of 3
Supporting Information:
In order to test nanofabricated array structures in a microfluidic channel, a
prototype flow system was assembled
1
. A 75 mm diameter silicon wafer was used as a
master for molding a polydimethyl siloxane (PDMS) channel. The silicon was thermally
oxidized, and photolithographically patterned using Shipley 1813 photoresist purchased
from MicroChem (Newton MA, USA) and a transparency mask purchased from Digital
PageWorks (Cambridge MA, USA). The test pattern simply consisted of two 2 mm
square wells connected by a 200
μ
m wide channel. This pattern was transferred into the
thermal oxide using buffered hydrofluoric acid purchased from J. T. Baker (Phillipsburg
NJ, USA). The photoresist was removed using a Caros acid solution mixed from
hydrogen peroxide and sulfuric acid (J. T. Baker). The silicon was then bulk etched
using potassium hydroxide (J. T. Baker) forming a 10
μ
m ridge for the microfluidic
channel mold. Dow Corning Sylgard 184 Polydimethlsiloxane (PDMS) was mixed and
deaired according to manufacturer specifications
2
and poured over the bulk
micromachined silicon master with hollow cylindrical glass wells placed at the extents of
the silicon ridge. The well walls are approximately 1 mm thick, 5 mm in diameter, and
20 mm in height. The entire assembly was cured at 150
o
C for 1 hour, and then the PDMS
was peeled away from the silicon master with the glass wells intact. This technique
resulted in a PDMS structure approximately 2-5 mm thick with a 10
μ
m deep channel
between the two wells. This molded PDMS microchannel was then placed over dimple
arrays fabricated on silicon wafers for fluid flow experiments.
The room temperature fluidic sealing of PDMS to silicon, and the simple method
of removal and reapplication of the channel to silicon made alignment of the 200
μ
m
wide channel to the 10
μ
m wide dimple arrays a matter of trial and error. The
nanostructured arrays were easily observed in an optical microscope and it was noted that
only a minor change in contrast was observed between arrays observed in air and arrays
observed with PDMS channels over them, which was attributed to the nearly ideal
transparency of the channel material. Wells were filled using a variety of liquids, some
of which filled the channel through capillary action, and some of which required
pneumatic actuation to fill the channel. Fluids tested included air, water, ethyl alcohol,
methyl alcohol, isopropyl alcohol (alcohols obtained from J. T. Baker), and index
matched fluids (Series A and AA, obtained from Cargille Labs, Cedar Grove NJ, USA).
These fluids provided a wide range of refractive index as well as wetting properties. The
channel and arrays were typically rinsed with solvent and dried in nitrogen between fluid
flow experiments.
Figure 1a-c shows optical micrographs of dimple arrays under a PDMS channel
with a 2-propanol/water interface over the top of the arrays. In 1a, the dimple arrays are
covered by a channel filled with air, in 1b the water covers half of the arrays, and in 1c
the arrays are completely covered in water. The entire prototype flow system is depicted
schematically in 1d. One array in particular shows a striking change from blue to orange
and this change is captured with the alcohol/air
interface covering half of this array in 1b.
For this setup the position of the air/water interface was controlled pneumatically, and the
arrays were able to switch back and forth between the two colors easily.
After the air/liquid demonstration shown in Figure 1, a row of arrays was
fabricated using parameters such that a striking change (between orange and blue) would
result when using the air/water system. These arrays were then placed under the fluidic
channel described above with the row of arrays positioned lengthwise along the channel
(rather then vertically across the channel, as depicted in Figure 1). Video 1 shows this
row of dimple arrays undergoing several changes between air and water. Acquisition rate
for digital video was set at about 100 milliseconds, or 10 frames per second. Each dimple
array is 10
μ
m on a side.
Due to the similar refractive indices of most liquids, a difference in color was not
easily observable for changes between water, ethanol, methanol, and 2-propanol, but
significant changes were easily observed for fluid–to–fluid changes using index matched
fluids. The color of the arrays used in Video 1 changed from orange to red when the
Cargille Series A fluid replaced the water in the channel. Video 2 illustrates a
demonstration of liquid–to–liquid operation of the channel. In this case a clear interface
exists between water and Cargille AA fluid, so no mixing of the water and Cargille AA
fluid occurs during the color change. Use of
these arrays as a detection mechanism is
more obvious in Video 3. The fluid–to–fluid interface is diffuse when using a Cargille
AA–to–Cargille A system because these fluids do mix when flowing through the
microchannel. A change in color for all fields shown indicates that Cargille A has
replaced the Cargille AA fluid in the arrays even though an interface is not clearly
observable. The color change
is such that each field approximately assumes the color of
the field to the right of it after about one minute. The change is complete after about 2
PDMS
Si
Outlet
Well
Inlet
Well
Dimple Arrays
Air/Water Interface
(d)
(a)
(b)
(c)
Figure 1a-c shows color change in dimple arrays of varying depth with an air/2-
propanol
interface. Each dimple array is 10
μ
m on a side. Figure 1d shows a schematic plan view
of the entire fluid flow prototype including inlet and outlet wells, PDMS with molded
microfluidic channel, and silicon with milled dimple arrays.
minutes. The only other indication of flow in the channel is a particle that passes just
below the fields (from left to right) in the final seconds.
The time for the arrays to switch back to their original color was observed to vary
when the arrays changed from water to air. We attribute this variation in switching time
to buildup of water vapor in the channel. Even though PDMS is highly hydrophobic,
over time the channels accumulated particulate contaminants that we believe contributed
to moisture buildup in the channel. Also, silicon with native oxide is hydrophilic, which
further contributes to moisture buildup. Due to these effects, the variation in time for
switching colors from water to air is attributable to varying conditions related to thermal
evaporation.
The length of time required for the arrays to change color was not constant for
different liquids. For all cases where a liquid / air interface was used, the arrays changed
color instantaneously (to the resolution of this technique) when switching from air to
liquid. When switching from liquid back to air, the switching time ranged from nearly
instantaneous to several days. When switching from water or 2-propanol to air, the time
was always less than one minute. Using methanol, the switching time was observed to go
on for very long periods of time even after the PDMS channel was removed from the
silicon; in fact, the only ways to get the arrays to switch back to their original color were
to either raise the temperature to 90
o
C for several seconds, or replace the methanol with
another liquid with a thorough rinsing. We believe that this variation in time is due to
surface energy differences between the liquids in the arrays and not due to thermal
evaporation. However, since elevated temperature was able to induce the change in all
cases, switching may be accomplished using a thermally resistive heater integrated
underneath the optical plane.
References:
1. Duffy, D.C., McDonald, J.C., Schueller, O.J.A., and Whitesides, G.M., “Rapid
Prototyping of Microfluidic Systems in Poly(dimethylsiloxane)”,
Anal. Chem.
,
70
, 4974,
(1998).
2. Product Information, SYLGARD® 184 Silicone Elastomer, Dow Corning, 1998.