Flow Characterization of a Liquid Micromixer Tony Hyun

Flow Characterization of a Liquid Micromixer
Tony Hyun Kim, Troy V. Tamas
6.152J/3.155J Nano/Micro Processing Technology, Spring 2009
Abstract: Measurements of fluid velocity and diffusion coefficient were performed in a channel of a microfluidic
mixer. Both are shown to be in good agreement with theory, with Umax(meas)=1.780.11 mm/s [c.f. Umax(theory)=1.84];
-6
2
-6
2
and Dmeas=1.510 cm /s [c.f. Dtheory = 210 cm /s].

Introduction

Fluid Velocity

A microfluidic device was created in
order to mix the contents of two
reservoirs through a 200um-wide,
30mm-long diffusion channel.

The operation of the micromixer was
visually inspected via an optical
microscope. The flow velocity and the
diffusion coefficients were in good
agreement with predicted values.

Fabrication

Theory

Theory

Assume laminar flow, for which:

The two fluids mix as they flow down
the channel, showing increase in
diffusion length scale, w.

12L
P
Q
3
WH

Q: flow rate
: viscosity
L,W,H: channel

Maximum fluid
velocity is:
2
U max

H

P
8L

Experimental Setup
Input reservoirs were loaded with
22.50.7 mm of water.
Microbeads were added to each
reservoir.
The motion of the microbeads in the
channel was filmed using a PC-based
microscope.
t
t +t
0

0

Spin on 50-micron thick SU-8
film.

Si

UV light

Si
develop

w

Si

PDMS
Si

Develop with PM-acetate and
silanize to discourage
bonding with PDMS.
Pour PDMS on master wafer.
Al rods define the reservoir
locations.
Remove PDMS mold.

PDMS

Velocity was measured by analyzing
snapshots of the fluid flow, as shown
above.

Maximum Fluid Velocity:
tube

glass

Place on glass slide and
bond with ashing. Insert tube
into reservoirs.

Umax=1.780.11 mm/s (measured)
Umax=1.84 mm/s

(theoretical)

D~210-6 cm2/s

(theoretical)

Diffusion coefficient:

Actual diffusion was constrained by an
opposite wall. (See below)
restricted diffusion (lower D)

w(t ) 2
D
t
4 ln 2

Experimental Setup
Two reservoirs were filled with water
and fine-tuned to have equal heights.

Conclusion

Red dye was added to one reservoir
to track diffusion

Microfluidic mixer was fabricated.

LEFT: The boundary between
the two fluids was set at the
midpoint of the channel by
adjusting the heights of the two
input reservoirs.

Flow characterization experiments
were performed, yielding:
Fluid velocity
Diffusion coefficient

RIGHT: Blue intensity profile
across the lateral channel
position, at 200 pixels above
bottom of image.

Image analysis technique:

Results

(measured)

Theoretical model was based on diffusion into
an infinitely sized sample.

Using Uavg = 2/3*Umax, and the
position along the channel, duration
of diffusion is determined.

ABOVE: By considering two consecutive snapshots of the flow (with
known time difference between them) the velocity of particle flow can be
determined.

Pixel values were calibrated by
measuring known feature sizes.

D=1.510-6 cm2/s

Why is the diffusion coefficient lower than
expected?

C

x
Prepare master template by
contact photolithography.

Results
Diffusion Coefficient:

dimensions
LEFT: Device mask for
the micromixer. The top
two reservoirs are
connected to the
bottom, output
reservoir through a
long, narrow channel.

Diffusion Coefficient

Filter/smooth profile
Differentiate the profile
Measure half-width half-maximum
of differential to obtain the diffusion
length scale, w

Measurements show good
agreement with theory device
behaves as expected.

Acknowledgements
We would like to thank Scott Poesse for his
guidance in lab; as well as the other students in
our group: Edison, Cankutan, Rebecca and
Michael.
Contact: [email protected], [email protected]