FOCUS
www.rsc.org/loc | Lab on a Chip
PDMS absorption of small molecules and
consequences in microfluidic applications
Michael W. Toepke and David J. Beebe*
DOI: 10.1039/b612140c
Microfluidic devices made out of polydimethylsiloxane (PDMS) have many
physical properties that are useful for cell culture applications, such as
transparency and gas permeability. Another distinct characteristic of PDMS is its
ability to absorb hydrophobic small molecules. Partitioning of molecules into
PDMS can significantly change solution concentrations and could potentially alter
experimental outcomes. Herein we discuss PDMS absorption and its potential
impact on microfluidic experiments.
Microfluidic devices have found use in
cell culture applications due to the
unique physical conditions that can be
achieved within small channels. Precise
patterning of cells,1 control over the
cellular and subcellular microenvironments,1,2 and gentler cell handling3 have
all been demonstrated using microfluidic
devices. Ease of patterning via soft
lithography,4 facile sealing, and optical
transparency have led a large number of
researchers to use polydimethylsiloxane
(PDMS) to fabricate their microfluidic
devices.
The network polymer structure of
PDMS makes it highly permeable relative to other materials, a property that
can be put to use in cell culture applications to supply oxygen and remove
carbon dioxide. Diffusion of water vapor
through PDMS has also been observed,
which may be useful for some applications,5 but could lead to unwanted drying
and convective flow within sample
volumes. The porous nature of PDMS
also enables small molecules to diffuse
into the bulk polymer, which has allowed
PDMS to be used in microextraction
applications for removing trace organic
compounds from aqueous samples.6,7
Silicone implanted into dogs was found
to absorb the equivalent of 0.7% of its
weight in small molecules, with lipids
comprising approximately two-thirds of
the absorbed mass.8 Similar uptake of
small molecules by PDMS may affect the
outcome of microfluidic experiments in
Department of Biomedical Engineering,
University of Wisconsin—Madison, Madison,
WI, 53706, USA. E-mail: [email protected]
1484 | Lab Chip, 2006, 6, 1484–1486
drug discovery, proteomic analysis, and
cell culture applications where species of
interest are present in micro- and nanomolar concentrations.
Fig. 1 shows fluorescence images of
a PDMS microfluidic channel after a
series of filling steps with Nile red (MW
318 g mol21), a hydrophobic fluorophore. For the experiment, the channel
was filled with 1 mM Nile red and
incubated for one minute before the
solution was aspirated out of the channel
from the exit. Next, the channel was
loaded with DI water, which was allowed
to stand for one minute and then
aspirated in preparation for the next
sample volume. A fluorescence image of
the channel after one filling with Nile red
is shown in Fig. 1(a). The majority of the
fluorescence appears as a halo around
the inlet port of the channel, with
essentially no signal from the main
channel. The channel was filled in
approximately one second, implying that
absorption can occur rapidly given that
little of the fluorophore reached the
main channel. Figs. 1(b)–(e) show the
fluorescence in the channel after several
Fig. 1 Fluorescence images of Nile red absorbed into a PDMS channel after (a) one, (b) two, (c)
six, (d) ten, (e) twenty fills, and (f) rinsing the channel with Alconox and DI water after the
twentieth filling. (g) Cross-section and (h) substrate of PDMS channel used in (f). (i) A
polystyrene substrate used with a PDMS channel of the same geometry and exposed to twenty
fillings with Nile red and no rinse steps.
This journal is ß The Royal Society of Chemistry 2006
Fig. 2 Fluorescence images of a PDMS channel (a) filled with pH 2 quinine solution, (b) after
rinsing with pH 2 water, (c) incubated with pH 7 quinine solution for five minutes, (d) 10 s after
replacing the solution in the channel with pH 2 water, (e) five minutes after the addition of pH 2
water, and (f) rinsing the channel with pH 2 water.
subsequent filling steps (two, six, ten, and
twenty). The fluorescence from the channel increased with each loading, indicating the accumulation of fluorophore. The
same channel can be seen in Fig. 1(f)
after rinsing with 200 channel volumes
of an aqueous Alconox solution and
200 channel volumes of DI water; the
negligible change in the fluorescence is
indicative of the strong interaction of the
dye with the polymer. A cross-section of
the same PDMS device after the rinsing
sequence (Fig. 1(g)) shows bright fluorescent signal more than 100 mm into the
walls of the PDMS channel, revealing
that the dye interaction with the PDMS
is not limited to the surface of the
channel. Fig. 1(h) and (i) show the top
view of PDMS and polystyrene substrates that were exposed to 20 channel
fillings with Nile red. The PDMS substrate shows a fluorescent signal after
having been rinsed with Alconox and
water. Although it was not rinsed after
exposure to Nile red, the polystyrene
substrate showed no visible fluorescence.
The absorption of small molecules into
PDMS is an equilibrium process, analogous to liquid–liquid extraction, and
therefore can be characterized by partition coefficients. Octanol/water partition
coefficients are often used to predict
PDMS absorption of small molecules
for
microextraction
applications.9
Partition coefficients and kinetics for
PDMS absorption of several pharmaceutical compounds, including ciprofloxacin
and paclitaxel, have already been
reported, and their values can vary
greatly depending on factors such as pH
and the ions in solution.7 The importance
of pH and counterion pairing can make
predicting PDMS absorption of small
molecules difficult, especially for complex mixtures such as cell culture media.
The pH dependence of small molecule
absorption is demonstrated in Fig. 2 with
quinine (MW 324 g mol21), a drug that
fluoresces at pH 2 but not pH 7. Fig. 2(a)
shows the fluorescence image of a microfluidic channel filled with a pH 2 quinine
solution (MW 324 g mol21). The fluorescent signal disappeared upon rinsing
with pH 2 solution (Fig. 2(b)), indicating
that little or no quinine was absorbed by
the PDMS. Next, the channel was filled
with a pH 7 quinine solution (Fig. 2(c))
and allowed to incubate for five minutes
at room temperature. The channel was
subsequently rinsed with pH 7 water and
then filled with pH 2 water. A faint
fluorescent signal was initially seen after
the pH was changed from 7 to 2
(Fig. 2(d)) and the fluorescent intensity
continued to increase over time until
reaching equilibrium after five minutes
(Fig. 2(e)). Finally, the channel was
rinsed with pH 2 solution (Fig. 2(f))
and the fluorescent signal was lost,
showing that essentially all of the quinine
had repartitioned back into solution
from the PDMS. The reversible nature
of absorption shows that it is important
to rinse channels out with medium prior
to cell seeding if channels are preconditioned with any chemical treatment.
Absorbance of small molecules into
PDMS microfluidic devices may have a
This journal is ß The Royal Society of Chemistry 2006
profound effect on the outcome of drug
screening studies, particularly when fixed
volumes of drug are used to determine
dose-response. Consider an assay performed with a hydrophobic pharmaceutical compound, such as quinine, in a
PDMS channel that is 250 mm wide and
100 mm tall. A partition coefficient of 26
has previously been reported for microextraction of quinine from phosphate
buffer solution (PBS) using PDMS
fibers.7 If one were to try to hold the
quinine concentration within the channel
at 2 mM, the 100 mm-thick layer of
PDMS surrounding the channel would
have to absorb the equivalent of
100 channel volumes of 2 mM quinine.
One channel volume of solution would
contain 81 pg of quinine and the
surrounding PDMS would be capable
of absorbing 2 mg assuming 0.4 wt%
small molecule uptake, indicating that
the PDMS would have sufficient absorbance capacity. If the channel were only
filled with 2 mM quinine solution one
time, the solution concentration would
be on the order of 20 nM, or 100 times
lower than the intended concentration.
Such a discrepancy would significantly
alter the results of drug discovery
screens, as apparent binding constants
of compounds that interact strongly with
PDMS would be significantly reduced.
Further reducing channel volumes
with the aim of decreasing sample consumption will only further exacerbate the
problem by increasing the surface area to
volume ratio; for example, reducing the
channel width to 125 mm would require
the absorption of the equivalent of
160 channel volumes to achieve equilibrium between the PDMS and the
solution in the channel. Paclitaxel, with
a partition coefficient of 51, would need
almost twice the excess required for
quinine. While partitioning could be
useful for sample collection/purification,
such large excesses may negate any
sample conservation that would be
expected from performing assays in
microfluidic devices. PDMS absorption,
in addition to adsorption,10 has the
potential to alter solution concentrations
of cell culture components such as
certain essential amino acids, vitamins,
steroids, hormones, neurotransmitters,
eicosanoids, and growth factors. The loss
of such molecules could lead to reduced
growth or alter phenotypic behavior in
Lab Chip, 2006, 6, 1484–1486 | 1485
cell culture, which might then be incorrectly attributed to other factors.
Some steps can be taken to modify
PDMS channels so that adsorption and
absorption are reduced. Low molecular
weight silicone molecules have been
removed from bulk PDMS through
aging11 and extraction,12 which reduced
hydrophobic recovery after plasma treatment and increased solvent resistance.
Excess crosslinking agent13 and zeolite
filler have been incorporated into PDMS
in order to reduce water evaporation.14
Unwanted adsorption of molecules onto
the surface of PDMS can be reduced by
first adsorbing BSA15 or by chemically
modifying the PDMS surface with a
silane.16 A method to coat the surface
of PDMS channels with a thin polymer
film by absorbing a photoinitiator into
PDMS has been developed, but the
technique has yet to be optimized for
preventing absorption.17
Ultimately, alternative fabrication
materials or coatings may be required
for microfluidic cell culture applications.
A variety of different materials have
already been developed for microfluidic
applications;18 however, finding a material that meets the diverse set of requirements needed for cell culture could be
difficult. Optical transparency, gas permeability, minimal investment in new
equipment, and resistance to absorption/
adsorption are just some of the desirable
traits for new materials. A fluorinated
polymer with properties similar to
PDMS, but likely with fewer solute
interactions, has been developed.19
Alternatively, microfluidic devices made
out of materials that more closely mimic
in vivo conditions are also being developed.20 While other materials may eventually be adopted, polystyrene has long
been the material of choice for cell
culture applications; however, prototyping polystyrene devices is both more
1486 | Lab Chip, 2006, 6, 1484–1486
difficult and expensive than soft lithography-based procedures. Therefore,
PDMS can still play a key role in
microfluidic device prototyping and
characterization, but absorption should
be taken into consideration because the
composition of solutions may be significantly altered.
Acknowledgements
This work was supported by funding
from the Army Breast Cancer Research
Program
(W81XWH-04-1-0572),
National Institutes of Health (K25
CA104162), and NHGRI training grant
5T32HG002760 (M.W.T.). We thank
Greg Cox at Molecular Probes
Invitrogen Detection Technologies for
valuable collaboration.
References
1 S. Takayama, J. C. McDonald, E. Ostuni,
M. N. Liang, P. J. A. Kenis, R. F.
Ismagilov and G. M. Whitesides, Proc.
Natl. Acad. Sci. U. S. A., 1999, 96,
5545–5548.
2 S. Takayama, E. Ostuni, P. LeDuc and
G. M. Whitesides, Nature, 2001, 411, 1018;
E. M. Lucchetta, J. H. Lee, L. A. Fu,
N. H. Patel and R. F. Ismagilov, Nature,
2005, 434, 1134–1138.
3 H. C. Zeringue, J. J. Rutledge and
D. J. Beebe, Lab Chip, 2005, 5, 86–90.
4 Y. Xia and G. M. Whitesides, Angew.
Chem., Int. Ed., 1998, 37, 550–575;
D. C. Duffy, J. C. McDonald, O. J. A.
Schueller and G. M. Whitesides, Anal.
Chem., 1998, 70, 4974–4984; B.-H. Jo,
L. M. Van Lerberghe, K. M. Motsegood
and D. J. Beebe, J. Microelectromech.
Syst., 2000, 9, 76–81.
5 G. C. Randall and P. S. Doyle, Proc. Natl.
Ac ad. Sc i. U. S. A., 200 5, 1 02,
10813–10818; E. Verneuil, A. Buguin and
P. Silberzan, Europhys. Lett., 2004, 68,
412–418; C. L. Hansen, S. Classen,
J. M. Berger and S. R. Quake, J. Am.
Chem. Soc., 2006, 128, 3142–3143.
6 E. Baltussen, P. Sandra, F. David and
C. Cramers, J. Microcolumn Sep., 1999,
11, 737–747.
7 G. Theodoridis, M. Aikaterini Lontou,
F. Michopoulos, M. Sucha and
T. Gondova, Anal. Chem. Acta, 2004,
516, 197–204.
8 J. W. Swanson and J. E. Lebeau,
J. Biomed. Mat. Res., 1974, 8, 357–367.
9 B. Shurmer and J. Pawliszyn, Anal. Chem.,
2000, 72, 3660–3664.
10 J. E. Butler, E. P. Lu, P. Navarro and
B. Christiansen, J. Mol. Recognit., 1997,
10, 36–51; J. E. Butler, P. Navarro and
J. Sun, Immunol. Invest., 1997, 26, 39–54;
G. Ocvirk, M. Munroe, T. Tang,
R. Oleschuk, K. Westra and D. J.
Harrison, Electrophoresis, 2000, 21,
107–115.
11 D. T. Eddington, J. P. Puccinelli and
D. J. Beebe, Sens. Actuators, B, 2006,
B114, 170–172.
12 J. N. Lee, C. Park and G. M. Whitesides,
Anal. Chem., 2003, 75, 6544–6554.
13 W.-J. Chang, D. Akin, M. Sedlak,
M. R. Ladisch and R. Bashir, Biomed.
Microdev., 2003, 5, 281–290.
14 M. Bennett, B. J. Brisdon, R. England and
R. W. Field, J. Membr. Sci., 1997, 137,
63–88.
15 E. Ostuni, C. S. Chen, D. E. Ingber and
G. M. Whitesides, Langmuir, 2001, 17,
2828–2834.
16 S. Jon, J. Seong, A. Khademhosseini,
T.-N. T. Tran, P. E. Laibinis and
R. Langer, Langmuir, 2003, 19, 9989–9993.
17 S. Hu, X. Ren, M. Bachman, C. E. Sims,
G. P. Li and N. L. Allbritton, Anal. Chem.,
2004, 76, 1865–1870.
18 C. H. Ahn, J. W. Choi, G. Beaucage,
J. H. Nevin, J.-B. Lee, A. Puntambekar
and J. Y. Lee, Proc. IEEE, 2004,
92, 154–173; J. de Jong, B. Ankoné,
R. G. H. Lammertink and M. Wessling,
Lab Chip, 2005, 5, 1240–1247; A. de Mello,
Lab Chip, 2002, 2, 31N–36N; H. Becker
and L. E. Locascio, Talanta, 2002, 56,
267–287; T. D. Boone, Z. H. Fan,
H. H. Hooper, A. J. Ricco, H. Tan and
S. J. Williams, Anal. Chem., 2002, 74,
78A–86A.
19 J. P. Rolland, R. M. Van Dam,
D. A. Schorzman, S. R. Quake and
J. M. DeSimone, J. Am. Chem. Soc.,
2004, 126, 2322–2323; Y. Huang,
P. Castrataro, C.-C. Lee and S. R.
Quake, Lab Chip, submitted.
20 M. Cabodi, N. W. Choi, J. P. Gleghorn,
C. S. D. Lee, L. J. Bonassar and
A. D. Stroock, J. Am. Chem. Soc., 2005,
127, 13788–13789; A. Paguirigan and
D. J. Beebe, Lab Chip, 2006, 6, 407–413.
This journal is ß The Royal Society of Chemistry 2006