Design of Ba Passive Microfluidic Mixer for BioMEMS Application
Fei Mi, Shuhan Kan, Advisor: Prof. Xingguo Xiong
Department of Biomedical Engineering, University of Bridgeport, Bridgeport, CT 06604
Results and Discussions
Abstract
Bio-MEMS (Bio-Microelectromechanical Systems) have been widely used for disease
diagnosis and treatment. In bio-MEMS devices, the mixing of different microfluid is
frequently needed. However, such mixing has been very challenging due to the fact
that microfluid is generally laminar flow. As a result, MEMS mixers which can enhance
the mixing of different microfluid are in pressing need. In this poster, the design and
simulation of a passive microfluidic mixer for Bio-MEMS (bio-microelectromechanical
systems) application is proposed. The proposed MEMS mixer utilizes re-convergent
channels to introduce turbulence, hence enhancing the mixing of two different
microfluidic flows. The mixing of microfluid in the proposed mixer is analyzed. Based
on the theoretical analysis, the dimension parameters of the mixer is decided. ANSYS
simulation is used to verify the effectiveness of the MEMS mixer device. The
fabrication flow of the MEMS mixer is also suggested. The proposed MEMS mixer can
be used for lab-on-a-chip, digital microfluidics, micro-drug delivery system and other
bio-MEMS applications.
ANSYS FEM simulation is used to verify the function of the MEMS mixer. The
velocities of both of species 1 and species 2 at inlets are set to be 1000μm/s
(boundary condition). Density of species 1 and 2 are set as 1g/cm3 and 1.5 g/cm3
respectively. Hand calculation shows the Reynolds number of the microfluid flows are
under 10, which belongs to laminar flow. The meshed ANSYS model of the MEMS
mixer is shown in Fig. 2. The contour plot of the flow density is shown in Fig. 3. From
Fig. 3, we can clearly see how both microfluid flows are mixed rapidly along the
channels. The blue and red color represent microfluid species 1 and 2 respectively.
Once both fluids are thoroughly mixed, its color becomes uniform yellow.
Introduction
Bio-MEMS devices need to manipulate different microfluids, such as blood, saliva,
urine samples. The mixing of different microfluid samples are frequently needed.
However, the mixing of microfluid samples is very challenging because most microfluid
is laminar flow instead of turbulent flow. The mixing of laminar flow solely relies on
diffusion, which is very slow. In order to improve the mixing efficiency, specially
designed MEMS mixers are needed. MEMS mixers can be divided into two categories:
passive and active mixers. In passive mixers, no active power is required. Passive
micro-mixers just use geometrical shape or fluid characteristics, to generate and mix
different fluids to get the effect of chaotic advection. Active micro-mixers use exterior
input energy to produce pulses to increase the rate and degree of mixing, such as
piezoelectric vibration, acoustic wave, heat and magnetic field. In this project, we
proposed a passive multi-folded MEMS mixer design, and use ANSYS to simulate its
effectiveness in mixing microfluid flows. By introducing split-and-recombination in the
mixer design, the microfluids can be effectively mixed. MEMS passive mixer has the
advantages of small size, low cost, no energy consumption and high reliability. It can
be used for various biomedical applications.
Fig. 2 Meshed ANSYS model of mixer Fig. 3 Contour plot of density distribution
The flow velocity vector plot of the microfluid along MEMS mixer is shown in Fig. 3.
From Fig. 3, we can clearly see that turbulence (change of flow velocity direction) is
successfully induced at each turn corner along the mixer, which is very helpful for the
mixing of laminar flow. The local view of the velocity turbulences are shown in Fig. 5.
Passive Micro-mixer Design
The proposed passive micro-mixier design is shown in Figure 1. With 3 inlets (1
vertical inlet and 2 horizontal inlets) and 1 outlet, it can effectively mix up to 3 different
microfluid flows. Microfluid flows from 3 inlets are first mixed into one channel, then
divided into 2 branches and then further divided into four sub-branches. After that, the
flows from 4 sub-branches are mixed back into two branches, and finally rejoined back
into one channel and flow out of the outlet. By introducing many rounds of turn-overs,
turbulence is induced each time when microfluid flow changes its flow direction. The
total length of the micro-mixer is about 20480 µm. The channel width are also
dynamically adjusted to adapt to different amount of microfluid flow along the channel.
For example, if multiple branches are joined together, the corresponding channel is
enlarged to accommodate the increased microfluid flow. If a channel is divided into
multiple branches, the corresponding branches are narrowed to adapt to the reduced
microfluid flow.
Fig. 4 ANSYS fluid velocity vector plot
Fig. 5 Local view of induced tubulence
The mixing can be verified not only from density contour plot, but also the density path
plot along the cross section of the inlet and outlet. The ANSYS simulated density
distribution plot along the cross section of the first top vertical channel section near the
inlets is shown in Fig. 6. We can see the microfluid is not well mixed yet at this time,
because there are more species 1 in the middle, and more species 2 in both sides
close to the channel sidewalls. After passing the MEMS mixer, the density path plot
along cross section of outlet is shown in Fig. 7. It shows the microfluids (species 1 and
2) are now thoroughly mixed, and a uniform density distribution is achieved.
Table 1. Micro-mixer design parameters
Parameter
Dimension
Inlet Channels
80 µm
Secondary Vertical
Channel
240 µm
Horizontal Channel
after secondary
channel
120 µm
Central/ Passing
Rectangular
Channels
60 µm
Horizontal Channel
Near Outlet
120 µm
Outlet Channel
240 µm
Figure 1. Structure of passive micro-mixer
Assume the microfluid flow in vertical inlet (labeled as fluid A) contains a solution like
protein, and the concentration is set to 1. For the microfluid flow in two horizontal
inlets (labeled as fluid B) is the same and its concentrations is set to be 0 (i.e. pure
water). Based on microfluid dynamics analysis, we can estimate the mixing efficiency
as well as the intensity of segregation and variation coefficient of the microfluid flows
inside the mixer. In additional, Reynolds number is also a very important factor of fluid
flow measuring the ratio between inertial forces and viscous in particular flow. For
Reynolds number below 2100, we can get laminar flow in which the fluid molecules
travel along well-ordered non-intersecting paths or layers. When Reynolds number is
over 4000, we can get turbulent flow that fluid molecules from adjacent layer become
totally mixed. For microfluidics, generally we get laminar flow.
Reynolds Number is calculated as:
The equation of Mixing Efficiency:
Re  Vref Lref

  1 


n

i 1
 Ci  C 
n

2

/ C   100%


where Ci is the mole percentage of the No. i dot in a certain area, n is the number of
style points in a certain cross section, C  is the ideal mole percentage, Ci is a number
in range from 0 to 1. When Ci is 0.5, it indicates A fluid and B fluid are mixed
completely.
2
2
The Intensity of Segregation and Variation Coefficient: Vc   0 xv I s   0  i
 i and  0 are the standard deviation of the mass fraction of species at the inlet and
outlet of micromixer, xv is the mass fraction of species at the inlet.
Fig. 6 Density distribution near inlet
Furthermore, we also simulated the case
when the density of species 2 is changed
from 1g/cm3 to 35g/cm3, the result is shown
in Fig. 8. Compared with result in Fig. 3, we
can see that as the initial density different
among input microfluid becomes larger, it
takes longer distance for both microfluids to
thoroughly mix with each other. This
indicates it’s more challenging to mix two
microfluids with larger difference in its
properties.
Fig. 7 Density distribution at outlet
Fig. 8 Density plot (ρ2=35g/cm3)
Device Fabrication
The proposed passive MEMS mixer is to be fabricated with Si bulk-micromachining
and bonding techniques. The fabrication flow is shown in Fig. 9.
1. Preparing for the silicone wafer;
2. Si thermal oxidation to grow 0.2µm SiO2,
3. Photolithography, etching SiO2 to open
windows,
4. Use SiO2 as etching mask, KOH anisotropic
etching to etch down microchannels,
5. Remove all SiO2 with buffered HF solution,
6. Bond Si wafer with Pyrex #7740 glass by Siglass
anodic
bonding
technique, the
microchannels are sealed with glass top cover.
The MEMS miccromixer device fabrication is Fig. 9 Fabrication flow of mixer
completed.
Conclusion and Future Work
In this project, the design and analysis of a passive MEMS mixer is proposed. The
passive MEMS mixer utilizes multiple turns to improve the mixing efficiency of laminar
microfluid flows. ANSYS simulation is used to verify the function of the mixer. The
proposed MEMS mixer can be used for various bio-MEMS applications. In the future,
we will further analyze how other microfluid properties (e.g. viscosity, temperature)
affect the mixing result of the mixer.