View Article Online / Journal Homepage / Table of Contents for this issue

PAPER

www.rsc.org/loc | Lab on a Chip

A novel high aspect ratio microfluidic design to provide a stable and uniform
microenvironment for cell growth in a high throughput mammalian cell
culture array
Paul J. Hung,{ Philip J. Lee,{ Poorya Sabounchi, Nima Aghdam, Robert Lin and Luke P. Lee

Published on 02 November 2004. Downloaded on 03/06/2015 07:11:41.

Received 14th July 2004, Accepted 17th September 2004

First published as an Advance Article on the web 2nd November 2004
DOI: 10.1039/b410743h
We present a high aspect ratio microfluidic device for culturing cells inside an array of
microchambers with continuous perfusion of medium. The device was designed to provide a
potential tool for cost-effective and automated cell culture. The single unit of the array consists of
a circular microfluidic chamber 40 mm in height surrounded by multiple narrow perfusion
channels 2 mm in height. The high aspect ratio (y20) between the microchamber and the
perfusion channels offers advantages such as localization of the cells inside the microchamber as
well as creating a uniform microenvironment for cell growth. Finite element methods were used to
simulate flow profile and mass transfer of the device. Human carcinoma (HeLa) cells were
cultured inside the device with continuous perfusion of medium at 37 uC and was grown to
confluency. The microfluidic cell culture array could potentially offer an affordable platform for a
wide range of applications in high throughput cell-based screening, bioinformatics, synthetic
biology, quantitative cell biology, and systems biology.

A. Introduction
Mammalian cell culture has played a major role in the
development of biotechnology, with applications such as
screening novel drugs to large scale production of protein
pharmaceuticals. Advances have been made in designing
automated cell culture systems capable of improving throughput and reducing process cost.1,2 However, the high cost and
large size of these systems limits application to large, dedicated
research facilities. It is therefore highly desirable to produce a
miniaturized, inexpensive platform for cell-culture-on-a-chip
to make high throughput cell experiments accessible in
laboratory, clinical, and field settings. The advantages of
microfabrication and microfluidic technologies are ideally
suited to address this issue.3 Since living cells are themselves
microscale systems, it follows that microfabricated devices
can more precisely control the culture environment. The
design of a microscale cell culture device presents a number
of unique challenges, many of which have been recently
discussed.4 Previous demonstrations of cell culture on
microfabricated devices include growth of hepatocytes,5,6
lung cells,7 and insect cells8 in both silicon and PDMS
substrates. These works validated the biocompatibility,
nutrient supply, and growth characteristics of cells within
microfabricated devices, but have yet to realize a high
throughput array that can replicate the main functionalities
of traditional cell culture.
Our current work addresses the development of a novel high
aspect ratio microfluidic cell culture array capable of providing
a stable and uniform microenvironment for cell growth. With
heterogeneous integration of a temperature control unit such
{ These authors contributed equally to this work.

44 | Lab Chip, 2005, 5, 4448

as an ITO heater, the device could potentially realize an

automated cost-effective cell culture platform without the large
robotic systems adapted by current practices. The microfluidic
device was designed to replicate the major processes in
traditional cell culture, making it adaptable to a large number
of applications. In this paper, 1 6 5 arrays were used for
device characterization to decrease the complexity of data
processing and time of optical monitoring. A 10 6 10 array
was also fabricated for increased throughput and initial
characterization of fluid flow through multiplexed arrays.

B. Device design and fabrication

The microfluidic cell culture system is designed to replicate the
major processes applied in standard eukaryotic culture
techniques. The current device is capable of providing a closed
environment for cell growth and manipulation without the
need of an incubator (Fig. 1a). The bonded PDMS culture
chambers maintain a sterile environment while permitting gas
exchange with the atmosphere. Temperature can be controlled
with a transparent ITO heater without hindering monitoring
of cells via optical microscopy. While humidity control of the
surrounding air would prevent evaporation from the microchannels, we observed that the continuous perfusion of
medium was sufficient to prevent the device from drying out.
A preliminary study of the flow profile inside a 10 6 10 array
was conducted to test the feasibility of splitting and
recombining flows in multiplexed arrays (Fig. 1b).
The single unit of the array consists of a circular
microfluidic chamber, a set of perfusion channels surrounding
the main chamber, and four ports for fluidic access to the
chamber. The microchambers are designed to have the same
cell growth area as a typical well in a 1,536 well microtiter
This journal is The Royal Society of Chemistry 2005

Published on 02 November 2004. Downloaded on 03/06/2015 07:11:41.

View Article Online

Fig. 1 Picture of the proposed system. (a) A 10 6 10 microfluidic cell culture array was bonded to a coverglass and mounted on a transparent
ITO heater. Stainless steel plugs with soft tubings connected the device to a syringe pump for fluidic control. Cell growth was monitored using an
optical microscope. (b) Stable gradient generation across the columns of the array. A blue and yellow dye was diffusively mixed and flowed from
top to bottom, creating 10 different concentrations.

plate. The left and right ports are designed to provide

continuous perfusion of the medium to the chamber for
sustaining cell growth. The top and bottom ports are used to
load cells and reagents for cell-based assays. Each perfusion
channel is 2 mm high and 2 mm wide compared to the main
culture chamber, which is 1 mm in diameter and 40 mm in
height. Fig. 2a and Fig. 2b are scanning electron microscope
pictures of a single unit of the device before bonding, and a
closer view of the high aspect ratio design, respectively. The
perfusion channels serve two main purposes. First, since they
are much smaller than the size of the cells (y10 mm in
diameter), they effectively prevent cells from being flushed
away or from migrating outside the chamber. Secondly, the
multiple perfusion channels provide uniform nutrient access
inside the microchamber as will be discussed later.
The microfluidic cell culture array was fabricated by using
soft-lithography technology and replicate molding (Fig. 3).
SU-8 negative photoresist (Microchem Corportion) was used
as the master template. First, the SU-8 2002 was patterned on
a silicon wafer to define the 2 mm high perfusion channels. A
40 mm SU-8 2050 layer was then spin coated on top of the
perfusion channels. Because the SU-8 2050 is substantially
thicker than SU-8 2002, the surface was planarized after spin

coating. The cell culture chamber and other channels were then
photolithographically defined. PDMS (Sylgard 184, Dow
Corning Corporation) was prepared with a 10 : 1 ratio
between the base and the curing agents. The PDMS was then
poured on the 2-level SU8 mold. The mold was degassed in a
vacuum chamber for 10 min before curing in a 70 uC oven for
4 h. The device dyes were then cut by a razor blade and the
fluidic connection ports were punched using an 18 gauge flat
tip needle. The device was then irreversibly bonded to a
coverglass (Fisher Scientific) after oxygen plasma treatment
(PlasmaTherm Etcher, 50 W, 2 Torr, 40 s) on both the bottom
of the device and the glass slide. 20 gauge stainless steel
connectors (Instech Laboratories) and soft tubings (Cole
Parmer Corporation) were used to provide fluidic connections
to a syringe pump (Cole Parmer 74900).

C. Simulation and characterization

The high aspect ratio of the microfluidic design creates a large
fluidic resistance in the perfusion channels. This has the effect of
providing a uniform flow profile at the outlet of each perfusion
channel. To support this claim that the high aspect ratio design
would help create a uniform culture microenvironment, we

Fig. 2 High aspect ratio design. (a) SEM picture of a single unit of the arrayed device before bonding. Multiple perfusion channels surround the
main culture chamber. The microchamber is 40 mm in height with a diameter of 1 mm. Each culture unit has four fluidic access paths. (b) SEM
image of perfusion channel dimensions. Each perfusion channel had a width of 2 mm and height of 2 mm.

This journal is The Royal Society of Chemistry 2005

Lab Chip, 2005, 5, 4448 | 45

View Article Online

Published on 02 November 2004. Downloaded on 03/06/2015 07:11:41.

the 2D simulations by coarsening and refining the grid spacing

and the time steps by a factor of two and repeating the
calculation.
As the boundary condition, a fully developed velocity profile
was imposed at the inlet with maximum velocity corresponding
to the inlet flow rate (0.2 mL min21). Zero traction and no slip
boundary conditions were set at the outlet and on the walls,
respectively. The flow profiles under various flow rates were
simulated. Due to the large fluidic resistance of the perfusion
channels compared to the microchamber, most of the fluidic
pressure drops on these narrow channels, helping the fluid flow
inside the microchamber to be uniform (Fig. 4). This analysis
indicates that the fluid velocity varies less than 5% of the inlet
flow rate over 90% of the culture area.
For the mass transfer simulation, the carrier liquid is
modeled as glucose solution at 293 K with a solute diffusion
coefficient D 5 1 6 10210 m2 s21. In the nutrient transport
simulation, a constant concentration of nutrient (C 5
25 mM L21) in the carrier liquid is supposed on inlet. The
medium turn around rate is 8 min21 under 0.2 mL min21
flow rate and the mass transfer was verified with food dye at
different time points (Fig. 5).

D. Results

Fig. 3 Flow diagram of fabrication process.

developed a finite element model for simulation of steady

state velocity profile and time-dependent mass transfer
characteristics of the device. The finite element model was
built using the FEMLAB2 software package (version 3.0,
Comsol, Stockholm, Sweden) on a PC platform.
First, the NavierStokes equations were solved for the
carrier liquid to obtain the flow field. This field was then fixed
and the time-dependent simulation of the convectiondiffusion
equation was performed. The dimensions of the domains are
equal to the values chosen in the preceding section. The 2D
domain also contains a longer part for the perfusion channels
to accommodate the fluidic resistance in 3D. The entire
domain was meshed using four-noded tetrahedron finite
elements and a time step of 0.5 s was used in all species
transport simulations. The solutions were found to be
practically independent of the spatial and temporal discretization with the grids and time step used. This was assessed for
46 | Lab Chip, 2005, 5, 4448

Demonstration of cell culture in the device was performed

using the human HeLa cell line. Proper sizes of elastic tubings
were used to interface the microfluidic device with the syringe
pump for fluidic control. The device and the tubings were
sterilized under UV light before use. The device was initially
prepared by filling with PBS to remove dead volume and
bubbles inside the device. HeLa cell suspension was obtained
at a concentration of 106 cells mL21 and introduced into the
microfluidic device via the top port. The cells were allowed to
settle to the bottom of the chamber for 2 h at 37 uC. After
confirming that cells were attached to the bottom of the
chamber, CO2 independent medium (Gibco, NY) was continuously flowed through the perfusion channels using a
syringe pump. During the continuous perfusion of the
medium, the device was placed inside an incubator for
humidity and temperature control (37 uC). Since the PDMS
was gas permeable,9 O2 diffusion from the air was sufficient for
cell culture. Due to the continuous flow of fresh medium into
the culture chambers, a CO2 buffering system for pH was not
necessary for cell survival. Cell growth was monitored daily
using a phase-contrast microscope. At a medium flow rate of
0.13 mL min21, the cells were able to grow to confluency after
8 days (Fig. 6). Cell viability inside the microchamber was
checked by flowing calcein AM (2 mM) from the loading port
to label living cells, indicating a viability of over 95%.

E. Discussion
The ability to control the cell culture environment on the
microscale offers many opportunities for improving biomedical and biotechnological research. The first generation of
microfabricated cell culture devices focused on patterned
surfaces and simple microchannels. In order to enhance
control over the culture environment, we utilized a two-level
This journal is The Royal Society of Chemistry 2005

Published on 02 November 2004. Downloaded on 03/06/2015 07:11:41.

View Article Online

Fig. 4 Finite element simulation of fluid velocity inside the chamber. Cross sections are plotted along the axis of perfusion (AA9 solid line) and
the perpendicular axis (BB9 dashed line). Due to the high aspect ratio between the perfusion channels and the cell culture chamber, the velocity
profile inside the microchamber is uniform within 5% of the inlet velocity over 90% of the culture area.

Fig. 5 Mass transfer. (a) Simulation of fluid flow through the microfluidic cell culture chamber. Medium is introduced from the perfusion inlet at
0.2 mL min21. The entire chamber is evenly filled, suggesting mass transfer can be assumed uniform for cells at different locations in the chamber.
Images depicted after 2, 5, and 15 s of flow. (b) Flow of colored dye at the corresponding time points verifies the mass transfer simulation.

lithography process. This solved the issue of non-uniform mass

transfer in microfluidic channels caused by the parabolic
laminar flow profile. The high aspect ratio design for the
microfluidic device offers many advantages for cell culture,
including size exclusion of cells, promoting a uniform fluidic
environment, ability to generate arrays of chambers, ease of
fabrication, and manipulation of reagents without an extensive
valve network. We hope to further develop this technology in
order to provide a complete miniaturized cell culture system
for high throughput cell-based assays.
We envision such a platform to have applications in many
different disciplines. More cost efficient drug screening
tools10 and rapid bioreactor optimization11 could benefit the
This journal is The Royal Society of Chemistry 2005

pharmaceutical industry. A portable cell culture array could be

deployed in the field or clinic for point-of-care diagnostics of
infectious agents or use in personal medicine. For example,
clinical cell culture is used for the detection and identification
of viruses, such as the causative agent of severe acute
respiratory syndrome (SARS).12 Doctors can also derive
important information about individual patients from cultured
biopsies, such as for optimization of chemotherapy regimens.13
The ability to perform inexpensive high throughput experiments could also have a big impact on biological research,
where thorough characterizations of experimental conditions
are currently limited. For example, integration with microfluidic gradient generators14 enables providing a different
Lab Chip, 2005, 5, 4448 | 47

Published on 02 November 2004. Downloaded on 03/06/2015 07:11:41.

View Article Online

Fig. 6 Cell growth inside the microfluidic cell culture array. Mammalian HeLa cells (human cell line derived from cervical cancer) were used to
characterize growth inside microfluidic chambers. Cells were loaded from the top channel on day 0. Buffered culture medium supplemented with
serum was continuously supplied from the perfusion inlet. Cell growth was recorded daily. At day 7.5, cells were grown nearly to confluency.

culture condition in each chamber of an array. Using this

concept, we have fabricated a PDMS cell culture device that
can potentially assay 100 different conditions on a single chip.
Initial characterization of a ten channel splitter indicated that
the flow was uniform within each column, and a different
reagent concentration could be introduced to different
elements of the array (Fig. 1b). There was no cross flow
between columns due to the high fluidic resistance imposed by
the perfusion channels. Individually addressing each column
and row could be attained through the implementation of a
microfluidic valve network.15 By improving the efficiency of
fluidic delivery, it is possible to introduce a level of
quantitative control to experiments that are traditionally
qualitative. There is also great promise in adapting microfluidic cell culture for research in tissue engineering.16
In this paper, we show the advantages of the high aspect
ratio design between the culture chamber and perfusion
channels and demonstrated the capability of culturing human
tissue culture cells in an arrayed PDMS microdevice. Ongoing
research is being conducted to fully miniaturize the system into
a portable assay device. This might be accomplished through
the integration of electro-osmotic pumps,17 optical sensors,18
and microphysiometers.19 The microfluidic cell culture array
can impact on a wide range of applications in high throughput
cell-based screening, bioinformatics, synthetic biology, quantitative cell biology, and systems biology in the near future.
Paul J. Hung,{ Philip J. Lee,{ Poorya Sabounchi, Nima Aghdam,
Robert Lin and Luke P. Lee
Berkeley Sensor & Actuator Center, Department of Bioengineering,
University of California, Berkeley, California, USA.
E-mail: lplee@socrates.berkeley.edu; Fax: 1-510-642-5835;
Tel: 1-510-642-5855

48 | Lab Chip, 2005, 5, 4448

References
1 S. A. Sundberg, Curr. Opin. Biotechnol., 2000, 11, 47.
2 W. G. Lachnita, U. Warrior, S. Gopalakrishnan, M. E. Schurdale
and D. J. Burns, Drug Discov. Today, 2001, 6, 1200.
3 T. H. Park and M. L. Shuler, Biotechnol. Prog., 2003, 19, 243.
4 G. M. Walker, H. C. Zeringue and D. J. Beebe, Lab Chip, 2004, 4,
91.
5 E. Leclerc, Y. Sakai and T. Fujii, Biotechnol. Prog., 2004, 20, 750.
6 M. J. Powers, K. Domansky, M. R. Kaazempur-Mofrad,
A. Kalezi, A. Capitano, A. Upadhyaya, P. Kurzawski,
K. E. Wack, D. B. Stolz, R. Kamm and L. G. Griffith,
Biotechnol. Bioeng., 2002, 78, 257.
7 K. Viravaidya and M. L. Shuler, Biotechnol. Prog., 2004, 20, 590.
8 G. Walker, M. Ozers and D. Beebe, Biomed. Microdev., 2002, 4,
161.
9 S. Charati and S. Stern, Macromolecules, 1998, 31, 5529.
10 J. Khandurina and A. Guttman, Curr. Opin. Chem. Biol., 2002, 6,
359.
11 Y. Kostov, P. Harms, L. Randers-Eichhorn and G. Rao,
Biotechnol. Bioeng., 2001, 72, 346.
12 C. Drosten, S. Gunther, W. Preiser, S. van der Werf, H. R. Brodt,
S. Becker, H. Rabenau, M. Panning, L. Kolesnikova, R. A. Fouchier,
A. Berger, A. M. Burguiere, J. Cinatl, M. Eickmann, N. Escriou,
K. Grywna, S. Kramme, J. C. Manuguerra, S. Muller, V. Rickerts,
M. Sturmer, S. Vieth, H. D. Klenk, A. D. Osterhaus,
H. Schmitz and H. W. Doerr, N. Engl. J. Med., 2003, 348,
1967.
13 W. T. Bellamy, Drugs, 1992, 44, 690.
14 N. Li Jeon, H. Baskaran, S. K. Dertinger, G. M. Whitesides, L. Van
der Water and M. Toner, Nat. Biotechnol., 2002, 20, 826.
15 T. Thorsen, S. J. Maerkl and S. R. Quake, Science, 2002, 298,
580.
16 H. Andersson and A. van den Berg, Lab Chip, 2004, 4, 98.
17 S. Attiya, A. B. Jemere, T. Tang, G. Fitzpatrick, K. Seiler,
N. Chiem and D. J. Harrison, Electrophoresis, 2001, 22, 318.
18 R. Ulber, J. G. Frerichs and S. Beutel, Anal. Bioanal. Chem., 2003,
376, 342.
19 F. Hafner, Biosens. Bioelectron., 2000, 15, 149.

This journal is The Royal Society of Chemistry 2005