microfluidic development of bubble-templated microstructured … · 2012-11-01 · this thesis...
TRANSCRIPT
Microfluidic Development of Bubble-templated Microstructured
Materials
By
Jai Il Park
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Chemistry
University of Toronto
© Copyright by Jai Il Park, 2010
II
Microfluidic Development of Bubble-templated
Microstructured Materials
Jai Il Park
Doctor of Philosophy
Department of Chemistry
University of Toronto
2010
Abstract
This thesis presented a microfluidic preparation of bubbles-templated micro-size materials.
In particular, this thesis focused on the microfluidic formation and dissolution of CO2
bubbles. First, this thesis described pH-regulated behaviours of CO2 bubbles in the
microfluidic channel. This method opened a new way to generate small (<10 µm in
diameter) with a narrow size distribution (CV<5%). Second, the microfluidic dissolution of
CO2 bubbles possessed the important feature: the local change of pH on the bubble surface.
This allowed us to encapsulate the bubbles with various colloidal particles. The bubbles
coated with particles showed a high stability against coalescences and Ostwald ripening.
The dimensions and shapes of bubbles with a shell of colloidal particle were manipulated
III
by the hydrodynamic and chemical means, respectively. Third, we proposed a microfluidic
method for the generation of small and stable bubbles coated with a lysozyme-alginate shell.
The local pH decrease at the periphery of CO2 bubbles led to the electrostatic attraction
between lysozyme on the bubble surface and alginate in the continuous phase. This
produced the bubbles with a shell of biopolymers, which gave a long-term stability (up to a
month, at least) against the dissolution and coalescence. Fourth, we presented a single-step
method to functionalize bubbles with a variety of nanoparticles. The bubbles showed the
corresponding properties of nanoparticles on their surface. Further, we explored the
potential applications of these bubbles as contrast agents in ultrasound and magnetic
resonance imaging.
Key words: microfluidics, microreactors, microbubble, polymer particles, carbon dioxide,
colloidal assembly, Pickering emulsions, nanoparticles, ultrasound imaging, magnetic
resonance imaging.
IV
Acknowledgements
First of all, I cannot express fully my gratitude to my Ph.D. supervisor, Prof. Eugenia
Kumacheva. With her great enthusiasm, she continuously teaches, advises and supports me
throughout my graduate life. This thesis would not have been completed without her
encouragement. I am also thankful to Prof. Mitchell A. Winnik and Prof. Michael K.
Georges for being my supervisory committee members and for providing me with valuable
advice.
I would also like to thank all current and former members in Prof. Kumacheva’s group
and my friends in the chemistry department at University of Toronto. It was a truly
experience of a lifetime working with such nice, kind and intelligent people: Dr. Zhihong
Nie, Dr. Alla Petukhova, Dr. Wei Li, Dr. Kun Liu, Dr. Jesse Greener, Dr. Ryan Simms, Dr.
Hung Pham, Dr. Nana Zhao, Dr. Hong Zhang, Dr. Daniele Fava, Dr. Mallika Das, Dr.
Hyunwoo Kim, Dr. Jungho Kim, Dr. Yonghwan Cho, Dr. Sungyeon Choi, Dr. Kyungtaek
Kim, Dr. Sunmok So, Dr. Minseok Seo, Dr. Jeffrey Haley, Dr. Jinshe Song, Mr. Ilya
Gourevich, Mr. Ivan Gorelikov, Ms. Lindsey Fiddes, Ms. Anna Lee, Mr. Ethan Tumarkin,
Mr. Stanislav Dubinski, Mr. Andrew Paton, Ms. Siyon Chung, Ms. Neta Raz, and Ms. Lsan
Tzadu.
I also acknowledge Prof. Axel Günther and Prof. Arron Wheeler for being one of my
committee members and giving valuable suggestions for my completion of Ph.D. I also
wish to acknowledge my collaborators Prof. Howard A. Stone, Prof. Bernard P. Binks, Prof.
Stefan A. F. Bon, Prof. Greg Staniz, Dr. Naomi Matsuura, Mr. Ross Williams, and Ms.
Wendy Oakden who provided valuable input for my different research projects.
My very special thanks belong to my wife, Min Hee, Han, and my 2 year-old son, Shin
Young, Park for their being with me and giving me limitless love. I also give my full
appreciation to my parents, parents-in-law, sister, brothers-in-law, uncles, aunts, cousins
and grandmothers for their constant love and support.
I am thankful to University of Toronto for the financial support.
V
Preface
This thesis was written in part based on a series of papers (see the list below), which
have been published in (or submitted to) peer-reviewed scientific journals. As identified by
primary authorship, all manuscripts were written by Jai Il Park with critical comments and
revision by Prof. Eugenia Kumacheva and corresponding collaborators. Jai Il Park
contributed to the papers by designing and carrying out key experiments, data analysis and
interpretation, and manuscript writing. The contributions of other authors are provided in
detail below.
Chapter 3. Microfluidic Generation and Dissolution of CO2 bubbles
The results in this chapter are from manuscripts published in J.I. Park, Z.H. Nie, A.
Kumachev, E. Kumacheva. A Microfluidic Route to Small CO2 Microbubbles with
Narrow Size Distribution. Soft Matter 2010, 6, 630.
Contribution: Z. H. Nie conducted preliminary work on the generation and dissolution of
CO2 bubbles. He also contributed to data analysis and provided critical comments. A.
Kumachev helped with making devices and conducting experiments.
Chapter 4. Assembly of Colloidal Particles at Gas-Water and Water-Oil Interfaces
The results in this chapter are from manuscripts published in J.I. Park, Z.H. Nie, A.
Kumachev, A. I. Abdelrahman, H. A. Stone, B. P. Binks, E. Kumacheva. A Microfluidic
Approach to Chemically Driven Assembly of Colloidal Particles at Gas-Liquid
Interfaces. Angewandte Chemie International Edition 2009, 48, 5300.,1 and Z.H. Nie, J.I.
Park (co-first author), W. Li, S. Bon, E. Kumacheva. An "Inside-Out" Microfluidic
Approach to Monodisperse Emulsions Stabilized by Solid Particles. Journal of the
American Chemical Society 2008, 130, 16508.2
Contribution: 1Z. H. Nie designed and conducted several key experiments. A. I.
Abdelrahman provided carboxylated polystyrene particles. A. Kumachev contributed to
make devices and conduct preliminary experiments. Prof. H. A. Stone designed a
theoretical model for the deposition of particles on the surface of bubbles. Prof. B. P. Bink
gave a critical input to manuscript writing. 2J.I. Park conducted several key experiments
VI
and finalized the manuscript. W. Li helped to obtain the confocal microscope images. Prof.
S. Bon initiated the project and provided guidance.
Chapter 5. Bubbles Encapsulated with a Shell of Biopolymers
The results in this chapter are from manuscripts published in J.I. Park, E. Tumarkin, E.
Kumacheva. Small, Stable, and Monodispersed Bubbles Encapsulated with
Biopolymers. Macromolecular Rapid Communications 2010, 31, 222.
Contribution: E. Tumarkin conducted several key experiments and helped to polish the
manuscript.
Chapter 6. A Single-step Microfluidic Route to Producing Multifunctional
Microbubbles The results in this chapter are from manuscripts submitted to J.I. Park, D.
Jagadeesan, R. Williams, W. Oakden, S. Chung, G. Stanisz, E. Kumacheva. A Single-step
Microfluidic Route to Producing Multifunctional Microbubbles. Journal of the
American Chemical Society 2010.
Contribution: D. Jagadeesan and S. Chung conducted experiments and gave valuable
suggestions to the manuscript. R. Williams performed ultrasound imaging experiments,
data analysis and provided critical inputs to the manuscript. W. Oakden conducted
magnetic resonance imaging experiments and analyzed the data. G. Stanisz gave valuable
direction to magnetic resonance imaging experiments.
VII
Publications during PhD Study
Peer-reviewed Papers:
1. E. Tumarkin, Z. Nie, J.I. Park. Greener, E. Kumacheva, Microfluidic Mimicking
Solubility Pump. 2010, in preparation. 2. J.I. Park, D. Jagadeesan, R. Williams, W. Oakden, S. Chung, G. Stanisz and E.
Kumacheva. A Single-step Microfluidic Route to Producing Multifunctional
Microbubbles. Journal of the American Chemical Society 2010, submitted.
3. J.I. Park, A. Saffari, S. Kumar, A. Günther, E. Kumacheva. Microfluidic Synthesis
of Polymer and Inorganic Particulate Materials. Annual Review of Materials
Research 2010, 40, 415, invited review article.
4. J.I. Park, E. Tumarkin, E. Kumacheva. Small, Stable, and Monodispersed
Bubbles Encapsulated with Biopolymers. Macromolecular Rapid
Communications 2010, 31, 222.
5. J.I. Park, Z.H. Nie, A. Kumachev, E. Kumacheva. A Microfluidic Route to Small
CO2 Microbubbles with Narrow Size Distribution. Soft Matter 2010, 6, 630.
6. J.I. Park, Z.H. Nie, A. Kumachev, A. I. Abdelrahman, H. A. Stone, B. P. Binks, E.
Kumacheva. A Microfluidic Approach to Chemically Driven Assembly of
Colloidal Particles at Gas-Liquid Interfaces. Angewandte Chemie International
Edition 2009, 48, 5300, cover page.
7. S. Dubinsky, J.I. Park, I. Gourevich, H.K.C. Chan, M. Deetz, E. Kumacheva,
Towards Controlling the Surface Morphology of Macroporous Copolymer
Particles. Macromolecules 2009, 42, 1990.
8. Z.H. Nie, J.I. Park (co-first author), W. Li, S. Bon, E. Kumacheva. An "Inside-
Out" Microfluidic Approach to Monodisperse Emulsions Stabilized by Solid
Particles. Journal of the American Chemical Society 2008, 130, 16508.
VIII
Oral Presentations in Conferences:
1. J.I. Park, E. Tumarkin, E. Kumacheva. Small, Stable, and Monodisperse Bubbles
Encapsulated with Biopolymers, Canadian Chemistry Conference, Toronto,
Canada, 2010.
2. J.I. Park, Z.H. Nie, Kumachev. A. I. Abdelrahman, B. P. Binks, H. A. Stone, E.
Kumacheva, A Microfluidic Approach to Chemically Driven Assembly of
Colloidal Particles at Gas-Liquid Interfaces, MicroTas, Jeju, Korea, 2009.
3. J.I. Park, Z.H. Nie, E. Kumacheva, Using Microfluidics to Study Dissolution of
Carbon Dioxide, Canadian Chemistry Conference, Hamilton, Canada, 2009
IX
Table of Content Chapter 1 .......................................................................................................................... 1
Introduction to Microfluidic Synthesis of Particulate Materials .................................... 1
1.1 Overview of Microfluidics .................................................................................... 1
1.2 Different Modes of Microfluidic Synthesis of Polymer Particles ........................... 2
1.3 Microfluidic Synthesis of Polymer Particles ......................................................... 5
1.3.1 Microfluidic Synthesis of Solid Polymer Microspheres .............................. 6
1.3.2 Microfluidic Synthesis of Spherical Polymer Microgels ............................. 8
1.3.3 Microfluidic Synthesis of Polymer Particles with Different Morphologies 10
1.3.4 Microfluidic Synthesis of Non-spherical Polymer Particles ...................... 14
1.3.5. Microfluidic Synthesis of Bubble-templated Polymer Particles ............... 17
1.3.6 Productivity of Continuous Microfluidic Reactors .................................... 18
1.4 Outlook .............................................................................................................. 20
References ................................................................................................................ 21
Chapter 2 ........................................................................................................................ 30
Materials and Methods .................................................................................................. 30
2.1 Materials ............................................................................................................ 30
2.2 Methods.............................................................................................................. 31
2.2.1 Mask Design ............................................................................................ 31
2.2.2 Microfabrication of Negative Masters ...................................................... 32
2.2.3 Fabrication of Microfluidic Devices ......................................................... 32
2.2.4 Microfluidic Experiments ......................................................................... 33
2.2.5 Photopolymerization Experiments ............................................................ 34
2.3. Characterization ................................................................................................. 34
2.3.1 Optical and Fluorescence Microscopy Imaging ........................................ 34
2.3.2 Size Distribution of Bubbles, Droplets and Particles ................................. 35
2.3.3 Scanning and Transmission Electron Microscopy imaging ....................... 35
2.3.4 Laser Confocal Fluorescence Microscopy imaging ................................... 35
2.3.5 Contact angle measurement ...................................................................... 36
2.3.6 Circular Dichroism Spectroscopy ............................................................. 36
X
2.3.7 Fluorescence Spectroscopy ...................................................................... 37
References ................................................................................................................ 37
Chapter 3 ........................................................................................................................ 40
Microfluidic Generation and Dissolution of CO2 bubbles ............................................ 40
3.1 Introduction ........................................................................................................ 40
3.2 Results and Discussions ...................................................................................... 42
3.2.1 Generation of CO2 bubbles ....................................................................... 42
3.2.2 Dissolution of CO2 Bubbles ..................................................................... 49
3.2.3 Generation of Small Bubbles .................................................................... 55
3.3 Conclusion ......................................................................................................... 56
References ................................................................................................................ 57
Chapter 4 ........................................................................................................................ 60
Assembly of Colloidal Particles at Gas-Water and Water-Oil Interfaces .................... 60
4.1 Introduction ........................................................................................................ 60
4.2 Results and Discussion ....................................................................................... 63
4.2.1 Synthesis of Colloidal Particle Armoured Bubbles ................................... 63
4.2.1.1 Experimental Design ..................................................................... 63
4.2.1.2 Generation of Armoured Bubbles .................................................. 64
4.2.1.3 Effect of Flow Rate on the Dimensions of Armoured Bubbles ....... 68
4.2.1.4 Effect of pH and Particle Concentration on the Dimension and
Morphologies of Armoured Bubbles.......................................................... 69
4.2.1.5 Generality of the Assembly of Colloidal Particles at Gas-Water
Interface .................................................................................................... 73
4.2.2 Synthesis of Colloidal Particle-Coated Droplets ....................................... 74
4.2.2.1 Experimental Design ..................................................................... 75
4.2.2.2 Microfluidic Control of Particle Coverage at Water-Oil Interfaces . 76
4.3. Conclusion......................................................................................................... 81
References ................................................................................................................ 82
XI
Chapter 5 ........................................................................................................................ 85
Bubbles Encapsulated with a Shell of Biopolymers ...................................................... 85
5.1 Introduction ........................................................................................................ 85
5.2 Results and Discussions ...................................................................................... 88
5.2.1 Preparation of Bubbles Encapsulated with a Mixture of Biopolymers ....... 88
5.2.2. Long-term Stability of Bubbles ............................................................... 91
5.2.3 Control of Bubble Dimensions ................................................................. 93
5.2.4 Characteristic Properties of the Bubbles ................................................... 97
5.3 Conclusion ....................................................................................................... 100
References .............................................................................................................. 101
Chapter 6 ...................................................................................................................... 104
A Single-step Microfluidic Route to Producing Multifunctional Microbubbles ........ 104
6.1 Introduction ...................................................................................................... 104
6.2 Results and Discussions .................................................................................... 106
6.2.1 Experimental Design .............................................................................. 106
6.2.2 Long-term Stability of Biopolymer Encapsulated-Bubbles Functionalized
with NPs ......................................................................................................... 109
6.2.3 Control of the Dimension of Biopolymer Encapsulated-Bubbles
Functionalized with NPs ................................................................................. 110
6.2.4 Control Experiments ............................................................................... 111
6.2.5 Characterization of Biopolymer Encapsulated-Bubbles Functionalized with
various NPs ..................................................................................................... 111
6.2.6 Control over the Amount of NPs on the Bubble Surface ......................... 113
6.2.7 Properties of NP-functionalized Bubbles ................................................ 113
6.2.8 Application of NP-functionalized Bubbles in US Imaging ...................... 114
6.2.9 Application of NP-functionalized Bubbles in MRI ................................. 116
6.3 Conclusions ...................................................................................................... 117
References .............................................................................................................. 118
Chapter 7 ...................................................................................................................... 121
XII
Conclusion and Outlook .............................................................................................. 121
Appendix ...................................................................................................................... 125
XIII
List of Figures
Chapter 1 Introduction to Microfluidic Synthesis of Particulate Materials
Figure 1.1 Synthesis of particulate materials (A) in a batch reactor and (B) in a flow-
based microfluidic (MF) reactor. Adapted with permission from reference 9.
Copyright 2010, Annual Reviews…………………………………………..…2
Figure 1.2 Schematic illustrations of (A) single-phase and (B) multiphase formats for the
MF synthesis of polymer particles synthesized (C) in droplets (i), and (D) at the
interface between the two phases (ii). Adapted with permission from reference
9. Copyright 2010, Annual Reviews…………………………………………..3
Figure 1.3 Rigid polymer particles synthesized in continuous multiphase MF reactors.
(A) SEM images of poly(tripropyleneglycol diacrylate) particles. (B) Size
distribution of the particles shown in panel (A). d denotes particle diameter. (C)
Optical fluorescence microscopy images of poly(tripropylene glycol diacrylate)
microspheres loaded with CdSe quantum dots. (D) Optical polarization
microscopy images of poly(tripropyleneglycol diacrylate) microspheres with 5–
20 wt. % of liquid crystals made of 4-cyano-4’-pentylbiphenyl. Panels (A–D)
are reproduced with permission from reference 20. Copyright 2005, Wiley-
VCH. (E) SEM image of porous poly(glycidyl methacrylate–co-ethylene glycol
dimethacrylate). Panel (E) is adapted with permission from reference 67.
Copyright 2009, American Chemical Society. (F) Confocal microscopy image
of poly(tripropyleneglycol diacrylate)/poly(urethane) particles with an
interpenetrating network (IPN) structure. Panel (F) is adapted with permission
from reference 68. Copyright 2008, American Chemical Society……………7
Figure 1.4 Microgel particles synthesized in continuous multiphase MF reactors. (A)
Optical microscopy images of the formation of droplets of poly(N-
isopropylacrylamide) microgel particles (left) and the resulting microgel
particles labeled with fluorescein isothiocyanate (right). The inset shows a
schematic of the microfluidic reactor (OF, outer fluid; MF, middle fluid).
Reproduced with permission from reference 55. Copyright 2007, Wiley-VCH.
XIV
(B) Optical microscopy images of ionically cross-linked alginate (left), k-
carrageenan (middle), and carboxymethylcellulose (right) microgels. Scale bar:
100 μm. Adapted with permission from reference 56. Copyright 2006,
American Chemical Society……………………………………………………9
Figure 1.5 MF synthesis of polymer particles with different morphologies. (A) Schematic
of the microfluidic synthesis of Janus microgels. The inset shows an optical
fluorescence image of poly(acrylamide) particles encapsulating silica particles
labeled with two different dyes in distinct compartments. Adapted with
permission from reference 72. Copyright 2006, American Chemical Society.
(B) Optical microscopy image of bicolored poly(isobornyl acrylate) particles
containing carbon black and titanium oxide in two distinct hemispheres. The
inset shows electrically actuated particles displaying their white-colored side in
the display panel. Adapted with permission from reference 79. Copyright 2006,
Wiley-VCH. (C) SEM images of microcapsules prepared by microfluidic
interfacial polymerization. The inset shows a high magnification image of the
interior of the capsular wall. Adapted with permission from reference 54.
Copyright 2005, American Chemical Society. (D) Optical microscopy image of
the colloidal crystalline array confined in polymeric shells. The inset shows the
particles with three colloidal crystal cores. Scale bars: 200 μm. Adapted with
permission from reference 99. Copyright 2008, American Chemical Society..11
Figure 1.6 MF synthesis of particles with non-spherical shapes. (A,B) SEM images of
poly(tripropyleneglycol diacrylate) particles with (A) disk and (B) rod shapes.
Insets show the relationship between the particle shape and the channel
dimensions. Adapted with permission from references 20 and 103. Copyright
2005, Wiley-VCH. Copyright 2005, American Chemical Society. (C) SEM
image of poly(tripropyleneglycol diacrylate) particles with bowl-like shapes.
The inset shows the precursor droplet with a small fraction of nonpolymerizable
liquid (indicated by the arrow). Scale bar: 40 μm. Adapted with permission
from reference 85. Copyright 2005, American Chemical Society. (D–F) SEM
images of poly(ethylene glycol diacrylate) particles with various shapes
synthesized in one-phase MF synthesis using continuous-flow lithography.
XV
Insets show the corresponding photomasks. Scale bars: 10 μm. Adapted with
permission from reference 15. Copyright 2006, Macmillan…………….…..15
Chapter 3 Microfluidic Generation and Dissolution of CO2 bubbles
Figure 3.1 Schematic of the MF device and the dissolution of CO2 bubbles. Inset shows a
zoomed in schematic of the orifice and the generation of bubbles……….…42
Figure 3.2 Schematic of pH-dependent dissolution of CO2 bubble leading to bubble
shrinkage……………………………………………………………………...43
Figure 3.3 Representative optical microscopy images taken immediately after the
generation of CO2 bubbles (A,B) and N2 bubbles (C,D) and in the downstream
channel 107 mm (A’-D’) away from the orifice. The initial pH value of the
continuous phase was 1.5 (A,A’,C,C’) and 13.2 (B,B’,D,D’). The bubbles were
generated at PCO2=27.6 kPa and QL=6 mL/h. Scale bar is 200 μm………….44
Figure 3.4 Variation in initial (●, ), and final (○,) volumes of CO2 (●,○) and N2
(,) bubbles plotted as a function of the initial pH of the continuous aqueous
phase……………………………………………………………………….….46
Figure 3.5 pH-dependence of the relative frequency of bubble generation with respect to
the frequency of the formation of bubbles at pH=13.2 for CO2(■) and N2()
bubbles. The bubbles were generated at P= 27.6 kPa, QL=6 mL/h……….…47
Figure 3.6 (A-C) Optical microscopy images of CO2 bubbles taken at varying distances
away from the orifice in the device at (A) pH=1.5, (B) pH=9, and (C) pH=13.2.
Scale bar is 50 μm. (D) Variation in the volume of CO2 bubbles examined at
different pH values, plotted vs the distance, d, away from the orifice……....48
Figure 3.7 (A) Comparison of theoretical (■) and experimental () amounts of dissolved
CO2 (mol/L), (B) Relative change in bubble volume plotted as a function of pH
of the continuous phase. PCO2=27.6 kPa, QL=6 mL/h……………………....52
Figure 3.8 Effect of the flow rate of the continuous phase, QL, on (A) the initial and (B)
final volume of CO2 bubbles generated at different pH values of the continuous
phase. pH=1.5 (,), pH=5(,), pH=7 (,), pH=9 (,), pH=11
XVI
(,), and pH=13.2 (,). PCO2=27.6 kPa. The lines are given for eye
guidance……………………………………………………………………....53
Figure 3.9 Variation in the final diameter of CO2 bubbles, plotted as a function of pH and
QL (each represents the QL range from 5 to 7 mL/h, respectively)………….54
Figure 3.10 (A) Generation of 30 μm-diameter CO2 bubbles. (B) Bubbles with a
stabilized diameter of 8 μm in the downstream channel. The bubbles were
generated in the microfluidic device with the length, width, and height of the
orifice of 60, 22 and 40 μm, respectively, at PCO2=55.2 kPa, QL=12 mL/h, and
pH=13.2……………………………………………………………………….55
Chapter 4 Assembly of Colloidal Particles at Gas-Water and Water-Oil Interfaces
Figure 4.1 (A) A schematic of a MF T-junction bubble generator. The widths of the main
and the side channels are 220 and 40 μm, respectively. The height of the
channels is 130 μm. (B) A schematic of the formation of colloidal particle shell
during the dissolution of CO2 bubbles……………………………………….62
Figure 4.2 Progression of the plugs to spherical armoured bubbles. Optical microscopy
images of bubbles at a distance of 0 (A), 30 (B), 80 (C) and 150 (D) mm from
the T-junction. Scale bars are 200 μm. Bubbles generated at pH=14, PCO2=34.5
kPa, QL=10.5 mL/h, initial particle concentration, Cp,=1.5 wt. %, and
28.8oC…………………………………………………………………………63
Figure 4.3 (A) Optical microscopy image of armoured bubbles generated as in Figure 4.2
and collected at the outlet of the MF device. Scale bar is 200 μm (B) Optical
microscopy image of close-packed crystalline shell of armoured bubble. Scale
bar is 25 μm (C) SEM image (side view) of the shell of the microtomed
armored bubble infiltrated with poly(ethylene glycol) diacrylate. Scale bar is 10
μm. The inset shows an LCFM image of the bubble coated with HY-labeled
PS-co-PAA particles. Scale bar is 25 μm……………………………………..64
Figure 4.4 Optical microscopy images of bubbles and plugs flowing through the
microchannels. (A) Dissolution of CO2 plugs in particle-free at pH=14. (B)
Plugs of gaseous N2 formed in a dispersion of anionic PS-co-PAA particles. (C)
XVII
Bubbles of CO2 generated in a dispersion of PS-co-PAA particles containing 2
wt. % of the non-ionic surfactant (Triton X-100). (D) Plugs of N2 formed in a
dispersion of 700 nm diameter cationic PMMA-co-P4VP particles. (E) Bubbles
of CO2 dispersed in an aqueous dispersion of 700 nm diameter PMMA-co-
P4VP particles. In all experiments, the bubbles were generated at 28.8oC,
pH=14, PCO2/N2=34.5kPa, QL=10.5mL/h, and Cp=1.5wt. %. Optical microscope
images of collected bubbles (F),(G) from (D),(E), respectively. Scale bars: 200
μm………………………………………………………………………….…..65
Figure 4.5 Variation in contact angle (θ) of an aqueous solution of NaOH in air on the
PS-co-PAA film, measured at different pH values..........................................66
Figure 4.6 (A) Effect of the flow rate, QL, of the continuous phase on the fractional
volume change (ΔV/V0) of CO2 plugs and the final diameter Df of armoured
bubbles in the microchannel. The produced armoured bubbles at QL= 8.5 (B)
and 10.5 mL/h (C). Scale bars are 50 μm. The insets show the initial plugs of
CO2 bubbles. Scale bars are 200 μm. Bubbles were generated at pH=14,
PCO2=34.5 kPa and Cp=1.5 wt. %....................................................................67
Figure 4.7 (A) Effect of the initial pH of the continuous phase on the fractional volume
reduction (ΔV/V0) of CO2 plugs and the final diameter, Df, of armoured bubbles
in the microchannel. The produced armoured bubbles at initial pH=14 (B) 8 (C)
and 5 (D). Scale bars are 100 μm. The insets show the bubbles before the exit of
the microchannel. Scale bars are 200 μm. Bubbles were generated at PCO2=34.5
kPa, QL=9.0 mL/h and Cp=1.5 wt. %..............................................................68
Figure 4.8 (A) Effect of the initial concentration of particles, Cp, of the continuous phase
on the fractional volume reduction, ΔV/V0, of CO2 plugs and the final diameter,
Df, of armoured bubbles in the microchannel. The produced armoured bubbles
at initial Cp =0.1 (B) 1.5 (C) and 5 wt. % (D). Scale bars are 100 μm. The insets
show the bubbles before the exit of the microchannel. Scale bars are 200 μm.
Bubbles were generated at PCO2=34.5 kPa, QL=10.0 mL/h and pH=14…….70
Figure 4.9 Generation of bubbles with various types of colloidal armour. (A) Optical
fluorescence microscopy image of bubbles coated with a shell of 2.8 μm
diameter HY-labelled PS-co-PAA particles (Cp=0.5 wt. %) and 3.5 μm
XVIII
diameter PS-co-PAA particles (Cp=1.0 wt. %), PCO2=41.4 kPa, QL=12 mL/h,
pH=14. Scale bar is 50 μm. (B) Optical microscopy image of bubbles coated
with 3 μm diameter carboxylated silica particles. PCO2=82.7 kPa, QL =24 mL/h,
pH=14, and Cp=1.5 wt. %. Scale bar is 100 μm. The inset shows a fluorescence
microscopy image of the surface of an armoured bubble coated with 3 μm
diameter carboxylated silica particles and 2.8 μm diameter HY-labelled PS-co-
PAA polymer particles in the continuous phase in the weight ratio 3:1,
respectively. Scale bar is 25 μm. (C) LCFM image of bubbles encapsulated
with 20 nm diameter carboxylated silica nanoparticles loaded with CdSe/ZnS
QDs, λex=480 nm. PCO2=55.2 kPa, QL=23 mL/h, pH= 10, and Cp =0.12 wt. %.
Scale bar is 100 μm. (D) Fluorescence optical microscopy image of armoured
bubbles engulfed with the FITC–BSA shell, λex=495 nm. The bubbles were
generated at PCO2=44.8 kPa, QL=13 mL/, pH=7, and a protein concentration of
0.02 wt. %. Scale bar is 100 μm………………………………………………73
Figure 4.10 (A) Generation of droplets from the water-ethanol mixture (85/15, v/v)
containing 4 wt. % of poly (DVB-co-MAA) particles. The flow rates of the
droplet and continuous phases are 0.5 and 3.5 mL/h, respectively. Scale bar is
200 μm. (B) Collected particle-coated droplets at the exit of the microchannel.
Scale bar is 150 μm……………………………………………………………75
Figure 4.11 A diagram of the surface coverage of water-ethanol droplets with 3.5 μm-
diameter poly(DVB-co-MAA) particles. The dashed line shows the theoretical
conditions for the complete coverage of the droplets with a monolayer of 2D
hexagonally packed particles. Filled circles show the experimental data
points…………………………………………………………….……………76
Figure 4.12 Plug-shaped particle-covered droplets. Scale bar is 100 µm. Inset shows the
corresponding elliptical droplets flowing in the microchannel. Scale bar is 200
μm. Cp=14 wt. %. The flow rate of the droplet and continuous phases are 0.5
and 5.5 mL/h, respectively…………………………………………………..77
Figure 4.13 (A-C) Optical and (D-F) LCFM images of the water-ethanol droplets
armoured with a shell of poly(DVB-co-MAA) particles at the not-complete
(A,D) and complete (B,E) surface coverage, and at the excess of particles in
XIX
the droplet interior (C,F). In (C) an excess of particles appears as the large dark
region on the background of the droplet coated with a monolayer of particles. In
(D-F) poly(DVB-co-MAA) microbeads were labelled with anthryl methacrylate.
LCFM images show the plane located in the centre of the droplets. Cp =8 wt. %.
λex=380 nm. Scale bars in (A-E) and in (F) are 50 and 100 μm, respectively.
Scale bars in insets are 5 μm…………………………………………………78
Figure 4.14 SEM (A) and LCFM (B) images of poly(TPGDA) particles armoured with
poly(DVB-MAA) particles. Cp=14 wt. %. Scale bars are 40 μm in (A) and 60
μm in (B)……………………………………………………………………...79
Chapter 5 Bubbles Encapsulated with a Shell of Biopolymers
Figure 5.1 Schematic of co-adsorption of lysozyme and alginate on the surface of CO2
bubble during its dissolution…………………………………………………88
Figure 5.2 A) Schematic of a MF flow-focusing bubble generator, B) Representative
optical microscopy images of the generation (top) and dissolution of CO2
bubbles in a MF channel (bottom). The image is taken 250 mm away from the
orifice of the MF device. The width and height of the orifice are 50 and 120 μm,
respectively. PCO2=48.3 kPa, QL=6 mL/h. Clys=Calg=0.2 wt. %, pH=12. Scale
bar is 200 μm………………………………………………………………….89
Figure 5.3 A-C) Optical microscopy images of the bubbles at the exit of the microchannel
(inset shows the image of initial bubble at the orifice, scale bar: 100 μm) (A),
after 0.3(B), 2 (C), 24 (D) and 720 h (E) storage. The bubbles were formed at
PCO2=48.3 kPa and QL=6 mL/h. Clys=Calg=0.2 wt. %. Scale bars: 50 μm. D)
Change in the diameter of bubbles plotted as a function of time; the dashed line
separates on-chip (left) and off-chip (right) bubble storage…………………91
Figure 5.4 N2 bubbles after 24 h storage. Scale bar: 200 μm. N2 bubbles generated at
PN2=48.3 kPa, QL=6 mL/h, pH=12, Clys=Calg=0.2 wt. %...............................92
Figure 5.5 A) Effect of QL of the continuous aqueous phase on the initial (Di) and final
dimension (Df) of bubbles. PCO2=48.3 kPa, Clys=Calg=0.2 wt. %, B-D) Optical
microscopy images of the bubbles after 24 h (insets show the image of initial
XX
bubble at the orifice, scale bars: 100 μm) at QL=4.5 (B), 5.5 (C) and 6.5 mL/h
(D). Scale bars: 15 μm……………………………………………………….93
Figure 5.6 A) Effect of Calg on the initial and final dimensions of bubbles, PCO2=48.3 kPa,
QL=6 mL/h. Clys=0.2 wt. %, B-D) Optical microscopy images of the bubbles
after 24h (insets show the image of initial bubble at the orifice, scale bars: 100
μm) at Calg=0 (B, scale bar: 30 μm), 0.15 (C) and 0.2 wt. % (D). Scale bars: 15
μm……………………………………………………………………………94
Figure 5.7 A) Effect of Clys on the initial and final dimensions of bubbles, PCO2=48.3 kPa,
QL=6 mL/h. Calg=0.2 wt. %, B-D) Optical microscopy images of the bubbles
after 24h (insets show the image of initial bubble at the orifice, scale bars: 100
μm) at Clys=0 (B), 0.05 (C) and 0.15 wt. % (D). Scale bars: 15 μm…………95
Figure 5.8 The preparation of 7 μm-diameter bubbles produced at PCO2=72.4 kPa,
QL=10.5 mL/h. Clys=0.05 wt. % and Calg=2 wt. %. Optical microscopy image of
the bubbles after 24 (A). (inset: the initial CO2 bubble at the orifice. scale bar:
80 μm) and 720 h (B). Scale bars: 15 μm…………………………………….95
Figure 5.9 A) LCFM image of the bubble encapsulated with a lysozyme-alginate shell
and stored for 10 days (PCO2=48.3 kPa, QL=6 mL/h, Clys=Calg=0.2 wt. %). The
focal plane is located at 8 μm below the surface of the bubble. B) SEM images
of the bubble produced under the same conditions as in A). In A) and B) the
scale bar is 7 μm. C) A high magnification SEM image of the surface of the
bubble shown in (B). The scale bar is 2 μm. D) SEM image of the fractured
bubble surface. The bubble was stored for 28 days. The scale bar is 4 μm….97
Figure 5.10 Comparison of CD spectra of the native lysozyme solution (Clys=0.003
wt. %) (━) and the bubbles encapsulated with a lysozyme-alginate shells
(,). The bubbles were stored for 5 days () and 28 days ().
Conditions of bubble formations: PCO2=48.3 kPa, QL=5 mL/h, Clys=Calg=0.2
wt. %...............................................................................................................98
Figure 5.11 A) Polarization optical microscopy image of the bubbles coated with a
lysozyme-alginate shell. The bubbles were produced under PCO2=48.3 kPa,
QL=6 mL/h, Clys=Calg=0.2 wt. % and stored for 28 days. The scale bar is 50 μm.
B) Fluorescence intensity profile of 50 μM ThT solution (), freshly prepared
XXI
lysozyme (Clys=0.003 wt. %) dissolved in 50 μM ThT solution (), freshly
prepared alginate (Calg=0.003 wt. %) dissolved in 50 μM ThT solution () and
the dispersion of biopolymer-encapsulated bubbles in 50 μM ThT solution ().
The bubbles were stored for 28 days………………………………………...99
Chapter 6 A Single-step Microfluidic Route to Producing Multifunctional
Microbubbles
Figure 6.1 Schematics of the microfluidic (MF) generation of multifunctional bubbles.
(A) Schematic of a MF reactor. The height of the MF device is 120 µm. The
width of the orifice and the length of the downstream microchannel are 50 µm
and 260 mm, respectively. The top and the bottom insets show optical
microcopy images of the bubbles at the beginning and the end of the process,
respectively. Bubbles were generated at PCO2=48.3 kPa and QL=9.5 mL/h.
Scale bars in insets are 200 µm. (B) Schematic of the formation of NP
functionalized bubbles stabilized with a mixed lysozyme-alginate layer….106
Figure 6.2 Optical microscopy images of the bubbles functionalized with Fe3O4 NPs
after different storage times: (A) 3 sec, (B) 1 h and (C) 2000 h. The scale
bars in (A), (B) and (C) are 50, 15 and 15 µm, respectively. Inset in (A)
shows the bubble imaged immediately after its generation in the orifice of
the MF device. Scale bar is 100 µm. Bubbles were generated at PCO2=48.3
kPa and QL=9.5 mL/h………………………………………………………108
Figure 6.3 (A) Variations in the initial,Di, and final dimensions, Df, of microbubbles
are plotted as a function of the flow rate of the continuous phase, QL. (B-D)
Representative optical microscope images of the initial (insets) and final
dimension of bubbles coated with the biopolymers and Fe3O4 NPs at
different QL s (A) 6.5, (B) 7.5 and (C) 8.5 mL/h. Clysozyme=0.05 wt. %.
Calginate=0.1 wt. %, and CFe3O4 dispersion = 1 wt. %. PCO2=48.3 kPa. Scales
bars are 15 µm (100 µm, insets)…………………………………………109
Figure 6.4 Optical microscope images of (A) aggregated bubbles generated with
lysozyme and Fe3O4 NPs. Scale bar is 15 µm and (B) coalesced N2 bubble
generated with lysozyme, alginate and Fe3O4 NPs. Scale bar is 200 µm…110
XXII
Figure 6.5 Scanning transmission electron microscopy (STEM) images of the bubbles
coated with the lysozyme-alginate shell and (A) Fe3O4 NPs, (B) Au NPs, and
(C) SiO2-encapsulated CdSe/ZnS NPs, Scale bars are 6 µm. Insets in (A-C)
show corresponding high magnification images of the surface of the bubbles.
Scale bars in insets are 150 nm. (D-F) Energy dispersive X-ray (EDX)
spectrometry line scanning profiles for the system shown as a red line in (A),
(B) and (C). Bubbles were generated at PCO2 = 48.3 kPa and QL = 8.5
mL/h………………………………………………………………………111
Figure 6.6 (A-C) STEM images of the surface of bubbles coated with Fe3O4 NPs at
surface density of (A) 1.5x105, (B) 6.6x105 and (C) 1.5x106 NPs/µm2. Scale
bars are 300 nm. Insets show the corresponding bubbles. Scale bars are 3 µm.
Bubbles were generated at PCO2=48.3 kPa and and QL=8 mL/h……….112
Figure 6.7 Properties of NP-coated bubbles. (A) Magnetic actuation of bubbles
functionalized with Fe3O4 NPs. Scale bar is 50 µm. Bubbles were generated at
PCO2=48.3 kPa and and QL=8.5 mL/h. (B) Extinction spectra of Au NPs (red
spectrum, top) and of the bubbles coated with these Au NPs (blue spectrum,
bottom). Bubbles were generated at PCO2=48.3 kPa and and QL =8.5 mL/h.
(C) Confocal fluorescence microscopy image of bubbles carrying SiO2-
encapsulated CdSe/ZnS NPs. Scale bar is 30 µm. λex=364 nm. Inset shows an
image of the individual bubble. Scale bar is 5 µm. Bubbles were generated at
PCO2=48.3 kPa and and QL=9.5 mL/h…………………………………….113
Figure 6.8 In-vitro US imaging of the dispersion of dispersion of biopolymer-coated
bubbles at 90 % receive gain. The dispersion is placed in the Opticell chamber.
(A) NP-free bubbles, (B-D) bubbles coated with Fe3O4 NPs (B), Au NPs (C)
and SiO2-encapsulated CdSe/ZnS NPs (D). (E) US signal enhancement over
background for the systems shown in A-D. The concentration of bubbles in all
systems was 104 bubbles/mL……………………………………………….115
Figure 6.9 (A) In-vitro MRI images (top-view) of the dispersions of biopolymer-
encapsulated bubbles coated with different amounts of Fe3O4 NPs. The images
were obtained at 6.9 ms echo time. (B) Variation in T2* relaxation rate plotted
XXIII
as a function of surface density of Fe3O4 NPs on the surface of bubbles. The
concentration of bubbles in all systems was 104 bubbles/mL……………116
XXIV
Abbreviation
Alginate Alginic acid sodium salt BSA Bovine albumin serum CV Coefficient of variation DMPA 2,2-dimethoxy-2-phenyl-acetophenone DVB Divinylbenzene FITC Fluorescein isothiocyanate 4VP 4-vinylpyridine HMPP 2-hydroxy-2- methylpropiophenone HY Hostasol Yellow LCFM Laser confocal fluorescence microscopy Lysozyme Hen egg white lysozyme MF Microfluidic MAA Methacrylic acid MMA Methyl methacrylate MRI Magnetic resonance imaging NP Nanoparticle PAA Poly(acrylic acid) PDMS Poly(dimethylsiloxane) PEGDA Poly(ethylene glycol) diacrylate PS Poly(styrene) SEM Scanning electron microscopy STEM Scanning Transmission electron microscopy ThT Thioflavin-T TPGDA Tripropyleneglycol diacrylate TX-100 Triton X-100 US Ultrasound
1
Chapter 1
Introduction to Microfluidic Synthesis of Particulate Materials
1.1 Overview of Microfluidics
Microfluidic devices manipulate small (from 10–9 to 10–18 L) volumes of fluids, using
channels with dimensions from tens to hundreds of micrometers.1 Research in the field of
microfluidics covers physics, chemistry, engineering, materials science, and biology.2-6 In
particular, the last decade has witnessed rapid advancement in the area of materials
chemistry. Continuous microfluidic (MF) synthesis of colloidal materials offered a number
of advantages in comparison with conventional batch synthesis, such as a significant
reduction of reagent consumption, fast and controlled heat and mass transfer, and the
capability of performing complex parallel or consecutive reactions in a highly controlled
manner.7 Moreover, the capability to effectively screen chemical formulations and the
ability to produce products of chemical reactions with enhanced properties, e.g., a narrow
molecular mass distribution of polymer molecules or highly monodisperse colloidal
particles, made MF synthesis very beneficial in the development of new materials.8,9 The
dawbacks in MF preparation of materials include low productivity, the use of fast chemical
reactions and the difficulty in interfacing conventional equipment with MF devices. The
purpose of this chapter is to review recent progress in the MF synthesis of polymeric
particulate materials. This chapter focuses on the synthesis of particles conducted in the
continuous mode, that is, in the MF reactor, in contrast to numerous comprehensive papers
describing MF synthesis combined with the batch synthesis.7, 10-14
2
1.2 Different Modes of Microfluidic Synthesis of Polymer Particles
Figure 1.1 schematically shows a conventional batch reactor and a MF reactor.
Compared to their batch counterparts (Figure 1.1A), MF reactors (Figure 1.1B) offer
superior control over broad reaction conditions such as mixing, residence times and
reaction temperatures. These properties result from rapid heat and mass transfer, short
mixing times and well-defined temperature profiles, which are the consequences of small
reaction volumes in a flow-based reaction format. The flow-based concept can also be
incorporated with in-flow analysis. With respect to the production of particulate materials,
MF reactors provide a continuous-flow alternative to conventional batch synthesis for
preparing micrometer and nanometer-sized particles in a single or in multi-phase solution.
Figure 1.1 Synthesis of particulate materials (A) in a batch reactor and (B) in a flow-based, microfluidic (MF) reactor. Adapted with permission from reference 9. Copyright 2010, Annual Reviews.
3
MF synthesis of particulate materials can be realized in a single- or a multi-phase flow
as shown in Figure 1.2. Single-phase MF synthesis (Figure 1.2A) occurs either in a single
liquid or in a mixture of miscible liquids. The reagents required for the reaction can be
supplied to the reactor simultaneously, or added in the course of reaction.8, 15-17 In the first
case, the reaction is generally preceded by a mixing step, which rapidly and uniformly
distributes reactant molecules within the microchannel cross-section. Since the reactant
flow is considered to be of laminar nature, mixing across the channel occurs either by
diffusion, or convection. The synthetic process can be triggered in a particular location of
the reactor, e.g., by applying a particular temperature profile along the microchannel length,
by radiative heating, or by diffusion of reactants to the area of the reaction. The particles
are generated in the continuous phase and are carried towards the outlet of the MF reactor.
Figure 1.2 Schematic illustrations of (A) single-phase and (B) multiphase formats for the MF synthesis of polymer particles synthesized (C) in droplets (i), and (D) at the interface between the two phases (ii). Adapted with permission from reference 9. Copyright 2010, Annual Reviews.
4
Multi-phase MF synthesis of polymeric particles starts by forming droplets of a gas or
a liquid which serve to induce convective mixing, to confine the synthesis to miniature
compartments or even to define particle size and shape. An analogy to chemical synthesis
can be drawn by applying individual reaction compartments to their path through the
microreactor. The multiphase flows of interest here are sometimes referred to as segmented
flows and either consists of continuously alternating liquid droplets and liquid segments, or
of liquid segments and gas bubbles. Regular and stable multiphase flows can be obtained
by varying reactant flow rates, viscosities and microchannel sizes/configurations.18,19
Generally, polymeric particles can be synthesized in one of two locations schematically
shown in Figure 1.2B: i) inside droplets (Figure 1.2C),20-22 or ii) at the interface between
the phases (Figure 1.2D).23,24
Compared to a single-phase approach, droplet-based synthesis is characterized by rapid
heat and mass transfer, and by efficient mixing. The effect of axial dispersion is either
eliminated or minimized.18, 25 Generally, droplets are generated in one of four types of MF
droplet generators, namely, terrace-like devices,26, 27 T-junctions or Y-junctions,28, 29 flow-
focusing droplet or bubble generators,30-32 and capillary-based co-flow devices.33, 34
The size of droplets can be controlled by tuning the dimensions of microchannels,
macroscopic properties of liquids and the flow rates of liquids. Regardless of the geometry
of MF reactors, the control of wetting of microchannel with liquids is critical in the
generation of droplets: the continuous fluid should wet the channel walls preferentially over
the dispersed phase.35, 36
5
Both single- and multi-phase MF syntheses of polymer particles provide the following
advantages: i) the ability to control particle dimensions (and hence the corresponding size-
dependent properties of the particles), ii) the ability to generate particles with a variety of
shapes that cannot be implemented in conventional (non-microfluidic) synthesis, iii) the
capability to control particle morphologies by the hydrodynamic means, iv) the ability to
control particle morphologies by the uniform supply of energy to precursor droplets and v)
a simple production of composite particles.
1.3 Microfluidic Synthesis of Polymer Particles
Polymer particles with narrow size distributions find a broad range of applications in
chromatography as ion exchange resins, toners, spacers, and calibration standards.37-39
Advanced applications of monodisperse polymer particles include optical data storage,40,41
biometrics,42,43 stabilization of Pickering systems,44 production of dielectric resonators,45,46
encapsulation of drugs and biological molecules,47 medical diagnostics48 and photonics.49
Control of particle compositions, dimensions, and internal structure is highly desirable and
sometimes it is vital in realizing a particular application.
Conventional methods of particle synthesis, such as emulsion, dispersion, precipitation,
suspension and miniemulsion polymerizations either lack control of particle size,
morphology, and shapes, or require multi-step processes, or are material-specific.50-52
Sometimes, lack of control of particle size necessitates post-synthetic fractionation of
polymer beads, which is both time- and material-consuming.
6
MF synthesis of polymer particles enables control of particle size, shape, composition
and internal structure. Polymer particles have been synthesized in both a single-phase15,16
and a multi-phase system.24, 53 Single-phase MF generation of polymer particles utilized
UV-illumination onto the continuous stream of a polymerizable liquid through a patterned
mask.15 Another example of the single-phase MF production of polymeric particles
includes the nanoprecipitation of block copolymers, e.g., poly(lactic-co-glycolic acid)-b-
poly(ethylene glycol), which was induced by the controlled mixing between two miscible
fluids.16 Multi-phase MF generation of polymer particles relies on the (i) emulsification of
liquid monomers or polymers and (ii) the solidification or gelation of the precursor droplets
by the chemical or physical means.
Polymerization mechanisms for the production of particles include
polycondensation24,54 and radical polymerization.21,55 In addition to chemical methods,
ionic crosslinking,56,57 thermosetting,58 or solvent extraction and evaporation from the
droplets transform precursor droplets into polymer particles.59,60
The following sections describe the recent accomplishments in the MF production of i)
rigid polymer particles with narrow size distributions, ii) gel particles (microgels), iii)
particles with controlled internal structures, iv) particles with non-spherical shapes, and v)
particle with gaseous core. In addition, the productivity of MF synthesis of polymeric
particles is discussed in section 1.3.6.
1.3.1 Microfluidic Synthesis of Solid Polymer Microspheres
Historically, droplet-based MF emulsification of monomers and subsequent post-
polymerization of monomer droplets in a batch process was proposed in 2002.26 In 2005,
7
continuous on-chip production of monodisperse solid polymer particles with various
compositions was realized by emulsifying monomers mixed with a photoinitiator and
various organic or inorganic additives and solidifying the precursor droplets by photo-
initiated polymerization (Figure 1.3A).20,21 The diameter of particles generated by the MF
synthesis was in the range from tens to hundreds of micrometers. In contrast to
conventional methods, the narrow size distribution of the precursor droplets was preserved
by polymerizing “one particle at a time”, that is, when droplets moving through the
downstream microchannel were separated with a well-defined gap of the continuous phase.
Particle polydispersity (coefficient of variation, CV, defined by a standard deviation
divided by the diameter of the droplets or particles) was typically below 5-6% (Figure
1.3B).
Figure 1.3 Rigid polymer particles synthesized in continuous multiphase MF reactors. (A) SEM images of poly(tripropyleneglycol diacrylate) particles. (B) Size distribution of the particles shown in panel (A). d denotes particle diameter. (C) Optical fluorescence microscopy images of poly(tripropylene glycol diacrylate) microspheres loaded with CdSe quantum dots. (D) Optical polarization microscopy images of poly(tripropyleneglycol diacrylate) microspheres with 5–20 wt. % of liquid crystals made of 4-cyano-4’-pentylbiphenyl. Panels (A–D) are reproduced with permission from Ref. 20. Copyright 2005, Wiley-VCH. (E) SEM image of porous poly(glycidyl methacrylate–co-ethylene glycol dimethacrylate). Panel (E) is adapted with permission from Ref. 67. Copyright 2009, American Chemical Society. (F) Confocal microscopy image of poly(tripropyleneglycol diacrylate)/poly(urethane) particles with an interpenetrating network (IPN) structure. Panel (F) is adapted with permission from Ref. 68. Copyright 2008, American Chemical Society.
8
MF synthesis allowed efficient production of rigid polymer microspheres with various
compositions by emulsifying mixtures of several monomers or mixtures of monomers with
inorganic nanoparticles, small molecules, or solvents and by conducting subsequent
continuous solidification of the precursor droplets.20,61-63 In this method, the requirements
to the liquid-to-be-emulsified included: (i) the absence of precipitation in the inlet tubing
and (ii) the appropriate values of viscosity and interfacial tension with the continuous phase.
Examples of composite rigid polymer particles produced by the MF synthesis included
copolymer particles,61,62 and polymer particles with inorganic nanoparticles20,63,64 and liquid
crystals.20,65 Porous polymer particles were prepared by mixing a monomer with a porogen
solvent, polymerizing the monomer and removing the solvent.20,66,67 Several examples of
composite rigid polymer particles are presented in Figure 1.3 (C-E).
Polymer particles with an interpenetrating polymer network (IPN) structure (Figure
1.3F) were synthesized by performing multi-step reactions in the precursor droplets: heat
generated in the exothermic photo-initiated free radical polymerization of the diacrylate
monomer triggered the polycondensation of the poly(urethane) oligomer with
diethanolamine.68
1.3.2 Microfluidic Synthesis of Spherical Polymer Microgels
Polymer microgel particles produced in continuous MF reactors have attracted a lot of
interest owing to the ability to encapsulate in their interior biological molecules and
cells.56,57,69,70 In chemically induced gelation, the MF emulsification of aqueous solutions of
e.g., acrylamides or poly(ethylene glycol) diacrylate was followed by the rapid on-chip
photoinitiated polymerization or crosslinking reactions, which yielded microgel
9
particles.55,71-73 Figure 1.4A shows the microgels of poly(N-isopropylacrylamide)
synthesized by redox polymerization.55 The proposed approach was based on generating
core-shell droplets composed of the aqueous core containing the redox initiator and the
aqueous shell containing a mixture of monomers with the reaction accelerator. Crossliking
of the microgels was ensured by complete mixing between the two aqueous phases.
Photoinitiated MF synthesis also allowed for the continuous production of fluorescent and
magnetic microgels.72,73
Figure 1.4 Microgel particles synthesized in continuous multiphase MF reactors. (A) Optical microscopy images of the formation of droplets of poly(N-isopropylacrylamide) microgel particles (left) and the resulting microgel particles labeled with fluorescein isothiocyanate (right). The inset shows a schematic of the microfluidic reactor (OF, outer fluid; MF, middle fluid). Reproduced with permission from reference 55. Copyright 2007, Wiley-VCH. (B) Optical microscopy images of ionically cross-linked alginate (left), k-carrageenan (middle), and carboxymethylcellulose (right) microgels. Scale bar: 100 μm. Adapted with permission from reference 56. Copyright 2006, American Chemical Society.
Physically-induced gelation was used to form microgels from biopolymers.74 The
methods that were used to gel precursor droplets included thermosetting (of e.g., agarose)20
or ionic crosslinking (of e.g., alginate, κ-carrageenan, or carboxymethyl cellulose).56,69,75
Since ionic gelation led to the rapid build-up in the viscosity of the droplet phase,
10
emulsification of the gelling solution was problematic.74 One of the ways to circumvent this
problem has utilized the diffusion of the cross-linking agent, e.g., Ca+2 or Fe+3 ions, from
the continuous phase into the precursor droplets. Examples of particles generated by
diffusion-controlled ionic crosslinking are shown in Figure 1.4B.56,75 Alternatively,
ionically-mediated gelation was achieved by generating precursor droplets
compartmentalizing a crosslinking agent in an inactive state and a polymer, and extracting
from the continuous phase a reagent that triggered gelation of the polymer in the droplets .
For example, droplets containing sodium alginate and CaCO3 underwent gelation due to
diffusion of acetic acid from the continuous phase into the droplets, thereby reducing pH
and liberating Ca+2 ions which crosslinked the alginate molecules.69,75 In another method,
coalescence of precursor droplets of polymer solution and droplets containing a
crosslinking agent also yielded alginate microgels.57
1.3.3 Microfluidic Synthesis of Polymer Particles with Different
Morphologies
Microfluidic synthesis can be used to generate particles with a variety of internal
structures that can be controlled by the hydrodynamic means. One of the examples is the
synthesis of Janus particles, that is, particles with two compartments having distinct
compositions. These particles are produced by generating precursor droplets from two
parallel co-flowing streams of two immiscible monomers and performing subsequent
polymerization.76,77 The volume fraction of each compartment of the particles can be
precisely controlled by varying the flow rate ratio of the constituent monomers.76-78
11
Furthermore, since at low Reynolds number (Re<1) convective transport across the
interface between the two adjacent streams is suppressed, Janus droplets can be formed
from the miscible solutions of the same monomer, which contain distinct additives, e.g.,
colloid particles labeled with different dyes.72,79 Rapid photopolymerization of the
monomer results in the formation of Janus particles.72,76,77 Figure 1.5A illustrates the
formation of Janus microgel particles from the aqueous solutions of acrylamide containing
silica particles labeled with rhodamine isothiocyanate (red color) and fluorescein
isothiocyanate (green color).72
Figure 1.5 MF synthesis of polymer particles with different morphologies. (A) Schematic of the microfluidic synthesis of Janus microgels. The inset shows an optical fluorescence image of poly(acrylamide) particles encapsulating silica particles labeled with two different dyes in distinct compartments. Adapted with permission from reference 72. Copyright 2006, American Chemical Society. (B) Optical microscopy image of bicolored poly(isobornyl acrylate) particles containing carbon black and titanium oxide in two distinct hemispheres. The inset shows electrically actuated particles displaying their white-colored side in the display panel. Adapted with permission from reference 79. Copyright 2006, Wiley-VCH. (C) SEM images of microcapsules prepared by microfluidic interfacial polymerization. The inset shows a high magnification image of the interior of the capsular wall. Adapted with permission from reference 54. Copyright 2005, American Chemical Society. (D) Optical microscopy image of the colloidal crystalline array confined in polymeric shells. The inset shows the particles with three colloidal crystal cores. Scale bars: 200 μm. Adapted with permission from reference 99. Copyright 2008, American Chemical Society.
12
The surface of each compartment of Janus particles can be selectively functionalized
with e.g., biological molecules.76 Alternatively, the constituent phases of the Janus particles
generated by MF synthesis can be loaded with two distinct species.79 For example,
poly(isobornyl acrylate) Janus particles were synthesized with two halves loaded with
carbon black and titanium oxide. In an alternating electric field, the particles were rotated
thereby giving the ability to change the color (black or white) of the display panel. These
experimental results showed potential application of such particles as a constituent of
electronic paper (Figure 1.5B).80
MF synthesis and assembly also proved to be useful in the generation of polymer
capsules. Polymer capsules find applications in cosmetics, nutrition, agricultural,
pharmaceutical and food industries and in biomedical engineering.81,82 Polymer particles
with a core-shell structure (capsules) have been produced by a number of MF methods that
are described below.
One of the most general approaches to polymer capsules is the formation of double
emulsions, e.g., water-in-oil-in-water or oil-in-water-in-oil droplets and subsequent
solidification of the shell of these droplets to yield core-shell particles.33,83,84 Double
emulsions have been produced in MF devices with different geometries such as a two
consecutive T-junctions or flow-focusing devices,83,84 a single flow-focusing device,85 a
combination of flow-focusing and T-junction devices,86 and a capillary-based device.33,55
Polymerization of the shell of these droplets led to the formation of capsules with rigid or
soft polymer shell.33,55,85 Alternatively, core-shell droplets containing a solution of diblock
copolymer or colloidal particles in the shell transformed into capsules when the solvent was
evaporated from the shell.87,88 Manipulation of the hydrodynamic conditions during the
formation of double emulsions, namely, the variation of the flow rates of the liquids
13
allowed control of the size of particles, shell thicknesses and the number of cores per
capsule.85,87,89
Another strategy used for the generation of capsules relies on reactions occurring at the
interface between the dispersed and the continuous phases.24,54,90 For example, interfacial
polymerization between aqueous 1,6-diaminohexane (a droplet phase) and adipoyl chloride
dissolved in a mixture of solvents (a continuous phase) yielded monodisperse capsules with
a nylon-6,6 membrane.24 Interfacial polymerization was also used to fabricate polyamide
capsules with a fibrous shell structure by emulsifying a mixture of acid chlorides in the
aqueous solution of polyethyleneimine and carrying out an interfacial condensation reaction
(Figure 1.5C).54
Diffusion-controlled reactions have also generated polymer capsules.56,91 For example,
diffusion of Ca2+ ions from the non-polar continuous phase into the aqueous droplets of
sodium alginate resulted in the limited crosslinking of the aliginate molecules and the
formation of microgel capsules.56 The thickness of the gelled shell depended on the time of
residence of the microgels in the microchannels and the concentration of the crosslinking
agent in the continuous phase. Diffusion of photo-activated initiators from the continuous
oil phase into the aqueous droplets of poly(N-isopropylacrylamide) also resulted in the
formation of gel capsules with a thin shell membrane.91
Finally, capsules have been generated by assembling particles or depositing polymeric
molecules at the interface between the dispersed phase and the continuous phase.92-96 For
example, the close-packed assembly of polymer particles at the surface of droplets resulted
in the formation of Pickering emulsions that were resistant to coalescence and Ostwald
Ripening.92,94 Sequential adsorption of poly(methacrylic acid) and poly(N-
14
vinylpyrrolidone) to the surface of liquid crystal of 4’-pentyl-4-cyanobiphenyl as droplets
produced capsules with the shell comprising alternating polymer layers.96
Another group of particles with interesting, non-conventional morphologies included
polymer microspheres with a supracolloidal structure.97-99 These particles were formed by
generating core-shell droplets compartmentalizing colloidal beads in the aqueous core and
conducting continuous polymerization of the shell composed of a photo-curable monomer.
The resulting composite microspheres are shown in Figure 1.5D. Due to the confinement-
induced crystallization of the microbeads in the particle core, the composite capsules
exhibited distinct diffraction patterns in the visible range that were independent of the
orientation of the spherical surface.
1.3.4 Microfluidic Synthesis of Non-spherical Polymer Particles
Many interesting applications of polymer particles are governed by their shapes.100,101
Single- and multi-phase MF synthesis allows the preparation of particles with non-spherical
shapes that would otherwise be difficult to realize in conventional methods such as
emulsion or suspension polymerizations. This is due to the minimization in interfacial
energy which makes droplets acquire a spherical shape in the conventional methods.
One of the ways to tune the shapes of particles generated by the multi-phase MF
synthesis is the polymerization of droplets under confinement.20,21,72,73,101 The shape of
particles is determined by the relationship between the diameter, d, of unperturbed droplets
(d=(6V/π)1/3 where V is the volume of the droplet) and the dimensions of microchannels.
For example, when w>d and h<d where w and h are the width and the height of the
microchannel, respectively, droplets acquire a discoid shape (Figure 1.6A). On the other
15
hand, when w<d and h<d, the droplets assume a rod-like shape (Figure 1.6B).102,103
Subsequent continuous polymerization yields disk-shaped and rod-shaped polymer
particles.20,21,72,103 Geometrical confinement combined with controlled coalescence between
aqueous droplets of sodium alginate and CaCl2 yielded spherical, discoid, rod-like and
thread-like alginate gel particles.104
Figure 1.6 MF synthesis of particles with non-spherical shapes. (A,B) SEM images of poly(tripropyleneglycol diacrylate) particles with (A) disk and (B) rod shapes. Insets show the relationship between the particle shape and the channel dimensions. Adapted with permission from references 20 and 103. Copyright 2005, Wiley-VCH. Copyright 2005, American Chemical Society. (C) SEM image of poly(tripropyleneglycol diacrylate) particles with bowl-like shapes. The inset shows the precursor droplet with a small fraction of nonpolymerizable liquid (indicated by the arrow). Scale bar: 40 μm. Adapted with permission from reference 85. Copyright 2005, American Chemical Society. (D–F) SEM images of poly(ethylene glycol diacrylate) particles with various shapes synthesized in one-phase MF synthesis using continuous-flow lithography. Insets show the corresponding photomasks. Scale bars: 10 μm. Adapted with permission from reference 15. Copyright 2006, Macmillan.
In another MF approach, semi-spherical and bowl-like polymer particles were
generated from the pre-cursor multi-phase droplets containing a monomer and a non-
polymerizable liquid (Figure 1.6C).78,85 The droplets were formed by breaking a coaxial jet
of silicone oil and a photo-curable monomer to form multi-phase droplets with a range of
16
morphologies. Following the polymerization of the monomer in the droplets, the oil phase
was removed. Since the volume fraction of each liquid was determined by its relative flow
rate, the morphologies of the precursor multi-phase droplets and the corresponding particles
were accurately described by a phase-like diagram.85
Toroidal particles were generated by generating droplets compartmentalizing
solutions of different polymers (e.g. poly(ether sulfone), polysulfone, or poly(methyl
methacrylate)) and removing the solvent from the droplets.105 The diffusion of the solvent
proceeded at spatially different rates across the surface of droplets. The difference in the
solvent diffusion rate led to a non-uniform solidification between the circumferential and
central regions of the droplets and resulted in the formation of the donut-shaped toroidal
particles.
Single-phase MF synthesis of polymer particles employed continuous flow projection
photolithography.15 Site-specific photomask-limited UV-irradiation of the poly(ethylene
glycol diacrylate) pre-polymer solution flowing through a microchannel led to the site-
specific photo-initiated polymerization, thereby creating polymer particles. The shapes of
the particles (e.g., triangles, rings, cubes and polygons) were precisely determined by the
features of the photomasks (Figure 1.6D-F). A similar approach has been applied to
synthesize long flattened particles encoded with multiple probes for molecular
identification.106 Amphiphilic non-spherical microparticles were also created using
multiphase stratified flows composed of a hydrophilic monomer and a hydrophobic
monomer stream.107 The particles showed the ability to self-assemble in water or at the oil-
water interface. Furthermore, three-dimensional and multifunctional particles were
synthesized by controlling fluid flows, channel topologies and intensity of UV
irradiation.108-110
17
1.3.5. Microfluidic Synthesis of Bubble-templated Polymer Particles
Micrometer-size bubbles have numerous biomedical applications such as ultrasound
contrast agents,111,112 targeted drug delivery vehicles,113,114 and tumor/thrombus demolition
materials.115-116 Bubbles are also used in the preparation of various porous materials, which
have a range of applications including the pharmaceutical, food and the cosmetic
industries.117 Currently, it is a challenge to generate and store bubbles with high uniformity
in size and long-term usability. Conventional methods of preparation of bubbles involve
sonication, mechanical agitation, or shear mixing of liquids containing amphiphilic
molecules like lipids, proteins, or surfactants.118-120 Bubbles produced by these methods
tend to have a broad distribution of sizes. In addition, bubbles stabilized by amphiphilic
molecules are prone to dissolution and coarsening via Ostwald ripening: excess Laplace
pressure across the air-water interface hampers the long-term stability of bubbles.121
MF approaches enable the production of monodisperse bubbles with highly controlled
dimensions.31,122,123 The polydispersity of these bubbles is typically less than 5%, however,
the long-term stability of bubbles remains a problem. Recently, it has been shown that the
stability of bubbles can be significantly enhanced by forming close-packed layers of
colloidal particles at the gas-water interface124,125 or by choosing appropriate
surfactants.126,127 MF production of particle-stabilized bubbles was obtained by
hydrodynamic shear flows92 and by chemical means.128 A mixture of lipids was chosen as a
shell material to improve the stability of bubbles.127 Depending on the type of materials
dispersed in the oil phase, gas-in-oil-in-water emulsions generated using a MF platform
could be used as templates for bubbles with a particle shell and drug-encapsulated
bubbles.129,130 For example, hydrophobic silica or magnetic nanoparticles were dispersed in
18
a volatile organic solvent. Upon the evaporation of solvent, the nanoparticles formed a stiff
multilayer shell at the gas-water interface, which protected the bubbles against dissolution
and coarsening.129 Consecutive adsorption of protein and polysaccharide layers on the
surface of bubbles produced using a MF device created monodisperse bubbles with long-
term stability.131 Furthermore, bubbles with a hydrogel shell were produced by
photopolymerizing gas-in-water-soluble monomer-in-oil emulsions.86
1.3.6 Productivity of Continuous Microfluidic Reactors
Although continuous MF synthesis of polymer particles is an efficient way for the
production of high value particles, future application of this method depends on the ability
to increase the productivity of MF synthesis. Typically, the productivity of a single planar
MF reactor does not exceed several mL/h.132 The simplest way to overcome this limitation
is to increase the number of reactors working in parallel in the approach called “numbering
out”.133 To avoid high cost and practical complications associated with linking many tubes,
connectors, and pumps, it is imperative to supply liquids in multiple parallel reactors using
only a few pumps.134
The integration of multiple droplet generators in a single chip causes complex fluid
behaviors, which originate from the feedback between adjacent individual generators
sharing the same liquid supply.135 Another problem originates from the lack of fabricating
multiple MF reactors with identical geometries.136 Consequently, the size distribution of
droplets broadens and polydispersity of the resulting particles increases.
The fabrication of a MF droplet generator using a silicon membrane containing a few
hundred thousands microchannels increases the efficiency of MF emulsification.137 A high
19
throughput of 60–70 mL/h was achieved; however, it was not as simple as in a common
MF droplet generator to control the size of droplets and create droplets with different
shapes and morphologies. A different type of MF reactor was implemented in a monolithic
module combining up to 256 cross-junction droplet generators.79,134 The device generated
monodisperse droplets (CV<4%) with the productivity of up to 320 mL/h. The generation
of Janus droplet was also feasible in the same type of reactor. The production of polymer
particles in this device was achieved by photopolymerization in the drain tubing or by
thermally initiated polymerization in a batch reactor.
Although the productivity of droplet and particle formation greatly increased by using
the above-mentioned approaches, the synthesis lacked the ability to control the dimensions
of the droplets137 thereby necessitating post-fractionation of the microbeads.79,134
Furthermore, device failures, e.g. clogging of a particular microchannel, could result in the
re-fabrication of the entire MF reactor.
Recently, a multiple modular microfluidic (M3) reactor for the synthesis of polymer
particles was proposed, which contained 8 modules accommodating 128 parallel
reactors.136 Several important features were implemented in the M3 reactor. First, one could
easily replace one of the fouled modules with the new one if the component failure took
place. Second, cross-talk between adjacent reactors was suppressed by elongating the
hydrodynamic path for the droplet phase prior to its entrance to the droplet generator. These
features allowed high throughput in droplet and particle production of up to 50 mL/h and
the generation of particles with polydispersity not exceeding 5 %. Third, the integration of a
polymerization compartment into each module allowed in situ particle synthesis, thereby
suppressing coalescence and/or Ostwald ripening of the precursor droplets.
20
1.4 Outlook
Future MF reactors have the potential to differentiate themselves from conventional
batch reactors through a number of beneficial features. In order to accomplish this goal, MF
platforms will need to become robust, well-integrated, scalable, stand-alone units (ideally,
without external pumps) and will require virtually no infrastructure investment such as
fume hoods, or glove boxes. These systems will be modular, computer-controlled and will
be interfaced with standard laboratory equipment and preparatory procedures. Currently,
continuous MF synthesis of a variety of colloidal particles has been successfully
demonstrated at a proof-of concept stage. The next major directions in the development of
MF synthesis of particulate materials include the followings.
i) MF synthesis produces high-value particles. Although the properties and “proof-of-
concept” applications of these particles have been demonstrated, to the best of our
knowledge, no application has been realized beyond the lab bench. Future work should be
directed towards the synthesis of materials with truly unique features that are critically
important in specific applications. Search for such applications is vital for biological
research. Currently, the encapsulation of cells appears to be the most promising application
of the particles generated by MF processes.
ii) A lot of work should be done to make MF synthesis cost-effective, robust and
reliable. For example, MF reactors have to be fabricated from materials that can
accommodate multi-hour MF synthesis. An appropriate selection of reactor material is one
of the most significant factors. Along with chemical stability, surface properties and ease of
fabrication of the device have to be taken in account. Other considerations are pertinent to a
particular mechanism of particle synthesis. For example, photoinitiated MF synthesis of
21
polymer particles requires transparency of the rector material in the wavelength range used
for photoinitiation.
iii) MF synthesis allows the high throughput screening of formulations used for the
synthesis of polymer colloids, which is achieved by varying the flow rates of the liquid
reagents supplied to the reactor. More work is needed in this field in order to fully realize
the exploratory side of continuous MF synthesis of particulate materials.
iv) The location and time of chemical reactions can be defined in MF reactor due to
time-to-distance transformation characteristics of MF synthesis. This feature enables multi-
step reactions to be conducted in a highly controlled manner. Such reactions triggered “on
demand” can benefit the synthesis of colloidal particles.
v) MF synthesis will greatly benefit from in-line chemical and physical
characterization of reaction products. Since microliter volumes are often sufficient for
analysis, MF synthesis is particularly advantageous for rapidly screening different reaction
conditions. As a result, the variables of reactions can be much more rapidly explored.
References
(1) G. M. Whitesides, Nature 2006, 442, 368.
(2) P. Garstecki, M. J. Fuerstman, G. M. Whitesides, Nature Physics 2005, 1, 168.
(3) H. Song, D. L. Chen, R. F. Ismagilov, Angewandte Chemie-International Edition
2006, 45, 7336.
(4) M. Prakash, N. Gershenfeld, Science 2007, 315, 832.
(5) S. E. Chung, W. Park, S. Shin, S. A. Lee, S. Kwon, Nature Materials 2008, 7, 581.
(6) J. D. Wan, W. D. Ristenpart, H. A. Stone, Proceedings of the National Academy of
Sciences of the United States of America 2008, 105, 16432.
22
(7) J. L. Steinbacher, D. T. McQuade, Journal of Polymer Science Part a-Polymer
Chemistry 2006, 44, 6505.
(8) Y. J. Song, J. Hormes, C. Kumar, Small 2008, 4, 698.
(9) J. I. Park, A. Saffari, S. Kumar, A. Gunther, E. Kumachev, Annual Review of
Materials Research 2010, 40.
(10) C. H. Chen, A. R. Abate, D. Y. Lee, E. M. Terentjev, D. A. Weitz, Advanced
Materials 2009, 21, 3201.
(11) J. H. Gao, H. W. Gu, B. Xu, Accounts of Chemical Research 2009, 42, 1097.
(12) Y. Yin, A. P. Alivisatos, Nature 2005, 437, 664.
(13) M. A. Lopez-Quintela, Current Opinion in Colloid & Interface Science 2003, 8, 137.
(14) Y. W. Jun, J. S. Choi, J. Cheon, Angewandte Chemie-International Edition 2006, 45,
3414.
(15) D. Dendukuri, D. C. Pregibon, J. Collins, T. A. Hatton, P. S. Doyle, Nature
Materials 2006, 5, 365.
(16) R. Karnik, F. Gu, P. Basto, C. Cannizzaro, L. Dean, W. Kyei-Manu, R. Langer, O.
C. Farokhzad, Nano Letters 2008, 8, 2906.
(17) T. L. Sounart, P. A. Safier, J. A. Voigt, J. Hoyt, D. R. Tallant, C. M. Matzke, T. A.
Michalske, Lab on a Chip 2007, 7, 908.
(18) A. Gunther, K. F. Jensen, Lab on a Chip 2006, 6, 1487.
(19) J. Atencia, D. J. Beebe, Nature 2005, 437, 648.
(20) S. Q. Xu, Z. H. Nie, M. Seo, P. Lewis, E. Kumacheva, H. A. Stone, P. Garstecki, D.
B. Weibel, I. Gitlin, G. M. Whitesides, Angewandte Chemie-International Edition 2005, 44,
724.
(21) D. Dendukuri, K. Tsoi, T. A. Hatton, P. S. Doyle, Langmuir 2005, 21, 2113.
(22) I. Lee, Y. Yoo, Z. Cheng, H. K. Jeong, Advanced Functional Materials 2008, 18,
4014.
(23) E. Stride, K. Pancholi, M. J. Edirisinghe, S. Samarasinghe, Journal of the Royal
Society Interface 2008, 5, 807.
(24) S. Takeuchi, P. Garstecki, D. B. Weibel, G. M. Whitesides, Advanced Materials
2005, 17, 1067.
23
(25) L. Shui, J. C. T. Eijkel, A. van den Berg, Advances in Colloid and Interface Science
2007, 133, 35.
(26) S. Sugiura, M. Nakajima, M. Seki, Industrial & Engineering Chemistry Research
2002, 41, 4043.
(27) S. Sugiura, M. Nakajima, M. Seki, Journal of the American Oil Chemists Society
2002, 79, 515.
(28) T. Thorsen, R. W. Roberts, F. H. Arnold, S. R. Quake, Physical Review Letters
2001, 86, 4163.
(29) T. Nisisako, T. Torii, T. Higuchi, Lab on a Chip 2002, 2, 24.
(30) S. L. Anna, N. Bontoux, H. A. Stone, Applied Physics Letters 2003, 82, 364.
(31) P. Garstecki, I. Gitlin, W. DiLuzio, G. M. Whitesides, E. Kumacheva, H. A. Stone,
Applied Physics Letters 2004, 85, 2649.
(32) A. M. Ganan-Calvo, J. M. Gordillo, Physical Review Letters 2001, 87.
(33) A. S. Utada, E. Lorenceau, D. R. Link, P. D. Kaplan, H. A. Stone, D. A. Weitz,
Science 2005, 308, 537.
(34) C. Cramer, P. Fischer, E. J. Windhab, Chemical Engineering Science 2004, 59,
3045.
(35) J. D. Tice, H. Song, A. D. Lyon, R. F. Ismagilov, Langmuir 2003, 19, 9127.
(36) W. Li, Z. H. Nie, H. Zhang, C. Paquet, M. Seo, P. Garstecki, E. Kumacheva,
Langmuir 2007, 23, 8010.
(37) M. P. Merkel, V. L. Dimonie, M. S. Elaasser, J. W. Vanderhoff, Journal of Polymer
Science Part a-Polymer Chemistry 1987, 25, 1219.
(38) T. Sugimoto, Advances in Colloid and Interface Science 1987, 28, 65.
(39) J. Ugelstad, A. Berge, T. Ellingsen, R. Schmid, T. N. Nilsen, P. C. Mork, P.
Stenstad, E. Hornes, O. Olsvik, Progress in Polymer Science 1992, 17, 87.
(40) E. Kumacheva, O. Kalinina, L. Lilge, Advanced Materials 1999, 11, 231.
(41) B. J. Siwick, O. Kalinina, E. Kumacheva, R. J. D. Miller, J. Noolandi, Journal of
Applied Physics 2001, 90, 5328.
(42) H. H. Pham, I. Gourevich, J. K. Oh, J. E. N. Jonkman, E. Kumacheva, Advanced
Materials 2004, 16, 516.
24
(43) H. H. Pham, I. Gourevich, J. E. N. Jonkman, E. Kumacheva, Journal of Materials
Chemistry 2007, 17, 523.
(44) B. Liu, W. Wei, X. Z. Qu, Z. H. Yang, Angewandte Chemie-International Edition
2008, 47, 3973.
(45) I. Gourevich, L. M. Field, Z. X. Wei, C. Paquet, A. Petukhova, A. Alteheld, E.
Kumacheva, J. J. Saarinen, J. E. Sipe, Macromolecules 2006, 39, 1449.
(46) Z. X. Wei, L. Gourevich, L. Field, N. Coombs, E. Kumacheva, Macromolecules
2006, 39, 2441.
(47) M. D. Bhavsar, M. M. Amiji, Expert Opinion on Drug Delivery 2007, 4, 197.
(48) S. Derveaux, B. G. Stubbe, K. Braeckmans, C. Roelant, K. Sato, J. Demeester, S. C.
De Smedt, Analytical and Bioanalytical Chemistry 2008, 391, 2453.
(49) X. L. Xu, S. A. Asher, Journal of the American Chemical Society 2004, 126, 7940.
(50) A. Azioune, E. Bourgeat-Lami, R. Soula, P. Ni, S. Bousalem, C. Burguiere, A.
Elaissari, Colloidal Polymers: synthesis and characterization, Marcel Dekker Ltd New
York 2003.
(51) M. Antonietti, K. Landfester, Progress in Polymer Science 2002, 27, 689.
(52) S. Omi, G. H. Ma, M. Nagai, Macromolecular Symposia 2000, 151, 319.
(53) C. Serra, N. Berton, M. Bouquey, L. Prat, G. Hadziioannou, Langmuir 2007, 23,
7745.
(54) E. Quevedo, J. Steinbacher, D. T. McQuade, Journal of the American Chemical
Society 2005, 127, 10498.
(55) J. W. Kim, A. S. Utada, A. Fernandez-Nieves, Z. B. Hu, D. A. Weitz, Angewandte
Chemie-International Edition 2007, 46, 1819.
(56) H. Zhang, E. Tumarkin, R. Peerani, Z. Nie, R. M. A. Sullan, G. C. Walker, E.
Kumacheva, Journal of the American Chemical Society 2006, 128, 12205.
(57) S. Sugiura, T. Oda, Y. Izumida, Y. Aoyagi, M. Satake, A. Ochiai, N. Ohkohchi, M.
Nakajima, Biomaterials 2005, 26, 3327.
(58) S. Iwamoto, K. Nakagawa, S. Sugiura, M. Nakajima., AAPS PharmSciTech 2002, 3,
72.
(59) LH. Hung, A. Lee, Journal of Medical and Biological Engineering 2007, 27, 1.
(60) J. Zhu, R. C. Hayward, Angewandte Chemie-International Edition 2008, 47, 2113.
25
(61) P. C. Lewis, R. R. Graham, Z. H. Nie, S. Q. Xu, M. Seo, E. Kumacheva,
Macromolecules 2005, 38, 4536.
(62) M. Zourob, S. Mohr, A. G. Mayes, A. Macaskill, N. Perez-Moral, P. R. Fielden, N.
J. Goddard, Lab on a Chip 2006, 6, 296.
(63) C. H. Yang, K. S. Huang, Y. S. Lin, K. Lu, C. C. Tzeng, E. C. Wang, C. H. Lin, W.
Y. Hsu, J. Y. Chang, Lab on a Chip 2009, 9, 961.
(64) C. A. Serra, Z. Q. Chang, Chemical Engineering & Technology 2008, 31, 1099.
(65) C. Ohm, C. Serra, R. Zentel, Advanced Materials 2009, 21, 4859.
(66) S. Dubinsky, H. Zhang, Z. H. Nie, I. Gourevich, D. Voicu, M. Deetz, E.
Kumacheva, Macromolecules 2008, 41, 3555.
(67) S. Dubinsky, J. I. Park, I. Gourevich, C. Chan, M. Deetz, E. Kumacheva,
Macromolecules 2009, 42, 1990.
(68) W. Li, H. H. Pharn, Z. Nie, B. MacDonald, A. Guenther, E. Kumacheva, Journal of
the American Chemical Society 2008, 130, 9935.
(69) W. H. Tan, S. Takeuchi, Advanced Materials 2007, 19, 2696.
(70) P. Panda, S. Ali, E. Lo, B. G. Chung, T. A. Hatton, A. Khademhosseini, P. S. Doyle,
Lab on a Chip 2008, 8, 1056.
(71) W. J. Jeong, J. Y. Kim, J. Choo, E. K. Lee, C. S. Han, D. J. Beebe, G. H. Seong, S.
H. Lee, Langmuir 2005, 21, 3738.
(72) R. F. Shepherd, J. C. Conrad, S. K. Rhodes, D. R. Link, M. Marquez, D. A. Weitz, J.
A. Lewis, Langmuir 2006, 22, 8618.
(73) D. K. Hwang, D. Dendukuri, P. S. Doyle, Lab on a Chip 2008, 8, 1640.
(74) E. Tumarkin, E. Kumacheva, Chemical Society Reviews 2009, 38, 2161.
(75) H. Zhang, E. Tumarkin, R. M. A. Sullan, G. C. Walker, E. Kumacheva,
Macromolecular Rapid Communications 2007, 28, 527.
(76) Z. H. Nie, W. Li, M. Seo, S. Q. Xu, E. Kumacheva, Journal of the American
Chemical Society 2006, 128, 9408.
(77) N. Prasad, J. Perumal, C. H. Choi, C. S. Lee, D. P. Kim, Advanced Functional
Materials 2009, 19, 1656.
(78) T. Nisisako, T. Torii, Advanced Materials 2007, 19, 1489.
26
(79) T. Nisisako, T. Torii, T. Takahashi, Y. Takizawa, Advanced Materials 2006, 18,
1152.
(80) D. Graham-Rowe, Nature Photonics 2007, 1, 248.
(81) C. S. Peyratout, L. Dahne, Angewandte Chemie-International Edition 2004, 43,
3762.
(82) J. Dusseault, F. A. Leblond, R. Robitaille, G. Jourdan, J. Tessier, M. Menard, N.
Henley, J. P. Halle, Biomaterials 2005, 26, 1515.
(83) S. Okushima, T. Nisisako, T. Torii, T. Higuchi, Langmuir 2004, 20, 9905.
(84) M. Seo, C. Paquet, Z. H. Nie, S. Q. Xu, E. Kumacheva, Soft Matter 2007, 3, 986.
(85) Z. H. Nie, S. Q. Xu, M. Seo, P. C. Lewis, E. Kumacheva, Journal of the American
Chemical Society 2005, 127, 8058.
(86) J. Wan, A. Bick, M. Sullivan, H. A. Stone, Advanced Materials 2008, 20, 3314.
(87) E. Lorenceau, A. S. Utada, D. R. Link, G. Cristobal, M. Joanicot, D. A. Weitz,
Langmuir 2005, 21, 9183.
(88) D. Lee, D. A. Weitz, Advanced Materials 2008, 20, 3498.
(89) R. K. Shah, J. W. Kim, J. J. Agresti, D. A. Weitz, L. Y. Chu, Soft Matter 2008, 4,
2303.
(90) J. L. Steinbacher, R. W. Y. Moy, K. E. Price, M. A. Cummings, C. Roychowdhury,
J. J. Buffy, W. L. Olbricht, M. Haaf, D. T. McQuade, Journal of the American Chemical
Society 2006, 128, 9442.
(91) C. H. Choi, J. H. Jung, D. W. Kim, Y. M. Chung, C. S. Lee, Lab on a Chip 2008, 8,
1544.
(92) A. B. Subramaniam, M. Abkarian, H. A. Stone, Nature Materials 2005, 4, 553.
(93) S. Abraham, E. H. Jeong, T. Arakawa, S. Shoji, K. C. Kim, I. Kim, J. S. Go, Lab on
a Chip 2006, 6, 752.
(94) Z. H. Nie, J. I. Park, W. Li, S. A. F. Bon, E. Kumacheva, Journal of the American
Chemical Society 2008, 130, 16508.
(95) S. Abraham, Y. H. Park, J. K. Lee, C. S. Ha, I. Kim, Advanced Materials 2008, 20,
2177.
(96) C. Priest, A. Quinn, A. Postma, A. N. Zelikin, J. Ralston, F. Caruso, Lab on a Chip
2008, 8, 2182.
27
(97) S. H. Kim, S. J. Jeon, G. R. Yi, C. J. Heo, J. H. Choi, S. M. Yang, Advanced
Materials 2008, 20, 1649.
(98) S. M. Yang, S. H. Kim, J. M. Lim, G. R. Yi, Journal of Materials Chemistry 2008,
18, 2177.
(99) S. H. Kim, S. J. Jeon, S. M. Yang, Journal of the American Chemical Society 2008,
130, 6040.
(100) H. Onoe, K. Matsumoto, I. Shimoyama, Small 2007, 3, 1383; J. A. Champion, Y. K.
Katare, S. Mitragotri, Proceedings of the National Academy of Sciences of the United States
of America 2007, 104, 11901.
(101) G. A. Gross, C. Hamann, M. Gunther, J. M. Kohler, Chemical Engineering &
Technology 2007, 30, 341.
(102) M. Seo, Z. H. Nie, S. Q. Xu, M. Mok, P. C. Lewis, R. Graham, E. Kumacheva,
Langmuir 2005, 21, 11614.
(103) M. Seo, Z. H. Nie, S. Q. Xu, P. C. Lewis, E. Kumacheva, Langmuir 2005, 21, 4773.
(104) K. Liu, H. J. Ding, J. Liu, Y. Chen, X. Z. Zhao, Langmuir 2006, 22, 9453.
(105) B. G. Wang, H. C. Shum, D. A. Weitz, Chemphyschem 2009, 10, 641.
(106) D. C. Pregibon, M. Toner, P. S. Doyle, Science 2007, 315, 1393.
(107) D. Dendukuri, T. A. Hatton, P. S. Doyle, Langmuir 2007, 23, 4669.
(108) D. Dendukuri, S. S. Gu, D. C. Pregibon, T. A. Hatton, P. S. Doyle, Lab on a Chip
2007, 7, 818.
(109) J. H. Jang, D. Dendukuri, T. A. Hatton, E. L. Thomas, P. S. Doyle, Angewandte
Chemie-International Edition 2007, 46, 9027.
(110) K. W. Bong, D. C. Pregibon, P. S. Doyle, Lab on a Chip 2009, 9, 863.
(111) J. R. Lindner, Nature Reviews Drug Discovery 2004, 3, 527.
(112) E. G. Schutt, D. H. Klein, R. M. Mattrey, J. G. Riess, Angewandte Chemie-
International Edition 2003, 42, 3218.
(113) K. Ferrara, R. Pollard, M. Borden, Annual Review of Biomedical Engineering 2007,
9, 415.
(114) Y. Y. Liu, H. Miyoshi, M. Nakamura, Journal of Controlled Release 2006, 114, 89.
28
(115) C. A. Molina, M. Ribo, M. Rubiera, J. Montaner, E. Santamarina, R. Delgado-
Mederos, J. F. Arenillas, R. Huertas, F. Purroy, P. Delgado, J. Alvarez-Sabin, Stroke 2006,
37, 425.
(116) E. C. Unger, T. Porter, W. Culp, R. Labell, T. Matsunaga, R. Zutshi, Advanced
Drug Delivery Reviews 2004, 56, 1291.
(117) E. Dickinson, Current Opinion in Colloid & Interface Science, 15, 40.
(118) U. Farook, E. Stride, M. J. Edirisinghe, European Biophysics Journal with
Biophysics Letters 2009, 38, 713.
(119) Q. Xu, M. Nakajima, S. Ichikawa, N. Nakamura, P. Roy, H. Okadome, T. Shiina,
Journal of Colloid and Interface Science 2009, 332, 208.
(120) E. Stride, M. Edirisinghe, Soft Matter 2008, 4, 2350.
(121) E. Dickinson, R. Ettelaie, B. S. Murray, Z. P. Du, Journal of Colloid and Interface
Science 2002, 252, 202.
(122) P. Garstecki, G. M. Whitesides, Physical Review Letters 2006, 97.
(123) P. Garstecki, M. J. Fuerstman, H. A. Stone, G. M. Whitesides, Lab on a Chip 2006,
6, 437.
(124) B. P. Binks, R. Murakami, Nature Materials 2006, 5, 865.
(125) U. T. Gonzenbach, A. R. Studart, E. Tervoort, L. J. Gauckler, Angewandte Chemie-
International Edition 2006, 45, 3526.
(126) E. Dressaire, R. Bee, D. C. Bell, A. Lips, H. A. Stone, Science 2008, 320, 1198.
(127) M. Borden, Soft Matter 2009, 5, 716.: E. Talu, K. Hettiarachchi, R. L. Powell, A. P.
Lee, P. A. Dayton, M. L. Longo, Langmuir 2008, 24, 1745.
(128) J. I. Park, Z. Nie, A. Kumachev, A. I. Abdelrahman, B. R. Binks, H. A. Stone, E.
Kumacheva, Angewandte Chemie-International Edition 2009, 48, 5300.
(129) M. H. Lee, V. Prasad, D. Lee, Langmuir, 26, 2227.
(130) K. Hettiarachchi, A. P. Lee, S. Zhang, S. Feingold, P. A. Dayton, Biotechnology
Progress 2009, 25, 938.
(131) J. I. Park, E. Tumarkin, E. Kumacheva, Macromolecular Rapid Communications
2010, 31, 222.
(132) W. Engl, R. Backov, P. Panizza, Current Opinion in Colloid & Interface Science
2008, 13, 206.
29
(133) Ehrfeld. W, Hessel. V, L. H., Microreactors-New Technology for Modern
Chemistry, Wiley-VCH, Weinheim 2000.
(134) T. Nisisako, T. Torii, Lab on a Chip 2008, 8, 287.
(135) V. Barbier, H. Willaime, P. Tabeling, F. Jousse, Physical Review E 2006, 74.
(136) W. Li, E. W. K. Young, M. Seo, Z. Nie, P. Garstecki, C. A. Simmons, E.
Kumacheva, Soft Matter 2008, 4, 258.
(137) I. Kobayashi, S. Mukataka, M. Nakajima, Langmuir 2005, 21, 7629.
(138) W. Li, J. Greener, D. Voicu, E. Kumacheva, Lab on a Chip 2009, 9, 2715.
30
Chapter 2
Materials and Methods
This chapter describes the materials and methods used in the present research from
Chapter 3 to 6.
2.1 Materials
Gaseous N2 (purity 99.99 %) and CO2 (purity 99.8 %) were purchased from BOC
Canada. Sodium hydroxide (NaOH, EMD Canada) and hydrochloric acid (HCl, EMD
Canada) were used to adjust the pH. Deionized water was purified by using a Mlli-Q Plus
purification system (Millipore Corp.). Ethanol was received from Commercial Alcohols
Inc.. Hexadecane, tri(propylene glycol) diacrylate (TPGDA), poly (ethylene glycol)
diacrylate (PEGDA, Mn~700 g/mol), divinylbenzene (DVB, technical grade, 80%), styrene,
acrylic acid (AA), methacrylic acid (MAA, 99%), methyl methacrylate (MMA, 99%) and
4-vinylpyridine (4VP, 95%) were purchased from Sigma-Aldrich, Canada. 2-(6-
methacryloyloxyhexyl)-thioxantheno[2,1,9-dej]isoquinoline-1,3-dione (Hostasol Yellow,
HY) dye monomer1 was synthesized and provided by Dr. Hung Pham. Initiators, hydroxy-
2-methylpropiophenone (HMPP), 2,2-dimethoxy-2-phenyl-acetophenone (DMPA), and
2,2′-azobis(2-methylpropionamidine) dihydrochloride (V50) were purchased from Sigma-
Aldrich, Canada. 2,2’-azobis(2-methylbutyronitrile) (AMBN) was purchased from Wako
Pure Chemical Industries Ltd.. Water soluble stabilizers, polyvinylpyrrolidone (PVP360,
Mw~360,000 g/mol), Triton X-305 (TX-305, 70% solution in water) and Triton X-100 (TX-
31
100) were purchased from Sigma-Aldrich, Canada. Biopolymers, alginic acid sodium salt
(alginate), hen egg white lysozyme (lysozyme) and bovine albumin serum were purchased
from Sigma-Aldrich, Canada. Fluorescein isothiocyanate (FITC)-conjugated bovine
albumin serum2 was provided by Dr. Raheem Peerani. Thioflavin-T (ThT), gold (III)
chloride trihydrate and sodium citrate (Sigma-Aldrich, Canada) were used as received.
Poly(dimethylsiloxane) (PDMS, Sylgard® 184) and photoresist resin (SU-8 50, 25) were
purchased from Dow Corning (USA), and Microchem Co. (MA, USA), respectively.
Carboxylated 3 μm-diameter SiO2 particles were purchased from Micromod
Partikeltechnologie GmbH, Germany. 100 nm-diameter Fe3O4 particles were purchased
from Chemicell, Germany. The synthesis of 3.5 µm-diameter poly(styrene-co-AA)
particles was carried out by Mr. Ahmed I. Abdelrahman.3,4 Cationic poly(MMA-co-4PV)
particles with a diameter of 0.7 µm were synthesized using a multi-stage surfactant-free
emulsion polymerization by Dr. Alla Petukhova.5 Carboxylated SiO2-encapsulated
CdSe/ZnS particles were synthesized and provided by Mr. Ivan Gorelikov.6 Poly(DVB-co-
MAA) particles with a diameter of 3.5 µm-diameter were synthesized by precipitation
polymerization.7 Gold nanoparticles with a diameter of 12 nm-diameter were synthesized
by a citrate reduction method.8
2.2 Methods
2.2.1 Mask Design
We designed masks using FreeHand® (Adobe Systems Inc., USA) software, and
printed the designed pattern on transparencies using laser printers with a resolution of 2500
dpi.
32
2.2.2 Microfabrication of Negative Masters
Masters were prepared in the cleanroom facility at Bahen Center, University of
Toronto. We fabricated the negative masters with microchannels of SU-8 photoresist (SU-8
25 or 50) on 3-inch silicon (Si) wafer substrates using photolithography.9,10 Briefly, the Si
wafers were washed several times with acetone and methanol. The surface of the wafers
was dried by blowing dry N2 gas. We spin-coated the Si wafers with SU-8 photoresist at a
typical spin rate of 1200 rpm in order to achieve 120 µm height features. After the spin-
coating process, the photoresist layer was thermally baked to evaporate the solvent, γ-
butyrolactone. The photoresist layer covered with a mask was exposed to and cured by UV-
light (Karl Suss MA6 mask aligner, λ=365~405 nm) to create the microchannel. Typical
exposure time was approximately 50 sec. Following UV-exposure, a post-bake process was
conducted to enhance the crosslinking of the UV-exposed areas of the photoresist. The
post-baked Si wafers with the crosslinked patterns of microchannel were immersed for ca.
10 min in the developer solution (1-methoxy-2-propanol acetate) to remove the non-
crosslinked regions of the photoresist. The Si wafers with the microchannel pattern were
rinsed several times with isopropanol and methanol and dried under a gentle stream of
nitrogen.
2.2.3 Fabrication of Microfluidic Devices
Microfluidic (MF) devices were fabricated in poly(dimethylsiloxane) (PDMS)
elastomer using a standard soft-lithographic procedure.11 The PDMS elastomer was
prepared from Sylgard® 184 (Dow Corning Cop., USA). The Sylgard® 184 base polymer
33
contains vinyl-terminated dimethylsiloxane oligomers, platinum catalyst, and silica filler
(dimethylvinylated and trimethylated silica).12 The base polymer was mixed with the curing
agent containing a cross-linker (dimethyl-methylhydrogen siloxane) and an inhibitor
(tetramethyltetravinyl cyclotetrasiloxane) at 10:1 (w/w) ratio. Air was removed from the
mixture under vacuum for 20 min. The mixture of the prepolymer mixture was poured onto
the master and baked at 75 °C in an oven for more than 12 hours. After curing, the replica
was peeled from the master, and holes were created by flat-tipped needle at the designated
positions corresponding to the inlets and outlets. The replica and a substrate (a plane PDMS
sheet) were oxidized in a plasma cleaner chamber (PDC-3XG, Harrick, USA) for 120 sec.
The plasma-treated replica and the substrate were brought in contact and sealed
immediately. Polytetrafluoroethylene tubings (Small Parts, USA) were placed into the
holes and sealed with epoxy glue (Lepage 12, Henkel Canada Co., Brampton, ON).
2.2.4 Microfluidic Experiments
Typically, gases (CO2 and N2) were supplied to the MF device via
polytetrafluoroethylene tubing (Small Parts, USA) attached to a Bellofram pressure
regulator. The applied pressure was from 27.6 to 82.7 kPa. Syringe pumps (Harvard
Apparatus, USA, PHD 2000 series) at flow rates ranging from 4 to 24 mL/h were used to
introduce liquids into the microchannel. To ensure stable bubble or droplet generation, after
introducing the liquids or gases in the microchannel, the MF device was equilibrated for at
least, 3 minutes.
34
2.2.5 Photopolymerization Experiments
To image the cross-section of bubbles coated with poly(styrene-co-AA) particles, 0.5
mL of the diluted bubble dispersion were mixed with 0.2 mL of 70 wt. % aqueous solution
of PEGDA containing the photoinitiator, HMPP. The dispersion was exposed to a UV-lamp
for 20-30 sec (UVAPRINT 40C/CE, Dr. K. Hönle GmbH UV-Technologie, Germany, λ =
330 ~ 380 nm, intensity of 200 mW/cm2). The UV-cured polymer embedded with the
armoured bubbles was microtomed (Leica, UCT Ultramicrotome) and mounted on a sample
grid before imaging.
Photopolymerization of monomer droplets of TPGDA loaded with the photoinitiator,
DMPA, and HY-labelled poly(DVB-co-MAA) particles was carried out immediately after
the release of droplets from the MF device. The droplets were collected in a vial and
exposed to UV irradiation for 1 min (UVAPRINT 40C/CE, Dr. K. Hönle GmbH UV-
Technologie, Germany, λ = 330 ~ 380 nm, intensity of 200 mW/cm2).
2.3. Characterization
2.3.1 Optical and Fluorescence Microscopy Imaging
An Olympus BX51 microscope (Olympus, USA) with a high-speed camera
(Photometrics CoolSNAR ES) was used to image the generation of bubbles, droplets and
particles in the MF device. We also used the same setup to image bubbles and droplets on a
glass slide. The fluorescence mode was used to acquire images of bubbles, droplets and
particles coated with fluorescence dye-labelled polymer beads.
35
2.3.2 Size Distribution of Bubbles, Droplets and Particles
The dimensions of bubbles, droplets and particles were analyzed using Image Pro Plus
(Media Cybernetics, USA) software. The polydispersity of bubbles or droplets was
characterized as the coefficient of variance (CV) as CV (%) = (σ/D)x100 where σ is the
standard deviation of the size of bubbles or droplets and D is the mean diameter of bubbles
or droplets. The volume, V, of spherical bubbles was calculated as V = (4/3)π(D/2)3. When
the value of D exceeded the height, h, of the microchannels, the bubbles formed discoids.
Their volume was approximated as V = (π /12)[2D3-(D-h)2(2D + h)].14
2.3.3 Scanning and Transmission Electron Microscopy imaging
We used scanning electron microscopy (SEM, Hitachi S-570 or Hitachi S-3400N) and
scanning transmission electron microscopy (STEM, Hitachi S-5200) to image the surface
and internal morphologies of bubbles and polyTPGDA particles coated with polymer beads.
The same microscopes were used to image the surface structure of the bubbles encapsulated
with a mixture of lysozyme, alginate and nanoparticles. The samples were washed with
deionized water five times and dried or frozen at -20 oC on the SEM or TEM grids prior to
imaging.
2.3.4 Laser Confocal Fluorescence Microscopy imaging
Laser confocal fluorescence microscopy (LCFM, Zeiss LSM 510 Meta microscope
equipped with an Axiovert 200M microscope, λex= 364 or 380 nm) was used to image
36
bubbles encapsulated with HY-labelled polymer beads, or a mixture of lyzoyme and
alginate carrying SiO2-encapsulated CdSe/ZnS nanoparticles. We also used LCFM to
image poly TPGDA particles loaded with HY-labelled polymer beads.
2.3.5 Contact angle measurement
A Drop Shape Analysis System (DSA100, KRÜSS, USA) was used to examine the
pH-dependent wetting properties of the polymer particles. The contact angles between
poly(styrene-co-AA) film on a glass substrate and an aqueous solution of NaOH at various
pHs were recorded. We also measured the contact angles between poly(DVB-co-MAA)
film and a mixture of water and ethanol (85:15, v/v ratio) or TPGDA monomer. A 3 μL-
droplet was deposited on the surface of the polymer film where advancing contact angles
were acquired.
2.3.6 Circular Dichroism Spectroscopy
Circular dichroism (CD) spectra were obtained using a JASCO J-710
spectropolarimeter in the range 200-250 nm, in order to examine the conformational
changes for lysozyme following their deposition on the bubble surface. The bubbles
encapsulated with a mixture of lysoyzme and alginate were washed with water five times.
The bubble dispersion was further diluted approximately 30 times with a 50 mM of
trishydroxymethylaminomethane-HCl buffer (pH=8) and transferred to a quartz cuvette
with 1 cm path length.
37
2.3.7 Fluorescence Spectroscopy
A fluorescence spectrophotometer (Cary Eclipse, Varian) was used to carry out ThT
binding tests. Aliquots of biopolymer-encapsulated bubble dispersions were added to 2 mL
of 50 μM ThT solution at pH=8. The intensity of fluorescence emission was recorded at
λex=412 nm. In corresponding control experiments, we examined fluorescence intensity of
freshly prepared a 50 μM ThT solution and a 0.003 wt. % of lysozyme or alginate dissolved
in 50 μM ThT solution at pH=8. In addition, the intensity profiles of photoluminescence
emission were measured at λex=364 nm for dispersions of biopolymer-engulfed
nanoparticle-free bubbles and biopolymer-engulfed bubbles functionalized with SiO2-
encapsulated CdSe/ZnS nanoparticles.
References
(1) F. Tronc, M. Li, J. P. Lu, M. A. Winnik, B. L. Kaul, J. C. Graciet, J. Polym. Sci. Pol.
Chem. 2003, 41, 766.
(2) G. Hungerford, J. Benesch, J. F. Mano, R. L. Reis, Photochemical &
Photobiological Sciences 2007, 6, 152.
(3) J. S. Song, L. Chagal, M. A. Winnik, Macromolecules 2006, 39, 5729.
(4) J. S. Song, F. Tronc, M. A. Winnik, Polymer 2006, 47, 817.
(5) A. Petukhova, A. S. Paton, I. Gourevich, S. V. Nair, H. E. Ruda, A. Shik, E.
Kumacheva, Canadian Journal of Chemistry-Revue Canadienne De Chimie 2010, 88, 298.
(6) I. Gorelikov, N. Matsuura, Nano Letters 2008, 8, 369.
(7) W. H. Li, H. D. H. Stover, J. Polym. Sci. Pol. Chem. 1999, 37, 2899.
(8) B. V. Enustun, J. Turkevich, Journal of the American Chemical Society 1963, 85,
3317.
(9) G. M. Wallraff, W. D. Hinsberg, Chemical Reviews 1999, 99, 1801.
38
(10) http://www.microchem.com/products/su_eight.htm.
(11) Y. N. Xia, G. M. Whitesides, Annual Review of Materials Science 1998, 28, 153.
(12) A. Olah, H. Hillborg, G. J. Vancso, Applied Surface Science 2005, 239, 410.
39
40
Chapter 3
Microfluidic Generation and Dissolution of CO2 bubbles
Reproduced with permission from Soft Matter, 2010, 6, 630. Copyright 2010 RSC
Publishing.
This chapter describes a microfluidic strategy for the generation of small (<10 μm)
microbubbles with a narrow size distribution. The proposed approach exploits the
following features: (i) the generation of bubbles from gaseous CO2 and (ii) the controlled
dissolution of these bubbles until they reach a desired size. We investigate the role of
various factors on the formation and the extent of bubble dissolution and show that the final
dimensions of bubbles are determined by the flow rate of the continuous aqueous phase and
the acid-base equilibria established in the continuous phase flowing within the
microchannels.
3.1 Introduction
Bubbles with dimensions smaller than 10 μm have a broad range of biomedical
applications, including ultrasound imaging, gene therapy and targeted site-specific drug
delivery.1-3 Control of the size and the size distribution of bubbles is critical in their ability
to safely circulate in blood vessels and their reflectivity to ultrasound wave.4 Generally,
bubbles with dimensions in the size range specified above are generated by sonication and
high shear emulsification.5-7 These methods simply dispersing gas or liquid in a suspension
41
containing a suitable coating material by applying high intensity ultrasound or shear stirring.
Since the experimental variables cannot be precisely managed, these methods provide
insufficient control over bubble dimension and size distribution.
Microfluidic (MF) generation of bubbles has recently emerged as a means for
producing bubbles with extremely narrow size distribution.8-12 Bubbles were generated in
three types of MF devices, namely, a T-junction,11,12 a planar flow-focusing device9 and a
capillary flow-focusing device.8 Although the mechanism of bubble generation in these MF
devices was different depending on the geometry of the MF device, the size of bubbles was
precisely controlled by the periodic breakup of the gaseous thread and it was tuned by
varying the flow rate of the continuous liquid phase and the pressure of the gas.9
Typically, the size of bubbles produced in MF devices was in the range from 5 to 200
μm.9-12 Smaller bubbles were generated in a capillary flow-focusing device8 or in a planar
flow-focusing device by reducing the dimensions of the nozzle.13 For example, bubbles
with diameters of ca. 5 μm were produced by using an orifice with a width of 7 μm and a
height of 25 μm. It is noted that coaxial electrohydrodynamic atomisation has been lately
developed for the controlled production of small bubbles (ca. 6 μm). In this technique, a
coaxial jet of gas and liquid mixed with a surfactant is formed and then atomised to
generate uniform bubbles by applying the difference in an electrical potential.1 The
dimension of bubbles is determined by the flow rate ratio of gas and liquid, and the applied
voltage.
In this chapter, a different approach to the generation of small (<10 μm) bubbles is
described. Our strategy includes (i) MF generation of bubbles of CO2 mixed with a minute
amount of low-soluble impurity gases in an aqueous phase and (ii) controllable in situ
dissolution of CO2 under conditions of varying acidity of the continuous medium while the
42
size of bubbles is stabilized. In this chapter, we show that the final dimensions of bubbles
are determined by their initial dimensions (controlled by the gas pressure, the flow rate of
the continuous phase, and the initial pH of the liquid phase) and the acid-base equilibria
established in the continuous phase in the microfluidic device.
3.2 Results and Discussions
The materials and methods described in the present chapter are provided in Chapter 2.
3.2.1 Generation of CO2 bubbles
Figure 3.1 shows a schematic of the MF device used in the present work. The device
contains a flow-focusing bubble generator9 and a serpentine downstream channel with a
length of 230 mm. A study of the formation and dissolution of CO2 bubbles was performed
in the device with the orifice length, width, and height of 30, 33, and 115 μm, respectively.
Small (<10 μm in diameter) bubbles were generated in the MF device with the orifice
length, width, and height of 60, 22, and 40 μm, respectively.
43
Figure 3.1 Schematic of the MF device and the dissolution of CO2 bubbles. Inset shows a zoomed in schematic of the orifice and the generation of bubbles. CO2 gas (99.8% purity) and an aqueous continuous phase (2 wt. % Triton X-100) were introduced into the central and the outer channels of the MF flow-focusing device, respectively. The pressure of the gas (PCO2) was fixed as 27.6 kPa (for the generation of small bubbles it was 55.2 kPa). The flow rate (QL) of the continuous aqueous phase varied from 5 to 12 mL/h. The value of pH of the continuous phase was controlled by varying the concentrations of HCl or NaOH in the range 1.5≤pH≤13.2 At the exit of the microchannel, the bubbles were exposed to an ambient pressure of 101.3 kPa.
Figure 3.2 schematically shows the shrinkage of CO2 bubbles. The diffusion of CO2
from the bubbles into the continuous phase was followed by the chemical reaction with
H2O or OH− ions.14 The process of CO2 dissolution is described by Henry’s law (eq. 3.1)
and pH-dependent chemical reactions (eq. 3.2-4):15
[CO2]l kH PCO2 (eq. 3.1)
pH<10: CO2 + H2O HCO3- + H+ (K = 4.4 x 10-7) (eq. 3.2)
pH>10: CO2 + OH- HCO3- (K = 3.2x 107) (eq. 3.3)
HCO3- + OH- CO3
2- + H2O (K = 3.5x 103) (eq. 3.4)
44
Figure 3.2 Schematic of pH-dependent dissolution of CO2 bubble leading to bubble shrinkage.
where [CO2]l, kH, and PCO2 are the concentration of molecularly dissolved CO2, Henry’s
law constant (for CO2 at 25oC kH=3.2x10-4 mol/(L kPa)), and the applied inlet gas pressure,
respectively, and K is the equilibrium constant. We note the important role of the acidity of
the medium: the equilibrium constant at pH>10 is 14 orders of magnitude larger than that at
pH<10. Thus, the change in the acidity of the continuous phase can be used to control the
extent of CO2 dissolution and hence the ultimate dimensions of bubbles flowing through
the MF device.
Figure 3.3A,B shows the representative optical microscopy images of the CO2 bubbles
generated in the MF device in acidic (pH=1.5) and basic (pH=13.2) aqueous solutions,
respectively. The bubbles formed in the geometrically controlled regime. In this regime, the
formation of bubble is determined by blockage of the orifice and the downstream channel
by the gas thread, which subsequently causes the increase of pressure behind the orifice and
break up of the bubble. Since the unperturbed dimension of the bubbles was larger than the
height of the microchannel, they had a disc-like shape.9,16 Polydispersity (CV) of the
bubbles was maintained below 5%.
45
Two features are clearly shown from Figure 3.2A,A’,B,B’: (i) in the basic solution the
bubbles underwent a stronger reduction in size than in the acidic solution and (ii) upon the
formation, the bubbles had a significantly larger initial size at higher pH values. The former
effect was expected: the reduction in the size of the bubbles was caused by the dissolution
of CO2 in the aqueous continuous phase, accompanied by the reaction between CO2 and
H2O (at pH=1.5) or between CO2 and OH- ions (at pH=13.2).15 A greater extent of
shrinkage of the bubbles at higher pH values occurred due to the stronger mass transfer of
CO2 into an aqueous phase.14 However, we expected that a larger initial size of the bubbles
in the basic solution was counter-intuitive: at higher pH values, the initial size of CO2
bubbles would rapidly decrease with increasing pH values.
Figure 3.3 Representative optical microscopy images taken immediately after the generation of CO2 bubbles (A,B) and N2 bubbles (C,D) and in the downstream channel 107 mm (A’-D’) away from the orifice. The initial pH value of the continuous phase was 1.5 (A,A’,C,C’) and 13.2 (B,B’,D,D’). The bubbles were generated at PCO2=27.6 kPa and QL=6 mL/h. Scale bar is 200 μm.
46
We hypothesized that a pH-dependent formation of CO2 bubbles was affected by their
dissolution in the downstream channel and the related change in resistance to flow. Thus
the size of a newly formed ‘‘successor’’ bubble was affected by the pH-dependent
dissolution of its predecessors. Such coupling or ‘‘memory’’ was earlier reported for the
generation and flow of foams in a MF device, at which at constant input parameters (the
flow rate of the continuous phase and the pressure of the gas) the inflow gas flow rate and
bubble volume showed periodic oscillations affected by the presence of pre-formed bubbles
flowing in the downstream channel.17
To verify this hypothesis, we examined the formation of N2 bubbles in the same MF
bubble generator and under the same hydrodynamic conditions as those used for producing
CO2 bubbles: we note that compared to the solubility in water of CO2 of ca. 0.15 wt%, the
solubility of N2 in water is only ca. 0.005 wt% at 101.3 kPa and 25 oC.18 Figure 3.3C,D
illustrates the generation of N2 bubbles under acidic (pH=1.5) and basic (pH=13.2)
conditions; Figure 3.3C’,D’ shows the corresponding N2 bubbles in the downstream
channel. Since the pressure in the microchannel reduced along its length,19,20 the bubbles
slightly expanded. Thus, we conclude that no noticeable effect of the acidity of the medium
on the size of N2 bubbles was observed.
Furthermore, we examined the effect of pH of the continuous phase on the initial and
final (stabilized) volume of CO2 and N2 bubbles at various pH values is shown in Figure
3.4. For CO2 bubbles, the changes in their initial and final volumes were in anti-correlation:
a higher pH value led to a stronger increase in the initial volume and a stronger reduction in
the final bubble size. On the other hand, both the initial and final volumes of N2 bubbles
showed no significant change at 1.5≤pH≤13.2.
47
Figure 3.4 Variation in initial (●, ), and final (○,) volumes of CO2 (●,○) and N2 (,) bubbles plotted as a function of the initial pH of the continuous aqueous phase.
Based on the results presented in Figure 3.3 and 3.4, we attributed the effect of pH of
the continuous phase on the original dimensions of CO2 bubbles to their pH-dependent
dissolution in the downstream channel. The dissolution of CO2 led to the reduction in
pressure in the MF device: it increased the inflow rate of the gaseous CO2 to compensate
for the drop in pressure. Thus a stronger dissolution of bubbles in the downstream channel
at high pH values led to the formation of larger bubbles. These results were qualitatively
agreed to the observations by other research groups on increasing initial volume of bubbles
when the hydrodynamic pressure in the microchannel was reduced.20
In addition, we found the effect of pH on the frequency (f) of formation of CO2 bubbles
by the ratio f/fpH=13.2 where f and fpH=13.2 are the frequencies of bubble formation at a
particular pH and at pH=13.2, respectively (the value of pH=13.2 was the most basic
condition used in the experiments). The frequency of the formation of bubbles was
estimated using experimental results as follows.
48
where Pc is the pressure drop in the downstream channel, Vc is the volume of the channel, a
is a dimensionless parameter, μ is the viscosity of the continuous phase, L, w, and h are the
length, the width, and the height of the channel, respectively, Lnb is the length of the
channel excluding the dimensions of the CO2 bubbles, and Di is the diameter of the i-th
bubble.19,20 Figure 3.5 shows that the frequency of the formation of CO2 bubbles increased
with increasing pH. In contrast, for N2 bubbles, the frequency of bubble generation was
practically constant regardless of the value of pH of the continuous phase owing to the
invariance in the volume of bubbles with the changing acidity of the medium.
Figure 3.5 pH-dependence of the relative frequency of bubble generation with respect to the frequency of the formation of bubbles at pH=13.2 for CO2(■) and N2() bubbles. The bubbles were generated at P= 27.6 kPa, QL=6 mL/h.
2
2
2
2
CO
CCO
CCO
CCO )()(P
PPVP
PPQf L −
∝−
∝
31
L
3nbL
C
)(
wh
DLQa
whLQaP
ni
ii∑
=
=
−==
µµ
(eq. 3.5)
(eq. 3.6)
49
3.2.2 Dissolution of CO2 Bubbles
In the downstream channel, CO2 bubbles underwent uniform, pH-dependent
dissolution, until their dimensions were stabilized. Figure 3.6A-C shows the snapshots of
the individual bubbles flowing in the continuous phase with varying initial pH values at the
distance, d, of 0.05, 12 and 107 mm from the orifice, respectively. The dissolution of CO2
led to the decrease in the initial value from pH=13.2 to 12.5. When the initial value of pH
was 11 the value of pH decreased to 9 with CO2 dissolution. Similarly, at initial pH ranging
from 5 to 9 the value of pH reduced to 4.5. The dissolution of CO2 bubbles did not alter the
initial value of pH=1.5. The polydispersity (CV<5%) was similar to that of the initial
bubbles. Due to the narrow distribution in bubble sizes, the difference in the Laplace
pressure was reduced, thereby leading to the uniform dissolution of all the bubbles.21 In
addition, Ostwald ripening during the dissolution of bubbles was suppressed due to the
presence of low-soluble gases in the gas phase, which provided an osmotic stabilizing
effect.22
Figure 3.6 (A-C) Optical microscopy images of CO2 bubbles taken at varying distances away from the orifice in the device at (A) pH=1.5, (B) pH=9, and (C) pH=13.2. Scale bar is 50 μm. (D) Variation in the volume of CO2 bubbles examined at different pH values, plotted vs the distance, d, away from the orifice.
50
Although larger initial CO2 bubbles formed at high pH values, they shrank to smaller
size than the bubbles shrunken at low pH values. This trend is shown in Figure 3.6D, where
the variation in bubble volume is plotted as a function of the distance, d, from the orifice.
For example, the initial volume of bubbles generated at pH=13.2 was 1.7-fold larger than
the volume of bubbles produced at pH=1.5, however the final volume of bubbles generated
at pH=13.2 was only 0.2 of the stabilized volume of bubbles formed at pH=1.5. Change in
pH also affected the kinetics of bubble dissolution. Increase in pH resulted in more rapid
shrinkage of the bubbles and the stabilization of their size within a shorter distance from the
orifice. However, at d~107 mm, the dimensions of bubbles were stabilized irrespective of
the pH of the continuous phase.
In principle, the stabilization of bubble size could be attributed to one of two
mechanisms: the saturation of the aqueous continuous phase with dissolved CO2, or the
increase in the concentration of impurities of low water soluble gases in the bubbles due to
CO2 dissolution; the latter effect could suppress further CO2 dissolution owing to the low
partitioning parameter between gaseous and aqueous phases.22
In order to determine the mechanism controlling the final dimension of bubbles, we
determined the experimental and theoretical amount of dissolved CO2 in the continuous
phase. The concentration of dissolved CO2 was estimated by using Henry’s law (eq. 3.1)
and acid-base equilibria reactions (eq. 3.2-3).14,15 In [CO2]l = kH PCO2 (eq. 1), the value of
kH varies on the concentration of the ionic species in water and it is estimated as
kH=kH010-0.138I where kH0 is Henry’s law constant for CO2 in pure water (kH0=3.3 x 10-4
mol/(L kPa)) and I is the ionic strength, I=0.5∑i
ii zC 2 and Ci is the concentration of ions
with charge z.15
51
At pH<5 the total concentration of dissolved CO2 is dominated by Henry’s law (eq. 1)
because the effect of reactions are negligible.23 At pH ≥5 the total concentration of
dissolved CO2 was estimated by adding the unreacted amount of CO2 (determined by
Henry's law(eq. 1)) and reacted CO2 (from the chemical reactions (eq. 3.2-3)).15 Below we
show the detailed calculation of the total concentration of dissolved CO2 for pH=13.2,
kH=3.2 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. The calculations of the amount of dissolved
CO2 for other pH values are given in Appendix. Based on (eq. 3.1), we calculated [CO2]l
~0.009 mol/L. Then, using the reaction (eq. 3.3), we determined the concentration of
reacted CO2 by writing acid-base equilibria as follows.
CO2(l) + OH- HCO3- K = 3.2 x 107
Initial concentration 0.009 0.16
Reacted concentration -x -x x
Final concentration 0.009-x 0.16-x x
x~0.009 mol/L
The total amount of dissolved CO2 (both unreacted and reacted) is [CO2]l + x, that is,
0.018 mol/L. Using the same approach we found that for pH=11, the concentration of
dissolved CO2 was ca. 0.01 mol/L. For 5≤pH<10, by combining Henry’s law (eq. 3.1) and
reaction (eq. 3.2), we determined the total concentration of dissolved CO2 to be
approximately 0.009 mol/L. For pH=1.5, we used Henry’s law (eq. 3.1) and found the total
concentration of dissolved CO2 to be ca. 0.009 mol/L. Experimentally, to find the actual
concentration of dissolved CO2 we defined the volume of ca. 6x10-6 mL as a unit volume
containing a single initial CO2 bubble (imaged immediately after the orifice). By applying
52
the same unit volume to the position of the bubble the downstream channel at which the
bubble size was stabilized, we used the relative change in bubble volume to find the amount
of dissolved CO2 per unit volume, and obtained the actual total concentration of dissolved
CO2.
Figure 3.7A shows that in the range 1.5≤pH≤13.2 the estimated and experimentally
found concentrations of dissolved CO2 were very close. Such agreement indicated that the
dissolution of CO2 bubbles and stabilization of the dimension of bubbles were controlled by
the saturation of the aqueous continuous phase. Figure 3.7B shows the fractional change in
the volume of bubbles, ΔV/V0, plotted as a function of initial pH of the continuous phase,
where ΔV is the difference between the initial and final (stabilized) volumes of bubbles and
V0 is the initial volume of bubbles. With increasing pH values the extent of shrinking of
bubbles increased, featuring the change in the slope at pH~10. Such change was in
agreement with the difference in the equilibrium constants for reactions (eq. 3.2-3): the
equilibrium constants for reaction (eq. 3.3) (pH>10) are over 10 orders of magnitude larger
than that for reaction (eq. 3.2). At pH=13.2, the strong shift of equilibrium towards the
products led to the 97.5 % volume loss of CO2 from the bubbles, whereas at pH=1.5 it was
81.6 %.
53
Figure 3.7 (A) Comparison of theoretical (■) and experimental () amounts of dissolved CO2 (mol/L), (B) Relative change in bubble volume plotted as a function of pH of the continuous phase. PCO2=27.6 kPa, QL=6 mL/h.
Next, we examined the effect of flow rate, QL, of the continuous phase on the initial
and final volume of CO2 bubbles and the relative change in bubble volume (Figure 3.8). In
Figure 3.8A, at a particular pH value, with increasing QL the initial volume of bubbles
decreased in agreement with the earlier results on the formation of bubbles from gases with
low solubility.9 The final volume of bubbles also decreased with increasing value of QL,
presumably, due to the stronger flux of HCO3- or CO3
2- ions from the surface of CO2
bubbles occurring at higher flow rate of the continuous phase (Figure 3.8B).14 We note a
relatively narrow range of flow rates studied for PCO2=27.6 kPa: at QL<5, the gas thread
did not break up into bubbles, whereas at QL>7 mL/h the gas thread was completely
retracted in the upstream channel.
54
Figure 3.8 Effect of the flow rate of the continuous phase, QL, on (A) the initial and (B) final volume of CO2 bubbles generated at different pH values of the continuous phase. pH=1.5 (,), pH=5(,), pH=7 (,), pH=9 (,), pH=11 (,), and pH=13.2 (,). PCO2=27.6 kPa. The lines are given for eye guidance.
To provide guidance in generating small bubbles by the “dissolution” approach, the
results of dissolution experiments were summarized as a three-dimensional graph (Figure
3.9). The final stabilized bubble diameter was controlled by varying the pH value and the
flow rate of the continuous phase (at constant PCO2). The initial size of the bubbles was not
used as a variable: it was controlled by tuning the values of pH and QL. The smallest, 47
μm-diameter bubbles were produced at a flow rate of 7 mL/h and pH=13.2, by 97.5%
fractional dissolution of CO2. It should be emphasized that this graph presents the results
obtained in the microfluidic device with particular dimensions used in the present work.
55
Figure 3.9 Variation in the final diameter of CO2 bubbles, plotted as a function of pH and QL (each represents the QL range from 5 to 7 mL/h, respectively).
In addition, we estimated the diffusion time, t, of CO2 at the scale of the initial bubble
size using the relation t ~D2/DCO2, D is the bubble diameter and DCO2 is the diffusion
coefficient of CO2 in water (DCO2~1.94x104 μm2 s-1).24 For pH=13.2, PCO2=27.6 kPa,
QL=6 mL/h, and the average initial bubble diameter of 200 μm, the diffusion time was ca. 2
sec.
3.2.3 Generation of Small Bubbles
Finally, we exploited the strong, uniform, and rapid dissolution of CO2 achieved in
microfluidic channels for the preparation of small, monodisperse bubbles with dimensions
below 10 μm.
56
Following guidance provided in Figure 3.8, we generated CO2 bubbles at pH=13.2 and
flow rate of the aqueous solution of QL=12 mL/h. Following the dissolution of bubbles with
an initial diameter D0=30 μm, we achieved a bubble size of 8 μm at a polydispersity of
4±1%. The estimated frequency of bubble generation was ca. 3x104 bubbles/sec.
Figure 3.10 (A) Generation of 30 μm-diameter CO2 bubbles. (B) Bubbles with a stabilized diameter of 8 μm in the downstream channel. The bubbles were generated in the microfluidic device with the length, width, and height of the orifice of 60, 22 and 40 μm, respectively, at PCO2=55.2 kPa, QL=12 mL/h, and pH=13.2.
3.3 Conclusion
The approach based on the MF formation of CO2 bubbles mixed with minute amounts
of low water soluble gases, followed by their uniform and controllable dissolution provides
a route to generation of small, monodisperse bubbles with controllable size. In the present
work, by generating 30 μm-size CO2 bubbles we produced bubbles with dimensions of 8
μm. Sub-micron size CO2 bubbles can be generated by reducing the dimensions of
microchannels and exploiting higher flow rates of the continuous aqueous phase, which
would allow for the generation of smaller initial bubbles and a stronger flux of HCO3- and
CO3-2 ions from the bubble surface. Under the latter conditions, following complete
dissolution of CO2 the ultimate size of bubbles will be determined by the minute
57
concentration of weakly-soluble gases, e.g. perfluorocarbon, that are either exist as
impurities, or are intentionally introduced in with CO2.22
In the present work, in order to focus on the effect of pH and hydrodynamic
conditions on bubble dimensions, bubbles were not stabilized against dissolution and
coalescence. However, one can improve their stability by encapsulating them with suitable
polymers,25,26 colloidal particles,27-29 or surfactants.30
This work has several other important implications. Gases with high solubility in the
continuous phase, e.g., ammonia or CO2 have strong environmental impact, due to their
high solubility and reactivity in water.31 They are also used in heterogeneous gas-liquid
chemical reactions.32 It can be expected that microfluidic generation of monodisperse
bubbles of highly soluble, reactive gases, accompanied with a study of the variation in
bubble dimensions would enable detailed studies of gas dissolution, as well as the kinetics
of gas-liquid reactions.
Secondly, microbubbles of CO2 are used for enzyme inactivation,33 as ultrasound
contrast agents,34 and as templates for the synthesis of inorganic microspheres or porous
materials.35 Current methods of the production of CO2 bubbles provide poor control of
bubble dimensions.33-35 Controlled generation and dissolution of CO2 bubbles could solve
the problem associated with bubble disparity in size and enable the study of interactions of
CO2 with the surrounding liquid medium.
References
(1) E. Stride, M. Edirisinghe, Soft Matter 2008, 4, 2350.
58
(2) K. Ferrara, R. Pollard, M. Borden, Annual Review of Biomedical Engineering 2007,
9, 415.
(3) S. R. Sirsi, M. A. Borden, Bubble Science, Engineering and Technology 2009, 1, 3.
(4) I. Lentacker, S. C. De Smedt, N. N. Sanders, Soft Matter 2009, 5, 2161.
(5) Y. Z. Zhao, H. D. Liang, X. G. Mei, M. Halliwell, Ultrasound in Medicine and
Biology 2005, 31, 1237.
(6) W. L. Nyborg, Ultrasound in Medicine and Biology 2001, 27, 301.
(7) B. B. Jiang, C. Y. Gao, J. C. Shen, Colloid and Polymer Science 2006, 284, 513.
(8) A. M. Ganan-Calvo, J. M. Gordillo, Physical Review Letters 2001, 87.
(9) P. Garstecki, I. Gitlin, W. DiLuzio, G. M. Whitesides, E. Kumacheva, H. A. Stone,
Applied Physics Letters 2004, 85, 2649.
(10) M. Yasuno, S. Sugiura, S. Iwamoto, M. Nakajima, A. Shono, Aiche Journal 2004,
50, 3227.
(11) J. H. Xu, S. Li, G. G. Chen, G. S. Luo, Aiche Journal 2006, 52, 2254.
(12) P. Garstecki, M. J. Fuerstman, H. A. Stone, G. M. Whitesides, Lab on a Chip 2006,
6, 437.
(13) K. Hettiarachchi, E. Talu, M. L. Longo, P. A. Dayton, A. P. Lee, Lab on a Chip
2007, 7, 463.
(14) T. Madhavi, A. K. Golder, A. N. Samanta, S. Ray, Chemical Engineering Journal
2007, 128, 95.
(15) P. V. Danckwerts, Gas-liquid Reactions, McGraw-Hill Book Company, New York
1970.
(16) P. Garstecki, M. J. Fuerstman, G. M. Whitesides, Physical Review Letters 2005, 94.
(17) J. P. Raven, P. Marmottant, Physical Review Letters 2006, 97.
(18) D. W. Oxtoby, N. H. Nachtrieb, Principles of Modern Chemistry, Harcourt Brace
Jovanovich College Publishers, Orlando 1990.
(19) M. J. Fuerstman, A. Lai, M. E. Thurlow, S. S. Shevkoplyas, H. A. Stone, G. M.
Whitesides, Lab on a Chip 2007, 7, 1479.
(20) M. T. Sullivan, H. A. Stone, Philosophical Transactions of the Royal Society a-
Mathematical Physical and Engineering Sciences 2008, 366, 2131.
59
(21) E. Lorenceau, Y. Y. C. Sang, R. Hohler, S. Cohen-Addad, Physics of Fluids 2006,
18.
(22) E. G. Schutt, D. H. Klein, R. M. Mattrey, J. G. Riess, Angewandte Chemie-
International Edition 2003, 42, 3218.
(23) J. N. Butler, Carbon dioxide equilibria and their applications, Lewis Publisher,
Michigan 1981.
(24) A. Tamimi, E. B. Rinker, O. C. Sandall, Journal of Chemical and Engineering Data
1994, 39, 330.
(25) F. Cavalieri, A. El Hamassi, E. Chiessi, G. Paradossi, R. Villa, N. Zaffaroni,
Biomacromolecules 2006, 7, 604.
(26) D. G. Shchukin, K. Kohler, H. Mohvald, G. B. Sukhorukov, Angewandte Chemie-
International Edition 2005, 44, 3310.
(27) J. I. Park, Z. Nie, A. Kumachev, A. I. Abdelrahman, B. R. Binks, H. A. Stone, E.
Kumacheva, Angewandte Chemie-International Edition 2009, 48, 5300.
(28) B. P. Binks, R. Murakami, Nature Materials 2006, 5, 865.
(29) U. T. Gonzenbach, A. R. Studart, E. Tervoort, L. J. Gauckler, Angewandte Chemie-
International Edition 2006, 45, 3526.
(30) E. Dressaire, R. Bee, D. C. Bell, A. Lips, H. A. Stone, Science 2008, 320, 1198.
(31) R. Pierantozzi, Kirk-Othmer Encyclopedia of Chemical Technology 2003, 4, 803.
(32) C. D. Cooper, Kirk-Othmer Encyclopedia of Chemical Technology 2007, 26, 667.
(33) M. Shimoda, Y. Yamamoto, J. Cocunubo-Castellanos, H. Tonoike, T. Kawano, H.
Ishikawa, Y. Osajima, Journal of Food Science 1998, 63, 709.
(34) P. Redondo, J. Cabrera, Seminars in Cutaneous Medicine and Surgery 2005, 24,
175.
(35) G. Hadiko, Y. S. Han, M. Fuji, M. Takahashi, Materials Letters 2005, 59, 2519.
60
Chapter 4
Assembly of Colloidal Particles at Gas-Water and Water-Oil
Interfaces
Reproduced with permission from Angewandte Chemie-International Edition 2009, 48,
5300. Copyright 2009, Wiley-VCH, and Journal of the American Chemical Society 2008,
130, 16508. Copyright 2008, American Chemical Society.
Small colloidal particles with size range from a few nanometers to micrometers and
appropriate wettability strongly adsorb to gas–liquid and liquid-liquid interfaces. Such
colloidal particles can be used as stabilizers of bubbles and droplets. Interesting materials,
including colloidosomes and anisotropic and porous particles, have been prepared by
assembling particles at fluid interfaces. In this chapter, we describe a microfluidic strategy
for the continuous, single-step production of particle-coated bubbles with a predetermined
size and narrow size distribution. In addition, this chapter demonstrates a microfluidic
“inside-out” approach for the generation of monodisperse water-in-oil and oil-in-water
Pickering emulsions in which the coverage of droplets with particles can be precisely
controlled by hydrodynamic means.
4.1 Introduction
Particle-stabilized bubbles and droplets (Pickering emulsion) have recently seen a
surge in interest owing to their high stability against coalescences and Ostwald ripening.
61
These systems had a broad range of applications in the fabrication of functional materials, e.
g., hollow permeable capsules,1 anisotropic structures2 and hybrid supracolloidal
assemblies.3 Adsorption of particles at a particular fluid interface occurs when they are not
completely wet by any of the adjacent phases.4 The wettability of the surface of particles
determines their location at the fluid interface. The hydrophilic particles largely reside in
water. On the other hand, they favourably stay in gas or oil when the particles are
hydrophobic. More importantly, the particles attached to the interface build up the energy
barrier required to remove them into one of the fluid phases. 4,5 This results in irreversible
adsorption of particles at the interface in comparison with conventional surfactants which
adsorb and desorb on the time scale of a few tens of miliseconds.6 Consequently, particle-
stabilized bubbles and droplets are stable against coalescence. In addition, Ostwald ripening
is suppressed by reducing the Laplace pressure-induced dissolution of bubbles and
droplets.4 This effect is at maximum when the ratio of the radius of particle to that of
bubble or droplet is approximately 0.1, which allows bubble or droplet to adopt faceted
polyhedral shape. At this shape, the Laplace pressure is eliminated by reducing the mean
surface curvature close to zero.7
Current production of particle-stabilized bubbles and droplets relies on injection
methods, or shear of a mixture of two immiscible fluids.8-10 These methods generate
polydisperse bubbles/droplets in an uncontrollable manner, thus hampering the fabrication
of materials with hierarchical periodic structures and the related interesting optical and
mechanical properties. Recently, hydrodynamic flow was used for the controlled assembly
of colloidal particles at gas-liquid and liquid-liquid interfaces.11 The strategy relied on the
shear-assisted delivery of colloidal particles to the surface of bubbles or droplets. The
62
produced bubbles or droplets with colloidal armour had a narrow size distribution, however
the method had a low productivity of ca. 10 bubbles (droplets)/s.
In this chapter, we present a microfluidic (MF) method for the controlled production of
particles-coated bubbles12and droplets.13 First, we describe a chemically mediated MF-
based approach for the continuous production of bubbles encapsulated with a shell of
colloidal particles (armoured bubbles) (Chapter 4.2.1).12 The strategy relies on several
events, occurring in succession: i) a MF generation of monodisperse CO2 bubbles in a
dispersion of anionic particles, ii) dissolution of CO2, resulting in the shrinkage of the
bubbles and a local increase in the acidity of the solution around the bubbles, and iii)
adsorption of the particles to the gas-liquid interface, driven by the chemically induced
change in the surface energy of the particles. The dimensions and shapes of bubbles with
colloidal armours are precisely controlled by hydrodynamic and chemical means. In
addition, the generality of the MF strategy is demonstrated by encapsulating bubbles with a
variety of particles and their mixtures.
Secondly, we describe a MF “inside-out” approach to the generation of monodisperse
water-in-oil and oil-in-water Pickering emulsions. The emulsions were used for the
generation of supracolloidal polymer microspheres (Chapter 4.2.2).13 This method
addresses the challenges frequently occurring in conventional methods of the preparation of
Pickering emulsions, that is, an insufficient control of droplet dimensions and an excess of
particles in the continuous media. The proposed method minimizes waste of particles, due
to their introduction in the droplet phase, and allows control of the coverage of droplets
with particles by tuning the concentration of particles or the flow rates of the liquids.
63
4.2 Results and Discussion
The materials and methods described in the present chapter are provided in Chapter 2.
4.2.1 Synthesis of Colloidal Particle Armoured Bubbles
4.2.1.1 Experimental Design
We used a planar MF T-junction device to produce plugs of CO2 bubbles in an aqueous
dispersion of anionic polymer particles (Figure 4.1A).14 A dispersion of poly(styrene-co-
acrylic acid) (PS-co-PAA) particles15 in an aqueous NaOH solution was supplied to the
microchannel at a typical flow rate, QL, from 7.5 to 24 mL/h. Gaseous CO2 comprising 0.2
vol.% of inert gases was introduced into the orthogonal channel by applying CO2 pressure
(PCO2) as 34.5 kPa.
Figure 4.1 (A) A schematic of a microfluidic (MF) T-junction bubble generator. The widths of the main and the side channels are 220 and 40 μm, respectively. The height of the channels is 130 μm. (B) A schematic of the formation of colloidal particle shell during the dissolution of CO2 bubbles.
64
In this experimental design, we hypothesized the dissolution of CO2 bubble would increase
the acidity of the medium adjacent to the bubbles. This would result in the formation of
bubbles armoured with PS-co-PAA particles due to the reduction in the surface energy of
the particles (Figure 4.1B, see below).
4.2.1.2 Generation of Armoured Bubbles
Figure 4.2 illustrates the formation of CO2 bubbles, their time-dependent dissolution
and adsorption of the particles to the gas-water interface. The thread of CO2 periodically
broke up to release gaseous plugs with a polydispersity of 2-5% (Figure 4.2A). Owing to
the rapid dissolution of CO2, the bubbles underwent a dramatic decrease in volume and
acquired a spherical shape (Figure 4.2B). While the bubbles were shrinking, the particles in
the continuous phase adsorbed to the gas-water interface, due to the local increase in acidity
in the area adjacent to the bubbles (Figure 4.2C). The bubbles were completely covered
with the particles in the downstream channel (Figure 4.2D). The entire process took place
around 2 sec.
Figure 4.2 Progression of the plugs to spherical armoured bubbles. Optical microscopy images of bubbles at a distance of 0 (A), 30 (B), 80 (C) and 150 (D) mm from the T-junction. Scale bars are 200 μm. Bubbles generated at pH=14, PCO2=34.5 kPa, QL=10.5 mL/h, initial particle concentration, Cp=1.5 wt. %, and 28.8oC.
65
Figure 4.3A shows a typical image of the bubbles collected at the outlet of the MF
device. A uniform dissolution and mass transfer achieved in the microchannels yielded
armoured bubbles with a narrow size distribution. On the bubble surface the particles
formed a close-packed 2D crystalline shell (Figure 4.3B). Imaging of microtomed bubbles
using Scanning Electron Microscopy (SEM) confirmed that the particles formed a
monolayer-thick shell (Figure 4.3C). Labeling of the polymer particles with a fluorescent
dye Hostasol Yellow (HY) allowed visualization of the colloidal shell using Laser Confocal
Fluorescence Microscopy (LCFM) (Figure 4.3C, inset).
Figure 4.3 (A) Optical microscopy image of armoured bubbles generated as in Figure 4.2 and collected at the outlet of the MF device. Scale bar is 200 μm (B) Optical microscopy image of close-packed crystalline shell of armoured bubble. Scale bar is 25 μm (C) SEM image (side view) of the shell of the microtomed armored bubble infiltrated with poly(ethylene glycol) diacrylate. Scale bar is 10 μm. The inset shows an LCFM image of the bubble coated with HY-labeled PS-co-PAA particles. Scale bar is 25 μm.
The armoured bubbles were stable to coalescence and Ostwald ripening, and had
narrow polydispersity of 2-5%, similar to the initially generated bubbles. The frequency of
the generation of armoured bubbles was up to ca. 700 bubbles/s; however the productivity
of the method was increased up to ca. 3000 bubbles/s by using a MF flow-focusing bubble
generator (Appendix, Figure A4.1).16
We conducted a series of control experiments. In a particle-free environment (Figure
4.4A) the shrinkage of CO2 plugs was similar to that in the particle suspension (Figure 4.2).
66
This indicated that at this pH value the size of bubbles was determined by the dissolution of
CO2 and not by the formation of the colloidal shell. The role of CO2 and carboxylated
particles was confirmed by showing that the armoured bubbles did not form in experiments
conducted with PS-co-PAA particles and N2 bubbles (Figure 4.4B), or with PS-co-PAA
particles and CO2 bubbles in the presence of surfactants (Figure 4.4C). Surfactants changed
the delicate balance in the surface energy of the particles at the gas-liquid interface. In
another series of control experiments, N2 and CO2 bubbles were not fully covered with
cationic poly(methyl methacrylate-co-4-vinyl pyridine) (PMMA-co-P4VP) beads (Figure
4.4D-E, respectively). These bubbles were prone to coalescence and they acquired a broad
size distribution when collected at the outlet of the MF device (Figure 4.4F-G).
Figure 4.4 Optical microscopy images of bubbles and plugs flowing through the microchannels. (A) Dissolution of CO2 plugs in particle-free at pH=14. (B) Plugs of gaseous N2 formed in a dispersion of anionic PS-co-PAA particles. (C) Bubbles of CO2 generated in a dispersion of PS-co-PAA particles containing 2 wt. % of the non-ionic surfactant (Triton X-100). (D) Plugs of N2 formed in a dispersion of 700 nm diameter cationic PMMA-co-P4VP particles. (E) Bubbles of CO2 dispersed in an aqueous dispersion of 700 nm diameter PMMA-co-P4VP particles. In all experiments, the bubbles were generated at 28.8oC, pH=14, PCO2/N2=34.5 kPa, QL=10.5 mL/h, and Cp=1.5 wt. %. Optical microscope images of collected bubbles (F),(G) from (D),(E), respectively. Scale bars: 200 μm.
67
We explained the formation of armoured bubbles as follows. The dissolution of CO2 in
an aqueous phase was followed by the reactions.17
pH<10: CO2 + H2O HCO3- + H+ (K = 4.4 x 10-7) (eq. 4.1)
pH>10: CO2 + OH- HCO3- (K = 3.2x 107) (eq. 4.2)
These reactions (eq. 4.1-2) led to a local decrease in the pH of the liquid adjacent to
the surface of bubbles, the protonation of the carboxylic groups on the surface of PS-co-
PAA particles and an increase of the contact angle (θ) of water on the polymer surface, as
illustrated in Figure 4.5.
Figure 4.5 Variation in contact angle (θ) of an aqueous solution of NaOH in air on the PS-co-PAA film, measured at different pH values.
The protonation of the surface carboxylic groups on the particles in the region of the
reduced pH favoured adsorption of the microbeads to the gas-water interface. The energy
barrier required for the removal of particles from the interface was given by
( )2LG,2 cos1 θγπ ±= paE (eq. 4.3)
where γG,L is the interfacial tension of the gas-liquid interface and ap is the radius of the
particles. The sign inside the bracket is negative for removal of particles into the water
68
phase, and positive for removal of particles into the gas phase.4 We stress that the particles
were hydrophobized only in the vicinity of the bubbles. In the rest of the continuous phase,
the particles retained their colloidal stability.18
4.2.1.3 Effect of Flow Rate on the Dimensions of Armoured Bubbles
Figure 4.6 shows the effect of the flow rate of the continuous phase, QL, on the final
diameter, Df, of the armoured bubbles at initial pH=14. At the largest values of QL of 10.5
mL/h, the fractional reduction in volume, ΔV/V0 (V0 and ΔV are the initial volume and the
change in volume of the plugs, respectively), of the plugs was up to 99.5%. This indicated
almost all the CO2 left the gaseous plugs (Figure 4.6A). Upon increasing QL, the Df of the
armoured bubbles decreased (Figure 4.6B,C). This effect occurred due to the generation of
smaller initial gaseous plugs16 (Figure 4.6B,C insets) and a more efficient dissolution of
CO2 achieved at high values of QL.19
Figure 4.6 (A) Effect of the flow rate, QL, of the continuous phase on the fractional volume change (ΔV/V0) of CO2 plugs and the final diameter Df of armoured bubbles in the microchannel. The produced armoured bubbles at QL= 8.5 (B) and 10.5 mL/h (C). Scale bars are 50 μm. The insets show the initial plugs of CO2 bubbles. Scale bars are 200 μm. Bubbles were generated at pH=14, PCO2=34.5 kPa and Cp=1.5 wt. %.
69
4.2.1.4 Effect of pH and Particle Concentration on the Dimension and
Morphologies of Armoured Bubbles
Figure 4.7A shows the effect of the initial pH of the continuous phase on the Df of the
bubbles and the fractional change in volume ΔV/V0, of gaseous plugs at an initial particle
concentration, CP, of 1.5 wt %. With increasing pH a more rapid dissolution of CO2 yielded
smaller bubbles that were uniformly covered with particles (Figure 4.6B-D). At pH=14 the
value of ΔV/V0 reached 99.5% while it was 40 % at pH=5 (Figure 4.7A). In the
experiments conducted at pH=14 in a particle-free environment, the CO2 plugs underwent a
similar change in volume, suggesting that at this pH value the size of bubbles was not
affected by adsorption of particles. Thus, we conclude that the final volume of the bubbles
was determined by the volume of the remaining inert gas.
Figure 4.7 (A) Effect of the initial pH of the continuous phase on the fractional volume reduction (ΔV/V0) of CO2 plugs and the final diameter, Df, of armoured bubbles in the microchannel. The produced armoured bubbles at initial pH=14 (B) 8 (C) and 5 (D). Scale bars are 100 μm. The insets show the bubbles before the exit of the microchannel. Scale bars are 200 μm. Bubbles were generated at PCO2=34.5 kPa, QL=9.0 mL/h and Cp=1.5 wt. %.
70
For pH<10, the bubbles retained their plug-like shape and were only partly covered
with particles (Figure 4.7C,D, insets). This effect occurred because the weaker dissolution
of CO2 at this range of pHs did not produce small enough surface area of bubbles to be
fully covered with the particles.19,20 Owing to the incomplete coverage with the particles,
the plugs were prone to coalescence. When collected at the outlet, they formed large, non-
spherical armoured bubbles due to the jamming of the particles at the interface (Figure
4.7C,D).
Figure 4.8 shows that with increasing value of Cp the final size of the armoured
bubbles increased (the initial dimensions of the CO2 plugs did not depend on CP). Rapid
formation of the colloidal shell counteracted the shrinkage of the plugs, thereby providing
additional control over the dimensions of armoured bubbles (Figure 4.8A). At CP <1 wt. %
the bubbles were poorly covered with particles and at the outlet of the MF device they
coalesced to form large armoured bubbles with a broad size distribution (Figure 4.8B). For
the concentration of particles in the range 1.5≤CP /wt. %≤5.0 (1 wt. % ~ 4.3 x 108
particles/mL) the colloidal armour suppressed coalescence of the bubbles, whose
polydispersity remained below 5% (Figure 4.8C,D).
71
))()(4()(
))()(4()(
22
223
tnatRqdt
tdn
tnatRqdt
tdR
ppbpp
ppbbb
αππ
βαππ
−=
−−=
Figure 4.8 (A) Effect of the initial concentration of particles, Cp, of the continuous phase on the fractional volume reduction, ΔV/V0, of CO2 plugs and the final diameter, Df, of armoured bubbles in the microchannel. The produced armoured bubbles at initial Cp =0.1 (B) 1.5 (C) and 5 wt. % (D). Scale bars are 100 μm. The insets show the bubbles before the exit of the microchannel. Scale bars are 200 μm. Bubbles were generated at PCO2=34.5 kPa, QL=10.0 mL/h and pH=14.
We developed a model that rationalized our observations and accounted for the
shrinkage of bubbles and the adsorption of particles to the gas-liquid interface. These two
processes were coupled: dissolution of CO2 occurred only through uncovered regions of the
gas-liquid interface.
For a bubble of radius Rb(t) and the number of adsorbed particles np(t), the uncovered
bubble area was described as (4πRb(t)2−απap2np(t)) where α depends on the contact angle.
The two equations have the form
(eq. 4.4)
(eq. 4.5)
72
BtdntdR
p
b −=)()( 3
22
330 4
peq
eq
aB
R
RR=
−
where the constant β≈1 accounts for the arrest of CO2 dissolution once the particles are
close-packed.7 The fluxes of gas and particles, qb and qp respectively, depend on the speed
U and the radius (a) of the bubble. For large Peclet numbers, URb(t)/D>>1, where D is the
diffusion coefficient of dissolved CO2, we expect qb and qp to scale similarly with U and
Rb(t). For such transport processes qp∝ Cp, analogous to Fick’s law. The bubbles shrink
faster under high pH conditions, since the diffusion of dissolved CO2 in the liquid boundary
layer is accompanied by the reaction with OH- ions. Thus, we expect qb to increase with
increasing pH. Neglecting the weak dependence of qb on β, the ratio of the two equations
(eq. 4.4-5) yields
where B is constant at a fixed pH or Cp. It decreases as Cp increases and it increases as the
pH increases. Thus, for the initial conditions Rb(0)=R0 and np(0)=0, the equilibrium bubble
radius Req is given by
Thus, the value of Req decreases monotonically with increasing B. Since the value of B
increases with high pH values, it can be expected that the bubble radius will decrease as the
pH increases, which is in agreement with the experimental trends shown in Figure 4.7A.
Further, based on eq. 4.7, the bubble radius should increase as the Cp increases (decreasing
B), which is consistent with experimental observations (Figure 4.8A).
(eq. 4.7)
(eq. 4.6)
73
4.2.1.5 Generality of the Assembly of Colloidal Particles at Gas-Water
Interface
The generality of our approach for the production of particle-coated bubbles was
demonstrated by coating bubbles with different types of anionic particles and mixtures of
anionic particles. Figure 4.9A shows bubbles encapsulated with a mixture of HY dye-
labelled and dye-free PS-co-PAA microspheres with similar dimensions. Figure 4.9B
illustrates bubbles with a shell of carboxylated silica particles and a mixture of
carboxylated silica particles and HY dye-labelled PS-co-PAA (Figure 4.9B inset). Bubbles
encapsulated with 20 nm-diameter carboxylated silica nanoparticles capped CdSe/ZnS
quantum dots (QDs) are shown in Figure 4.9C. We also encapsulated bubbles with the
protein bovine serum albumin labelled with fluorescein isothiocyanate (FITC-BSA), which
was dissolved in an aqueous phase at pH=7. Following the dissolution of CO2, the bubbles
were coated with protein particles, which were 1-2 μm in size (Figure 4.9D). We believe
that this observation indicates that in the acidic environment adjacent to the shrinking
bubbles the protein molecules reached their isoelectric point of pH=4.8 and aggregated to
form clusters precipitating on the bubble surface.21
74
Figure 4.9 Generation of bubbles with various types of colloidal armour. (A) Optical fluorescence microscopy image of bubbles coated with a shell of 2.8 μm diameter HY-labelled PS-co-PAA particles (Cp=0.5 wt. %) and 3.5 μm diameter PS-co-PAA particles (Cp=1.0 wt. %), PCO2=41.4 kPa, QL=12 mL/h, pH=14. Scale bar is 50 μm. (B) Optical microscopy image of bubbles coated with 3 μm diameter carboxylated silica particles. PCO2=82.7 kPa, QL =24 mL/h, pH=14, and Cp=1.5 wt. %. Scale bar is 100 μm. The inset shows a fluorescence microscopy image of the surface of an armoured bubble coated with 3 μm diameter carboxylated silica particles and 2.8 μm diameter HY-labelled PS-co-PAA polymer particles in the continuous phase in the weight ratio 3:1, respectively. Scale bar is 25 μm. (C) LCFM image of bubbles encapsulated with 20 nm diameter carboxylated silica nanoparticles loaded with CdSe/ZnS QDs, λex=480 nm. PCO2=55.2 kPa, QL=23 mL/h, pH= 10, and Cp =0.12 wt. %. Scale bar is 100 μm. (D) Fluorescence optical microscopy image of armoured bubbles engulfed with the FITC–BSA shell, λex=495 nm. The bubbles were generated at PCO2=44.8 kPa, QL=13 mL/, pH=7, and a protein concentration of 0.02 wt. %. Scale bar is 100 μm.
4.2.2 Synthesis of Colloidal Particle-Coated Droplets
Research activities in the field of Pickering emulsions have led to the production of a
wide range of supracolloidal materials.1-3 For a particular system, the concentration of
particles required for the efficient stabilization of droplets depends on the ratio between the
dimensions of droplets and particles, and the number of particles at the fluid interface.20,26
75
Current methods for producing particle-coated emulsions generate droplets with a broad
distribution in sizes,8-10 which hampers the precise rationalization of the amount of particles
required for the effective system. We demonstrate a MF “inside-out” approach to the
production of monodisperse water-in-oil and oil-in-water Pickering emulsions, as well as
the supracolloidal polymer microspheres.13 The MF emulsification of a dispersion of
colloidal particles in the particle-free continuous phase allowed the control of the coverage
of the droplets with a layer of solid particles in a pre-determined way.
4.2.2.1 Experimental Design
The generation of particle-coated droplets was conducted in a MF flow-focusing
device (Figure 4.10A).16 Droplets of a water-ethanol (85/15 v/v) mixture containing 3.5
μm-diameter poly(divinylbenzene-co-methacrylic acid) (poly(DVB-co-MAA)) particles22
was emulsified in hexadecane as the continuous phase. The generation of droplets occurred
at a flow-focusing regime, where the dispersed thread was focused and broken up in the
orifice by the continuous phase.16 The polydispersity of produced droplets was below 5 %.
The value of the contact angle (θ) between the polymer film derived from the particles and
the water-ethanol mixture was 82.2±2.1o. Thus, the microbeads migrated from the interior
of the droplets to the surface of droplets. Figure 4.10B shows the typical microscope image
of the collected particle-coated droplets at the exit of the outlet.
76
Figure 4.10 (A) Generation of droplets from the water-ethanol mixture (85/15, v/v) containing 4 wt. % of poly (DVB-co-MAA) particles. The flow rates of the droplet and continuous phases are 0.5 and 3.5 mL/h, respectively. Scale bar is 200 μm. (B) Collected particle-coated droplets at the exit of the microchannel. Scale bar is 150 μm.
4.2.2.2 Microfluidic Control of Particle Coverage at Water-Oil Interfaces
Since the diameter of particles used in the present work was significantly smaller than
the diameter of droplets, we estimated the surface coverage of the droplets,δ, was
simplified as
δ = Ap/Ad = Cp⋅ρd⋅ad/(4ρp⋅ap) (eq. 4.8)
where Ad and Ap are the surface areas of the droplet and of the droplet coated with particles,
respectively and ρd and ρp are the densities of the droplet and the particles, respectively.
In Figure 4.11, the dashed line shows the estimated variation in the concentration of
particles required to achieve complete coverage of the droplets with different sizes. We
77
assumed that (i) all particles migrate from the droplet interior to the droplet surface and (ii)
at the interface the particles form a hexagonal lattice with a packing density of 0.906. The
surface area of the droplet coated with particles, Ap, is proportional to the volume of the
droplet, ad3. On the other hand, the surface area of the droplet, Ad, increases with ad
2. Thus,
droplets with larger dimensions require a lower concentration of particles to achieve the
complete coverage of particles. In Figure 4.11, below the dashed line, the amount of
particles is not sufficient for the complete coverage of the droplets; whereas above the line,
the particles exist in excess to form a multilayer shell or to remain in the droplet interior.
Figure 4.11 A diagram of the surface coverage of water-ethanol droplets with 3.5 μm-diameter poly(DVB-co-MAA) particles. The dashed line shows the theoretical conditions for the complete coverage of the droplets with a monolayer of 2D hexagonally packed particles. Filled circles show the experimental data points.
Experimentally, we controlled the surface coverage of the droplets by varying
independently the value of Cp from 4 to 16 wt. % and the values of ad from 40 to 80 μm (by
tuning the ratio of flow rates of the droplet-to-continuous phases).23 Filled symbols in
Figure 4.11 show the experimental results. Droplets with the surface coverage of δ ≥0.7
78
(ad>63 μm, Cp=8 wt. %) were stable to coalescence, whereas droplets with δ <0.7 (ad≤63
μm, Cp=8 wt. %) were prone to coalescence when collected at the exit of the MF device.
The attachment of particles from the droplet phase to the liquid-liquid interface and the
formation of the close-packed crystalline shell occurred rapidly, being assisted by the
hydrodynamic flow.11 Owing to the very rapid particle jamming at the interface at high
values of Cp, we were able to produce non-spherical particle-coated droplets with the shape
determined by the geometry of the microchannel (Figure 4.12).24,25
Figure 4.12 Plug-shaped particle-covered droplets. Scale bar is 100 µm. Inset shows the corresponding elliptical droplets flowing in the microchannel. Scale bar is 200 μm. Cp=14 wt. %. The flow rate of the droplet and continuous phases are 0.5 and 5.5 mL/h, respectively.
In Figure 4.13, representative optical microscopy (top) and LCFM (bottom) images
show that for Cp=8 wt. % and increasing the size of the droplets, a transition occurred from
the non-complete coverage to the complete coverage, and then to the excess-of-particles
regime. The non-complete coverage of the surface of droplets with polymer particles is
79
illustrated in Figure 4.13A,D. Figure 4.13 B,E shows the complete coverage of the droplet
with particles. With the values of Cp and ad predicted by the diagram in Figure 4.11, the
microbeads formed a close-to-hexagonal lattice on the droplet surface (Figure 4.13B, inset).
In the excess-of-particles regime, the microbeads were concentrated at the interior of
droplet, which was visualized by the darker concentrated region of droplet (Figure 4.13C).
In the LCFM image, this region appeared brighter in the interior of the droplets (Figure
4.13F).
Figure 4.13 (A-C) Optical and (D-F) LCFM images of the water-ethanol droplets armoured with a shell of poly(DVB-co-MAA) particles at the not-complete (A,D) and complete (B,E) surface coverage, and at the excess of particles in the droplet interior (C,F). In (C) an excess of particles appears as the large dark region on the background of the droplet coated with a monolayer of particles. In (D-F) poly(DVB-co-MAA) microbeads were labelled with anthryl methacrylate. LCFM images show the plane located in the centre of the droplets. Cp =8 wt. %. λex=380 nm. Scale bars in (A-E) and in (F) are 50 and 100 μm, respectively. Scale bars in insets are 5 μm.
The MF approach was also used for the preparation of oil-in-water Pickering emulsions
and the corresponding supracolloidal structures. The droplet phase contained a mixture of
tripropylene glycol diacrylate, TPGDA, a photoinitiator, 2,2-dimethoxy-2- phenyl-
80
acetophenone (4 wt. %), and poly(DVB-co-MAA) particles at Cp=14 wt. %. Following the
deposition of particles at the surface of droplets, they were collected at the exit of the MF
device and exposed for 1 min to UV-irradiation (400W, λ=330-380 nm).
Photopolymerization of TPDGA yielded large polymer microspheres armoured with a shell
of poly(DVB-co-MAA) particles. Figure 4.14A shows the surface of the supracolloidal
microsphere produced at Cp =14 wt% (approximately 40% higher than required to achieve
the complete surface coverage of δ of 0.906). The average diameter of poly(DVB-co-
MAA) particles protruded on the surface of microsphere was 3.0 μm (the original diameter
of the particles was 3.5 μm). The center-to-center distance between the particles was 4.2
μm of (Figure 4.14A). This indicated that (i) a larger fraction of the microbead surface
resided in the droplets than in the continuous phase and (ii) approximately 0.7 μm mean
distance existed between the surfaces of the particles immersed in the droplet phase.26
Inspection of the colloidal shell using LCFM revealed that poly(DVB-co-MAA) particles
formed a multilayer on the surface of the microspheres (Figure 4.14B), presumably due to
the long range attraction between the microparticles in the TPGDA medium.27
Figure 4.14 SEM (A) and LCFM (B) images of poly(TPGDA) particles armoured with poly(DVB-MAA) particles. Cp=14 wt. %. Scale bars are 40 μm in (A) and 60 μm in (B).
81
4.3. Conclusion
In summary, we have demonstrated a MF approach to the preparation of particle-
encapsulated bubbles and droplets. Our method yielded bubbles with high stability against
coalescence and Ostwald ripening. It increased productivity in comparison with other MF
approaches. It demonstrated the ability to form bubbles with precisely controlled
dimensions. The armoured bubbles generated by the proposed strategy have potential
applications in producing 3-D foams with hierarchical order and ultralight 2-D coatings
with precisely controlled pore sizes. Furthermore, standard methods for the generation of
small bubbles are either cost-inefficient,28 or lacks control over bubble size distribution. In
our approach, these problems are solved by (i) producing large monodisperse bubbles from
gaseous mixtures and (ii) controllably removing one of the components of the mixture to
reach the targeted bubble size. The described strategy can be extended in several ways.
First, the deposition of cationic particles can be achieved by dissolving, e.g. NH3 and
increasing the pH in the region adjacent to the surface of the bubbles. Second, an increase
in the productivity of the process can be achieved by using parallel integrated MF bubble
generators.29,30 Third, controllable transfer of one of the components of the gaseous
mixture to the continuous phase can provide the ability to control the composition of the
bubbles and of the surrounding medium and to activate a particular reaction at a particular
timing, on demand.
A MF “inside-out” approach for the preparation of particle-encapsulated droplets has
the following features: (i) high efficiency and a minimized quantity of particles needed for
the generation of Pickering emulsions , (ii) a narrow size distribution of droplets, which
provides a route for easier analysis of the dynamics of formation and buckling of Pickering
82
emulsions, as well as colloid crystallization on curved fluid interfaces, and (iii) the ease of
the control over the coverage of droplets with particles by manipulating the concentration
of particles or the hydrodynamic conditions of fluids. We anticipate that the present
approach will pave a way to the controlled preparation of colloidal materials with complex
structural hierarchy.31
References
(1) A. D. Dinsmore, M. F. Hsu, M. G. Nikolaides, M. Marquez, A. R. Bausch, D. A.
Weitz, Science 2002, 298, 1006.
(2) B. Liu, W. Wei, X. Z. Qu, Z. H. Yang, Angewandte Chemie-International Edition
2008, 47, 3973.
(3) T. Chen, P. J. Colver, S. A. F. Bon, Advanced Materials 2007, 19, 2286.
(4) B. P. Binks, Current Opinion in Colloid & Interface Science 2002, 7, 21.
(5) P. Pieranski, Physical Review Letters 1980, 45, 569.
(6) J. Eastoe, A. Rankin, R. Wat, C. D. Bain, International Reviews in Physical
Chemistry 2001, 20, 357.
(7) M. Abkarian, A. B. Subramaniam, S. H. Kim, R. J. Larsen, S. M. Yang, H. A. Stone,
Physical Review Letters 2007, 99.
(8) B. P. Binks, T. S. Horozov, Angewandte Chemie-International Edition 2005, 44,
3722.
(9) U. T. Gonzenbach, A. R. Studart, E. Tervoort, L. J. Gauckler, Angewandte Chemie-
International Edition 2006, 45, 3526.
(10) R. G. Alargova, D. S. Warhadpande, V. N. Paunov, O. D. Velev, Langmuir 2004,
20, 10371.
(11) A. B. Subramaniam, M. Abkarian, H. A. Stone, Nature Materials 2005, 4, 553.
(12) J. I. Park, Z. Nie, A. Kumachev, A. I. Abdelrahman, B. R. Binks, H. A. Stone, E.
Kumacheva, Angewandte Chemie-International Edition 2009, 48, 5300.
83
(13) Z. H. Nie, J. Il Park, W. Li, S. A. F. Bon, E. Kumacheva, Journal of the American
Chemical Society 2008, 130, 16508.
(14) T. Thorsen, R. W. Roberts, F. H. Arnold, S. R. Quake, Physical Review Letters
2001, 86, 4163.
(15) J. S. Song, L. Chagal, M. A. Winnik, Macromolecules 2006, 39, 5729.
(16) P. Garstecki, I. Gitlin, W. DiLuzio, G. M. Whitesides, E. Kumacheva, H. A. Stone,
Applied Physics Letters 2004, 85, 2649.
(17) P. V. Danckwerts, Gas-Liquid Reactions, McGraw-Hill, New York 1970.
(18) B. P. Binks, R. Murakami, S. P. Armes, S. Fujii, A. Schmid, Langmuir 2007, 23,
8691.
(19) T. Madhavi, A. K. Golder, A. N. Samanta, S. Ray, Chemical Engineering Journal
2007, 128, 95.
(20) K. Golemanov, S. Tcholakova, P. A. Kralchevsky, K. P. Ananthapadmanabhan, A.
Lips, Langmuir 2006, 22, 4968.
(21) X. M. Qi, S. J. Yao, Y. X. Guan, Biotechnology Progress 2004, 20, 1176.
(22) W. H. Li, H. D. H. Stover, Journal of Polymer Science Part a-Polymer Chemistry
1999, 37, 2899.
(23) J. D. Tice, H. Song, A. D. Lyon, R. F. Ismagilov, Langmuir 2003, 19, 9127.
(24) D. Dendukuri, K. Tsoi, T. A. Hatton, P. S. Doyle, Langmuir 2005, 21, 2113.
(25) S. Q. Xu, Z. H. Nie, M. Seo, P. Lewis, E. Kumacheva, H. A. Stone, P. Garstecki, D.
B. Weibel, I. Gitlin, G. M. Whitesides, Angewandte Chemie-International Edition 2005, 44,
724.
(26) T. S. Horozov, R. Aveyard, B. P. Binks, J. H. Clint, Langmuir 2005, 21, 7405.
(27) M. E. Leunissen, A. van Blaaderen, A. D. Hollingsworth, M. T. Sullivan, P. M.
Chaikin, Proceedings of the National Academy of Sciences of the United States of America
2007, 104, 2585.
(28) K. Hettiarachchi, E. Talu, M. L. Longo, P. A. Dayton, A. P. Lee, Lab on a Chip
2007, 7, 463.
(29) W. Li, J. Greener, D. Voicu, E. Kumacheva, Lab on a Chip 2009, 9, 2715.
(30) T. Nisisako, T. Torii, Lab on a Chip 2008, 8, 287.
84
(31) S. M. Yang, S. H. Kim, J. M. Lim, G. R. Yi, Journal of Materials Chemistry 2008,
18, 2177.
85
Chapter 5
Bubbles Encapsulated with a Shell of Biopolymers
Reproduced with permission from Macromolecular Rapid Communications, 2010, 31,
222. Copyright 2010, Wiley-VCH.
This chapter describes a microfluidic route to producing small (<10 μm) bubbles with a
narrow size distribution and a long-term stability. The bubbles were encapsulated with a
protein-polysaccharide shell. The strategy includes the following events occurring in
sequence: i) microfluidic generation of bubbles comprising a mixture of CO2 and a minute
amount of gases with limited solubility. The bubbles were dispersed in an aqueous solution
of lysozyme and alginate; (ii) the dissolution of CO2 leading to the shrinkage of bubbles
and an increase in acidity of the medium surrounding the bubbles; (iii) the deposition of
lysozyme at the gas-water interface and electrostatically-driven complexation of alginate
with lysozyme on the surface of bubbles.
5.1 Introduction
Bubbles with a diameter in the range from 1 to 100 μm that are encapsulated with
biopolymers (e.g, proteins, polysaccharides, or lipids) find applications in biomedical
research,1,2 and in the food and cosmetics industries.3,4 Recently, interesting applications in
simultaneous ultrasonic imaging and targeted drug delivery have emerged for small (<10
μm) bubbles.5 Ultrasound imaging originates from the ability of bubbles to improve the
86
scattering of applied ultrasound wave.6 The drug delivery function is achieved by
intravenously injecting bubbles carrying a drug in the shell and applying a focused
ultrasound excitation at a particular site. The disruption of the bubbles leads to the release
of the drug.7 The requirements to bubbles used for biomedical purposes include long-term
stability, biocompatibility and control over bubble dimensions and size distribution. The
most challenging requirement is the long-term stability of bubbles. Generally, bubbles
continuously change their dimensions due to the intrinsic kinetic and thermodynamic
instabilities. Coalescence of bubbles occurs when the thin film separating two adjacent
bubbles ruptures.10 Dissolution of bubbles occurs because surface tension at the gas-liquid
interface generates the excess Laplace pressure, ΔP = 2σ/r, where ΔP is the difference in
pressure inside and outside the bubble, σ is the surface tension at the gas-liquid interface,
and r is the radius of bubble. Since the value of ΔP rapidly increases as r decreases, the
chemical potential of the bubbles increases and so does the solubility of the gas. Therefore,
bubbles with smaller sizes have a higher propensity to dissolution. Furthermore, bubbles
with a broad size distribution undergo Ostwald ripening: due to the higher solubilty, gas
molecules from smaller bubbles dissolve and diffuse into larger bubbles. This leads to the
growth of larger bubbles at the expense of the shrinkage of smaller bubbles.11
Generally, the stabilization of bubbles is achieved by adsorbing surfactants to their
surface, in order to achieve electrostatic or steric repulsion between the bubbles and to
decrease the value of interfacial tension (σ). Although conventional low-molecular weight
surfactants provide a barrier to coalescence, they do not prevent bubble dissolution due to
their inability to form a robust surface layer on the bubble surface. In addition, surfactants
rapidly desorb from the gas-liquid interface.12 (we note that it was recently reported that
87
patches of sucrose stearate crystallized on the bubble surface can stop gas dissolution,
thereby providing long term bubble stability.13)
Alternatively, bubbles can be stabilized with lipids or biopolymers such as proteins or
polysaccharides.14 Typically, bubbles encapsulated with biological species are prepared by
sonication or mechanical agitation.8,9 These methods produce bubbles with dimensions in
the size range from one micrometer to tens of micrometers and a broad distribution of
size.1,8,9
Due to their amphiphilic nature, proteins rapidly adsorb to the gas-water interface
where they undergo partial unfolding and form the interfacial layer with high a mechanical
strength.14,15 In practice, however, protein layers alone do not provide a sufficiently strong
barrier to the dissolution of bubbles, unless protein unfolding is enhanced by adding
chemical agents (e.g. reducing agents) or by applying heat.11,14,16,17
Significant enhancement of bubble stability is achieved by co-adsorbing on the bubble
surface of a mixture of proteins and polysaccharides. Electrostatically driven complexation
between cationic proteins and anionic polysaccharides increases the thickness and the
strength of the adsorbed layer at the gas-water interface.18,19 In this approach, the value of
pH and ionic strength of the medium, as well as the charge distribution on the biopolymer
molecules, play an important role in bubble stabilization.20
In this chapter, a microfluidic (MF) route to the generation of small, stable,
monodisperse bubbles encapsulated with a lysozyme-alginate shell is described. The
approach relies on three events occurring in sequence. (i) A MF generation of bubbles from
gaseous CO2 containing a small amount of gas with a low solubility in water. The bubbles
are produced in an aqueous solution of lysozyme and sodium alginate. (ii) A rapid and
uniform dissolution of CO2 leading to the reduction in bubble size and the decrease in pH
88
of the aqueous medium in the vicinity of the bubbles. (iii) The deposition of positively
charged lysozyme at the gas-water interface and its complexation with anionic sodium
alginate.
We show that a MF method provides the ability to generate bubbles with a controlled
size, narrow size distribution, and a high (> 1 month) stability when stored in a sealed
container. At a constant gas pressure the ultimate size of the bubbles is determined by the
flow rate of the continuous phase and the concentration of lysozyme and alginate. The
method has the productivity of up to 4x104 bubbles/min and in order to generate small
bubbles, it does not require a significant reduction of microchannel dimensions.
5.2 Results and Discussions
The materials and methods described in the present chapter are listed in Chapter 2.
5.2.1 Preparation of Bubbles Encapsulated with a Mixture of Biopolymers
Figure 5.1 shows the schematic of the formation of monodisperse bubbles encapsulated
and stabilized with a biopolymer shell. In Step 1, bubbles are generated from the gaseous
CO2 mixed with 0.2 vol % of the gas with low solubility in water. The bubbles are
generated in an aqueous solution of lysozyme and alginate at an initial pH of 12. At this pH,
both lysozyme and alginate are negatively charged.21 In Step 2, CO2 dissolves in the
continuous aqueous phase and the pH value of the solution surrounding the bubbles reduces.
The lysozyme molecules acquire a positive net charge (for this protein the isoelectric point
pI =11)22,23 and adsorb to the negatively charged surface of the bubbles.24 In Step 3, the
89
anionic alginate molecules (pKa= 3.5)25 form a complex with the positively charged
lysozyme on the surface of bubble. The complexation yields a biopolymer shell
encapsulating the shrunken bubbles. The ultimate size of the bubbles is controlled by their
original size, the extent of CO2 dissolution, the stabilizing effect of the biopolymer shell,
and the fraction of the low-water soluble gas in the gaseous CO2.
Figure 5.1 Schematic of co-adsorption of lysozyme and alginate on the surface of CO2 bubble during its dissolution.
Figure 5.2A shows a schematic of the planar MF flow-focusing device used to form
biopolymer-encapsulated bubbles. The CO2 bubbles were generated in the geometrically
controlled regime and had a discoid-like shape since the initial diameter of unperturbed
bubble was larger than the height of microchannel. Polydispersity of the bubbles was
below 5%.26 Figure 5.2B shows typical optical microscopy images of the bubbles, which
were acquired immediately after bubble formation in the aqueous solution of lysozyme and
sodium alginate (Figure 5.2B, top) and following bubble shrinkage in the downstream
microchannel ca. 3 sec after their formation (ca. 100 mm from the orifice of the bubble
90
generator) (Figure 5.2B, bottom). At this point, the dimensions of the bubbles were
stabilized due to the saturation of the continuous aqueous phase with CO2.27 By comparing
the initial and the stabilized volume of the bubbles, we estimated that the fractional
change in bubble volume was approximately 80 vol %.
Figure 5.2 A) Schematic of a MF flow-focusing bubble generator, B) Representative optical microscopy images of the generation (top) and dissolution of CO2 bubbles in a MF channel (bottom). The image is taken 250 mm away from the orifice of the MF device. The width and height of the orifice are 50 and 120 μm, respectively. PCO2=48.3 kPa, QL=6 mL/h. Clys=Calg=0.2 wt. %, pH=12. Scale bar is 200 μm.
Under strongly basic conditions used in the present work, the dissolution of CO2 is
governed by Henry’s law (eq. 5.1) and chemical reactions (eq. 5.2-3):28
[CO2]l kH PCO2 (eq. 5.1)
pH>10: CO2 + OH- HCO3- (K = 3.2x 107) (eq. 5.2)
HCO3- + OH- CO3
2- + H2O (K = 3.5x 103) (eq. 5.3)
where K is the equilibrium constant and [CO2]l, kH, and PCO2 are the concentration of
molecularly dissolved CO2, Henry’s law constant (for CO2 at 25oC kH=3.2x10-4 mol/(L
kPa)), and the gas pressure, respectively.
91
The dissolution of CO2 led to the acidification of the liquid surrounding bubbles.29
The aqueous solution collected at the exit of the MF device acquired pH of
approximately 8. The bubbles were collected at the outlet of the MF device and stored
in a sealed container.
5.2.2. Long-term Stability of Bubbles
The dimensions of collected bubbles in a sealed container further reduced within 2
h, due to the change of CO2 pressure to ambient condition (Figure 5.3A-C). In the
microchannel, CO2 bubbles experienced a pressure of 48.3 kPa, while partial CO2
pressure at 1 atm (outside the device) was 0.04 kPa.30 At this point, the fractional
change in bubble volume was 99.8 %, which implied that all CO2 was removed from the
bubbles and their size was determined by the remaining volume of low water soluble
impurities. Following this initial shrinkage, there was no noticeable change in the
dimensions of the bubbles during up to 4 weeks of storage in a sealed contained (Figure
5.3C-F).
To explore the role of pH reduction induced by CO2 dissolution, we conducted a
control experiment by generating N2 bubbles with a narrow size distribution in the
aqueous solution of lysozyme and alginate. After 24 h, the bubbles of N2 showed a broad
distribution of sizes dominated by their coalescence (Figure 5.4). This result suggested that
the adsorption of lysozyme governed by the amphiphilic nature of this protein was not
sufficient for bubble stabilization. This indicated the importance of the change in pH caused
by the dissolution of CO2.
92
Figure 5.3 A-C) Optical microscopy images of the bubbles at the exit of the microchannel (inset shows the image of initial bubble at the orifice, scale bar: 100 μm) (A), after 0.3(B), 2 (C), 24 (D) and 720 h (E) storage. The bubbles were formed at PCO2=48.3 kPa and QL=6 mL/h. Clys=Calg=0.2 wt. %. Scale bars: 50 μm. D) Change in the diameter of bubbles plotted as a function of time; the dashed line separates on-chip (left) and off-chip (right) bubble storage.
93
Figure 5.4 N2 bubbles after 24 h storage. Scale bar: 200 μm. N2 bubbles generated at PN2=48.3 kPa, QL=6 mL/h, pH=12, Clys=Calg=0.2 wt. %.
5.2.3 Control of Bubble Dimensions
Control over bubble dimensions was achieved (i) by varying the flow rate (QL) of the
continuous aqueous phase at constant PCO2; (ii) by changing the concentration of alginate
(Calg), and (iii) by concentration of lysozyme (Clys) (Figure 5.5-7).
Figure 5.5A shows the effect of QL on the initial and final diameters (Di and Df,
respectively) of the bubbles (the value of Df was defined as the diameter of bubbles after
their 24 h storage in a sealed container). The value of Di decreased with increasing QL,
which subsequently, led to the reduction in Df (Figure 5.5B-D).31
94
Figure 5.5 A) Effect of QL of the continuous aqueous phase on the initial (Di) and final dimension (Df) of bubbles. PCO2=48.3 kPa, Clys=Calg=0.2 wt. %, B-D) Optical microscopy images of the bubbles after 24 h (insets show the image of initial bubble at the orifice, scale bars: 100 μm) at QL=4.5 (B), 5.5 (C) and 6.5 mL/h (D). Scale bars: 15 μm
Figure 5.6 shows the variation in the values of Di and Df plotted as a function of Calg.
The initial size of the CO2 bubbles decreased with increasing Calg, due to the increase in the
viscosity of the continuous phase.31.32 As expected, the smaller Di resulted in the smaller Df
(Figure 5.6B-D). We note that the encapsulation of bubbles with lysozyme alone (at Calg= 0
wt.%) did not protect the bubbles from coalescence and/or Ostwald ripening (Figure 5.6B).
Presumably, lysozyme did not form a dense layer on the surface of bubbles, owing to
repulsive interactions between the cationic protein molecules.33
95
Figure 5.6 A) Effect of Calg on the initial and final dimensions of bubbles, PCO2=48.3 kPa, QL=6 mL/h. Clys=0.2 wt. %, B-D) Optical microscopy images of the bubbles after 24h (insets show the image of initial bubble at the orifice, scale bars: 100 μm) at Calg=0 (B, scale bar: 30 μm), 0.15 (C) and 0.2 wt. % (D). Scale bars: 15 μm
The effect of the variation in Clys on the initial and final dimensions of the bubbles is
shown in Figure 5.7A. We note that the bubbles coated with alginate only (at Clys= 0 wt. %)
completely dissolved after 24 h storage (Figure 5.7B). The value Di of the bubbles
decreased when Clys increased, owing to the higher viscosity of the continuous phase.31,34
however, the final size of bubbles, Df, increased (Figure 5.7B-D). The latter effect was
caused by the higher surface concentration of lysozyme at the gas-liquid interface35 and the
formation of a more robust barrier to the dissolution of bubbles.
96
Figure 5.7 A) Effect of Clys on the initial and final dimensions of bubbles, PCO2=48.3 kPa, QL=6 mL/h. Calg=0.2 wt. %, B-D) Optical microscopy images of the bubbles after 24h (insets show the image of initial bubble at the orifice, scale bars: 100 μm) at Clys=0 (B), 0.05 (C) and 0.15 wt. % (D). Scale bars: 15 μm
Figure 5.8 The preparation of 7 μm-diameter bubbles produced at PCO2=72.4 kPa, QL=10.5 mL/h. Clys=0.05 wt. % and Calg=2 wt. %. Optical microscopy image of the bubbles after 24 (A). (inset: the initial CO2 bubble at the orifice. scale bar: 80 μm) and 720 h (B). Scale bars: 15 μm.
97
Overall, these results indicate the crucial role of both lysozyme and alginate in the
stabilization and control of the final dimension of bubbles. At the optimized concentration,
the lysozyme shell controlled partial CO2 dissolution thereby controlling the value of Df,
while alginate contributed in the formation of the robust shell protecting bubbles from
coalescence. Figure 5.8A shows 7 μm-diameter biopolymer-encapsulated bubbles
generated from the bubbles with the initial diameter of 135 μm (inset) at Clys=0.05 wt.
% and Calg=0.2 wt. %. The 7 μm-diameter bubbles retained their size distributions for
more than a month (Figure 5.8B).
5.2.4 Characteristic Properties of the Bubbles
The biopolymer-encapsulated bubbles were imaged using laser confocal fluorescence
microscope (LCFM) and scanning electron microscope (SEM). In the first series of
experiments, we used the autofluorescence of lysozyme (attributed to the presence of
tryptophan).36 A typical LCFM image of the encapsulated bubble in Figure 5.9A shows a
well-defined bright fluorescent shell with the thickness of ca. 1 μm, which surrounded a
dark gaseous core. The cryo-SEM imaging carried at -20oC revealed the surface and
internal structure of the biopolymer shell (Figure 5.9B-D). A few micrometer-thick strands
surrounding the bubble appear due to the complexation of lysozyme and alginate in the
continuous phase (Figure 5.9B). The images of both the surface of the shell and of the inner
layer, adjacent to the bubble surface, showed the presence of aggregates of lysozyme
(Figure 5.9C, D).37 In particular, the lysozyme layer was composed of 500~700 nm-size
spherical aggregates.
98
Figure 5.9 A) LCFM image of the bubble encapsulated with a lysozyme-alginate shell and stored for 10 days (PCO2=48.3 kPa, QL=6 mL/h, Clys=Calg=0.2 wt. %). The focal plane is located at 8 μm below the surface of the bubble. B) SEM images of the bubble produced under the same conditions as in A). In A) and B) the scale bar is 7 μm. C) A high magnification SEM image of the surface of the bubble shown in (B). The scale bar is 2 μm. D) SEM image of the fractured bubble surface. The bubble was stored for 28 days. The scale bar is 4 μm.
We examined the change in the secondary structure of lysozyme deposited on the
surface of a bubble. Figure 5.10 shows the CD (circular dichroism) spectra of the aqueous
solution of lysozyme and the bubbles encapsulated with a lysozyme-alginate shell. All
spectra were acquired in tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl)
buffer solution at pH=8. Although the intensity profile differed compared to the native
lysozyme, the dispersion of bubbles stored for 5 days showed the minimum negative peak
99
at ca. 207 nm which is characteristic of the secondary structure of lysozyme.16,38 After 28
day storage, the CD spectra showed the shift in the minimum negative peak to around 215
nm, which occurs when the conformation of the protein changes to a β-sheet rich
structure.39
Figure 5.10 Comparison of CD spectra of the native lysozyme solution (Clys=0.003 wt. %) (━) and the bubbles encapsulated with a lysozyme-alginate shells (,). The bubbles were stored for 5 days () and 28 days (). Conditions of bubble formations: PCO2=48.3 kPa, QL=5 mL/h, Clys=Calg=0.2 wt. %.
We further investigated on the organization of the β-sheet structure using cross-
polarization microscopy. The Maltese-cross pattern was observed for the lysozyme-alginate
shells following 28 day bubble storage (Figure 5.11A). This pattern was previously
observed for spherical aggregates with the radial arrangement of amyloid fibrils whose
internal structure is largely cross β-sheet structure, at low pH or elevated temperature
conditions.40 However, since in our work the dimensions of the spherical aggregates in the
bubble shell were only 500~700 nm (as in Figure 5.9D), we assumed that the spherical
aggregates consists of short fragments of amyloid fibrils.41 To verify the existence of
100
fragments of amyloid fibrils in the biopolymer shell of the bubbles stored for 28 days, we
performed a Thioflavin-T (ThT) dye binding test. The binding of the dye molecule to
amyloid fibril results in the enhancement of fluorescence intensity by restricting the free
rotation of the dye. 42 An aliquot of the bubble dispersion was added into the 50 μM ThT
solution at pH=8. In comparison with control experiments with a freshly prepared
lysozyme-ThT solution, alginate-ThT solution and a stock ThT solution, we observed the
increase in fluorescence intensity (Figure 5.11B), which confirmed the existence of
fragmented amyloid fibrils on the bubble surface.40,41
Figure 5.11 A) Polarization optical microscopy image of the bubbles coated with a lysozyme-alginate shell. The bubbles were produced under PCO2=48.3 kPa, QL=6 mL/h, Clys=Calg=0.2 wt. % and stored for 28 days. The scale bar is 50 μm. B) Fluorescence intensity profile of 50 μM ThT solution (), freshly prepared lysozyme (Clys=0.003 wt. %) dissolved in 50 μM ThT solution (), freshly prepared alginate (Calg=0.003 wt. %) dissolved in 50 μM ThT solution () and the dispersion of biopolymer-encapsulated bubbles in 50 μM ThT solution (). The bubbles were stored for 28 days.
5.3 Conclusion
The potential applications of bubbles in biomedical field are continually growing as
novel formulations and methods appear. Bubbles in an appropriate size range (below 10
μm) provide a unique range of responses to ultrasound, which makes them useful for
101
ultrasound contrast imaging, identifying molecular expression and targeting drugs to
specific tissue sites.1,2,7 Advances in the understanding of physicochemical properties has
led to the recent development of bubbles encapsulated with various shell materials
including polymer brushes, polyelectrolyte multilayers, and nanoparticles.7
In the proposed approach, a uniform dissolution of the CO2 bubbles generated by the
MF approach provides the ability to produce small (~7 μm) bubbles and encapsulate them
with a protein–polysaccharide shell. The polymers used for bubble encapsulation are
biocompatible and have good mechanical properties, which is crucial for the above-
mentioned biomedical applications.1,7 The bubbles have a controllable size, a narrow size
distribution, and a long-term stability. Further reduction in bubble size to 2-3 μm can be
achieved in the microfluidic devices with ca. 20 % smaller size of features of the MF
device. Further increase in productivity of the device can be achieved by using multiple
modular MF devices.43 Moreover, the described method will allow further functionalization
of the biololymer shell by conjugating it with e.g., peptides or incorporating in it
therapeutic agents, magnetic or fluorescent species.17,44,45
References
(1) E. Stride, M. Edirisinghe, Soft Matter 2008, 4, 2350.
(2) M. A. Borden, Soft Matter 2009, 5, 716.
(3) J. D'Arrigo, Stable Gas-in-Liquid Emulsions: Production in Natural Waters and
Artificial Media, Elsevier Science Pub. Co., New York 1986.
(4) Y. Y. Shen, M. L. Longo, R. L. Powell, Journal of Colloid and Interface Science
2008, 327, 204.
102
(5) I. Lentacker, S. C. De Smedt, J. Demeester, V. Van Marck, M. Bracke, N. N.
Sanders, Advanced Functional Materials 2007, 17, 1910.
(6) I. Lentacker, S. C. De Smedt, N. N. Sanders, Soft Matter 2009, 5, 2161.
(7) S. R. Sirsi, M. A. Borden, Bubble Science, Engineering & Technology 2009, 1, 3.
(8) M. W. Grinstaff, K. S. Suslick, Proceedings of the National Academy of Sciences of
the United States of America 1991, 88, 7708.
(9) K. Bjerknes, K. Dyrstad, G. Smistad, I. Agerkvist, Drug Development and
Industrial Pharmacy 2000, 26, 847.
(10) N. Duerr-Auster, R. Gunde, R. Mader, E. J. Windhab, Journal of Colloid and
Interface Science 2009, 333, 579.
(11) E. Dickinson, R. Ettelaie, B. S. Murray, Z. P. Du, Journal of Colloid and Interface
Science 2002, 252, 202.
(12) E. A. Disalvo, Advances in Colloid and Interface Science 1988, 29, 141.
(13) E. Dressaire, R. Bee, D. C. Bell, A. Lips, H. A. Stone, Science 2008, 320, 1198.
(14) B. S. Murray, Current Opinion in Colloid & Interface Science 2007, 12, 232.
(15) P. J. Wilde, Current Opinion in Colloid & Interface Science 2000, 5, 176.
(16) B. S. Murray, Current Opinion in Colloid & Interface Science 2002, 7, 426.
(17) F. Cavalieri, M. Ashokkumar, F. Grieser, F. Caruso, Langmuir 2008, 24, 10078.
(18) E. Dickinson, E. Izgi, Colloids and Surfaces a-Physicochemical and Engineering
Aspects 1996, 113, 191.
(19) R. A. Ganzevles, R. Fokkink, T. Van Vliet, M. A. C. Stuart, H. H. J. Jongh, Journal
of Colloid and Interface Science 2008, 317, 137.
(20) E. Dickinson, Soft Matter 2008, 4, 932.
(21) A. Soltys, M. Lazarz, E. Chibowski, Colloids and Surfaces a-Physicochemical and
Engineering Aspects 1997, 127, 163.
(22) M. R. H. Krebs, G. L. Devlin, A. M. Donald, Biophysical Journal 2007, 92, 1336.
(23) W. Norde, J. P. Favier, Colloids and Surfaces 1992, 64, 87.
(24) M. Takahashi, Journal of Physical Chemistry B 2005, 109, 21858.
(25) T. Harnsilawat, R. Pongsawatmanit, D. J. McClements, Food Hydrocolloids 2006,
20, 577.
(26) P. Garstecki, H. A. Stone, G. M. Whitesides, Physical Review Letters 2005, 94.
103
(27) J. I. Park, Z. H. Nie, A. Kumachev, E. Kumacheva, Soft Matter, 6, 630.
(28) P. V. Danckwerts, Gas-liquid Reactions, McGraw-Hill Book Company, New York
1970.
(29) J. I. Park, Z. Nie, A. Kumachev, A. I. Abdelrahman, B. R. Binks, H. A. Stone, E.
Kumacheva, Angewandte Chemie-International Edition 2009, 48, 5300.
(30) A. Persat, R. D. Chambers, J. G. Santiago, Lab on a Chip 2009, 9, 2437.
(31) P. Garstecki, I. Gitlin, W. DiLuzio, G. M. Whitesides, E. Kumacheva, H. A. Stone,
Applied Physics Letters 2004, 85, 2649.
(32) C. H. Choi, J. H. Jung, Y. W. Rhee, D. P. Kim, S. E. Shim, C. S. Lee, Biomedical
Microdevices 2007, 9, 855.
(33) Y. Desfougeres, V. Lechevalier, S. Pezennec, F. Artzner, F. Nau, Journal of
Agricultural and Food Chemistry 2008, 56, 5120.
(34) J. Lefebvre, Rheologica Acta 1982, 21, 620.
(35) B. S. Murray, Langmuir 1997, 13, 1850.
(36) T. Imoto, J. A. Rupley, F. Tanaka, L. S. Forster, Proceedings of the National
Academy of Sciences of the United States of America 1972, 69, 1151.
(37) K. Nayebzadeh, J. S. Chen, E. Dickinson, T. Moschakis, Langmuir 2006, 22, 8873.
(38) F. Tanaka, L. S. Forster, P. K. Pal, J. A. Rupley, J. Biol. Chem. 1975, 250, 6977.
(39) S. Goda, K. Takano, Y. Yamagata, R. Nagata, H. Akutsu, S. Maki, K. Namba, K.
Yutani, Protein Science 2000, 9, 369.
(40) M. R. H. Krebs, K. R. Domike, D. Cannon, A. M. Donald, Faraday Discussions
2008, 139, 265.
(41) C. Ha, C. B. Park, Biotechnology and Bioengineering 2005, 90, 848.
(42) M. Biancalana, S. Koide, Biochimica Et Biophysica Acta-Proteins and Proteomics,
1804, 1405.
(43) W. Li, J. Greener, D. Voicu, E. Kumacheva, Lab on a Chip 2009, 9, 2715.
(44) N. Gaponik, I. L. Radtchenko, G. B. Sukhorukov, A. L. Rogach, Langmuir 2004, 20,
1449.
(45) H. A. Wege, A. K. F. Dyab, O. D. Velev, V. N. Paunov, Physical Chemistry
Chemical Physics 2007, 9, 6300.
104
Chapter 6
A Single-step Microfluidic Route to Producing Multifunctional
Microbubbles
This chapter describes a single-step microfluidic approach to producing small and
stable bubbles functionalized with nanoparticles (NPs). The strategy includes the following
events occurring in sequence: (i) a microfluidic generation of bubbles from a mixture of
CO2 and a minute amount of gases with low solubility in water, in an aqueous solution of a
protein, a polysaccharide and NPs; (ii) rapid dissolution of CO2 leading to the shrinkage of
bubbles and an increase in acidity of the medium in the vicinity of the bubbles; and (iii) co-
deposition of the biopolymers and NPs at the gas-liquid interface. The proposed approach
yields microbubbles with a narrow size distribution, long term stability and multiple
functions originating from the attachment of metal oxide, metal, or semiconductor NPs onto
the bubble surface. We show the potential applications of these bubbles in ultrasound and
magnetic resonance imaging.
6.1 Introduction
Ultrasound (US) imaging is non-invasive, safe and cost-effective.1 The limitation of
this method is its low contrast, in comparison with other imaging techniques such as X-ray
tomography or magnetic resonance imaging (MRI).2 To address this shortcoming,
microbubbles are utilized as US imaging contrast agents. Microbubbles enhance
backscattered acoustic signal, due to the large impedance mismatch between the bubbles
and the living tissue, and enable resonant scattering.1,3,4 Further improvement in the
105
accuracy of clinical assessment of a disease can be achieved by combining US with MRI or
fluorescence imaging.5-10 These imaging modalities are realized by immobilizing magnetic
or semiconductor nanoparticles (NPs) on the bubble surface.9,10 Attachment of NPs to the
bubble surface also increases the contrast in US imaging, owing to enhanced asymmetric
bubble oscillations.1,11 In addition to diagnostics, microbubbles carrying drug-loaded NPs
have promising therapeutic applications: bubbles could be disrupted by a localized US to
release therapeutic agents at the target site.12-15
Currently, microbubbles coated with NPs are produced in a multi-step procedure by
sonicating a mixture of NPs and oil in the presence of a gas and subsequently transferring
the resulting bubbles into an aqueous medium.9,16 Alternatively, NPs are attached to the
surface of bubbles in a layer-by-layer (LbL) deposition process.10,17 These time-consuming
processes yield bubbles with dimensions in the size range between one and tens of
micrometers and a broad distribution of sizes.18 Recently, a microfluidic (MF) strategy
enabled the synthesis of monodisperse NP-loaded bubbles via the formation of double
emulsions.19 The method required very careful tuning of the surface energies at the gas-
liquid and liquid-liquid interfaces.19,20 In addition, the dimensions of bubbles exceeded 10
µm, the size utilized in US diagnostics or therapeutics.21
This chapter decribes a new MF approach to producing NP-coated microbubbles with a
narrow size distribution and a long-term stability. The proposed approach exploits the
following events, occurring concurrently within 3 sec: (i) the MF generation of
monodisperse bubbles from a mixture of CO2 and a minute amount of a gas with a low
solubility in water, in an aqueous solution of lysozyme, alginate and anionic NPs; (ii) the
controllable dissolution of CO2 leading to bubble shrinkage and an increase in acidity of the
medium in the neighbourhood of the bubbles. The decrease in pH renders positive net
106
charge to lysozyme; and (iii) the adsorption of cationic lysozyme to the negatively charged
surface of bubbles followed by the deposition of anionic alginate and NPs onto the
lysozyme layer. This rapid and simple process generates bubbles with polydispersity not
exceeding 6 % and stability of at least, 3 months. We show the ability to produce 5 μm-
diameter bubbles, however for convenience of optical imaging we present most of the
results for bubbles with a mean diameter of approximately 10 μm. Bubbles with such
dimensions, comparable to red blood cells, can safely pass through the microvasculature
without diffusing across the endothelium.22 We demonstrate the generality of the MF
approach by attaching to the bubble surface metal oxide, metal and semiconductor NPs.
Furthermore, we show the applications of the NP-functionalized bubbles as imaging agents
in US, fluorescence and MRI.
6.2 Results and Discussions
The materials and methods described in the present chapter are listed in Chapter 2 and
Appendix.
6.2.1 Experimental Design
Figure 6.1A shows the schematic of the MF device used in the present work. A MF
flow-focusing bubble generator23 was followed by a serpentine downstream microchannel
channel (Appendix, Figure A6.1). The CO2 gas mixed with 0.2 vol % of N2, O2, He, and
Ne supplied under pressure PCO2 of 48.3 kPa to the central channel. An aqueous solution
107
containing a mixture of lysozyme, alginate and anionic NPs (Fe3O4, Au or SiO2-
encapsulated CdSe/ZnS NPs) was introduced into the two side channels as a continuous
phase using a syringe pump at the flow rate, QL, varying from 6.5 to 9.5 mL/h. The value of
pH=12 of the continuous phase was achieved by adding to it a 1M NaOH solution.
Formation of bubbles occurred via a periodic breakup of the gas thread in the orifice of
the MF device.23 Since the unperturbed diameter of the bubbles immediately after their
formation was larger than the height of the downstream channel, they acquired a discoid
shape (Figure 6.1A, top inset). In the downstream channel the dimensions of the
bubbles dramatically reduced, due to the dissolution of CO2, and acquire a spherical
shape (Figure 6.1A, bottom inset).
Figure 6.1 Schematics of the microfluidic (MF) generation of multifunctional bubbles. (A) Schematic of a MF reactor. The height of the MF device is 120 µm. The width of the orifice and the length of the downstream microchannel are 50 µm and 260 mm, respectively. The top and the bottom insets show optical microcopy images of the bubbles at the beginning and the end of the process, respectively. Bubbles were generated at PCO2=48.3 kPa and QL=9.5 mL/h. Scale bars in insets are 200 µm. (B) Schematic of the formation of NP functionalized bubbles stabilized with a mixed lysozyme-alginate layer.
108
Figure 6.1B illustrates the sequence of events leading to the generation of small and
stable NP-functionalized bubbles encapsulated with the mixture of biopolymers. Rapid and
uniform shrinkage of bubbles is driven by the dissolution of CO2 and chemical reactions
of CO2 with OH- ions (see below). The dissolution of CO2 leads to the decrease in pH of
the aqueous medium in the vicinity of the gas-liquid interface (Figure 6.1B(1)). As a result
of the change in acidity, lysozyme molecules in the neighborhood of the bubbles gain a net
positive charge and adsorb to the negatively charged bubble surface (Figure 6.1B(2)). (The
isoelectric point of lysozyme is ~11).24,25 Anionic NPs and anionic alginate molecules
deposit on the cationic lysozyme-coated bubble (Figure 6.1B(3)). The formation of the
lysozyme-alginate shell on the surface of bubbles is described in greater detail in Chapter 5.
An important feature of our work was the ability to control the dimensions of
bubbles, which was achieved by generationg bubbles with a well-defined size,
controllable dissolution of CO2 and stabilization of the bubbles with a biopolymer layer.
The dissolution of CO2 in the microchannel was governed by Henry’s law (eq. 6.1), which
was combined with chemical reactions occurring at pH>10 (eq. 6.2- 3):26
[CO2]l kH PCO2
CO2 + OH- HCO3-
HCO3- + OH- CO3
2- + H2O
where K1 = 3.2x107 and K2= 3.5x103 are the equilibrium constants and [CO2]l, kH, and
PCO2 are the concentration of molecularly dissolved CO2, Henry’s law constant (for
CO2 at 25oC kH=3.2x10-4 mol/(L kPa)), and the gas pressure, respectively. The extent of
dissolution of bubbles in the MF device was controlled by the saturation of the continuous
aqueous phase, 27 and in the present work the dissolution led to ~80% of the reduction in
K1
K2
(eq. 6.1)
(eq. 6.2)
(eq. 6.3)
109
bubble volume. Bubbles were generated at a frequency of 700 bubbles/sec. The
bioencapsulated NP-functionalized bubbles were collected in a 2 mL container, where
their dimensions further decreased due to the reduction of external pressure to 1
atm.28,29 The container was sealed for long-term bubble storage.
6.2.2 Long-term Stability of Biopolymer Encapsulated-Bubbles
Functionalized with NPs
Figure 6.2A-C shows typical optical microscopy images taken at different times
after the generation of bioecapsulated bubbles loaded with Fe3O4 NPs. The
polydispersity of the bubbles did not exceed 6%. Following bubble formation, within 3
sec their mean diameter reduced from 150 to 40 μm (on-chip), and subsequently, to 5
μm within 1 h of off-chip storage. At this point, the fractional reduction in the bubble
volume reached 99.9 %, which implied that all CO2 and 0.1% of the low-soluble gases
were removed from the bubbles. No further change in bubble dimensions was observed
after 2000 h storage.
Figure 6.2 Optical microscopy images of the bubbles functionalized with Fe3O4 NPs after different storage times: (A) 3 sec, (B) 1 h and (C) 2000 h. The scale bars in (A), (B) and (C) are 50, 15 and 15 µm, respectively. Inset in (A) shows the bubble imaged immediately after its generation in the orifice of the MF device. Scale bar is 100 µm. Bubbles were generated at PCO2=48.3 kPa and QL=9.5 mL/h.
110
6.2.3 Control of the Dimensions of Biopolymer Encapsulated-Bubbles
Functionalized with NPs
The final dimensions of bubbles - defined as the diameter of bubbles after 1 h
storage – were tuned by varying the flow rate of the continuous aqueous phase (QL),
while maintaining PCO2 as 48.3 kPa. Figure 6.3A shows the variation in the initial (Di)
and the final (Df) dimensions of NP-coated bubbles, plotted as a function of the QL of
the continuous phase. The value of QL affected the final size of bubbles in two ways.
With increasing value of QL the original bubble size reduced, thereby resulting in the
decrease of Df. Second, the extent of dissolution of CO2 increased at higher values of
QL,27,30 also leading to smaller bubble size (Figure 6.3B-D).
Figure 6.3 (A) Variations in the initial,Di, and final dimensions, Df, of microbubbles are plotted as a function of the flow rate of the continuous phase, QL. (B-D) Representative optical microscope images of the initial (insets) and final dimension of bubbles coated with the biopolymers and Fe3O4 NPs at different QL s (A) 6.5, (B) 7.5 and (C) 8.5 mL/h. Clysozyme=0.05 wt. %. Calginate=0.1 wt. %, and CFe3O4 dispersion = 1 wt. %. PCO2=48.3 kPa. Scales bars are 15 µm (100 µm, insets).
111
6.2.4 Control Experiments
In the control experiments conducted in the absence of lysozyme, the bubbles were
completely dissolved within a few minutes after their generation. Without alginate in
the continuous phase the bubbles aggregated due to the insufficient electrostatic and
steric stabilization (Figure 6.4A). The importance of local acidification of the medium
was examined by generating N2 bubbles at initial pH=12. These bubbles did not have a
dense lysozyme-alginate shell and within 1 h they underwent coalescence (Figure
6.4B).
Figure 6.4 Optical microscope images of (A) aggregated bubbles generated with lysozyme and Fe3O4 NPs. Scale bar is 15 µm and (B) coalesced N2 bubble generated with lysozyme, alginate and Fe3O4 NPs. Scale bar is 200 µm.
6.2.5 Characterization of Biopolymer Encapsulated-Bubbles
Functionalized with various NPs
The NP-coated bubbles were collected at the outlet of the MF device, washed several
times with de-ionized water, dried and imaged using scanning transmission electron
microscopy (STEM). The STEM images of the bubbles coated with Fe3O4, Au and SiO2-
encapsulated CdSe/ZnS NPs are shown in Figure 6.5A-C, respectively. We note that the
CdSe/ZnS NPs were encapsulated within a SiO2 shell, in order to suppress the cytotoxicity
of these NPs.31 Upon drying, the bubbles maintained a spherical shape and featured
112
wrinkles on their surface (see e.g., Figure 6.5A). In the high magnification images (insets in
Figure 6.5A-C), the NPs appeared bright on the dark background of the lysozyme-alginate
biopolymer shell. The Au and SiO2-encapsulated CdSe/ZnS NPs were well-separated,
however Fe3O4 NPs tend to form clusters in the biopolymer shells.
Figure 6.5 Scanning transmission electron microscopy (STEM) images of the bubbles coated with the lysozyme-alginate shell and (A) Fe3O4 NPs, (B) Au NPs, and (C) SiO2-encapsulated CdSe/ZnS NPs, Scale bars are 6 µm. Insets in (A-C) show corresponding high magnification images of the surface of the bubbles. Scale bars in insets are 150 nm. (D-F) Energy dispersive X-ray (EDX) spectrometry line scanning profiles for the system shown as a red line in (A), (B) and (C). Bubbles were generated at PCO2 = 48.3 kPa and QL = 8.5 mL/h.
In each system, the presence of NPs on the surface of bubbles was verified by energy
dispersive X-ray (EDX) analysis. Figure 6.5D-F shows the results of EDX analysis of the
surface of bubbles, based on the major constituent of the NPs. The EDX line profiles
featured strong EDX signals of Fe, Au and Si (Appendix, Figure A6.2-4).
113
6.2.6 Control over the Amount of NPs on the Bubble Surface
Control over the amount of NPs deposited on the surface of bubbles was achieved by
varying the concentration of NPs in the continuous phase. Figure 6.6A-C shows the STEM
images of the surface of bubbles coated with a different amount of Fe3O4 NPs. When the
initial concentration of the NPs in the continuous phase was 1.0x1012, 4.0x1012 or 1.0x1013
NPs/mL, their density on the surface of bubbles was 1.5x105, 6.6x105, and 1.5x107
NPs/µm2, respectively (Figure 6.6A-C).
Figure 6.6 (A-C) STEM images of the surface of bubbles coated with Fe3O4 NPs at surface density of (A) 1.5x105, (B) 6.6x105 and (C) 1.5x106 NPs/µm2. Scale bars are 300 nm. Insets show the corresponding bubbles. Scale bars are 3 µm. Bubbles were generated at PCO2=48.3 kPa and and QL=8 mL/h.
6.2.7 Properties of NP-functionalized Bubbles
The NP-functionalized bubbles had a narrow size distribution with polydispersity not
exceeding 6 %. The bubbles coated with Fe3O4 NPs were readily aligned in chains when
manipulated by an external magnetic field (Figure 6.7A).32 The chains of bubbles moved
towards a magnet at a velocity of ~20 µm/s. The bubbles did not lose their ability for
magnetic actuation for at least, 3 month-long storage (Appendix, Figure A6.5).
Figure 6.7B shows the extinction spectra of the aqueous dispersion of bubbles coated
with Au NPs and of the original citrate-stabilized Au NPs in an aqueous solution. The
114
incorporation of Au NPs within a biopolymer shell led to a ~13 nm red shift of the surface
plasmon resonance bands, due to interparticle electromagnetic coupling.33,34
Figure 6.7C shows a typical confocal fluorescence microscopy image of the bubbles
coated with SiO2-encapsulated CdSe/ZnS NPs. The shell of bubbles showed strong
fluorescence when excited at 364 nm (Appendix, Figure A6.6). Figure 6.7C, inset shows a
gaseous core (dark) and a bright fluorescent shell of an individual bubble.
Figure 6.7 Properties of NP-coated bubbles. (A) Magnetic actuation of bubbles functionalized with Fe3O4 NPs. Scale bar is 50 µm. Bubbles were generated at PCO2=48.3 kPa and and QL=8.5 mL/h. (B) Extinction spectra of Au NPs (red spectrum, top) and of the bubbles coated with these Au NPs (blue spectrum, bottom). Bubbles were generated at PCO2=48.3 kPa and and QL =8.5 mL/h. (C) Confocal fluorescence microscopy image of bubbles carrying SiO2-encapsulated CdSe/ZnS NPs. Scale bar is 30 µm. λex=364 nm. Inset shows an image of the individual bubble. Scale bar is 5 µm. Bubbles were generated at PCO2=48.3 kPa and and QL=9.5 mL/h.
6.2.8 Application of NP-functionalized Bubbles in US Imaging
Aqueous dispersions of the bubbles were introduced in the Opticell chamber (Thermo
Scientific Inc.) and imaged in a bubble-specific, nonlinear imaging mode using a clinical
US system (iU22, Philips, Appendix). The nonlinear US signal was caused by asymmetric
(compression vs. expansion) oscillations of the bubbles in response to US excitation. Figure
6.8A-D shows in-vitro US images of the dispersion of NP-free and NP-functionalized
bubbles. A stronger contrast in US images in Figure 6.8B-D signified a stronger nonlinear
signal enhancement for the NP-functionalized bubbles, in comparison with NP-free bubbles
115
in Figure 6.8A. Quantitatively, signal enhancement for the NP-coated bubbles is presented
in Figure 6.8E. The signals of the bubble dispersions were compared to the background
signal of water. The enhancement was calculated as the ratio of the integrated power
measured in the chamber filled with the dispersion of bubbles to the integrated power
measured for the chamber containing pure water. The bubbles functionalized with Fe3O4,
Au and SiO2-encapsulated CdSe/ZnS NPs showed 24.1, 32.7 and 34 dB signal
enhancements, respectively, in comparison with 21.7 dB signal enhancement measured for
the NP-free bubbles (Figures 6.8B-D). We ascribe the increase in the US signal to the
enhanced non-linearity in oscillations of the NP-coated bubbles, owing to the increased
microbubble resistance to compression. This effect is strongly influenced by the surface
coverages of bubbles with NPs.11,16,35 The stronger nonlinear US signal measured for the
bubbles carrying Au or SiO2-encapsulated CdSe/ZnS NPs, in comparison with the bubbles
carrying Fe3O4 NPs supports this explanation. Based on the STEM image analysis, the
surface coverages of the bubbles with Au and SiO2-encapsulated CdSe/ZnS NPs were
three- and five-fold higher, respectively, than that for the bubbles coated with Fe3O4 NPs.
We note that the role of different types of NPs on US signal is unknown. We conclude that
the deposition of NPs onto the bubble surface increased the detectability of bubbles with
contrast-specific US pulse sequences which were designed to reject the linear signal
components arising from the tissue and bubbles and to preserve the nonlinear components
from the bubbles.1,34
The NP-coated bubbles did not lose their echogenicity under continuous exposure to low-
power US pulses (130 kPa at 5 MHz) which suggested an excellent stability of the
lysozyme-alginate shell with embedded NPs. Under increased peak negative US pressure of
1100 kPa, the bubbles were disrupted. This result implied that the NP-loaded bubbles can
116
be controllably disrupted with US and the NPs can be released and deposited at the target
sites under US image guidance.
Figure 6.8 In-vitro US imaging of the dispersion of dispersion of biopolymer-coated bubbles at 90 % receive gain. The dispersion is placed in the Opticell chamber. (A) NP-free bubbles, (B-D) bubbles coated with Fe3O4 NPs (B), Au NPs (C) and SiO2-encapsulated CdSe/ZnS NPs (D). (E) US signal enhancement over background for the systems shown in A-D. The concentration of bubbles in all systems was 104 bubbles/mL.
6.2.9 Application of NP-functionalized Bubbles in MRI
We examined the application of the Fe3O4 NP-coated bubbles in MRI, a method
providing a high spatial resolution. The NP-coated bubbles with the surface density of 0,
1.5x105, 6.6x105 and 1.5x106/µm2 were suspended in water and placed in 1.8 mL
Eppendorf tubes. The dispersions were imaged at 3.0 Tesla with a 2D coronal fast field
echo sequence (Philips Healthcare, Andover, MA, Appendix). Figure 6.9A shows that the
bubbles carrying a higher amount of Fe3O4 NPs exhibited enhanced negative contrast,
117
reflected as the darker area in the center of the Eppendorf tubes.36,37 Figure 6.9B shows the
increase in relaxation rate (1/T2*) for dispersions containing bubbles with a higher surface
density of Fe3O4 NPs. The NPs modified the MRI signal by locally perturbing the magnetic
field and thereby leading to the negative signal enhancement by increasing the T2*
relaxation rate of the nearby water molecules.37,38 These results indicate that the Fe3O4 NP-
coated bubbles can serve as effective MRI contrast agents.
Figure 6.9 (A) In-vitro MRI images (top-view) of the dispersions of biopolymer-encapsulated bubbles coated with different amounts of Fe3O4 NPs. The images were obtained at 6.9 ms echo time. (B) Variation in T2* relaxation rate plotted as a function of surface density of Fe3O4 NPs on the surface of bubbles. The concentration of bubbles in all systems was 104 bubbles/mL.
6.3 Conclusions
In conclusion, we have developed a single-step, simple MF method for producing
microbubbles functionalized with various types of NPs. The method has the productivity of
4x104 bubbles/min. The higher productivity can be achieved by using multiple modular MF
devices. The bubbles exhibit small size, a low polydispersity and long-term stability. The
functionalization of the surface of bubbles rendered the bubbles with plasmonic,
fluorescence and magnetic properties. The NP-coated bubbles showed enhanced
performance in US imaging. We also demonstrated multiple imaging modalities for NP-
118
functionalized bubbles, such as fluorescence and MRI. We envisage that these
multifunctional bubbles will have important applications in medical diagnostics, where
multiple imaging methods using a single contrast agent will be advantageous. The proposed
method can also be used in triggered site-specific release of drugs by using magnetic
actuation and focused US exposure.
References
(1) E. G. Schutt, D. H. Klein, R. M. Mattrey, J. G. Riess, Angewandte Chemie-
International Edition 2003, 42, 3218.
(2) D. Cosgrove, European Journal of Radiology 2006, 60, 324.
(3) I. Lentacker, S. C. De Smedt, N. N. Sanders, Soft Matter 2009, 5, 2161.
(4) J. R. Lindner, Nature Reviews Drug Discovery 2004, 3, 527.
(5) R. Dharmakumar, D. B. Plewes, G. A. Wright, Phys. Med. Biol. 2005, 50, 4745.
(6) Y. M. Huh, Y. W. Jun, H. T. Song, S. Kim, J. S. Choi, J. H. Lee, S. Yoon, K. S.
Kim, J. S. Shin, J. S. Suh, J. Cheon, Journal of the American Chemical Society 2005, 127,
12387.
(7) I. L. Medintz, H. T. Uyeda, E. R. Goldman, H. Mattoussi, Nature Materials 2005, 4,
435.
(8) W. J. M. Mulder, R. Koole, R. J. Brandwijk, G. Storm, P. T. K. Chin, G. J. Strijkers,
C. D. Donega, K. Nicolay, A. W. Griffioen, Nano Letters 2006, 6, 1.
(9) F. Yang, Y. X. Li, Z. P. Chen, Y. Zhang, J. R. Wu, N. Gu, Biomaterials 2009, 30,
3882.
(10) H. Ke, Z. W. Xing, B. Zhao, J. R. Wang, J. B. Liu, C. X. Guo, X. L. Yue, S. Q. Liu,
Z. Y. Tang, Z. F. Dai, Nanotechnology 2009, 20.
(11) E. Stride, K. Pancholi, M. J. Edirisinghe, S. Samarasinghe, Journal of the Royal
Society Interface 2008, 5, 807.
119
(12) F. Cavalieri, I. Finelli, M. Tortora, P. Mozetic, E. Chiessi, F. Polizio, T. B. Brismar,
G. Paradossi, Chemistry of Materials 2008, 20, 3254.
(13) I. Lentacker, S. C. De Smedt, J. Demeester, V. Van Marck, M. Bracke, N. N.
Sanders, Advanced Functional Materials 2007, 17, 1910.
(14) A. F. H. Lum, M. A. Borden, P. A. Dayton, D. E. Kruse, S. I. Simon, K. W. Ferrara,
Journal of Controlled Release 2006, 111, 128.
(15) M. Postema, O. H. Gilja, Current Pharmaceutical Biotechnology 2007, 8, 355.
(16) F. Yang, L. Li, Y. X. Li, Z. P. Chen, J. R. Wu, N. Gu, Phys. Med. Biol. 2008, 53,
6129.
(17) F. Cavalieri, M. Ashokkumar, F. Grieser, F. Caruso, Langmuir 2008, 24, 10078.
(18) E. Stride, M. Edirisinghe, Soft Matter 2008, 4, 2350.
(19) M. H. Lee, V. Prasad, D. Lee, Langmuir 2010, 26, 2227.
(20) H. C. Shum, J. W. Kim, D. A. Weitz, Journal of the American Chemical Society
2008, 130, 9543.
(21) K. Hettiarachchi, E. Talu, M. L. Longo, P. A. Dayton, A. P. Lee, Lab on a Chip
2007, 7, 463.
(22) S. Mayer, P. A. Grayburn, Progress in Cardiovascular Diseases 2001, 44, 33.
(23) P. Garstecki, I. Gitlin, W. DiLuzio, G. M. Whitesides, E. Kumacheva, H. A. Stone,
Applied Physics Letters 2004, 85, 2649.
(24) M. Takahashi, Journal of Physical Chemistry B 2005, 109, 21858.
(25) M. R. H. Krebs, G. L. Devlin, A. M. Donald, Biophysical Journal 2007, 92, 1336.
(26) P. V. Danckwerts, Gas-liquid Reactions, McGraw-Hill Book Company, New York
1970.
(27) J. I. Park, Z. H. Nie, A. Kumachev, E. Kumacheva, Soft Matter 2010, 6, 630.
(28) CO2 bubbles experienced a pressure of 48.3 kPa in the microchannel while partial
CO2 pressure at 1 atm (outside the device) is 0.04 kPa (see ref. 29).
(29) A. Persat, R. D. Chambers, J. G. Santiago, Lab on a Chip 2009, 9, 2437.
(30) J. I. Park, E. Tumarkin, E. Kumacheva, Macromolecular Rapid Communications
2010, 31, 222.
(31) S. T. Selvan, T. T. Tan, J. Y. Ying, Advanced Materials 2005, 17, 1620.
120
(32) R. M. Erb, H. S. Son, B. Samanta, V. M. Rotello, B. B. Yellen, Nature 2009, 457,
999.
(33) Z. D. Lu, J. Goebl, J. P. Ge, Y. D. Yin, J. Mater. Chem. 2009, 19, 4597.
(34) D. M. Zhang, O. Neumann, H. Wang, V. M. Yuwono, A. Barhoumi, M. Perham, J.
D. Hartgerink, P. Wittung-Stafshede, N. J. Halas, Nano Letters 2009, 9, 666.
(35) E. Stride, Cerebrovasc Diseases 2009, 27, 1.
(36) J. W. M. Bulte, D. L. Kraitchman, Current Pharmaceutical Biotechnology 2004, 5,
567.
(37) J. H. Gao, G. L. Liang, J. S. Cheung, Y. Pan, Y. Kuang, F. Zhao, B. Zhang, X. X.
Zhang, E. X. Wu, B. Xu, Journal of the American Chemical Society 2008, 130, 11828.
(38) A. P. Pathak, Wiley Interdisciplinary Reviews-Nanomedicine and
Nanobiotechnology 2009, 1, 84.
121
Chapter 7
Conclusion and Outlook
The main goal of this thesis was to develop microfluidic (MF) methods for the
preparation of microbubble-templated materials in a continuous fashion and explore their
applications. In particular, this work pioneered a pH-regulated MF formation and
dissolution of CO2 bubbles. The MF dissolution of CO2 bubbles led to two important
results: the ability to generate small (<10 µm in diameter) bubbles and the capability to
produce bubbles encapsulated with various types of colloidal particles and biopolymers. In
the second approach we exploited a local decrease in pH at the length scale of an individual
bubble. In addition, this thesis described a MF method for the synthesis of Pickering
emulsions and polymer particles with supracolloidal structure.
In Chapter 3, we investigated a pH-dependent MF formation and dissolution of CO2
bubbles. We found the initial formation of CO2 bubbles was affected by the extent of the
bubble dissolution in the downstream channel. The degree of CO2 dissolution was precisely
controlled by the flow rate of the continuous aqueous phase and the acid-base equilibria in
the continuous phase in the microchannel. The proposed approach provided a new route to
generation of small bubbles (<10 µm in diameter) with a narrow size distribution (CV<5%).
Although our focus was limited to the CO2-water reactions, these experiments can be
extended to other soluble and reactive gases such as NH3 and SO2, which would enable
systematic kinetic studies of various gas-liquid reactions. Due to the environmental
importance of such gases, this research will have significant impact on science and industry.
122
Chapter 4 described a MF method for the preparation of bubbles and droplets with a
shell of colloidal particles (colloidal armour). First, we developed a chemically mediated
MF approach for the production of bubbles encapsulated with various colloidal armours.
The strategy utilized the following events occurring in the microchannel: i) monodisperse
CO2 bubbles were generated in a dispersion of anionic particles, ii) the dissolution of CO2
bubble led to the increase in the acidity of the solution at the periphery of the bubbles, and
iii) this increased the hydrophobicity on the particle surface and resulted in the adsorption
of particles to the gas-water interface. The size and shapes of bubbles with a shell of
colloidal particle were precisely controlled by the flow rate and pH of the continuous phase,
respectively. The generality of the MF approach was demonstrated by depositing various
particles and their mixtures on the surface of bubbles.
The armoured bubbles generated by the MF method can be used as templates for
fabricating acoustic insulators, separation membranes, light-weight structures, and scaffolds
for tissue engineering.
In Chapter 4, we also presented a MF “inside-out” method for the preparation of liquid
droplets coated with colloidal particles (Pickering emulsions). We demonstrated the
utilization of this system for the synthesis of supracolloidal polymer microparticles. The
proposed method has several advantages over the conventional methods of the preparation
of Pickering emulsions. First, it allows the precise control of the dimensions of droplets.
Second, it minimized the waste of particles by introducing them into the droplet phase.
Third, it provides the ability to control the surface coverage of droplets with particles.
The proposed method for the generation of Pickering emulsion can be applied for the
controlled encapsulation of reagents, drugs, and biological species.
123
In Chapter 5, a MF route to the generation of small, stable, and monodisperse bubbles
encapsulated with a lysozyme-alginate shell was proposed. The dissolution of CO2 bubbles
resulted in the increase of the acidity around the bubble surface. This led to the deposition
of positively charged lysozyme on the negatively charged surface of bubbles. The alginate
molecules in the continuous phase formed a complex with the positively charged lysozyme
on the bubble surface. As a result, the bubbles were coated with a biopolymer shell, which
provided a long-term stability (up to a month, at least) against bubble dissolution and
coalescence. The dimensions of bubbles were controlled by the hydrodynamic means and
by tuning the concentrations of lysozyme and alginate in the continuous phase.
In Chapter 6, we presented a MF single-step functionalization of bubbles with various
types of nanoparticles (NPs). The MF dissolution of CO2 bubbles resulted in the co-
deposition of biopolymers and NPs on the bubble surface due to the electrostatic
interactions. The presence of NPs on the surface of bubbles rendered the bubbles with
plasmonic, fluorescence and magnetic properties. The amount of NPs on the bubble surface
was controlled by tuning the initial concentration of NPs in the continuous phase.
Furthermore, we explored the applications of these NP-coated bubbles as contrast agents in
in-vitro medical imaging. First, we demonstrated the application of bubbles with a shell of
biopolymers and NPs as ultrasound (US) contrast agents. The bubbles encapsulated with
biopolymers and NPs showed improved contrast in US imaging, in comparison with the
NP-free bubbles. Second, the incorporation of Fe3O4 NPs on the surface of bubbles allowed
us to use them as effective magnetic resonance imaging (MRI) agents. The contrast in MRI
was enhanced as the concentration of Fe3O4 NPs on the surface of bubbles increased.
The bubbles prepared by the approaches presented in Chapter 5 and 6 can find
applications in medical diagnostics, drug delivery and gene therapy. For instance, drug or
124
nucleic acid-loaded NPs can be deposited on the surface of bubbles along with magnetic
NPs. This would allow the release of the loaded agents at a target site by using magnetic
actuation and focused US exposure. In addition, the localization of NPs on the bubble
surface would benefit the increase of the drug payload that is delivered at the target site.
125
Appendix
Appendix to Chapter 3
Estimation of the Amount of Dissolved CO2 and the Change of pH of the
Continuous Phase Following the Dissolution of CO2
For pH=13.2, kH=3.2 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. Thus, [CO2]l kH PCO2,
[CO2]l ~0.009 mol/L (Henry’s law). The total concentration of dissolved CO2 can be
estimated by adding the dissolved unreacted amount of CO2 (determined by Henry's law)
and reacted CO2 (following eq. 3.3).
CO2 (l) + OH- HCO3- K = 3.2 x 107
Initial concentration 0.009 0.16
Reacted concentration -x -x x
Final concentration 0.009 -x 0.16- x x
x ~ 0.009 mol/L
The total amount of dissolved CO2 (both unreacted and reacted) is [CO2]l + x, that is, 0.018
mol/L.
For the estimation of the change in the value of pH after the dissolution of CO2 in the
microchannel, we consider the reaction (eq. 3.4) to determine the final concentration of OH-.
HCO3- + OH- CO3
2- + H2O K = 3.5x 103
Initial concentration 0.009 0.15
Reacted concentration -y -y y
Final concentration 0.009-y 0.15-y y
y~0.009 mol/L
Therefore, the final [OH-] ~ 0.14 mol/L and pH=-log(10-14/[OH-]) ~ 13.1
126
For pH=11, kH=3.3 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. Thus, [CO2]l = kH PCO2, [CO2]l
~0.009 mol/L (Henry’s law). The total concentration of dissolved CO2 can be estimated by
adding the dissolved unreacted amount of CO2 (determined by Henry's law) and reacted
CO2 (following eq. 3.3).
CO2 (l) + OH- HCO3- K = 3.2 x 107
Initial concentration 0.009 0.001
Reacted concentration -x -x x
Final concentration 0.009 -x 0.001- x x
x ~ 0.001 mol/L
The total amount of dissolved CO2 (both unreacted and reacted) is [CO2]l + x, that is, 0.01
mol/L.
For the estimation of the change in the value of pH after the dissolution of CO2 in the
microchannel, we consider the reaction (eq. 3.4) to determine the final concentration of OH-.
HCO3- + OH- CO3
2- + H2O K = 3.5x 103
Initial concentration 0.001 0.00003
Reacted concentration -y -y y
Final concentration 0.001-y 0.00003-y y
y~0.00002 mol/L
Therefore, the final [OH-] ~ 0.00001 mol/L and pH=-log(10-14/[OH-]) ~ 9
For pH=9, kH=3.3 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. Thus, [CO2]l = kH PCO2, [CO2]l
~0.009 mol/L (Henry’s law). The total concentration of dissolved CO2 can be estimated by
adding the dissolved unreacted amount of CO2 (determined by Henry's law) and reacted
CO2 (following eq. 3.2).
CO2 + H2O HCO3- + H+ K = 4.4 x 10-7
Initial concentration 0.009
Reacted concentration -x x x
Final concentration 0.009 -x x x
x ~ 0.00006 mol/L
127
The total amount of dissolved CO2 (both unreacted and reacted) is [CO2]l + x, that is,
0.00906 mol/L.
Here, the final concentration of [H+] ~ x ~ 0.00906. Therefore, the final pH=-log[H+] ~ 4.2.
For pH=7, kH=3.3 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. Thus, [CO2]l = kH PCO2, [CO2]l
~0.009 mol/L (Henry’s law). The total concentration of dissolved CO2 can be estimated by
adding the dissolved unreacted amount of CO2 (determined by Henry's law) and reacted
CO2 (following eq. 3.2).
CO2 + H2O HCO3- + H+ K = 4.4 x 10-7
Initial concentration 0.009
Reacted concentration -x x x
Final concentration 0.009 -x x x
x ~ 0.00006 mol/L
The total amount of dissolved CO2 (both unreacted and reacted) is [CO2]l + x, that is,
0.00906 mol/L.
Here, the final concentration of [H+] ~ x ~ 0.00906. Therefore, the final pH=-log[H+] ~ 4.2.
For pH=5, kH=3.3 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. Thus, [CO2]l = kH PCO2, [CO2]l
~0.009 mol/L (Henry’s law). The total concentration of dissolved CO2 can be estimated by
adding the dissolved unreacted amount of CO2 (determined by Henry's law) and reacted
CO2 (following eq. 3.2).
CO2 + H2O HCO3- + H+ K = 4.4 x 10-7
Initial concentration 0.009
Reacted concentration -x x x
Final concentration 0.009 -x x x
x ~ 0.00006 mol/L
The total amount of dissolved CO2 (both unreacted and reacted) is [CO2]l + x, that is,
0.00906 mol/L.
Here, the final concentration of [H+] ~ x ~ 0.00906. Therefore, the final pH=-log[H+] ~ 4.2.
128
For pH=1.5, kH=3.3 x 10-4 mol/(L kPa) and PCO2=27.6 kPa. The total concentration of
dissolved CO2 is dominated by Henry’s law, [CO2]l = kH PCO2 ~0.009 mol/L. For the
estimation of the change in the value of pH after the dissolution of CO2 in the microchannel,
we consider the reaction (eq. 3.2)
CO2 + H2O HCO3- + H+ K = 4.4 x 10-7
Initial concentration 0.032
Reacted concentration -x x x
Final concentration 0.009 x 0.032+x
x ~ 0.0000001 mol/L
Here, the final concentration of [H+] ~ x ~ 0.032. Therefore, the final pH=-log[H+] ~ 1.5.
129
Appendix to Chapter 4
A4.1 Generation of Armoured Bubbles in a Microfluidic Flow-Focusing
Bubble Generator
Figure A4.1 Production of armoured bubbles in a microfluidic (MF) flow-focusing bubble
generator. Generation of the CO2 plugs (A), their dissolution (B), and armouring of the
shrunken bubbles with PS-co-PAA particles in the downstream channel (C). In (A-C), scale
bars are 200 μm. (D) Armoured bubbles collected at the outlet of the MF device. The scale
bar is 100 μm. (E) Magnified image of the bubbles in (D), showing the close-packed PS-co-
PAA particles on the bubble surface. Scale bar is 50 μm. Bubbles generated at PCO2=48.5
kPa, QL=10.4 mL/h, Cp=1.0 wt% and pH=14.
130
Appendix to Chapter 6
A6.1 Experimental Design
The dimensions of the MF bubble generator are shown in Figure S1. The continuous
phase contained 0.05 wt % of lysozyme, 0.1 wt % of alginate (the mixture of lysozyme and
alginate is referred as biopolymers) and 1 wt % of NP dispersion. The concentrations of
NPs were controlled by the number of centrifugation cycles of NP dispersions at 8000 rpm
for 30 min.
Figure A6.1 Schematic of the MF flow-focusing bubble generator. Wg= 105 μm, Wa=75
μm, Wo=50 μm, Wc=240 μm, and Lc=260 mm. The height of the microchannels is 120 μm.
A6.2 Inductively Coupled Plasma Atomic Emission Spectroscopy
Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was used to
determine the concentration of Fe3O4 NPs on the surface of bubble. The amount of Fe in the
bubbles was estimated as follows: Cbubbles(Fe) = Cinitial(Fe) - Cfinal(Fe) where Cbubbles(Fe) is the
concentration of Fe in the suspension of bubbles (g/mL), Cinitial(Fe) and Cfinal(Fe) are the initial
and final concentrations of Fe in the continuous phase (g/mL), respectively. The number of
NPs per bubble was estimated as Cbubbles(Fe) x MWFe3O4 /(3 x MWFe x d x V x N) where
MWFe3O4 and MWFe are the molecular weight of Fe (55.8 g/mol) and Fe3O4 (231.5 g/mol),
131
respectively. d and V are the density (~1.25 g/mL), volume (~4.2x10-18 mL) of the NP,
respectively. N is the number concentration of bubbles (~106 bubbles/mL).
A6.3 In vitro Ultrasound Imaging
10 mL of the bubble suspension with concentration of approximately 104 bubbles/mL were
introduced into Opticell chamber (Thermo Scientific Inc.). The bubbles were exposed to a
focused ultrasound pulse (5 MHz) using L9-3 transducer and iU22 ultrasound system
(Philips). In vitro US images were obtained using a bubble-specific, nonlinear imaging
mode. The signal enhancement values were obtained by examining a 5 mm x 5 mm region
of interest and recording the average integrated power within this region for the contrast-
specific (non-linear mode) image. The enhancement was calculated as the ratio of the
integrated power measured in the chamber filled with bubbles to the integrated power
measured in the fashion when the chamber contained only water.
A6.4 In vitro Magnetic Resonance Imaging
Dispersions of bubbles with bubble concentration of approximately 104 bubbles/mL and
with varying amount of Fe3O4 NPs on the bubble surface were introduced into 2 mL
Eppendorf tubes. The tubes were placed in a plastic holder and submerged in water to
eliminate unwanted external susceptibility effects arising from air-water interfaces, which
would also cause a signal decrease on T2-weighted images. Imaging was conducted using a
Philips Achieva 3.0Tesla MRI scanner (Philips Healthcare, Andover, MA) with a 2D
coronal fast field echo sequence (parameters: 128mm field of view, 1mm x 1mm x 5mm
resolution, repetition time 100 ms, flip angle 15̊ , and four different echo times: 4.6, 6.9, 9.2,
132
and 20.7ms). T2* values were fit to this data using MatLab (The MathWorks, Inc., Natick,
MA).
A6.5 Energy Dispersive X-ray (EDX) Line Scan Results for Nanoparticle-
functionalized Bubbles
In addition to the elements of interest, in the control experiments the energy dispersive
X-ray (EDX) scans were carried out for V. Figure A6.2 shows the EDX line scan results for
bubbles encapsulated with a biopolymer shell and functionalized with Fe3O4 nanoparticles
(NPs). We note that the average intensity of EDX signal from V (run in control
experiments) is significantly weaker than that from Fe.
Figure A6.2 (A) Scanning transmission electron microscope (STEM) image of bubbles engulfed with the biopolymer shell and Fe3O4 NPs. Scale bar is 6 μm. (B-F) EDX line scan results for (B) Fe, (C) O, (D) C, (E) N, and (F) V.
133
Figure A6.3 shows the EDX line scan results for bubbles encapsulated with a
biopolymer shell and functionalized with Au NPs. The average intensity of the EDX signal
from V was substantially lower than that from Au.
Figure A6.3 (A) STEM image of bubbles encapsulated with a biopolymer shell and functionalized with Au NPs. Scale bar is 6 μm. (B-F) EDX line scan results for (B) Au, (C) O, (D) C, (E) N, and (F) V.
134
Figure A6.4 shows the EDX line scan results for bubbles engulfed with a biopolymer
shell and functionalized with SiO2-encapsulated CdSe/ZnS NPs. The average intensity of
EDX signal from V as a control was weaker than those from Si, Cd, Se, Zn and S.
Figure A6.4 (A) STEM image of bubbles coated with a biopolymer shell and SiO2-encapsulated CdSe/ZnS NPs. Scale bar is 6 μm. (B-J) EDX line scan results for (B) Si, (C) O, (D) C, (E) N, (F) Cd, (G) Se, (H) Zn, (I) S, and (J) V.
135
A6.6 Long-term Magnetic Actuation of Bubbles Encapsulated with the
Biopolymer Shell and Functionalized with Fe3O4 NPs
Bubbles maintained the ability of magnetic actuation even 3 months after the
preparation as illustrated in Figure A6.5
.
Figure A6.5 Optical microscope image of magnetically actuated bubbles encapsulated with a biopolymer shell and functionalized with Fe3O4 NPs. The image is taken 3 months after the preparation of the bubbles. The bubbles were generated at PCO2=48.3 kPa and QL=8.5 mL/h. Scale bar is 50 μm.
136
A6.7 Photoluminescence (PL) Emission Measurements for Bubbles
Encapsulated with a Biopolymer Shell and Functionalized with SiO2-
capped CdSe/ZnS NPs
Figure A6.6 shows the PL spectrum of the aqueous dispersion of bubbles engulfed with a
biopolymer shell and coated with SiO2-encapsulated CdSe/ZnS NPs. For comparison we
also show the spectra of NP-free bubbles coated with the biopolymer layer and SiO2-
capped CdSe/ZnS NPs. For the NP-coated bubbles the maximum in PL emission of
appeared at λmax=635 nm. The loading of the NPs into the biopolymer shell led to the slight
red shift (~ 5 nm) of the λmax, in comparison with the emission of SiO2-encapsulated
CdSe/ZnS NPs, due to the resonance energy transfer induced by close proximity of the NPs
on the bubble surface.1 NP-free bubbles did not show the characteristic emission peak of the
NPs at 635 nm.
Figure A6.6 FL emission intensity profiles of aqueous dispersions of SiO2-capped CdSe/ZnS NPs (red), bubbles coated with the NPs and biopolymers (blue), bubbles coated with the biopolymers (black). λex = 364 nm.
137
References (1) C. R. Kagan, C. B. Murray, M. G. Bawendi, Physical Review B 1996, 54, 8633.