University of Southern California Dissertations and Theses

Selective deposition of polymer coatings onto structured surfaces

(USC Thesis Other)

Selective deposition of polymer coatings onto structured surfaces

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Selective Deposition of Polymer Coatings
onto Structured Surfaces

Doctor of Philosophy Dissertation
Benny Chen
May, 2016

University of Southern California
Mork Family Department of Chemical Engineering and Materials Science
Los Angeles, CA, USA

Committee Members
Dr. Malancha Gupta (Chair)
Dr. Ellis Meng
Dr. Noah Malmstadt

Executive Summary
In this document, the numerous interesting properties and applications that can be
achieved using complex geometries such as pillar arrays, microbead layers, and porous media,
are described. These properties include, but are not limited to, increased adhesion, optical effects,
self-cleaning properties, and improved separations. Additional interesting properties are
achievable if these complex surfaces are not limited to a single conformation. In order to
manipulate pillar surfaces and alter their conformations, we have self-assembled numerous
hierarchal microstructure systems composed of pillars made from elastomeric materials and
stabilized them using polymer coatings that can act as an adhesive. By using different strategies
to destabilize the polymer adhesive, a multitude of methods for manipulating the pillars were
achieved. The two methods of applying polymer coatings onto these complex geometries are
described, where a vapor phase polymerization system (section 1.4) is used to coat pillar surfaces
in section 2 and porous media in section 4, and a simpler solution casting method is used to
apply polymer on pillar surfaces and microbeads in section 3. Sections 2 and 4 also describe two
methods for patterning polymer on complex geometries, including a photoresist method and the
use of an inhibitor that prevents polymer from depositing is selected locations.
An additional method of applying scaffolded polymer coatings onto complex geometries
is described in section 5, where low vapor pressure liquids are used to direct the deposition of
vapor deposited polymer. We demonstrate two methods of directing the liquid onto surfaces: The
first uses high surface tension liquids to allow for patterning and control over the radial
dimension of the polymer canopies and the second uses low surface tension liquids to allow for
ease of fabrication over larges areas in addition to providing control over the height of the
canopy (section 5.2).

Acknowledgements

The author wishes to express gratitude to his co-authors, Scott Seidel, Hiroki Hori,
Carson T. Riche, Marcus Lehmann, Philip Kwong, and Robert Frank-Finney, as well as the rest
of the Gupta Group for their assistance in the development of this dissertation. The author would
also like to recognize the members who served on the author’s qualifying and dissertation
committees for their guidance. Additionally, the author thanks Dr. Brian J. Jordan, Dr. Vincent
M. Rotello, Dr. Bob Auger, and Dr. Malancha Gupta, for their professional mentorship. Finally,
the author wishes to thank friends and family whose assistance was vital in the personal
development of the author.
5

Contents
1. Introduction ......................................................................................................................................... 7
1.1 Pillar Surfaces ............................................................................................................................ 7
1.2 Capillary Force Manipulation .................................................................................................... 8
1.3 Adhesion .................................................................................................................................. 11
1.4 Initiated Chemical Vapor Deposition ...................................................................................... 13
1.5 References ................................................................................................................................ 14
2. Self-Assembly of Pillars Modified with Vapor Deposited Polymer Coatings .............................. 16
2.1 Introduction .............................................................................................................................. 16
2.2 Results and Discussion ............................................................................................................ 17
2.3 Conclusion ............................................................................................................................... 25
2.4 Experimental ............................................................................................................................ 26
2.5 Acknowledgment ..................................................................................................................... 28
2.6 References ................................................................................................................................ 28
3. Responsive Polymer Welds via Solution Casting for Stabilized Self-Assembly .......................... 29
3.1 Introduction .............................................................................................................................. 29
3.2 Results and Discussion ............................................................................................................ 31
3.3 Conclusion ............................................................................................................................... 42
3.4 Experimental ............................................................................................................................ 42
3.5 Acknowledgements .................................................................................................................. 45
3.6 References ................................................................................................................................ 45
4. Patterned Fluoropolymer Barriers for Containment of Organic Solvents within Paper-Based
Microfluidic Devices ................................................................................................................................. 46
4.1 Introduction .............................................................................................................................. 46
4.2 Results and Discussion ............................................................................................................ 49
4.3 Conclusion ............................................................................................................................... 65
4.4 Experimental ............................................................................................................................ 66
4.5 Acknowledgements .................................................................................................................. 72
4.6 References ................................................................................................................................ 72
5. Fabricating Polymer Canopies onto Structured Surfaces Using Liquid Scaffolds ..................... 74
5.1 Introduction .............................................................................................................................. 74
6

5.2 Results and Discussion ............................................................................................................ 76
5.3 Conclusion ............................................................................................................................... 86
5.4 Experimental ............................................................................................................................ 87
5.5 Acknowledgments.................................................................................................................... 90
5.6 References ................................................................................................................................ 91
6. Conclusions and Future Work ............................................................................................................. 93

7

1. Introduction

1.1 Pillar Surfaces

Structured surfaces have received substantial amounts of attention due to their ability to
exhibit interesting properties. One particular geometry that has been exploited to generate
interesting properties are pillar arrays. By tailoring the geometry of micro or nano-sized pillars,
many studies have been able to produce enhanced properties, including adhesion,
1,2
optical
effects,
3
and hydrophobicity.
4
Although pillar surfaces can exhibit interesting properties, other
benefits can be realized if the pillars are not restricted to a single conformation. For example,
Reddy et. al has shown that surfaces with switchable adherence levels can be fabricated using
pillar surfaces composed of shape memory thermoplastic elastomers.
5
Lower adhesion was
achieved by deforming the pillars on the surface through the application of heat and pressure, but
a heating process could revert the pillars back to an upright conformation and thus increase the
adhesion of the surface.
5
Scanning electron micrograph images depicting the different pillar
conformation changes that yield different degrees of adhesion are shown in Figure 1.

b) a)
8

Figure 1. Scanning electron micrographs that show micropillars in a) an upright conformation exhibiting enhanced
adhesion and b) a deformed conformation exhibiting decreased adhesion.
5
The labels “a)” and “b)” were inserted
into the original image.

Other interesting property changes can be achieved by manipulating the conformation of pillar
surfaces. For example, Chandra and Yang has demonstrated that color changes can be achieve by
collapsing pillars exhibiting Bragg diffraction of light so that they randomly scatter light
creating a whitening effect (Figure 2).
3
In this case, the pillars were collapsed by applying and
evaporating a solvent off the surface, inducing clustering of the ordered pillars on the surface.

Figure 2. (Left) Image of nanopillar surfaces exhibiting Bragg diffraction creating a coloring effect and (Right)
collapsed pillars exhibiting the random scattering of light resulting in a whitening effect.
3

1.2 Capillary Force Manipulation

Due to the many interesting property changes that can be achieved using adjustable pillar
surfaces, many studies have investigated the ability to manipulate pillar surfaces. One interesting
method for manipulating pillar surfaces is through the use of capillary forces induced by solvent
evaporation. Chandra and Yang has performed extensive research on this topic, including the
derivation of models and criteria for self-assembling microstructures from capillary force
9

induced clustering,
3
and the confirmation of lateral capillary meniscus forces responsible for
collapsing 2D arrays of high aspect ratio structures during solvent evaporation
6
in contrast to
previous work that explained the phenomena using Laplace pressures due to isolated capillary
bridges.
7,8,9,10,11
A more recent study by Duan and Berggren has demonstrated the ability to
produce more complicated arrays of microstructures by strategically ordering pillar arrays to
control the sequence in which they collapse during solvent evaporation.
12
Some example of the
microstructures that were self-assembled in this study is shown in Figure 3.
10

Figure 3. (Left) Schematics depicting the ordered arrays of pillars that were sequentially collapsed to produce the
(Right) microstructures shown in the scanning electron micrographs.
12

11

1.3 Adhesion

Depending on the geometry of a pillar array and the material composing the pillars on a
surface, it is possible for a pillar array to resist self-assembly and revert back to an upright
conformation after self-assembly when capillary forces no longer exist. Matsunaga et. al has
performed a study on the importance of adhesion for microstructure formation by applying
different self-assembled monolayers onto their pillars to dictate the level of adhesion. Depending
on the affinity between the functionalities on the self-assembled monolayers, different degrees of
successful self-assembly between pillar structures were observed. Figure 4 shows optical
microscope images of pillars modified with various functionalities that have been treated with
ethanol to induce capillary force self-assembly. Depending on the functional groups on the self-
assembled monolayer, different quantities of microstructures were stabilized through adhesion
after self-assembly, where the stronger the affinity between the functional groups the greater
quantity of microstructures were stabilized.
13
This study has brought attention to how adhesion is
an important limiting factor of microstructure self-assembly. In cases where pillars are not
permanently deformed during self-assembly, even if capillary forces are capable of drawing a
cluster of pillars together into a microstructure formation, if the adhesive force present is too
weak to prevent the pillars from reverting back to an upright position after solvent evaporation,
then adhesion becomes the limiting factor during microstructure formation.
12

Figure 4. Optical microscope images of pillar arrays modified with various self-assembled monolayers that have
been rinsed with ethanol to induce capillary force self-assembly of microstructures. The degree of successful self-
assembly is dependent on the affinity between the functional groups on the self-assembled monolayer, where shorter
carbon chains yield decreased successful self-assembly unless the terminating functional groups are capable
hydrogen bonding.

In order to provide an adhesive force into our system, we introduce polymer which can
act as an adhesive layer by forming a polymer weld between pillars when they come into contact
during self-assembly. Polymer welds form when polymer between or on two surfaces is soften
with a solvent and mobilized so that the chains intertwine and entangle resulting in adhesion or
cohesion between the two surfaces.
14
For example, polymer welds have been used to bond
13

microfluidic devices.
15,16
In order to introduce polymer to form a weld between pillars, we have
demonstrated two different techniques which can be used to successfully self-assemble polymer
welded microstructures. The first technique that will be discussed introduces polymer by first
applying a polymer coating using initiated chemical vapor deposition (iCVD) on the pillars prior
to applying a solvent to induce self-assembly. The second technique is a more facile approach,
where a solution containing the welding polymer is directly applied to a pillar surface to induce
self-assembly and form polymer welds. Each technique has its own inherent advantages and
disadvantages which will be discussed in their respective sections.
1.4 Initiated Chemical Vapor Deposition

The iCVD process is a vapor phase process that can apply a wide range of polymeric
coatings using mild reaction conditions. In the iCVD process, initiator and monomer reactants
are flown into a reactor chamber as shown in Figure 5. Initiator molecules come into contact
with the heated filament array causing it to thermally decompose into free radicals. These free
radical molecules can then combine with adsorbed monomer molecules on the stage and begin
free radical polymerization. Typically the samples of interest that are being coated are situated
on the stage where a recirculating chiller is used to control the temperature of the samples to help
dictate adsorption of molecules onto the samples. The solventless nature of the technique makes
it a powerful tool for applying conformal coatings onto surfaces with complex features or
intricate porosities due to the freedom from common solvent related issues such as solubility,
surface tension issues leading to defects or clogging, or environmental incompatibility. These
advantages have resulted in the successful conformal coating of non-planar substrates such as
electrospun fiber mats
17,18
and silicon nanotrenches.
19
Additionally, the iCVD process has been
shown to produce a variety of functional coatings such as light responsive
20
and temperature
14

responsive films,
21
as well as polymers with click-active moieties,
22
which provides an
abundance of chemical flexibility.

Figure 5. Schematic showing the iCVD process, where initiator and monomer are flown into a reactor where
initiator molecules can be decomposed into radicals which can then combine with adsorbed monomer on a surface to
begin free radical polymerization.

1.5 References

(1) Huber, G. ; Mantz, H.; Spolenak, R.; Meche, K.; Jabobs, K.; Gorb, S. N.; Arzt, E. Proceedings of the National
Academy of Sciences 2005, 102, 16293-16296.
(2) Lee, H.; Lee, B. P.; Messersmith, P. B. Nature 2007, 448, 338-342.
(3) Chandra, D.; Yang, S. Acc. Chem. Res. 2010, 43, 1080–1091.
(4) Sun, T.; Feng, L.; Gao, X.; Jiang, L. Acc. Chem. Res. 2005, 38, 644-652.
(5) Reddy, S.; Arzt, E.; del Campo, A. Adv. Mater. 2007, 19, 3833–3837.
(6) Chandra, D.; Yang, S. Langmuir 2009, 25, 10430-10434.
(7) Chandra, D.; Taylor, J. A.; Yang, S. Soft Matter 2008, 4, 979.
(8) del Campo, A.; Greiner, C. J. Micromech. Microeng. 2007, 17, R81.
(9) Segawa, H.; Yamaguchi, S.; Yamazaki, Y.; Yano, T.; Shibata, S.; Misawa, H. Appl. Phys. A: Mater. Sci.
Process. 2006, 83, 447.
(10) Chakrapani, N.; Wei, B.; Carrillo, A.; Ajayan, P. M.; Kane, R. S. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
4009.
(11) Lee, T. W.; Mitrofanov, O.; Hsu, J. W. R. Adv. Funct. Mater. 2005, 15(10), 1683.
(12) Duan, H.; Berggren, K. K. Nano Letters 2010, 10, 3710-3716.
(13) Matsunaga, M.; Aizenberg, M.; Aizenberg, J. J. Am. Chem. Soc. 2011, 133, 5545–5553.
I-I
M
2I•
M M
I-I
I-M-M-M•
initiator
monomer
vacuum
pump
I•
Stage
Filament array
15

(14) Fisher, L. W. In Selection of Engineering Materials and Adhesives; CRC Press, Taylor & Francis Group: Boca
Raton, 2005; p 473-477.
(15) Shah, J. J.; Geist, J.; Locascio, L. E.; Gaitan, M.; Rao, M. V.; Vreeland, W. N. Anal.Chem. 2006, 78, 3348-
3353.
(16) Klank, H.; Kutter, J. P.; Geschke, O. Lab Chip 2002, 2, 242-246.
(17) Ma, M.; Mao, Y.; Gupta, M.; Gleason, K. K.; Rutledge, G. C. Macromolecules 2005, 38, 9742-9748.
(18) Ma, M.; Gupta, M.; Li, Z.; Zhai, L.; Gleason, K. K.; Cohen, R. E.; Rubner, M. F.; Rutledge, G. C. Adv. Mater.
2007, 19, 255-259.
(19) Tenhaeff, W. E.; McIntosh, L. D.; Gleason, K. K. Adv. Funct. Mater. 2010, 20, 1144-1151.
(20) Haller, P. D.; Flowers, C. A.; Gupta, M. Soft Matter 2011, 7, 2428-2432.
(21) Alf, M. E.; Godfrin, P. D.; Hatton, T. A.; Gleason, K. K. Macromol. Rapid Commun.2010, 31, 2166-2172.
(22) O’Shaughnessy, W. S.; Marı´-Buye´, N.; Borro´s, S.; Gleason, K. K. Macromol. Rapid Commun. 2007, 28,
1877-1882.
16

2. Self-Assembly of Pillars Modified with Vapor Deposited Polymer
Coatings

2.1 Introduction

Nature has demonstrated that interesting surface properties can be achieved by
optimizing surface chemistry and surface roughness. For example, the
superhydrophobicity of the Lotus leaf is achieved through a combination of nanometer
and micron sized structures on the leaf and a low surface energy wax coating.
4

Microstructured and nanostructured pillars on the feet of geckos enable these lizards to
climb up walls.
1
These biological surfaces have inspired researchers to design periodic
structures that exhibit self-cleaning
23,24
and adhesive
25,26,27
properties.
Capillary forces can be used to self-assemble pillars, needles, and nanotubes into
arrays of hierarchical structures.
28,3,29,30,31,32
This self-assembly process is controlled by a
competition between elastic forces and capillary forces. The shape and periodicity of the
self-assembled structures can be tuned by changing the aspect ratio, spacing, tilt, and
elastic modulus of the pillars.
33,12
In most studies, the material used to fabricate the pillars
determines both the elastic modulus and the surface energy of the pillars. For example,
Yang and coworkers fabricated pillars using different compositions of copolymers and
both the elastic modulus and surface energy changed with composition.
34

In our document, we separate the effect of surface chemistry from the effect of the
elastic modulus by using initiated chemical vapor deposition (iCVD) to coat elastomeric
poly(dimethylsiloxane) (PDMS) pillars with thin layers of polymer coatings. In the iCVD
17

process, monomer and initiator vapors are introduced into a vacuum chamber where a heated
filament array decomposes the initiator into free radicals. The free radicals and monomer
molecules adsorb onto the surface of a cooled substrate where polymerization occurs via a free
radical chain mechanism.
35,36,37,38
The key advantage of using iCVD over liquid-phase processing
is there are no solvent tension effects such as wetting or de-wetting and therefore the process can
be used to uniformly coat nonplanar, curved, and porous substrates. For example, iCVD has been
used to deposit uniform fluoropolymer coatings onto electrospun fiber mats
39,40
and
membranes.
41

In this paper, we show that polymer coatings deposited via iCVD can be used to self-
assemble PDMS pillars into perfect clusters and light-sensitive coatings can be used to pattern
the location of self-assembly. Aizenberg and coworkers recently modified the surfaces of epoxy
pillars with short and long chain self-assembled monolayers (SAMs) in order to study the role of
adhesion in capillary-forced self-assembly.
42
They found that the chain length and functionality
affected the adhesion between the pillars and thereby affected the stability and reversibility of the
resulting clusters. Our study confirms their observation that adhesion is important in determining
the stability of self-assembled microstructures. However, our method introduces the use of
polymer welds to stabilize the microstructures. These polymer welds can be exploited to tune the
response of the microstructures to various solvents.
2.2 Results and Discussion

Polymer coatings were deposited onto elastomeric PDMS pillars using iCVD as shown in
Figure 6a. The coated pillars were self-assembled using water as the solvent. In order for
capillary meniscus forces to collapse the pillars, the water must be able to penetrate between the
18

pillars. As shown in Figure 6b, the uncoated PDMS pillars are hydrophobic therefore
water cannot penetrate between the pillars to cause collapse. Deposition of
poly(hydroxyethyl methacrylate) (PHEMA) or poly(methacrylic acid) (PMAA) causes
the pillars to become hydrophilic and water easily penetrates between the pillars. One
distinct advantage of the iCVD process is that different layers of polymer coatings can be
stacked onto the pillars without removal of the substrate from the reaction chamber. The
hydrophilic pillars can be reverted back to a hydrophobic state by the addition of a thin
layer of poly(ortho-nitrobenzyl methacrylate) (PoNBMA) onto the PHEMA or PMAA
layer.

Figure 6. a) Schematic of the initiated chemical vapor deposition process used to deposit thin layers of polymer onto
PDMS pillars. b) The uncoated PDMS pillars are hydrophobic and water cannot penetrate between the pillars. Water
wets pillars coated with PHEMA and PMAA. The hydrophilic pillars can be reverted back to a hydrophobic state by
the addition of a thin layer of PoNBMA.
The self-assembly of pillars coated with PHEMA and PMAA was examined by
submerging the pillars in water and allowing the water to evaporate. Pillars with a
diameter of 22 μm, edge-to-edge spacing of 18 μm, and heights of 34, 54, and 65 μm
19

were tested. The height of the pillars had a large effect on the self-assembly process. In order to
collapse the pillars, the capillary force must be greater than the elastic restoring force. The force
required to bend a pillar scales as F~Er
4
s/h
3
where E is Young’s modulus, r is the radius of the
pillar, s is the distance between the pillars, and h is the height of the pillar.
33
If the pillars are
below a critical height, the capillary force will not be strong enough to bend the pillars far
enough to cause physical contact. Figure 7 shows a time series of the collapse of 34 μm tall
pillars coated with a thin layer of PHEMA. Although the pillars begin to bend into
microstructures as the water evaporates and menisci form, the capillary force is not strong
enough to cause physical contact between the pillars and the pillars are fully restored to their
upright position after complete evaporation.

Figure 7. Microscope images of 34 μm tall pillars coated with a layer of PHEMA. a) Dry pillar array. b) Pillar array
submerged in water. c) Menisci start to form as water evaporates. d) The pillars begin to bend into microstructures
but do not make physical contact. e) The pillars begin to restore into their original formation. f) The pillars are
restored to their upright position after complete evaporation.
0s 160s 210s
320s 350s 380s
a) b) c)
d) e) f)
80μm
20

When the pillar height is increased to 54 μm, evaporation yields perfect clusters of four pillars
(Figure 8). The arrays are defect-free over large areas for both PHEMA and PMAA coatings. At
a pillar height of 65 μm, there were several clusters of six pillars in addition to clusters of four
pillars.
An adhesive force is required to stabilize the microstructures during self-assembly.
Therefore two criteria must be satisfied to form stable clusters: the height of the pillars must be
great enough to cause physical contact and the adhesive force must be strong enough to resist the
elastic restoring force. The relationship between pillar height and cluster size has been previously
confirmed.
6,34
The novelty of our paper is the demonstration that polymer welds can be formed
through solvent bonding to provide an adhesive force that overcomes the elastic restoring force
thereby enabling the formation of stable clusters. In the absence of adhesive forces, it is
expected that the collapsed pillars should revert back to their upright position after complete
evaporation and disappearance of the capillary forces. Close-up SEM images of the pillars
(Figures 8e and 8f) show that our pillars remain collapsed after complete evaporation. Visible
polymer welds connect the edges of adjacent pillars. These polymer weld forms when a solvent
softens and mobilizes polymer chains and enables them to interdiffuse.
14
These types of polymer
welds are typically exploited to bond polymer pieces composed of the same material together.
For example, microfluidic devices can be made by bonding two pieces of poly(methyl
methacrylate) together using acetone
15
and ethanol
16
as the solvent. In our system, water is able
to soften the PMAA and PHEMA coatings while capillary forces bring the pillars into contact
which allows the polymer chains on adjacent pillars to form a polymer weld. For thin coatings,
polymer welds are formed between adjacent pillars (Figure 8e) while for thick coatings, polymer
welds form between the pillars and within the center of the cluster (Figure 8f).
21

Figure 8. Large arrays of stable microstructures were formed by coating 54 μm tall pillars with a) 130 nm of
PHEMA, b) 800 nm of PHEMA, c) 80 nm of PMAA, and d) 250 nm of PMAA. e,f) Close-up views of samples c
and d show polymer welds between the pillars.

22

Stabilization by solvent bonding allows us to tailor the response of the self-
assembled microstructures to different solvent environments. If the polymer coating is
incompatible in the solvent, the pillars will remain collapsed. If the polymer coating is
compatible in the solvent, the microstructures will open and revert back to an upright
position while submerged in the solvent and then collapse back into a cluster as the
solvent evaporates. The amount of cycles of opening and closing depends on the solvent.
Since the polymer coating becomes mobile during submersion, bare PDMS can be
exposed after repeated use. The cycling can be extended by using organic solvents
because of their ability to wet PDMS.
The time required for the microstructures to revert back to an upright position can
be tuned by blending solvents. As shown in Table 1, the opening time of the pillars can
be controlled by submerging microstructures formed using PHEMA coatings in mixtures
of acetonitrile in methanol (MeOH). The fastest rate was achieved using 75% w/w
acetonitrile in MeOH. At this composition, the Hansen solubility parameter of the blend
is similar to the Hansen solubility parameter of PHEMA (acetonitrile
43
: δd = 15.3, δp =
18.0, δh = 6.1 MPa
1/2
, MeOH
43
: δd = 15.1, δp = 12.2, δh = 22.24 MPa
1/2
, and PHEMA
44
: δd
= 15.14 + 0.68, δp = 11.87 + 0.33, δh = 18.84 + 0.37 MPa
1/2
). The amount of time
required for the microstructures to open can also be tuned by changing the thickness of
the coating. As shown in Table 1, increasing the thickness of the coating lengthens the
opening time of the microstructures. Decreasing the thickness of PHEMA to
approximately 100 nm allows the microstructures to open almost instantaneously (<2
seconds) when submerged in MeOH or acetonitrile.

23

Table 1. Opening Time in Blends of Acetonitrile in Methanol (% w/w)

opening time (s)
PHEMA coating thickness (nm) MeOH 25% 50% 75% acetonitrile
900 1190 1070 920 15 260
1900 4150 2990 2170 170 560

We examined whether we could control the location of self-assembly by patterning a
hydrophobic polymer layer onto a hydrophilic layer. Figure 9a shows a schematic of our
patterning process. First a layer of PHEMA was deposited onto the PDMS pillars and then a
layer of PoNBMA was deposited onto the PHEMA layer. The hydrophobic PoNBMA layer was
then selectively removed by exposure to UV-light through a mask. Exposure to UV-light cleaves
the nitrobenzyl moieties and converts the exposed PoNBMA into PMAA which can be removed
in pH 8 buffer solution.
20
The exposed pillars become hydrophilic due to the underlying PHEMA
coating whereas the unexposed pillars remain hydrophobic due to the hydrophobic PoNBMA
coating. Figure 9b shows that self-assembly only occurs in the hydrophilic regions.
24

Figure 9. a) Schematic of the process used to control the location of self-assembly. b) Stereoscope image showing
square regions of collapsed pillars.

It may be possible to use these microstructures for synthesis within flow systems.
For example, catalysts and solid reactants can be entrapped within and released from the
microstructures by flowing different solvents through a channel. Figure 10 shows our
ability to capture beads within our microstructures. When a compatible solvent was
applied to the surface, the microbeads were observed to eject from the microstructure
allowing them to flow more freely.
25

Figure 10. SEM image of a polystyrene bead captured within a microstructure.

2.3 Conclusion

We demonstrated that iCVD can be used to deposit thin layers of polymer coatings onto
pillars and these coatings stabilize self-assembled microstructures by providing an adhesive force
through solvent bonding. Furthermore, the location of self-assembly can be controlled through
the use of a light-responsive coating. The coating process described in this paper is substrate-
independent and therefore can be applied to pillars composed of any material (epoxy, PMMA,
carbon nanotubes, etc.). Although micron-sized pillars were coated in this paper, the coating
process can be extended to nanometer-sized pillars since iCVD has been shown to be effective
for coating extremely small dimensions.
45
Future work will determine if temperature-responsive
poly(N-isopropylacrylamide) (PNIPAAm) coatings can be used to pattern the location of self-
assembly using heat.
46

26

The microstructures formed in this study have several potential applications in fields as
diverse as microfluidics and drug delivery because the response of the microstructures can be
tuned by the thickness of the polymer coating and the solubility parameter of the solvent.
Aizenberg and coworkers have suggested that clusters of pillars might be useful for capturing
and releasing particles for controlled drug release.
42
Our microstructures are fabricated using
biocompatible materials such as PDMS and PHEMA and the tunable time release can be used to
control drug doses over a period of time or to sequentially administer multiple drugs.
2.4 Experimental

Polydimethylsiloxane (PDMS) pillars were made using soft lithography.
47,48
SU-8 2050
photoresist (MicroChem) was patterned onto a silicon wafer using UV-light. The pattern
consisted of arrays of circular wells. Sylgard® 184 (10:1 base: crosslinker ratio) was poured onto
the SU-8 mold and cured in an oven at 60°C overnight. The surface of the SU-8 was treated
with trichloro(1H,1H,2H,2H-perfluorooctyl) silane (Sigma Aldrich) to allow the cured PDMS
pillars to be easily released.
The PDMS pillars were placed inside a custom designed reaction chamber (GVD
Corporation, 250 mm in diameter, 48 mm in height). For poly(2-hydroxyethyl methacrylate)
(PHEMA) depositions, 2-hydroxyethylmethacrylate (99%, Sigma Aldrich) was heated to 65°C
and flowed into the reactor at a rate of 0.5 sccm, the reactor pressure was maintained at 0.14
Torr, and the reactor stage was kept at 35°C. For poly(methacrylic acid) (PMAA) depositions,
methacrylic acid (99%, Sigma Aldrich) was heated to 60°C and flowed into the reactor at a rate
of 20 sccm, the reactor pressure was maintained at 0.76 Torr, and the reactor stage was kept at
25°C. For poly(o-nitrobenzyl methacrylate) (PoNBMA) depositions, o-nitrobenzyl methacrylate
27

(95%, Polysciences) was heated to 110°C and flowed into the reactor at a rate of 0.1 sccm, the
reactor pressure was maintained at 0.12 Torr, and the reactor stage was kept at 20°C. For all
depositions, tert-butyl peroxide (TBPO) (Aldrich, 98%) was kept at room temperature and
flowed into the reactor at a rate of 0.8 sccm. A nichrome filament array (80% Ni, 20% Cr,
Omega Engineering) was resistively heated to 200°C and the distance between the filament array
and the substrate was kept constant at 32 mm. Film thicknesses on reference silicon wafers were
determined using interferometry. Contact angle goniometry (ramé-hart Model 290-F1) was used
to study the wetting properties of the coated pillars.
Pillars coated with PHEMA and PMAA were collapsed into microstructures by
dispensing 5 μL of deionized water onto the coated PDMS pillars and allowing the water to
evaporate in a covered petri dish at room temperature. Patterned pillar samples were fabricated
by coating PDMS pillars with PHEMA followed by PoNBMA. The pillars were then exposed to
365 nm UV light for 1 hr under a transparency mask. After exposure, the samples were
developed for 30 s in pH8 buffer (BDH) and allowed to dry in a covered petri dish. The
tunability of the microstructures was tested using ~0.5 cm
2
microstructure arrays. Residual
solvent was removed by placing the samples in a vacuum chamber overnight. The
microstructures were then submerged in acetonitrile (Malinckrodt Chemicals), methanol
(Malinckrodt Chemicals), or a blend of both. Polystyrene beads (20 μm in diameter, 2.5% w/v in
water, Polysciences) were mixed at a 1:2 ratio with a 0.25% v/v solution of Triton X-100 (Sigma
Aldrich) in methanol. In order to capture the beads within the microstructures, 5 L of this
solution was dispensed onto a 77 m tall needle array (0.5 cm
2
) coated with 500 nm of PHEMA.
Images were taken using an optical microscope (Meiji ML8000), a stereoscope (National Optical
420 Stereo Zoom), and a low-vacuum scanning electron microscope (JEOL-7001). A thin gold
28

coating was sputtered onto the surface of the samples prior to imaging with the scanning electron
microscope.
2.5 Acknowledgment

This work was supported by the James H. Zumberge Faculty Research and Innovation
Fund, the Annenberg Graduate Fellowship (B.C.), and the Viterbi School of Engineering
Fellowship (S.S.).

2.6 References

(23) Patankar, N.A. Langmuir 2004, 20, 7097-7102.
(24) Otten, A.; Herminghaus, S. Langmuir 2004, 20, 2405-2408.
(25) Geim, A. K. ; Dubonos, S. V.; Grigorieva, I. V.; Novoselov, K. S.; Zhukov, A. A.; Shapoval, S. Y. Nature
Materials 2003, 2, 461-463.
(26) Greiner, C.; del Campo, A.; Arzt, E. Langmuir 2007, 23, 10235-10243.
(27) Lee, H.; Lee, B. P.; Messersmith, P. B. Nature 2007, 448, 338-341.
(28) Wu, D.; Chen, Q.-D.; Xu, B.-B.; Jiao, J.; Xu, Y.; Xia, H.; Sun, H.-B. Applied Physics Letters 2009, 95, 091902.
(29) De Volder, M.; Park, S. J.; Tawfick, S. H.; Vidaud, D. O.; Hart, J. J. Micromech. Microeng. 2011, 21, 04033.
(30) De Volder, M.; Tawfick, S. H.; Park, S. J.; Copic, D.; Zhao, Z.; Lu, W.; Hart, J. Adv. Mater. 2010, 22, 4384-
4389.
(31) Lim, X.; Foo, H. W. G; Chia, G. H.; Sow, C. H.; ACS Nano 2010, 4, 1067-1075.
(32) Futaba, D. N; Hata, K.; Yamada, T.; Hikaoka, T.; Hayamizu, Y.; Kakudate, Y.; Tanaike, O.; Hatori, H.;
Yumura, M.; Iijima, S. Nat. Mater. 2006, 5, 987-994.
(33) Pokroy, B.; Kang, S. H.; Mahadevan, L.; Aizenberg, J. Science 2009, 323, 237-240.
(34) Chandra, D.; Yang, S. Acc. Chem. Res. 2010, 43, 1080-1091.
(35)
(36) Gupta, M.; Gleason, K. K. Langmuir 2006, 22, 10047-10052.
(37) Chan, K.; Gleason, K. K. Langmuir 2005, 21, 8930-8939.
(38) Xu, J.; Gleason, K. K. Chemistry of Materials 2010, 22, 1732-1738.
(39) Ma, M.; Mao, Y.; Gupta, M.; Gleason, K. K.; Rutledge, G. C. Macromolecules 2005, 38, 9742-9748.
(40) Ma. M.; Gupta, M.; Li, Z.; Zhai, L.; Gleason, K. K.; Cohen, R. E.; Rubner, M. F.; Rutledge, G. C. Advanced
Materials 2007, 19, 255-259.
(41) Gupta, M.; Kapur, V.; Pinkerton, N. M.; Gleason, K. K. Chemistry of Materials 2008, 20, 1646-1651.
(42) Matsunaga, M.; Aizenberg, M.; Aizenberg, J. J. Am. Chem. Soc. 2011, 133, 5545-5553.
(43) Hansen, C. In The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient: Their
Importance In Surface Coating Formation; Danish Technical Press: Copenhagen, 1967; p 18-19.
(44) Çaykara, T.; Özyürek, C.; Kantoğlu, Ö.;Güven, O. J. Polym. Sci. Phys. Ed. 2002, 40, 1995-2003.
(45) Ince, G. O.; Demirel, G.; Gleason, K. K.; Demirel, M. C. Soft Matter 2010, 6, 1635-1639.
(46) Alf, M. E.; Godfrin, P. D.; Hatton, T. A.; Gleason, K. K. Macromol. Rapid Commun. 2010, 31, 2166-2172.
(47) Zhao, X.-M.; Xia, Y.; Whitesides, G. M. J. Mater. Chem. 1997, 7, 1069-1074.
(48) Xia, Y.; Whitesides, G. M. Annu. Rev. Mater. Sci. 1998, 28, 153-184.
29

3. Responsive Polymer Welds via Solution Casting for Stabilized Self-
Assembly

3.1 Introduction

The ability to control the formation and stabilization of complex structured arrays is
important for device fabrication. For example, capillary forces induced by molten alloys resulted
in the self-assembly and bonding of microscale components to form electrical connections.
49

Biomimetic mushroom-shaped adhesive microstructures, inspired by beetles, have been
fabricated by molding polymer within a template.
50
Photonic crystals have been fabricated from
ordered arrays of spherical metal nanoparticles.
51
Other interesting applications can be realized if
the structures are not restricted to a single configuration. For example, ultrathin optical devices
can be created by transitioning between a periodic array of pillars exhibiting Bragg diffraction
and self-assembled clusters of pillars that scatter light randomly to create a whitening effect,
3

similar to that of the white beetle.
52
Additionally, the reversible adhesion of gecko feet has been
replicated in biomimetic studies.
53,54
Matsunaga et al. has developed actuatable microstructures using self-assembled
monolayers (SAMs) as a reversible adhesive layer,
13
and we have recently demonstrated that
polymer coatings deposited via vapor phase polymerization (VPP) can be combined with
appropriate solvents to control the assembly and disassembly of microstructures.
55
The formation
of SAMs requires the use of specific chemistries between the substrate and corresponding head
groups, while VPP requires specialized and expensive equipment (mass flow controllers, pumps,
and heat exchangers), limiting these techniques to a few laboratories. In this paper, we
30

demonstrate for the first time a cheaper and simpler solution casting method to form
polymer welds that stabilize self-assembled features. A polymer solution is applied to a
system of discrete objects and then allowed to evaporate. During evaporation, capillary
forces pull the objects together and induce self-assembly. As the polymer solution
becomes increasingly concentrated, the polymer chains interdiffuse and solidify to form a
polymer weld that stabilizes the self-assembled structures. An added benefit of our
method compared to the VPP technique is that virtually any polymer can be used as long
as a suitable solvent is utilized. Responsive systems can be formed using polymers that
are known to experience drastic property changes due to environmental stimuli such as
temperature,
56
pH,
57
light,
58
and electrical signal.
59
We demonstrate the versatility of our
technique by forming pH-responsive and temperature-responsive microstructures from
arrays of poly(dimethylsiloxane) (PDMS) pillars and by adhering and selectively
releasing poly(styrene) microbeads within a microfluidic channel. A diverse set of
polymers can be solution casted including biocompatible materials such as alginate and
poly(2-hydroxyethyl methacrylate) which allows our technique to be extended to
biomedical applications such as tissue engineering and drug delivery. Furthermore, the
simplicity of our experimental procedure allows it to be used universally in every
laboratory.

31

3.2 Results and Discussion

A simple polymer solution casting technique was used to self-assemble pillars into
microstructures stabilized by polymer welds. The welds are formed by taking advantage of
capillary forces combined with interdiffusion and solidification of the polymer during solvent
evaporation. The general process is outlined in Figure 11. As the solvent evaporates, the liquid
level falls below the height of the pillars, resulting in capillary forces that pull the pillars into
contact.

Figure 11. Schematic showing the self-assembly of pillars and the formation of polymer welds by solution casting.

32

The formation of microstructures requires that the capillary forces be strong
enough to cause the pillars to make contact. The capillary force on an individual pillar in
a system of four pillars is given by
6

𝐹 𝑐𝑎𝑝 =
𝜋𝛾 𝑑 2
𝑐𝑜𝑠 2
𝜃 2
(√
2
( 𝑝 − 2𝛿 )
2
− 𝑑 2
+ √
1
2( 𝑝 − 2𝛿 )
2
− 𝑑 2
) ( 1)
where γ is the liquid-vapor surface tension, d is the diameter of the pillars, θ is the contact angle
of the solution on the pillar material, p is the distance between the centers of adjacent non-
diagonal pillars, and δ is the deflection distance of the pillar. As the pillars deflect, the capillary
forces are resisted by the net elastic restoring force of the pillars. The elastic restoring force of a
single pillar is given by
60

𝐹 𝑒𝑙𝑎𝑠𝑡𝑖𝑐 =
3√ 2𝜋𝐸 𝑑 4
𝛿 64ℎ
3
( 2)
where E is the elastic modulus of the pillar material and h is the height of the pillar. If the
capillary force is large enough to overcome the elastic restoring force, adjacent pillars will be
brought into contact with one another. Once the solvent completely evaporates, capillary forces
are no longer present to overcome the elastic restoring force; therefore, an adhesive force must
be introduced to stabilize the microstructures. In our system, the adhesive force is provided by a
polymer weld that forms at the interface between the pillars while they are in contact. Figure 12
shows our ability to form stable microstructures by solution casting nonpolar, polar, and charged
polymers of various molecular weights (MW) onto an array of PDMS pillars. 10 μL quantities of
0.5 % w/w poly(styrene) (4,000 MW) in acetone, poly(methyl methacrylate) (50,000 MW) in
acetone, poly(2-hydroxyethyl methacrylate) (20,000 MW) in methanol, and alginate in water
were solution casted onto 0.7 × 0.7 cm arrays of PDMS pillars. The PDMS pillars were 60 μm in
33

height and 22 μm in diameter with a spacing of 18 μm between the edges of the pillars as
measured from scanning electron microscopy (SEM) images.

Figure 12. SEM images of PDMS microstructures stabilized by polymer welds formed from a) poly(styrene), b)
poly(methyl methacrylate), c) poly(2-hydroxyethyl methacrylate), and d) alginate. All scale bars represent 100 μm
in length.

Depending on the substrate and solvent combination used during solution casting, wettability
may be an issue. In order to overcome surface tension effects, we irradiated our PDMS pillars
with air plasma using a hand-held corona generator to increase hydrophilicity prior to wetting the
pillars with aqueous solutions.
34

We can control the quality of adhesion via the solution casting method. The
polymer weld must resist the total elastic restoring force of all the pillars forming the
microstructure. Therefore, higher-ordered microstructures require a stronger polymer
weld to counteract the increased elastic restoring force due to additional contributing
pillars and longer deflection distances. For example, we have observed that during self-
assembly, a microstructure formed by a cluster of four pillars will detach into two
clusters of two pillars when a polymer weld is too weak. The sizes of microstructures
formed are indicative of the relative strength of the polymer welds. We tested the
sensitivity of the polymer welds to changes in concentration and molecular weight as
shown in Figure 13. Starting near solution saturation, 10 μL quantities of 500,000 MW
poly(methyl methacrylate) in acetone at weight percentages of 5 × 10
-1
down to 5 × 10
-5

were applied onto the arrays of PDMS pillars. Solutions with weight percentages of 5 ×
10
-1
and 5 × 10
-2
consistently produced arrays of microstructures predominantly
consisting of clusters of four pillars. At 5 × 10
-3
% w/w, the adhesion weakened and we
observed several detached clusters of four pillars yielding clusters of two pillars while
some pillars remained singular, failing to form microstructures of any order. At lower
dilutions, the adhesion was too weak to stabilize clusters of any size. The increased
adhesion at greater polymer concentrations is likely due to the formation of thicker welds,
which has been confirmed by SEM. These experiments were repeated using lower
molecular weights (50,000 MW and 5,000 MW) at the same dilutions. As shown in
Figure 13, at 5 × 10
-2
% w/w the 50,000 MW solution continued to produce clusters of
four pillars; whereas, the 5,000 MW solution produced clusters of two and four pillars.
Diluting the poly(methyl methacrylate) concentration below 5 × 10
-3
% w/w yielded less
35

clusters for both 5,000 MW and 50,000 MW compared to 500,000 MW. Further dilution failed
to produce any significant number of structures. These results indicate that solutions with lower
molecular weight polymers produce less microstructures compared to solutions with higher
molecular weight polymers at similar weight percentages.

Figure 13. Microscope images of PDMS microstructures showing increasing cluster formation with greater
molecular weight and concentration.

SEM images of these structures revealed that thicker welds are formed when larger molecular
weights are used. The formation of thicker welds is likely due to increased viscosities at higher
molecular weights
61
resulting in greater polymer accumulation at the pillar interface upon
evaporation, similar to dip coating processes where greater viscous forces lead to thicker films.
62

Generally, polymer welds are formed by placing a solvent between two surfaces of the
same material and applying pressure to form a weld composed of the same material as the
surface. For example, poly(methyl methacrylate) microfluidic channels can be bonded by placing
acetone
63
or a mixture of dichloroethane and ethanol
64
between the surfaces. In our system, two
36

similar surfaces are adhered together by a dissimilar polymer. The polymer must be compatible
with the substrate in order to form a strong polymer weld. To examine this, we attempted to use
our solution casting technique on PDMS pillars coated with a low surface energy fluoropolymer,
poly(1H,1H,2H,2H-perfluorodecyl acrylate-co-ethylene glycol diacrylate) (P(PFDA-co-EGDA)).
These coated pillars resisted the formation of polymer welds in all tested polymer-solvent
systems which included 0.5 % w/w solutions of poly(styrene) in acetone, poly(methyl
methacrylate) in acetone, poly(2-hydroxyethyl methacrylate) in methanol, and alginate in water.
A possible concern is that the fluorinated coating may change the capillary forces and thereby
prevent the pillars from making contact. However, we used optical microscopy to confirm that
the pillars made contact and then separated as capillary forces diminished in each case. The
fluoropolymer coating prevents the diffusion of the solvent into the PDMS;
65
therefore, the
PDMS remains immobilized during the welding process. C. Y. Yue has shown that the adhesion
of dissimilar polymer welds such as poly(vinyl chloride)-poly(methyl methacrylate) and
poly(vinyl chloride)-poly(carbonate) are significantly enhanced when both interfaces are
mobilized during the solvent welding process.
66
By preventing polymer interdiffusion with the
PDMS substrate, adhesion depends only on the interaction between the P(PFDA-co-EGDA)
surface and the solution casted polymer, which is weak due to the fluorinated pendant groups in
the coating.
A major advantage of polymer welded microstructures is the ability to use specific
external stimuli to reverse the self-assembly. The microstructures remain intact when washed
with an incompatible solvent and disassemble when the polymer weld is dissolved by a
compatible solvent. For example, microstructures formed using poly(styrene) remained stable
when washed with methanol, but disassembled when washed with acetone. However, polymer
37

weld dissolution is not restricted to weak intermolecular interactions. For example, electrostatic
interactions can be used to control the dissolution of a weld with pH-sensitive moieties. This
concept is demonstrated in Figure 14 where reversible pH-responsive microstructures were
created by solution casting poly(methacrylic acid) (100,000 MW) in methanol onto shorter 54
μm tall PDMS pillars. Shorter pillar heights were used in order to increase the elastic restoring
force which resulted in faster disassembly rates. To test the dissolution of our polymer welds at
low pH, we submerged the stabilized microstructures in pH 4 buffer for 3 hours, after which the
polymer welds remained intact, allowing the microstructures to remain assembled. When these
microstructures were submerged in neutral water, the poly(methacrylic acid) dissociated causing
the welds to dissolve and the microstructures to revert back to an upright position. The
disassembly process using neutral water occurred within 30 minutes, but could be expedited to
approximately 1 minute by submerging the structures in pH 8. After disassembly, the pillars
reassembled back into microstructures during evaporation of the buffer. Prevention of
reassembly can be achieved by washing the samples with copious amounts of pH 8 buffer to
remove the polymer.
Certain devices, such as microelectronics,
67,68
are sensitive to the presence of solvents,
and therefore, a solventless method for actuation may be desired.

When a linear polymer is
heated far beyond its glass transition temperature, the chains become mobilized and the modulus
of the polymer decreases rapidly.
69
We used this concept to create temperature-responsive
microstructures that do not require solvents for actuation. These microstructures were fabricated
by solution casting poly(vinyl methyl ether) in water onto the shorter 54 μm PDMS pillars in a
refrigerator set to 4°C (Figure 14). Since poly(vinyl methyl ether) has a low glass transition
temperature (~-31°C),
70
the formation of stable microstructures is difficult at room temperature
38

due to residual solvent weakening the weld. Conducting the self-assembly process at lower
temperatures increases the modulus of the polymer so that the newly formed welds can stabilize
the microstructures while residual solvent evaporates. After the residual solvent is removed, the
polymer weld is strong enough so that the microstructures remain stable when brought to room
temperature. When the microstructures were heated to 30°C, they remained stable; whereas, far
beyond the glass transition temperature at 60°C, they disassembled within 20 minutes. A much
faster disassembly time of 5 minutes was achieved at 90°C. After disassembly, the
microstructures could be reformed via the addition of 10 μL of water in a refrigerator (4°C) to
recreate the solution casting process.

Figure 14. Schematics showing microstructures that respond to variations in a) pH value and b) temperature.
39

The formation of a polymer weld by evaporating a polymer solution at an interface is not
restricted to pillars. Theoretically, any combination of compatible interfaces that can be brought
into contact by capillary forces can be stabilized by our welding process. For example, we used
the solution casting process to self-assemble microbeads within an array of microstructures by
simultaneously collapsing pillars and capturing beads (Figure 15a). This assembly was
performed by applying a suspension of polystyrene microbeads in an alginate-water solution
onto an array of pillars, forming welds between the bead and pillar interfaces upon evaporation.
Additionally, we used polymer welds to create responsive microbead release systems.
Poly(styrene) microbeads were self-assembled and welded onto flat PDMS and within
microfluidic channels by evaporating aqueous solutions of alginate containing a suspension of
the microbeads. The formation of polymer welds between beads and various substrates is similar
to the formation of welds between pillars, where capillary forces draw the evaporating polymer
solution towards the interfaces of adjacent beads where the polymer interdiffuses, solidifies, and
forms a weld. Prior to solution casting, the PDMS surface was treated with air plasma to enhance
wetting of the solution on the substrate. Depending on the concentration of the polymer solution,
the polymer welds can form at different locations around the beads as shown in Figure 15b,c.
Adjacent beads were welded to each other and to the underlying PDMS substrate by pipetting 15
μL of an aqueous mixture containing 0.3 % w/w alginate and 1.7 % w/w 20 μm poly(styrene)
beads. However, when the alginate concentration was diluted to 0.03 % w/w, the microbeads
were welded to adjacent beads but not to the PDMS substrate.
40

Figure 15. SEM images showing variations of alginate weld formation around poly(styrene) microbeads. a)
Microbeads were captured and welded within microstructures. On flat PDMS, welds were formed b) underneath the
beads and between adjacent beads at 0.3 % w/w c) whereas welds were formed only between adjacent beads at 0.03
% w/w.

In order to ensure the release of microbeads in our system, a weld separating the bead from the
substrate is necessary. If microbeads make contact with the substrate, they could become adhered
to it through adhesive forces such as Van der Waals attraction. For example, our poly(styrene)
microbeads that were directly in contact with the PDMS could not be released by dissolving the
alginate with water due to the attraction of the beads to the surface. However, microbeads that
were separated from the PDMS surface by alginate were easily removed from the surface by the
addition of water.
A demonstration of the use of solution casting to adhere and selectively release beads
within microfluidic channels is shown in Figure 16. Microbeads were self-assembled by
injecting and then evaporating an aqueous solution of 0.6 % w/w alginate containing a 6 % w/w
suspension of 20 μm poly(styrene) beads into a microfluidic channel with a cross sectional area
41

of 300 × 1000 µm. A polymer concentration of 0.6 % w/w was used to ensure that the beads
were welded to the underlying PDMS. When an incompatible solvent such as decane or hexane
was flowed through the channel, the particles remained adhered to the PDMS due to the inability
of the solvents to dissolve the polymer welds. However, when a compatible solvent such as
water was flowed through the channel, the polymer welds dissolved and the beads travelled
freely down the channel. Images showing the details of the release process are shown in Figure
16. When water was flowed into the channel, the polymer welds were clearly visible at the
interfaces of the hexagonally packed poly(styrene) beads. As the water incorporated itself into
the polymer welds, the welds began to swell causing greater separation between the beads until
finally the alginate welds dissolved, releasing the beads.

Figure 16. a) Schematic showing controlled release of self-assembled microbeads within a microfluidic channel
using different solvents. b) Time series of microscope images showing the release of poly(styrene) microbeads
welded with alginate by flowing water within a microfluidic channel.
42

3.3 Conclusion

We have demonstrated a simple, robust solution casting technique to stabilize capillary-
force directed self-assembled systems using polymer welds. A wide variety of polymer solutions
were used to stabilize arrays of PDMS microstructures and to control the adhesion of microbeads
within microfluidic channels. The strength of the polymer welds was systematically tuned by
either increasing the molecular weight of the polymer or by increasing the solution
concentration. Utilizing different polymers created reversible, environmentally-responsive
systems. Temperature, pH, and solubility were used to control the assembly, disassembly, and
reassembly of the microstructures. The simplicity of our experimental procedure allows it to be
used universally and tailored for specific applications due to the vast number of commercially
available polymers. For example, the use of aqueous solutions of biocompatible polymers such
as alginate and poly(2-hydroxyethyl methacrylate) allows our technique to be used in biomedical
applications such as drug delivery and tissue engineering.

3.4 Experimental

PDMS pillars and microchannels were fabricated using standard photolithography and
soft lithography techniques. A master mold was fabricated by spin coating SU-8 2050 photoresist
(MicroChem) for the pillar arrays and SU-8 50 photoresist (MicroChem) for the microchannels
onto a silicon wafer and exposing it to UV-light through an emulsion transparency mask
(CAD/Art Services, Inc.). Afterwards, trichloro(1H,1H,2H,2H-perfluorooctyl) silane (Sigma
Aldrich) was deposited onto the mold using a desiccator to ensure the easy release of cured
PDMS. Sylgard
®
184 was mixed at a 10:1 base to cross-linker ratio, poured onto the mold, and
43

thermally cured at 60°C for 4 h. The microchannels were assembled by oxidizing the channel
piece and a flat slab of PDMS with a corona generator (BD20-AC, Electro-Technic Products,
Inc.), bringing both layers together in intimate contact, and curing the device in an oven at 65°C
for 4 h.
The PDMS pillars were self-assembled by pipetting 10 μL of polymer solution onto a 0.7
× 0.7 cm array. The pillars were 60 μm in height and 22 μm in diameter with a spacing of 18 μm
between the edges of the pillars. The polymer solutions were composed of 0.5 % w/w polymer
solutions of alginic acid sodium salts (Sigma Aldrich) in deionized water, poly(methyl
methacrylate) (Varian GPC standards 5,000, 50,000, and 500,000 MW) in acetone, poly(styrene)
(Fluka GPC standard 4,000 MW) in acetone, and poly(2-hydroxylethyl methacrylate) (Sigma
Aldrich 20,000 MW) in methanol. Optical microscopy images were taken using a Meiji ML8000
microscope and SEM images were taken using a JEOL 7001 scanning electron microscope. Prior
to SEM imaging, all samples were coated with a thin layer of gold via sputtering.
In order to test different degrees of compatibility between the polymer weld and the
substrate, arrays of PDMS pillars were coated with a thin layer of poly(1H,1H,2H,2H-
perfluorodecyl acrylate-co-ethylene glycol diacrylate) in a custom designed initiated chemical
vapor deposition (iCVD) chamber (GVD Corp, 250 mm diameter, 48 mm height) as described
previously.
65
While maintaining the pressure at 65 mTorr, 1H,1H,2H,2H-perfluorodecyl acrylate
(SynQuest, 97%), ethylene glycol diacrylate (Monomer-Polymer, 90%), and di-tert-butyl
peroxide (Sigma-Aldrich, 98%) were flowed into the chamber at 0.3, 0.8, and 1.2 sccm,
respectively. The di-tert-butyl peroxide initiator was thermally decomposed into free radicals by
a heated wire array (220°C), the monomer and initiator radicals adsorbed to the cooled substrate
(30°C), and polymerized via a free radical chain mechanism.
44

pH-responsive microstructures were formed using a 0.35 % w/w aqueous solution of
poly(methacrylic acid) (Polysciences 100,000 MW). Prior to forming the solution, 500 mg of
poly(methacrylic acid) was soaked in 10 mL of deionized water for 2 minutes and filtered using
grade 413 filter paper (VWR). The resultant solution was then dried and redissolved in deionized
water to form 0.35 % w/w solution. The pH-responsive microstructures were self-assembled by
pipetting 10 μL of the solution onto PDMS pillars that were 54 μm tall, 22 μm in diameter, and
separated by 18 μm of spacing between the edges of the pillars. In order to test the responses of
the microstructures at different pH values, the structures were treated with a corona generator
and then wet with 20 μL of either deionized water, pH 8 buffer (BDH), or pH 4 buffer (BDH).
Temperature-responsive microstructures were formed by pipetting 20 μL of a 1 % w/w solution
diluted from a 50 % w/w stock aqueous solution of poly(vinyl methyl ether) (Sigma Aldrich)
onto pillars treated with a corona generator. Fast and slow disassembly of the poly(vinyl methyl
ether) stabilized microstructures was performed by placing them into a preheated oven set to
90°C for 5 minutes and 60°C for 20 minutes, respectively. Stability at lower temperatures was
tested in an oven set at 30°C for 20 minutes.
The capture of microbeads within microstructures was performed by pipetting 15 μL of
an aqueous mixture containing 0.03 % w/w alginate and 1.7 % w/w 20 μm polystyrene
microbeads onto a 0.7 × 0.7 cm array of pillars with a height of 60 μm. Microbeads were adhered
onto flat PDMS by pipetting 15 μL of an aqueous mixture containing 0.3 % w/w or 0.03 % w/w
alginate and 1.7 % w/w 20 μm polystyrene microbeads onto a 0.7 × 0.7 cm flat PDMS substrate.
The adhesion of the microbeads was tested under an optical microscope while pipetting 50 μL of
water onto the surface of the sample. SEM images were taken after sputtering a thin layer of gold
onto the samples. The self-assembly of microbeads within a microfluidic channel was performed
45

by slowly injecting an aqueous mixture containing 0.6 % w/w alginate and 6 % w/w
poly(styrene) microbeads into the channel by hand. The solvent was allowed to evaporate at
ambient conditions. The controlled release of the microbeads was performed by injecting the
desired solvents at 2 mL h
-1
(U = 6.67 mm s
-1
) into the channel using syringe pumps. Images
were collected at 10.26 frames per second on a Nikon (Tokyo, Japan) TI-E inverted microscope.

3.5 Acknowledgements

This work was supported by the Annenberg Graduate Fellowship (B.C.) and a fellowship
from the Chevron Corporation (USC−CVX UPP) (C. T. R.).

3.6 References

(49) Morris, C. J.; Parviz, B. A. J. Micromech, Microeng. 2008, 18, 015022.
(50) Gorb, S.; Varenberg, M.; Peressadko, A.; Tuma, J. J. R. Soc. Interface 2007, 4, 271-275.
(51) Wang, Z. L.; Chan, C. T.; Zhang, W. Y.; Ming, N. B.; Sheng, P. Phys. Rev. B 2001, 64, 113108.
(52)Vukusic, P.; Hallam, B.; Noyes, J. Science 2007, 315, 348.
(53) Geim, A. K.; Dubonos, S. V.; Grigorieva, I. V.; Novoselov, K. S.; Zhukov, A. A.; Shapoval, S. Y. Nat. Mater.
2003, 2, 461–463.
(54) Reddy, S.; Arzt, E.; del Campo, A. Adv. Mater. 2007, 19, 3833–3837.
(55) Chen, B.; Seidel, S.; Hori, H.; Gupta, M. ACS Appl. Mater. Interfaces 2011, 3, 4201–4205.
(56) Malmstadt, N.; Hoffman, A. S.; Stayton, P. S. Lab Chip 2004, 4, 412-415.
(57) Lynn, D. M; Amiji, M. M.; Langer, R. Angew. Chem. 2011, 113, 1757-1760.
(58) Haller, P. D.; Flowers, C. A.; Gupta, M. Soft Matter 2011, 7, 2428-2432.
(59) Kwon, I. C.; Bae, Y. H.; Kim, S. W. Nature 1991, 354, 291-293.
(60) Beer, F. B.; Johnston, E. R. Mechanics of Materials, 2nd ed.; McGraw-Hill: New York, 1992.
(61) Murtaza, S.; Naeem, A.; Farina, K.; Dilara, B.; and Tasleem, S. J. Chem. Soc. Pak. 2004, 26, 3. 202-206.
(62) L. E. Scriven. MRS Proceedings 1988, 121, 717-729.
(63) Shah, J. J.; Geist, J.; Locascio, L. E.; Gaitan, M.; Rao, M. V.; Vreeland, W. N. Anal. Chem. 2006, 78, 3348–
3353.
(64) Lin, C. H.; Chao, C. H.; Lan, C. W. Sensor Actuat B-Chem, 2007, 121, 698-705.
(65) Riche, C.T.; Marin, B.C.; Malmstadt, N.; Gupta. M. Lab Chip 2011, 11, 3049-3052.
(66) Yue, C. Y.; Int. J. Adhesion 1988, 8, 47-53.
(67) Kim, W.; Javey, A.; Vermesh, O.; Wang, Q.; Li, Yiming.; Dai, Hongjie. Nano Lett. 2003, 3 (2), 193-198.
(68) Kröger, S.; Turner, A. P. F. Anal. Chim. Acta 1997, 347, 9-18.
(69) Nielsen, L. E. Mechanical Properties of Polymers; Reinhold Publishing Corporation New York, NY, 1962;
pp108-111.
(70) Arndt, K. F.; Schmidt, T.; Reichelt, R. Polymer 2001, 42, 6785-6791.
46

4. Patterned Fluoropolymer Barriers for Containment of Organic Solvents
within Paper-Based Microfluidic Devices

4.1 Introduction

Paper-based microfluidic devices are an attractive option for rapid low-cost diagnostics
due to numerous advantages including portability,
71
ease of use,
72
and limited requirements for
operation.
73
Recent developments have expanded the available operations for paper-based
microfluidic platforms to include mixing,
74
separation of analytes,
75,76
biosensing,
77
and fluid
manipulation using integrated valves and fluidic diodes.
78
Paper-based microfluidic devices
operate using capillary action to drive liquid flow which is typically directed by wax barriers.
79,80

Attempts at using other media to direct the flow of liquids such as hydrophobic photoresists,
81

polymer yarns,
82
and silk
83
have also been explored. However, these methods use hydrophobic
barriers that are vulnerable to effects such as dissolution, solvent diffusion, or swelling by
organic solvents that ultimately restricts the possible operating liquids to aqueous solvents.
Expanding the operating liquids used in paper-based microfluidic devices to include
organic solvents could unlock several advantages over traditional aqueous solvents. For example,
utilizing organic solvents could expand the applications of these devices for the processing and
analysis of water insoluble chemicals, including chemical warfare agents and pesticides, such as
organophosphates,
84,85,86
and pharmaceutical drugs.
87
By utilizing lipophobic barriers that are
compatible with organic solvents, these assays, which may require sequential or timed multi-step
reactions, interfacial reactions, or separation of other hydrophobic contaminants, could
potentially be translated onto paper-based microfluidic devices by directing and manipulating
liquid flow. The use of organic solvents would also allow for additional modification and tuning
47

of both the mobile and stationary phases, which has been challenging due to the difficulty of
creating lipophobic barriers,
88
yielding highly adaptable platforms.
Fluoropolymers have been known to exhibit desirable properties as a barrier material,
including lipophobicity,
89,90
as well as excellent mechanical, thermal, and chemical stability.
91,92

For example, fluoropolymers have been used as the bulk material of microfluidic channels to
increase chemical robustness.
93,94
Additionally, we have recently shown that initiated chemical
vapor deposition (iCVD) can be used to deposit fluoropolymer coatings of poly(1H,1H,2H,2H-
perfluorodecyl acrylate) (PPFDA) onto poly(dimethylsiloxane) microfluidic devices to act as
lipophobic barriers that prevent the diffusion of organic solvents.
95
The patterning of
fluoropolymers has led to advancements in fields such as biosensing,
92
pH-sensing,
96
and optical
device fabrication.
97
However, currently reported techniques for patterning fluoropolymers, such
as directed plasma exposure
98
or ion bombardament,
92
and microcontact printing,
96
are generally
limited to planar substrates. In this paper, we demonstrate a method for patterning a thin
conformal PPFDA coating onto porous substrates in order to replace traditional wax barriers
used in paper-based microfluidic devices, allowing for the containment of both aqueous and
organic solvents without dissolution or swelling of the barrier. The fluoropolymer was patterned
using iCVD combined with transition metal salts, which are able to selectively inhibit the
deposition of polymer.
99
The iCVD technique is a solvent-free coating process that uses a heated
filament to decompose initiator molecules into radical species that react with monomer
molecules to begin free radical polymerization. The solventless nature of the iCVD process
alleviates common issues caused by solvents during polymer processing such as dewetting and
clogging of porous substrates, and has been previously used to conformally coat a variety of non-
planar substrates, including electrospun fiber mats
100
and silicon nanotrenches.
101
We evaluated
48

the resolution of the following three patterning methods for the application of the transition metal
salt inhibitor: painting, spray coating, and selective wetting through the use of a photoresist.
Previous attempts at patterning polymers onto porous substrates via iCVD using physical
masking have proven unsuccessful due to the large mean-free path of the reactant vapors.
99

While other patterning methods have been successfully combined with iCVD, such as electron-
beam lithography
102,103
and capillary force lithography,
104
these methods are also generally
limited to planar substrates. In contrast, the use of transition metal salts allows us to pattern the
fluoropolymer barrier coating throughout the depth of the porous chromatography paper by
selectively inhibiting the deposition of polymer.
We also demonstrate that the fluoropolymer barrier coatings can be used to improve
separation processes within paper-based microfluidic devices. The fluoropolymer barrier
coatings enable the modification of the mobile phase using various organic solvents to tune the
separation of water insoluble analytes. Additionally, the use of organic operating liquids allows
for modification of the stationary phase with organic polymer coatings that are incompatible with
aqueous systems due to their hydrophobicity. The use of polymer coatings to modify stationary
phases in column chromatography has shown an abundance of utility, such as the ability to
facilitate hydrophobic interaction chromatography
105,106
and size-based separation,
107,108
but has
only recently been demonstrated for aqueous-based paper-based microfluidic applications using
electrostatic interactions.
76
In this paper, we use the iCVD process to modify the channels of our
devices with polymer coatings to further tune the Rf values using interactions such as hydrogen
bonding and π-stacking. The iCVD process is ideal for modifying paper-based microfluidic
devices due to its ability to conformally coat porous substrates without compromising their
49

morphology;
76.99,109
as maintaining the porous nature of the paper is essential for creating paper-
based microfluidic devices due to its critical role in facilitating capillary-force driven flow.

4.2 Results and Discussion

Paper-based microfluidic channels were fabricated by patterning fluoropolymer PPFDA
barrier coatings onto chromatography paper using the iCVD process in combination with
copper(II) chloride (CuCl2), as shown in Figure 17. A solution of CuCl2, which acted to inhibit
polymer deposition,
99
was applied using three different techniques: painting, spray coating, and
selective wetting through the use of a hydrophobic photoresist. After the transition metal salt was
applied, the sample was coated using the iCVD process with 50 nm of PPFDA (as measured on a
reference silicon wafer using in-situ interferometry). After the deposition of the PPFDA barrier
coating, the channels were washed with water and methanol to remove the salt. The PPFDA
coating was conformal around the paper fibers and did not occlude the pores of the paper, as
shown in the scanning electron micrographs in Figure 17b.
50

Figure 17. a) Schematic representation of the device fabrication process. CuCl 2 is applied to the chromatography
paper by painting, spray coating, or photolithography, yielding selective deposition of PPFDA by iCVD in the areas
free of CuCl 2. The salt is then removed to yield the final device. b) Comparative SEM micrographs demonstrating
no morphology changes to the porous paper substrate after applying the PPFDA coating via iCVD.

CuCl
2
=
paper =
PPFDA =
wash with water and
methanol
barrier region
channel region
coat with PPFDA
barriers
apply CuCl
2
solution
20 μm
Paper Coated with PPFDA Native Paper
a)
b)
51

Additionally, no change in the porosity of the paper could be detected gravimetrically due to the
low thickness of the polymer coating. The resolution of each salt application technique was
evaluated by comparing the dimensions of a resultant fluoropolymer pattern to an intended
design. A tapered triangular channel was used as the intended design to evaluate the resolution as
a function of channel width. For all three methods of salt application, the resultant fluoropolymer
pattern was larger than the intended design and the deviation was found to be independent of
channel width over the range tested. Painting was performed by brushing a solution of CuCl 2 in
methanol and diethyl ether onto the paper. Methanol was selected to ensure the dissolution of the
CuCl2 salt, while diethyl ether was selected to provide a fast evaporation rate for greater control
over the application of the salt. Even while demonstrating extreme care when painting, this
technique was the least consistent, yielding an average deviation of 0.80 ± 0.26 mm (Table S1)
between the intended channel width and the actual channel width.

52

Table S1. Dimensions of the intended regions, patterned regions, and deviations (Δ) for painting.
Painting

Sample 1 Sample 2 Sample 3
Distance
From
Apex (mm)
Intended
Width
(mm)
Pattern
Width
(mm)
Δ1
(mm)
Pattern
Width
(mm)
Δ2
(mm)
Pattern
Width
(mm)
Δ3
(mm)
1.00 0.29 1.07 0.78 1.52 1.23 1.16 0.87
1.50 0.37 1.17 0.80 1.50 1.13 1.15 0.78
2.00 0.43 1.15 0.72 1.31 0.88 1.49 1.06
2.50 0.50 1.73 1.23 1.57 1.07 1.16 0.66
3.00 0.57 1.50 0.93 1.56 0.99 1.04 0.47
3.50 0.65 1.81 1.16 1.61 0.96 1.07 0.42
4.00 0.73 1.37 0.64 1.66 0.93 1.20 0.47
4.50 0.80 1.41 0.61 1.60 0.80 1.32 0.52
5.00 0.89 1.45 0.56 1.57 0.68 1.17 0.28
Average Δ: 0.80 Standard Deviation Δ: 0.26
When attempting to pattern large areas or multiple devices simultaneously, a technique such as
spray coating is preferable since it is capable of rapidly applying a salt solution over large areas.
We tested this technique by spraying a solution of CuCl2 in methanol and diethyl ether onto the
paper through a physical mask. The resultant deviation between the intended channel width and
the actual channel width was found to be 0.46 ± 0.10 mm (Table S2), which is a significant
improvement when compared to painting.
53

Table S2. Dimensions of the intended regions, patterned regions, and deviations (Δ) for spray
coating.
Spray Coating

Sample 1 Sample 2 Sample 3
Distance
From
Apex (mm)
Intended
Width
(mm)
Pattern
Width
(mm)
Δ1
(mm)
Pattern
Width
(mm)
Δ2
(mm)
Pattern
Width
(mm)
Δ3
(mm)
1.00 0.29 0.73 0.44 0.54 0.25 0.65 0.36
1.50 0.37 0.84 0.47 0.77 0.40 0.76 0.39
2.00 0.43 0.80 0.37 0.96 0.53 0.84 0.41
2.50 0.50 0.95 0.45 1.11 0.61 1.03 0.53
3.00 0.57 0.93 0.36 1.17 0.60 1.04 0.47
3.50 0.65 1.04 0.39 1.19 0.54 0.96 0.31
4.00 0.73 1.25 0.52 1.30 0.57 1.18 0.45
4.50 0.80 1.39 0.59 1.35 0.55 1.15 0.35
5.00 0.89 1.47 0.58 1.40 0.51 1.38 0.49
Average Δ: 0.46 Standard Deviation Δ: 0.10
Although the spray coating technique was able to produce samples quickly and with improved
resolution, it should be noted that the technique required significantly more CuCl 2 due to loss of
the solution onto the mask and to the surrounding area. For applications that require higher
resolutions, a photolithographic approach
110
can be used. In this method, patterns were produced
by first coating the entire paper with a hydrophobic photoresist using the iCVD process. The
conformal nature of the iCVD process ensured that the photoresist did not occlude the pores of
54

the paper, allowing for the subsequent fluoropolymer coating to penetrate the entire depth of the
paper. The photoresist was then selectively exposed to UV light to generate patterned
hydrophilic areas, which were then wet with an aqueous solution of CuCl 2. Using this
photolithographic process, we were able to produce patterns that deviated from the intended
channel width by 0.33 ± 0.10 mm (Table S3).
Table S3. Dimensions of the intended regions, patterned regions, and deviations (Δ) for
photolithography.
Photolithography

Sample 1 Sample 2 Sample 3
Distance
From
Apex (mm)
Intended
Width
(mm)
Pattern
Width
(mm)
Δ1
(mm)
Pattern
Width
(mm)
Δ2
(mm)
Pattern
Width
(mm)
Δ3
(mm)
1.00 0.29 0.42 0.13 0.73 0.44 0.62 0.33
1.50 0.37 0.61 0.24 0.82 0.45 0.66 0.29
2.00 0.43 0.67 0.24 0.90 0.47 0.68 0.25
2.50 0.50 0.79 0.29 0.92 0.42 0.80 0.30
3.00 0.57 0.81 0.24 0.97 0.40 0.83 0.26
3.50 0.65 0.85 0.20 1.09 0.44 1.02 0.37
4.00 0.73 0.91 0.18 1.17 0.44 1.13 0.40
4.50 0.80 1.01 0.21 1.25 0.45 1.09 0.29
5.00 0.89 1.35 0.46 1.34 0.45 1.26 0.37
Average Δ: 0.33 Standard Deviation Δ: 0.10
55

While the deviations in the painting and spray coating methods are likely the result of the CuCl 2
solution bleeding outside the intended area, the use of a hydrophobic photoresist prevents
bleeding. The deviation associated with the use of a photoresist is instead attributed to
undercutting of the photoresist during exposure. It is important to note that other patterning
techniques for applying the salt solutions may yield better resolution or scalability, such as inkjet
printing.
111
However, this method requires surfactants and optimization to prevent clogging of
printer heads. Since the painting technique required less salt than spray coating and is less
cumbersome than photolithography, we used the painting method for the remainder of our
studies.
During the deposition of the PPFDA barrier coatings, two criteria must be met to
effectively fabricate the channels: 1) the PPFDA coating in the barrier region must be conformal
around the paper fibers, otherwise the uncoated areas may cause solvent to bleed through the
barrier; and 2) the deposition of PPFDA in the channel region must be inhibited in order to
prevent unpredictable wetting and decreased device performance. While the first criterion is
more easily satisfied at higher coating thicknesses, the second criterion is more easily satisfied at
lower coating thicknesses since the ability of transition metal salts to inhibit the deposition of
iCVD coatings has been shown to decrease as greater amounts of polymer are deposited. This
decrease in inhibition was hypothesized to be caused by the formation of a layer of deactivated
precursor molecules that shields the transition metal salt from deactivating additional precursor
molecules.
99
In order to systematically determine the PPFDA coating thicknesses that fulfilled
both criteria, X-ray photoelectron spectroscopy (XPS) was used to analyze the amount of
PPFDA in both the barrier region (Figure 18a) and in the channel region (Figure 18b) as a
function of the coating thickness as measured on a reference silicon wafer. The XPS spectra of
56

the paper surface show an increasing fluorine intensity at 686 eV with increasing coating
thickness in both the barrier and channel regions which is indicative of an increasing
amount of PPFDA on the surface.

Figure 18. XPS spectra showing the chemical composition of a) the barrier regions and b) the channel regions of
devices with different thicknesses of deposited PPFDA.

250 350 450 550 650 750
Binding Energy (eV)
a)
C
250 350 450 550 650 750
Binding Energy (eV)
F O
b)
440 nm PPFDA
50 nm PPFDA
10 nm PPFDA
Unmodified chromatography paper
640 nm PPFDA
440 nm PPFDA
50 nm PPFDA
10 nm PPFDA
Unmodified chromatography paper
640 nm PPFDA
57

XPS probes approximately the top 5 nm of the surface and therefore the weight fraction of the
PPFDA on the surface relative to cellulose can be estimated by comparing the fluorine to carbon
ratio on the surface of the devices to a surface of homopolymer PPFDA (Table 1).
Table 4. Weight fraction of PPFDA in both the barrier and channel regions of the paper devices.
Polymer Thickness on
Reference Silicon
Weight Fraction of PPFDA in
Barrier Regions
Weight Fraction of PPFDA in
Channel Regions
10 0.68 0.03
50 0.99 0.08
440 0.98 0.17
640 0.99 0.85

If the coating was too thin, areas of cellulose remained exposed as indicated by a low PPFDA
weight fraction. For example, when 10 nm of coating was deposited, the PPFDA weight fraction
within the barriers was 0.68, resulting in barriers that were unable to contain hexane within the
channel (Figure 19a). When the coating thickness was increased to 50 nm or more, the PPFDA
weight fraction remained steady at approximately 0.98-0.99 indicating at least 5 nm of conformal
coating, yielding barriers that could contain hexane (Figure 19b). However, as we deposited
thicker coatings, the transition metal salt was unable to conformally inhibit the deposition of
polymer, resulting in a significant increase in the PPFDA weight fraction within the channel. As
a result, 640 nm thick coatings showed non-uniform wetting of hexane within the channel region
(Figure 19c).
58

Figure 19. Images of paper-based microfluidic devices after applying organic solvents containing dye for
visualization. Fluoropolymer barriers a) are unable to contain hexane with a 10 nm thick PPFDA coating, b) are able
to successfully contain hexane with a 50 nm thick PPFDA coating, and c) have non-uniform wetting with a 640 nm
thick PPFDA coating. d) Fluoropolymer barriers (left) made with 50 nm thick PPFDA coating succeed at containing
a wide variety of organic solvents whereas traditional wax barriers (right) fail.

The water contact angles, as shown in Table 2, in both the barrier and channel regions are
consistent with the PPFDA weight fraction as determined by XPS. The top and bottom sides of
the paper show similar contact angles even at low deposition thicknesses, confirming that the
PPFDA coating is conformal through the depth of the paper.
hexane ethyl acetate acetone methanol
d)
b) c)
a)
1 cm
10 nm 50 nm 640 nm
barrier region channel region
PPFDA Coating Thickness
wax barrier fluoropolymer barrier
Applied Solvent
59

Table 2. Water contact angles on both the top and bottom in both the barrier and channel regions of paper devices.
Polymer Thickness on
Reference Silicon
Contact Angle in Barrier
Regions
Contact Angle in Channel
Regions
10
Top 124.5 ± 0.9 0.0 ± 0.0
Bottom 125.9 ± 4.0 0.0 ± 0.0
50
Top 148.1 ± 1.2 0.0 ± 0.0
Bottom 145.4 ± 3.0 0.0 ± 0.0
440
Top 147.3 ± 1.0 106.0 ± 7.4
Bottom 149.9 ± 4.9 110.9 ± 4.1
640
Top 154.0 ± 1.2 154.2 ± 0.6
Bottom 155.6 ± 2.6 154.9 ± 2.9
Based on our observations above, we deposited approximately 50 nm of PPFDA on all
further paper-based microfluidic devices to effectively contain organic solvents. Although the
channel regions contained a small amount of PPFDA, the resultant devices exhibited the same
solvent wicking behavior as uncoated paper, indicating that this low amount of PPFDA did not
significantly affect device performance. These results are in agreement with the contact angle
data in Table 2, where the channels regions of devices coated with 50 nm or less PPFDA
exhibited a contact angle of 0°. In contrast, when 440 nm of polymer was applied, we see that at
even with a surface PPFDA weight of 0.17, water imbibition is no longer reliable within the
device. A comparison between the ability of the PPFDA barriers and traditional wax barriers to
contain organic solvents with contrasting polarities (hexane, ethyl acetate, acetone, and
methanol) is shown in Figure 19d. In all cases, the tested solvents were able to penetrate the
60

traditional wax barriers, leading to undesirable flows. In contrast, the PPFDA barriers were able
to successfully contain this wide variety of solvents. A full list of tested solvents can be found in
Table S4.

61

Table S4. Ability of wax barriers and fluoropolymer barriers to contain solvents.
Solvent
Contained by
Wax Barriers
Contained by
Fluoropolymer Barriers
Acetone No Yes
Acetonitrile Yes Yes
Aqueous sulfuric acid (5 M) Yes Yes
Aqueous sodium hydroxide (5 M) Yes Yes
Butanol No Yes
Chloroform No Yes
Cyclohexane No Yes
Diethyl ether No Yes
Dimethyl formamide Yes Yes
Dimethyl sulfoxide Yes Yes
Ethanol No Yes
Ethyl acetate No Yes
Hexane No Yes
Isopropanol No Yes
Methanol No Yes
Tetrahydrofuran No Yes
Toluene No Yes
Water No Yes
62

The ability to use organic solvents in paper-based microfluidic devices also allows for a
variety of operations that are difficult or impossible with aqueous solvents, such as the
chromatographic separation of lipophilic analytes. A common metric used to describe
chromatographic separation is the retardation factor (Rf), which is defined as the distance
travelled by the analyte relative to the distance travelled by the mobile phase. Higher R f values
exist when there is greater affinity between the analyte and the mobile phase, while lower Rf
values exist when there is greater affinity between the analyte and the stationary phase. Thus by
modifying either the mobile or stationary phase of the system, we can tune the R f value of
analytes, allowing for separation of a multi-component system. A simple demonstration of how
the mobile phase can be tailored to yield specific degrees of separation is shown in Figure 20a
where various compositions of methanol and water were used to control the R f values of the
lipophilic dyes Sudan Black B and Nile Red. Due to the greater affinity of Sudan Black B to
methanol compared to Nile Red, as the relative amount of water increased, the R f value of Nile
Red decreased more rapidly than Sudan Black B, resulting in a larger separation between the
analytes. Although two lipophilic dyes were examined as model analytes, the ability to use
organic solvents to affect the separation of a mixture of analytes is applicable to a wide variety of
systems. Additionally, channels can be modified with functional polymer coatings that may be
incompatible with aqueous systems due to their hydrophobicity in order to further tune the R f
value of analytes. For example, we modified the channels by pre-coating chromatography paper
with copolymers composed of 4-vinyl pyridine (4VP) and ethylene glycol dimethacrylate
(EGDMA) prior to depositing the fluoropolymer barriers to tune the Rf value of Sudan Black B
in hexane.
63

Figure 20. Graphs plotting retardation factors of a) Sudan Black B and Nile Red in solvent blends of water and
methanol on cellulose patterned with fluoropolymer barrier coatings, and b) Sudan Black B in hexane on channels
coated with copolymers composed of 4VP and EGDMA on devices patterned with fluoropolymer barrier coatings.

The mole fraction of EGDMA in the copolymer coating was systematically varied between 0.04
and 1.00, as determined by XPS (Table 3).
0.00
0.20
0.40
0.60
0.80
1.00
0.00 0.20 0.40 0.60 0.80 1.00
R
f
H
2
O wt. Fraction in Methanol
Sudan Black B
Nile Red
a)
b)
0.00
0.20
0.40
0.60
0.80
1.00
0.00 0.20 0.40 0.60 0.80 1.00
R
f
EGDMA Mole Fraction
64

Table 3. Mole fraction of EGDMA in copolymer coating and corresponding Rf value of Sudan Black B with hexane as the
mobile phase.
Mole Fraction of EGDMA Rf
0.04 0.28 ± 0.01
0.53 0.50 ± 0.02
0.64 0.71 ± 0.02
1.00 0.87 ± 0.01

In addition to providing functionality, the EGDMA cross-linker also prevented dissolution of the
copolymer coatings. The lowest Rf value of Sudan Black B was measured on an unmodified
cellulose channel (0.22 ± 0.02), which we hypothesize to be due to the ability of cellulose to act
as both a proton donor (OH--N) and acceptor (O--HN), resulting in greater affinity to Sudan
Black B. The effect of the EGDMA mole fraction in the copolymer coating on the R f values of
Sudan Black B is shown in Figure 20b. When the stationary phase was modified with copolymer
coatings composed mostly of 4VP (EGDMA mole fraction of 0.04), the R f value increased
slightly, which is likely due to weaker hydrogen bonding interactions between the 4VP moieties
and Sudan Black B (N--HN) compared to cellulose and Sudan Black B, as well the inability of
4VP moieties to act as a proton donor. However, the weaker interactions may be offset by the
presence of π-stacking interactions, leading to only a small net change in the R f value. As the
mole fraction of EGDMA increased, the R f value monotonically increased, which we attribute to
decreased attraction between Sudan Black B and the copolymer coating due to a reduction of π-
stacking interactions. The effect of the copolymer-modified stationary phase on the Rf value of
65

Nile Red showed a similar trend and thus resulted in minimal separation of the two dyes.
However, the behavior of the Rf value of a given analyte depends on the specific interactions
between the analyte and the stationary phase and so the improvement in separation for a given
system would depend on the type of polymer coating used to modify the stationary phase.
Nevertheless, these results show that modifying the stationary phase with polymer coatings
allows the Rf value of analytes to be tuned over a wide range, which can improve the separation
of multi-component systems.
4.3 Conclusion

We have demonstrated the ability to use organic solvents within paper-based microfluidic
devices by patterning fluoropolymer barriers using iCVD in conjunction with transition metal
salts that inhibit polymer deposition. The patterning resolution of three different methods of
applying the salt were compared, and it was revealed that selective wetting through use of a
hydrophobic photoresist yielded the highest resolution. XPS data was used to determine the
amount of fluoropolymer coating necessary for optimal performance of the device. The efficacy
of the fluoropolymer barrier coatings relied on the conformality of the coating around the paper
fibers. However, depositing excess polymer jeopardized the ability of the transition metal salt to
inhibit polymer deposition within the intended channel regions. We also demonstrated the
additional utility of using organic solvents in paper-based microfluidic applications by separating
lipophilic dyes by controlling the composition of the operating solvent and by application of
organic polymer coatings within the channels. Although we provide example applications using
simple straight channels, the generality of our technique can be applied to more complex devices
and broaden the range of available applications for paper-based microfluidic devices by
expanding the possible operating liquids. Additionally, our technique has the potential to
66

advance current fields that employ the use of hydrophobic or lipophobic regions such as
biosensing, chemical detection, and optics.
4.4 Experimental

1H,1H,2H,2H-perfluorodecyl acrylate (PFDA) (SynQuest, 97 %), di-tert-butyl peroxide
(Sigma Aldrich, 98 %), 4-vinyl pyridine (Sigma Aldrich, 95 %), ethylene glycol dimethacrylate
(Sigma Aldrich, 98 %), ortho-nitrobenzyl methacrylate (oNBMA) (Polysciences, 95 %),
copper(II) chloride (Sigma Aldrich, reagent grade), chromatography paper (Whatman, No 1), pH
8 buffer (BDH, pH 8 ± 0.02), acetone (Macron, 99.5 %), acetonitrile (Mallinckrodt, 99.5 %),
butanol (Mallinckrodt, 99.4 %), chloroform (Mallinckrodt, 99.9 %), cyclohexane (EMD, 99.99
%) diethyl ether (BDH, 99 %), dimethyl formamide (EMD, 99.8 %), dimethyl sulfoxide
(Mallinckrodt, 99.8 %), ethanol (Koptec, 200 proof), ethyl acetate (Mallinckrodt, 99.5 %),
hexane (EMD, 98.5 %), isopropanol (Macron, ACS grade), methanol (Macron, absolute),
tetrahydrofuran (Mallinckrodt, 99.0 %), toluene (J. T. Baker, 99.7 %), blue food coloring
(McCormick), Sudan Black B (Sigma Aldrich, Biological Stain Commission certified), Nile Red
(Sigma Aldrich, microscopy grade), sulfuric acid (EMD, ACS grade), and sodium hydroxide
(Mallinckrodt, 98.8 %) were used as received without further purification.
To fabricate PPFDA barrier coatings, CuCl2 was applied to chromatography paper and
subsequently coated in a custom designed iCVD reactor chamber (GVD Corp, 250 mm diameter,
48 mm height). The deposition of PPFDA was performed at a constant pressure of 40 mTorr
while the samples were maintained at 30 °C using a backside recirculating chiller. The monomer,
1H,1H,2H,2H-perfluorodecyl acrylate, and initiator, di-tert-butyl peroxide, flow rates were 0.4
and 3.9 sccm, respectively. During the deposition, a nichrome filament array (80 % Ni, 20 % Cr,
67

Omega Engineering) inside the reactor was resistively heated to 250 °C to decompose the
initiator molecules into free radicals. Polymerization occurred on the surface of the substrate via
free radical polymerization. Polymer film thickness measurements on silicon were obtained
using a 633 nm helium-neon laser interferometer (Industrial Fiber Optics).
The CuCl2 was applied prior to deposition of PPFDA using painting, spray coating, or
selective wetting through use of a hydrophobic photoresist. Following the deposition of PPFDA,
the CuCl2 salt was removed by washing the samples with water, followed by methanol. The
subsequent methanol wash was used to remove CuCl2 salt, but also served to alleviate wrinkling
effects that occurred during the water evaporation. Following this step with an additional wash
with acetone resulted in paper that was virtually wrinkle-free after solvent evaporation. The
samples were then allowed to dry in ambient conditions prior to analysis. The resolution of the
patterning technique for each method of salt application was determined by comparing the width
of an isosceles triangular mask with a base of 5 mm and a height of 50 mm to that of a final
device dyed with a 50:1 by volume mixture of methanol and blue food coloring for visualization
at nine equally distributed intervals from 1 mm to 5 mm from the apex along a line perpendicular
to the base (Figure 21). The reported deviation is an average of these nine measurements across
three samples per method with ± values representing one standard deviation.
68

Figure 21. Image defining the variables used to perform resolution analysis on intended
isosceles triangle patterns.
Painted samples were patterned by applying approximately 80 μL/cm
2
of a 1:1 by volume
mixture composed of diethyl ether and 4 M CuCl2 in methanol onto chromatography paper using
a standard 3/0 round paintbrush (Princeton Art & Brush Co.). A dark outline of the intended
channel area was situated below the paper to act as a guide to be traced. Spray coated samples
were patterned by spraying a 4:1 by volume mixture composed of diethyl ether and 2 M CuCl 2 in
methanol 50 times over physically masked chromatography paper using a household hand-
operated spray bottle (FamilyMaid). After every 10 sprays, the samples were dried with a heat
gun to evaporate excess methanol and ether. The selective wetting through the use of a
hydrophobic photoresist was performed by coating chromatography paper with approximately 25
nm of the photoresist poly(ortho-nitrobenzyl methacrylate) (PoNBMA) using the iCVD process
with a constant pressure of 50 mTorr while the samples were maintained at 20 °C. The
monomer, ortho-nitrobenzyl methacrylate, and initiator, di-tert-butyl peroxide, flow rates were
0.05 sccm and 0.7 sccm, respectively. After the polymer was deposited, the paper was selectively
exposed to 365 nm UV light (UVP, UVL-21) through a mask for 90 minutes. Afterwards, the
sample was submerged in pH 8 buffer for 60 min followed by a water rinse to remove the
apex
1 mm from apex 5 mm from apex
intended
width
patterned
width
69

photoresist from the exposed areas. The paper was allowed to dry under ambient conditions, after
which 40 μL/cm
2
of a 4 M aqueous solution of CuCl2 was pipetted onto the exposed area.
In order to compare the fluoropolymer barriers to traditional wax barriers, wax toner was
printed onto chromatography paper using a Xerox Phaser 8560N printer and subsequently melted
through the depth of the paper using an oven set at approximately 180 °C for 3 minutes as
described in previous studies.
79,80
Channels measuring 1 cm by 5 cm with PPFDA barriers were
made by applying the CuCl2 using the painting method. The evaluation of the ability of PPFDA
barriers and traditional wax barriers to contain various solvents was performed by pipetting
approximately 150 µL of a 0.2 mg/mL solution of Sudan Black B in each solvent into the
channels followed by visual detection of whether the solvent bled through the barrier or
exhibited irregular wetting behavior within the channel. Sudan Black B was dissolved in the
solvents to provide greater contrast between the wetted and non-wetted areas.
The channel regions of devices measuring 1 cm by 5 cm were modified with polymer
coatings by pre-coating chromatography paper with copolymers composed of 4-vinyl pyridine
and ethylene glycol dimethacrylate using the iCVD process. For all depositions, the di-tert-butyl
peroxide flow rate was 0.7 sccm, the filament temperature was 250 °C, and the samples were
kept at 20 °C. The reaction conditions for each polymer coating are summarized in Table 4.
PPFDA barriers were then subsequently patterned onto the coated paper as described above by
applying CuCl2 using the painting method.
The Rf values were determined by spotting a 0.4 % by weight methanolic solution of
either Sudan Black B or Nile Red 1 cm from the bottom edge of the device. After the analyte
dried, the device was inserted vertically into a glass chamber filled with solvent to a height of
approximately 0.5 cm. The chamber was then immediately covered to reduce evaporation. When
70

the mobile phase travelled at least 2 cm, the sample was removed from the chamber and the
distances travelled by both the mobile phase and the analyte were measured to calculate the R f
value. The reported Rf values were an average of triplicate measurements using a new device for
each measurement, with ± values representing a single standard deviation.

Table 4. iCVD experimental conditions
EGDMA Flow Rate
(sccm)
4VP Flow Rate
(sccm)
Pressure
(mTorr)
Mole Fraction of
EGDMA in Coating
0.2 7.8 500 0.04
0.2 5.8 325 0.53
0.2 3.5 200 0.64
0.2 0.0 50 1.00

X-ray photoelectron spectroscopy (XPS) was performed using a Surface Science Instruments M-
Probe spectrometer with a monochromatic Al Kα X-ray source. Survey spectra were averaged
over 5 scans and were acquired at binding energies between 1 and 1000 eV with a resolution of 1
eV. Data analysis was performed using the ESCA25 Analysis Application software (V5.01.04).
The relative amount of PPFDA on the surface of the paper devices was determined by comparing
the relative carbon to fluorine atomic ratios according to Equation S1.
The relative amount of PPFDA on the surface of the paper devices was determined by
comparing the relative carbon to fluorine atomic ratios, according to Equation S1. This
71

calculation determines the weight fraction of PPFDA assuming cellulose (monomer unit:
(C6H10O5)) makes up the remaining portion of the surface, where #C and #F represent the atomic
percentages of carbon and fluorine in the sample, respectively, and #CPPFDA and #FPPFDA
represent the atomic percentages of carbon and fluorine in the reference homopolymer PPFDA,
as measured on paper coated with more than 2 µm of polymer to ensure at least 5 nm of
coverage. The measured experimental ratios of carbon, oxygen, and fluorine in the reference
PPFDA sample were 38 %, 6 %, and 56 %, which are in close agreement with the theoretical
values of 40.625 %, 6.25 %, and 53.125 %, respectively. The weight fraction of PPFDA is
determined by subtracting the weight fraction of cellulose, represented by the fractional term in
Equation S1, from unity. The numerator represents the relative weight of cellulose. The relative
amount of carbon associated with PPFDA is subtracted from the total atomic percentage of
carbon and divided by six to account for the number of carbons per cellulose moiety, and is then
multiplied by the molecular weight of a cellulose moiety (162.14). The denominator represents
the total relative weight of both cellulose and PPFDA, where the factor of 13 accounts for the
number of carbons per PFDA moiety, and the value 518.17 is the molecular weight of a PFDA
moiety.

The mole fraction of EGDMA in the copolymer coating was determined by comparing the
relative carbon to nitrogen atomic ratios, according to Equation S2, as measured on a reference
silicon wafer. #C and #N represent the atomic percentages of carbon and nitrogen in the sample,
respectively, and #CP4VP and #NP4VP represent the atomic percentages of carbon and nitrogen in
%𝑃𝑃𝐹𝐷𝐴 = 1 −
162.14 ×
#𝐶 −
#𝐶 𝑃𝑃𝐹𝐷𝐴 #𝐹 𝑃𝑃𝐹𝐷𝐴 × #𝐹 6
162.14 ×
#𝐶 −
#𝐶 𝑃𝑃𝐹𝐷𝐴 #𝐹 𝑃𝑃𝐹𝐷𝐴 × #𝐹 6
+ 518.17 ×
#𝐶 𝑃𝑃𝐹𝐷𝐴 #𝐹 𝑃𝑃𝐹𝐷𝐴 × #𝐹 13
( S1)
72

homopolymer P4VP, as measured experimentally. The measured experimental ratios of carbon
and nitrogen in the homopolymer sample of P4VP were 90 %, and 10 %, which were in close
agreement with the theoretical values of 87.5 % and 12.5 %, respectively. The numerator
represents the relative number of EGDMA moieties; the relative amount of carbon associated
with 4VP is subtracted from the total atomic percentage of carbon and divided by 10 to account
for the number of carbons per EGDMA moiety. The denominator represents the total number of
both EGDMA and 4VP moieties, where the factor of 7 accounts for the number of carbons per
4VP moiety.

4.5 Acknowledgements

This work was supported by the National Science Foundation Division of Civil,
Mechanical, and Manufacturing Innovation Award Number 1069328, the Natural Sciences and
Engineering Research Council of Canada Scholarship (P.K.), and the Alfred Mann Institute at
the University of Southern California (P.K.). We thank the Molecular Materials Research Center
of the Beckman Institute of the California Institute of Technology for use of their XPS.

4.6 References

(71) Nie, Z.; Nijhuis, C. A.; Gong, J.; Chen, X.; Kumachev, A.; Martinez, A. W.; Narovlyansky, M.; Whitesides, G.
M. Lab Chip 2010, 10, 477-483.
(72) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 19606-19611.
(73) Martinez, A. W.; Phillips, S. T.; Whitesides, G. M.; Carrilho, E. Anal. Chem. 2010, 82, 3-10.
(74) Osborn, J. L.; Lutz, B.; Fu, E.; Kauffman, P.; Stevens, D. Y.; Yager, P. Lab Chip 2010, 10, 2659-2665.
%𝐸𝐺𝐷𝑀𝐴 =
#𝐶 −
#𝐶 𝑃 4𝑉𝑃
#𝑁 𝑃 4𝑉𝑃
× #𝑁 10

#𝐶 −
#𝐶 𝑃 4𝑉𝑃
#𝑁 𝑃 4𝑉𝑃
× #𝑁 10
+
#𝐶 𝑃 4𝑉𝑃
#𝑁 𝑃 4𝑉𝑃
× #𝑁 7
( S2)
73

(75) Yang, X.; Forouzan, O.; Brown, T. P.; Shevkoplyas, S. S. Lab Chip 2012, 12, 274-280.
(76) Kwong, P.; Gupta, M. Anal. Chem. 2012, 84, 10129-10135.
(77) Su, S.; Ali, M. M.; Filipe, C. D. M.; Li, Y.; Pelton, R. Biomacromolecules 2008, 9, 935-941.
(78) Chen, H.; Cogswell, J.; Constantine, A.; Faghri, M. Lab Chip 2012, 12, 2909-2913.
(79) Lu, Y.; Shi, W.; Jiang, L.; Qin, J.; Lin, B. Electrophoresis 2009, 30, 1497-1500.
(80) Carrilho, E.; Martinez, A. W.; Whitesides, G. M. Anal. Chem. 2009, 81, 7091-7095.
(81) Martinez, A. W.; Phillips, S. T.; Carrilho, E.; Thomas, S. W.; Sindi, H.; Whitesides, G. M. Anal. Chem. 2008,
80, 3699-3707.
(82) Owens, T. L.; Leisen, J.; Beckham, H. W.; Breedveld, V. ACS Appl. Mater. Interfaces 2011, 3, 3796-3803.
(83) Bhandari, P.; Narahari, T.; Dendukuri, D. Lab Chip 2011, 11, 2493-2499.
(84) Han, S.; Xue, Z.; Wang, Z.; Wen, T. B. Chem. Commun. 2010, 46, 8413-8415.
(85) O’Brien, R. D. J. Biol. Chem. 1956, 219, 927-931.
(86) Kim, J.-W.; Rainina, E. I.; Mulbry, W. W.; Engler, C. R.; Wild, J. R. Biotechnol. Prog. 2002, 18, 429-436.
(87) World Health Organization Basic Tests for Pharmaceutical Dosage Forms; 1991.
(88) You, I.; Yun, N.; Lee, H. ChemPhysChem 2013, 14, 471-481.
(89) Yabu, H.; Takebayashi, M.; Tanaka, M.; Shimomura, M. Langmuir 2005, 21, 3235-3237.
(90) Mao, J.; Ni, P.; Mai, Y.; Yan, D. Langmuir 2007, 23, 5127-5134.
(91) Patri, M.; Hande, V. R.; Phadnis, S.; Deb, P. C. Polym. Adv. Technol. 2004, 15, 622-627.
(92) Jung, J.-C.; Hwang, I.-T.; Kuk, I.-S.; Choi, J.-H.; Oh, B.-K.; Lee, Y.-M. ACS Appl. Mater. Interfaces 2013, 5,
2155-2160.
(93) Rolland, J. P.; Van Dam, R. M.; Schorzman, D. A.; Quake, S. R.; DeSimone, J. M. J. Am. Chem. Soc. 2004,
126, 2322-2323.
(94) A. Taberham, M. Kraft, M. Mowlem and H. Morgan, J. Micromech. Microeng., 2008, 18, 064011.
(95) Riche, C. T.; Marin, B. C.; Malmstadt, N.; Gupta, M. Lab Chip 2011, 11, 3049-3052.
(96) Prissanaroon, W.; Brack, N.; Pigram, P. J.; Hale, P.; Kappen, P.; Liesegang, J. Synth. Met. 2005, 154, 105-108.
(97) Tamura, H.; Kojima, R.; Usui, H. Appl. Opt. 2003, 42, 4008-4015.
(98) Vargo, T. G.; Calvert, J. M.; Wynne, K. J.; Avlyanov, J. K.; MacDiarmid, A. G.; Rubner, M. F. Supramol. Sci.
1995, 2, 169-174.
(99) Kwong, P.; Flowers, C. A.; Gupta, M. Langmuir 2011, 27, 10634-10641.
(100) Ma, M.; Gupta, M.; Li, Z.; Zhai, L.; Gleason, K. K.; Cohen, R. E.; Rubner, M. F.; Rutledge, G. C. Adv. Mater.
2007, 19, 255-259.
(101) Tenhaeff, W. E.; McIntosh, L. D.; Gleason, K. K. Adv. Funct. Mater. 2010, 20, 1144-1151.
(102) Im, S. G.; Kim, B.-S.; Lee, L. H.; Tenhaeff, W. E.; Hammond, P. T.; Gleason, K. K. Macromol. Rapid
Commun. 2008, 29, 1648-1654.
(103) Mao, Y.; Felix, N. M.; Nguyen, P. T.; Ober, C. K.; Gleason, K. K. J. Vac. Sci. Technol. B 2004, 22, 2473-
2478.
(104) Im, S. G.; Bong, K. W.; Kim, B.-S.; Baxamusa, S. H.; Hammond, P. T.; Doyle, P. S.; Gleason, K. K. J. Am.
Chem. Soc. 2008, 130, 14424-14425.
(105) Chang, J.-P.; El Rassi, Z.; Horváth, C. J. Chromatogr. 1985, 319, 396-399.
(106) Hjertén, S.; Mohammad, J.; Eriksson, K.-O.; Liao, J.-L. Chromatographia 1991, 31, 85-94.
(107) Desilets, C. P.; Rounds, M. A.; Regnier, F. E. J. Chromatogr. 1991, 544, 25-39.
(108) Sing, Y. L. K.; Kroviarski, Y.; Cochet, S.; Dhermy, D.; Bertrand, O. J. Chromatogr. 1992, 598, 181-187.
(109) Haller, P. D.; Flowers, C. A.; Gupta, M. Soft Matter 2011, 7, 2428-2432.
(110) Brown, A. A.; Azzaroni, O.; Huck, W. T. S. Langmuir, 2009, 25, 1744-1749.
(111) Thomas, S. W.; Chiechi, R. C.; LaFratta, C. N.; Webb, M. R.; Lee, A.; Wiley, B. J.; Zakin, M. R.; Walt, D. R.;
Whitesides, G. M. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 9147-9150.
74

5. Fabricating Polymer Canopies onto Structured Surfaces Using Liquid
Scaffolds
5.1 Introduction

Modifying surfaces with polymers can provide desirable functionalities such as
antimicrobial properties,
112
biocompatibility,
113
and environmental responsiveness.
114
Applying
polymers to the surfaces of complex geometries such as trenches and pores can allow the effects
of the functionality to be enhanced by the greater surface to volume ratio. For example,
electrospun fiber mats can be coated with polymer to make superhydrohyphobic surfaces,
115

microtrenches can be modified to immobilize nanoparticles,
116
and porous media can be coated
to facilitate the separation of analytes
117
and to control drug delivery.
118

Chemical vapor deposition is an attractive option for forming coatings on complex
geometries due to numerous advantages derived from its solventless nature. In contrast, typical
solution-based coating processes such as spin coating and spray coating are subject to surface
tension effects,
119
ultimately limiting their application on complex geometries. Many chemical
vapor deposition techniques have been used to produce functional polymer coatings such as
plasma enhanced chemical vapor deposition
120
and oxidative chemical vapor deposition.
121
One
specific technique that has seen recent advancement is initiated chemical vapor deposition
(iCVD). The iCVD technique is a solventless vacuum process that forms polymer coatings
directly onto substrates. In this process, monomer and initiator molecules are flown into a reactor
chamber where the initiator comes into contact with a heated filament array, which decomposes
75

the initiator into radicals. These radicals can then adsorb onto the surface of the substrate and
react with adsorbed monomer to initiate free radical polymerization. The iCVD process has been
used to produce functional coatings that lead to the biocompatibility of medical implants
122
and
facilitate the separation of molecules.
123,124
Additionally, the iCVD process can be used to apply
conformal coatings onto complex geometries such as porous media,
12,13,125
micropillar arrays,
126

and forests of carbon nanotubes with diameters on the order of hundreds of nanometers.
127
The
iCVD process has also recently been used to deposit polymers onto low vapor pressure liquids
such as ionic liquids and silicone oils to expand the structural control of polymer deposition to
include free standing films,
128
gels,
129
particles,
130
gradient films,
131
and microstructured films.
132

In this study, we explore a liquid scaffolding technique that is used in conjunction with
the iCVD process to fabricate polymer canopies onto complex geometries. We demonstrate the
capabilities of this technique by directly depositing a variety of polymer canopies over
micropillar and microstructure arrays. Unlike previous demonstrations of using liquids with the
iCVD process, the liquid scaffolding technique focuses on using the liquid as a template which
simultaneously masks the surface while providing structural support for the deposited polymer.
After the polymer film is deposited, the liquid is removed to reveal regions that were masked in
the process as well as a polymer film that retains the shape of the liquid surface. We explore two
methods, the Droplet Method and the Inverted Method, for manipulating the scaffolding liquids
such that a variety of polymer canopies can be made using a range of liquids. Unlike physically
placing a polymer film on top of the substrate, our fabrication methods directly form canopies on
the surfaces allowing for better mechanical stability. Additionally, the Inverted Method has the
added capability of controlling the height of the canopy relative to the substrate features. The
formation of polymer layers over complex geometries is of interest in fields such as drug release,
76

where the polymer can act as a tunable gatekeeper.
133,134,135
Additionally, liquid scaffolding can
potentially be applied to other geometries such as microfluidic channels, allowing for the
fabrication of polymer membranes within the channels to extend their utility. For example,
polycarbonate membranes used within microfluidic channels can prevent the cross-
contamination between two regions while allowing for the delivery of reactants across the
membrane.
136
Although we only demonstrate a few examples of liquid manipulation for
generating scaffolding templates, there are many additional strategies that can be employed to
control liquids on complex geometries, including the controlled formation of microdrops on top
of microposts,
137
the controlled infiltration of liquid within textured surfaces,
138,139
and the
capillary rise of liquids between various pillar geometries.
140,141
Additionally, the methods
presented are not limited to iCVD systems since the generality of the strategy can be easily
translated to other vapor deposition techniques.

5.2 Results and Discussion

Our group has previously demonstrated the ability to deposit polymers onto low vapor
pressure liquids using the iCVD process.
142,17,18,19,20,21
The surface interactions between the
liquid and the polymer determines the initial morphology of the deposited polymer. The surface
interaction can be quantified using the spreading coefficient (S):
S = γ
LV
∗ ( 1 + cos θ)− 2 γ
PV

where γ LV is the liquid−vapor surface tension, θ is the advancing contact angle of the liquid on
the polymer, and γPV is polymer−vapor surface tension.
143
When the spreading coefficient is
77

positive, it is energetically favorable for the polymer to spread over the surface resulting in
thermodynamically stable films, whereas a negative spreading coefficient represents systems
where it is favorable for the polymer to reduce contact with the liquid surface resulting in
particles. Additional work has demonstrated that it is possible to deposit polymer films on liquids
when the system spreading coefficient is negative by using a cross-linker.
21

We were able to control the fabrication of polymer canopies on top of micropillar arrays
by exploiting systems with positive spreading coefficients or by using cross-linkers with systems
that have negative spreading coefficients. The fabrication was performed by applying liquid onto
the micropillar arrays and then depositing a polymer film using the iCVD process. The liquid
acts as a scaffold, serving as a temporary supporting template for the polymer to deposit onto so
that upon its removal, the polymer retains the shape of the liquid surface. We focused on two
methods for directing liquids on the pillars, each having its own inherent advantages that allow
for the fabrication of canopies using liquids with either high or low surface tensions. The Droplet
Method (Figure 1a) exploits liquids that rest in the Cassie-Baxter state on top of the pillars,
allowing the iCVD process to deposit a polymer film underneath the droplet. This method
typically requires liquids with high surface tensions to generate the Cassie-Baxter state, where
the size of the canopy is dictated by the volume of the liquid and the position of the canopy is
dictated by the location of the droplet. The Inverted Method (Figure 1b) utilizes the inversion of
pillars onto a thin layer of low surface tension liquid such that the iCVD process deposits a
polymer film in the spaces between the pillars on top of the liquid. The height of the canopy can
be easily controlled using the Inverted Method by altering the thickness of the liquid. Due to the
ease of generating thin layers of liquids over large areas, this technique is preferred for
fabricating canopies over larger areas.
78

Figure 1. Schematics showing the fabrication of polymer canopies via a) the Droplet Method
and b) the Inverted Method.

When using the Droplet Method, it is critical to select systems that allow the liquid to
maintain a Cassie-Baxter state, otherwise the liquid will penetrate between the pillars,
compromising the ability of the liquid to act as the desired template for canopy fabrication. Many
studies have documented the requirements for maintaining a Cassie-Baxter state, where factors
such as surface energy and surface roughness have a major influence on droplet
stability.
144,145,146,147
Taking these factors into consideration, we chose 1-ethyl-3-
methylimidazolium tetrafluoroborate ([emim][BF4]) as our scaffolding liquid because its high
a)
b)
Inverted Method
Droplet Method
remove
liquid
remove film
over liquid
apply pillars
remove
liquid
= PDMS = scaffolding liquid = polymer = silicon wafer
apply liquid
mask
79

surface tension (55.6 mN/m)
32
allows it to easily exist in the Cassie-Baxter state on our PDMS
pillar array (height = 60 μm, diameter = 22 μm, and pitch = 18 μm). Canopies were fabricated by
applying a droplet onto a pillar array followed by the deposition of a poly(1H,1H,2H,2H-
perfluorodecyl acrylate–co-ethylene glycol diacrylate) (P(PFDA-co-EGDA)) copolymer coating
(S = 18 mN/m), resulting in the formation of a polymer film over the exposed liquid and solid
surfaces. The thickness of the coating was 1 μm as measured on a reference silicon wafer. The
film residing on top of the droplet was then peeled off with a pair of tweezers such that a
subsequent solvent wash removed the [emim][BF4], leaving behind a P(PFDA-co-EGDA)
canopy on top of the pillar array. An example of a polymer canopy generated with a 5 μL droplet
is depicted in Figure 2. SEM micrographs reveal that the canopy forms at the upper rim of the
pillars (Figure 2a), leaving the tops of the pillars as native PDMS. A cross-sectional SEM
micrograph (Figure 2b) of the sample after it had been cut with a razorblade confirms that the
polymer forms a canopy structure over the pillar arrays where the shorter pillars in the front of
the image are pillars with their tops severed during the cutting process. A range of droplet sizes
from 5 μL to 100 μL were used to fabricate canopies of varying areas, where larger liquid
volumes yielded canopies with larger radii. Using [emim][BF4] droplet volumes of 5 μL, 10 μL,
and 100 μL, we were able to fabricate canopies of approximately 2 mm, 3 mm, and 6 mm in
diameter, respectively. Droplet volumes much larger than 100 μL were difficult to use due to the
tendency of the droplet to roll off the surface of the pillar array. The resultant canopies are
depicted in Figures 2c−e, where the insets show optical microscopy images of the canopies.
Figure 2e also demonstrates the ability to selectively place the liquid droplets to pattern the
location of the canopies. In this case, we deposited the 10 μL droplets of [emim][BF 4] in a 2 × 2
80

array before depositing a P(PFDA-co-EGDA) coating. After removing the liquid, the resultant
canopies remained at the locations once occupied by the droplets.

Figure 2. a) Top down and b) cross-sectional SEM micrographs of P(PFDA-co-EGDA) canopies
fabricated using [emim][BF4] with the Droplet Method. SEM micrographs showing canopies
made with c) a 5 μL droplet, d) a 100 μL droplet, and e) 10 μL droplets in a 2 × 2 array. Optical
microscopy images depicting each canopy are provided as an inset in each image.

15 μm 15 μm
b) a)
c) d) e)
20 μm 20 μm 20 μm
2 mm 2 mm 2 mm
81

The SEM images in Figure 2 show that the canopies are wrinkled. We hypothesize that
the wrinkled texture of the canopies is due to anisotropic swelling, which has been shown to be
responsible for the wrinkling of other films that are anchored to a solid.
148
Our previous studies
of iCVD polymerization onto liquids have shown that there is only surface polymerization in the
cases where the monomer is insoluble in the liquid whereas there is both surface and bulk
polymerization in the cases where the monomer is soluble in the liquid.
31
Since 1H,1H,2H,2H-
perfluorodecyl acrylate (PFDA) is insoluble in [emim][BF4] whereas ethylene glycol diacrylate
(EGDA) is soluble in [emim][BF4], deposition of P(PFDA-co-EGDA) on [emim][BF4] by iCVD
results in the formation of heterogeneous layered films, which was confirmed in our previous
work.
149
During the deposition of polymer, EGDA is able to absorb into the [emim][BF4] and
polymerize so that the resultant heterogeneous polymer film is comprised of a layer of P(PFDA-
co-EGDA) and a layer of polymerized EGDA containing integrated [emim][BF4]. After the
washing step, the removal of [emim][BF4] likely leads to a volume change within the polymer
matrix resulting in the wrinkling of the canopy.
Energy dispersive spectroscopy (EDS) was used to characterize the compositions of the
canopy and the tops of the pillars. A scan of the canopy region showed an elemental composition
of 66.6% C, 11.1% O, 17.8% F, and 4.5% Si. The high fluorine content is attributed to the
poly(1H,1H,2H,2H-perfluorodecyl acrylate) PPFDA within the P(PFDA-co-EGDA) canopy
while the silicon is attributed to the PDMS from the underlying substrate due to the sampling
depth of the analysis method which can exceed 1 μm. In contrast, a scan of the top of the pillars
showed an elemental composition of 52.4% C, 18.1% O, 4.8% F, and 24.7% Si which is similar
to the expected elemental composition of PDMS (50% C, 25% O, 25% Si). The low fluorine
content and the relative decrease in oxygen content relative to PDMS suggest that the canopy
82

material exists in small quantities on the top of the pillars. Since we have previously shown that
PFDA does not polymerize within the bulk of [emim][BF4] due to insolubility,
31
the fluorine
content is attributed to polymer which was displaced onto the pillar surface during the washing
step.
Fabricating canopies using the Droplet Method requires the liquid scaffold to be in a
Cassie-Baxter state, limiting the selection of liquids to those with high surface tensions. High
surface tension liquids can make it difficult to fabricate canopies over large areas since larger
droplet volumes tend to roll off the surface. Furthermore, the Cassie-Baxter state limits control
over the height of the canopy since the liquid is required to rest on top of the pillars. In order to
expand the viable liquids that can be used to generate canopies and introduce a method of control
over the canopy height, we explored an alternative strategy for manipulating the scaffolding
liquid. In this second method (Inverted Method), the pillars are inverted onto a thin layer of
liquid that was spin coated onto a silicon wafer before applying the iCVD process (Figure 1b). A
solvent wash was then used to remove any remaining liquid, revealing a polymer canopy. In
contrast to the Droplet Method, the liquids used in the Inverted Method have low surface
tensions, allowing them to be easily spin coated into thin layers. For this reason, we chose
silicone oil as our model liquid for the Inverted Method. However, since silicone oil is
chemically similar to our PDMS pillar material, we precoated the pillars with 100 nm of PPFDA
to prevent any liquid from diffusing into and swelling the pillars. PPFDA was selected because it
has previously been shown to act as an effective barrier coating against organic liquids.
150,13

Using precoated pillars with the Inverted Method, we successfully fabricated P(PFDA-co-
EGDA) (S = 13 mN/m) canopies with a range of heights by spin coating liquid layers of varying
thicknesses. The thickness of the coating was 1 μm as measured on a reference silicon wafer.
83

SEM micrographs of the resultant canopies are shown in Figure 3 with heights of 58 μm (Figure
3a), 50 μm (Figure 3b), and 46 μm (Figure 3c) with cross-sectional views confirming the canopy
structure of the film. In comparison to the Droplet Method canopies, the relative smoothness of
the Inverted Method canopies is attributed to the insolubility of either monomer within the
silicon oil, preventing the effects of anisotropic swelling. An example of an optical microscopy
image of the canopy with a height of 58 μm is depicted in Figure 3d, demonstrating the ability to
fabricate canopies over large areas. Although the depicted canopy has an approximate area of a
square centimeter, the method can easily be used to make larger canopies by increasing the size
of the pillar array and the area of the liquid layer.

Figure 3. SEM micrographs of P(PFDA-co-EGDA) canopies with a height of a) 58 μm, b) 50
μm, c) and 46 μm fabricated using silicone oil with the Inverted Method, where cross-sectional
10 μm
20 μm
b)
2 mm
d)
10 μm
30 μm
c)
10 μm
a)
40 μm
84

depictions are shown as insets. d) An optical microscopy image of the 58 μm canopy is also
provided, demonstrating the large area of the canopy.
EDS analysis was performed on the canopy and the tops of the pillars. A scan of the
canopy region showed a composition of 64.1% C, 12.8% O, 18.8% F, and 4.3% Si, which is
consistent with the Droplet Method canopy, where the high fluorine content is attributed to the
PPFDA within the P(PFDA-co-EGDA) canopy and the silicon content is attributed to the
underlying PDMS substrate. A scan of the tops of the pillars showed a composition of 56.5% C,
10.7% O, 7.3% F, and 25.5% Si. As expected, there is a significant increase of silicon content
which suggests that the canopy does not form on top of the pillar. The relative decrease in
oxygen content and increase in carbon and fluorine compared to native PDMS can be explained
by the 100 nm precoating of PPFDA which was deposited onto the pillars prior to applying the
Inverted Method.
We have demonstrated thus far that canopies can be fabricated using systems with
positive spreading coefficients. However, it may be desirable to fabricate canopies comprised of
polymers that yield negative spreading coefficients. Attempts at fabricating canopies using
systems with negative spreading coefficients, such as poly(2-hydroxyethyl methacrylate) on
[emim][BF4] or silicone oil, were unsuccessful since the polymer formed particles instead of a
film at the liquid surface. However, we were able to successfully fabricate canopies by adding a
cross-linker to these systems since the cross-linker can covalently attach growing polymer chains
as they deposit on the liquid surface, leading to the formation of a film over the liquid surface.
25

Using EGDA as a cross-linking agent, we successfully fabricated poly(2-hydroxyethyl
methacrylate-co-ethylene glycol diacrylate) (P(HEMA-co-EGDA)) canopies using both the
Droplet Method (S = -22 mN/m) and the Inverted Method (S = -68 mN/m). SEM micrographs of
85

the resultant P(HEMA-co-EGDA) canopies with cross-sectional insets are shown in Figure 4a
and 4b, respectively, where the height of the Inverted Method canopy is 50 μm tall.

Figure 4. SEM micrographs of P(HEMA-co-EGDA) canopies fabricated using a) the Droplet
Method and b) the Inverted Method, where insets show a cross-sectional view of each sample.

Micropillar arrays function well as a model substrate for demonstrating the Droplet
Method and the Inverted Method, but either technique can easily be extended to other
geometries. In order to show this concept, we applied both methods on an alternative geometry
by converting the pillars into hierarchal microstructures (Figure 5a) using a technique we have
previously developed.
151
In this technique, we cast a solution of poly(methyl methacrylate) in
acetone onto the pillar surface. Capillary forces produced from solvent evaporation bring the
pillars into contact forming microstructures which are then stabilized by poly(methyl
methacrylate) welds that form at the interface between the pillars. The successful fabrication of
P(PFDA-co-EGDA) polymer canopies on these microstructures using the Droplet Method and
the Inverted Method are shown in Figure 5b and Figure 5c, respectively, where the height of the
Inverted Method canopy is 46 μm tall.
15 μm 15 μm
a)
15 μm 20 μm
b)
86

Figure 5. SEM micrographs of a) PDMS microstructures stabilized with PMMA, and P(PFDA-
co-EGDA) canopies over microstructures fabricated using b) the Droplet Method and c) the
Inverted Method.

5.3 Conclusion

We have demonstrated a novel scaffolding technique where low vapor pressure liquids
are used in conjunction with the iCVD process to simultaneously act as a mask and temporary
supporting template which allows for the fabrication of polymer canopies over micropillar
arrays. Two strategies for applying the liquid scaffolds were investigated: 1) the Droplet Method,
which uses high surface tension liquids, was used to tune the size of the canopies by selectively
by controlling the volume of the liquid droplet, and 2) the Inverted Method, which was used to
fabricate canopies with a range of heights over large areas using thin layers of low surface
tension liquids with varying thicknesses. Systems with negative spreading coefficients that
would generally not yield successful fabrication of canopies were overcome through the
utilization of a cross-linker, demonstrating that a range of polymer canopies can be synthesized
using either method. The generality of each method was also shown by fabricating canopies over
an alternative microstructure geometry. Additionally, the methods shown are not restricted to
iCVD and can be translated to other vapor deposition systems.
20 μm
a)
20 μm
b)
20 μm
c)
87

5.4 Experimental

[emim][BF4] (97%, Sigma-Aldrich), PFDA (97%, SynQuest), di-tert-butyl peroxide
(98%, Sigma-Aldrich), EGDA (97%, Monomer-Polymer), silicone oil (100 cSt, Sigma-Aldrich),
2-hydroxyethyl methacrylate (98%, Sigma-Aldrich), trichloro(1H,1H,2H,2H-perfluorooctyl)
silane (97%, Sigma Aldrich), glycerol (EMD Chemicals), diiodomethane (99%, Sigma-Aldrich),
methanol (Macron, absolute), hexane (98%, Sigma-Aldrich), and poly(methyl methacrylate)
(500,000 MW, Varian GPC standards) were used as received without further purification.
Micropillar arrays were fabricated by pouring Sylgard 184 mixed at a base to cross-linker
weight ratio of 10:1 into a master mold before curing the mixture at 65 °C for 24 hours. The
mold was fabricated using standard photolithography by spin coating SU-8 2050 photoresist
(MicroChem) before exposing it to UV light through an emulsion transparency mask (CAD/Art
Services, Inc.). Prior to molding PDMS pillars, the mold was treated with
trichloro(1H,1H,2H,2H-perfluorooctyl) silane in a desiccator to reduce the adhesion of cured
PDMS, ensuring the easy release of the micropillar arrays. PDMS microstructures were self-
assembled by pipetting 10 μL of a 0.5 % w/w solution of poly(methyl methacrylate) in acetone
onto 0.7 cm × 0.7 cm PDMS pillar arrays and allowing the solvent to evaporate overnight prior
to use.
All polymer depositions were performed within a custom iCVD reactor chamber (GVD
Corp, 250 mm diameter, 48 mm height) equipped with a nichrome filament array (80% Ni, 20%
Cr, Omega Engineering) maintained at 240 °C. Di-tert-butyl peroxide initiator molecules were
flown in at room temperature through a mass flow controller at a rate of 1.0 sccm. All monomer
inlet lines were heated to temperatures 20 °C greater than the monomer temperatures to prevent
condensation within the lines. P(PFDA-co-EGDA) films were synthesized by maintaining a
88

pressure of 70 mTorr and a stage temperature of 30 °C while heating the PFDA and EGDA
monomers to a temperature of 50 °C to achieve flowrates of 0.3 sccm and 1.2 sccm, respectively.
P(HEMA-co-EGDA) films were synthesized by maintaining a pressure of 50 mTorr and stage
temperature of 25 °C while heating the 2-hydroxyethyl methacrylate monomer to a temperature
of 25 °C and heating the EGDA monomer to a temperature of 30 °C, to achieve flowrates of 0.6
sccm and 0.9 sccm, respectively. Each polymer film was grown to a 1 μm thickness as measured
on a reference silicon wafer using an in situ 633 nm helium−neon laser interferometer (Industrial
Fiber Optics). Polymer canopies were imaged using a SEM (Topcon Aquila) and a stereo
microscope (National). SEM micrographs were acquired using a 15 kV accelerating voltage on
samples that were sputter coated with platinum to prevent charging. EDS was performed using a
SEM (JEOL JSM-7001F) with an EDS detector (EDAX Apollo X).
The [emim][BF4] used in the Droplet Method was evacuated for 24 hours in a desiccator
to remove residual water prior to dispensing it onto PDMS micropillar surfaces using a
micropipette. After depositing the polymer coating, the polymer film residing on top of the
[emim][BF4] droplet was removed using a pair of tweezers to peel the film off the liquid. The
[emim][BF4] droplet was then removed by submerging the pillars in a methanol bath for 15
minutes and allowing the solvent to evaporate for two hours before imaging.
The silicone oil heights used in the Inverted Method were obtained by spin coating 1 mL
of silicone oil (1000 cSt) onto a 3.5 cm × 3.5 cm wafer using speeds of 4000 rpm, 3000 rpm, and
2500 rpm with an acceleration of 100 rpm/s for 30 seconds to obtain thicknesses that produced
canopies with heights of 58 μm, 50 μm, and 46 μm, respectively. After spin coating, all silicone
oil layers were allowed to rest on a hot plate set at a constant temperature of 60 °C for 60
minutes to increase film uniformity. PDMS pillars coated with PPFDA were then gently inverted
89

onto the silicone oil layers using a pair of tweezers and covered with a glass slide to prevent the
pillars from floating. The inverted samples were then placed under vacuum using the iCVD
reactor chamber where the silicone oil was once again heated to 60 °C for an hour to reduce any
non-uniform wetting that may have occurred during the inversion of the pillars onto the silicon
oil. Afterwards, the samples were cooled to the appropriate reactor stage temperature for 20
minutes before depositing a polymer coating. After the coating was applied, the PDMS
substrates were carefully removed from the silicone oil with a pair of tweezers and submerged in
hexane for 15 minutes to wash away any residual silicone oil.
Spreading coefficients were calculated using a goniometer (Rame
́ -Hart Model 290-F1) to
measure advancing contact angles, equilibrium contact angles, liquid−vapor surface tensions,
and polymer−vapor surface tensions. Advancing contact angles were measured using the tilting
base method. Liquid−vapor surface tensions were measured using the pendant drop method,
while polymer−vapor surface tensions were calculated using the acid-base method with
equilibrium contact angles from water, glycerol, and diiodomethane. All measurements were
performed five times.

Corresponding Author
*E-mail: malanchg@usc.edu.
Notes
The authors declare no competing financial interest.

90

5.5 Acknowledgments

This work was supported by the National Science Foundation Division of Civil, Mechanical, and
Manufacturing Innovation Award Number 1069328.

91

5.6 References

112
Fuchs, A D.; Tiller, J. C. Contact-Active Antimicrobial Coatings Derived from Aqueous Suspensions. Angew.
Chem. Int. Ed. 2006, 45, 6759−6762.
113
Wu, Z.; Tong, W.; Jiang W.; Liu, X.; Wang, Y.; Chen, H. Poly(N-vinylpyrrolidone)-Modified
Poly(dimethylsiloxane) Elastomers As Anti-Biofouling Materials. Colloid Surface B. 2012, 96, 37−43.
114
Schmidt, S.; Motschmann, H.; Hellweg, T.; von Klitzing, R. Thermoresponsive Surfaces by Spin-Coating of
PNIPAM-co-PAA Microgels: A Combined AFM and Ellipsometry Study. Polymer 2008, 49, 749−756.
115
Ma, M.; Gupta, M.; Li, Z.; Zhai, L.; Gleason, K. K.; Cohen, R. E.; Rubner, M. F.; Rutledge, G. C. Decorated
Electrospun Fibers Exhibiting Superhydrophobicity. Adv. Mater. 2007, 19, 255–259.
116
Subramani, C.; Ofir, Y.; Patra, D.; Jordan B. J.; Moran, I. W.; Park, M. H.; Carter, K. R.; Rotello, V. M.
Nanoimprinted Polyethyleneimine: A Multimodal Template for Nanoparticle Assembly and Immobilization. Adv.
Funct. Mater. 2009, 19, 2937–2942.
117
Chang, J.-P.; El Rassi, Z.; Horva
́ th, C. Silica-Bound Polyethyleneglycol As Stationary Phase for Separation of
Proteins By High-Performance Liquid Chromatography. J. Chromatogr. 1985, 319, 396−399.
118
Gao, Q.; Xu, D.; Sun, Y.; Li, X. pH-Responsive Drug Release from Polymer-Coated Mesoporous Silica Spheres.
J. Phys. Chem. C 2009, 113, 12753–12758.
119
Tenhaeff, W. E.; Gleason, K. K.; Initiated and Oxidative Chemical Vapor Deposition of Polymeric Thin Films:
iCVD and oCVD. Adv. Funct. Mater. 2008, 18, 979–992.
120
Susut, C.; Timmons, R. B. Plasma Enhanced Chemical Vapor Depositions to Encapsulate Crystals in Thin
Polymeric Films: a New Approach to Controlling Drug Release Rates. Int. J. Pharm. 2005, 288, 253–261.
121
Baxamusa, S. H.; Im, S. G.; Gleason, K. K. Initiated and Oxidative Chemical Vapor Deposition: a Scalable
Method for Conformal and Functional Polymer Films on Real Substrates. Phys. Chem. Chem. Phys. 2009, 11, 5227–
5240.
122
Bose, R. K.; Nejati, S.; Stufflet, D. R.; Lau, K. K. S. Graft Polymerization of Anti-Fouling PEO Surfaces by
Liquid-Free Initiated Chemical Vapor Deposition. Macromolecules 2012, 45, 6915−6922.
123
Kwong, P.; Gupta, M. Vapor Phase Deposition of Functional Polymers onto Paper-Based Microfluidic Devices
for Advanced Unit Operations. Anal. Chem. 2012, 84, 10129−10135.
124
Chen, B.; Kwong. P.; Gupta, M. Patterned Fluoropolymer Barriers for Containment of Organic Solvents within
Paper-Based Microfluidic Devices. ACS Appl. Mater. Interfaces 2013, 5, 12701−12707.
125
Kwong, P.; Seidel, S.; Gupta, M. Effect of Transition Metal Salts on the Initiated Chemical Vapor Deposition of
Polymer Thin Films. J. Vac. Sci. Technol. A 2015, 33, 031504.
126
Chen, B.; Seidel, S.; Hori, H.; Gupta, M. Self-Assembly of Pillars Modified with Vapor Deposited Polymer
Coatings. ACS Appl. Mater. Interfaces 2011, 3, 4201–4205.
127
Lau, K. K. S.; Bico, J, Teo, K. K. B. K.; Chhowalla, M.; Amaratunga, G. A. J.; Milne, W. I.; McKinley, G. H.;
Gleason, K. K. Superhydrophobic Carbon Nanotube Forests. Nano Lett. 2003, 3 (12), 1701−1705.
128
Frank-Finney, R. J.; Haller, P. D.; Gupta, M. Ultrathin Free-Standing Films Deposited onto Patterned Ionic
Liquids and Silicone Oil. Macromolecules 2012, 45, 165−170.
129
Frank-Finney, R. J.; Bradley, L. C.; Gupta, M. Formation of Polymer-Ionic Liquid Gels Using Vapor Phase
Precursors. Macromolecules 2013, 46, 6852−6857.
130
Haller, P. D.; Gupta, M. Synthesis of Polymer Nanoparticles via Vapor Phase Deposition onto Liquid Substrates.
Macromol. Rapid Commun. 2014, 35, 2000−2004.
131
Bradley, L. C.; Gupta, M. Copolymerization of 1-Ethyl-3-Vinylimidazolium Bis(trifluoromethylsulfonyl)imide
via Initiated Chemical Vapor Deposition. Macromolecules 2014, 47, 6657−6663.
132
Bradley, L. C.; Gupta, M. Microstructured Films Formed on Liquid Substrates via Initiated Chemical Vapor
Deposition of Cross-Linked Polymers. Langmuir 2015, 31, 7999–8005.
133
Berg, M. C.; Zhai, L.; Cohen, R. E.; Rubner, M. F. Controlled Drug Release from Porous Polyelectrolyte
Multilayers. Biomacromolecules 2006, 7, 357–364.
134
Wu, J.; Sailor, M. J. Chitosan Hydrogel-Capped Porous SiO 2 as a pH Responsive Nano-Valve for Triggered
Release of Insulin. Adv. Funct. Mater. 2009, 19, 733–741.
135
Zhang, Y.; Ang, C. Y.; Li, M.; Tan, S. Y.; Qu, Q.; Luo, Z.; Zhao, Y. Polymer-Coated Hollow Mesoporous Silica
Nanoparticles for Triple-Responsive Drug Delivery. ACS Appl. Mater. Interfaces 2015, 7, 18179–18187.
136
Ismagilov, R. F.; Ng, J. M. K.; Kenis, P. J. A.; Whitesides, G. M. Microfluidic Arrays of Fluid-Fluid Diffusional
92

Contacts as Detection Elements and Combinatorial Tools. Anal. Chem.2001, 73, 5207–5213.
137
Krumpfer, J. W.; Bian, P.; Zheng, P.; Gao, L.; McCarthy, T. J. Contact Angle Hysteresis on Superhydrophobic
Surfaces: An Ionic Liquid Probe Fluid Offers Mechanistic Insight. Langmuir 2011, 27, 2166–2169.
138
Seiwert, J.; Clanet, C.; Quéré, D. Coating of a Textured Solid. J. Fluid Mech. 2011, 669, 55–63.
139
Smith, J. D.; Dhiman, R.; Anand, S.; Reza-Garduno, E.; Cohen, R. E.; McKinley, G. H.; Varanasi, K. K. Droplet
Mobility on Lubricant-Impregnated Surface. Soft Matter, 2013, 9, 1772–1780.
140
Princen, H. M. Capillary Phenomena in Assemblies of Parallel Cylinders: I. Capillary Rise between Two
Cylinders. J. Colloid Interface Sci. 1969, 30, 69-75.
141
Princen, H. M. Capillary Phenomena in Assemblies of Parallel Cylinders: II. Capillary Rise in Systems with
More than Two Cylinders. J. Colloid Interface Sci. 1969, 30, 359-371.
142
Haller, P. D.; Frank-Finney, R. J.; Gupta, M. Vapor-Phase Free Radical Polymerization in the Presence of an
Ionic Liquid. Macromolecules, 2011, 44 (8), 2653–2659.
143
Haller, P. D.; Bradley, L. C.; Gupta, M. Effect of Surface Tension, Viscosity, and Process Conditions on Polymer
Morphology Deposited at the Liquid-Vapor Interface. Langmuir 2013, 29, 11640−11645.
144
Extrand, C. W. Criteria for Ultralyophobic Surfaces. Langmuir 2004, 20, 5013-5018.
145
Patankar, N. A. Transition between Superhydrophobic States on Rough Surfaces. Langmuir 2004, 20, 7097-7102.
146
Lee, J. B.; Gwon, H. R.; Lee, S. H.; Cho, M. Wetting Transition Characteristics on Microstructured Hydrophobic
Surfaces. Mater. Trans., JIM 2010, 51, 1709-1711.
147
Bhushan, B.; Nosonovsky, M.; Jung, Y. C. Towards Optimization of Patterned Superhydrophobic Surfaces. J. R.
Soc. Interface 2007, 4, 643–648.
148
Yang, S.; Khare, K.; Lin. P. C. Harnessing Surface Wrinkle Patterns in Soft Matter. Adv. Funct. Mater. 2010, 20,
2550–2564.
149
Bradley, L. C.; Gupta, M. Formation of Heterogeneous Polymer Films via Simultaneous or Sequential
Depositions of Soluble and Insoluble Monomers onto Ionic Liquids. Langmuir, 2013, 29, 10448–10454.
150
Riche, C. T.; Marin, B. C.; Malmstadt, N.; Gupta, M. Vapor Deposition of Cross-linked Fluoropolymer Barrier
Coatings onto Pre-Assembled Microfluidic Devices. Lab Chip, 2011, 11, 3049–3052.
151
Chen, B.; Riche, C. T.; Lehmann, M.; Gupta, M. Responsive Polymer Welds via Solution Casting for Stabilized
Self-Assembly. ACS Appl. Mater. Interfaces 2012, 4, 6911−6916.

93

6. Conclusions and Future Work

Controlling the properties of structured surfaces is important for a multitude of
applications such as optics, adhesives, separations, and assay development. Many methods for
modifying and manipulating structured surfaces using polymers that were applied using a variety
of techniques. The utility of these methods were demonstrated by controlling the actuation of
microstructures, dictating the mobility of microbeads, containing organic solvents within paper-
based microfluidic, and fabrication of polymer canopy structures.
The most recent development in this work, pertaining to the fabrication of polymer
structures using liquid scaffolding, has potential for additional applications. By applying liquid
scaffolding to microfluidic devices, it may be possible to directly fabricate membranes within
microfluidic channels, allowing for membrane based operations to be used within the devices.
Liquid scaffolding also has the potential to generate mechanical testing devices for vapor
deposited films, which can be difficult using standard techniques due to the brittle nature of the
films. These types of Issues could potentially be solved by depositing an identical film as a
canopy onto strategically spaced micropillars. A liquid can then be applied to the micropillars to
induce capillary forces resulting in pillar actuation. Since the motion of the pillars is due to a
balance of forces, real-time observation of the pillars can then be used to analyze the force
applied to the film when it cracks to determine mechanical properties.

Abstract (if available)

Abstract In this document, the numerous interesting properties and applications that can be achieved using complex geometries such as pillar arrays, microbead layers, and porous media, are described. These properties include, but are not limited to, increased adhesion, optical effects, self-cleaning properties, and improved separations. Additional interesting properties are achievable if these complex surfaces are not limited to a single conformation. In order to manipulate pillar surfaces and alter their conformations, we have self-assembled numerous hierarchal microstructure systems composed of pillars made from elastomeric materials and stabilized them using polymer coatings that can act as an adhesive. By using different strategies to destabilize the polymer adhesive, a multitude of methods for manipulating the pillars were achieved. The two methods of applying polymer coatings onto these complex geometries are described, where a vapor phase polymerization system (section 1.4) is used to coat pillar surfaces in section 2 and porous media in section 4, and a simpler solution casting method is used to apply polymer on pillar surfaces and microbeads in section 3. Sections 2 and 4 also describe two methods for patterning polymer on complex geometries, including a photoresist method and the use of an inhibitor that prevents polymer from depositing is selected locations. ❧ An additional method of applying scaffolded polymer coatings onto complex geometries is described in section 5, where low vapor pressure liquids are used to direct the deposition of vapor deposited polymer. We demonstrate two methods of directing the liquid onto surfaces: The first uses high surface tension liquids to allow for patterning and control over the radial dimension of the polymer canopies and the second uses low surface tension liquids to allow for ease of fabrication over larges areas in addition to providing control over the height of the canopy (section 5.2).

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Vapor phase deposition of dense and porous polymer coatings and membranes for increased sustainability and practical applications

PDF

Scale-up of vapor-phase deposition of polymers: towards large-scale processing

PDF

The patterning of polymer thin films on porous substrates via initiated chemical vapor deposition

PDF

Fabrication of polymer films on liquid substrates via initiated chemical vapor deposition: controlling morphology and composition

PDF

Formation of polymer gels, films, and particles via initiated chemical vapor deposition onto liquid substrates

PDF

Understanding structural changes of polymer films during vapor phase deposition

PDF

Vapor phase deposition of polymers in the presence of low vapor pressure liquids

PDF

Inorganic/organic hybrid materials and grafted coatings via vapor phase deposition processes

PDF

Simultaneous monomer deposition and polymerization at low substrate temperatures for the formation of porous polymer membranes

PDF

Controlling polymer film patterning, morphologies, and chemistry using vapor deposition

PDF

Fabrication of functional porous membranes via polymerization of solid monomer by a vapor-phase initiator

PDF

Novel processing of liquid substrates via initiated chemical vapor depostion

PDF

Fabrication of silicon-based membranes via vapor-phase deposition and pyrolysis of organosilicon polymers

PDF

Synthesis of high-quality nanoparticles using microfluidic platforms

PDF

Strategies for improving mechanical and biochemical interfaces between medical implants and tissue

PDF

Fabrication and study of organic and inorganic optoelectronics using a vapor phase deposition (VPD)

PDF

Real time surface analysis of complex oxide thin films during pulsed laser deposition

PDF

Slurry based stereolithography: a solid freeform fabrication method of ceramics and composites

PDF

Development of implantable Parylene-based MEMS technologies for cortical applications

PDF

Influence of polymer structure on electronic properties and performance as donor materials in bulk heterojunction solar cells

Asset Metadata

Creator Chen, Benny (author)

Core Title Selective deposition of polymer coatings onto structured surfaces

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Chemical Engineering

Publication Date 02/22/2016

Defense Date 07/23/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag coatings,OAI-PMH Harvest,polymers,thin films,vapor deposition

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Gupta, Malancha (committee chair), Malmstadt, Noah (committee member), Meng, Ellis (committee member)

Creator Email bennyc617@gmail.com,bennyche@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-211056

Unique identifier UC11278265

Identifier etd-ChenBenny-4124.pdf (filename),usctheses-c40-211056 (legacy record id)

Legacy Identifier etd-ChenBenny-4124.pdf

Dmrecord 211056

Document Type Dissertation

Format application/pdf (imt)

Rights Chen, Benny

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA