micromachines
Article
A Simple Method for Fabrication of Microstructures
Using a PDMS Stamp
Hun Lee 1 , Domin Koh 1 , Linfeng Xu 1 , Sindhu Row 2 , Stelios T. Andreadis 2 and Kwang W. Oh 1, *
1
2
*
Sensors and MicroActuators Learning Laboratory (SMALL), Department of Electrical Engineering,
University at Buffalo, State University of New York, Buffalo, NY 14260, USA; [email protected] (H.L.);
[email protected] (D.K.); [email protected] (L.X.)
Department of Chemical and Biological Engineering, University at Buffalo, State University of New York,
Buffalo, NY 14260, USA; [email protected] (S.R.); [email protected] (S.T.A.)
Correspondence: [email protected]; Tel.: +1-716-645-1025
Academic Editor: Nam-Trung Nguyen
Received: 14 June 2016; Accepted: 19 September 2016; Published: 1 October 2016
Abstract: We report a simple method to fabricate PDMS (polydimethylsiloxane) microwell arrays on
glass by using a PDMS stamp to study cell-to-cell adhesion. In the cell-to-cell study, a glass substrate
is required since glass has better cell attachment. The microwell arrays are replicated from an SU-8
master mold, and then are transferred to a glass substrate by lifting the PDMS stamp, followed by
oxygen plasma bonding of the PDMS stamp on the glass substrate. For the cell-to-cell adhesion,
four different types of PDMS arrays (e.g., rectangle, bowtie, wide-rhombus, and rhombus) were
designed to vary the cell-to-cell contact length. The transfer success rates of the microwell arrays
were measured as a function of both the contact area of the PDMS and the glass substrate and the
different ratios between the base polymers and the curing agent. This method of generating the
microwell arrays will enable a simple and robust construction of PDMS-based devices for various
biological applications.
Keywords: PDMS microwell; PDMS microwell transfer; PDMS stamp; cell-to-cell study
1. Introduction
The microwell array has become an essential analysis tool for various biological and chemical
applications [1]. It is important to develop efficient fabrication methods for the microwell arrays, which
can offer lower costs and simpler production methods [1–3]. Particularly, PDMS (polydimethylsiloxane)
microwell arrays have been employed in manipulating and culturing cells for studying the individual
behavior of cells in a microenvironment. Due to the fact that PDMS can be easily fabricated by using
soft-lithography techniques and because it possesses many advantageous properties such as optical
transparency, a non-toxic nature, and biocompatibility, it is considered to be a good material for various
biological experiments [4,5]. In the early stages of cell-to-cell interaction studies, Petri dishes were
used to form cell-to-cell contact [5]. In cell-to-cell studies, cells must be attached on the substrate, and
glass is a good substrate for attaching cells because cells like hard surfaces [6].
However, the ability to observe the cells that grow in the dish is limited due to the difficulty
in growing and pairing cells for cell-to-cell study. Under conventional cell culture conditions, it is
difficult to manipulate the cell position, spreading, shape and density. For a controlled cell culture
environment, a PDMS stamping (or imprinting) method has been widely used since it can easily print
micropatterns on a substrate of interest [7–9]. This technology can offer a cell culture environment
with well-controlled sizes, shapes, and positions on a substrate, thus providing a useful tool for cell
studies. Gray et al. demonstrated a surface patterning method with cell-adhesive molecules using
the PDMS stamping method for cell-to-cell contact studies [10]. Despite its huge potential, the contact
Micromachines 2016, 7, 173; doi:10.3390/mi7100173
www.mdpi.com/journal/micromachines
Micromachines 2016, 7, 173
2 of 9
PDMS
stamping method for cell-to-cell contact studies [10]. Despite its huge potential, the contact
Micromachines 2016, 7, 173
2 of 9
printing method by the PDMS stamp should be carefully controlled to avoid deformation of patterns
limiting the robust fabrication of the device for testing. With a similar approach, Nelson et al. utilized
printing micropatterns
method by thebased
PDMSon
stamp
shouldstamp
be carefully
controlled
to substrate
avoid deformation
of patterns
agarose
the PDMS
method
on a glass
for studying
cell-tolimiting
the
robust
fabrication
of
the
device
for
testing.
With
a
similar
approach,
Nelson
et
al.
utilized
cell contact [11]. Its limitation is that, using this method, creating thick and high density patterns is
agarose micropatterns
based
on the
PDMS
stamp
method
on a glass
substrate
for studying
cell-to-cell
challenging
and, also, air
bubbles
can
be easily
trapped
during
the filling
process
of the agarose.
contact
[11].
limitation aissimple
that, using
this method,
thick
highmicrowell
density patterns
is
Here
we Its
demonstrate
fabrication
method creating
for creating
theand
PDMS
arrays by
challenging
and,
also,
air
bubbles
can
be
easily
trapped
during
the
filling
process
of
the
agarose.
transferring the patterns of the PDMS stamp onto a glass substrate. This method overcomes the
Here we
a simple
fabrication
method
forsince
creating
the cured
PDMSPDMS
microwell
arrays by
limitations
of demonstrate
the conventional
PDMS
imprinting
method
it uses
micropatterns
transferring
the
patterns
of
the
PDMS
stamp
onto
a
glass
substrate.
This
method
overcomes
only. Cured PDMS allows negligible volume shrink in microwell transfer because it is solid and the
limitations
the conventional
PDMS imprinting
method
it uses
curedtreatment
PDMS micropatterns
only.
fact
that noofextra
chemical treatment
is necessary,
just since
oxygen
plasma
for the bonding
Cured PDMS
allowsprocess,
negligible
volume
shrinkand
in microwell
transferofbecause
it is solid
the fact
during
the transfer
allows
low-cost
quick fabrication
the devices.
Theand
handling
ofthat
the
no
extra
chemical
treatment
is
necessary,
just
oxygen
plasma
treatment
for
the
bonding
during
device is easy and volume shrink is not as severe as when using agarose, without deformation of the
transfer process, allows low-cost and quick fabrication of the devices. The handling of the device is
pattern.
easy and volume shrink is not as severe as when using agarose, without deformation of the pattern.
2. Methods/Experiment
2. Methods/Experiment
2.1. Microarrays and Microfluidic Channel Design
2.1. Microarrays and Microfluidic Channel Design
For cell-to-cell interaction study using the suggested method, four different types of microwell
For cell-to-cell interaction study using the suggested method, four different types of microwell
arrays were designed to vary cell-to-cell contact length and to study the cell-to-cell interaction study.
arrays were designed to vary cell-to-cell contact length
and to study the cell-to-cell interaction study.
All types microwell have the same area, 12502 μm2, but different cell-to-cell contact length. The
All types microwell have the same area, 1250 µm , but different cell-to-cell contact length. The diameter
diameter of each microwell is in Figure 1. Pairs of cells will be trapped in each well like circles in each
of each microwell is in Figure 1. Pairs of cells will be trapped in each well like circles in each
well in Figure 1. By changing the width of the centre of well, the contact area of each cell can be easily
well in Figure 1. By changing the width of the centre of well, the contact area of each cell can be
controlled.
easily controlled.
The size of the PDMS microwell array stamp was 5 mm × 5 mm and the microwell arrays of each
The size of the PDMS microwell array stamp was 5 mm × 5 mm and the microwell arrays of
pattern has a 10 μm spacing (s in Figure 2) between patterns which is the width of the PDMS
each pattern has a 10 µm spacing (s in Figure 2) between patterns which is the width of the PDMS
microwell. The heights of all patterns are 20 μm (h in Figure 2). The PDMS patterns with 20 μm
microwell. The heights of all patterns are 20 µm (h in Figure 2). The PDMS patterns with 20 µm
thickness were designed and fabricated on the glass substrate for the cell-to-cell adhesion studies. By
thickness were designed and fabricated on the glass substrate for the cell-to-cell adhesion studies.
conventional soft-lithography, four different types of microwell arrays (e.g., rectangle, bowtie, wideBy conventional soft-lithography, four different types of microwell arrays (e.g., rectangle, bowtie,
rhombus, and rhombus) were replicated from a master mould as shown in Figure 1. The density of
wide-rhombus, and rhombus) were replicated from a master mould as shown in Figure 1. The density
patterns and contact surface area between the patterns and the glass are listed in Table 1.
of patterns and contact surface area between the patterns and the glass are listed in Table 1.
Figure 1. Schematic illustration of the four different types of microwell arrays to trap two cells for
Figure 1. Schematic illustration of the four different types of microwell arrays to trap two cells for
cell-to-cell contact. Each circle in each well represents the trapped cell in each well. The pattern size for
cell-to-cell contact. Each circle in each well represents the trapped cell in each well. The pattern size
all the shapes is fixed at 1250 µm2 . (a) Bowtie; (b) Rectangle; (c) Rhombus; (d) Wide-rhombus shape.
for all the shapes is fixed at 1250 μm2. (a) Bowtie; (b) Rectangle; (c) Rhombus; (d) Wide-rhombus
shape.
Micromachines 2016, 7, 173
3 of 9
Micromachines 2016,
7, 173
Table
1. PDMS (polydimethylsiloxane) microwell arrays for cell-to-cell adhesion.
3 of 9
Table 1. PDMS (polydimethylsiloxane)
microwell
for cell-to-cell
Shape
Rectangle
Bowtie arraysWide
Rhombusadhesion.
Rhombus
2)
Shape
937.11
Density of pattern (#/mm
2)
Density
of
pattern
(#/mm
Contact area of PDMS
41.43
Contact
area of PDMS (polydimethylsiloxane)
(%)
(polydimethylsiloxane)
(%)
Rectangle
797.20
937.11
50.18
41.43
Bowtie641.46
Wide Rhombus 946.84
Rhombus
797.20
641.46
946.84
40.82
50.18 59.91 59.91
40.82
2.2. Fabrication
2.2.
Fabrication
The fabrication
fabrication process
process of
of PDMS
PDMS microwell
microwell arrays
The
arrays on
on aa glass
glass substrate
substrate is
is illustrated
illustrated in
in Figure
Figure 2.
2.
The
20-μm-thick
master
mould
(h
in
Figure
2)
was
initially
fabricated
through
photolithography
The 20-µm-thick master mould (h in Figure 2) was initially fabricated through photolithography
process using
using an
an SU-8
SU-8 negative
negative photoresist
photoresist (SU-8
(SU-8 2015,
2015, Micro-Chem
Micro-Chem Corp.,
process
Corp., Newton,
Newton, MA,
MA, USA)
USA) [12],
[12],
and then
thenititwas
was
treated
with
hexamethyldisilazane
(Sigma
Aldrich,
Saint Louis,
MO,inUSA)
in a
and
treated
with
hexamethyldisilazane
(Sigma
Aldrich,
Saint Louis,
MO, USA)
a vacuum
vacuum
chamber
for
1
h
to
easily
peel
off
the
PDMS
stamp
from
the
mold
(Figure
2a).
PDMS
materials
chamber for 1 h to easily peel off the PDMS stamp from the mold (Figure 2a). PDMS materials
were made
made using
using pre-polymer
pre-polymer and
and curing
curing agent
agent (Sylgard
(Sylgard 184,
184, Dow
Dow Corning
were
Corning Co.,
Co., Midland,
Midland, MI,
MI, USA),
USA),
varying
the
mixing
ratio
from
5:1
to
20:1,
from
more
stiff
to
less
stiff.
To
mould
the
PDMS
against
the
varying the mixing ratio from 5:1 to 20:1, from more stiff to less stiff. To mould the PDMS against
master,
the
mixture
of
pre-polymer
and
curing
agent
was
carefully
poured
onto
the
SU-8
master
the master, the mixture of pre-polymer and curing agent was carefully poured onto the SU-8 master
mould and
min
(Figure
2b).
TheThe
molded
PDMS
replica
was was
peeled
off and
◦ Cfor
mould
and cured
cured at
at65
65°C
for3030
min
(Figure
2b).
molded
PDMS
replica
peeled
off both
and
sides
of
the
PDMS
was
bonded
irreversibly
on
two
glass
substrates
by
oxygen
plasma
treatment
using
both sides of the PDMS was bonded irreversibly on two glass substrates by oxygen plasma treatment
plasmaplasma
cleanercleaner
PDC-32G
(Harrick
Plasma,Plasma,
Ithaca, NY,
USA)
at USA)
poweratofpower
18 W (Figure
Then
the
using
PDC-32G
(Harrick
Ithaca,
NY,
of 18 W2c,d).
(Figure
2c,d).
PDMS was lifted from the bottom glass substrate which leaving PDMS microwell array on the
Then the PDMS was lifted from the bottom glass substrate which leaving PDMS microwell array on
substrate (Figure 2e). The process of PDMS microwell array transfer on glass substrate take about 40
the substrate (Figure 2e). The process of PDMS microwell array transfer on glass substrate take about
min excluding the fabrication process of the SU-8 master mold, which takes about 2 h including
40 min excluding the fabrication process of the SU-8 master mold, which takes about 2 h including
hexamethyldisilazane treatment.
hexamethyldisilazane treatment.
Figure 2.
onon
thethe
glass
slide.
(a) (a)
TheThe
microwell
arrays
are
2. Schematic
Schematicillustration
illustrationofofthe
thefabrication
fabricationprocess
process
glass
slide.
microwell
arrays
formed
on a silicon
by using
standard
photolithography.
(b) The PDMS (polydimethylsiloxane)
are formed
on awafer
silicon
wafer
by using
standard photolithography.
(b) The PDMS
is
poured onto the moldisand
cured.onto
(c) The
off from the
mold.
both
(polydimethylsiloxane)
poured
the replicated
mold and PDMS
cured. is
(c)peeled
The replicated
PDMS
is Then
peeled
off
sides
withofoxygen
to be bonded
to glass
substrates,
as shown
(d).
from of
thePDMS
mold. were
Thentreated
both sides
PDMSplasma
were treated
with oxygen
plasma
to be bonded
toinglass
(d)
Both sides
PDMSinare
bonded
by sides
two glass
slides,are
then
top glass
removed
with
substrates,
as of
shown
(d).
(d) Both
of PDMS
bonded
bysubstrate
two glassisslides,
then
topPDMS
glass
microwell
array
left
on
the
bottom
glass
slide.
(e)
By
applying
the
mechanical
force
to
the
top
glass
substrate is removed with PDMS microwell array left on the bottom glass slide. (e) By applying the
slide,
the patterns
can
easily
transferred
to the bottom
mechanical
force to
thebetop
glass
slide, the patterns
can beglass.
easily transferred to the bottom glass.
2.3. Cell-to-Cell Study
studies using
using the
the microwell
microwell arrays,
arrays, human
human BM-MSCs
BM-MSCs (bone
(bone marrow
marrow mesenchymal
mesenchymal
For the cell studies
stem cells) between passage 4 and 8 were
were trypsinized
trypsinized from
from normal
normal culture
culture conditions
conditions (Dulbecco’s
(Dulbecco’s
modified Eagle
Eaglemedium
mediumcontaining
containing10%
10%fetal
fetalbovine
bovineserum)
serum)
then
plated
onto
microwell
arrays
modified
then
plated
onto
thethe
microwell
arrays
on
on glass
the glass
substrate
were placed
in plates.
six-wellFor
plates.
For the experiments,
cells were
the
substrate
whichwhich
were placed
in six-well
the experiments,
cells were trypsinized
trypsinized
counted
usinghemacytometer.
a standard hemacytometer.
The
cellsused
were
in each
and
countedand
using
a standard
The cells were
then
inthen
eachused
pattern;
cellspattern;
settled
cells
settledinto
by the
gravity
into the
microwell
arrays,
and were
allowed
to h.
attach
for 48 this,
h. Following
this,
by
gravity
microwell
arrays,
and were
allowed
to attach
for 48
Following
cells trapped
cells trapped in the microwell arrays were fixed using 4% paraformaldehyde, blocked with 5% goat
serum for 1 h and exposed to primary antibody αSMA (α-smooth muscle actin, Sigma-aldrich, St.
Louis, MI, USA) in 1:100 dilution. After incubation overnight at 4 °C, corresponding secondary
Micromachines 2016, 7, 173
4 of 9
Micromachines
2016, 7,arrays
173
in the microwell
were fixed using 4% paraformaldehyde, blocked with 5% goat serum for4 1ofh9
and exposed to primary antibody αSMA (α-smooth muscle actin, Sigma-aldrich, St. Louis, MI, USA)
antibodies
were used
1:200 dilution
for 1 at
h 4at◦ C,
room
temperature.
Samples antibodies
were counterstained
in 1:100 dilution.
Afterin
incubation
overnight
corresponding
secondary
were used
with
Hoechst
33342
dye
(EMD
Millipore
Laboratory
Chemicals,
Billerica,
MA,
USA;
10 mg/mL;
in 1:200 dilution for 1 h at room temperature. Samples were counterstained with Hoechst
333421:200
dye
dilution;
5 min at
room temperature)
nuclei. MA,
Images
were
obtained 1:200
usingdilution;
a Zeiss Axio-observer
(EMD
Millipore
Laboratory
Chemicals,for
Billerica,
USA;
10 mg/mL;
5 min at room
(Zeiss,
Pleasanton,
CA, Images
USA) and
Image
J software
(NIH,(Zeiss,
Bethesda,
MA, USA,
temperature)
for nuclei.
wereanalyzed
obtained by
using
a Zeiss
Axio-observer
Pleasanton,
CA,
https://imagej.nih.gov/ij/).
USA) and analyzed by Image J software (NIH, Bethesda, MA, USA, https://imagej.nih.gov/ij/).
3. Result/Discussion
Result/Discussion
The image of four different shapes of microwell arrays transferred onto the glass substrate using
the stamping method is shown in Figure 3. One pair of cells is expected to be trapped in each of the
microwells because the diameter of each cell is approximately 20 µm
μm (the size of the well is 1250 μm
µm22,,
2
625 μm
µm for
for each
each cell,
cell, so
so itit is
is hard
hard to
to trap
trap more
more than
than two cells in each well).
well). To
To transfer the microwell
arrays from the PDMS stamp to the glass substrate, homemade tearing equipment was used to exert
a force onto the top glass substrate. It is important to keep the peeling direction the same and while
the force was continually applied, the stamp was peeled away as shown in Figure 4. As a result, the
microwell arrays
fractured
and
torn
along
a mechanically
weak
corner
and
arrayson
onthe
thePDMS
PDMSstamp
stampwere
were
fractured
and
torn
along
a mechanically
weak
corner
stable
PDMS
patterns
were transferred
from the
PDMS
the bottom
substrate.
and stable
PDMS
patterns
were transferred
from
thestamp
PDMStostamp
to theglass
bottom
glass Naturally,
substrate.
the
PDMSthe
microwell
arrays were
transferred
onto theonto
topthe
glass
due todue
thetofact
that that
the
Naturally,
PDMS microwell
arrays
were transferred
topsurface
glass surface
the fact
bonding
force
between
the
PDMS
and
the
bottom
glass
is
larger
than
that
which
is
between
the
PDMS
the bonding force between the PDMS and the bottom glass is larger than that which is between the
microstructure
and theand
topthe
glass.
PDMS microstructure
top glass.
Figure 3.
of of
thethe
microwell
arrays
on the
The microwell
arrays have
a 20-μm3. Images
Images
microwell
arrays
onglass
the substrate.
glass substrate.
The microwell
arrays
have
thick
and 10 μm
among
the patterns.
Trapped
cells incells
microwells
were were
fluorescent-dyed
and
a 20-µm-thick
andgap
10 µm
gap among
the patterns.
Trapped
in microwells
fluorescent-dyed
showed
on the
sideside
in (a,b).
(a) Rectangle.
Dashed
lineline
is illustrated
on on
thethe
bottom
and
it it
is is
a
and showed
on right
the right
in (a,b).
(a) Rectangle.
Dashed
is illustrated
bottom
and
side-view
ofof
the
PDMS
pattern
glass
slide.
The
h ofh the
PDMS
is 20
andand
s iss10
(b) (b)
Bowtie;
(c)
a side-view
the
PDMS
pattern
glass
slide.
The
of the
PDMS
is μm
20 µm
is μm;
10 µm;
Bowtie;
Rhombus;
(d) (d)
Wide-rhombus
shape.
(c)
Rhombus;
Wide-rhombus
shape.
Micromachines 2016, 7, 173
Micromachines 2016, 7, 173
5 of 9
5 of 9
Figure 4. The illustration of peeling off PDMS after bonding. (a) Bonding two glass substrates to each
Figure 4. The illustration of peeling off PDMS after bonding. (a) Bonding two glass substrates to each
side of the PDMS layer. (b) Peeling off to transfer patterns by applying force in the same direction. In
side of the PDMS layer. (b) Peeling off to transfer patterns by applying force in the same direction.
this illustration, the patterns are transferred on the top slide but it is just for convenience (transferring
In this illustration, the patterns are transferred on the top slide but it is just for convenience (transferring
to either the top or bottom glass does not matter).
to either the top or bottom glass does not matter).
As shown in Figure 5a, the transfer rate was measured as a function of the contact area of the
shown
5a, the transfer
ratestamp
was measured
as of
a function
the contact
areawhich
of the
PDMSAs
stamp
to in
theFigure
glass substrate
using the
with a size
5 mm × 5ofmm.
Error bars,
PDMS
stamp
to
the
glass
substrate
using
the
stamp
with
a
size
of
5
mm
×
5
mm.
Error
bars,
which
are the range between the maximum value and minimum value, were obtained through repeating
aresame
the range
between
maximum
value
and minimum
value,The
were
obtained
repeating
the
experiment
10 the
times;
the points
indicate
average values.
transfer
ratethrough
is defined
as the
the
same
experiment
10
times;
the
points
indicate
average
values.
The
transfer
rate
is
defined
transferred number of patterns divided by the initial number of PDMS patterns. Since PDMSasisthe
a
transferred
number
of
patterns
divided
by
the
initial
number
of
PDMS
patterns.
Since
PDMS
is
flexible material, the variation in the transfer rate is large. However, the average was above 50%, ina
flexible
in the
ratewas
is large.
the average
in
the
rangematerial,
of 40.82%the
to variation
59.91%, and
the transfer
maximum
76% atHowever,
the low contact
area ofwas
theabove
PDMS50%,
to the
the
range
of
40.82%
to
59.91%,
and
the
maximum
was
76%
at
the
low
contact
area
of
the
PDMS
to
the
glass. It shows that the patterns with a smaller contact area are more easily transferred since the crack
glass. It shows
thatthus,
the patterns
withstarts
a smaller
contact
area are
morethe
easily
transferred
since the crack
resistance
is low;
the crack
more
uniformly
along
edge
of the patterns
at a
resistance
is
low;
thus,
the
crack
starts
more
uniformly
along
the
edge
of
the
patterns
at
a
mechanically
mechanically unstable point. However, with increasing the contact area, the success
rate was
unstable point.
However,
with
increasing
the area
contact
area,
successbonding
rate wasforce;
gradually
decreased
gradually
decreased
because
a higher
contact
leads
to athe
stronger
this causes
the
because
a higher
tomore
a stronger
force;
thisthe
causes
PDMS
stamp
to becontact
broken.area
As aleads
result,
PDMSbonding
residue is
left on
glass.the PDMS stamp to be
broken.
As a result,
morerequired
PDMS residue
is left
on the
glass.
Moreover,
the force
to lift the
PDMS
was
measured as a function of the contact area
Moreover,
the
force
required
to
lift
the
PDMS
was
measured
as aafunction
of theiscontact
and
and the result is shown in Figure 5b. As the contact area is higher,
greater area
bondedarea
to the
the
result
is
shown
in
Figure
5b.
As
the
contact
area
is
higher,
a
greater
area
is
bonded
to
the
glass,
and
glass, and therefore a stronger force is required to lift it from the PDMS which can break the PDMS
therefore
a
stronger
force
is
required
to
lift
it
from
the
PDMS
which
can
break
the
PDMS
stamp
itself.
stamp itself.
Thestiffness
stiffnessof
ofthe
thePDMS
PDMSstamp
stampcan
canbe
becontrolled
controlledby
byadjusting
adjustingthe
themixing
mixingratio
ratiobetween
betweenthe
the
The
pre-polymer
and
the
curing
agent.
Higher
ratios
of
the
curing
agent
lead
to
the
increased
stiffness
pre-polymer and the curing agent. Higher ratios of the curing agent lead to the increased stiffness
andhardness
hardnessofofthe
the
PDMS,
mainly
caused
by increase
an increase
in cross-linking
The mixing
and
PDMS,
mainly
caused
by an
in cross-linking
[13]. [13].
The mixing
ratios ratios
(pre(pre-polymer:curing
varied
5:120:1
to 20:1
transfer
rate
wasmeasured
measuredasasdescribed
described
polymer:curing
agent)agent)
varied
fromfrom
5:1 to
andand
thethe
transfer
rate
was
previously. Figure
Figure66shows
showsthe
thetransfer
transferrate
rateas
asaafunction
functionof
ofthe
theratio
ratiobetween
betweenthe
thepre-polymer
pre-polymerand
and
previously.
thecuring
curingagent.
agent.The
The
error
bars
were
obtained
through
repeating
the same
experiment
10 times.
the
error
bars
were
obtained
through
repeating
the same
experiment
10 times.
The
The
contact
area
of
the
PDMS
to
the
glass
was
fixed
at
60%.
Young’s
modulus
is
defined
by
Hooke’s
contact area of the PDMS to the glass was fixed at 60%. Young’s modulus is defined by Hooke’s
law
E =where
σ/ε, where
is an applied
stress
ε is a resultant
strain
the
samesmaller
strain, strains
smaller
Elaw
= σ/ε,
σ is anσapplied
stress and
ε isand
a resultant
strain [14].
For[14].
the For
same
strain,
strains indicate
higher stiffness
of the material.
Atcross-linking
the low cross-linking
density
therate
transfer
indicate
a higherastiffness
of the material.
At the low
density (20:1),
the(20:1),
transfer
was
rate
was
increased
to
85%
since
the
PDMS
can
be
easily
elongated,
causing
the
crack
at
the
corner
of
increased to 85% since the PDMS can be easily elongated, causing the crack at the corner of the
the
patterns.
Also,
a
smaller
amount
of
curing
agent
in
the
PDMS
has
a
better
transfer
rate
must
be
patterns. Also, a smaller amount of curing agent in the PDMS has a better transfer rate must be caused
caused
by cohesion
of the different
A smaller
of agent
curingmust
agentlead
must
to
by
cohesion
of the different
polymerpolymer
mixture.mixture.
A smaller
amountamount
of curing
to lead
lower
lower
cohesion
of
the
polymer,
which
allows
for
easier
breaking.
cohesion of the polymer, which allows for easier breaking.
Micromachines 2016, 7, 173
Micromachines 2016, 7, 173
6 of 9
6 of 9
Micromachines 2016, 7, 173
6 of 9
Figure
Figure 5.
5. Homemade
Homemade equipment
equipment for
for lifting
lifting PDMS
PDMS (on
(on the
the left
left side)
side) and
and its
its power
power range
range (with
(with relation
relation
of
force
to
PDMS
transfer
success
rate).
Error
bars
are
the
range
between
the
maximum
value
and
Figure
5.
Homemade
equipment
for
lifting
PDMS
(on
the
left
side)
and
its
power
range
(with
relation
of force to PDMS transfer success rate). Error bars are the range between the maximum value and
minimum
value;
the
points
indicate
average
values.
The
force
is
applied
through
the
clamp
which
is
of force to value;
PDMSthe
transfer
rate).
Errorvalues.
bars are
theforce
range
betweenthrough
the maximum
value
and
minimum
pointssuccess
indicate
average
The
is applied
the clamp
which
connected
to
a
point
on
the
glass
substrate,
as
shown
in
Figure
4.
(a)
The
relationship
between
transfer
minimum
value;
points
indicate
values.
The force
is applied
the clamp between
which is
is
connected
to athe
point
on the
glass average
substrate,
as shown
in Figure
4. (a)through
The relationship
success
rates
terms
ofthe
contact
area
of the
glass
percentage).
Thisbetween
shows
highera
connected
to ainpoint
glass
asPDMS
shown
inthe
Figure
4. (in
(a)
The (in
relationship
transfer
transfer
success
rateson
in
terms
ofsubstrate,
contact
area
of theto
PDMS
to the
glass
percentage).
This ashows
success
rate
the of
contact
area
is of
smaller,
but but
the
success
rate
is is
decreased
asasthe
area
successsuccess
rateswhen
inrate
terms
contact
area
thesmaller,
PDMS
to the
glass
(inrate
percentage).
This
shows
a higher
higher
when
the
contact
area
is
the
success
decreased
thecontact
contact
area
increases.
As
the
contact
area
increases,
the
bonding
force
between
the
PDMS
and
the
glass
becomes
increases.
Aswhen
the contact
area increases,
the bonding
force
between
PDMS andasthe
glass
becomes
success rate
the contact
area is smaller,
but the
success
ratethe
is decreased
the
contact
area
stronger
and
leaves
large
residue.
(b)The
The
relationship
between
the
applied
forces
to
lift
PDMS
stronger
aa large
residue.
(b)
between
the
applied
forces
liftglass
thethe
PDMS
to
increases.and
Asleaves
the contact
area
increases,
therelationship
bonding
force
between
the
PDMS
and to
the
becomes
to
the
contact
area
ofathe
PDMS
to the
A higher
contact
area
requires
more
applied
to the
lift
the
contact
area
of the
PDMS
to the
glass.
A higher
contact
area
requires
more
applied
to lift
stronger
and
leaves
large
residue.
(b)glass.
The
relationship
between
the
applied
forces
toforce
liftforce
the
PDMS
the
PDMS
since
the
strength
isglass.
stronger
a higher
contact
area. more applied force to lift
PDMS
since
the
bonding
stronger
at higher
a at
higher
contact
area.
to the
contact
area
ofbonding
the strength
PDMS
toisthe
A
contact
area
requires
Transfer
Transfer
raterate
(%) (%)
the PDMS since the bonding strength is stronger at a higher contact area.
100
100
80
80
60
60
40
40
20
20
0
5:1
10:1
20:1
5:1
10:1
20:1
0 of base polymer and curing agent
Ratio
Ratio of base polymer and curing agent
Figure 6. Rate of microwell arrays transferred to the glass as a function of the ratio between the prepolymer
and the
curing agent.
A lower amount
of the
curing
agent in of
the PDMS
mixture
shows
a
Figure 6.
to to
the
glass
as aasfunction
between
the preFigure
6. Rate
Rateofofmicrowell
microwellarrays
arraystransferred
transferred
the
glass
a functionthe
of ratio
the ratio
between
the
better
transfer
rate.
Error
bars
are
the
range
between
the
maximum
value
and
minimum
value;
the
polymer andand
the the
curing
agent.
A lower
amount
of the
curing
agent
in in
thethe
PDMS
mixture
shows
pre-polymer
curing
agent.
A lower
amount
of the
curing
agent
PDMS
mixture
showsa
points
indicate rate.
average
values.
better
transfer
Error
bars
are
the
range
between
the
maximum
value
and
minimum
value;
the
a better transfer rate. Error bars are the range between the maximum value and minimum value;
points
indicate
average
values.
the
points
indicate
average
values.
Next, we characterized the uniformity in the height of the microwells. As shown in Figure 7, the
PDMS
patterns
from the stamp
after the transfer
andmicrowells.
the uniformity
of the in
pattern
Next,
we characterized
the uniformity
in the were
heightcut,
of the
As shown
Figureheight
7, the
Next,
we
characterized
the
uniformity
in
the
height
of
the
microwells.
As
shown
in Figurethe
7,
was
observed.
of theafter
transferred
patterns
the the
PDMS
stamp was
measured
PDMS
patternsUniformity
from the stamp
the transfer
werefrom
cut, and
uniformity
of the
pattern by
height
the
PDMS
patterns
from
the
stamp
after
the
transfer
were
cut,
and
the
uniformity
of
the
pattern
height
following
definition:
U = (H
−H
min)/Hmean ×
100, where
max PDMS
is the maximum
Hmin by
is the
was observed.
Uniformity
ofmaxthe
transferred
patterns
fromHthe
stamp washeight,
measured
the
was
observed.
Uniformity
of
the
transferred
patterns
from
the
PDMS
stamp
was
measured
by
the
minimum
and H
ismax
the−average
thewhere
height.HA
U value means
more
uniform
following height
definition:
Umean
= (H
Hmin)/Hvalue
mean × of
100,
maxsmaller
is the maximum
height,
Hmin
is the
followingThe
definition:
U = uniformity
(Hmax − Hmin
× 100,
Hmax is the
maximum
height, Hofminthe
is
mean
patterns.
calculated
of )/H
the
patterns
bywhere
this Adefinition
15.3%.
minimum height
and Hmean
is the average
value
of the height.
smaller Uwas
value
meansBecause
more uniform
the
minimum
height
and
H
is
the
average
value
of
the
height.
A
smaller
U
value
means
more
mean
flexible
of PDMS, the
height of of
thethe
transferred
not uniform
but this of
is still
patterns.nature
The calculated
uniformity
patterns PDMS
by thismicrowells
definitioniswas
15.3%. Because
the
uniform
patterns.
The
calculated
uniformity
of
the
patterns
by
this
definition
was
15.3%.
Because
of
enough
for trapping
cells the
in aheight
cell-to-cell
since the
height
of the cellisisnot
20 μm
and the
flexible nature
of PDMS,
of thestudy
transferred
PDMS
microwells
uniform
butwell
thisheight
is still
the
flexible
nature
of
PDMS,
the
height
of
the
transferred
PDMS
microwells
is
not
uniform
but
this
is
is
close to
(usually
higher
than 20 μm).
enough
forthis
trapping
cells
in a cell-to-cell
study since the height of the cell is 20 μm and the well height
still enough
for
trapping
cells
in
a
cell-to-cell
study
since
the
height
of
the
cell
is
20
µm
and
the
well
Weto
observed
the trapped
cells on
the bowtie microwell arrays as shown in Figure 8. The capture
is close
this (usually
higher than
20 μm).
height
is
close
to
this
(usually
higher
than
20 The
µm).cell-to-cell adhesion could be examined by staining
efficiency
for cell pairs
was 29%
on on
average.
We observed
the trapped
cells
the bowtie
microwell arrays as shown in Figure 8. The capture
cells
using
fluorescent
dye,
targeting
various
cadherin
molecules
control by
BM-MSCs
efficiency for cell pairs was 29% on average. The cell-to-cell
adhesion
couldthat
be examined
staining
differentiation
towards smooth
lineage.
Specifically,
studied levels
Cad-11 (cadherincells using fluorescent
dye, muscle
targeting
various
cadherinwemolecules
thatofcontrol
BM-MSCs
differentiation towards smooth muscle lineage. Specifically, we studied levels of Cad-11 (cadherin-
Micromachines 2016, 7, 173
7 of 9
We observed the trapped cells on the bowtie microwell arrays as shown in Figure 8. The capture
efficiency for cell pairs was 29% on average. The cell-to-cell adhesion could be examined by staining
cells using fluorescent
Micromachines
2016, 7, 173 dye, targeting various cadherin molecules that control BM-MSCs differentiation
7 of 9
Micromachines 2016, 7, 173
7 of 9
towards smooth muscle lineage. Specifically, we studied levels of Cad-11 (cadherin-11) and its effect on
11)
and its effect
on smooth
muscle genes
αSMA
(α-smooth
muscle actin),
CNN-1 (calponin)
and
smooth
genes
αSMA (α-smooth
muscle
actin),
CNN-1 (calponin)
and MYH11
(myosin heavy
11) and muscle
its effect
on smooth
muscle genes
αSMA
(α-smooth
muscle actin),
CNN-1 (calponin)
and
MYH11
(myosin
heavy
chain).
Moreover,
this
tool
is
expected
to
be
used
in
further
studies
of
three
chain).
this tool
is expected
to be this
usedtool
in further
studies
neighboring
in
MYH11Moreover,
(myosin heavy
chain).
Moreover,
is expected
to of
bethree
usedor
infour
further
studies ofcells
three
or
four neighboring
cells
in microwells;
therefore,
it may help
to investigate
biological
questions such
microwells;
therefore,
it
may
help
to
investigate
biological
questions
such
as
whether
cadherin-11
is
a
or four neighboring cells in microwells; therefore, it may help to investigate biological questions such
as
whether
cadherin-11
is
a
master
regulator
of
BM-MSCs
for
smooth
muscle
differentiation.
master
regulator
of BM-MSCs
for smooth
muscle
differentiation.
as whether
cadherin-11
is a master
regulator
of BM-MSCs
for smooth muscle differentiation.
Figure
7. Image
on the
the
and not
transferred
Figure
Image of
of microwell
microwell array
array
the PDMS
PDMS stamp,
stamp, with
with
the transferred
transferred part
part
Figure 7.
7. Image
of
microwell
array on
on the
PDMS
stamp,
with the
transferred
part and
and not
not transferred
transferred
part.
(a)
Cross-section
view
of
the
microwell
array
on
the
PDMS
stamp.
After
the
transfer,
the PDMS
PDMS
part.
part. (a)
(a) Cross-section
Cross-section view
view of
of the
the microwell
microwell array
array on
on the
the PDMS
PDMS stamp.
stamp. After
After the
the transfer,
transfer, the
the PDMS
stamp
is
cut
to
calculate
the
uniformity
of
the
height.
(b)
Top
view
of
the
microwell
array.
The
left
stamp
ofof
thethe
height.
(b)(b)
TopTop
view
of the
microwell
array.
The left
stamp is
is cut
cuttotocalculate
calculatethe
theuniformity
uniformity
height.
view
of the
microwell
array.
Thepart
left
part
the microwell
that
not transferred
the PDMS
microwell
that
the
is
theis
arrayarray
that is
notis
to theto
microwell
array,array,
whichwhich
meansmeans
that the
array
part
ismicrowell
the microwell
array
that
istransferred
not transferred
toPDMS
the PDMS
microwell
array,
which
means
that
the
array
remained
on
the PDMS
stamp
andright
the right
part
istransferred
the transferred
microwell
array,
which
means
remained
on
the
PDMS
stamp
and
the
part
is
the
microwell
array,
which
means
the
array remained on the PDMS stamp and the right part is the transferred microwell array, which means
the
array
removed
the stamp
by being
peeled
array
waswas
removed
fromfrom
the stamp
by being
peeled
off. off.
the array was removed from the stamp by being peeled off.
Figure 8.
Image of
of the
Cells are
are loaded
Figure
8. Image
the trapped
trapped cells
cells on
on the
the bowtie-shaped
bowtie-shaped microwell
microwell arrays.
arrays. Cells
loaded into
into
Figure 8. Image
ofone
the pair
trapped
cells
on the bowtie-shaped
microwell
arrays.
Cells are
loaded into
microwells
so
that
of
cells
is
cultured
within
the
microwell
arrays,
resulting
in
contacting
microwells so that one pair of cells is cultured within the microwell arrays, resulting in contacting aa
microwells
so that one The
paircell-to-cell
of cells is cultured
within
the microwell
arrays, resulting
in contacting
a
single
neighboring
adhesion
can be
examined
by immunostaining
various
cadherin
single
neighboringcell.
cell. The
cell-to-cell
adhesion
can
be examined
by immunostaining
various
single
neighboring
cell.
The
cell-to-cell
adhesion
can
be
examined
by
immunostaining
various
moleculesmolecules
which control
(bone marrow
stem cells)
differentiation
towards
cadherin
whichBM-MSCs’
control BM-MSCs’
(bone mesenchymal
marrow mesenchymal
stem
cells) differentiation
cadherinmuscle
molecules
whichThe
control
BM-MSCs’
(bone
marrow
mesenchymal
stem cells)
differentiation
smooth
lineage.
number
of
cells
trapped
within
the
bowtie-shaped
microwell
arrays
towards smooth muscle lineage. The number of cells trapped within the bowtie-shaped microwell
towards
smooth muscle lineage. The number of cells trapped within the bowtie-shaped microwell
was
measured.
arrays was measured.
arrays was measured.
Another possible application of the method is the fabrication of glass-PDMS-glass microfluidic
Another possible application of the method is the fabrication of glass-PDMS-glass microfluidic
devices. Since glass is a good substrate for studying cells, it is thus advisable to have a microfluidic
devices. Since glass is a good substrate for studying cells, it is thus advisable to have a microfluidic
device made by glass only, but it requires extensive work and expensive equipment. Therefore,
device made by glass only, but it requires extensive work and expensive equipment. Therefore,
instead of using only glass, a device made by top and bottom glass slides sandwiching a PDMS
instead of using only glass, a device made by top and bottom glass slides sandwiching a PDMS
microfluidic channel layer (glass-PDMS-glass configuration) is easier and cheaper, and also is a good
Micromachines 2016, 7, 173
8 of 9
Another possible application of the method is the fabrication of glass-PDMS-glass microfluidic
devices. Since glass is a good substrate for studying cells, it is thus advisable to have a microfluidic
device made by glass only, but it requires extensive work and expensive equipment. Therefore, instead
of using only glass, a device made by top and bottom glass slides sandwiching a PDMS microfluidic
channel layer (glass-PDMS-glass configuration) is easier and cheaper, and also is a good environment
for microfluidic cell studies. For the fabrication of the glass-PDMS-glass configuration device, instead
of transferring microwell arrays, we transfer the microfluidic channel pattern on a glass substrate and
seal the channel with another glass substrate on top of the transferred microfluidic channel pattern.
Proving the reliability of the proposed method in the glass-PDMS-glass configuration for studying
cells in microfluidic systems requires good experimental results [15].
4. Conclusions
We have demonstrated the simple fabrication method of the microwell arrays on the glass
substrate using the conventional photolithography technique. For cell-to-cell adhesion to smooth
muscle cells, four different types of patterns can be successfully created on the glass substrate by tearing
the patterns from the PDMS stamp to the glass substrate. The pattern transfer rate was investigated as
the function of the contact area of the PDMS stamp to the glass substrate and the ratios between the
pre-polymer and the curing agent. The maximum transfer rate of 76% was achieved at the ratio of 10:1
as the contact area decreased. At the ratio of 20:1, the transfer rate increased to 85%.
Acknowledgments: This work was supported by the National Science Foundation grants (Grant Nos.
NSF-CBET-1403086 and NSF-CBET-1337860).
Author Contributions: H.L, D.K., L.X., and K.W.O. had developed the proposed fabrication process and
performed experiments to prove its reliability. S.R. and S.T.A. had performed cell-to-cell experiment using
the PDMS microwell fabricated by the proposed method.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Goral, V.N.; Zhou, C.F.; Lai, F.; Yuen, P.K. A continuous perfusion microplate for cell culture dagger. Lab Chip
2013, 13, 1039–1043. [CrossRef] [PubMed]
Charnley, M.; Textor, M.; Khademhosseini, A.; Lutolf, M.P. Integration column: Microwell arrays for
mammalian cell culture. Integr. Biol. 2009, 1, 625–634. [CrossRef] [PubMed]
Jain, T.; McBride, R.; Head, S.; Saez, E. Highly parallel introduction of nucleic acids into mammalian cells
grown in microwell arrays. Lab Chip 2009, 9, 3557–3566. [CrossRef] [PubMed]
Choi, J.H.; Lee, H.; Jin, H.K.; Bae, J.S.; Kim, G.M. Micropatterning of neural stem cells and Purkinje neurons
using a polydimethylsiloxane (PDMS) stencil. Lab Chip 2012, 12, 5045–5050. [CrossRef] [PubMed]
Nelson, C.M.; Liu, W.F.; Chen, C.S. Manipulation of cell-cell adhesion using bowtie-shaped microwells.
Methods Mol. Biol. 2007, 370, 1–9. [PubMed]
Johnson, L.A.; Rodansky, E.S.; Haak, A.J.; Larsen, S.D.; Neubig, R.R.; Higgins, P.D. Novel Rho/MRTF/SRF
inhibitors block matrix-stiffness and TGF-β-induced fibrogenesis in human colonic myofibroblasts.
Inflamm. Bowel Dis. 2014, 20, 154–165. [CrossRef] [PubMed]
Charest, J.L.; Jennings, J.M.; King, W.P.; Kowalczyk, A.P.; Garcia, A.J. Cadherin-mediated cell-cell contact
regulates keratinocyte differentiation. J. Investig. Dermatol. 2009, 129, 564–572. [CrossRef] [PubMed]
Tamura, T.; Sakai, Y.; Nakazawa, K. Two-dimensional microarray of HepG2 spheroids using
collagen/polyethylene glycol micropatterned chip. J. Mater. Sci. Mater. Med. 2008, 19, 2071–2077. [CrossRef]
[PubMed]
Schmid, H.; Wolf, H.; Allenspach, R.; Riel, H.; Karg, S.; Michel, B.; Delamarche, E. Preparation of metallic films
on elastomeric stamps and their application for contact processing and contact printing. Adv. Funct. Mater.
2003, 13, 145–153. [CrossRef]
Gray, D.S.; Liu, W.F.; Shen, C.J.; Bhadriraju, K.; Nelson, C.M.; Chen, C.S. Engineering amount of cell-cell
contact demonstrates biphasic proliferative regulation through RhoA and the actin cytoskeleton. Exp. Cell Res.
2008, 314, 2846–2854. [CrossRef] [PubMed]
Micromachines 2016, 7, 173
11.
12.
13.
14.
15.
9 of 9
Nelson, C.M.; Chen, C.S. VE-cadherin simultaneously stimulates and inhibits cell proliferation by altering
cytoskeletal structure and tension. J. Cell Sci. 2003, 116, 3571–3581. [CrossRef] [PubMed]
Lee, H.; Lee, K.; Ahn, B.; Oh, K.W. A new fabrication process for SU-8 thick photoresist structures by
simultaneously removing edge-bead and air bubbles. J. Micromech. Microeng. 2011, 21, 125006. [CrossRef]
Kim, M.; Huang, Y.; Choi, K.; Hidrovo, C.H. The improved resistance of PDMS to pressure-induced
deformation and chemical solvent swelling for microfluidic devices. Microelectron. Eng. 2014, 124, 66–75.
[CrossRef]
Johnston, I.D.; McCluskey, D.K.; Tan, C.K.L.; Tracey, M.C. Mechanical characterization of bulk Sylgard 184 for
microfluidics and microengineering. J. Micromech. Microeng. 2014, 24, 035017. [CrossRef]
Mojica, W.D.; Oh, K.W.; Lee, H.; Furlani, E.P.; Sykes, D.; Sands, A.M. Microfluidics enables multiplex
evaluation of the same cells for further studies. Cytopathology 2015, 27, 277–283. [CrossRef] [PubMed]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).