) | Flow-focusing PDMS microfluidic device | Pipette tips. | None | 79 | ||
| Continuous: 5%, 10% w/v PEG-rich (35 kDa) | |||||||||
| / | Emulsion: 12.8%, 16%, 20%, 25.6% (w/v) DEX (500 kDa) | 27 | |||||||
| Continuous: 5%,10%, 20%(w/v) PEG (35 kDa) | |||||||||
| / | Emulsion: 16% (w/v) DEX | 80 | |||||||
| Continuous: 10% (w/v) PEG | |||||||||
| Microneedle-assisted gravity-driven hydrostatic pressure method | ∼0.150 | Emulsion: 20% (w/v) DEX (500 kDa) | Metallic microneeddle inserted 3D Flow-focusing PDMS microfluidic device | Plastic syringe tubes. | None | 81 | |||
| Continuous: 10% (w/v) PEG (35 kDa) | |||||||||
| Or reverse. | |||||||||
| Pressure-/flow rate-controlled method | ∼0.103 | Emulsion: DEX (16% (w/v)) | Flow-focusing PDMS/glass hybrid microfluidic device | Pressure controller (pressure-driven flow injection). | A high-precision microfluidic pressure control system with a flow unit platform including a flow-board and two in-line flow units is needed. | 82 | |||
| Continuous: PEG (10% (w/v)) | |||||||||
| Fracture-based variable-width microchannel method | ∼0.012 to ∼0.103 | Emulsion: DEX-rich (500 kDa) (6.4%, 16%) (w/v) | PDMS device installed in a Micro-Vice | Emulsion: pressure pump; | Stretching device to apply mechanical strain. | 83 | |||
| Continuous: Syringe pump. | |||||||||
| Continuous: PEG-rich (35 kDa)(5%,10%) (w/v) | |||||||||
| Assembled evaporation pump method | / | Emulsion: DEX (20 kDa, 40 kDa, 100 kDa, 500 kDa) | Flow-focusing PDMS/glass hybrid microfluidic device | Suction of the evaporation pump | None | 84 | |||
| Continuous: PEG (20 kDa, 500 kDa) | |||||||||
| Transient Oil-medium | Oil-droplet chopper method | ∼0.064 to 0.164 | Emulsion: DEX-rich (40 kDa) | Flow-focusing glass device with a theta-shape injection capillary | Syringe Pumps. | None | 51 | ||
| Continuous: PEG-rich (8 kDa) | |||||||||
| Transient double emulsion method | ∼0.064 to 0.164 | Emulsion: DEX-rich (40 kDa) or PEG-rich(8 kDa) | Flow-focusing glass device | Syringe Pumps. | None | 85 | |||
| Middle: Silicone oil or PDMS oil mixture | |||||||||
| Continuous: PEG-rich(8 kDa) or DEX-rich(40 kDa) | |||||||||
| Oil-mediated method (passive flow focusing) | ∼0.116 to 0.521 | Emulsion: DEX-rich (10 kDa) with PAH | Typical multi-inlet flow-focusing PDMS device. | / | None | 76 | |||
| Middle-1: Pure PEG-rich (8 kDa) | |||||||||
| Middle-2: PEG-rich (8 kDa with PSS | |||||||||
| Continuous: HFE 7500 oil. | |||||||||
| Others | Water-head-driven microfluidic oscillators method | ∼0.1 | Emulsion: 20% (w/w)DEX(550 kDa) | Three-layer PDMS microfluidic device | Height differences of the oscillator, input and output well | None | 86 | ||
| Continuous: 10% (w/w) PEG (35 kDa) | |||||||||
| Interface fingering instability method | / | Emulsion: 3 wt. % DEX (500 kDa) | Vertically orientated Hele-Shaw cell | Manually. | None | 87 | |||
| Continuous: 30 wt. % PEG (8 kDa)DEX (500 kDa,1–30 wt. %) or DEX (10 kDa, 5 wt. %) PEG (8 kDa, 0–40 wt. %) or PEG (6 kDa, 30 wt. %) | |||||||||
TABLE IV.
Advantages and disadvantages of the all-aqueous single-emulsion droplets generation methods.
| Method | Advantages | Disadvantages | Reference | |||
|---|---|---|---|---|---|---|
| Active methods | Mechanical perturbation | Vibrating the microtubing | Mechanical vibration | Easy to implement; device is easy to generate; low cost; good control over the flow morphology | Low-generation frequency; require lots of external components; narrow range of generation conditions | 28 59 60 |
| Mechanical shaking | Easy to implement; device is easy to generate; low cost | Low-generation frequency; require external orbital shaker; narrow range of generation conditions | 61 | |||
| Acoustic actuation (audio speaker) | The device is easy to fabricate, and the whole experimental setup is easy to assemble; applicable to ATPS with lower interfacial tension; all external components are off-the-shelf | Require lots of external components; low uniformity | 62 | |||
| Acoustic actuation (rupture of interfacial fingers) | Device is easy to generate; relative high-generation frequency in existing active method; prepared the smallest ATPS droplets so far; low cost | Require external mechanical vibration | 63 | |||
| Varying channel dimension | Piezo-electric bending disk | Relative high-generation frequency in active methods; much wider range of generation conditions | Require lots of external components; low uniformity; the two-layer PDMS-glass microfluidic device is relative complex | 64 65 66 | ||
| Pneumatic solenoid valve | Easy to scaling up | The three-layer PDMS device is relative complex; droplet size is relative big; low-generation frequency | 25 | |||
| Braille pin actuation | Applicable to ATPS with lower interfacial tension; the device can be fabricated by BDLL procedure involving only a single exposure and no alignment steps | Low-generation frequency; droplet size is relative big | 67 | |||
| Piezo-electric buzzer | Relative high-generation frequency in active methods; much wider range of generation conditions | Require lots of external components; low uniformity | 68 | |||
| “On-off” flow control | Pulsating inlet pressure | Simplicity; easy to scaling up | Require external components; low-generation frequency as the disperse phase applied pressure “off” time toff has to be sufficiently long to enable the droplet generation | 69 | ||
| Square-wave-changing injection method | Easy to implement | Low frequency; low uniformity | 70 | |||
| Electrical actuation | Electrohydrodynamic (EHD) method | Modifiable via post-processing of the droplets; high controllability for the droplet numbers | Electric field may cause material degradation, impair the bioactivity of cells and biomolecules, imposing restrictions on biomedical applications; not applicable for ATPS with similar electrophoretic mobility; low-generation frequency | 71 72 73 74 | ||
| Electrospray method | Simple microfluidic device; applicable to viscous liquids; controllable ATPS droplets size and structure; wide droplet radius distribution | Electric field may cause material degradation, impair the bioactivity of cells and biomolecules, imposing restrictions on biological and biomedical applications | 75 | |||
| Oil-mediated method (active pico-injection) | High-generation frequency | The oil phase should be removed after the ATPS droplets are generated. The applied electric field may cause material degradation, impair the bioactivity of cells and biomolecules, imposing restrictions on biological and biomedical applications. | 76 | |||
| Passive method | Membrane emulsification | Hollow-fiber membrane emulsification method | Does not rely on any external trigger | The ATPS droplets size are very large; poor-wettability requirement for the membrane material; low-generation frequency; difficult to observe and measure as the membrane is not transparent | 58 | |
| Shirasu porous glass membrane emulsification method | No need external components; obtain small ATPS droplets | Both droplet-in-droplet and Janus structure can be obtained, without precise controllability | 78 | |||
| Precise control of liquid pressure | Weak gravity-driven hydrostatic pressure method | Easy to implement; take liquid-filled pipette tips replace the syringe pump; no need external components and complex chip fabrication; low cost | A narrow range of flow conditions; low-generation frequency; the applied hydrostatic pressures decrease as the decreasing of the liquid column height in the pipette tip | 27 79 80 | ||
| Microneedle-assisted gravity-driven hydrostatic pressure method | Easy to implement; no need external components and complex chip fabrication; relative high-generation frequency; small ATPS droplets; low cost | Low uniformity | 81 | |||
| Pressure-/flow rate-controlled method | Relative high-generation frequency; the droplet size is relative small | The flow control system is complex; low uniformity | 82 | |||
| Fracture-based variable-width microchannel method | Produce small ATPS droplets; useful for ATPS of low solute concentration with low interfacial tension | Low uniformity; low-generation frequency; the PDMS device is relative complex | 83 | |||
| Assembled evaporation pump method | No need external components; easy to scale-up | Low-generation frequency; the flow rates of disperse and continuous phases cannot be controlled independently | 84 | |||
| Transient oil-medium | Oil droplet chopper method | No need external components and complex chip fabrication; wide range of flow conditions; Selectively combining the benefits of both passive and active methods; high-generation frequency; droplet size with uniformity; high degree of controllability: the droplet size and generation frequency can be controlled independently; low cost | The oil phase should be removed after the ATPS droplets are generated | 51 | ||
| Transient double emulsion method | No need external components and complex chip fabrication; low cost; the generated ATPS droplets with controllable stability; on-demand release the ATPS droplets from the shell phase; both dextran-in-PEG and PEG-in-dextran ATPS droplets can be generated | The droplet size is big; the oil phase should be removed after the ATPS droplets are generated | 85 | |||
| Oil-mediated method (passive flow focusing) | No need external components and complex chip fabrication; low cost; high-generation frequency | The oil phase should be removed after the ATPS droplets are generated | 76 | |||
| Others | Water-head-driven microfluidic oscillators method | No need off-chip controller | The microfluidic device is complex and difficult to implement; droplet size is relative big; Low-generation frequency | 86 | ||
| Interface fingering instability method | Relative high-generation frequency; easy to implement and easy to scaling up; low cost | The ATPS droplets sizes are large; low uniformity | 87 | |||
A. Active generation of all-aqueous single emulsion droplets
Mechanical and electrical energies are the two most commonly used energy types to facilitate ATPSs droplets generation. Mechanical approaches can tune the hydrodynamic pressure to vary the flow velocity, and electrical approaches can apply electrical forces to the aqueous interface.51
1. Mechanical perturbation
According to the actuation mechanism, mechanical perturbation methods can be subdivided into three categories: vibrating the flexible microtubing, tuning the channel dimension, and switching “on–off” the valve control.
a. Vibrating the microtubing
By attaching an external component to the flexible microtubing of the dispersed phase liquid, the applied mechanical wave (with different frequency and amplitude) can perturb the inlet pressure of the dispersed phase liquid to facilitate ATPS single emulsion droplets generation. The external component can be a mechanical vibrator as in the mechanical vibration method,28,59,60 an orbital shaker as in the mechanical shaking method,61 or an audio speaker as in the acoustic actuation method.62,63 With these methods, syringe pumps are used to drive the dispersed and continuous phases into a co-flow glass microfluidic device, with an external component attaching to the flexible microtubing of the dispersed phase. Advantages include the low cost, the easy fabrication of the device (e.g., co-flow glass capillary microfluidic device), and ease of implementation. However, they involve many external components and their range of generation conditions is narrow.
Mechanical Vibration
For the mechanical vibration method, the mechanical vibrator is attached to the microtubing to directly perturb the inlet pressure [Fig. 5(a)]. Glass capillary devices with a co-flow geometry59 or a geometry combing co-flow and flow-focusing structure28,60 can be used. A sinusoidal-wave generator is used to control the vibration frequency in the range between 0.1 and 5000 Hz and the vibration amplitude by tuning the input voltage. The frequency of ATPS droplet generation (9–35 Hz) is determined by the sinusoidal-wave generator, which is the same as the vibration frequency. The ATPS droplet size (radius being from 20.5 to 100 μm) can be tuned by manipulating the vibration frequency, vibration amplitude, and flow rates of both the inner and outer phase liquids.59 However, the droplet generation is limited to the condition of a sufficiently high amplitude and a narrow range of low frequency.
FIG. 5.
Vibrating the microtubing. (a) Mechanical vibration. A mechanical vibrator was connected to the microtubing of the dispersed phase to provide mechanical vibration. Reproduced from A. Sauret and H. C. Shum, Appl. Phys. Lett. 100, 154106 (2012), with the permission of AIP Publishing.59 (b) Mechanical shaking. An orbital shaker was attached to the microtubing of the dispersed phase to provide mechanical vibration. Reproduced from Shum et al., Biomicrofluidics 6, 012808 (2012), with the permission of AIP Publishing.61 (c). Acoustic actuation (audio speaker). An audio speaker was attached to the injection microtube for the dispersed phase. Reproduced with permission from De Lora et al., ACS Appl. Bio Mater. 2, 4097 (2019). Copyright 2019 American Chemical Society.62 (d) Acoustic actuation (rupture of interfacial fingers). The vibration-induced local rupture of an ATPS jet interface. Reprinted with permission from Mak et al., Langmuir 34, 926 (2018). Copyright 2017 American Chemical Society.63
Mechanical Shaking
For the mechanical shaking method, an orbital shaker is used instead of a mechanical vibrator. The mechanical shaking, with a tunable frequency and constant amplitude, is introduced by the shaker to provide transverse and longitudinal perturbing for ATPS droplet generation [Fig. 5(b)].61 The ATPS droplet size can be tuned by controlling the shaking frequency and flow rates. However, the generation frequency is low (6–8 Hz) with a narrow operation range.
Acoustic Actuation
To increase the ATPS droplet generation frequency, acoustic actuation is introduced to perturb the inlet pressure of the dispersed phase. In the acoustic actuation method, lots of off-the-shelf acoustic components (such as an audio speaker, an amplifier, and a sinusoidal waveform source) are introduced to construct and assemble the experimental setup to facilitate the ATPS droplet generation [Fig. 5(c)].62 The sinusoidal waveform source (either a function/arbitrary waveform generator or a smartphone with an application for wave-function generation) is connected to an amplifier and then to an audio speaker, which is attached to the microtubing to provide a mechanical wave. By tuning the acoustic modulation frequency and the flow rates, ATPS droplets with smaller sizes (radius from 50 to 125 μm) and improved generation frequency (10–60 Hz) are generated. However, this method requires several external components and the generated droplets have poor uniformity in size with the CV value of ∼18.3%. Then, the rupture of interfacial finger technique63 is proposed to produce ATPS droplets with a higher generation frequency, smaller droplet size, and better monodispersity. In this technique, a louder speaker is used to provide the mechanical vibration with a higher frequency and a lower amplitude. The W/W interface folds and forms a protruded liquid finger, which can break up to form ATPS droplets when the aspect ratio of the finger exceeds a threshold value.63 By tuning the perturbation frequency from 200 to 1000 Hz, relatively high uniform (CV of ∼3%) ATPS droplets with small sizes (radius from 1.5 to 8.5 μm) and relatively high-generation frequencies (200–1000 Hz) can be achieved [Fig. 5(d)].
b. Varying channel dimension
In these methods, syringe pumps are used to drive the liquids, and the mechanical vibration is applied to the dispersed phase by integrating external components into Polydimethylsiloxane (PDMS) poly (methyl methacrylate) (PMMA) microfluidic devices to vary the channel dimension. The actuation external components include piezo-electric bending disk,64–66 pneumatic solenoid valve,25 Braille pin (BP) actuator,67 and piezo-electric buzzer.68 Typically, the integration of external components into the microfluidic device makes the device complex and difficult to fabricate.
Piezo-electric Bending Disk
By this method,64–66 the two-layer three-dimensional (3D) flow-focusing PDMS microdevice, embedded with a piezo-electric bending disk, is fabricated through soft lithography [Fig. 6(a)]. The piezoelectric bending disk is placed on top of the central inlet channel of the PDMS device. A sinusoidal alternating current (AC) voltage, generated by a function generator and amplified via a linear amplifier, is applied to the piezoelectric bending disk. The piezo-electric bending disk contracts and relaxes periodically, resulting in the contraction and relaxation of the channel width, and ATPS droplets are periodically released. The droplet size is controlled by the actuation frequency of the piezo-electric bending disk and the flow rates of both liquid phases. By piezo-electric bending disk actuation, the droplet generation frequency (up to 50 Hz) is relatively higher and they enable a wider range of flow conditions for droplet generation in comparison to other active methods. However, the uniformity of droplets is poor (CV < 10%), and the 3D microfluidic device is relatively complex.65
FIG. 6.
Varying channel dimension. (a) Piezo-electric bending disk. The PDMS flow-focusing microfluidic device, with a piezo-electric bending disk integrated on top of the PDMS central channel. Reproduced with permission from Ziemecka et al., Lab Chip 11, 620 (2011). Copyright 2011 Royal Society of Chemistry.65 (b) Pneumatic solenoid valve. Schematic of the droplet microfluidic system, chip design, droplet generation unit, and the actual image of the microfluidic chip. Reproduced with permission from Liu et al., Small 14, 1801095 (2018). Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.25 (c). Braille pin actuation. Braille pin (BP) was integrated under the narrow orifice. CH1, CH2, OH: channel 1, 2 and orifice height, respectively. Reproduced with permission from Lai et al., Lab Chip 11, 3551 (2011). Copyright 2011 Royal Society of Chemistry.67 (d). Piezo-electric buzzer. Image of the 3D flow-focusing microfluidic device embedded with a piezo-electric buzzer. Reproduced with permission from Tarameshlou et al., Int. J. Polym. Mater. Polym. Biomater. 63, 884 (2014). Copyright 2014 Taylor & Francis Group, LLC.68
Pneumatic Solenoid Valve
The 3D microfluidic device is fabricated by three pieces of PDMS sheets bonded together, which is composed of a control unit at the top layer, a droplet generation unit and a collection unit in the middle layer, and a flat plane at the bottom layer [Fig. 6(b)].25 The pneumatic single-layer membrane valves are integrated into the PDMS device and located beside to the disperse phase flow channel for ATPS droplet generation. When the pneumatic solenoid valves are switched off, the dispersed channel keeps its original shape without deformation and the dispersed dextran phase is continuously pumped into the junction of the device. When valves are switched on, they compress the disperse channel, by which only a small amount of dextran phase is pumped into the junction to break up into ATPS droplets. Varying the switching period of the solenoid valve changes both the size of ATPS droplets and the distance between droplets. Although this method is easy to scale up, the three-layer PDMS microfluidic device is complex, the size of the droplets is large (radius from 50 to 90 μm) with low uniformity, and the generation frequency is low (∼12.5 Hz).
Braille Pin Actuation
This method has produced ATPS droplets with low interfacial tension (∼0.01 mN m−1). A PDMS flow-focusing microfluidic device, with bell-shaped cross-section and rounded multi-level micro-channels, is fabricated by backside diffused light lithography (BDLL) using a photomask with different widths and disconnected regions. Furthermore, BDLL fabrication endows the device with narrow orifices between multi-level channels [Fig. 6(c)].67 The Braille pin is integrated under the narrow orifice to deform microchannels for actuating the dispersed-phase flow in a frequency ranging from 0.83 to 2.5 Hz. By decreasing the Braille actuation frequency, both the ATPS droplet volume (0.9–3.4 nl) and the inter-droplet spacing (200–1400 μm) increase. Braille pin actuation method applies to ATPSs with ultralow interfacial tension, and the BDLL enables the fabrication of complex PDMS microdevice with multi-levels and a narrow orifice. However, the droplet-generation frequency is low (0.83–2.5 Hz), and the droplet size is relatively large (radius of ∼60–93 μm corresponding to the volume of 0.9–3.4 nl).
Piezo-electric Buzzer
In the piezo-electric buzzer method,68 the micro-channels are designed by AutoCAD and engraved on two PMMA layers by CO2 laser micromachining. Then, the two PMMA engraved sheets are sandwiched by two metal layers and fastened by screws [Fig. 6(d)]. A piezo-electric buzzer is integrated into the flow-focusing microfluidic device to mechanically force the disperse phase liquid for ATPS droplet generation. A function generator is used to switch the “on” and “off” of the piezoelectric buzzer. The size of ATPS droplets can be controlled by varying the flow rates and the piezoelectric buzzer frequency. The produced ATPS droplets have a radius ranging from 9.5 to 56 μm, CV of <10%, and generation frequency of 2–40 Hz.
c. “On–off” flow control
In these methods, external components are used to switch on and off the driven force for controlling the injection of the dispersed phase or the outer phase. Although these methods have simple microfluidic devices and are easy to implement, the generation frequency of the all-aqueous emulsion droplets is ultralow [typically O(0.1) to O(1) Hz] and the generated droplets have poor uniformity.
Pulsating Inlet Pressure
In this method, a PDMS/glass hybrid microfluidic device with a flow-focusing geometry is used. The dispersed dextran phase is injected by a pulsating inlet pressure that is controlled by a three-way pneumatic solenoid vale, while the continuous polyethylene glycol (PEG) phase is injected via a syringe pump with a constant flow rate. A 250-μl pipette tip, filled with the disperse dextran phase, is vertically inserted into the dispersed inlet of the microfluidic device, and the other end of the pipette tip is connected to a Tygon tubing for compressed air transportation. A three-way solenoid valve is used to switch on and off the compressed air for controlling the injection of the dispersed dextran phase [Fig. 7(a)].69 When the inlet pressure is switched on by opening the compressed air, it pushes the dispersed dextran phase into the cross-junction of the microfluidic device. When the inlet pressure is switched off, the dextran jet is cut off by the outer continuous phase liquid. ATPS droplets are generated continuously by periodically modulating the on and off states of the dispersed-phase inlet pressure while keeping the continuous-phase flow rate constant. The droplet size can be controlled by manipulating the time for inlet pressure on (ton) and off (toff), and the continuous phase flow rate. This method is simple, easy to scale up, and applicable to on-demand ATPS droplet formation. However, the droplet generation frequency is ultralow (typically ranging from O(0.1) to O(1) droplets per second) due to the requirement of a sufficiently long time (toff) for switching off the inlet pressure.
FIG. 7.
“On–off” flow control. (a) Pulsating inlet pressure. The inner dextran phase is injected by an applied pulsating pressure (via a pressure regulator and a three-way pneumatic solenoid valve), while the outer PEG phase is introduced by a syringe pump. Reproduced with permission from Moon et al., Lab Chip 15, 2437 (2015). Copyright 2015 Royal Society of Chemistry.69 (b) Square-wave-changing injection. A microflow controller is used to precisely control the flow rates of the fluids. Reproduced with permission from Liu et al., Biomed. Microdevices 19, 55 (2017). Copyright 2017 Springer Science Business Media New York.70
Square-wave-changing Injection
A modified co-flow PDMS microfluidic device, integrated with a microflow controller, is used for ATPS droplet generation. The microfluidic device involves three inlets, a narrow convergence channel, followed by an open space [Fig. 7(b)]. The narrow converging channel enables a dextran-rich (DEX-rich) jet, while the subsequent open space results in a significant decrease in the jetting velocity, which pinches off ATPS droplets from the jet. The integrated microflow controller is used to obtain a constant injection pressure for the inner disperse phase (dextran-rich phase) and a periodically square-wave-changing injection pressure for both the middle PEG (without H2O2) and outer PEG (with H2O2) phases. The periodically square-wave-changing injection pressure of the PEG phases facilitates the stable dextran-in-PEG jet break up into ATPS droplets.70 Square-wave-changing injection method combines a modified flow-focusing [as shown in Fig. 7(b)] microfluidic device structure with the periodically changing injection pressure for the middle and outer phases to facilitate the ATPS droplet generation. By using this method, ATPS droplets (32.5–55.5 μm in radius) can be produced; however, the generation frequency is ultralow (typically ∼0.01–0.1 Hz).
2. Electrical actuation
In electrical actuation, electric forces are usually applied to the aqueous interface to facilitate the all-aqueous droplet generation. For electrical actuation methods, liquids are driven by syringe pumps, a direct current (DC) power supply is needed. The electric field may cause material degradation and damage the bioactivity of biomolecules, thus affecting their applications in biomedicine.
a. Electrohydrodynamic (EHD) method
Electrohydrodynamic (EHD) method71–74 exploits electrostatic force as a trigger to generate ATPS droplets. It is applicable to ATPSs with different electrophoretic mobilities, such as tetrabutylammonium bromide (TBAB) and ammonium sulfate (AS). A PDMS/glass hybrid microfluidic device with a T-junction geometry is fabricated by soft lithography. The stainless steel tubes are inserted into the outlet of the device to serve as both electrodes and outlet ports. A DC power supply is used to generate the potential difference between the electrodes. During the operation, one branch of the T-junction microchannel is filled with the continuous TBAB-rich phase while the other branch is occupied by both the dispersed AS-rich phase and the continuous TBAB-rich phase [Fig. 8(a-i)]. When the DC power supply is switched on, the AS-rich dispersed phase with a higher electrophoretic mobility moves faster to the positively charged electrode than the TBAB-rich phase. Afterward, when the electric pulse signal is switched off, the lengthened AS-rich stream is “chopped” into a droplet by the shearing force of the continuous phase liquid [Fig. 8(a-ii)].72,73 The size and generation frequency of ATPS droplets are controlled by changing the magnitude and duration of the DC electric pulse. When the applied voltage is increased, the deformation of the AS-rich phase becomes larger and a smaller volume of the AS-rich phase is detached to form a smaller droplet. As only one ATPS droplet is produced during an electrical pulse period, the generation frequency of ATPS droplets increases with the decrease in the duration of DC electrical pulse. ATPS droplets with a radius of ∼58.4 μm and CV of ∼6.5% are fabricated and the generation frequency can be up to ∼5 Hz. The electrohydrodynamic method works for generating ATPSs with different electrophoretic mobilities but is not applicable for generating ATPS with similar electrophoretic mobility, and the generation frequency is low.
FIG. 8.
Electrical actuation. (a) Electrohydrodynamic method. (i) Schematic of the microfluidic device (upper) and the microscope image of the region in the dotted-line box at the T-junction (lower). Reproduced with permission from Song et al., J. Chromatogr. A 1162, 180 (2007). Copyright 2007 Elsevier B.V.72 (ii) The applied forces for ATPS droplet generation in EHD method. Reproduced from Choi et al., AIP Conf. Proc. 1027, 1006 (2008), with the permission of AIP Publishing.73 (b) Electrospray. Experimental setup for ATPS emulsion droplets generation in electrospray approach. The disperse phase is positively charged and injected into a glass capillary, forced through an oppositely charged electrode, and finally pulled into the continuous phase bath. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75 (c). Active pico-injection. Reproduced with permission from Nan et al., Microsyst. Nanoeng. 6, 70 (2020). Copyright 2020 Springer Nature.76
b. Electrospray
In the electrospray method, the liquid jet (such as dextran) is positively charged and sprayed into the air by a DC power supply, whose cathode is connected to a metallic needle and anode is connected to a metallic ring. The dextran jet readily breaks up into discrete aqueous droplets due to the combined effects of high surface tension of the air–water interface and the applied electrostatic force. Once generated, dextran droplets are forced to pass through the negatively charged metallic ring and finally fall into a rotating bath containing the continuous PEG phase for the formation of ATPS droplets [Fig. 8(b)].75 The generation frequency and size of ATPS droplets can be controlled by changing the electrical field strength, the nozzle diameter, and the injection rate of dextran. Using electrospray, the microfluidic device is simple to fabricate (typically a simple taper glass capillary), a wide range of droplet size (ranging from 22.5 to 1000 μm in radius) can be achieved and this method is applicable to viscous aqueous liquids. However, the applied electric field may cause material degradation and impair the bioactivity of cells and biomolecules, so as to impose restrictions on biological and biomedical applications.
c. Active pico-injection
Recently, an active pico-injection method is proposed to upgrade the generation frequency of all-aqueous emulsion droplets.76 This method involves two steps: (1) the PEG-in-HFE 7500 oil emulsion droplets are generated at the cross-junction when the inner PEG-rich phase meets the HFE 7500 oil phase, and (2) when the PEG-in-HFE 7500 oil (W/O) single emulsion droplets flow by the pico-injector, a high voltage is applied to the electrodes to destabilize the water–oil interface, thus introducing the dextran-rich (DEX-rich) solution into the shell drop to form the DEX-in-PEG-in-HFE 7500 oil (W/W/O) double emulsion droplets. It relies on accurate control over the applied electric field to achieve a stable injection of an identical volume of DEX-rich solution into each droplet [Fig. 8(c)]. After generation, the double emulsion droplets are transferred into a petri dish filled with the PEG-rich phase to remove the outer oil phase by evaporation and release the encapsulated DEX-rich core to form DEX-in-PEG droplets. In this method, a volatile oil (HFE 7500 oil) phase and an electric field are used to facilitate the ATPS droplets generation. The generation frequency of the DEX-in-PEG (W/W) emulsion droplets is the same as the PEG-in-HFE 7500 oil (W/O) and the maximum generation frequency can reach ∼2.4 kHz.
B. Passive generation of all-aqueous single emulsion droplets
Compared with active methods, passive methods typically have a narrow generation range, but have the following advantages: low cost, easy to implement, no external components, simple and easy device fabrication (typically using PDMS, glass, or PDMS/glass hybrid microfluidic devices with flow-focusing geometry). According to the mechanism, the passive methods are classified into four groups: membrane emulsification,58,77,78 precise control of liquid pressure,27,79–84 transient oil medium,51,76,85 and others (such as interfacial fingering instability).86,87
In 2015, membrane emulsification was first proposed for the passive generation of all-aqueous droplets. Although the membrane emulsification method makes it possible to prepare all-aqueous emulsion droplets passively, the generation frequency is low, the uniformity is poor, and the controllability is poor. To generate highly uniform all-aqueous emulsion droplets passively, the weak gravity-driven hydrostatic pressure method was proposed in 2016. The method uses liquid-filled pipette tips to replace the syringe pump to provide a slow flow rate in facilitating the generation of all-aqueous emulsion droplets. In addition to the gravity-driven hydrostatic pressure method, some other strategies (such as pressure-/flow rate-controlled method, fracture-based variable-width method, and assembled evaporation pump method) are also proposed to generate ATPS droplets passively by providing a slow flow rate. Although these methods improved the droplet uniformity, the generation frequency was still low due to their slow flow rate. To widen the ATPS droplet generation condition, and improve the generation frequency and uniformity, Zhou et al. proposed a simple “oil droplet chopper” method in 2017 to produce highly uniform ATPS droplets with a wide range of droplet sizes and high-generation frequency by introducing an oil phase into ATPS. After that, a “transient double emulsion method” was also proposed for the on-demand generation of all-aqueous emulsion droplets.
1. Membrane emulsification
Both porous hollow-fiber membrane58,77 and Shirasu porous glass (SPG) membrane78 have been used to generate all-aqueous emulsion droplets.
a. Hollow-fiber membrane emulsification
As the first passive method, the hollow-fiber membrane emulsification method (proposed in 2015) made the passive preparation of ATPS droplets possible. It exploits the porous polypropylene microfiltration hollow-fiber membrane for ATPS droplet generation. The two ends of the membrane are connected to two metal tubes, which are fixed in a transparent acrylic shell with a through-hole joint to form an experimental device with a core-shell concentric tube structure [Fig. 9(a-i)].77 With this technique, the dispersed phase is injected into the hollow fiber, while the continuous phase is injected into the acrylic shell and then permeates through the membrane wall [Figs. 9(a-ii) and 9(a-iii)]. The permeating continuous phase forms an annular outer phase flow around the dispersed phase thread inside the hollow fiber. The flow rate of the continuous phase is constantly accelerated along the entire membrane length as the continuous phase continuously permeates into the hollow fiber, while the dispersed phase flow rate keeps constant, which facilitates the ATPS droplet generation [Fig. 9(a-ii)].58 The generation frequency and size of ATPS droplets can be controlled by varying the hollow-fiber dimensions and the flow rates of both disperse and continuous phases. Typically, the droplets have diameters 20%–60% larger than the inner hollow fiber diameter. To obtain a smaller droplet size, the dispersed phase is diluted. Then when the two out-of-equilibrium phases contact each other inside the hollow fiber membrane and generate the ATPS droplets, the mass transport between the droplet and the continuous phase toward thermodynamic equilibrium occurs and the droplet shrinks to a smaller size [Fig. 9(a-iii)].58 With the hollow-fiber membrane emulsification method, the produced ATPS droplet sizes are relatively large (with radius 320–610 μm), with relatively high uniformity (CV: 0.75%–3%). However, the generation frequency is low, and it is difficult to select the membrane material because of the poor water-wettability requirement and difficult to observe the droplet generation process as the membrane is not transparent.
FIG. 9.
Membrane emulsification. (a) Hollow-fiber membrane emulsification. (i) Experimental setup. (i) Reproduced with permission from Breisig et al., J. Membr. Sci. 467, 109 (2014). Copyright 2014 Elsevier B.V.77 (ii) ATPS droplet generation from phases in equilibrium. (iii) Smaller ATPS droplet generation from ATPS phases not in equilibrium. (ii) and (iii) Reproduced from H. Breisig and M. Wessling, Biomicrofluidics 9, 044122 (2015), with the permission of AIP Publishing.58 (b) Shirasu porous glass (SPG) membrane emulsification. (i) Two kinds of W/O emulsion droplets generation. (ii) Mix of the two sets of W/O emulsion droplets. (iii) Water extraction due to the osmotic pressure difference, yielding ATPS droplets. (iv) and (v) Evolution of the two sets of W/O droplets. Reproduced with permission from Akamatsu et al., Langmuir 35, 9825 (2019). Copyright 2019 American Chemical Society.78
Shirasu Porous Glass Membrane Emulsification
To generate smaller ATPS droplets, Akamatsu et al. proposed a passive two-step Shirasu porous glass (SPG) membrane emulsification method in 2019.78 It involves two steps: (1) two sets of W/O emulsion droplets are first prepared using SPG membrane emulsification [Fig. 9(b–i)], and (2) then, the two sets of W/O emulsion droplets are mixed [Figs. 9(b-ii) and 9(b-iv)] by which water is extracted from one set of W/O emulsion droplets into the other set of W/O emulsion droplets due to the osmotic pressure difference. The water extraction leads to the increased concentration of polymers in the W/O droplets. When the concentration exceeds the threshold of phase separation, ATPS-droplet-in-oil W/W/O double emulsion droplets are formed [Figs. 9(b-iii) and 9(b-v)]. The ATPS droplet generation is independent of the solute type and the SPG membrane size, while the solute concentration in the second set of W/O emulsion droplets is the critical parameter, which provides the driving force for water extraction. Only when the solute concentration is high enough, can the resulting osmotic pressure difference be sufficiently high for the extraction of enough water to produce ATPS droplets. The ATPS droplet size is determined by the SPG membrane pore size. The SPG membrane emulsification method can be used to fabricate ATPS droplets with small size (radius of 5–25 μm). However, the fabricated ATPS droplets may have a core-shell structure or a Janus structure, which cannot be precisely controlled.
2. Precise control of liquid pressure
To generate all-aqueous emulsion droplets spontaneously in the dripping mode by passive methods, a low flow velocity u is required for ATPS with an ultralow interfacial tension. Gravity-driven hydrostatic pressure,27,79–81 high-precision flow control system,82,83 and assembled evaporation pump84 are used to precisely control the flow rates.
a. Weak gravity-driven hydrostatic pressure
Weak gravity-driven hydrostatic pressure method27,79,80 is a completely passive method, as no external components (or even a syringe pump) are required for ATPS droplets generation. Both dispersed and continuous phases are injected into the microfluidic device by the vertically inserted liquid-filled pipette tips [Fig. 10(a-i)]. The ultra-low flow condition can be precisely controlled by adjusting the column height of the hydrostatic fluid in the pipette tips. The glass-PDMS hybrid microfluidic device with a flow-focusing geometry is fabricated by the standard soft lithography method. Both phases are fed by the solution-filled 250-μl pipette tips under the gravity-driven hydrostatic pressure, which are vertically inserted into the inlets of the microfluidic device [Figs. 10(a-i) and 10(a-ii)].79,81 The applied hydrostatic pressure can be easily controlled by varying the fluid column height in the pipette tip. The droplet size can be tuned by changing the interfacial tension and viscosity of the ATPS, or the fluid column height filled in the pipette tips. This method is simple and easy to implement. However, it has two limitations: (i) the hydrostatic pressure is decreasing with time due to the decrease in column height in the pipette tip as the fluid is injected out, and thus, the generated ATPS droplets are poor in the uniformity of size; (ii) the inherent generation frequency is low (∼12.5 Hz) due to the low flow rate by the weak gravity-driven hydrostatic pressures; and (iii) the total throughput is limited by the volume of liquids stored in the pipette tips.
FIG. 10.
Precise control of liquid pressure. (a) Weak gravity-driven hydrostatic pressure. (i) Image of the experimental setup, including a microfluidic device with two prefilled pipette tips vertically inserted into the inlets. Reprinted with permission from Moon et al., Anal. Chem. 88, 3982 (2016). Copyright 2016 American Chemical Society.79 (ii) The PDMS-glass flow-focusing microfluidic device. Reprinted with permission from Jeyhani et al., J. Colloid Interface Sci. 553, 382 (2019). Copyright 2019 Elsevier, Inc.81 (b) Microneedle-assisted gravity-driven hydrostatic pressure. (i) Schematic of the experimental platform consists of two plastic syringe tubes, a microfluidic device, and an adjustable lab jack used to adjust the effective column heights. (ii) The microneedle-inserted microfluidic device. Reprinted with permission from Jeyhani et al., J. Colloid Interface Sci. 553, 382 (2019). Copyright 2019 Elsevier, Inc.81 (c) Pressure-/flow rate-controlled. Schematic of the experimental setup, including a flow control system and a PDMS/glass microfluidic device. Reproduced with permission from Y. Chao and H. C. Shum, Chem. Soc. Rev. 49, 114 (2020). Copyright 2019 Royal Society of Chemistry.56 (d) Fracture-based variable-width microchannel. (i) Schematic of the PDMS microfluidic device. (ii) Image of the experimental setup, with a PDMS microfluidic device mounted on the microscope stage by a micro-vice. Reproduced with permission from Choi et al., Soft Matter 15, 4647 (2019). Copyright 2019 Royal Society of Chemistry.83 (e) Assembled evaporation pump. (i) Schematic of the experimental setup. (ii) Microfluidic device with eight generation units for high-generation frequency. Reproduced with permission from Zhang et al., J. Mater. Sci. 54, 14905 (2019). Copyright 2019 Springer.84
Microneedle-Assisted Gravity-Driven Hydrostatic Pressure
To solve the unsteady injection pressure and low-generation frequency issues existing in the weak gravity-driven hydrostatic pressure method, the microneedle-assisted gravity-driven hydrostatic pressure method is proposed by the same research group.81 To maintain a steady injection pressure, pipette tips are replaced by plastic BD syringe tubes that have wider inner diameters and larger volumes (60 ml). Furthermore, the hydrostatic pressure from the effective fluid column height is controlled by using an adjustable lab jack [Figs. 10(a-i) and 10(b-i)]. To increase the generation frequency, the conventional flow-focusing microfluidic device is modified by inserting a metal microneedle through the dispersed phase inlet channel into the flow-focusing junction, forming a 3D flow-focusing microfluidic device. The simple modification of the microfluidic device facilitates the ATPS droplet generation with a wide range of sizes.81 By using the microneedle, the injection area is reduced, and thus, smaller ATPS droplets can be generated with a higher generation frequency. In addition, constant and higher effective liquid column height and hydrostatic pressure can be obtained by using an adjustable lab jack, leading to constant and higher ATPS droplet generation frequency, which is up to ∼850 Hz (compared with ∼15 Hz for weak gravity-driven hydrostatic pressure method). For the microneedle-assisted hydrostatic pressure method, the device is simple and easy to implement, small droplet size (5 –65 μm in radius) with relatively high-generation frequency (∼850 Hz) can be obtained. However, as generated in the jetting regime, the ATPS droplets have low uniformity in size (CV of ∼10%).
Pressure-/Flow Rate-Controlled Method
Pressure-/flow rate-controlled method82 exploits a high-precision pressure control system to provide injection flow rates lower than syringe pump-driven method but higher than pipette tip-driven method. In this way, ATPS droplets with small sizes and relatively high-generation frequencies can be fabricated by controlling the flow rates in the jetting regime rather than in the dripping regime. The experimental setup is composed of a flow control system and a flow-focusing PDMS/glass hybrid microfluidic device [Fig. 10(c)]. The flow control system consists of a high-precision pressure control system and a flow unit platform. Precisely controlled pressurized air drives the fluids into the microfluidic device to generate ATPS droplets. Both droplet size and generation frequency can be controlled by changing the inlet pressures. This method can generate relatively small ATPS droplets (7–14 μm in radius) with relatively high-generation frequency (∼300 Hz). However, as generated in the jetting regime, the droplets have poor uniformity in size.
Fracture-Based Variable-Width Microchannel
For the fracture-based variable-width microchannel method, the ATPS droplet generation condition is expanded by varying the central inlet channel width.83 By applying mechanical strain, the fracture-based inlet central microchannel width can be tuned in the range of 1–10 μm, which widens the range of the ATPS droplets generation conditions. The three-layer microfluidic device is fabricated by an h-PDMS (“hard” PDMS) layer and a PDMS substrate layer, which consists of three inlets and one outlet [Fig. 10(d-i)]. The microfluidic device is installed in a micro-vice to apply mechanical strain to the PDMS substrate [Fig. 10(d-ii)]. Although the variable central inlet channel width does not affect the droplet size, it can expand the ATPS droplet generation conditions. This method is applicable to ATPSs with low interfacial tension (∼0.012 to ∼0.103 mN m−1) and low solute concentration. ATPS droplets with relatively small sizes can be fabricated (radius of ∼5–50 μm). However, the microfluidic device structure is relatively complex, the generation frequency is low (∼2–20 Hz), and the resultant ATPS droplets have low uniformity in size (CV of 5%–20%).
Assembled Evaporation Pump
In this method, an assembled evaporation pump replaces the syringe pump to provide stable and ultra-low flow rates for ATPS droplets generation.84 The experimental setup is constructed by a hybrid glass-PDMS microdevice, whose inlets are connected to solution-filled wide-mouth bottles, and the outlet is connected to an assembled evaporation pump. The evaporation pump is constructed by connecting an evaporation tube (EPt) and an output tube (OPt) at each end of a cache chamber using two connectors, and the other end of the Opt is sealed by a plug [Fig. 10(e)]. The whole evaporation pump is filled with ethyl acetate (EA) and blocked with a sealing plug. When the EA in the EPt penetrates the wall and evaporates, a negative pressure will generate at the open end of the OPt. This negative pressure will suck the inner and outer phases flow into the microdevice. Due to the function of the cache chamber, the evaporation area and the negative pressure remain constant, which enables flow rates stable. The EPt length is used to control the flow rate of the evaporation pump, which can be varied from 0.033 to 0.41 μl min−1, as the evaporation tube length changes from 500 μm to 1.0 cm. The droplet size (22–46.5 μm in radius) can be controlled by tuning the flow rate of the evaporation pump, the molecular weight of dextran, and the concentration of PEG solution. Because both the dispersed and continuous phases are driven by the same negative pressure, the flow rate of the two phases cannot be changed independently. Therefore, the generated ATPS droplet size shows a small variation when changing the flow rates of the evaporation pump. The generation frequency is low due to the low flow rates of the evaporation pump. Although a 3D PDMS microfluidic device with eight junctions is fabricated to improve the generation frequency, its maximum frequency is only ∼60 Hz (∼7–8 Hz for each unit) [Fig. 10(e-ii)]. Using assembled evaporation pump, constant and low flow rate can be obtained for ATPS droplet generation spontaneously without any external components.
3. Transient oil-medium
In the transient oil-medium methods, an additional oil phase is used to facilitate ATPS droplets formation, which is then removed by on-chip separation,51 dewetting,85 or evaporation76 after ATPS droplets are generated. Note that the generation of ATPS droplets can be synchronized with the generation of oil droplet choppers,51 the water-in-oil-in-water (W/O/W) double emulsion droplets85 or the water-in-water-in-oil (W/W/O) double emulsion droplets.76 Therefore, the generation of low-interfacial-tension ATPS droplets can be manipulated by controlling the generation of high-interfacial-tension oil-in-water droplets. The transient oil-medium methods have the following advantages: the simple device (flow-focusing glass or PDMS device), low cost, a wide range of generation conditions, and high-generation frequency. However, the added oil phase must be removed after droplet generation.
a. Oil-droplet chopper
To assist the ATPS droplet generation, an additional dispersed oil phase is introduced in the “oil droplet chopper” method.51 Due to the high interfacial tension of water–oil systems, oil-in-water droplets can be easily produced. The generated oil droplets perturb and distort the aqueous interface when they pass through the narrow orifice of the collection capillary. As each oil droplet perturbs the aqueous interface to break off one aqueous droplet, the all-aqueous droplet generation is synchronized with the oil droplet chopper. The glass capillary microfluidic device is fabricated by coaxially aligning a theta-shaped injection capillary and a rounded collection capillary inside a square glass capillary in which the right collection capillary is connected to a theta-shaped glass capillary whose two side channels are aligned vertically in position. As such, the generated all-aqueous droplets and oil droplet choppers can be separated due to the density difference, where the lighter oil droplet choppers and heavier all-aqueous droplets vent into the upper and lower channels, respectively [Figs. 11(a-i)–11(a-iv)]. Four generation regimes are identified: jet regime when the oil phase is not introduced; non-uniform all-aqueous droplets when fo (oil chopper generation frequency) is smaller than flow (the lower critical frequency flow); uniform all-aqueous droplets when flow < fo < fup, where fup is the upper critical frequency; formation of a continuous thread when fo > fup. It demonstrates that the generation of all-aqueous droplets is out of synchronization with the generation of oil choppers in the non-uniform regime, the generation of both is synchronized in the uniform regime, and the generation of oil choppers is too fast for the aqueous interface to break off in the thread regime. In the synchronized uniform regime, the generation frequency and droplet size can be independently tailored by simply adjusting the flow rates of the dispersed oil-chopper and aqueous phases, respectively. All-aqueous emulsion droplets with a wide range of droplet size (radius: 5–180 μm), high uniformity (CV: 0.75%–2.45%), and a maximum generation frequency (∼2137 Hz) have been generated by this method. In summary, the “oil droplet chopper” method combines the advantages of active and passive methods for ATPS droplet generation: no need of external components, simple microfluidic device, ease of implementation, a wide range of generation conditions, and independently tunable generation frequency and droplet size.
FIG. 11.
Transient oil-medium methods. (a) Oil droplet chopper. (i) Schematic of the glass capillary microfluidic device for the ATPS droplet generation, transportation, and the on-chip separation of the ATPS droplets and oil droplet choppers. (ii) The ATPS droplet generation at the junction. (iii) The transport of ATPS and oil droplets in the collection capillary. (iv) Separation of ATPS droplets and oil droplet choppers into the lower and upper channel in the θ-shaped capillary, respectively. Reproduced with permission from Zhou et al., Lab Chip 17, 3310 (2017). Copyright 2017 Royal Society of Chemistry.51 (b) Transient double emulsion. (i) Transient water-in-oil-in-water (W1/O/W2) double emulsion droplets are generated in a glass capillary microfluidic device. (ii) The destabilization process of the transient double emulsion. (iii) Images of the four generation regimes of the ATPS droplet generation. Reproduced with permission from Zhou et al., Lab Chip 21, 2684 (2021). Copyright 2021 Royal Society of Chemistry.85 (c) Oil-mediated method. (i) The oil-mediated passive flow-focusing method for water-in-water-in-oil double emulsion generation. (ii) Transferring process of the double emulsions into a PEG-rich bath for releasing the inner core droplets from the double emulsions. Reproduced with permission from Nan et al., Microsyst. Nanoeng. 6, 70 (2020). Copyright 2020 Springer Nature.76
Transient Double Emulsion
The transient double emulsion method is proposed to generate ATPS droplets with controllable and prolonged stability. In this method, a middle oil phase is introduced to facilitate the generation of ATPS droplets and control their generation frequency and size. It involves two steps: the transient water-in-oil-in-water double emulsion droplets are generated in a glass capillary microfluidic device [Fig. 11(b-i)], and then, the unstable transient double emulsion dewets into oil-in-water and water-in-water ATPS droplets, driven by the negative spreading parameter of the middle oil phase [Fig. 11(b-ii)]. The demulsification process of the transient double emulsion can also be divided into two steps: first, the aqueous core droplet moves toward the shell interface between the middle oil phase and outer aqueous phase within a time t1 until the three phases come into contact; then, the middle oil phase dewets the aqueous core droplet in a time t2. The total demulsification time t equals t1 + t2. Time t1 is determined by the viscosity of the middle oil phase, the density difference between the inner aqueous and the middle oil phase, and the relative size of the aqueous core droplet inside the middle shell phase. The dewetting time t2 is related to the spreading parameter. Since the droplet movement time t1 is much larger than the dewetting time t2, the demulsification time is dominated by time t1. Therefore, the demulsification time can be controlled by using the middle oil phase with different viscosities.85 Four generation regimes can occur: the oil-chopper, thread, double emulsion with one core, double emulsion with multiple cores regimes [Fig. 11(b-iii)]. In the double emulsion with one core regime, the generation frequency of the ATPS droplets is the same as that of the transient double emulsion droplets, which increases with the increase in the flow rate of any phase. The droplet size of the ATPS droplets can also be controlled by tuning the flow rates. The radius increases by increasing the flow rate of the inner phase and decreases with the increase in the middle oil phase flow rate or the increase in the outer phase flow rate. By choosing the middle oil phase with different viscosities, the stability of the ATPS droplets can be precisely controlled. To obtain ATPS droplets immediately after generation, low-viscous oil should be used as the middle oil phase. High-viscous oil is needed to encapsulate the aqueous core droplet for prolonged storage and on-demand release. For an even longer stabilization time, a PDMS oil mixture with high viscosity is used as the middle oil phase, and the transient double emulsion droplets are encapsulated into microfibers. In this way, the transient double emulsion can maintain stability for several weeks and can be used to obtain the ATPS droplets after washing away the microfiber shell.
Oil-Mediated Method
The additional oil phase can be used as the continuous phase to rapidly generate water-in-water-in-oil double emulsion droplets and then removed to release the inner cores for the generation of ATPS droplets [Fig. 11(c-i)]. An oil evaporation-based method is used for removing the oil phase and releasing the inner cores to generate ATPS droplets [Fig. 11(c-ii)].76
4. Others
a. Water-head-driven microfluidic oscillator
In the water-head-driven microfluidic oscillator method, ATPS droplets are generated with on-chip microfluidic oscillators driven by water-head pressure.86 The PDMS water-head-driven microfluidic oscillator, fabricated by multilayer soft lithography, consists of a top layer, a 30-μm-thick membrane middle layer, and a bottom layer, thereby forming microfluidic valves, membranes serving as mechanical capacitors, and microchannels serving as fluidic resistors [Fig. 12(a-i)]. Microfluidic valves cut off flows through the elastic deformation of the valve membranes, and mechanical capacitors release the pressure through the capacitor membranes. For ATPS droplets generation, the dispersed phase is regulated by the valve, whose opening and closing are controlled by the water-head pressure-driven microfluidic oscillator. Pressure sensors are connected to the microfluidic device via microtubing for measuring pressure profiles [Fig. 12(a-ii)]. To obtain constant input pressure PI and output pressure PO, the height difference between the input well and the reference well, and the height difference between the output well and the reference well should be kept constant. The droplet size can be controlled by adjusting the switching period T and flow rate Q. Although the water-head-driven microfluidic oscillator method does not require any off-chip controller, the PDMS microfluidic device structure is complex, the droplet generation process is difficult to manipulate, the droplet size is big (radius of ∼90–160 μm), and the generation frequency (∼0.1 Hz) is low.
FIG. 12.
Other methods. (a) Water-head-driven microfluidic oscillator. (i) Schematic of the microfluidic device. (ii) Image of the integrated oscillator and valve. Reprinted with permission from V. B. Dang and S. J. Kim, Lab Chip 17, 286 (2017). Copyright 2017 Royal Society of Chemistry.86 (b) Interface fingering instability. (i) Schematic of the experimental setup. (ii) Schematic of the destabilization process: water diffusion, dense layer formation, fingering emergence, and fingers breakup. (iii) Phase diagram based on the initial concentration of dextran (ΦDEX) and PEG (ΦPEG) solutions. Reproduced with permission from Chao et al., Langmuir 34, 3030 (2018). Copyright 2018 American Chemical Society.87
b. Interfacial fingering instability
In the interface fingering instability method, all-aqueous droplets are generated by the spontaneous breakup of the instability fingers at the non-equilibrium water–water interface, when a less dense aqueous phase is placed on a denser aqueous phase in a vertical Hele–Shaw cell.87 The experimental setup consists of a vertically orientated Hele–Shaw cell, a light-emitting diode (LED) panel, and a long-range camera [Fig. 12(b-i)].87 The evolution of the non-equilibrium aqueous interface into all-aqueous emulsion droplets includes three stages [Fig. 12(b-ii)]: water diffuses from the upper dextran phase into the lower PEG phase, and a denser dextran layer is formed (stage I); fingering instability emerges and keeps growing (stage II); finally, the fingers break into an array of droplets (stage III). Here, the initial dextran concentration (ΦDEX) and PEG concentration (ΦPEG) are the critical parameters that affect the fingering instability and all-aqueous droplet generation [Fig. 12(b-iii)]. The non-equilibrium effect plays a critical role in all-aqueous droplets generation in the interface fingering instability method. This method is robust, easy to implement, and the generation frequency is high. However, the resultant droplet size is relatively large (radius of ∼250 to ∼750 μm), and the size uniformity is poor.
IV. ALL-AQUEOUS COMPLEX EMULSION DROPLETS GENERATION
In addition to all-aqueous single emulsion droplets, all-aqueous complex emulsion droplets with core-shell and Janus structures are also useful for various biomedical applications. For the generation of all-aqueous core-shell emulsion droplets, both phase-separated (phases in equilibrium) and single-phased (phases not in equilibrium) ATPSs have been used: typically, phase-separated ATPSs are used in the one-step methods, and the single-phased ATPSs are used in the two-step methods, where the phase-separation process is exploited to form the core-shell emulsion structures. For all-aqueous Janus emulsion droplets, the single-phased ATPSs are used by which in-droplet phase separation produces the Janus structure.
A. Generation of all-aqueous double-droplet systems
All-aqueous double-droplet systems refer to droplets consisting of two immiscible aqueous compartments dispersed in a third immiscible aqueous solution. The configurations of the double-droplet systems include complete engulfing for the core-shell structure, non-engulfing for two separated droplets, and partial engulfing for the Janus structure, as determined by the spreading coefficient.88 The spreading coefficient is defined as S = γjk − γij – γik, where γjk, γij, and γik are the interfacial tensions between different phases with i ≠ j ≠ k = o, i, m, and subscripts i, o, and m indicating the inner, outer, and middle phases, respectively.88,89 All-aqueous core-shell droplets have been prepared by the breakup of water-in-water-in-water (W/W/W) compound aqueous jet using a one-step method or by the phase separation of the single-phase droplets using a two-step method. All-aqueous Janus droplets have also been generated by the two-step method with phase separation. We summarize the detailed generation of all-aqueous core-shell droplets in Table V.
TABLE V.
Summary of the ATPS core-shell droplets generation methods. wt. %: Weight of polymer to the weight of the solution.
| Method | ATPSs | Device | CV | Year | Reference | |
|---|---|---|---|---|---|---|
| One-step method | Mechanical vibration method (active–Mechanical) | Inner phase: 15 wt. % dextran solution (T-500, 500 kDa); Middle phase: 17 wt. % PEG (8 kDa); Outer phase: 15 wt. % dextran solution (T-500, 500 kDa). | Co-flow and flow-focusing glass microcapillary device | ∼10% | 2012 | 59 |
| Mechanical solenoid valve method (active–mechanical) | Inner phase: Sodium carboxymethylcellulose; Middle phase: A mixture solution of sodium alginate and dextran; Outer phase: PEG solution. | Tube-in tube co-flow glass microcapillary device | … | 2019 | 22 | |
| Electrohydrodynamic (EHD) method (active–electrical) | A validated 3D numerical model is presented. | 2019 | 74 | |||
| Electrospray method (active–electrical) | Inner phase: 10 wt. % dextran (T-500, 500 kDa); Outer phase: 8 wt. % PEG (8 kDa) | A co-flow glass microcapillary device | … | 2013 | 43 | |
| Spontaneous generation method (passive) | Inner phase: 10 wt. % PEG (10 kDa) solution; Middle phase: 5 wt. % of both cross-linkable dextran (20 kDa and 500 kDa) solution; Outer phase: 40 wt. % PEG (10 kDa) solution. | A three-dimensional PDMS device with two junctions. | … | 2017 | 100 | |
| Hybrid microfluidic method (passive) | Inner phase: Dextran-rich phase (500 kDa); Middle phase: PEG-rich phase (35 kDa); Outer phase: Dextran-rich phase. Phase separated ATPSs from PEG and dextran solutions: PEG solution: 5%, 10%, 20%; Dextran solution: 6.4%, 12.8%, 25.6%. | A PDMS-based hybrid microfluidic with a microneedle-assisted flow-focusing design. | <5%–6% | 2020 | 90 | |
| Two-step method | Mechanical vibration + phase separation | Inner phase: Aqueous solution with 5 wt. % dextran (T-500, 500 kDa) and 1 wt. % PEG (8 kDa); Outer phase: Aqueous solution with 8 wt. % PEG (8 kDa) and 20 wt. % glycerol. (Glycerol increases the viscosity and density of the outer phase) | Co-flow and flow-focusing glass microcapillary device | <4% | 2012 | 60 |
| Piezo-electric bending disk + phase separation | Innermost phase: 7 wt. % PEG (10 kDa) solution; Inner phase: 10 wt. % dextran (500 kDa) solution; Middle phase: 10 wt. % dextran (500 kDa) solution; Outer phase: 7 wt. % PEG (10 kDa) solution | A PDMS microfluidic device with three junctions. | … | 2011 | 66 | |
| Electrospray + phase separation | Inner phase: A single-phase solution with 5 wt. % dextran (T-500, 500 kDa) and 1 wt. % PEG (20 kDa); Outer phase: Aqueous solution with 8 wt. % PEG (8 kDa) | A tapered glass capillary. | … | 2015 | 75 | |
1. All-aqueous core-shell droplets
a. One-step methods
In the one-step fabrication method, phase-separated (phases in equilibrium) ATPSs are used. The all-aqueous W/W/W compound jet can break up into all-aqueous core-shell droplets actively with the help of extra perturbation (such as the mechanical vibration,59 mechanical solenoid valve,22 electrohydrodynamic,74 and electrospray43 methods) or passively with hybrid microfluidic methods.90
All-aqueous core-shell droplets have been directly generated in one step with the help of the mechanical perturbation, provided by a mechanical vibrator that is connected to the tubing of the inner phase. Under the mechanical perturbation, the dextran-in-PEG-in-dextran compound jet breaks up and generates all-aqueous core-shell droplets directly [Fig. 13(a)]. The generated all-aqueous core-shell droplets have a polydispersity of ∼10%. The size and the inner droplet number can be controlled by tuning the input voltage and frequency of the mechanical vibrator, and the flow rates of different phases.59
FIG. 13.
One-step methods for all-aqueous core-shell droplets generation. (a) Mechanical vibration. A mechanical vibrator is connected to the microtubing of the dispersed phase. Reproduced from A. Sauret and H. C. Shum, Appl. Phys. Lett. 100, 154106 (2012), with the permission of AIP Publishing.59 (b) Mechanical solenoid valve. A solenoid valve connects to the tubing of the shell phase. Reproduced with permission from Zhu et al., ACS Appl. Mater. Interfaces 11, 4826 (2019). Copyright 2019 American Chemical Society.22 (c) Electrohydrodynamic method. Simulated ATPS core-shell droplets generation by the electrohydrodynamic method. Reproduced with permission from Azizian et al., Chem. Eng. Sci. 195, 201 (2019). Copyright 2018 Elsevier Ltd.74 (d) Hybrid microfluidic method. The hybrid device is composed of a microneedle in the first inlet and an embedded glass capillary inside the microchannel between the first and second cross-junction for all-aqueous core-shell droplets formation. (ii) Schematic of the all-aqueous core-shell droplets generation. Reproduced with permission from Jeyhani et al., Small 16, 1906565 (2020). Copyright 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.90
An oscillating solenoid valve is used to control the injection of the middle phase in the mechanical solenoid-valve method [Fig. 13(b)].22 The solenoid valve is installed on the tubing connecting to the middle shell phase to close the microchannel periodically. The middle shell phase can flow continuously if the solenoid valve is off, while when the valve was on, the shell phase channel is closed, and the fluids are squeezed to produce the all-aqueous core-shell droplets. The core diameter and the shell thickness can be controlled by the flow rates and the oscillating frequency.
A validated 3D numerical model is presented for all-aqueous core-shell droplets generation using the electrohydrodynamic method [Fig. 13(c)].74 In this method, three co-flow phases are injected into the microchannel, where the middle phase serves as a covering shell for the inner droplet. When turning off the electric voltage, the middle-phase will be attracted toward the positively charged electrode. The shearing stress pinches off the compound liquid jet of the inner and middle phases, thereby forming the all-aqueous core-shell droplets.
Recently, a hybrid microfluidic method has been proposed to generate all-aqueous core-shell droplets based on ATPSs by introducing pressure-driven flows. This method uses a conventional microfluidic flow-focusing geometry coupled with a coaxial microneedle and a glass capillary embedded in PDMS flow-focusing junctions. The PDMS-based hybrid microfluidic device setup enables the controlled generation of core-shell droplets by injecting the phase-separated equilibrated ATPSs using pressure pumps [Fig. 13(d)].90
In addition, all-aqueous core-shell droplets have also been directly fabricated using the electrospray approach, where a cylindrical glass capillary with a tapered tip is coaxially inserted into another tapered square capillary, forming a co-flowing geometry microfluidic device.43 In the electrospray method, the outer PEG-rich phase of the dextran-in-PEG compound jet is charged by a high-voltage power supply. The compound dextran-in-PEG jet is forced to go through a ring-shaped counter electrode under electrostatic forces. Upon the breakup of the compound jet, the all-aqueous core-shell droplets finally fall into the continuous phase of a dextran solution or directly fall on the surface of a solid substrate.41
b. Two-step methods
Single-phased ATPSs are used in the two-step methods, where the phase-separation process is exploited to form the all-aqueous core-shell droplets. The all-aqueous single-phased droplets are generated (the first step) first, which are used as the templates for the following phase separation (the second step). So far, mechanical vibration,60 piezo-electric bending disk perturbation,66 and electrospray75 methods have been used to combine with phase separation strategy for all-aqueous core-shell droplets fabrication, which is summarized in Table V and Fig. 14. The size and polydispersity of resultant all-aqueous core-shell droplets are controlled by their single-phased droplet templates.
FIG. 14.
Two-step methods for all-aqueous core-shell droplets generation. (a) Mechanical vibration. (i) All-aqueous single-phase droplets generation using mechanical vibration method. (ii) Schematic (top) and snapshots (bottom) of the all-aqueous core-shell droplets formation induced by phase separation. Reproduced with permission from Y. Song and H. C. Shum, Langmuir 28, 12054 (2012). Copyright 2012 American Chemical Society.60 (b) Piezo-electric bending disk. (i) Generation of all-aqueous single-phase droplets using piezo-electric bending disk method. (ii) Phase separation of the all-aqueous single-phase into core-shell droplets when flowing through the microchannel. Reproduced with permission from Ziemecka et al., Soft Matter 7, 9878 (2011). Copyright 2011 Royal Society of Chemistry.66 (c) Electrospray. (i) All-aqueous single-phase droplets generation via electrospray. (ii) The transition from all-aqueous single-phase to core-shell droplets by phase separation. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75
A single-phase aqueous solution with 5 wt. % dextran and 1 wt. % PEG is used as the dispersed phase, while an aqueous solution with 8 wt. % PEG and 20 wt. % glycerol is used as the continuous phase. First, the all-aqueous single-phase droplets are generated in a co-flow glass capillary microfluidic device by mechanical perturbation, provided by a sinusoidal-wave generator controlled mechanical vibrator [Fig. 14(a-i)]. Then, water is extracted from the single-phased droplets to the surrounding continuous phase as the generated droplets move downstream, leading to the shrinking of the droplet size and increasing the concentration of both dextran and PEG in the droplet phase. When the concentrations reach the critical concentrations of phase separation, small PEG-rich droplets start to form inside the all-aqueous single-phased droplet templates. As water is continuously extracted from the droplet phase, the inner PEG-rich droplets grow bigger and then coalesce with each other, finally yielding the all-aqueous core-shell droplets having a single PEG-rich core [Fig. 14(a-ii)].60
All-aqueous core-shell droplets can also be fabricated by using a PDMS-on-glass microfluidic device that is integrated with a piezo-electric bending disk. The piezo-electric bending disk perturbation facilitates the generation of all-aqueous single-phase droplets, which are used as the template for the all-aqueous core-shell droplets generation by combining with phase separation. In this work, the microdevice consists of three flow-focusing junctions in series, in which four aqueous streams are injected, two aqueous solutions of 10 wt. % dextran and two aqueous solutions of 7 wt. % PEG [Fig. 14(b-i)]. The piezo-electric bending stirs the innermost PEG and dextran phases that meet at the first junction and facilitates the all-aqueous single-phase droplets generation at the most downstream junction. Then, the all-aqueous core-shell droplets with a PEG core and a dextran shell are yielded a few centimeters downstream by the phase separation between PEG and dextran [Fig. 14(b–ii)].66 In another study, a single-phase aqueous solution with 5 wt. % dextran and 1 wt. % PEG is injected into an 8 wt. % PEG continuous phase to generate the all-aqueous single-phase droplets by electrospray method [Fig. 14(c-i)], which are then phase separated into two immiscible phases, yielding the all-aqueous core-shell droplets with a PEG-rich core and a dextran-rich shell [Fig. 14(c-ii)].75
2. All-aqueous Janus droplets
All-aqueous Janus droplets can be fabricated using two-step methods with phase separation. That is, the single-phase droplets are first generated, which are subsequently triggered to separate into two immiscible phases and form the all-aqueous Janus structures.91,92
As shown in Fig. 15(a), as the concentration of PEG/dextran is lower than the critical concentration for phase separation, the aqueous droplets keep the single phase. After depositing on a glass slide, the single-phase droplets are exposed to the air in which the concentration of PEG/dextran increases as the water evaporates into the air. Finally, the single-phase droplets are separated into two immiscible phases when the concentrations are above the critical concentrations for phase separation. The evaporation of water is the driving force for phase separation. The volume ratio between the two compartments of the all-aqueous Janus droplets can be tuned by the initial concentration ratio of PEG and dextran in the aqueous solution.91 In another work, all-aqueous Janus droplets are fabricated through solvent extraction-induced phase separation of the single-phase droplets. The morphologies of the all-aqueous Janus droplets can be tuned by using different ATPSs with different mass ratios [Fig. 15(b)].92
FIG. 15.
All-aqueous Janus droplets from ATPSs. (a) All-aqueous Janus droplets fabricated by evaporation-triggered phase separation. Reproduced with permission from Yuan et al., Macromol. Chem. Phys. 218, 1600422 (2017). Copyright 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.91 (b) Phase separation-induced formation of all-aqueous Janus droplets from different ATPSs. Reproduced with permission from Cui et al., Langmuir 33, 12670 (2017). Copyright 2017 American Chemical Society.92
B. Generation of all-aqueous high-order core-shell droplets
All-aqueous high-order core-shell droplets consist of more than two liquid layers, which have been generated by both the one-step hybrid microfluidic method90 and the two-step osmosis-driven phase separation method.
1. Hybrid microfluidic method
A PDMS-based hybrid microfluidic device with four inlets and two glass capillaries can be used to fabricate all-aqueous triple-layer core-shell droplets. The device consists of three flow-focusing cross-junctions, with a microneedle inserted into the first inlet followed by the embedding of two glass capillaries inside the microchannel between different cross-junctions. The glass capillaries help to prevent wetting of the microchannel walls by the dispersed phases. This hybrid device enables the controlled generation of all-aqueous triple core-shell droplets in one step based on equilibrated ATPSs. As shown in Fig. 16(a), the innermost dextran-rich phase is first introduced through a microneedle, then meets the middle PEG-rich phase and the middle dextran-rich phase when flowing into the first and second glass capillary, respectively, and finally forms a dextran-in-PEG-in-dextran compound jet when finally meeting the outer PEG-rich phase at the third cross-junction. The compound jet finally breaks up into all-aqueous triple-layer core-shell droplets.90
FIG. 16.
Generation of all-aqueous high-order core-shell droplets. (a) Hybrid microfluidic method. Schematic of the hybrid device for generating all-aqueous triple droplets in one step. Reproduced with permission from Jeyhani et al., Small 16, 1906565 (2020). Copyright 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.90 (b) Osmosis-driven phase separation. (i) Schematic of the co-flow glass capillary microfluidic device for the initial single-phase droplets generation and the subsequent multiple steps of phase separation. (ii) Optical images showing the phase separation of the single-phase droplets: transformation from single to double, triple, and eventually quadruple emulsions. (iii) Phase diagram of the complexity of the resultant all-aqueous droplets according to PEG and salt concentrations. Scale bars, 200 μm. Reproduced with permission from Chao et al., Small 14, 1802107 (2018). Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.93
2. Osmosis-driven phase separation method
Chao et al. have demonstrated that all-aqueous high-order core-shell droplets can be generated by taking all-aqueous single-phase droplets as templates, followed by osmosis-induced phase separation.93 In the first step, a salt-rich ternary aqueous mixture with the lower initial concentration and a PEG-rich aqueous solution with the higher initial concentration are used as the dispersed and continuous phases, respectively, to generate the single-phase all-aqueous droplets in a simple co-flow glass microfluidic device [Fig. 16(b-i)]. The flow is stopped immediately after the single-phase all-aqueous droplets are generated, and the evolution process is observed in the channel. Because of the lower initial concentration of the salt-rich solution in the droplet phase and higher initial concentration of the PEG-rich solution in the continuous phase, water continuously diffuses from the droplet to the surrounding continuous phase, which leads to the concentration increase in the droplet phase, triggering the phase separation of the droplet. The continuous water extraction combined with the availability of PEG and salts in the ternary aqueous single-phase droplets results in repeated phase separation. As a consequence, the initial single-phase droplet phase separates sequentially into double, triple, and eventually quadruple all-aqueous droplets [Figs. 16(b-i) and 16(b-ii)]. The complexity of the resultant all-aqueous droplets is determined by various parameters, such as the osmotic pressure difference between the droplet and continuous phases, the osmolality, and the solute concentration of the initial single-phase droplet. It can be seen from Fig. 16(b-iii), the resultant multiple-emulsion droplets can have different degrees of complexity: “single emulsion,” “double emulsion,” “triple emulsion,” and “quadruple emulsion” as the initial composition changes. The osmosis-driven phase separation method is simple and easy to implement, which only requires two initial aqueous solutions to generate multilayer all-aqueous emulsion droplets, without the requirement of multiple inlets to feed different solutes.
C. All-aqueous complex emulsion droplets from aqueous three-phase systems
In addition to ATPSs, aqueous three-phase systems have also been used to prepare complex emulsion droplets using the phase separation-induced technique.92,94 All-aqueous complex emulsion droplets with adjustable morphologies have been generated by simply extracting water from single-phase homogeneous droplets made up of an aqueous three-phase system comprising PEG, DEX, and polyvinyl alcohol (PVA). By changing the mass ratio of the three components, all-aqueous emulsion droplets with five types of morphologies, that is, binary-core/shell, core/shell-single phase Janus, ellipsoid Janus, multicore-in-matrix, and single core-double shell morphologies, can be obtained (Fig. 17).92
FIG. 17.
Complex all-aqueous emulsion droplets from aqueous three-phase systems. (a) Droplet geometry map shows the effect of the mass ratio of PEG/PVA/DEX on droplet structures. Regime I: multicore-in-matrix three-component droplets; regimes II and III: core/shell-single-phase Janus droplets; regime IV: ellipsoid three-component Janus droplets; regime V: binary-core/shell three-component droplets. (b) Schematic of the five kinds of all-aqueous complex emulsion droplets. Reproduced with permission from Cui et al., Langmuir 33, 12670 (2017). Copyright 2017 American Chemical Society.92
V. STABILIZATION OF ALL-AQUEOUS DROPLETS AND FABRICATION OF BIOMATERIALS
The stability of all-aqueous droplets is of utmost importance to practical applications. Due to the ultralow interfacial tension and thick water–water interface, however, the generated all-aqueous droplets are with poor stability, especially in compositions near the critical point.49 The all-aqueous droplet tends to coalesce with the adjacent droplets into bigger droplets or remix with the continuous phase for compositions near the critical point on the binodal curve. At the critical point, the composition difference between the droplet phase and the continuous phase vanishes, the interfacial tension approaches zero and the two immiscible phases remix into one single phase.49,95 The thickness of the water–water interface ranges from tens to hundreds of nanometers. Thus, small surfactant molecules (typically several nanometers long) are large enough to settle at the O/W interface but unable to straddle the thick W/W interfacial zone.55,56 Several strategies, such as solidification, interfacial complexation, and Pickering effect, have been developed to stabilize all-aqueous droplets, and various biomaterials can be generated simultaneously.
A. Solidification of all-aqueous droplets and droplet-templated materials
To obtain stabilized all-aqueous droplets, the dispersed phase of all-aqueous single emulsions or the shell phase of all-aqueous double emulsions can be selectively solidified and meanwhile generate hydrogel microparticles or microcapsules, respectively. This is the so-called solidification stabilization approach. The solidification process can be triggered chemically, optically, or thermally, as summarized in Table VI and Fig. 18.75
TABLE VI.
Solidification stabilization approaches for all-aqueous droplets. wt. %: Weight of polymer to the weight of the solution. DEX: Dextran; PEG: Poly (ethylene glycol); HEMA: 2-hydroxyethyl methacrylate; PEGDA: poly (ethylene glycol) diacrylate.
| Solidification approach | ATPSs | Solidification process | ATPS droplet-based materials | Reference |
|---|---|---|---|---|
| Chemical polymerization | Emulsion: 15 wt. % DEX (500 kDa) with 1 wt. % sodium alginate; Continuous: 17 wt. % PEG (20 kDa); Collection: A collecting pool with 1 wt. % CaCl2 solution | Chemical cross-linking between sodium alginate and calcium chloride | Calcium alginate microparticles | 25 |
| Emulsion: (5, 10, or 15 wt. %) DEX (500 kDa) with 1.5 wt. % sodium alginate; Continuous: 8 wt. % PEG (20 kDa) with 1.5 wt. % CaCl2 | Chemical cross-linking between sodium alginate and calcium chloride | Calcium alginate microparticles | 75 | |
| Emulsion: 1% (w/v) sodium carboxymethylcellulose; Middle: 1% (w/v) sodium alginate with 15% (w/v) DEX; Continuous: 30% (w/v) PEG (The outlet of the device was connected to a calcium chloride solution filled container.) | Chemical cross-linking between sodium alginate and calcium chloride. | Cell-laden core-shell hydrogel microcapsules | 22 | |
| Emulsion: 16 wt. % DEX (500 kDa) with 1 wt. % sodium carbonate; Continuous: 17 wt. % PEG (8 kDa) with 1 wt. % CaCl2 | Chemical reaction between sodium carbonate with calcium chloride. | Calcium carbonate microparticles | 61 | |
| Emulsion: 10 wt. % DEX (70 kDa), alginate-Ph, and HRP in PBS; Middle: 7.5 wt. % PEG (100 kDa) in PBS; Continuous: 7.5 wt. % PEG (100 kDa) and 1 mM H2O2 in PBS | Chemical cross-linking between HRP and H2O2 | Cell-laden microgel | 70 | |
| (HRP-catalyzed reaction) | ||||
| Emulsion: 3% gelatin with genipin; Continuous: 20% Maltodextrin (The 3% gelatin/20% maltodextrin emulsions with a given amount of genipin were prepared via homogenization.) | Cooling and chemical cross-linking Cooling and cross-linking with genipin. | Enzyme-loaded gelatin microgels | 29 | |
| Emulsion: 360 mg/ml DEX (32–45 kDa) in PBS (pH = 7.4 or 4.0); Middle 1: 90 mg/ml of TA-PEG in PBS (pH = 7.4 or 4.0); Middle 2: 90 mg/ml of TN-PEG in PBS (pH = 7.4 or 4.0); Continuous: 3 wt. % Solsperse 19000 in tridecane, or 3 wt. % Span 80 in decane, or 3 wt. % Span 80 in tridecane, or 3 wt. % Span 80 in mineral oil, or 3 wt. % Solsperse 19000 in decane, or 3 wt. % Solsperse 19000 in mineral oil | Cross-end coupling reaction between tetra-PEG-NHS (TN-PEG) and tetra-PEG-NH2 (TA-PEG) to synthesize tetra-PEG hydrogel microparticles or microcapsules | Tetra-PEG hydrogel: Core-shell microcapsules (pH: 7.4–7.8) Crescent microparticle (pH ≤ 7.0) | 96 | |
| Emulsion: 10 wt. % DEX-GMA with 2% (v/v) TEMED in 0.22 M KCl; Continuous: 10 wt. % PEG with 185 mg/ml ammonium peroxydisulfate (APS) | APS diffuses to the DEX-GMA-rich phase and react with TEMED to form the radicals that start the cross-linking for gel formation. | Hydrogel microparticles | 65 | |
| Emulsion: 16% Starch + 2% NaOH + 2%–10% TSTP; Continuous: 28%–44% PEG (10 or 20 kDa) | Maintain at 30 °C for 6 min to prepare water-in-water emulsion | Starch microspheres | 97 | |
| Chemical cross-linking the starch by trisodium trimetaphosphate (TSTP) | ||||
| Photo polymerization | PEGDA/dextran and PEGDA/K3PO4 systems | UV polymerization of PEGDA | PEGDA microparticles | 61 |
| Emulsion: 10 wt. % PEGDA (700 kDa); Continuous: 30 wt. % K3PO4 with 0.05 wt. % Irgacure 2959 | UV polymerization of PEGDA | PEGDA microparticles | 75 | |
| Emulsion: 20 wt. % DEX (40 kDa); Middle: 40 wt. % PEGDA (700 kDa); Continuous: HFE 7500 oil | Selective photo-polymerization of PEGDA phase and removal of the dextran phase of the dextran-PEGDA core-shell droplets | PEGDA microparticles | 98 | |
| Emulsion: DEX solution; Middle: PEGDA solution; Continuous: Hexadecane | Selective photo-polymerization of PEGDA phase and removal of the dextran phase of the dextran-PEGDA core-shell droplets | Crescent-shaped PEGDA particle | 99 | |
| Emulsion: 80 wt. % HEMA + 20 wt. % PEGDA and 3.5 wt. % UV photo-initiator; Continuous: 40 wt. % Na2SO4 + 2 wt. % NaNO2 | Situ UV polymerization | Spherical p-HEMA microgel particles | 68 | |
| Emulsion: 10 wt. % PEG (10 kDa); Middle: 5 wt. % DEX-GPE (500 kDa) and 5 wt. % DEX-SH (20 kDa); Continuous: 40 wt. % PEG (10 kDa) | Photo polymerization the DEX shell using the thiol–yne click reaction via UV light focused through the microscope objective lens | Hydrogel microcapsules | 100 | |
| Thermal polymerization | Emulsion: 10 wt. % DEX (500 kDa) with 1.5 mg/ml Collagen (at 4 °C); Continuous: 8 wt. % PEG (20 kDa) | Thermal gelation (heated at 37 °C for 12 h) of 1.5 mg/ml collagen precursors | Collagen microparticles | 75 |
| Collagen microcapsules |
FIG. 18.
Solidification stabilization approaches. (a) Chemical cross-linking. (i) and (ii) Chemical cross-link of sodium alginate in the dispersed phase with calcium chloride in the continuous phase yielding calcium alginate microparticles. (i) Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75 (ii) Reproduced with permission from Liu et al., Small 14, 1801095 (2018). Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.25 (iii) Chemical cross-link of sodium alginate in the shell phase with calcium chloride in the collection solution yielding calcium alginate microcapsules. Reproduced with permission from Zhu et al., ACS Appl. Mater. Interfaces 11, 4826 (2019). Copyright 2019 American Chemical Society.22 (iv) Tetra-PEG microcapsules by cross-end coupling reaction. Reproduced with permission from Watanabe et al., Langmuir 35, 2358 (2019). Copyright 2019 American Chemical Society.96 (b) Photo polymerization. (i) Photo polymerization by exposing the PEGDA-in-K3PO4 emulsion droplets to UV light. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75 (ii) and (iii) Crescent-shaped PEGDA microparticles fabricated by selectively photo-polymerizing the PEGDA phase and washing away the dextran phase. (ii) Reproduced with permission from Liu et al., Angew. Chem., Int. Ed. 58, 547 (2019). Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.99 (iii) Reproduced with permission from Ma et al., Small 8, 2356 (2012). Copyright 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.98 (iv) Thiol–yne photo-cross-linking of the dextran phase. Reproduced with permission from Mytnyk et al., RSC Adv. 7, 11331 (2017). Copyright 2017 Royal Society of Chemistry.100 (c) Thermal gelation. Thermal gelation of the collagen precursors in the dextran-in-PEG droplets. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75
1. Chemical crosslinking
Chemical cross-linking reaction is a common strategy to solidify the disperse phase of all-aqueous single emulsions or the shell phase of all-aqueous double-emulsion droplets, yielding microparticles or microcapsules, respectively [Fig. 18(a) and Table VI]. Two reagents are involved for chemical cross-linking: one is added into the droplet phase, while the other is added into the continuous phase or the collection solution, respectively. The chemical cross-linking takes place when the two reagents meet at the water–water interfaces.
The most commonly used reagents for chemical cross-linking are sodium alginate and calcium chloride, due to their hydrophilic, biocompatible, and biodegradable features. Generally, sodium alginate and calcium chloride are added into the dispersed and continuous phases, respectively, for solidifying all-aqueous single-emulsion droplets [Figs. 18(a-i) (Ref. 75) and 18(a-ii) (Ref. 25)]. For all-aqueous double-emulsion droplets stabilization, the sodium alginate is added into the shell phase, while calcium chloride is added into the collection solution [Fig. 18(a-iii)].22 After all-aqueous droplets are generated and collected, the Ca2+ ions diffuse across the water–water interface and cross-link the sodium alginate in the dispersed phase of the all-aqueous single-emulsion droplets obtaining calcium alginate microparticles [Figs. 18(a-i) and 18(a-ii)] or cross-link the sodium alginate in the shell phase of the all-aqueous double-emulsion droplets yielding calcium alginate microcapsules [Fig. 18(a-iii)].
The cross-end coupling reaction between pentaerythritol tetra(succinimidyloxyglutaryl) (tetra-PEG-NHS, and TN-PEG) and pentaerythritol tetra(aminopropyl)polyoxyethylene (TA-PEG) has also been used to chemically cross-link the tetra-arm poly(ethylene glycol) (tetra-PEG), stabilize the all-aqueous droplets and fabricate the tetra-PEG core-shell microcapsules or tetra-PEG crescent microparticles [Fig. 18(a-iv)].96 Both tetra-PEG microcapsules with a core-shell structure and tetra-PEG crescent microparticles can be produced by adjusting the pH value. When manipulating pH value between 7.4 and 7.0, the cross-end coupling reaction of the tetra-PEG macromolecules is fast, and thus, tetra-PEG microcapsules with core-shell structure are produced. When pH is ≤7.0, the cross-end coupling reaction is slow, and thus, the core-shell structure will change to the Janus structure during the chemical reaction, and finally generate tetra-PEG crescent microparticles.
In addition, the precipitation reaction between sodium carbonate and calcium chloride,61 the H2O2 triggered horseradish peroxidase(HRP)-catalyzed cross-linking between alginate-Ph molecules,70 the react between ammonium peroxydisulfate (APS) and N,N,N′,N′-tetramethylethylenediamine (TEMED),65 and the cross-linking between starch and trisodium trimetaphosphate (TSTP)97 have also been used to stabilize the all-aqueous droplets (as summarized in Table VI).
2. Photo polymerization
In the photo polymerization stability approach, the photo-curable precursor in the disperse phase of all-aqueous single emulsions or the shell phase of all-aqueous double-emulsion droplets can be solidified by exposing them to a UV light, yielding microparticles or microcapsules, respectively [Fig. 18(b) and Table VI].
The most commonly used photo-curable precursor is poly-(ethylene glycol) diacrylate (PEGDA) for the all-aqueous system. PEGDA instead of PEG is used as the disperse phase of all-aqueous single emulsions or the shell phase of all-aqueous double-emulsion droplets. Spherical PEGDA microparticles [Fig. 18(b-i)]75 or crescent-shaped PEGDA microparticles [Figs. 18(b-ii) and 18(b-iii)]98,99 with a concave structure can be fabricated by photo polymerization under UV light exposing.
Other photo cross-linking strategies, such as thiol–yne photo cross-linking of the dextran shell phase, have been used to stabilize the all-aqueous double-emulsion droplets for producing core-shell microparticles [Fig. 18(b-iv)]100 or photo-curable 2-hydroxyethyl methacrylate (HEMA) and PEGDA monomers66 can be exploited to stabilize the all-aqueous emulsion droplets.
3. Thermal gelation
By adding thermally curable precursors, the all-aqueous droplets can be solidified by thermal gelation [Fig. 18(c) and Table VI]. In all-aqueous systems, the most commonly used thermal curable precursor is collagen, which is spontaneously partitioned into the dextran-rich phase in the PEG/dextran system and generates collagen microparticles after thermal gelation at 37 °C.75
4. All-aqueous droplets-templated materials
It shows that the all-aqueous droplets can be exploited as templates for biomaterial fabrication. In the solidification stabilization approach, the all-aqueous droplets can be converted to microparticles or microcapsules by adding different additives into different phases of the all-aqueous systems. Therefore, by taking all-aqueous droplets as templates, combined with droplet solidification, various kinds of microparticles and microcapsules have been generated (Table VI and Fig. 19).
FIG. 19.
All-aqueous droplet-based materials. (a) Microparticle fabrication by solidifying the dispersed phase of the all-aqueous single-emulsion droplets. (i) Schematic of the cell-loaden microparticle fabrication by solidifying the dispersed phase. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75 (ii) Cell-loaden alginate hydrogel microbeads formed by chemical cross-linking of the sodium alginate. Reproduced with permission from W. H. Tan and S. Takeuchi, Adv. Mater. 19, 2696 (2007). Copyright 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.101 (iii) Starch microspheres generated by chemical cross-linking. Reproduced with permission from Li et al., Carbohydr. Polym. 88, 912 (2012). Copyright 2012 Elsevier Ltd.97 (iv) Spherical microgel particles produced by photo polymerization of the PEGDA. Reproduced with permission from Tarameshlou et al., Int. J. Polym. Mater. Polym. Biomater. 63, 884 (2014). Copyright 2014 Taylor & Francis Group, LLC.68 (v) Gel microparticles with a socket structure, fabricated by selective photo polymerization. Reproduced with permission from Ma et al., Small 8, 2356 (2012). Copyright 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.98 (vi) Crescent-shaped hydrogel microparticle fabricated via cross-end coupling reaction. Reproduced with permission from Watanabe et al., Langmuir 35, 2358 (2019). Copyright 2019 American Chemical Society.96 (b) Microcapsules fabrication by solidifying the shell phase of the all-aqueous double-emulsion droplets. (i) Schematic of the collagen microcapsules fabrication based on PEG-in-dextran-in-PEG double-emulsion droplets. Reproduced with permission from Song et al., ACS Appl. Mater. Interfaces 7, 13925 (2015). Copyright 2015 American Chemical Society.75 (ii) Image of dextran-shell microcapsule with a PEG core, generated by photo polymerization of the dextran shell using the thiol–yne click reaction via UV light. Reproduced with permission from Mytnyk et al., RSC Adv. 7, 11331 (2017). Copyright 2017 Royal Society of Chemistry.100 (iii) Tetra-PEG hydrogel microcapsules generated by the end-coupling reaction. Reproduced with permission from Watanabe et al., Langmuir 35, 2358 (2019). Copyright 2019 American Chemical Society.96 (iv) Images of the calcium alginate microcapsules generated by chemical cross-linking. Reproduced with permission from Zhu et al., ACS Appl. Mater. Interfaces 11, 4826 (2019). Copyright 2019 American Chemical Society.22
When exploiting all-aqueous single-emulsion droplets as templates, one additive for solidification is added in the disperse phase, ensuring the transition of droplets to microparticles [Table VI and Fig. 19(a)]. The fabricated microparticles can have a spherical structure [Figs. 19(a-i)–19(a-iv)],68,75,97,101 a socket structure [Fig. 19(a-v)]98 or a crescent-shaped structure by controlling the chemical cross-ending coupling reaction at pH ≤ 7.0 [Fig. 19(a-vi)].96
When taking all-aqueous double-emulsion droplets as templates followed by solidifying the shell phase, various kinds of microcapsules can be fabricated, which are summarized in Fig. 19(b) and Table VI. By adding thermal-curable collagen in the shell phase of the PEG-in-dextran-in-PEG double-emulsion templates, the collagen microcapsules can be fabricated by thermal gelation of the shell phase and have been used for protein separation [Fig. 19(b-i)].75 Microcapsules, with a PEG core and a dextran shell, are fabricated by photo polymerizing the dextran phase by the thiol–yne click reaction via UV light [Fig. 19(b-ii)].100 Tetra-PEG hydrogel microcapsules can be generated by the end-coupling chemical reaction when keeping pH at 7.4–7.8 [Fig. 19(b-iii)].96 Calcium alginate microcapsules are generated by chemically cross-linking the sodium alginate shell with calcium chloride [Fig. 19(b-iv)].22 Crescent-shaped microparticles with a concave structure have also been fabricated by taking all-aqueous double-emulsion droplets as templates, followed by selectively photo-polymerizing the PEGDA shell and washing away the dextran core phase [Figs. 19(a-v) and 19(a-vi)].96,98
B. Interfacially based stabilization of all-aqueous droplets and interfaces-templated materials
In addition to the solidification stabilization strategy, all-aqueous droplets can also be stabilized by engineering the water–water interfaces and serve as templates to fabricate biomaterials with liquid cores, which are useful for active ingredients encapsulation. Recent research showed that both particle adsorption (Pickering emulsion effect) and complexation of polyelectrolytes (PEs) (interfacial complexation) on the water–water interfaces are effective strategies for all-aqueous droplet stabilization.
1. Stabilization with Pickering emulsion effect
Small molecules cannot stabilize the all-aqueous droplets because of their incapability of settlement at the thick water–water interface.55,56 Colloidal particles have been introduced to aggregate at the water–water interface to enhance the all-aqueous droplet stability, known as the Pickering emulsion effect stabilization approach, which has been studied extensively for water/oil systems in the last decade. The mechanism of particle retention at the W/W interface is due to the increase in the free energy (ΔG) when the interfacial area between the two aqueous phases is increased by the departure of the particle. By the adsorption of a spherical particle (with a radius of R) at the all-aqueous droplet surface, the area occupied by the particle is , where is the contact angle of the particles with the interface, which is determined by the difference of the interfacial tension of the particles with the two aqueous phases. ΔG equals to the reduction of the interfacial area multiplied with the interfacial tension95
| (3) |
The free energy change (ΔG) of adsorption of a particle (also known as adsorption energy) with other shapes (platelet, disk, rod, and fibril) in Pickering all-aqueous emulsions are summarized in Table VII. The adsorption energy depends on particle shape, particle size, the interfacial tension (γ) of the all-aqueous system, and the contact angle (θ) between the particles and the water–water interface.
TABLE VII.
Adsorption energy of differently shaped particles in Pickering effect stabilization approach.
| Particle shape | Adsorption energy | Parameters | References |
|---|---|---|---|
| Spherical | Where R is the radius of the particle, γ is the interfacial tension, and θ is the contact angle. | 55 and 110 | |
| Platelet/disk | Where R is the radius of the platelet or disk, γ is the interfacial tension, and θ is the contact angle. | 55 and 122 | |
| Rod | Where l is the length and b is the width of the rods, γ is the interfacial tension, and θ is the contact angle. | 122 | |
| Fibril | Where R is the radius and L is the length of the fibril, γ is the interfacial tension, and θ is the contact angle. | 10 |
Typically, for traditional water–oil systems, small nanoparticles (NPs) serve as good surfactants.102 However, because of the ultralow interfacial tension and larger interfacial thickness of the water–water interface, larger particles should be used to stabilize the all-aqueous droplets. Compared with water–oil systems, the Pickering effect is much smaller for the stabilization of all-aqueous droplets, due to the ultralow interfacial tension and thick interface feature of all-aqueous systems.95 It has been demonstrated that only when the diameter of the particle exceeds several hundred nanometers, can the adsorption energy be sufficient to stabilize the all-aqueous droplets.103 Therefore, the particle size is relatively large in Pickering all-aqueous emulsions. However, if the particle size is too large, the particles will not cover enough interfacial area, and the adsorption speed is too slow.95 Therefore, the optimum particle size for the stability of all-aqueous droplets ranges from several hundred nanometers to several micrometers (as shown in Table VIII).
TABLE VIII.
Pickering effect stabilization approach for all-aqueous droplets. DEX: Dextran; PEG: Poly (ethylene glycol); PEO: Poly (ethylene oxide); LBG: Locust bean gum; and AMP: Amylopectin.
| ATPSs | Type of stabilizer | Particle diameter | Stability Period | Year | Reference | |
|---|---|---|---|---|---|---|
| A | B | |||||
| Methylcellulose | DEX | Fat (0.39μm), quartz (1.1μm), whey protein (1μm), probiotics, monoglycerides (0.21μm). | Fat: 0.39 μm; Monog lyceride:0.21 μm; | Over 1 week. | 2008 | 133 |
| Quartz: 1.1 μm; Whey protein: 1 μm | ||||||
| PEO | DEX | Spherical latex particles (0.2, 0.5, 1.0 and 2.0μm), protein particle (β-lactoglobulin). | Latex particle: 0.2 to 2.0 μm | / | 2012 | 109 |
| DEX | PEG | Polydopamine NPs and crosslink PDP via poly(acrylic acid) and carbodiimide. | / | Over 16 weeks. | 2019 | 129 |
| DEX | PEG | Carboxylated particles. | 1, 6, 10 μm | / | 2018 | 24 |
| DEX | Fish gelatin | Latex microparticles. | 1 to 12 μm. | / | 2018 | 130 |
| Waxy corn starch | LBG | Whey protein microgel particles. | ∼150 nm. | 1 to 7 days. | 2016 | 116 |
| PEO | Dextran | Spherical, PH-sensitive polymer microgels. | 120 to 440 μm. | At least 1 week | 2015 | 117 |
| DEX | PEG | Fluorescent amine-modified polystyrene latex beads. | 100 nm. | / | 2018 | 108 |
| Gelatin | Oxidized starch | Polystyrene latex. | 262 or 313 nm. | / | 2009 | 107 |
| Starch | LBG | Silica nanoparticles. | 20 nm. | ∼1 month. | 2014 | 106 |
| PEG | Na2SO4 | Dichlorodimethylsilane-modified silica nanoparticles | 20 nm | Over one year. | 2019 | 105 |
| Gelatin | DEX | Zein particles. | 160 to 200 nm. | Several days. | 2017 | 132 |
| PEO | DEX | Protein particles (β-lactoglobulin particles). | 30 to 640 nm. | At least 1 week. | 2013 | 110 |
| Xyloglucan (XG) | (AMP) | Polysaccharide-coated protein particles (Milk protein β-lactoglobulin). | 500 nm. | 3 days (PH < 5). | 2016 | 111 |
| PEO | DEX | Fractal aggregates or microgels of protein particles | 300 nm. | More than 2 weeks. | 2017 | 112 |
| PEO | DEX | Protein particle (β-lactoglobulin) morphology (microgel, fractal aggregates, fibril). | Microgel: 300 nm; Fractal :300 nm; | Microgel: 1 day | 2016 | 104 |
| Fibril: d = 5 nm, L = 1to20 μm. | Fractal, Fibril: >1week. | |||||
| PEO | DEX | Protein particle (β-lactoglobulin or whey protein isolate particles). | / | / | 2017 | 113 |
| DEX | Fish gelatin | Nanoplates (Gibbsite nanoplates) | 1:Wide: ∼170 nm, thick: ∼7 nm; | Several weeks. | 2015 | 55 |
| 2: Wide: ∼700m, thick: ∼35 nm. | ||||||
| PEO | Pullulan | Montmorillonite platelets. | Diameter: 200–2000 nm; Thick: 1 nm. | At least 21 days. | 2017 | 114 |
| PEG | DEX | Diamond-shaped Poly (lactide) platelets with a range of different sizes. | Up to Ca. 9.5 μm in length | 2 days. | 2017 | 115 |
| PEO | DEX | Nanorods: in the form of Cellulose nanocrystals (CNCs) | 160 nm × 6 nm × 6 nm. | Months. | 2016 | 122 |
| DEX | PEG | Protein nanofibrils (lysozyme fibrils). | Length: 600 nm, thick: 15 nm. | Over 30 days. | 2016 | 10 |
| PEO | DEX | Cellulose nanocrystals | Length: 180 nm. | / | 2018 | 121 |
| PEG | DEX | Polymer – protein conjugate particles. | 200 to 500 nm. | Over 1 week. | 2017 | 103 |
| PEO | DEX | Protein microgel/ polysaccharide complexes. | / | At least 1 week. | 2018 | 118 |
| DEX | PEG | Liposome stabilized. (Lipid vesicles). | ∼100 nm. | / | 2015 | 123 |
| DEX | PEG | PEGylated liposomes (negatively charged). | ∼130 nm. | / | 2014 | 9 |
| DEX | PEG | Giant lipid vesicles (PEGylated lipid). | 10 to 30 μm. | / | 2011 | 124 |
| DEX | PEG | Triblock polymers containing PEGMA, BuMA, and DMAEMA blocks) | / | > 6 months | 2013 | 126 |
| DEX | PEG | Two different amphiphilic diblock copolymers (PEG-PCL, DEX-PCL). | / | / | 2010 | 125 |
| Sodium caseinate | Xanthan | Emulsion droplets. | / | / | 2013 | 119 |
| Sodium caseinate | Xanthan | Emulsion droplets. | / | ∼10 days. | 2014 | 120 |
| Gelatin | maltodextrin | Single-celled microorganisms. | / | / | 2014 | 127 |
| Casein | xanthan | Living bacteria. | / | ∼1 week. | 2015 | 128 |
The particle morphology is also important for the Pickering stabilization effect. Gonzale-Jordan et al. have studied the influence of particle morphology on Pickering stabilization behavior by using β-lactoglobulin protein particles with different morphologies (microgel, fractal aggregates, or fibril).104 The results showed that the fractal aggregates provide a better stability effect, compared to microgels and fibrils, because the fractals can cover the larger interfacial area with a relatively low mass.
In addition to the particle size and morphology, the surface wettability of particles to the two aqueous phases is also an important parameter that affects Pickering stability.50 Binks and Shi have studied the stabilization effect of various particles with different hydrophilicity/hydrophobicity, by using hydrophilic calcium carbonate particles, wax particles, and partially hydrophobic fumed silica nanoparticles modified by dichlorodimethylsilane. They have demonstrated that the PEG/Na2SO4 ATPS emulsions can keep stable for over one year when taking the 20 nm partially hydrophobic fumed silica nanoparticles as stabilizers.105
In Pickering emulsion effect stabilization approach, a wide range of biocompatible particle-based agents are used as stabilizers for all-aqueous droplets, which are summarized in Table VIII and Fig. 20, including spherical particles (carboxylated particles,24 silica particles,106 polystyrene latex,107,108 or protein particles104,109–113 [Fig. 20(a)]), anisotropic platelets (gibbsite nanoplates,55 montmorillonite platelets,114 or diamond-shaped polylactide platelets115 [Fig. 20(b)]), deformable particles (microgels116–118 or oil droplets119,120), nanorods (protein nanofibrils10 or cellulose nanocrystals121,122[Fig. 20(c)]), hybrid particles (polymer–protein conjugate particles103 or protein microgel–polysaccharides complexes118 [Fig. 20(d)]), liposomes (lipid vesicles123,124 or PEGylated liposome9 [Fig. 20(e)]), copolymers (diblock copolymer125 or triblock copolymer126 [Fig. 20(f)]), microorganisms,127 living bacteria,128 and other kind of particles.129–133
FIG. 20.
Pickering emulsion effect for all-aqueous droplets stabilization. (a) Spherical particles as stabilizers. (i) Carboxylated particles. Reproduced with permission from Abbasi et al., Langmuir 34, 213 (2018). Copyright 2017 American Chemical Society.24 (ii) Fluorescent amine-modified polystyrene latex beads. Reproduced with permission from Douliez et al., Angew. Chem., Int. Ed. 57, 7780 (2018). Copyright 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.108 (iii) Protein (β-lactoglobulin) particles. Reproduced with permission from Das et al., Mater. Horiz. 4, 1196 (2017). Copyright 2017 Royal Society of Chemistry.113 (b) Platelets as stabilizers. (i) Montmorillonite platelets. Reproduced with permission from Ganley et al., J. Colloid Interface Sci. 505, 139 (2017). Copyright 2017 Elsevier.114 (ii) Diamond-shaped polylactide platelets. Reproduced with permission from Y. Chao and H. C. Shum, Chem. Soc. Rev. 49, 114 (2020). Copyright 2019 Royal Society of Chemistry.56 (c) Nanorods as stabilizers. (i) A monolayer of protein (lysozyme) nanofibrils. Reproduced with permission from Song et al., Nat. Commun. 7, 12934 (2016). Copyright 2016 Nature Publishing Group.10 (ii) Cellulose nanocrystals. Reproduced with permission from Ben Ayed et al., Langmuir 34, 6887 (2018). Copyright 2018 American Chemical Society.121 (iii) Nanorods (in the form of cellulose nanocrystals). Reproduced with permission from Peddireddy et al., ACS Macro Lett. 5, 283 (2016). Copyright 2016 American Chemical Society.122 (d) Hybrid particles as stabilizers. (i) The polymer–protein conjugate particles with biocatalytic activity. Reproduced with permission from Xue et al., ACS Macro Lett. 6, 679 (2017). Copyright 2017 American Chemical Society.103 (ii) Protein microgel/polysaccharides complexes. Reproduced with permission from Khemissi et al., Langmuir 34, 11806 (2018). Copyright 2018 American Chemical Society.118 (e) Liposomes as stabilizers. (i) Lipid vesicles as stabilizers. Reproduced with permission from Cacace et al., Langmuir 31, 11329 (2015). Copyright 2015 American Chemical Society.123 (ii) PEGylated liposome stabilizers. Reproduced with permission from Dewey et al., Nat. Commun. 5, 4670 (2014). Copyright 2014 Nature Publishing Group.9 (f) Triblock copolymers as stabilizers. Reproduced with permission from Buzza et al., Langmuir 29, 14804 (2013). Copyright 2013 American Chemical Society.126
In addition to particles, diblock and triblock copolymers have also been used as stabilizers for all-aqueous droplets' stability. For example, all-aqueous droplets have been stabilized by using a triblock copolymer (Pp–Bb–Dd), which includes three blocks such as poly[poly(ethylene glycol) methyl ether methacrylate] (PEGMA), poly(n-butyl methacrylate) (BuMA), and poly[2-(dimethylamino) ethyl methacrylate] (DMAEMA). For the triblock, it has two hydrophilic outer blocks Pp and Dd connected by a middle hydrophobic block Bb in which the Pp block has an affinity to the PEG-rich phase, while the Dd block has an affinity to the dextran-rich phase. To get optimal stability, triblock copolymers with different chain lengths are fabricated. It is hypothesized that the PEG-dextran ATPS droplets are stabilized by a monolayer of triblock copolymers coating on the water–water interface, as shown in Fig. 20(f).126 Another hypothesis is that the triblock copolymers aggregate and adsorb on the water–water interface just like particles.134
2. Interfacial complexation stabilization
In this method, the all-aqueous droplets are stabilized by the complexation between a pair of oppositely charged species, such as polyelectrolytes and nanoparticles, which are usually added into the dispersed and continuous phases, respectively. When the species meet at the water–water interface, they form different kinds of microcapsules with an aqueous core, as shown in Table IX and Fig. 21.135
TABLE IX.
Interfacial complexation stabilization approach for all-aqueous droplets. PE+: Positive charged polyelectrolyte; PE−: Negative charged polyelectrolyte; PSS-1: Poly (sodium 4-styrene sulfonate); PSS-2: polystyrene sodium sulfate; PDADMAC: Poly (diallyldimethylammonium chloride); PAH: Poly (allylamine hydrochloride); FITC-PAH: Fluorescently labeled poly (allylamine hydrochloride).
| Methods | ATPSs | Polyelectrolytes | ATPS Interface-based biomaterial | Year | Reference |
|---|---|---|---|---|---|
| Interfacial complexation of the oppositely charged polyelectrolytes | Emulsion: 15 wt. % DEX with 0.5 wt. % PDADMAC (or 0.64% PSS-1); Continuous: 10 wt. % PEG with 0.64% PSS-1 (or 0.5 wt. % PDADMAC) | PE+: PDADMAC PE−: PSS-1 | Polyelectrolyte microcapsule | 2016 | 26 |
| Emulsion: 15 wt. % DEX and 0.5 wt. % PDADMAC (PE+); Middle: 17 wt. % PEG solution; Continuous: 17 wt. % PEG with 1 wt. % PSS-2 (PE−) | PE+: PDADMAC PE−: PSS-2 | Polyelectrolyte microcapsules | 2016 | 136 | |
| Emulsion: 15 wt. % DEX with 0.5 wt. % PDADMAC (+); Middle: 17 wt. % PEG; Continuous: 17 wt. % PEG with PSS-2 (−) with or without SiO2 NPs (−) | PE+: PDADMAC PE−: PSS-2 | Polyelectrolyte microcapsules | 2019 | 28 | |
| Emulsion: 20 wt. % DEX (450–600 kDa) with PDADMAC (PE+); Continuous: 15 wt. % PEG (20 kDa) with PSS-1 (PE−) and/or SiO2 NPs(−) | PE+: PDADMAC PE−: PSS-1 SiO2 NPs (−) | PE/PE microcapsule; PE/NP microcapsule; PE/(PE,NP) microcapsule (PE/NP AWE-somes) | 2017 | 137 | |
| Emulsion: 15% DEX (450–600 kDa) with PDADMAC (PE+); Continuous: 10% PEG (20 kDa) with SiO2 NPs (−) | PE+: PDADMAC SiO2 NPs (−) | PE/NP AWE-somes | 2017 | 138 | |
| Partitioning-dependent assembly of the polyelectrolytes | Emulsion: 10 wt. % DEX (500 kDa) with FITC-PAH (PE+); Continuous: 8 wt. % PEG (8 kDa) with PSS-1 (PE−) | PE+: FITC-PAH PE−: PSS-1 | Polyelectrolyte microparticles (pH > 7) | 2018 | 139 |
| Emulsion: 10 wt. % DEX (500 kDa) with FITC-PAH (PE+); Continuous: 8 wt. % PEG (8 kDa) with PSS-1 (PE−) | PE+: FITC-PAH PE−: PSS-1 | Polyelectrolyte microcapsules (pH = 7) | |||
| Partitioning-dependent assembly of Ag NPs and polyelectrolytes | Emulsion: 10 wt. % DEX(500 kDa) with Amino-Ag NPs (+); Continuous: 8 wt. % PEG (8 kDa) with PSS-1 (PE−) | Amino-Ag NPs (+) PE−: PSS-1 | Hybrid PE-NPs microcapsules | ||
| Interfacial assembly of polyelectrolytes | Emulsion: DEX (500 kDa) with PSS-1 (PE−); Continuous: PEG (8 kDa) with PAH (PE+) | PE+: PAH PE−: PSS-1 | Polyelectrolyte microparticles | 2016 | 140 |
| Emulsion: DEX (500 kDa) with PAH (PE+); Continuous: PEG (8 kDa) with PSS-1 (PE−) | PE+: PAH PE−: PSS-1 | Polyelectrolyte microcapsules | |||
| Interfacial complexation polyelectrolytes at the W/W interfaces layer-by-layer | Emulsion: DEX(500 kDa) with PAH (PE+); Continuous: PEG (8 kDa) with PSS-1 (PE−) | PE+: PAH PE−: PSS-1 | Multilayered polyelectrolyte microcapsules |
FIG. 21.
Interfacial complexation stabilization approach for all-aqueous droplets and the templated interface-based materials. (a) Interfacial complexation of oppositely charged polyelectrolytes in ATPS single-emulsion droplets. (i) ATPS single-emulsion droplets generation by electrospray. Two kinds of polyelectrolytes (PE1 and PE2) are added into the dispersed and the continuous phases, respectively. (ii) The complexation locations: continuous phase (left), water–water interface (middle), and inner phase (right), depending on the relative mass flux of the electrolytes. (iii) and (iv) Microcapsules fabricated by electrospraying the dispersed phase into the continuous phase bath. Reproduced with permission from Hann et al., ACS Appl. Mater. Interfaces 8, 25603 (2016). Copyright 2016 American Chemical Society.26 (b) Interfacial complexation of oppositely charged polyelectrolytes in the middle phase of all-aqueous double-emulsion droplets. (i) All-aqueous double emulsion generation. (ii) The polyelectrolyte microcapsule structure. Optical (iii) and confocal (iv) microscope images of the polyelectrolyte microcapsules. Reproduced with permission from Zhang et al., Angew. Chem., Int. Ed. 55, 13470 (2016). Copyright 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.136 (c) Polyelectrolyte/nanoparticle (PE/NP) hybrid microcapsules from interfacial complexation. (i) (PE−/NP+) Hybrid microcapsules. The fabrication process (upper), SEM images (lower left), and the surface (lower right) of the hybrid polyelectrolyte/nanoparticle [PSS-1(−)/Ag NPs(+)] microcapsules. Reproduced with permission from Ma et al., Soft Matter 14, 1552 (2018). Copyright 2018 Royal Society of Chemistry.139 (ii) (PE+/NPs−) Hybrid microcapsules. Initial placement of nanoparticles and polyelectrolytes in ATPS (upper left), and the schematic of the interface (upper right). Bottom: Fluorescently labeled components are included during the formation of microcapsules (insets are the corresponding bright-field images). Reproduced with permission from Hann et al., ACS Appl. Mater. Interfaces 9, 25023 (2017). Copyright 2017 American Chemical Society.138 (iii) [PE/(PE,NPs)] Hybrid microcapsules. Initial and final configuration of PE/PE, PE/NP, and PE/(PE,NP) microcapsules. Reproduced with permission from Hann et al., Phys. Chem. Chem. Phys. 19, 23825 (2017). Copyright 2017 Royal Society of Chemistry.137 (d) Partitioning-dependent interfacial polyelectrolyte microcapsules fabrication. Reproduced with permission from Ma et al., ACS Macro Lett. 5, 666 (2016). Copyright 2016 American Chemical Society.140 (e) Partitioning-dependent interfacial polyelectrolyte microgel particle fabrication. Reproduced with permission from Ma et al., ACS Macro Lett. 5, 666 (2016). Copyright 2016 American Chemical Society.140
Polyelectrolyte microcapsules can be fabricated by taking all-aqueous single-emulsion droplets as templates, combined with the interfacial complexation stabilization strategy. The two kinds of oppositely charged polyelectrolytes are added into the dispersed and the continuous phase, respectively. For example, a solution of 15% dextran solution with 0.5% poly(diallyldimethylammonium chloride) (PDDAMAC, positive charged polyelectrolyte, PE+) is served as the disperse phase and injected into the electrospray microfluidic device, and then sprayed into a bath with the continuous phase of 10% PEG with 0.64% poly(sodium 4-styrene sulfonate) (PSS-1, negative charged polyelectrolyte, PE−), producing the all-aqueous dextran-in-PEG droplets. When the PDADMAC (PE+) in the droplet phase and the PSS-1 (PE−) in the continuous phase meet at the water–water interface, they form a membrane shell coating the all-aqueous droplets, thus fabricating polyelectrolyte microcapsules. The relative mass fluxes of the two polyelectrolytes should be tuned to ensure that the polyelectrolytes meet at the interface, or otherwise, the polyelectrolytes will meet and complexate in the inner phase (internal complexation) or the continuous phase (external complexation), as shown in Fig. 21(a).26
It is demonstrated that polyelectrolyte microcapsules can also be fabricated by taking the transient all-aqueous double-emulsion droplets as templates, coupled with the complexation of oppositely charged polyelectrolytes in the middle phase. In this scheme, the two kinds of oppositely charged polyelectrolytes are added into the inner phase and the outer phase. The middle phase is used to separate the oppositely charged polyelectrolytes before the all-aqueous double-emulsion droplets generation and serves as a “complexation zone” after the droplet generation, as shown in Fig. 21(b).28,135,136 The “complexation zone” enables that the complexation location no longer relies on the relative fluxes of the two polyelectrolytes. For instance, dextran-in-PEG-in-PEG transient all-aqueous double-emulsion droplets are fabricated by taking a 15% dextran solution with 0.5% PDADMAC (PE+) as the inner phase, a 17% PEG solution as the middle phase, and a 17% PEG solution with 1 wt. % polystyrene sodium sulfate (PSS-2, PE) as the outer continuous phase. Then, the PDADMAC (PE+) in the inner phase and the PSS-2 (PE−) in the outer phase meet and complexate in the middle phase, finally forming the polyelectrolyte microcapsules [Fig. 21(b-i)].136
To improve the stability, permeability, and the shell mechanical strength of the resultant polyelectrolyte microcapsules, charged nanoparticles (e.g., negatively charged silica nanoparticles,28,137,138 positively charged amino-Ag nanoparticles139) are incorporated, and hybrid polyelectrolyte/nanoparticle (PE/NP) microcapsules28,137,138 and polyelectrolyte/(polyelectrolyte, nanoparticle) [PE/(PE,NP)] microcapsules137 can be fabricated, as shown in Fig. 21(c).
Hybrid PE/NP microcapsules are fabricated, when one of the polyelectrolytes is replaced by nanoparticles with the same charge, to improve the permeability and mechanical stability of the microcapsule shells, as shown in Fig. 21. When oppositely charged amino-Ag nanoparticle (NP+) and PSS-1 (PE−) are added into the inner 10% dextran and the outer 8% PEG phases, respectively, the NP+ and PE− meet and complexate at the water–water interface, fabricating the hybrid PE/NP microcapsules [Fig. 21(c-i)].139 By the interfacial complexation of the positively charged PDADMAC (PE+) in the disperse phase and the negatively charged silica nanoparticles (NP−) in the continuous phase of the all-aqueous single-emulsion droplets, hybrid PE/NP microcapsules can also be fabricated.138 In another work, with the positively charged polyelectrolyte PDADMAC (PE+) in the inner phase and negatively charged silica nanoparticles or negatively charged polyelectrolyte PSS-2 (PE−) in the continuous phase of the all-aqueous double-emulsion droplets, polyelectrolyte microcapsules (PE/PE microcapsules), and hybrid polyelectrolyte microcapsules (PE/NP microcapsules) are fabricated via the interfacial complexation between PE+ and PE− and between PE+ and NP− with opposite charges, respectively.28 Interestingly, an internal aqueous droplet of the same material as the continuous phase is spontaneously formed, which is encapsulated in the all-aqueous droplet with a membrane shell. This structure is referred to as all water emulsion microcapsules (AWE-somes), as shown in Fig. 21(c-ii).137 For excellent mechanical strength of the microcapsule shells, polyelectrolyte/(polyelectrolyte, nanoparticle) [PE1/(PE2,NP)] microcapsules are fabricated by the interfacial complexation between PE1 in the disperse phase with the oppositely charged PE2 and NPs in the continuous phase, forming AWE-somes as well [Fig. 21(c-iii)].137
In addition to polyelectrolyte microcapsules, polyelectrolyte microgel particles have also been fabricated by the interfacial complexation of oppositely charged polyelectrolytes. It has demonstrated that the partitioning coefficients of the polyelectrolytes in the two aqueous phases of all-aqueous systems change with the pH value (see Table X), which changes the assembly and the complexation location of the polyelectrolytes, converting the resultant structure from microcapsules to microgel particles, as shown in Figs. 21(d) and 21(e).139,140 In this study, the positively charged fluorescently labeled poly(allylamine hydrochloride) (FITC-PAH, PE+) is mixed with the dispersed dextran-rich phase, while the negatively charged PSS-1 (PE−) is added into the continuous PEG-rich phase. At pH = 7.0, the PSS (PE−) in the continuous phase has a higher affinity with the dispersed dextran-rich phase, as indicated by its higher partitioning coefficient in the dextran-rich phase (0.69) than in the PEG-rich phase (0.31), while the FITC-PAH (PE+) shows only a slightly higher preference to the continuous PEG-rich phase by its partitioning coefficient of 0.48 and 0.52 in the dextran-rich and PEG-rich phases, respectively. Therefore, the polyelectrolyte PSS-1 (PE−) migrates inward from the continuous PEG-rich phase to the dispersed dextran-rich phase due to its high partition preference to the dextran-rich phase, while the FITC-PAH (PE+) shows weak migration out of the dextran-rich droplet phase. Therefore, the high partitioning preference of the PSS-1 (PE−) in the dextran-rich phase and the weak partitioning preference of FITC-PAH (PE+) in the dextran-rich phase have the two polyelectrolytes meet and complexate at the water–water interface, yielding polyelectrolyte microcapsules at pH = 7.0 [Fig. 21(d)].139,140
TABLE X.
Partitioning coefficients of the polyelectrolytes in the DEX-rich and PEG-rich phases at different pH values.139–141 PE+: Positive charged polyelectrolyte; PE−: Negative charged polyelectrolyte; PAH: Poly (allylamine hydrochloride); PDADMAC: Poly (diallyldimethylammonium chloride); PEI: Poly (ethylene imine); PAA: Poly (acrylic acid); PSS-1: Poly (sodium-4-styrene sulfonate) or Polystyrene sodium sulfate; FITC-PAH: Fluorescently labeled poly (allylamine hydrochloride); BSA: Bovine serum albumin; Hb: Hemoglobin; and IgG: Immunoglobulin G.
| Partitioning coefficient | DEX-rich phase | PEG-rich phase | |
|---|---|---|---|
| pH = 7.0 | PAH (PE+) | 0.49 | 0.51 |
| PDADMAC (PE+) | 0.48 | 0.52 | |
| PEI | 0.53 | 0.47 | |
| PAA | 0.67 | 0.33 | |
| PSS-1 (PE−) | 0.69 | 0.31 | |
| FITC-PAH (PE+) | 0.48 | 0.52 | |
| BSA (PE−) | 0.96 | 0.04 | |
| Hb (PE+) | 0.70 | 0.30 | |
| IgG (PE−) | 0.73 | 0.27 | |
| pH = 9.0 | PSS-1 (PE−) | 0.73 | 0.27 |
| FITC-PAH (PE+) | 0.57 | 0.43 | |
| pH = 5.0 | PSS-1 (PE−) | 0.63 | 0.37 |
| FITC-PAH (PE+) | 0.40 | 0.60 | |
As the partitioning coefficient of the polyelectrolytes varied with the pH value, the partition coefficients of both PSS-1 (PE−) and FITC-PAH (PE+) in the dextran-rich phase change when pH changes from 7.0 to 9.0. The partitioning coefficient of FITC-PAH (PE+) in the dextran-rich phase increase from 0.48 (at pH = 7.0) to 0.57 (at pH = 9.0), indicating the preference of FITC-PAH is changed from the PEG-rich phase to the dextran-rich phase. Meanwhile, the partitioning coefficient of PSS-1 has rarely been changed and the PSS-1 continues to migrate toward the dextran-rich droplet phase. Therefore, when controlling the pH value at 9.0, the FITC-PAH (PE+) will keep in the dextran-rich phase, and meanwhile, the PSS-1 (PE−) will migrate from the continuous PEG-rich phase to the dextran-rich droplet phase for its high partitioning coefficient in the dextran-rich phase (0.73 at pH = 9.0), leading to the two polyelectrolytes meet and complexate in the dextran-rich phase and yielding polyelectrolyte microgel particles [Fig. 21(e)].139,140 Experimentally measured partitioning coefficients of the commonly used polyelectrolytes in the dextran-rich and PEG-rich phases at different pH values are summarized in Table X.139–141
In conclusion, the interfacial complexation can be used for all-aqueous droplets stabilization. Both all-aqueous single-emulsion and double-emulsion droplets can be used as templates for the polyelectrolyte microcapsules fabrication. In all-aqueous single-emulsion templates, the relative fluxes of the two polyelectrolytes should be balanced for controlling the complexation location at the W/W interface, while in the all-aqueous double-emulsion templates, the middle phase separates the oppositely charged polyelectrolytes and served as the “complexation zone,” without a need to manipulate the relative mass fluxes of the two polyelectrolytes. The interfacial complexation of PE+ and PE− fabricates polyelectrolyte (PE/PE) microcapsules, while the interfacial complexation between PE+ with NP− and between NP+ with PE− fabricate hybrid polyelectrolyte/nanoparticle (PE/NP) microcapsules. By combing the PE/PE and PE/NP microcapsule schemes, hybrid PE1/(PE2,NP) microcapsules are also fabricated to enhance the mechanical strength of the microcapsule shells. In addition to the polyelectrolyte microcapsules, polyelectrolyte microgel particles can also be fabricated by the interfacial complexation between the oppositely charged polyelectrolytes, based on the differentiated partitioning of the two polyelectrolytes in the two aqueous phases, which can be tuned by changing the pH value.
3. All-aqueous interfaces-based materials
As discussed above, the Pickering emulsion effect and interfacial complexation strategies have been exploited for all-aqueous droplets stabilization. The water–water interfaces can serve as templates for the fabrication of biomaterials with an aqueous core, which is useful for active ingredients encapsulation.
In the Pickering emulsion effect stabilization approach (Table VIII, Fig. 20), a stable particle-layer is formed at the water–water interface, containing droplets of one aqueous phase and surrounded by another aqueous phase, by particle adsorption and aggregation on the aqueous interface. The corresponding generated interface-based biomaterials include protein microparticles,112 particle-stabilized droplets,24,103,115,132 liquid-like emulsions,118 emulsion gels,121,127 vesicles,123,124 cell spheroids,113 polymersomes,125,126 colloidosomes,108,129,133 and fibrillosomes.10
In interfacial complexation stabilization approach (Table IX, Fig. 21), oppositely charged polyelectrolytes and nanoparticles complexate and form a membrane at the water–water interface, yielding a variety of biomaterials, including polyelectrolyte (PE/PE) microparticles,138,139 polyelectrolyte (PE/PE) microcapsules,26,28,135,138,139 hybrid polyelectrolyte/nanoparticle (PE/NP) microcapsules, or polyelectrolyte/nanoparticle [PE/NP or PE/(PE, NP)] all water emulsion microcapsules (PE/NP AWE-somes).136–138
VI. SUMMARY
The selective separation capability, rapid mass transfer, and all-aqueous properties make all-aqueous systems promising candidates for various biomedical applications. However, the ultra-low interfacial tension and relatively thick water–water interface make all-aqueous droplet generation and stabilization challenging, respectively. Various strategies have been proposed for all-aqueous droplets generation and stabilization in the last few decades.
In this review, we have first introduced the properties of all-aqueous systems and then summarized recently developed microfluidic technologies for all-aqueous droplets generation including single, core-shell, Janus, and other complex structures, with a focus on the all-aqueous single-emulsion droplet generation. Both active and passive approaches for all-aqueous single-emulsion droplets generation are summarized in detail, including the generation mechanisms, controlling methods, microfluidic devices, droplet properties (such as CV value, radius, and generation frequency), the advantages and limitations. Based on all-aqueous single-emulsion droplets, all-aqueous core-shell and Janus droplets can be fabricated by taking all-aqueous single-phase emulsions as templates and combining with phase separation by tuning the composition of all-aqueous systems.
The three most commonly used stabilization strategies (solidification, Pickering effect, and interfacial complexation) for all-aqueous droplets and their correspondingly resultant all-aqueous emulsion-based materials are summarized. In general, the solidification approaches stabilize the all-aqueous emulsion droplets by solidifying the inner or the middle phase chemically, optically or thermally; the Pickering effect stabilizes the all-aqueous droplets through the particle-layer formed by the particle adsorption on the water–water interface; the interfacial complexation approaches stabilize the all-aqueous emulsion droplets by forming a membrane at the water–water interface via interfacial complexation between oppositely charged polyelectrolytes or using one polyelectrolyte with oppositely charged nanoparticles. Based on these stabilization strategies, the all-aqueous droplets can be stabilized and various all-aqueous emulsion-based biomaterials are fabricated, such as microgel particles, microcapsules, polyelectrolyte microcapsules, hybrid polyelectrolyte microcapsules, AWE-somes, polymersomes, colloidosomes, and fibrillosomes.
ACKNOWLEDGMENTS
The financial support from the Research Grants Council of Hong Kong (Grant Nos. GRF 17205421, 17204420, 17210319, 17204718, and ECS 21213621) is gratefully acknowledged. This work was also supported in part by the Zhejiang Provincial, Hangzhou Municipal, and Lin'an County Governments.
Contributor Information
Pingan Zhu, Email: mailto:pingazhu@cityu.edu.hk.
Liqiu Wang, Email: mailto:lqwang@hku.hk.
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
DATA AVAILABILITY
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
References
- 1. Hatti-Kaul R., Mol. Biotechnol. 19, 269 (2001). 10.1385/MB:19:3:269 [DOI] [PubMed] [Google Scholar]
- 2. Walter H. and Johansson G., Methods in Enzymology: Aqueous Two-phase Systems ( Academic Press, San Diego, 1994), Vol. 228, p. 3. [Google Scholar]
- 3. SooHoo J. R. and Walker G. M., Biomed. Microdevices 11, 323 (2009). 10.1007/s10544-008-9238-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Münchow G., Hardt S., Kutter J. P., and Drese K. S., Lab Chip 7, 98 (2007). 10.1039/B612669N [DOI] [PubMed] [Google Scholar]
- 5. Porto T. S., Medeiros e Silva G. M., Porto C. S., Cavalcanti M. T. H., Neto B. B., Lima-Filho J. L., Converti A., Porto A. L. F., and Pessoa A., Chem. Eng. Process. 47, 716 (2008). 10.1016/j.cep.2006.12.004 [DOI] [Google Scholar]
- 6. Hahn T. and Hardt S., Anal. Chem. 83, 5476 (2011). 10.1021/ac201228v [DOI] [PubMed] [Google Scholar]
- 7. Hahn T., Münchow G., and Hardt S., J. Phys.: Condens. Matter 23, 184107 (2011). 10.1088/0953-8984/23/18/184107 [DOI] [PubMed] [Google Scholar]
- 8. Vijayakumar K., Gulati S., Demello A. J., and Edel J. B., Chem. Sci. 1, 447 (2010). 10.1039/c0sc00229a [DOI] [Google Scholar]
- 9. Dewey D. C., Strulson C. A., Cacace D. N., Bevilacqua P. C., and Keating C. D., Nat. Commun. 5(1), 4670 (2014). 10.1038/ncomms5670 [DOI] [PubMed] [Google Scholar]
- 10. Song Y., Shimanovich U., Michaels T. C. T., Ma Q., Li J., Knowles T. P. J., and Shum H. C., Nat. Commun. 7, 12934 (2016). 10.1038/ncomms12934 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Keating C. D., Acc. Chem. Res. 45, 2114 (2012). 10.1021/ar200294y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Zhang L., Cai L. H., Lienemann P. S., Rossow T., Polenz I., Vallmajo-Martin Q., Ehrbar M., Na H., Mooney D. J., and Weitz D. A., Angew. Chem. 128, 13668 (2016). 10.1002/ange.201606960 [DOI] [PubMed] [Google Scholar]
- 13. Zhou C., Zhu P., Tian Y., Xu M., and Wang L., ACS Nano 13, 6319 (2019). 10.1021/acsnano.9b00731 [DOI] [PubMed] [Google Scholar]
- 14. Zhu P., Chen R., Zhou C., Aizenberg M., Aizenberg J., and Wang L., Adv. Mater. 33, 2008558 (2021). 10.1002/adma.202008558 [DOI] [PubMed] [Google Scholar]
- 15. Zhu P. and Wang L., “ Microfluidics-enabled soft manufacture of materials with tailorable wettability,” Chem. Rev. (published online). 10.1021/acs.chemrev.1c00530 [DOI] [PubMed] [Google Scholar]
- 16. Tian Y., Zhu P., Tang X., Zhou C., Wang J., Kong T., Xu M., and Wang L., Nat. Commun. 8, 1080 (2017). 10.1038/s41467-017-01157-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Shi R., Tian Y., Zhu P., Tang X., Tian X., Zhou C., and Wang L., ACS Appl. Mater. Interfaces 12, 29747 (2020). 10.1021/acsami.0c04824 [DOI] [PubMed] [Google Scholar]
- 18. Zhu P., Chen R., Zhou C., Tian Y., and Wang L., Chem. Eng. J. 415, 128944 (2021). 10.1016/j.cej.2021.128944 [DOI] [Google Scholar]
- 19. Teixeira A. G., Agarwal R., Ko K. R., Grant-Burt J., Leung B. M., and Frampton J. P., Adv. Healthcare Mater. 7, 1701036 (2018). 10.1002/adhm.201701036 [DOI] [PubMed] [Google Scholar]
- 20. Soares R. R. G., Silva D. F. C., Fernandes P., Azevedo A. M., Chu V., Conde J. P., and Aires-Barros M. R., Biotechnol. J. 11, 1498 (2016). 10.1002/biot.201600356 [DOI] [PubMed] [Google Scholar]
- 21. Hardt S. and Hahn T., Lab Chip 12, 434 (2012). 10.1039/C1LC20569B [DOI] [PubMed] [Google Scholar]
- 22. Zhu K., Yu Y., Cheng Y., Tian C., Zhao G., and Zhao Y., ACS Appl. Mater. Interfaces 11, 4826 (2019). 10.1021/acsami.8b19234 [DOI] [PubMed] [Google Scholar]
- 23. Navi M., Abbasi N., Jeyhani M., Gnyawali V., and Tsai S. S. H., Lab Chip 18, 3361 (2018). 10.1039/C8LC00867A [DOI] [PubMed] [Google Scholar]
- 24. Abbasi N., Navi M., and Tsai S. S. H., Langmuir 34, 213 (2018). 10.1021/acs.langmuir.7b03245 [DOI] [PubMed] [Google Scholar]
- 25. Liu H. T., Wang H., Wei W. B., Liu H., Jiang L., and Qin J. H., Small 14, 1801095 (2018). 10.1002/smll.201801095 [DOI] [PubMed] [Google Scholar]
- 26. Hann S. D., Niepa T. H. R., Stebe K. J., and Lee D., ACS Appl. Mater. Interfaces 8, 25603 (2016). 10.1021/acsami.6b07939 [DOI] [PubMed] [Google Scholar]
- 27. Moon B. U., Hwang D. K., and Tsai S. S. H., Lab Chip 16, 2601 (2016). 10.1039/C6LC00576D [DOI] [PubMed] [Google Scholar]
- 28. Zou Y., Song J., You X., Yao J., Xie S., Jin M., Wang X., Yan Z., Zhou G., and Shui L., ACS Appl. Mater. Interfaces 11, 21227 (2019). 10.1021/acsami.9b02788 [DOI] [PubMed] [Google Scholar]
- 29. Beldengrün Y., Aragon J., Prazeres S. F., Montalvo G., Miras J., and Esquena J., Langmuir 34, 9731 (2018). 10.1021/acs.langmuir.8b01599 [DOI] [PubMed] [Google Scholar]
- 30. Znidarsic-Plazl P., Brecko Z., Vicente F. A., Serucnik M., Coutinho J. A., and Ventura S. P., New Biotechnol. 44, S19 (2018). 10.1016/j.nbt.2018.05.154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Lee S. Y., Khoiroh I., Ooi C. W., Ling T. C., and Show P. L., Sep. Purif. Rev. 46, 291 (2017). 10.1080/15422119.2017.1279628 [DOI] [Google Scholar]
- 32. De Barros D. P. C., Campos S. R. R., Azevedo A. M., Baptista A. M., and Aires-barros M. R., J. Chromatogr. A 1470, 50 (2016). 10.1016/j.chroma.2016.09.072 [DOI] [PubMed] [Google Scholar]
- 33. Asenjo J. A. and Andrews B. A., J. Chromatogr. A 1238, 1 (2012). 10.1016/j.chroma.2012.03.049 [DOI] [PubMed] [Google Scholar]
- 34. Asenjo J. A. and Andrews B. A., J. Chromatogr. A 1218, 8826 (2011). 10.1016/j.chroma.2011.06.051 [DOI] [PubMed] [Google Scholar]
- 35. Su C. K. and Chiang B. H., Process Biochem. 41, 257 (2006). 10.1016/j.procbio.2005.06.026 [DOI] [Google Scholar]
- 36. Ma Q., Song Y., Baier G., Holtze C., and Shum H. C., J. Mater. Chem. B 4, 1213 (2016). 10.1039/C5TB02187A [DOI] [PubMed] [Google Scholar]
- 37. Phong W. N., Show P. L., Chow Y. H., and Ling T. C., J. Biosci. Bioeng. 126, 273 (2018). 10.1016/j.jbiosc.2018.03.005 [DOI] [PubMed] [Google Scholar]
- 38. Hahn T. and Hardt S., Soft Matter 7, 6320 (2011). 10.1039/c1sm05309d [DOI] [Google Scholar]
- 39. Walter H., Partitioning in Aqueous Two-Phase Systems: Theory, Methods, Uses, and Applications to Biotechnology ( Elsevier, 2012). [Google Scholar]
- 40. Albertsson P.-Å., Nature 177, 771 (1956). 10.1038/177771a0 [DOI] [PubMed] [Google Scholar]
- 41. Albertsson P.-Å., Nature 182, 709 (1958). 10.1038/182709a0 [DOI] [PubMed] [Google Scholar]
- 42. Albertsson P.-Å., Partition of Cell Particles and Macromolecules: Separation and Purification of Biomolecules, Cell Organelles, Membranes, and Cells in Aqueous Polymer Two-Phase Systems and Their Use in Biochemical Analysis and Biotechnology, 3rd ed. ( Wiley, 1986). [Google Scholar]
- 43. Song Y., Sauret A., and Shum H. C., Biomicrofluidics 7, 061301 (2013). 10.1063/1.4827916 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Iqbal M., Tao Y., Xie S., Zhu Y., Chen D., Wang X., Huang L., Peng D., Sattar A., Shabbir M. A. B., Hussain H. I., Ahmed S., and Yuan Z., Biol. Proced. Online 18, 18 (2016). 10.1186/s12575-016-0048-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Benavides J., Rito-Palomares M., and Asenjo J. A., Comprehensive Biotechnology, 2nd ed. ( Elsevier B.V., 2011), Vol. 2, p. 697. [Google Scholar]
- 46. Yau Y. K., Ooi C. W., Ng E.-P., Lan J. C.-W., Ling T. C., and Show P. L., Bioresour. Bioprocess. 2, 49 (2015). 10.1186/s40643-015-0078-0 [DOI] [Google Scholar]
- 47. Silva D. F. C., Azevedo A. M., Fernandes P., Chu V., Conde J. P., and Aires-Barros M. R., J Chromatogr. A 1370, 115 (2014). 10.1016/j.chroma.2014.10.035 [DOI] [PubMed] [Google Scholar]
- 48. Ruthven M., Ko K. R., Agarwal R., and Frampton J. P., Analyst 142, 1938 (2017). 10.1039/C7AN00255F [DOI] [PubMed] [Google Scholar]
- 49. Esquena J., Curr. Opin. Colloid Interface Sci. 25, 109 (2016). 10.1016/j.cocis.2016.09.010 [DOI] [Google Scholar]
- 50. Dickinson E., Trends Food Sci. Technol. 83, 31 (2019). 10.1016/j.tifs.2018.11.004 [DOI] [Google Scholar]
- 51. Zhou C., Zhu P., Tian Y., Tang X., Shi R., and Wang L., Lab Chip 17, 3310 (2017). 10.1039/C7LC00696A [DOI] [PubMed] [Google Scholar]
- 52. Atefi E., Mann J. A., and Tavana H., Langmuir 30, 9691 (2014). 10.1021/la500930x [DOI] [PubMed] [Google Scholar]
- 53. Pécréaux J., Döbereiner H. G., Prost J., Joanny J. F., and Bassereau P., Eur. Phys. J. E 13, 277 (2004). 10.1140/epje/i2004-10001-9 [DOI] [PubMed] [Google Scholar]
- 54. Petelska A. D., Cent. Eur. J. Chem. 10, 16 (2012). 10.2478/s11532-011-0130-7 [DOI] [Google Scholar]
- 55. Vis M., Opdam J., Vant Oor I. S. J., Soligno G., Van Roij R., Tromp R. H., and Erné B. H., ACS Macro Lett. 4, 965 (2015). 10.1021/acsmacrolett.5b00480 [DOI] [PubMed] [Google Scholar]
- 56. Chao Y. and Shum H. C., Chem. Soc. Rev. 49, 114 (2020). 10.1039/C9CS00466A [DOI] [PubMed] [Google Scholar]
- 57. Lo H. Y., Liu Y., Mak S. Y., Xu Z., Chao Y., Li K. J., Shum H. C., and Xu L., Phys. Rev. Lett. 123, 134501 (2019). 10.1103/PhysRevLett.123.134501 [DOI] [PubMed] [Google Scholar]
- 58. Breisig H. and Wessling M., Biomicrofluidics 9, 044122 (2015). 10.1063/1.4929519 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Sauret A. and Shum H. C., Appl. Phys. Lett. 100, 154106 (2012). 10.1063/1.3702434 [DOI] [Google Scholar]
- 60. Song Y. and Shum H. C., Langmuir 28, 12054 (2012). 10.1021/la3026599 [DOI] [PubMed] [Google Scholar]
- 61. Shum H. C., Varnell J., and Weitz D. A., Biomicrofluidics 6, 012808 (2012). 10.1063/1.3670365 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62. De Lora J. A., Fencl F. A., MacIas Gonzalez A. D. Y., Bandegi A., Foudazi R., Lopez G. P., Shreve A. P., and Carroll N. J., ACS Appl. Bio Mater. 2, 4097 (2019). 10.1021/acsabm.9b00617 [DOI] [PubMed] [Google Scholar]
- 63. Mak S. Y., Chao Y., Rahman S., and Shum H. C., Langmuir 34, 926 (2018). 10.1021/acs.langmuir.7b02633 [DOI] [PubMed] [Google Scholar]
- 64. Geschiere S. D., Ziemecka I., Van Steijn V., Koper G. J. M., van Esch J. H., and Kreutzer M. T., Biomicrofluidics 6, 022007 (2012). 10.1063/1.3700117 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65. Ziemecka I., Van Steijn V., Koper G. J. M., Rosso M., Brizard A. M., Van Esch J. H., and Kreutzer M. T., Lab Chip 11, 620 (2011). 10.1039/C0LC00375A [DOI] [PubMed] [Google Scholar]
- 66. Ziemecka I., Van Steijn V., Koper G. J. M., Kreutzer M. T., and Van Esch J. H., Soft Matter 7, 9878 (2011). 10.1039/c1sm06517c [DOI] [Google Scholar]
- 67. Lai D., Frampton J. P., Sriram H., and Takayama S., Lab Chip 11, 3551 (2011). 10.1039/c1lc20560a [DOI] [PubMed] [Google Scholar]
- 68. Tarameshlou M., Jafari S. H., Rezaeian I., and Khonakdar H. A., Int. J. Polym. Mater. Polym. Biomater. 63, 884 (2014). 10.1080/00914037.2014.886225 [DOI] [Google Scholar]
- 69. Moon B. U., Jones S. G., Hwang D. K., and Tsai S. S. H., Lab Chip 15, 2437 (2015). 10.1039/C5LC00217F [DOI] [PubMed] [Google Scholar]
- 70. Liu Y., Nambu N. O., and Taya M., Biomed. Microdevices 19, 55 (2017). 10.1007/s10544-017-0198-8 [DOI] [PubMed] [Google Scholar]
- 71. Choi Y. H., Song Y. S., and Kim D. H., J. Chromatogr. A 1217, 3723 (2010). 10.1016/j.chroma.2010.04.015 [DOI] [PubMed] [Google Scholar]
- 72. Song Y. S., Choi Y. H., and Kim D. H., J. Chromatogr. A 1162, 180 (2007). 10.1016/j.chroma.2007.06.032 [DOI] [PubMed] [Google Scholar]
- 73. Choi Y. H., Song Y. S., and Kim D. H., AIP Conf. Proc. 1027, 1006 (2008). 10.1063/1.2964449 [DOI] [Google Scholar]
- 74. Azizian P., Azarmanesh M., Dejam M., Mohammadi M., Shamsi M., Sanati-Nezhad A., and Mohamad A. A., Chem. Eng. Sci. 195, 201 (2019). 10.1016/j.ces.2018.11.050 [DOI] [Google Scholar]
- 75. Song Y., Chan Y. K., Ma Q., Liu Z., and Shum H. C., ACS Appl. Mater. Interfaces 7, 13925 (2015). 10.1021/acsami.5b02708 [DOI] [PubMed] [Google Scholar]
- 76. Nan L., Cao Y., Yuan S., and Shum H. C., Microsyst. Nanoeng. 6, 70 (2020). 10.1038/s41378-020-0180-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 77. Breisig H., Hoppe J., Melin T., and Wessling M., J. Membr. Sci. 467, 109 (2014). 10.1016/j.memsci.2014.05.022 [DOI] [Google Scholar]
- 78. Akamatsu K., Kurita R., Sato D., and Nakao S. I., Langmuir 35, 9825 (2019). 10.1021/acs.langmuir.9b01320 [DOI] [PubMed] [Google Scholar]
- 79. Moon B. U., Abbasi N., Jones S. G., Hwang D. K., and Tsai S. S. H., Anal. Chem. 88, 3982 (2016). 10.1021/acs.analchem.6b00225 [DOI] [PubMed] [Google Scholar]
- 80. Mastiani M., Seo S., Jimenez S. M., Petrozzi N., and Kim M. M., Colloids Surf. A 531, 111 (2017). 10.1016/j.colsurfa.2017.07.083 [DOI] [Google Scholar]
- 81. Jeyhani M., Gnyawali V., Abbasi N., Hwang D. K., and Tsai S. S. H., J. Colloid Interface Sci. 553, 382 (2019). 10.1016/j.jcis.2019.05.100 [DOI] [PubMed] [Google Scholar]
- 82. Mastiani M., Seo S., Mosavati B., and Kim M., ACS Omega 3, 9296 (2018). 10.1021/acsomega.8b01768 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 83. Choi D., Lee E., Kim S. J., and Han M., Soft Matter 15, 4647 (2019). 10.1039/C9SM00469F [DOI] [PubMed] [Google Scholar]
- 84. Zhang Q., Chen J., and Gai H., J. Mater. Sci. 54, 14905 (2019). 10.1007/s10853-019-04001-w [DOI] [Google Scholar]
- 85. Zhou C., Zhu P., Han X., Shi R., Tian Y., and Wang L., Lab Chip 21, 2684 (2021). 10.1039/D1LC00351H [DOI] [PubMed] [Google Scholar]
- 86. Dang V. B. and Kim S. J., Lab Chip 17, 286 (2017). 10.1039/C6LC00911E [DOI] [PubMed] [Google Scholar]
- 87. Chao Y., Mak S. Y., Ma Q., Wu J., Ding Z., Xu L., and Shum H. C., Langmuir 34, 3030 (2018). 10.1021/acs.langmuir.7b04168 [DOI] [PubMed] [Google Scholar]
- 88. Guzowski J., Korczyk P. M., Jakiela S., and Garstecki P., Soft Matter 8, 7269 (2012). 10.1039/c2sm25838b [DOI] [Google Scholar]
- 89. Lee M. H., Hribar K. C., Brugarolas T., Kamat N. P., Burdick J. A., and Lee D., Adv. Funct. Mater. 22, 131 (2012). 10.1002/adfm.201101303 [DOI] [Google Scholar]
- 90. Jeyhani M., Thevakumaran R., Abbasi N., Hwang D. K., and Tsai S. S. H., Small 16, 1906565 (2020). 10.1002/smll.201906565 [DOI] [PubMed] [Google Scholar]
- 91. Yuan H., Ma Q., Song Y., Tang M. Y. H., Chan Y. K., and Shum H. C., Macromol. Chem. Phys. 218, 1600422 (2017). 10.1002/macp.201600422 [DOI] [Google Scholar]
- 92. Cui C., Zeng C., Wang C., and Zhang L., Langmuir 33, 12670 (2017). 10.1021/acs.langmuir.7b02888 [DOI] [PubMed] [Google Scholar]
- 93. Chao Y., Mak S. Y., Rahman S., Zhu S., and Shum H. C., Small 14, 1802107 (2018). 10.1002/smll.201802107 [DOI] [PubMed] [Google Scholar]
- 94. Song Q., Chao Y., Zhang Y., and Shum H. C., J. Phys. Chem. B 125, 562 (2021). 10.1021/acs.jpcb.0c09884 [DOI] [PubMed] [Google Scholar]
- 95. Nicolai T. and Murray B., Food Hydrocolloids 68, 157 (2017). 10.1016/j.foodhyd.2016.08.036 [DOI] [Google Scholar]
- 96. Watanabe T., Motohiro I., and Ono T., Langmuir 35, 2358 (2019). 10.1021/acs.langmuir.8b04169 [DOI] [PubMed] [Google Scholar]
- 97. Li B. Z., Wang L. J., Li D., Adhikari B., and Mao Z. H., Carbohydr. Polym. 88, 912 (2012). 10.1016/j.carbpol.2012.01.043 [DOI] [Google Scholar]
- 98. Ma S., Thiele J., Liu X., Bai Y., Abell C., and Huck W. T. S., Small 8, 2356 (2012). 10.1002/smll.201102715 [DOI] [PubMed] [Google Scholar]
- 99. Liu Q., Zhao M., Mytnyk S., Klemm B., Zhang K., Wang Y., Yan D., Mendes E., and van Esch J. H., Angew. Chem., Int. Ed. 58, 547 (2019). 10.1002/anie.201811374 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 100. Mytnyk S., Ziemecka I., Olive A. G. L., Van Der Meer J. W. M., Totlani K. A., Oldenhof S., Kreutzer M. T., Van Steijn V., and Van Esch J. H., RSC Adv. 7, 11331 (2017). 10.1039/C7RA00452D [DOI] [Google Scholar]
- 101. Tan W. H. and Takeuchi S., Adv. Mater. 19, 2696 (2007). 10.1002/adma.200700433 [DOI] [Google Scholar]
- 102. Miesch C., Kosif I., Lee E., Kim J. K., Russell T. P., Hayward R. C., and Emrick T., Angew. Chem., Int. Ed. 51, 145 (2012). 10.1002/anie.201106665 [DOI] [PubMed] [Google Scholar]
- 103. Xue L. H., Xie C. Y., Meng S. X., Bai R. X., Yang X., Wang Y., Wang S., Binks B. P., Guo T., and Meng T., ACS Macro Lett. 6, 679 (2017). 10.1021/acsmacrolett.7b00294 [DOI] [PubMed] [Google Scholar]
- 104. Gonzalez-Jordan A., Nicolai T., and Benyahia L., Langmuir 32, 7189 (2016). 10.1021/acs.langmuir.6b01993 [DOI] [PubMed] [Google Scholar]
- 105. Binks B. P. and Shi H., Langmuir 35, 4046 (2019). 10.1021/acs.langmuir.8b04151 [DOI] [PubMed] [Google Scholar]
- 106. Murray B. S. and Phisarnchananan N., Food Hydrocolloids 42, 92 (2014). 10.1016/j.foodhyd.2014.01.004 [DOI] [Google Scholar]
- 107. Firoozmand H., Murray B. S., and Dickinson E., Langmuir 25, 1300 (2009). 10.1021/la8037389 [DOI] [PubMed] [Google Scholar]
- 108. Douliez J. P., Martin N., Beneyton T., Eloi J. C., Chapel J. P., Navailles L., Baret J. C., Mann S., and Béven L., Angew. Chem., Int. Ed. 57, 7780 (2018). 10.1002/anie.201802929 [DOI] [PubMed] [Google Scholar]
- 109. Balakrishnan G., Nicolai T., Benyahia L., and Durand D., Langmuir 28, 5921 (2012). 10.1021/la204825f [DOI] [PubMed] [Google Scholar]
- 110. Nguyen B. T., Nicolai T., and Benyahia L., Langmuir 29, 10658 (2013). 10.1021/la402131e [DOI] [PubMed] [Google Scholar]
- 111. De Freitas R. A., Nicolai T., Chassenieux C., and Benyahia L., Langmuir 32, 1227 (2016). 10.1021/acs.langmuir.5b03761 [DOI] [PubMed] [Google Scholar]
- 112. Gonzalez-Jordan A., Benyahia L., and Nicolai T., Colloids Surf. A 532, 332 (2017). 10.1016/j.colsurfa.2017.04.073 [DOI] [Google Scholar]
- 113. Das A. A. K., Filby B. W., Geddes D. A., Legrande D., and Paunov V. N., Mater. Horiz. 4, 1196 (2017). 10.1039/C7MH00677B [DOI] [Google Scholar]
- 114. Ganley W. J., Ryan P. T., and van Duijneveldt J. S., J. Colloid Interface Sci. 505, 139 (2017). 10.1016/j.jcis.2017.05.062 [DOI] [PubMed] [Google Scholar]
- 115. Inam M., Jones J. R., Pérez-Madrigal M. M., Arno M. C., Dove A. P., and O'Reilly R. K., ACS Cent. Sci. 4, 63 (2018). 10.1021/acscentsci.7b00436 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 116. Murray B. S. and Phisarnchananan N., Food Hydrocolloids 56, 161 (2016). 10.1016/j.foodhyd.2015.11.032 [DOI] [Google Scholar]
- 117. Nguyen B. T., Wang W., Saunders B. R., Benyahia L., and Nicolai T., Langmuir 31, 3605 (2015). 10.1021/la5049024 [DOI] [PubMed] [Google Scholar]
- 118. Khemissi H., Bassani H., Aschi A., Capron I., Benyahia L., and Nicolai T., Langmuir 34, 11806 (2018). 10.1021/acs.langmuir.8b02383 [DOI] [PubMed] [Google Scholar]
- 119. Hanazawa T. and Murray B. S., Langmuir 29, 9841 (2013). 10.1021/la401595u [DOI] [PubMed] [Google Scholar]
- 120. Hanazawa T. and Murray B. S., Food Hydrocolloids 34, 128 (2014). 10.1016/j.foodhyd.2012.11.025 [DOI] [Google Scholar]
- 121. Ben Ayed E., Cochereau R., Dechancé C., Capron I., Nicolai T., and Benyahia L., Langmuir 34, 6887 (2018). 10.1021/acs.langmuir.8b01239 [DOI] [PubMed] [Google Scholar]
- 122. Peddireddy K. R., Nicolai T., Benyahia L., and Capron I., ACS Macro Lett. 5, 283 (2016). 10.1021/acsmacrolett.5b00953 [DOI] [PubMed] [Google Scholar]
- 123. Cacace D. N., Rowland A. T., Stapleton J. J., Dewey D. C., and Keating C. D., Langmuir 31, 11329 (2015). 10.1021/acs.langmuir.5b02754 [DOI] [PubMed] [Google Scholar]
- 124. Andes-Koback M. and Keating C. D., J. Am. Chem. Soc. 133, 9545 (2011). 10.1021/ja202406v [DOI] [PMC free article] [PubMed] [Google Scholar]
- 125. Zhang Y., Wu F., Yuan W., and Jin T., J. Controlled Release 147, 413 (2010). 10.1016/j.jconrel.2010.07.121 [DOI] [PubMed] [Google Scholar]
- 126. Buzza D. M. A., Fletcher P. D. I., Georgiou T. K., and Ghasdian N., Langmuir 29, 14804 (2013). 10.1021/la403356j [DOI] [PubMed] [Google Scholar]
- 127. Firoozmand H. and Rousseau D., Food Hydrocolloids 42, 204 (2014). 10.1016/j.foodhyd.2014.04.040 [DOI] [Google Scholar]
- 128. Hann S. D., Goulian M., Lee D., and Stebe K. J., Soft Matter 11, 1733 (2015). 10.1039/C4SM02267J [DOI] [PMC free article] [PubMed] [Google Scholar]
- 129. Zhang J., Hwang J., Antonietti M., and Schmidt B. V. K. J., Biomacromolecules 20, 204 (2019). 10.1021/acs.biomac.8b01301 [DOI] [PubMed] [Google Scholar]
- 130. Keal L., Colosqui C. E., Tromp R. H., and Monteux C., Phys. Rev. Lett. 120, 208003 (2018). 10.1103/PhysRevLett.120.208003 [DOI] [PubMed] [Google Scholar]
- 131. Simon K. A., Sejwal P., Gerecht R. B., and Luk Y. Y., Langmuir 23, 1453 (2007). 10.1021/la062203s [DOI] [PubMed] [Google Scholar]
- 132. Chatsisvili N., Philipse A. P., Loppinet B., and Tromp R. H., Food Hydrocolloids 65, 17 (2017). 10.1016/j.foodhyd.2016.10.036 [DOI] [Google Scholar]
- 133. Poortinga A. T., Langmuir 24, 1644 (2008). 10.1021/la703441e [DOI] [PubMed] [Google Scholar]
- 134. Atefi E., Fyffe D., Kaylan K. B., and Tavana H., J. Chem. Eng. Data 61, 1531 (2016). 10.1021/acs.jced.5b00901 [DOI] [Google Scholar]
- 135. Hann S. D., Stebe K. J., and Lee D., Langmuir 33, 10107 (2017). 10.1021/acs.langmuir.7b02237 [DOI] [PubMed] [Google Scholar]
- 136. Zhang L., Cai L. H., Lienemann P. S., Rossow T., Polenz I., Vallmajo-Martin Q., Ehrbar M., Na H., Mooney D. J., and Weitz D. A., Angew. Chem., Int. Ed. 55, 13470 (2016). 10.1002/anie.201606960 [DOI] [PubMed] [Google Scholar]
- 137. Hann S. D., Lee D., and Stebe K. J., Phys. Chem. Chem. Phys. 19, 23825 (2017). 10.1039/C7CP02809A [DOI] [PubMed] [Google Scholar]
- 138. Hann S. D., Stebe K. J., and Lee D., ACS Appl. Mater. Interfaces 9, 25023 (2017). 10.1021/acsami.7b05800 [DOI] [PubMed] [Google Scholar]
- 139. Ma Q., Yuan H., Song Y., Chao Y., Mak S. Y., and Shum H. C., Soft Matter 14, 1552 (2018). 10.1039/C7SM02275A [DOI] [PubMed] [Google Scholar]
- 140. Ma Q., Song Y., Kim J. W., Choi H. S., and Shum H. C., ACS Macro Lett. 5, 666 (2016). 10.1021/acsmacrolett.6b00228 [DOI] [PubMed] [Google Scholar]
- 141. Deng Y., Ma Q., Yuan H., Lum G. C., and Shum H. C., J. Mater. Chem. B 7, 3059 (2019). 10.1039/C8TB03074J [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data sharing is not applicable to this article as no new data were created or analyzed in this study.





















