Hierarchical Biomolecular Emulsions Using 3-D Microfluidics with

Subscriber access provided by Gothenburg University Library

Article

Hierarchical biomolecular emulsions using 3D microfluidics with uniform surface chemistry Zenon Toprakcioglu, Aviad Levin, and Tuomas P.J. Knowles Biomacromolecules, Just Accepted Manuscript • DOI: 10.1021/acs.biomac.7b01159 • Publication Date (Web): 29 Sep 2017 Downloaded from http://pubs.acs.org on September 29, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biomacromolecules is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

Hierarchical biomolecular emulsions using 3-D microfluidics with uniform surface chemistry Zenon Toprakcioglu,† Aviad Levin,† and Tuomas P. J. Knowles∗,†,‡ †Department of Chemistry, University of Cambridge, Lensfield Road, UK ‡Cavendish Laboratory, Department of Physics, J J Thomson Avenue, UK E-mail: [email protected]

Abstract Microfluidic devices can be used to produce single, double and higher order emulsions, where droplet sizes can be precisely controlled and modulated. Such emulsions have great potential for the storage and study of biomolecules, including peptides and proteins. However, advancement of this technique has remained challenging due to the tendency of various biomolecules to adhere to the surface of the formed channels, resulting in changes in surface wetting and fouling on the micrometer scale. Thus, precise control of surface wettability plays a crucial role in the processes that govern droplet formation. Here, we report an approach for producing both water-oil-water (w/o/w) and oil-wateroil (o/w/o) double emulsions without any need for surface modification, an enabling feature for biomolecular encapsulation. Using this strategy, we show that the number of monodisperse encapsulated internal droplets can be controlled systematically and 1

ACS Paragon Plus Environment

Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

reproducibly by suitable adjustment of the relevant flow rates, and ranges from 1 to 40 in the case of w/o/w emulsions. We further demonstrate that the number of internal droplets scales linearly with the reciprocal flow rate of the outer continuous phase, when the inner and middle phase flow rates are kept constant. We demonstrate that this approach is suitable for forming double emulsions where the inner phase consists of reconstituted silk protein solution whereby incubation of the internal droplets can be induced to form a gel resulting in silk fibroin micro-gels surrounded by an external oil shell. Finally, for o/w/o emulsions, we show that single or multiple monodisperse internal droplets can be encapsulated with a size that ranges over one order of magnitude, from ca. 10 µm to > 100 µm. Moreover, o/w/o emulsions where the middle phase consists of silk fibroin solution were prepared and by allowing the protein to aggregate, a coreshell structure was formed. This microfluidic strategy allows for multiple emulsions to be generated drop by drop for biomolecular solutions with potential applications in the biomedical and pharmaceutical fields.

2


Page 2 of 31

Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

Introduction Microfluidic devices offer a highly suitable platform for producing water-in-oil (w/o) or oil-in-water (o/w) emulsions. Control of the dispersed phase droplet sizes is well established and can readily be achieved by varying the channel widths and the ratio of flow between the dispersed and continuous phases. 1 Controlling the surface wettability, however, is crucial for droplet formation. 2 In order to produce droplets, the continuous phase must wet the channel surface, thus preventing the dispersed phase from coming into contact with the surface, i.e. to form a w/o emulsion, the surface of the device must be hydrophobic and vice versa when forming an o/w emulsion. For example, polydimethylsiloxane (PDMS), commonly used in microfluidic device fabrication, is intrinsically hydrophobic so that o/w emulsions are challenging to form without surface modification and even w/o droplets cannot easily be produced. Surface modification techniques such as oxygen plasma, or silane treatments 3,4 can render the microfluidic device hydrophilic or hydrophobic, depending on their intended use. Surface chemistry plays an even greater role when generating double emulsions. This is because the first and second junction of a microfluidic device need to display different surface properties. 5 For instance to produce w/o/w droplets, the first junction must be hydrophobic and the second one hydrophilic, (see figure 1c, d). Yet, such changes in surface chemistry on the micrometer scale can be challenging to achieve. 5,6 A variety of methods for producing double emulsions exist, such as making two devices; one that is rendered hydrophilic while the other is hydrophobic. The w/o emulsion is formed in the former and then transferred directly to the latter device, where the process of producing w/o/w double emulsions is finalised. 7,8 While such methods have been shown to allow the successful formation of double emulsion droplets, not only do they require surface treatment of the devices, 9,10 which is often temporary, 11 but may also present additional 3


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

difficulties, as both devices must be optimised to allow the sequential production of the droplets, adding another level of complexity to the process and to the potential towards scaling-up. Thus, there has been an increased interest in producing non-planar (3-D) devices, 2,12,13 where the geometry of the microfluidic device allows for the continuous phase, or the outer phase in the case of double emulsions, to completely engulf and surround the dispersed phase, illustrated in figure 2a. By applying such an approach, the need for surface modifications is eliminated, and instead, by utilising the properties of the inner, middle and outer phase solutions one can make w/o/w or o/w/o double emulsions using a single microfluidic device design. Double emulsions can potentially be beneficial to a variety of applications, such as in the biomedical and pharmaceutical fields, cosmetics or even in food products. 14–16 Therefore, allowing the scale-up of such platforms using reliable microfluidic devices that have high efficiencies is of increasing appeal. The ability to encapsulate drugs, 17–19 enzymes, 20,21 hormones 22 or even living cells 23–25 in the core of the double emulsion (i.e. the inner phase) and then control their release kinetics by altering the middle phase shell thickness is promising. 26–29 Additionally, double emulsions may have a significant role to play in the removal of toxic waste from non-potable water, as active matter can migrate from the outer to the inner phase of the emulsion. 30 Furthermore, biomimetic materials such as liposomes can be produced in this manner. 9 It has been reported that by using oleic acid, asolectin and cholesterol as the middle phase, w/o/w emulsions can be transformed into liposomes by solvent extraction. 9 Recently, however, microfluidics has been employed in the production of droplets containing protein monomers, and by utilising the self-assembly properties of these proteins, micro-gels can be formed. 31 Not only are these structures non-toxic and biodegradable, but they are easily synthesised and offer a facile route towards encapsulation and release

4


Page 4 of 31

Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

of small molecules with potential use in targeted drug delivery or even in gene therapy. 32 Studying conformational changes in proteins via the aggregation of soluble monomers to the formation of β-sheet rich supramolecular polymers, has been of major interest in the last few decades. Protein nanostructures such as amyloid fibrils, which can be formed as a result of aggregation, 31 play a vital role in different organisms and are biologically relevant as they have been associated with various neurodegenerative diseases such as Alzheimers or Parkinsons, thus understanding their properties is essential. 31,33 A protein that has remarkable properties and is of particular interest both for fundamental studies of protein self-assembly and for industrial applications, is silk. The silk produced in the glands of spiders has remarkable mechanical and biological properties and due to its biocompatibility can be used in different applications, such as drug delivery. 34 However, obtaining large quantities of silk protein from spiders is challenging due to the harvesting process, which involves extracting the silk from the glands of spider while it is still alive. An alternative is to use the commercially available cocoon of the Bombyx mori silkworm. Bombyx mori silk is comprised of native silk fibroin coated with sericin. The silk fibroin, which is the protein of interest, consists of heavy (Mw ∼390kDa) and light chains (Mw ∼26kDa) that are linked by disulfide bonds. 35 Silk fibroin is a hydrophobic block copolymer that self-assembles to forms β-sheet structures resulting in an extremely resilient and tough material, which due to its bioavailability is excellent for biomedical applications. 36 The purification process involving the removal of sericin results in a new type of fibroin, reconstituted silk fibroin (RSF), which displays different properties to native silk fibroin. Fibroin has been used to form gels using different conditions such as low pH 37–39 and has shown great potential towards biomedical applications. More recently, RSF has been used to produce micro-gels as it self-assembles to form nanofibrillar structures resulting in

5


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a protein capsule. Capillary microfluidics has been used to form RSF micro and nano-gels where the dispersed phase comprised of RSF and the continuous phase was poly(vinyl alcohol) (PVA). 40 In fact, by changing the concentration of fibroin and adjusting the ratio of the flow rate of the dispersed to the continuous phase nano-gels as small as 200 nm could be formed. Additionally, the release kinetics of small molecules encapsulated within the formed micro/nano-gels was monitored, and it was found that 40% of the encapsulated molecules are released within the first 24 hours. 40 However, RSF micro-gels have so far not been made using conventional microfluidics, which exhibit high reproducibility. More importantly, release studies involving silk fibroin solution as one of the phases in double emulsion formation are still at an early stage. 41 So, a controllable way of making double emulsions is essential for the use of these biocompatible materials for biomedical applications. 42,43 Producing double emulsions using PDMS based microfluidics containing protein, without any surface modifications, is particularly challenging due to the additional difficulty of the protein adhering to the surface of the device. In this work, we fabricate a 3-D microfluidic device through which w/o/w double emulsion production is optimised, allowing the formation of monodisperse internal droplets, whose number within the external droplet can be controlled by changing the flow rate of the outer phase while keeping the middle and inner phase flow rates constant. Additionally, it is possible to change the shell thickness by regulating the flow rate ratio of the three-phase system in a controlled manner. We further demonstrate that the size of the internal droplets can be varied by changing the ratio of the inner to middle phase flow rates, while maintaining a constant outer phase flow rate. Likewise, we show that our device is capable of forming w/o/w double emulsions where the inner phase contains reconstituted silk fibroin, and if left to incubate, the protein aggregates resulting in the formation of micro-gels. Furthermore, we demonstrate that using the same versatile

6


Page 6 of 31

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

device architecture, we can not only form w/o/w double emulsions, but also produce o/w/o droplets with no need for surface treatment or modification. Finally by using silk fibroin solution as the middle phase, we are able to generate core-shell structures that have encapsulated oil droplets.

Materials and Methods Device Fabrication

Two-level photolithography was used in the production of masters. The two-layered master was fabricated by firstly spin coating a 25 µm thick negative photo-resist (SU-8 3025, MicroChem) onto a silicon wafer, and then soft baking for 15 minutes at 95 ◦ C. The mask in figure 1a was placed on to the wafer, exposed under UV light and post baked for 5 minutes at 95 ◦ C. 44 A second 25 µm thick layer was spin coated onto the wafer and the second mask was correctly aligned (figure 1b) with the pattern formed from the first mask and UV exposed. Finally, the master was developed in Propylene glycol methyl ether acetate (PGMEA) (Sigma Aldrich) to remove any excess photo-resist. To produce the second master, the design shown in figure 1b was used and the same experimental procedure was employed using a 25 µm thick layer of photo-resist. Microfluidic devices were fabricated using a 10:1 ratio of pre-polymer PDMS to curing agent (Sylgard 184, DowCorning, Midland, MI) and cured for 3 hours at 65 ◦ C. The PDMS was cut and peeled off the masters and holes of 0.75 mm were punched on the PDMS molded after master 2. Following treatment with a plasma bonder (Diener Electronic, Ebhausen, Germany), the PDMS molded after master 1, was bonded on a glass slide with the channels facing upwards. The other PDMS mold was then placed on top and carefully aligned. A drop of water was applied in between the two PDMS layers 7


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a)

e)

b)

d)

c)

Inner Phase Middle Phase Outer Phase

Page 8 of 31

f)

Figure 1: (a-d:) Design of microfluidic devices used. (a-b:) Schemes of photomask designs used in the fabrication of the 3-D device. (a) Mask 1. Middle and Inner phase inlets and their respective channels. The middle phase channel width is 100 µm. The inner phase is 50 µm. The first junction is completely planar, or 2-D. (b) Mask 2. Outer phase inlet with its respective channels and outlet. The outlet channel is 200 µm, while the rest of the channels are 100 µm. Mask 1 is aligned with mask 2 to produce a two-layer master. (c) Schematic design of a microfluidic device for the productions of double emulsions. The device consists of 3 inlets, one for each of the three phases and an outlet, where the double emulsions are collected. (d) Magnified view of the two junctions where the double emulsions are formed. At the first junction, the inner and middle phase intersect, resulting in droplet formation. At the second junction, the outer phase engulfs the other two phases-leading to the production of double emulsions. (e) Schematic representation of the 3-D microfluidic channels. In the first junction droplets of inner phase surrounded by the middle phase are formed, while in the second junction, the outer phase encapsulates the other two phases forming a double emulsion. In the schematic shown above, there are three internal droplets within the external droplet. (f ) Image showing a perfectly aligned non-planar microfluidic device where the second junction is 3-D. The arrows represent the flows of the three different phases.

8


Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

to act as a lubricant so that they could be aligned with respect to each other. The device was baked at 65 ◦ C for 24 hours to ensure optimal bonding.

Double emulsion droplet formation

neMESYS syringe pumps (Cetoni, Korbussen, Germany) were used to control the flow rates within the channels. For the w/o/w double emulsions, deionised water was used as the inner phase, fluorinated oil (Fluorinert FC-40-Sigma Aldrich) containing 2% w/w fluorosurfactant (RAN biotechnologies) was used as the middle phase, while an aqueous solution of 3% w/w Sodium dodecylsulfate (SDS) (Sigma Aldrich, 3% w/w Tween 20 (Fisher Scientific) and 20% w/w Polyethylene glycol (PEG) (Fluka BioChemika) was used as the outer phase. For w/o/w double emulsions containing protein, a solution of reconstituted silk fibroin (Mindsets (UK) Limited) with 3% w/w Tween 20 and 100 µM Thioflavin T was used as the inner phase while the middle and outer phases remained the same. For the o/w/o double emulsions, FC-40 was used as the inner phase, an aqueous solution of 3% w/w SDS, 3% w/w Tween 20 and 20% w/w PEG was used as the middle phase, and FC-40 with 2% w/w fluorosurfactant was used as the outer phase. For o/w/o double emulsions containing protein, a solution of reconstituted silk fibroin (Mindsets (UK) Limited) with 3% w/w Tween 20 and 100 µM Thioflavin T was used as the middle phase while the inner and outer phases remained the same. Additionally, we employed the use of a fast camera (Mikrotron High Speed Cameras), which can reach speeds of 5000 fps, to track droplet formation.

Contact angle measurements

9


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The static contact angles for the different solutions used were measured by taking images and then anaylsing them using ImageJ. A 50 µL drop was placed on the PDMS surface and the angle between the surface and the tangent to the droplet was determined. It was found that deionised water has a contact angle of 105◦ ± 2◦ with PDMS while the aqueous solution containing both PEG and surfactants (Tween 20 and SDS) has a contact angle of 60◦ ± 2◦ with PDMS. Moreover, the contact angle of fluorinated oil with PDMS is 25◦ ± 2◦ . All static contact angle images are in figure S1.

Viscosity measurements

Viscosity measurements were conducted on a Physica MCR 501 Anton Paar Rheometer. Cone and plate geometry was used to ensure the shear stress remained constant throughout the sample. All measurements were taken at 25 ◦ C under oscillatory shear stress and are shown in figure S2. The viscosity of deionised water was found to be 0.8 ± 0.2 mPas, while the aqueous solution containing both PEG and surfactants has a viscosity of 34 ± 1 mPas. Finally, the viscosity of the fluorinated oil was determined to be 4 ± 0.2 mPas.

Silk fibroin preparation and purification

Bombyx mori silk cocoons (Mindsets (UK) Limited) were used to extract the silk fibroin protein by a well-established protocol. 36 Cocoons were cut into tiny pieces and boiled for 30 minutes with 0.02 M sodium carbonate; dissolving the sericin in the process. Undissolved fibroin was rinsed with deionised water three times and left to dry for 24 hours. Dried silk fibroin was dissolved in 9.3 M lithium bromide at 60 ◦ C for 4 hours. The

10


Page 10 of 31

Page 11 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

silk-LiBr solution was dialysed for 48 hours to remove any salts. Finally, in order to remove any solid impurities, the silk fibroin solution was centrifuged at 9000 r.p.m. at 4 ◦

C for 20 minutes.

Thioflavin T (ThT) assay

The structural change from silk fibroin monomers to fibrils was monitored and studied by looking at the change in intensity of ThT dye at 480 nm. ThT concentration of 100 µM was used for all experiments. The fluorescent signal was monitored microscopically using filters that allowed an excitation wavelength of 440 nm and an emission wavelength of 480 nm.

Results and Discussion Both double emulsions of w/o/w and o/w/o were formed using the same 3-D microfluidic device presented here. The ratio between the flow rates of the inner (Qin ), middle (Qmid ) and outer (Qout ) phases was optimised in order to control droplet size and/or number of encapsulated droplets. A top view image of the microfluidic device used is shown in figure 1f. The arrows represent the flow of the three phases. The first part of the microfluidic device, up to the second junction, is planar (2-D), while the second junction is 3-D. A schematic representation of our device is shown in figure 1e. At the first junction, inner phase droplets are formed in a medium of middle phase (effectively forming the first emulsion); while at the second junction, the outer phase engulfs the middle and inner phases, resulting in the double emulsion.

11


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Water/Oil/Water double emulsions The time evolution of droplet formation was monitored using a high speed-camera (figure 2). For Qout =500 µL/hr, Qmid =200 µL/hr and Qin =50 µL/hr, the external droplets were formed after 250 ms. The capillary number, Ca=µu/γ (µ is the viscosity of the outer phase, u is the mean speed of the outer phase fluid and γ is the interfacial tension between the middle and outer phases), corresponding to our measurements was found to be 90◦ during formation and detachment of the droplets (figure 2-left panel). As has been reported, 2 droplet formation can only occur when the contact angle is >90◦ . The argument presented refers to both planar and non-planar devices, but not to double emulsions. Our results demonstrate that this condition clearly still holds for double emulsions formed in 3-D microfluidic devices.

Figure 2: Time evolution of droplet formation. Upon formation, the contact angle is ca. 100◦ and remains essentially constant throughout the growth and detachment of the droplet. The droplets break off at around t = 250 ms leading to the double emulsion. Scale bar for all images is 100 µm.

The relationship between Qout and the number of internal droplets encapsulated was further investigated. This was determined by keeping Qin and Qmid constant, at 50 µL/hr and 200 µL/hr respectively. The Qout values ranged from 200 µL/hr up to 5000 µL/hr.

12


Page 12 of 31

Page 13 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

Monodisperse encapsulated internal droplets were formed, and these were counted by observing the images taken and analysing them frame by frame. For each case, the number of internal droplets (N) was taken to be the average of 10 different images. Figure 3a shows double emulsions with 8 internal droplets, while in figure 3b, N is 4. As expected, decreasing Qout leads to increasing N. This can be seen in figure 3d and in figure S2d. Our data further suggest that very large numbers of internal droplets can be encapsulated, since N increases asymptotically when Qout decreases. As the outer phase flow rate is decreased, the external droplet size increases, and so more droplets can be encapsulated. In practice, however, Qout cannot, be made arbitrarily small if w/o/w double emulsions are to be produced. If one were to continue decreasing Qout , a phase separation between the middle and outer phases would occur (figure 4d), which explains why N increases asymptotically. Effectively, one would obtain inner phase water droplets surrounded by an oil phase medium, forming water droplets dispersed in an oil droplet of effectively infinite size. Therefore, for a given device geometry, there is a practical upper limit to N. If we were to use smaller inner phase channels and larger outer phase channels, then N could theoretically be increased. A double logarithmic plot of N against Qout , shown in figure 3e, indicates a power law with an exponent close to -1. We were able to encapsulate up to 40 internal droplets by increasing Qmid with respect to Qin and decreasing Qout to as low as 200 µL/hr, (figure 3c). Droplets were collected, placed on a glass slide and imaged, as shown in figures 3f-h. In figures 3f and 3g all of the collected droplets either contained 2 or 3 internal droplets respectively, while for figure 3h, all but one of the double emulsion droplets does not contain 4 internal droplets. Encapsulation of multiple internal droplets, but without effective control over N, has been reported. 45 Here, we have shown that within the geometric limitations of our microfluidic device, we can predictably and reproducibly manipulate N by adjusting the flow rates and the

13


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a)

b)

Page 14 of 31

C)

e)

d)

1 1

f)

g)

h)

Figure 3: (a) w/o/w double emulsion droplets when Qout =700 µL/hr, Qmid =200 µL/hr and Qin =50 µL/hr with N = 7. (b) w/o/w double emulsion droplets when Qout =1500 µL/hr, Qmid =200 µL/hr and Qin =40 µL/hr with N = 4. (c) w/o/w double emulsion droplets when Qout =200 µL/hr, Qmid =300 µL/hr and Qin =40 µL/hr with 40 encapsulated internal droplets. The scale bar for all three images is 200 µm. (d) Dependence of the number of internal droplets (N) on outer phase flow rate (Qout ). N increases asymptotically as Qout decreases. (e) Log-log plot of N as a function of Qout . The gradient is approximately -1. (f ) Collected droplets consisting of 2 internal droplets, (g) 3 internal droplets and (h) 4 internal droplets. The scale bar for these images is 100 µm.

14


Page 15 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

geometry of the devices. The variation of internal droplet size as a function of flow rate was further studied. Internal droplets of 50 µm are shown in figure 4a while droplets as small as 35 µm are shown in figure 4b. The internal droplet size was changed by varying the oil phase, Qmid , from 200-1200 µL/hr, while keeping Qin and Qout constant at 20 µL/hr and 1000 µL/hr respectively. As Qmid is increased, and the ratio Qmid /Qin increases, the droplet size decreases as can be seen in figure 4c. The relationship between flow rate and droplet size is asymptotic, which suggests that as the middle phase is decreased with respect to the inner phase, droplets will not form and phase separation will occur. Moreover, as previously mentioned, if Qmid exceeds a certain value (in this case 1200 µL/hr when Qout is 1000 µL/hr), droplet formation cannot occur at the second junction as seen in figure 4d. Furthermore, by decreasing the ratio of Qmid /Qin , extremely large internal water droplets could be formed. This could be achieved using two different techniques; either by making water-in-oil droplets at the first junction followed by double emulsion formation in the second junction, or by carefully tuning Qin and Qmid , so that one can co-flow the inner and middle phases in the first junction, allowing the outer phase to encapsulate both phases at the second junction. In the former situation, Qin and Qmid were 100 µL/hr and 150 µL/hr respectively while Qout was kept at 1000 µL/hr. Figure 4e shows that large internal droplets, of around 85 µm can be produced using this technique. In the latter case, internal droplets of up to 150 µm can be produced, as seen in figure 4f, where Qin is 400 µL/hr, Qmid is 200 µL/hr and Qout is again 1000 µL/hr. Moreover, the thickness of the oil shell can be controlled directly during droplet formation. This may be of particular importance when the internal droplet is loaded with cargo molecules, as the release kinetics will vary considerably based on the thickness

15


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of the oil phase engulfed within the emulsion. In the co-flowing regime, shell thickness was found to be 11 ± 0.5 µm (figure 4f), while for the droplet regime (figure 4e), where Qmid /Qin is low, shell thickness was 22.2 ± 0.5 µm, with largest shell thickness determined at 38.6 ± 0.5 µm (figure 4b). Shell thickness could be further decreased down to 10.1 ± 0.5 µm by adjusting the flow rates. This leads us to believe that we could potentially obtain shell thicknesses of only a few micrometers just by manipulating the ratio of the three flow rates. Finally, the external droplet size dependence on the ratio of the flow rates Qmid /Qout was investigated. When Qmid /Qout is small, droplets form an elongated morphology (figure 5a inset). In order to work out the volume of these structures we assumed ellipsoidal geometry to model their shape, and determined the three axes a,b, and c for each ellipsoid (schematic shown in figure 5a). Since the droplets conserve their volume but eventually assume a spherical shape upon exiting the channel, we calculated the final diameter of the spheres from the volume of the corresponding ellipsoid. To explore this dependence, the value of Qin and Qmid were kept constant at 50 µL/hr and 200 µL/hr respectively, while Qout ranged from 200-5000 µL/hr. The same asymptotic behaviour is displayed on the graph of external droplet size against Qout (figure 5a) as has previously been seen for the internal droplet size and N (figures 4c and 3d) respectively. It has been shown 1 that droplet size as a function of the ratio between the dispersed phase (Qdis ) to the continuous phase (Qcon ) flow rate displays two regimes. In the first regime, where Qdis /Qcon < 1, droplet size stays approximately constant while in the second regime, where Qdis /Qcon > 1, the droplet size increases linearly with increasing dispersed phase flow rate. 1 Assuming that Qin plays no role on this relationship, in our system the dispersed phase flow rate is Qmid while the continuous phase flow rate is Qout . A double logarithmic plot of external droplet size against Qmid /Qout , figure 5b, shows that our

16


Page 16 of 31

Page 17 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

a)

b)

d)

e)

c)

f)

Figure 4: (a) w/o/w double emulsions where the internal droplet size is 50 µm. Qout =1000 µL/hr, Qmid =200 µL/hr, Qin =20 µL/hr (b) w/o/w double emulsions where the internal droplet size is around 35 µm. Qout =1000 µL/hr, Qmid =1000 µL/hr, Qin =20 µL/hr (c) Graph showing the size dependence of the internal droplet as a function of flow rate, Qmid . Internal droplet diameters were varied by more than 30%. (d-f:) w/o/w double emulsions. (d) Qout =1000 µL/hr, Qmid =1200 µL/hr, Qin =20 µL/hr. The middle flow rate is too high and does not allow for the outer phase to form the double emulsion. (e) Single internal droplet encapsulation by droplet formation in the first junction where Qout =1000 µL/hr, Qmid =150 µL/hr, Qin =100 µL/hr (f ) Co-flowing of the inner and middle phases resulting in a single internal droplet of around 150 µm, where Qout =1000 µL/hr, Qmid =200 µL/hr, Qin =400 µL/hr. Scale bar for all images is 200 µm. data display similar behaviour as that seen by Garstecki et al. for Qmid /Qout < 1. 1 In this regime, our data indicate a weak external droplet size dependency on this ratio, the gradient and hence the apparent exponent of the corresponding power law being approximately 0.25. We were unable to reach high Qmid /Qout values, however, as when the outer phase flow rate is not high enough for droplet formation, phase separation

17


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a)

b)

b)

Figure 5: (a) Graph showing the external droplet size dependence on outer phase flow rate, when Qin =50 µL/hr and Qmid =200 µL/hr. There is an asymptotic relation between the droplet size and Qout . Inset: Schematic representation of elongated droplets formed in a microfluidic channel and ellipsoid with its three main axes a, b, and c. (b) Log-log plot of the external droplet size as a function of Qmid /Qout . The gradient is 0.25 and most of the data points are in the regime where Qmid /Qout 1 we could not determine whether 3D double emulsion microfluidics exhibit the same linear behaviour (i.e. a scaling exponent of 1) as in the simple planar case.

Water/Oil/Water double emulsions with encapsulated silk fibroin Forming an emulsion where proteins are involved is particularly challenging due to the propensity of proteins to adhere to the surfaces of microfluidic devices. Thus, when attempting to form w/o emulsions, it is essential to render the surface of the channels more hydrophobic in order to avoid protein surface wetting. However, in double emulsion formation, controlling the surface properties of the two junctions can be challenging in

18


Page 18 of 31

Page 19 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

a)

e)

g)

Biomacromolecules b)

c)

d)

f)

h)

Figure 6: (a-f:) w/o/w double emulsion where the inner phase contains reconstituted silk fibroin. A concentration of 0.1 mg/mL was used. (a) Qout =500 µL/hr, Qmid =200 µL/hr, Qin =20 µL/hr. 4 internal droplets were encapsulated within the external droplet. (b) Qout =150 µL/hr, Qmid =200 µL/hr, Qin =20 µL/hr with 12 encapsulated internal droplets. It is clear that the protein does not wet the surface of the channels in the first junction. (c) Qout =400 µL/hr, Qmid =100 µL/hr, Qin =100 µL/hr. Double emulsions were generated via co-flowing the inner and middle phases in the first junction followed by encapsulation in the second junction. The concentration of protein used was 10 mg/mL. (d) Qout =350 µL/hr, Qmid =250 µL/hr, Qin =20 µL/hr. Double emulsion formation where the internal droplets are 15 µm in diameter. The ratio Qmid /Qin is kept > 10 so that the aqueous phase does not wet the surface. The concentration of protein used was 10 mg/mL. (e-h: left panels) Bright field images of double emulsions with two and three internal droplets respectively. (e-h: right panels) Fluorescent microscopy images of the same droplets as shown in the left panels. the case of protein solutions. Therefore, using a device that does not need surface modifications and yet allows double emulsion formation with protein as the inner phase

19


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

is important if we are to make double emulsions in a reproducible and facile manner. By correctly tuning the ratio Qmid /Qin , we were able to form double emulsions which contained reconstituted silk fibroin as the inner phase. The protein-containing droplets exhibit the same behaviour as water droplets, and parameters such as number of internal droplets or external droplet size can be clearly controlled as can be seen in figure 6a and 6b respectively. The values of protein concentration employed ranged across three orders of magnitude, from 0.01 mg/mL up to 10 mg/mL. We found that for protein concentrations greater than 1 mg/mL, double emulsions formation was possible by co-flowing the inner and middle phases in the first junction followed by encapsulation in the second junction, which can be seen in figure 6c. Using a protein concentration of 10 mg/mL, double emulsions with internal droplets as small as 15 µm were made, shown in figure 6d. This was achieved by employing a large value of Qmid /Qin so as to ensure that the protein would not wet the surface of the channels. Double emulsions containing a solution of reconstituted silk fibroin with ThT as the inner phase were produced. The droplets were incubated at 37 ◦ C for 2 days in order to promote aggregation and as a result internal droplets gelate and silk fibroin microgels are formed. As ThT is a fluorescent dye that mostly binds to β-sheet structures, protein-containing droplets that have undergone a structural transition can be detected. ThT-stained internal droplets were thus studied using fluorescent microscopy. To avoid droplet generation in the co-flowing regime, a concentration of 0.1 mg/mL silk fibroin was used. w/o/w emulsions with 2 and 3 internal droplets were imaged both in bright field and by using ThT filters. In the bright field images (left panels on figure 6e-6h) both internal and external droplets are clearly visible. However, in the images where we use of ThT filters (right panels on figure 6e-6h) only droplets containing aggregated silk fibroin can be detected.

20


Page 20 of 31

Page 21 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

A question that arises in double emulsions in which one of the aqueous phases contains dissolved macromolecules, is whether osmotic gradient effects may result in swelling or shrinking of the droplets. This has been studied in detail in the pioneering work of Mezzenga et al., 46 where it was shown that the size of the droplets is ultimately determined by a balance between the osmotic pressure and the Laplace pressure. Given the presence of silk protein and PEG in the aqueous phases employed in our study, one may expect such osmotic effects to play an important role. Yet, we observe no significant change in droplet size over the time scale of our observations (two days). We suggest this may be due to the large oil shell thickness the water molecules would have to diffuse through between aqueous phases, coupled with the low diffusion constant of water in the strongly hydrophobic fluorinated oil used in our emulsions. Micro-gel formation due to protein aggregation may further reduce the rate of water diffusion. 31 Thus, while our droplets might be expected to change in size so as to attain the necessary balance between osmotic and Laplace pressure, this probably occurs over time scales much longer than our experimental observations.

Oil/Water/Oil double emulsions Given the intrinsic hydrophobicity of PDMS, producing o/w/o double emulsions without surface treatment is particularly challenging. However, when the first junction is fabricated as a 3-D structure, the geometry of the microfluidic device favours the formation of such emulsions, as has been previously speculated. 2 Nevertheless, using the geometry of our device (i.e. the second junction being 3-D), o/w/o emulsions were generated by tuning the ratio between Qin , Qmid and Qout appropriately. Unlike w/o/w double emulsions, in o/w/o emulsions, there is less control over internal droplet formation. This is because after the first junction (see figure 1c, d), the inner oil phase tends to wet 21


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the hydrophobic surface of the channel, making it challenging to produce o/w droplets. Co-flowing the oil and water phases in the first junction, followed by encapsulation in the second junction seems to be the best approach of preparing double emulsions. However, by regulating Qmid /Qin , we have successfully encapsulated multiple, small, internal droplets within the external droplet, where instead of co-flowing the inner and middle phases, we form oil-in-water droplets at the first junction. Moreover, we believe that the role played by the surfactant in the middle phase is extremely vital for co-flowing to occur. It appears likely that the hydrophobic part of the amphiphilic surfactant adsorbs onto the surfaces of the hydrophobic PDMS channel, while the hydrophilic part remains in contact with the aqueous solution. Thus, the aqueous middle phase wets the surfactant-bearing PDMS channels and prevents the oil from doing so. This idea is supported by experiments conducted using lower amounts of surfactant. It was found that for surfactant concentrations of 1% w/w SDS and 1% w/w Tween 20 the oil was able to wet the surface despite following the same experimental procedure, i.e. flowing the middle phase before the inner one and having a high Qmid /Qin ratio. The same result was obtained when both surfactant concentrations were doubled. In fact, only when 3% w/w SDS and 3% w/w Tween 20 were used, did we observe double emulsion formation. Double emulsions consisting of multiple internal droplets were observed when Qmid /Qin was 10. By keeping Qout constant at 1000 µL/hr and ranging Qmid /Qin from 10 to 1, the size of the internal droplets increased to the point where only one droplet was encapsulated within the external droplet (when Qmid /Qin was equal to 2). Further increase of Qin resulted in augmentation of the internal droplet. The effect of flow rate on size and number of internal droplets is summarised in figure 7a-c and in S3. The largest droplets we were able to form were of diameter 114 ± 2 µm, while the smallest appear to be ∼10 µm. Furthermore, we were able to reach internal droplet diameters of 5 - 10 µm

22


Page 22 of 31

Page 23 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

by applying a large Qmid /Qin . The first emulsion, oil-in-water, is formed in the jetting regime. This multiphase flow is then encapsulated by the outer phase as seen in figure 7d. a)

e)

b)

c)

d)

f)

Figure 7: (a-d:) o/w/o double emulsions. (a) Qout =1000 µL/hr, Qmid =100 µL/hr, Qin =10 µL/hr. Internal droplet size is 10 µm. (b) Qout =1000 µL/hr, Qmid =100 µL/hr, Qin =40 µL/hr. Internal droplet size is 25 µm. (c) Qout =1000 µL/hr, Qmid =100 µL/hr, Qin =100 µL/hr. The inner and middle phases co-flow in the first junction. At the second junction, the outer phase encapsulates both the inner and middle phases resulting in internal droplets of 114 ± 2 µm (d) Qout =4000 µL/hr, Qmid =3000 µL/hr, Qin =100 µL/hr. The first o/w emulsion was formed in the jetting regime, producing oil droplets of 5-10 µm. (e-f:) Fluorescent microscopy images of o/w/o droplets consisting of 0.1 mg/mL silk fibroin solution with 100 µM ThT as the middle phase. For all images, the incubation temperature was 37 ◦ C. (e: left panel) Image taken after 12 hours of incubation. The white points indicate the formation of small aggregates and give an insight towards the first steps of aggregation whereas the black circles correspond to the internal oil droplets. (e: right panel) Image taken after 24 hours of incubation. Aggregation has proceeded resulting in larger clumps being visible. (f ) Fluorescent microscopy images of an o/w/o droplet trapped in a capillary. (f: left panel) Image taken after 24 hours of incubation. There are few clumps present, which suggests that the protein has almost fully aggregated. (f: right panel) Image taken after 48 hours of incubation. There are no clumps present, indicating that the protein has fully aggregated and the core-shell structure is stable.

Emulsions consisting of silk fibroin solution in the middle phase were formed. Avoid23


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ing surface wetting proved to be challenging when using high concentrations of protein, as the aqueous solution adhered to the channels of the device. However, it was determined that for protein concentration of 0.1 mg/mL, o/w/o double emulsions could be produced. ThT was added to the middle phase to detect aggregated protein. Using microscopy, silk fibroin aggregation was monitored at different stages of aggregation. Measurements were taken across a period of 2 days. Initially, after 12 hours of incubation at 37 ◦ C protein aggregates start to appear in the middle phase as small white points, where the black circles correspond to the non fluorescent internal oil droplets, as can be seen in figure 7e (left panel). After 24 hours of observation, silk fibroin further aggregates resulting in the formation of larger clumps, which are clearly seen in figure 7e (right panel), and they correspond to the brighter areas, with the darker areas again being the internal oil droplets. Droplets were unstable when incubated in bulk for more than 24 hours. In order to circumvent this problem, the use of capillaries was employed. Droplets were therefore trapped inside capillaries and imaged after 24 and 48 hours of incubation. After 24 hours few aggregate clumps were detected, which suggests that most of the silk fibroin has gelled, as can be seen in the left panel of figure 7f. However, after 48 hours, there are no clumps present (right panel of figure 7f), indicating that the aggregation is complete and that the core-shell structure has fully gelled. It is worth noting that for the case of o/w/o emulsions, since there is no osmotic gradient between the inner and outer phases, no droplet swelling or shrinking is expected.

Conclusions We have demonstrated that a double junction microfluidic device, where only the second junction is fabricated as a 3-D structure, is capable of producing both w/o/w and o/w/o emulsions without the need for surface modification. Furthermore, we have shown that 24


Page 24 of 31

Page 25 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

we can readily prepare monodisperse droplets and vary internal and/or external droplet size by adjusting the ratio between flow rates. For w/o/w emulsions, we have achieved full control over the number of encapsulated droplets, N, which for the first time can be varied in a predictable and reproducible manner. Moreover, in w/o/w emulsions we have shown that the shell thickness of the oil phase can be easily regulated. High speed imaging was used to study the evolution of droplet formation and monitor the dynamic contact angle, which was found to exceed 90◦ and remained essentially constant during droplet formation, growth and detachment. Additionally, we demonstrate that our nonsurface modified device is capable of forming double emulsions where the inner phase contains reconstituted silk fibroin solution. The protein does not adhere to the surface of the PDMS and thus, this device is ideal for practical applications such as protein encapsulation, and shows promise towards potential use in biomedical studies. Following incubation for 4 days, fluorescent microscopy using ThT as the fluorescent dye confirms that the internal droplets have gelled and formed micro-gels. Finally, we were able to generate o/w/o double emulsions with either single or multiple internal droplets showing a wide range of sizes, from 10 µm to 100 µm. Such small micro-droplets have potential as carriers for drug delivery, and in the biomedical industry in general, 14–16,23,24 where the smaller the carrier droplet, the better. Likewise, o/w/o emulsions where the middle phase consisted of silk fibroin solution were formed. Such bio-emulsions show great promise, as core-shell structures with a variable shell thickness can be tailored so that the release kinetics of valuable cargo molecules encapsulated within the internal droplets can be precisely controlled, making them ideal for drug delivery applications. The ability of this device to reproducibly form both w/o/w and o/w/o emulsions, combined with its simplicity and versatility, render it particularly promising for potential applications in the pharmaceutical, cosmetic and food industries due to its ease of production and

25


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

scale-up potential.

Supporting Information Figure S1. Images of the static contact angle between different solutions and PDMS. Figure S2a-c. Variation of viscosity as a function of shear rate for solutions used in double emulsion formation. Figure S2d. Dependence of the number of internal droplets (N) on outer phase flow rate (Qout ). Figure S3. o/w/o double emulsion droplet formation at different flow rates of the inner phase (Qin ).

Acknowledgments The research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Programme (FP7/2007-2013) through the ERC grant PhysProt (agreement n◦ 337969). Dr Aviad Levin would like to thank the FEBS Long-Term Fellowship for their support. The authors would like to thank Dr Marc Rodriguez-Garcia for helpful discussions and Alessio Caciagli for his help with all the viscosity measurements.

26


Page 26 of 31

Page 27 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

References (1) Garstecki, P.; Fuerstman, M. J.; Stone, H. a.; Whitesides, G. M. Lab Chip 2006, 6, 437–446. (2) Rotem, A.; Abate, A. R.; Utada, A. S.; Van Steijn, V.; Weitz, D. a. Lab Chip 2012, 12, 4263. (3) Tan, S. H.; Nguyen, N. T.; Chua, Y. C.; Kang, T. G. Biomicrofluidics 2010, 4 . (4) Mazutis, L.; Gilbert, J.; Ung, W. L.; Weitz, D. A.; Griffiths, A. D.; Heyman, J. A. Nat. Protoc. 2013, 8, 870–891. (5) Abate, A. R.; Thiele, J.; Weitz, D. A. Lab Chip 2010, 105, 1774–1776. (6) Romanowsky, M. B.; Heymann, M.; Abate, R.; Krummel, T.; Fraden, S.; Weitz, D. Lab Chip 2010, 10, 1521–1524. (7) Bashir, S.; Bashir, M.; Casadevall i Solvas, X.; Rees, J. M.; Zimmerman, W. B. J. Micromech. Microeng. 2015, 6, 1445–1458. (8) Eggersdorfer, M. L.; Zheng, W.; Nawar, S.; Mercandetti, C.; Ofner, A.; Leibacher, I.; Koehler, S.; Weitz, D. A. Lab Chip 2017, 17, 936–942. (9) Davies, R. T.; Kim, D.; Park, J. J. Micromech. Microeng. 2012, 22, 055003. (10) Nisisako, T.; Okushima, S.; Torii, T. Soft Matter 2005, 1, 23. (11) Bauer, W.-A. C.; Fischlechner, M.; Abell, C.; Huck, W. T. S. Lab Chip 2010, 10, 1814–9. (12) Chang, F.-c.; Su, Y.-c. J. Micromech. Microeng. 2008, 18, 1–8. 27


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(13) Cole, R. H.; Tran, T. M.; Abate, A. R. J. Visualized Exp. 2015, e53516. (14) Kim, S.-H.; Kim, J. W.; Cho, J.-C.; Weitz, D. a. Lab Chip 2011, 11, 3162–3166. (15) Giri, T. K.; Choudhary, C.; Ajazuddin,; Alexander, A.; Badwaik, H.; Tripathi, D. K. Saudi Pharm. J. 2013, 21, 125–141. (16) Lee, J.-s.; Kim, J.-w.; Han, S.-h.; Chang, I.-s.; Kang, H.-h.; Lee, O.-s.; Oh, S.-g.; Suh, K.-d. J. Cosmet. Sci. 2004, 55, 1–12. (17) Mishra, B.; Pandit, J. K. J. Controlled Release 1990, 14, 53–60. (18) Vasiljevic, D.; Parojcic, J.; Primorac, M.; Vuleta, G. Int. J. Pharm. (Amsterdam, Neth.) 2006, 309, 171–177. (19) Qi, X.; Wang, L.; Zhu, J. J. Pharm. Sci. 2011, 100, 2203–2211. (20) Cohen, S.; Yoshioka, T.; Lucarelli, M.; Hwang, L. H.; Langer, R. Pharm. Res. 1991, 8, 713–720. (21) Mastrobattista, E.; Taly, V.; Chanudet, E.; Treacy, P.; Kelly, B. T.; Griffiths, A. D. Chem. Biol. 2005, 12, 1291–1300. (22) Singh, M. N.; Hemant, K. S. Y.; Ram, M.; Shivakumar, H. G. Res. Pharm. Sci. 2010, 5, 65–77. (23) Lim, F.; Sun, a. Science 1980, 210, 908–910. (24) Dusseault, J.; Leblond, F. A.; Robitaille, R.; Jourdan, G.; Tessier, J.; Ménard, M.; Henley, N.; Hallé, J. P. Biomaterials 2005, 26, 1515–1522. (25) Chan, H. F.; Zhang, Y.; Ho, Y.-P.; Chiu, Y.-L.; Jung, Y.; Leong, K. W. Sci. Rep. 2013, 3, 3462. 28


Page 28 of 31

Page 29 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biomacromolecules

(26) Zhao, C. X.; Chen, D.; Hui, Y.; Weitz, D. A.; Middelberg, A. P. J. ChemPhysChem 2016, 17, 1553–1556. (27) Imlmal, E.; Rekax, C. J. Controlled Release 1995, 33 . (28) Laugel, C.; Baillet, a.; Ferrier, D.; Grossiord, J. L.; Marty, J. P. Int. J. Cosmet. Sci. 1998, 20, 183–91. (29) Zhao, C. X.; Chen, D.; Hui, Y.; Weitz, D. A.; Middelberg, A. P. ChemPhysChem 2017, 18, 1393–1399. (30) Boyadzhiev, L.; Bezenshek, E. J. Membr. Sci. 1983, 14, 13–18. (31) Shimanovich, U.; Efimov, I.; Mason, T. O.; Flagmeier, P.; Buell, A. K.; Gedanken, A.; Linse, S.; ˚ Akerfeldt, K. S.; Dobson, C. M.; Weitz, D. A.; Knowles, T. P. J. ACS nano 2014, 9, 43–51. (32) Kim, J.; Hwang, I.; Britain, D.; Chung, T. D.; Sun, Y.; Kim, D.-H. Lab Chip 2011, 11, 3941–8. (33) Zhou, X. M.; Shimanovich, U.; Herling, T. W.; Wu, S.; Dobson, C. M.; Knowles, T. P.; Perrett, S. ACS Nano 2015, 9, 5772–5781. (34) Lefèvre, T.; Rousseau, M.-E.; Pézolet, M. Biophys. J. 2007, 92, 2885–2895. (35) Vepari, C.; Kaplan, D. L. Prog. Polym. Sci. 2007, 32, 991–1007. (36) Rockwood, D. N.; Preda, R. C.; Y¨ ucel, T.; Wang, X.; Lovett, M. L.; Kaplan, D. L. Nat. Protoc. 2011, 6, 1612–1631. (37) Terry, A. E.; Knight, D. P.; Porter, D.; Vollrath, F. Biomacromolecules 2004, 5, 768–772. 29


Biomacromolecules

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 31

(38) Tabatabai, A. P.; Kaplan, D. L.; Blair, D. L. Soft Matter 2015, 11, 756–761. (39) Ayub, Z. H.; Arai, M.; Hirabayashi, K. Biosci., Biotechnol., Biochem. 1993, 57, 1910–1912. (40) Mitropoulos, A. N.;

Perotto, G.;

Kim, S.;

Marelli, B.;

Kaplan, D. L.;

Omenetto, F. G. Adv. Mater. 2014, 26, 1105–1110. (41) Elia, R.; Guo, J.; Budijono, S.; Normand, V.; Benczédi, D.; Omenetto, F.; Kaplan, D. L. J. Coat. Technol. Res. 2015, 12, 793–799. (42) Pritchard, E. M.; Kaplan, D. L. Expert Opin. Drug Delivery 2011, 8, 797–811. (43) Zhao, Z.; Li, Y.; Xie, M. B. Int. J. Mol. Sci. 2015, 16, 4880–4903. (44) Challa, P. K.; Kartanas, T.; Charmet, J.; Knowles, T. P. J. Biomicrofluidics 2017, 11, 014113. (45) Huang, S.-H.; Tan, W.-H.; Tseng, F.-G.; Takeuchi, S. J. Micromech. Microeng. 2006, 16, 2336–2344. (46) Mezzenga, R.; Folmer, B. M.; Hughes, E. Langmuir 2004, 20, 3574–3582.

30


Page 31 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

e)

Biomacromolecules

Inner Phase Middle Phase Outer Phase

31


Hierarchical Biomolecular Emulsions Using 3-D Microfluidics with

Recommend Documents