Separation Microfluidic Devices Fabricated by
Different Milling Processes
Inês M. Gonçalves
1,2
, Miguel Madureira, Inês Miranda
3
, Helmut Schütte
4
, Ana S. Moita
2,5 a
,
Graça Minas
3
, Stefan Gassmann
4
and Rui Lima
1,6
1
METRICS, University of Minho, Guimarães, Portugal
2
IN+, Instituto Superior Téncico, Universidade de Lisboa, Lisboa, Portugal
3
Center for MicroElectromechanical Systems (CMEMS-UMinho), University of Minho, Guimarães, Portugal
4
Jade University of Applied Science, Wilhelmshaven, Germany
5
CINAMIL, Departamento de Ciências Exatas e de Engenharia, Academia Militar,
Instituto Universitário Militar, Lisboa, Portugal
6
CEFT, Faculty of Engineering of the University of Porto, Porto, Portugal
Keywords: Separation Methods, Blood, Microfluidic Biomedical Devices, Micromilling.
Abstract: The diagnostic of several diseases can be performed by analysing the blood plasma of the patient. Despite of
the extensive research work, there is still the need to improve the current low-cost fabrication techniques and
devices for the separation of the plasma from the blood cells. Microfluidic biomedical devices have great
potential for that process. Hence, two microfluidic devices made by micromilling and sealed with or without
the solvent bonding technique were tested by means of a blood analogue fluid. A high-speed video microscopy
system was used for the visualization and acquisition of the analogue fluid flow. Then, the separation of
particles and plasma was evaluated using the software ImageJ. The device manufactured by the micromilling
process without bonding showed a significant reduction of the amount of cells between the entrance and the
exit of the microchannels. However, further analysis and optimizations of the microfluidic devices will be
conducted in future work.
1 INTRODUCTION
Several diseases can affect the components of the
blood, altering its rheological properties.
Consequently, blood plasma and cells are commonly
used for disease diagnosis and therapeutics (Yin et
al., 2013; Nam et al., 2015; Shamloo et al., 2016).
Different information can be obtained from different
components. For instance, the blood plasma is used to
evaluate the inflammatory response and in proteomic
studies, while the red blood cells are used to diagnose
certain diseases (Lee et al., 2014). Therefore, cell
separation is important in medicine and performed
millions of times per day over the world (Al-Fandi et
al., 2011; Bento et al., 2019, Catarino et al., 2019;
Pinho et al. 2020). Nevertheless, some limitations still
hinder the separation process. The most common
techniques used are filtration and centrifugation. The
former was the earliest to be used but has the
a
http://orcid.org/0000-0001-9801-7617
disadvantage of filter clogging that leads to a
reduction of the process efficiency (Lee et al., 2014).
Centrifugation is the most common method to
separate blood cells from plasma and makes use of
the density difference between the plasma and the
cellular components (Nam et al., 2019). One of the
problems that arise from this technique is the cell
damage due to the high shear forces which can lead
to cell lysis and contamination of the plasma (Geng et
al., 2013). Developments in the field of microfluidics
allowed the improvement of blood separation
techniques with the fabrication of smaller devices
with precise geometries, making the technology
suitable for automation, low-cost, portable, with
improved sterility and requiring less preparation time
and amount of sample and reagents (Bento et al.,
2019, Nam et al., 2019; Tsutsui and Ho, 2009;
Ookawara et al., 2010).
Gonçalves, I., Madureira, M., Miranda, I., Schütte, H., Moita, A., Minas, G., Gassmann, S. and Lima, R.
Separation Microfluidic Devices Fabricated by Different Milling Processes.
DOI: 10.5220/0010906500003123
In Proceedings of the 15th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2022) - Volume 1: BIODEVICES, pages 185-190
ISBN: 978-989-758-552-4; ISSN: 2184-4305
Copyright
c
2022 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
185
The design of the microfluidic devices has been
developed to better mimic the geometry of in vivo
microvascular networks. Different separation
mechanisms were also developed and can be divided
in active or passive methods (Dominical et al. 2015;
Shamloo et al., 2016). In the active methods, an
external force is applied to move the cells apart from
the plasma. Those forces can be gravitational,
acoustic, magnetic, optical, chemical, electronic or
electrical. The gravitational method is the most
popular but has some limitations in microfluidic
devices due to the interface tension, low Reynolds
number and laminar flow that occur in flows at those
dimensions (Xue et al., 2012). On the other hand,
passive techniques do not require an external force
and are preferred to the active techniques due to ease
of use and fabrication (Lee et al., 2011; Nam et al.,
2015, Catarino et al., 2019). The passive techniques
depend on hydrodynamic effects that occur in the
fluid when flowing through capillaries. If the
Reynolds number is in values of unity order, the
viscous forces balance the inertial forces and the red
blood cells have the tendency to move towards the
center of the channel, leaving a cell free layer (CFL)
near the walls (Shamloo et al., 2016). This
phenomenon leads to the reduction of the apparent
viscosity, a phenomenon called the Fåhræus-
Lindqvist effect, and to less hematocrit in small
vessels than in large ones (Yin et al., 2013, Losserand
et al., 2019; Medhi et al., 2019). The geometrical
designs used for this type of microchannels include
channel constriction, bending channel and bifurcated
channels. Bending channels make use of centrifugal
forces, constrictions in channels force particles to the
center of the channel while bifurcations make use of
the ZweifachFung effect. This effect consists in the
tendency of the cells to travel to the branch channel
of an asymmetric bifurcation with the higher flow rate
since the pressure is lower (Fekete, et al., 2012; Xue
et al., 2012; Meyari et al., 2020).
In this work a microfluidic device for separation
of cells and plasma was investigated. The device was
composed of a microchannel with a series of three
bifurcations and was fabricated by two different
methods. Two blood analogues were used to infer on
the efficiency of the cell clearance and the results
were analysed using ImageJ software. Though further
optimization of the design and fabrication process is
still required, a significant reduction of cells was
obtained with one of the microchannels.
2 MATERIALS AND METHODS
2.1 Microchannel Geometry
The proposed geometry was selected and optimized
by taking into account several previous flow
experiments performed by Lopes et al., 2015, Singhal
et al., 2016 and Madureira et al., 2018. Hence, the
proposed microfluidic device is composed by four
microchannels with the same geometry, but different
depths. The microchannels comprise three separation
segments connected in sequence. After the entrance
the channel width narrows to force the cells to the
center of the channel. Then, an intersection was
placed so the cells continue forward to the exit while
the plasma flows upward or downward through
branch channels. Those channels that will merge
again to form the main channel and the structure are
repeated two more times. A scheme of the
microchannel can be seen in Figure 1and the depth of
each channel is presented in Table 1.
Figure 1: Microchannel geometry with detail showing the
dimensions of the different regions of the segments.
Table 1: Depth of the different microchannels of the
microfluidic devices.
Microchannel
1 and 2
3
and
4
Manufacture 1
W
1
, W
2
, W
4
300 μm
150
μm
W
3
100 μm
50
μm
Manufacture 2
W
1
, W
2
, W
4
300 μm
150
μm
W
3
150 μm
50
μm
Outlet 2
Outlet 1
Outlet 3
Outlet 4
W
1
= 260 μm
W
2
= 150 μm
W
3
= 150 μm
W
4
= 360 μm
BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices
186
2.2 Fabrication Methods
2.2.1 Manufacture Process 1
One of the microfluidic devices was fabricated from
poly(methyl methacrylate) (PMMA) through a
micromilling process. The geometry of the
microchannels was outlined using a CAD software
and then the device was manufactured using a
micromilling machine (Minitech Mini-Mill/GX. Two
different micromilling tools were used for the
manufacture, one with 100 μm and the other with 1
mm in diameter. The geometry was converted to NC
code using the software Visual Mill so it could be read
by the milling machine. Also, several parameters,
such as, the angle, depth and velocity of the tool, were
defined using that software. All the parameters were
set to each of the used tools considering their different
properties. Starting the milling process, the reference
for the z-axis has to be set manually since the milling
equipment has no height control. A magnifying
camera was used to observe the first touch of the
milling tool on the material. To avoid the damaging
the milling tools due to the heat resulting from the
high rotation speed, the work piece and the tools were
cooled with a mixture of water and detergent during
the milling process. More details of the micromilling
process can be found elsewhere (Lopes et al., 2015;
Singhal et al., 2016; Madureira et al., 2018).
2.2.2 Manufacture Process 2
The second microchannel device was also
manufactured by the micromilling technique above
described. However, for this device, two PMMA
plates were milled and then bonded together using the
solvent bonding technique. The microchannels were
milled in one plate and the other plate worked as the
lid. After the milling and washing, the plates were
placed together and exposed to a solvent vapor which
creates an irreversible bonding between them. In the
end of the treatment, the plates were pressed to ensure
the bonding was maintained.
On both devices, stainless steel tubes with a length of
about 1 cm and an inner diameter of about 1 mm were
used as inlets and outlets of the microchannels. The
tubes were fixed to the Plexiglass (acrylic glass,
Plexiglas XT 0A000) using epoxy glue. The
microchannel devices were then cleaned with water
and detergent for 20 minutes using ultrasonication.
Lastly, the devices were dried with pressurized air
and sealed with an adhesive foil for PCR plates
(Polyester 50 μm ultra-clear, non-sterile, heat-
resistant).
2.3 Experimental Set-up
Flow visualization was performed using an inverted
microscope (IX71, Olympus, Japan) and a high-speed
camera (Fastcam SA3, Photron, USA). The
microchannels were placed on the microscope
connected to a syringe pump (CetoniNEMESYS
Syringe Pump) to control the fluid flow. The
experimental apparatus is presented in Figure 2.
Figure 2: Experimental set-up used to perform the
experimental microfluidic high-speed video recordings.
Two blood analogue solutions were made by adding
the surfactant Brij L4 to distilled water at a
concentration of 1 wt.%, as presented in (Lima et al.,
2020). The emulsion suspended phase is composed by
spherical micelles with sizes ranging between 10 and
20m. One of the solutions (Analogue 1) was then
filtered by two filters of 10 and 20 μm average pore
size while the other solution (Analogue 2) was just
filtered with the 20 μm filter. Consequently, the fluids
present spherical micelles with size ranging from 10
to 20 μm. Initially three different fluid flows, 60
μL/min, 100 μL/min and 150 μL/min, were compared
and 100 μL/min was set for the remaining
experiments.
2.4 Image Processing
The images from the microscope were obtained using
a high-speed camera using a frame rate of 2000 and
6000 frames/second. The image sequences obtained
were analysed using the minimum intensity feature of
the Z-Project from the ImageJ software (Abràmoff et
al., 2004, Carvalho et al. 2021). The intensity of the
pixels of a selected region was then evaluated and
converted to a grey scale using the plot profile
feature. High values correspond to brighter pixels
which indicate regions with less particle density while
lower values correspond to darker pixels and,
consequently, high particle density or channel walls.
Syringe Pump
High-Speed
Camera
Inverted
Microscope
Microfluidic
Device
Separation Microfluidic Devices Fabricated by Different Milling Processes
187
The grey value was measured transversely to the
entrance and exit of the channel, using the Plot Profile
feature of ImageJ. The average of the values obtained
was then calculated and normalized, considering the
maximum grey value obtained using water, to a scale
of 0 to 1 for result comparison.
3 RESULTS AND DISCUSSION
The grey value was measured for the microchannel
made by the first manufacture process using the
Analogue 1 at three different fluid flows. Preliminary
tests show that a greater variation between the entry
and the exit is noted for 100 μL/min. The variation is
similar for the other two flows, being lower for 60
μL/min (Figure 3).
Figure 3: Variation of average grey value at the entrance
and exit of the microchannel for different fluid flows.
Figure 4: Variation of average grey value at the entrance
and exit of the microchannel for manufacture process 1 at a
100 μL/min fluid flow.
Subsequently, the grey value variation was
measured using both analogue fluids for the
microchannel of the manufacture process 1 (Figure
4), and of the manufacture process 2, (Figure 5). The
variation of the grey value between the entry and exit
on both the microchannels was similar between the
analogues. For the first manufacture process the
variation was higher using the analogue 2, while for
the second process the variation was higher with
Analogue 1.
The influence of the depth of the microchannel
was also investigated. The variation of the grey
intensity on the different microchannels of the
devices of manufacture 1 and 2 is presented in Figures
6 and 7, respectively. In the case of the former, the
microchannels with lower depths showed higher
clearance of cells while in the later the contrary was
verified. Consequently, the tested deepness of the
microchannels do not have a significant impact on the
cell and plasma separation.
Figure 5: Variation of average grey value at the entrance
and exit of the microchannel for manufacture process 2 at a
100 μL/min fluid flow.
Figure 6: Variation of average grey value at the entrance
and exit of the different microchannels of manufacture
process 1 at a 100 μL/min fluid flow.
Figure 7: Variation of average grey value at the entrance
and exit of the different microchannels of manufacture
process 2 at a 100 μL/min fluid flow.
BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices
188
As expected, the grey value increased from the
entry to the exit, indicating a reduction of the particle
concentration throughout the microchannels.
Difference in particle size does not seem to affect the
clearance process since the results were similar using
both of the analogue fluids. On the other hand, the
fluid flow affects the efficiency of the microchannel.
For first manufacture process it was possible to obtain
high grey values corresponding to a reduction of the
particle concentration. The grey values were also near
the value obtained with water indicating a successful
separation of the particles and plasma. Nevertheless,
the detection is dependent on the resolution of the
acquired images. Consequently, the minimum value,
corresponding to the maximum amount of particles,
obtained in each measurement varied. Further
techniques and models for image acquisition and data
treatment need to be improved for a better evaluation
of the effectiveness of the device.
4 CONCLUSIONS
The present study shows the potentiality of a novel
geometry for a microfluidic device for the separation
of blood cells and plasma. The separation of the
microparticles from the base fluid was mainly due to
the successive contractions effect, where the particles
tend to migrate into the center whereas the base fluid
tend to flow towards the walls of the microchannel.
From the two manufacture processes tested, the
micromilling process without bonding showed
promising results with grey values close to the ones
obtained with water. The depth of the channels and
the size of the particles did not show a major influence
on the separation process. On the other hand, the fluid
flow affected the results obtained, being 100 μL/min
the flow that presented better results. Further image
analysis studies and optimization of the geometry to
improve the separation process will be performed in
the near future.
ACKNOWLEDGEMENTS
Authors acknowledge FCT Fundação para a Ciência
e a Tecnologia for partially supporting this
work through projects UIDB/00319/2020,
UIDB/04077/2020, UIDB/00690/2020, UIDB/
04436/2020 and PTDC/EME-SIS/30171/2017
NORTE-01-0145-FEDER-030171, and NORTE-01-
0145-FEDER-029394 funded by COMPETE2020,
NORTE 2020, PORTUGAL 2020, Lisb@2020 and
FEDER and for supporting the PhD Fellowship
(2020.08646.BD).
REFERENCES
Faustino, V.; Rodrigues, R.O.; Pinho, D.; Costa, E.; Santos-
Silva, A.; Miranda, V.; Amaral, J.S.; Lima, R. A (2019),
Microfluidic Deformability Assessment of Pathological
Red Blood Cells Flowing in a Hyperbolic Converging
Microchannel. Micromachines, 10, 645.
Yin, X., Thomas, T., Zhang, J. (2013) Multiple red blood
cell flows through microvascular bifurcations: Cell free
layer, cell trajectory, and hematocrit separation.
Microvasc. Res., 89:4756.
Shamloo, A., Vatankhah, P., Bijarchi, M.A. (2016)
Numerical optimization and inverse study of a
microfluidic device for blood plasma separation. Eur. J.
Mech. B/Fluids, 57: 3139.
Nam, J., Namgung, B., Lim, C.T., Bae, J.E., Leo, H.L.,
Cho, K.S., Kim, S. (2015) Microfluidic device for
sheathless particle focusing and separation using a
viscoelastic fluid. J. Chromatogr. A, 1406:244250.
Lee, M.G., Shin, J.H., Choi, S., Park, J.K. (2014) Enhanced
blood plasma separation by modulation of inertial lift
force. Sensors Actuators, B Chem., 190:311317.
Al-Fandi, M., Al-Rousan, M., Jaradat, M.A.K., Al-Ebbini,
L. (2011) New design for the separation of
microorganisms using microfluidic deterministic lateral
displacement. Robot. Comput. Integr. Manuf., 27:237
244.
Bento, D., Fernandes, C.S. Miranda, J.M., Lima, R. (2019)
In vitro blood flow visualizations and cell-free layer
(CFL) measurements in a microchannel network. Exp.
Therm. Fluid Sci., 109:109847.
Catarino, S. O., R. O. Rodrigues, D. Pinho, J. M. Miranda,
G. Minas, R. Lima. (2019) Blood Cells Separation and
Sorting Techniques of Passive Microfluidic Devices:
From Fabrication to Applications. Micromachines, 10,
593.
Pinho D, Carvalho V, Gonçalves IM, Teixeira S, Lima R.
(2020) Visualization and Measurements of Blood Cells
Flowing in Microfluidic Systems and Blood Rheology:
A Personalized Medicine Perspective Journal of
Personalized Medicine 10, 4, 249.
Nam, J., Yoon, J., Kim, J., Jang, W.S., Lim, C.S. (2019)
Continuous erythrocyte removal and leukocyte
separation from whole blood based on viscoelastic cell
focusing and the margination phenomenon. J.
Chromatogr. A, 1595:230239.
Geng, Z., Ju, Y., Wang, W., Li, Z. (2013) Continuous blood
separation utilizing spiral filtration microchannel with
gradually varied width and micro-pillar array. Sensors
Actuators, B Chem., 180:122129.
Tsutsui, H., Ho, C.M. (2009) Cell separation by non-inertial
force fields in microfluidic systems. Mech. Res.
Commun., 36:92103.
Ookawara, S., Oozeki, N., Ogawa, K., Löb, P., Hessel, V.
(2010) Process intensification of particle separation by
Separation Microfluidic Devices Fabricated by Different Milling Processes
189
lift force in arc microchannel with bifurcation. Chem.
Eng. Process. Process Intensif., 49:697703.
Dominical, V.M., Vital, D.M., O’Dowd, F., Saad, S.T.O.,
Costa, F.F., Conran, N. (2015) In vitro microfluidic
model for the study of vaso-occlusive processes. Exp.
Hematol., 43: 223228.
Xue, X., Patel, M.K., Kersaudy-Kerhoas, M., Desmulliez,
M.P.Y., Bailey, C., Topham, D. (2012) Analysis of
fluid separation in microfluidic T-channels. Appl. Math.
Model., 36:743755.
Lee, M.G., Choi, S., Park, J.K. Inertial separation in a
contraction-expansion array microchannel. J.
Chromatogr. A, 1218:41384143.
Losserand, S., Coupier, G., Podgorski, T. (2019) Migration
velocity of red blood cells in microchannels.
Microvasc. Res., 124:3036.
Lima, R.; Vega, E.J.; Moita, A.S.; Miranda, J.M.; Pinho,
D.; Moreira, A.L.N. (2020) Fast, flexible and low-cost
multiphase blood analogue for biomedical and energy
applications. Experiments in Fluids, 61, 231,
doi:10.1007/s00348-020-03066-7.
Medhi, B.J., Reddy, M.M., Singh, A. (2019) Particle
migration of concentrated suspension flow in
bifurcating channels. Adv. Powder Technol., 30:1897
1909.
Fekete, Z., Nagy, P., Huszka, G., Tolner, F., Pongrácz, A.,
Fürjes, P. (2012) Performance characterization of
micromachined particle separation system based on
Zweifach-Fung effect. Sensors Actuators, B Chem.,
162:8994.
Meyari, M., Salehi, Z., Zarghami, R., Saeedipour, M.
(2020) Numerical investigation of particle separation in
Y-shaped bifurcating microchannels. Particuology,
Corpus ID: 228863362
Lopes R., Rodrigues R. O., Pinho D., Garcia, V., Schutte
H., Lima R., Gassmann S., (2015). Low cost
microfluidic device for partial cell separation:
micromilling approach. in Proceedings of 2015 IEEE
International Conference on Industrial Technology,
Seville, Spain.
Singhal, J., Pinho, D., Lopes, R., Sousa, P., Garcia, V.,
Schütte, H., Lima, R., Gassmann, S. (2016) Blood Flow
Visualization and Measurements in Microfluidic
Devices Fabricated by a Micromilling Technique.
Micro Nanosyst, 7:148153.
Madureira, M., Faustino, V., Schütte, H., Gassmann, S.,
Lima, R. (2018) Blood cells separation in a T-shaped
microchannel fabricated by a micromilling technique.
J. Mech. Eng. Biomech., 2:4349.
Abràmoff, M.D., Magalhães, P.J., Ram, S.J. (2004) Image
processing with imageJ. Biophotonics Int., 11:3641.
Carvalho, V.; Gonçalves, I.M.; Souza, A.; Souza, M.S.;
Bento, D.; Ribeiro, J.E.; Lima, R.; Pinho, D. (2021)
Manual and Automatic Image Analysis Segmentation
Methods for Blood Flow Studies in
Microchannels. Micromachines, 12, 317.
BIODEVICES 2022 - 15th International Conference on Biomedical Electronics and Devices
190