Mixing of two uid streams in a microchannel using the

Taylor-Aris dispersion eect

Hengzi Wang, Pio Iovenitti, Erol Harvey and Syed Masood
Industrial Research Institute Swinburne,
Swinburne University of Technology, PO Box 218,
Hawthorn,Victoria 3122, Australia
ABSTRACT
Taylor-Aris dispersion is an unwanted eect in some applications, such as chromatography, because of its rapid
dispersion along channel axis direction to cause the diculties to separation. However, this eect can be used
in solving mixing problems. In this presentation, the authors studied mixing of two streams of food dye solution
in a microchannel. The two liquids sandwiched in the axis direction, and due to non-slip boundary condition,
the slugs of liquids stretched and thus increase their interfacial area. This phenomena was rstly studied by
Taylor about the dispersion of solute in a circular capillary, and then improved by Aris. Numerical analysis was
applied to design mixing section in microchannels, and simplied experiments were conducted to illustrate the
concept.
Keywords: Taylor-Aris dispersion, Microuidics, microchannel, MEMS, micromixer, passive mixing
1. INTRODUCTION
It is well known that the Reynolds number is low in typical microuidic channels, and the ow is laminar under
normal conditions, especially for liquids. Therefore, the mixing of two uid streams in a microchannel relies
mainly on diusion. For a typical microuidic device, the length scale is too large for a rapid diusion and
too small to include mechanical agitation. It is possible to achieve this by dynamic mixing with the assistance
of externally forced mass transport, i.e. Yang et al.
1
used ultrasonic waves to enhance mixing, and Knight
2
described fast mixing by forming and controlling nanoscale, submerged uids jets. However, it is much easier
to control mixing in a laminar ow using a static micromixer. The principle of a static micromixer can be
categorized into four types. (a). T-shape micromixer. The T-mixer simply combines two or more uid streams,
which ow parallel to each other in the microchannel, and mixing relies purely on molecular diusion.
36
The
T-mixer normally has a small channel width of the order of tens of microns and sucient channel length. (b).
Geometrically splitting and recombining substreams. In this way, large contact surfaces and small diusion
paths are generated (Ehrfeld et al.,
7
Schwesinger, et al.,
8
Koch et al.
9, 10
). (c). Chaotic Mixer. Stroock et
al.
11
presented this mixer to stretch and fold the streams by rotating the streamline inside the microchannel.
The circulation of the streamlines was due to the transverse pressure component created by the anisotropic
resistance to viscous ow. Johnson et al.
12
reported similar structures machined by Excimer laser to achieve
rapid mixing. However, the obvious drawback for these mixers is the dead volume created by these structures.
(d). Altering ow direction laterally. This method tries to create stirring (convective) eects by forcing one
uid stream into another (Liu et al.,
13
and He et al.
14
). Lius serpentine-shaped micromixer was more favorable
to high Reynolds numbers ( 70), and Hes in-situ micromixer could alter the ow direction, shrink the channel
and divide the main stream into substreams. However, Hes mixer is not likely to be used in a pressure-driven
environment, due to the large pressure drop in its packed columns. Taylor and Aris
15, 16
described the dispersion
of solute in a circular pipe along the longitudes (axial) direction rather than transverse direction, like many
T-type microchannel do. There were also extensive discussions about the Taylor-Aris dispersion in a rectangular
Send correspondence to Hengzi Wang: E-mail: hwang@swin.edu.au, Telephone: 61 3 9214 4332, Fax: 61 3 9214 5050
Biomedical Applications of Micro- and Nanoengineering, Dan V. Nicolau, Abraham P. Lee, Editors,
Proceedings of SPIE Vol. 4937 (2002) 2002 SPIE 0277-786X/02/$15.00
158
channel.
17, 18
From these discussions, it is generally accepted that the apparent diusion coecient D

(Taylors
dispersivity) in a rectangular channel can be described in equation (1),
D

= D +
1
210
u
2
W
2
D
f(
h
W
) (1)
where D is the molecular diusion coecient, W is the width of the channel, u is the mean velocity in the
channel, h is the height of the channel and f is the geometric function of a rectangular channel, decided by
the height to width aspect ratio h/W. It is surprisingly that the geometric function f is not unity,
18
instead,
f = 7.95 while h/w 0. For a Taylor dispersivity D

, the length between the dispersed leading and trailing

edges of slug, l
m
, shall be estimated by equation (2) developed by Einstein when he studied Brownian motion,
19
l
m
=

2dt (2)
Taylor dispersivity D

is much large than the molecular diusivity D, therefore, for a xed width d of a channel,
the mixing time t is reduced.
2. EXPERIMENTAL AND NUMERICAL DESIGNS
2.1. Numerical setup
While experiments can illustrate the dispersion a slug of solute in a microchannel and reveal that it is quicker
than pure diusion, numerical simulations can explain some details better and also save time. A straight channel
was modelled numerically using the computational uid dynamic (CFD) software package MemCFD
r
v2001.3d
from Coventor. Flow visualization was performed with Coventor visualizer.
20
MemCFD
r
uses a nite volume
method to solve the Navier-Stokes equations. The channels were meshed into 8-nodes hexahedral elements. The
simulation was run as transient (steady for studying obstacles in a microchannel), laminar, Newtonian, with 2
uids to evaluate the dispersion of one uid in another. The inlets were assigned ow rate boundary condition,
and all the channel walls were assigned wall boundary condition (velocity components: v
i
= 0, i = 1 3). Fluid
properties were calculated and consistent with the uids used in the experiments. In this case, low concentration
of food coloring dyes in water solutions were used.
The length, width and height of the channels were 2mm, 200 m and 100m respectively. Simulations were
performed on Win NT4.0 with Pentium III 800MHz CPU and 512MB memory.
2.2. Experimental Setups
Plane of uniformity
& Mask
Lens Array Homogenizer
Input Beam
248 nm (KrF) or 193nm (ArF)
Projection Lens
Workpiece step motion
Figure 1. Excimer micromachining
The microuidic channel was fabricated in the microtechnology laboratory in the Industrial Research Insti-
tute of Swinburne (Hawthorn, AU). An Excimer Laser (ExiTech) was used to create the channels. The standard
Proc. of SPIE Vol. 4937 159
Figure 2. Section of Y-channel machined in Polycarbonate by Excimer Laser
1sec
2sec
5sec
8sec
10sec
12sec
15sec
18sec
20sec
Figure 3. Taylor dispersion along a straight channel: numerical simulation for various time intervals
operation procedures can be found in early papers of the laboratory.
21
The UV laser beam, with a wavelength
of 248nm, projected through the designed features on a chrome-on-quartz mask, and then the laser beam is
focused to pattern the feature on the substrate (Figure 1 ). The Excimer laser machined structures have a
depth resolution of the order of 0.1m and spatial resolutions of the order of 1m or better.
Figure 2 shows the Excimer laser ablated Y-channel on a polycarbonate substrate. Then, the channel was
packaged in a lamination process, a thin PET foil coated with a melting adhesive layer was pressed by a heated
lamination roller onto the structure as the lid. Melinex
r
Polyester lm type 301 (Dupont Teijin lms), 30m
in thickness was used in our fabrication, and the lamination equipment was MEGA dry lm laminator model
305 (Mega Electronics).
The aqueous solutions, one was mixture of water, 2.4% food dye E102 and E122(Yellow), and another one was
mixture of water and 2.1% food dye E133 (blue)(Queen Fine Foods Pty Ltd, Australia), were introduced to the
channel by capillary eects, and the solute plug was pumped into the channel by a syringe. Flow rates were
measured by weighing the uid collected at the outlet of the channel. Viscosity of the food dye/water solution
was estimated approximately the same as water. Diusion coecient was calculated according to the chemical
groups.
22
160 Proc. of SPIE Vol. 4937
500m
500m
Figure 4. The solute dispersed at, (a) time t1 and (b) time t2
3. RESULTS AND DISCUSSION
Analytical results can evaluate Taylors dispersivity for rectangular channels with dierent depth to width aspect
ratio.
17
The transient behavior of most practical applications are much easier to interpret through numerical
simulations. In Figure 3, a time series of mass fraction of uid 1 dispersed in uid 2 for a time period of 20 sec-
onds. To calibrate with the conditions for analytical interpretation, the Reynolds number was low (Re = 0.005)
for diusion to sample all the streamlines.
16
It is quite clear that the length of the dispersed uid is longer
than a pure diusion, because of the convective spreading of uid physically. As Taylors dispersivity coecient
is a function of mean velocity, D

= f( u), it is quite intuitive to have a large ow rate for a quick mixing. This
is dierent from Taylors original study of dispersion of soluble matter in solvent owing slowly.
16
However,
in most practical applications, it is not necessary to have a slow motion. As a matter of fact, the velocity is
usually higher and has a parabolic shape. While velocity becomes larger, there will be less time for the diusion
to sample all the streamlines. In Figure 4, the green region was the mixed uids between uid 1 (yellow) and 2
(blue). And there were two transition regions between the yellow/green and green/blue, where the uids began
to diuse into each other. The dispersed solute was initially a narrow parabolic shape at time t1, and then
developed into a wider parabolic shape at time t2 in a short distance. Although there is lack of quantitative
measurement for evaluating the mixed volume of the two uids, the enlarged area of mixed green region and
lengthened parabolic shape indicated the enhance of mixing. Nevertheless, numerical simulation agreed with
experiments qualitatively and can be used to access the quality of the mixing. More structured mesh with
higher density of elements can reduce numerical diusion and improve the accuracy, whenever computer re-
source allows.
It was eective to inject one slug into the main stream of uids for certain analysis. It is ideal to create a series
of sandwiched slugs of uids proportional to each other according to specication of microuidic mixers. To do
it, a dynamic system is required, i.e. a membrane switch or a periodic micropump.
The concept of Taylor-Ariss dispersion can be shifted to a more broad way of interpretation, such as using
obstacles in the channel to disrupt ow and create transverse component of ow,
23
which is brought into the
main stream in the longitudinal direction. It suggested that convection happened in the both the longitudinal
direction and transverse direction. It is dierent from the conventional Taylor-Aris dispersion which require
slow motion and limited to one-dimension, the ow pattern in a microchannel with obstacles was not exactly
slug-like and can be two-dimension or three-dimension.
4. CONCLUSION AND FUTURE WORK
Simple microchannels were fabricated and dispersion of slugs of solute was experimentally and numerically
investigated. It is dicult to apply dynamic systems to create sandwiched slugs in the channel. Alternatively,
obstacles can be a passive approach to create convective eects, which in principle is similar to Taylor-Aris
dispersion.
Proc. of SPIE Vol. 4937 161
Optimising various geometric parameter of obstacles were conducted to create more eective convection, and
will be reported elsewhere. In the meantime, it will be a promising solution to microuidic application by
applying dynamic systems to introduce sandwiched slugs to create complete mixing.
ACKNOWLEDGMENTS
The authors wish to express their gratitude to the Cooperative Research Centre (CRC) for Microtechnology for
the support in this research project. Dr. Rowan Deam also provided very helpful discussions.
REFERENCES
1. Z. Yang, H. Goto, M. Matsumoto, and R. Maeda, Active micromixer for microuidic systems using lead-
zirconate-titanate(pzt)-generated ultrasonic vibration, Electrophoresis 21(1), pp. 116119, 2000.
2. J. B. Knight, A. Vishwanath, J. P. Brody, and R. H. Austin, Hydrodynamic focusing on a silicon chip:
Mixing nanoliters in microseconds, Physical Review Letters 80(17), pp. 38633866, 1998.
3. A. Kamholz and P. Yager, Theoretical analysis of molecular diusion in pressure driven laminar ow in
microuidic channels., Biophysical Journal 80(1), pp. 155160, 2001.
4. A. E. Kamholz, B. H. Weigl, B. A. Finlayson, and P. Yager, Quantitative analysis of molecular interaction
in a microuidic channel: the t-sensor, Analytical Chemistry 71(23), pp. 53405347, 1999.
5. S. C. Jacobson, T. E. McKnight, and J. M. Ramsey, Microuidic devices for electrokinetically driven
parallel and serial mixing, Analytical Chemistry 71(20), pp. 44554459, 1999.
6. A. E. Kamholz, E. A. Schilling, and P. Yager, Optical measurement of transverse molecular diusion in a
microchannel, Biophysical Journal 80(4), pp. 19671972, 2001.
7. H. L. Wolfgang Ehrfeld, Volker Hessel, Microreactors: New Technology for Modern Chemistry, John Wiley
& Sons, 2000.
8. N. Schwesinger, T. Frank, and H. Wurmus, Modular microuid system with an integrated micromixer,
Journal of Micromechanics and Microengineering 6(1), pp. 99102, 1996.
9. M. Koch, D. Chatelain, A. Evans, and A. Brunnschweiler, Two simple micromixers based on silicon,
Journal of Micromechanics and Microengineering 8(2), pp. 123126, 1998.
10. M. Koch, H. Witt, A. Evans, and A. Brunnschweiler, Improved characterization technique for micromix-
ers, Journal of Micromechanics and Microengineering 9(2), pp. 156158, 1999.
11. A. D. Stroock, S. K. W. Dertinger, A. Ajdari, I. Mezic, H. A. Stone, and G. M. Whitesides, Chaotic mixer
for microchannels, Science 295(5555), pp. 647651, 2002.
12. T. J. Johnson, D. Ross, and L. E. Locascio, Rapid microuidic mixing, Analytical Chemistry 74(1),
pp. 4551, 2002.
13. R. H. Liu, M. A. Stremler, K. V. Sharp, M. G. Olsen, J. G. Santiago, R. J. Adrian, H. Aref, and D. J.
Beebe, Passive mixing in a three-dimensional serpentine microchannel, Journal of Microelectromechanical
Systems 9(2), pp. 190197, 2000.
14. B. He, B. J. Burke, X. Zhang, R. Zhang, and F. E. Regnier, A picoliter-volume mixer for microuidic
analytical systems, Analytical Chemistry 73(9), pp. 19427, 2001.
15. R. Aris, On the dispersion of a solute in a uid owing through a tube, Proceeding of the Royal Society
of London, series A (Mathematical and Physical Science) 235, pp. 6777, 1956.
16. G. Taylor, Dispersion of soluble matter in solvent owing slowly through a tube, Proceeding of the Royal
Society of London, series A (Mathematical and Physical Science) 219, pp. 186203, 1953.
17. D. Dutta and D. T. Leighton, Dispersion reduction in pressure driven ow through microetched channels,
Analytical Chemistry 73(3), pp. 504513, 2001.
18. P. Chatwin and S. P.J., The eect of aspect ratio on longitudinal diusivity in rectangular channels,
Journal of Fluid Mechanics 120, pp. 347358, 1982.
19. A. Einstein, On the motion of small particles suspended in liquids, Annalen der Physik 17, pp. 549560,
1905.
20. CoventorWare, Coventorware users manual, 2001.1, 2001.
162 Proc. of SPIE Vol. 4937
21. E. Harvey, J. Hayes, B. Dempster, T. Mackin, and R. Scholten, Excimer laser ablation used for the
fabrication of micro-optic phase and diraction elements, in Proceedings of SPIE: Micro-Opto-Electro-
Mechanical System, pp. 152158, Society of Photo-Optical Instrumentation Engineers, (Glasgow. UK),
2000.
22. R. C. Reid, J. M. Prausnitz, and B. E. Poling, The Properties of Gases and Liquids, McGraw-Hill, Inc.,
4th ed., 1987.
23. H. Wang, P. Iovenitti, E. Harvey, and S. Masood, Optimizing layout of obstacles for enhanced mixing in
microchannels, Smart Materials and Structures 11(5), pp. 662667, 2002.
Proc. of SPIE Vol. 4937 163