Fabrication of Microfluidic Channels by Femtosecond Laser
Micromachining and Application in Optofluidics
João M. Maia
1,2
, Vítor A. Amorim
1,2
, D. Alexandre
2,3
and P. V. S. Marques
1,2
1
Department of Physics and Astronomy of Faculty of Sciences, University of Porto,
Rua do Campo Alegre 687, Porto, Portugal
2
CAP – Centre for Applied Photonics, INESC TEC, Rua Dr. Roberto Frias, Porto, Portugal
3
Department of Physics of School of Science and Technology, University of Trás-os-Montes e Alto Douro,
Quinta de Prados 5000-801, Vila Real, Portugal
Keywords: Optofluidics, Femtosecond Laser Direct Writing, Micromachining, Chemical Etching, Fused Silica.
Abstract: Micromachining with femtosecond laser can be exploited to fabricate optical components and microfluidic
channels in fused silica, due to internal modification of the glass properties that is induced by the laser
beam. In this paper, we refer to the formation of microfluidic channels, where an optimization of the
fabrication procedure was conducted by examining etch rate and surface roughness as a function of the
irradiation conditions. Microfluidic channels with high and uniform aspect ratio and with smooth sidewalls
were obtained, and such structures were successfully integrated with optical components. The obtained
results set the foundations towards the development of new optofluidic devices.
1 INTRODUCTION
The concept of lab-on-a-chip (LOC) revolves around
the idea of miniaturizing a complete laboratory into
a hand-sized chip, which allows preparation,
transport, reaction and analysis of reagents, without
loss of efficiency and measurement accuracy
(Dittrich, 1996).
These devices integrate multiple systems, such as
microfluidic channels (for fluid movement),
mechanical elements (valves and mixers for control
of reagent flow and reaction (Xu, 2013, Liu, 2013))
and optical elements (mirrors, lenses and
waveguides for in situ analysis (Hwang, 2009)).
Although conventional fabrication techniques,
such as photolithography and soft lithography,
enabled production of microfluidic devices (Vos,
2007, Wu, 2004), some problems remain. First, both
techniques are planar, thus hindering fabrication of
three-dimensional (3D) structures and limiting
integration of components and device
functionalities. Second, integration of microfluidic
systems with optical layers for on-chip optical
detection has been proven to be demanding and
complex.
In order to overcome these issues, femtosecond
laser micromachining has been adopted for the
production of LOC devices. This non-contact
technology relies on the laser ability to internally
modify the sample properties, due to a non-linear
absorption process (Osellame, 2012). The
modification is confined to the focal volume, hence
submicron resolutions are attainable. Given the
ultrashort duration of the laser pulse, the
modification is not a result of thermal diffusion, thus
formation of heat-affected zones is minimized
(Sugioka, 2014). In the case of picosecond pulses
the resulting structures are formed due to cumulative
thermal effects, which result in erratic and irregular
structures (Corbari, 2013).
For ultrashort pulses, the laser-matter interaction
can be described qualitatively in three steps: (i)
generation of free electron plasma, (ii) followed by
energy relaxation and (iii) modification of the
material properties. When fused silica is exposed to
a femtosecond laser beam, two important structural
modifications may occur, depending on the pulse
energy. For low energies, there is a smooth
modification of the refractive index that can be
exploited to produce optical waveguides, Bragg
grating waveguides, directional couplers, among
other devices (Zeil, 2013, Davis, 1996). For
106
Maia J., Amorim V., Alexandre D. and Marques P.
Fabrication of Microfluidic Channels by Femtosecond Laser Micromachining and Application in Optofluidics.
DOI: 10.5220/0006167401060113
In Proceedings of the 5th International Conference on Photonics, Optics and Laser Technology (PHOTOPTICS 2017), pages 106-113
ISBN: 978-989-758-223-3
Copyright
c
2017 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
moderate energies, birefringent sub-wavelength
nanogratings are formed. These structures are more
selective to an etching reaction, which allows
fabrication of microfluidic channels.
Although these two effects may be combined to
produce optofluidic devices (Bellini, 2010,
Applegate, 2006), some problems regarding reactant
detection are still unsolved. As the size of the
microfluidic systems decrease, so does detection
volume, usually in the range of 10
-9
to 10
-12
litres,
and surface-to-volume ratio, hence surface forces
(surface tension, van der Waals and surface
roughness) become dominant (Stone, 2004).
Therefore, the fabrication procedure of microfluidic
channels needs to be optimized, in order to form
highly sensitive optofluidic systems.
Figure 1: Illustration of the FLICE fabrication technique in
fused silica (Said, 2004).
2 MICROFLUIDIC CHANNEL
FABRICATION PROCEDURE
The microfluidic channels were produced using the
FLICE technique (Femtosecond Laser Irradiation
followed by Chemical Etching), which is depicted in
figure 1.
To machine fused silica, we use the 2
nd
harmonic
(515 nm) of a fiber-amplified laser from Amplitude
Systèmes. The laser has a pulse duration of 250 fs at
repetition rate of 500 kHz and maximum average
energy of 1.1 μJ. A schematic of the laser direct
writing unit is depicted in figure 2.
The first block is a power control unit, which is
followed by a beam expander system formed by two
planar-convex lenses. This expander increases the
beam spot size so that it equals the objective lens
entrance aperture. The beam polarization is
determined by a half-wave plate.
The laser beam is focused inside the sample by
an aspherical objective lens with 40x amplification
and numerical aperture equal to 0.55, which prevents
spherical aberrations.
Figure 2: Laser direct writing unit.
The objective is mounted in a piezoelectric stage
(PI P-725) that moves along Z, while the sample is
mounted in two air-bearing stages (Aerotech
ABL10100-LN) that move along the plane XY. This
way we can write 3D structures, by defining the
design geometry in a software program that controls
stage position, speed and pulse energy.
The channels were written by focusing the beam
inside the glass and by stacking multiple scans side
by side, starting from below (in order to avoid beam
propagation through areas already exposed). This
strategy allows design of different cross-section
geometries to be obtained.
After writing, the sample edges were polished to
enable cross-section characterization by microscopy.
Then, the etching reaction was performed by
immersion of the sample in a 10% hydrofluoric acid
(HF) solution. The reaction was performed in an
ultrasonic bath (Branson 2510 Ultrasonic Cleaner) at
frequency of 40 kHz to facilitate debris removal.
After etching the samples were rinsed in
isopropanol and then in deionized water. The
samples were characterized by direct observation at
an optical microscope and by performing SEM
microscopy.
3 OPTIMIZATION OF
FABRICATION PARAMETERS
The production of microfluidic channels was
optimized by studying two main parameters: etch
rate and surface roughness. The etch rate is
associated with the etching selectivity that should be
high, and with the aspect ratio (length/cross-section)
which should be uniform. The channel should also
present smooth sidewalls (low surface roughness), to
enable light coupling and to minimize optical losses
due to scattering.
Fabrication of Microfluidic Channels by Femtosecond Laser Micromachining and Application in Optofluidics
107
Multiple channels were fabricated while varying
beam polarization (parallel or perpendicular to the
scanning direction), pulse energy (60 nJ to 300 nJ),
scanning speed (100 μm/s to 500 μm/s), scanning
depth (50 μm to 150 μm) and scan separation (1 μm
to 15 μm).
3.1 Etch Rate Calibration
We started by analyzing the etch rate dependence on
the polarization angle. The results, shown in
figure 3, indicate a strong dependence on the
polarization angle, which results from the
nanogratings orientation.
Figure 3: Etch rate vs polarization angle of single-scan
channels fabricated 150 μm below surface at 500 μm/s and
with 80 nJ pulse energy. The figure inset shows the
nanograting orientation for writing with (A) parallel
polarization and (B) perpendicular polarization, alongside
the light propagation direction (k), the writing direction
(S), and the laser electric field vector (E) (Taylor, 2008).
For perpendicular polarization (90°), the etch rate
is maximum, because the nanogratings align with
the channel axis, enabling fast HF diffusion. For
parallel polarization (0° and 180°), the planes are
normal to the writing direction, hence the HF acid
encounters alternate layers of nanogratings and
pristine material, which reduce the etch rate. To take
advantage of a maximum etch rate, we then decided
to write the channels with perpendicular
polarization.
Analyzing the influence of pulse energy and
scanning speed, figure 4, we observed that, within
the tested range, the etch rate stays uniform. We also
determined that at pulse energies lower than 60 nJ,
there is no formation of nanogratings, and hence
there is no etching selectivity. For scanning speeds
higher than 500 μm/s, we expect the etch rate to start
decreasing, due to a lower contrast between
nanogratings and pristine volumes.
Figure 4: Etch rate as a function of the pulse energy and
scanning speed for channels fabricated with perpendicular
polarization and 150 μm below the silica surface.
The etch rate also depends on the separation
between scans. As shown in figure 5, the etch rate
starts increasing as the separation between scans
decreases, as expected.
Figure 5: Etch rate of channels as function of the
separation between scans along Y and Z. The channels
were fabricated with perpendicular polarization, 150 μm
below surface at 500 μm/s and with pulse energy of 80nJ.
For small overlap, the scans do not interact with
each other and the etch rate is similar to the one
obtained for single-scan channels. However,
decreasing the separation results in an increase of
the stress field around the laser-affected zones
(Champion, 2013), which modifies the irradiated
PHOTOPTICS 2017 - 5th International Conference on Photonics, Optics and Laser Technology
108
volume density, and consequently increases the etch
rate (Agarwal, 1997).
3.2 Tapering Effect
For long etching reactions we also observed that, for
writing with perpendicular polarization, the etch rate
decreases over time, figure 6. This phenomenon
occurs because the HF acid cannot reach the
channel’s end as easily.
Figure 6: Etch rate as a function of the etching reaction
time for channels written with perpendicular and parallel
polarization. The channels written with perpendicular
polarization were fabricated 150 μm below surface at
500 μm/s and with pulse energy of 80nJ. The channels
written with parallel polarization were fabricated 150 μm
below surface at 300 μm/s and with pulse energy of
300 nJ. For perpendicular polarization the etch rate
decreases over time, while for parallel polarization it
remains constant.
Alongside this effect, it also occurs lateral
etching, where the HF acid etches the pristine
material surrounding the channel. These two features
contribute to the formation of tapered channels,
figure 7, and affect significantly the obtained aspect
ratio.
Figure 7: Top view image of a tapered single-scan channel
fabricated with perpendicular polarization at a depth of
50 μm, with pulse energy of 60 nJ and scanning speed of
500 μm/s.
In order to prevent this issue, several solutions
can be adopted: (i) use of KOH at 80°C as an etchant
solution (Kiyama, 2009), (ii) design of a structure
that compensates and balances the etching reaction
(Vishnubhatla, 2009), and (iii) fabrication of vertical
holes connecting the channel to the sample surface
(Ho, 2013). Due to simplicity, we used the latter
technique, whose configuration is depicted in
figure 8.
Figure 8: Fabrication layout of vertical holes (Ho, 2013).
By fabricating holes 200 μm apart from each
other, we obtained channels with uniform aspect
ratio, as shown in figure 9, and with length over one
centimetre.
Figure 9: Top view images of microfluidic channels
fabricated buried in fused silica at different times of
etching reaction.
Figure 9 also reveals the time evolution of the
etching reaction. After 30 minutes, we observe that
the vertical holes have been fully etched, and that
the HF acid diffuses preferentially along the channel
axis, due to the favourable nanogratings alignment.
After 60 minutes, the entire channel has been etched,
but its dark colour reveals that some glass fibers are
still present. After 90 minutes the process is
completed.
Fabrication of Microfluidic Channels by Femtosecond Laser Micromachining and Application in Optofluidics
109
3.3 Surface Roughness
Regarding surface roughness, it depends mainly on
two factors: stress field distribution and scan
separation.
The formation of laser-affected zones is
associated with the generation of stress fields around
the irradiated volume. The stress fields induce
birefringence, which can be measured by
transmitting light through the sample.
It is reported that pulse energy, scan separation
and beam polarization affect the accumulated stress
(Champion, 2013). For high pulse energies and for
small separations the accumulated stress increases,
and at extreme conditions (pulse energy above
300 nJ and/or separation of tenths of nanometres)
the stress can be released. This relaxation process
generates cracks that propagate over a few microns
into the surroundings, figure 10.
Figure 10: Cross-section image of a microfluidic channel
presenting cracks. The channel was fabricated with
parallel polarization, 50 μm below surface at 200 μm/s and
with 300nJ pulse energy.
On the other hand, if the laser-affected zones are
separated by tenths of microns, we observe that,
after etching, these volumes do not merge together,
figure 11. By decreasing the scan separation, the
volumes start merging, but the sidewalls are still
strongly corrugated. Therefore, it is necessary to
continue decreasing the scan separation to obtain
smoother sidewalls, while avoiding, at the same
time, crack formation.
Figure 11: Cross-section image of microfluidic channels
fabricated with different scan separations along Y and Z:
1 μm and 15 μm (left) and 1 μm and 7 μm (right),
respectively. Both channels were fabricated with
perpendicular polarization, 150 μm below surface at
500 μm/s and with 80 nJ pulse energy.
3.4 Optimum Conditions
From this study, we found that the optimum
irradiation conditions to obtain long channels with
uniform aspect ratio and smooth sidewalls are: beam
polarization perpendicular to the scanning direction
(to obtain maximum etching selectivity), scanning
speed of 500 μm/s, pulse energy between 60 nJ and
80 nJ (to minimize stress accumulation), scan
separation along Y of 1 μm to 2 μm and along Z of
1 μm to 4 μm and scanning depth between 50 μm
and 150 μm.
Under these conditions, we obtained a maximum
etching selectivity of 140:1.
The surface roughness along Y varied between
100 nm and 200 nm. This value was obtained by
producing the channels at the silica surface and by
measuring the roughness using a profilometer
(Dektak XT from Bruker with stylus with 2 μm
radius and with 3 mg applied force). Regarding
surface roughness along Z we estimate it to be of
tenths of nanometres, given that similar values were
obtained by other groups who produced channels
under identical conditions (Ho, 2013). In addition, it
is expected that the surface roughness is lower along
Z than along Y due to the parallel alignment of the
nanogratings with the channel sidewalls.
Figure 12: Cross-section of a microfluidic channel
fabricated under optimum conditions.
4 ON-CHIP OPTICAL
DETECTION
4.1 Principle of Operation
The results presented in the previous section indicate
that the obtained channels may be integrated with
PHOTOPTICS 2017 - 5th International Conference on Photonics, Optics and Laser Technology
110
optical components for production of optofluidic
devices. In this paper, we present a possible on-chip
optical detection scheme that enables measurement
of refractive index and temperature of the fluid
circulating in the channel.
Figure 13: Optical detection scheme layout.
The proposed scheme, whose layout is shown in
figure 13, consists in placing a Bragg grating
waveguide (BGW1) a few microns from the
microfluidic channel. According to the Bragg law,
equation (1), light being reflected by the BGW
depends on the medium effective index (n

).
λ
2Λn

(1)
If the channel is a few microns from the BGW,
the medium effective index will depend on the fluid
present in the channel. Therefore, by changing the
fluid properties (concentration, temperature) or by
changing the fluid, we obtain a different Bragg
wavelength (λ
) and are able to sense the fluid
properties. A second BGW, which is surrounded
only by pristine, is fabricated with a different
modulation period (Λ) and is designed to serve as
reference, allowing to exclude thermal effects.
4.2 Proof of Concept
We started by fabricating microfluidic channels with
BGWs at increasing distances from the channel.
The channels were 1.1 cm long and 100 μm
below the glass surface (distance between silica
surface and top of the channel), and were fabricated
with perpendicular polarization, at 500 μm/s with
60 nJ pulse energy and with scan separation along Y
and Z equal to 1 μm and 2 μm, respectively. The
Bragg grating waveguides were produced with
parallel polarization, 125 μm below surface at
400 μm/s and with 250 nJ pulse energy. Two
gratings were produced, one to sense the channel
and designed to operate at 1550 nm and another to
operate at 1558 nm and to serve as a reference. After
writing, the glass facets were polished to improve
light coupling. The etching reaction lasted 90
minutes.
After the devices were fabricated, we observed at
an optical microscope that some of the BGWs had
been etched, figure 14, suggesting that a higher
control of the etching reaction is needed.
Figure 14: Top view of the optofluidic device.
Regarding the remaining devices, we
characterized them, by measuring the Bragg
wavelength when the channels were filled with air.
The measurements were performed without
polarization control. The results, shown in table 1,
indicate that the wavelength shift decreases as the
BGW-channel separation increases. Thus, we
conclude that to achieve higher sensitivities the gap
should be, at maximum, 5 μm.
Table 1: Wavelength shift for increasing separation
between BGW and microfluidic channel.
Separation
BGW-channel
3 μm 4 μm 5 μm 7 μ
m
Wavelength shift
(±0.05nm)
0.30 0.23 0.21 0.06
For the gap of 3 μm, we then filled the channel
with water and were able to measure a different
Bragg wavelength, figure 15.
The results obtained highlight one of the
advantages of the FLICE technique: it can be used to
integrate monolithically optical layers with
microfluidic channels.
Fabrication of Microfluidic Channels by Femtosecond Laser Micromachining and Application in Optofluidics
111
Figure 15: Optical spectrum of a BGW placed 3 μm from
the microfluidic channel, when the channel is filled with
air (red) and with water (blue). An initial spectrum
(black), taken before the etching reaction, is also included.
The inset picture shows that different Bragg wavelengths
are obtained for air and water.
5 CONCLUSIONS
In this project, we employed the FLICE technique to
study the fabrication process of microfluidic
channels. From this study, we were able to produce
centimetre long channels, with uniform aspect ratio
and with crack-free and smooth sidewalls.
This optimization enabled monolithic integration
of the channels with optical components. By
producing Bragg grating waveguides a few microns
from the microfluidic channels, we were able to
detect the fluid through evanescent coupling.
These results, although preliminary, are
important for the development of high-quality
optofluidic devices.
ACKNOWLEDGEMENTS
Project "NanoSTIMA: Macro-to-Nano Human
Sensing: Towards Integrated Multimodal Health
Monitoring and Analytics/NORTE-01-0145-
FEDER-000016" is financed by the North
Portugal Regional Operational Programme (NORTE
2020), under the PORTUGAL 2020 Partnership
Agreement, and through the European Regional
Development Fund (ERDF).
REFERENCES
Dittrich, P., Manz, A., 2006. Lab-on-a-chip: microfluidics
in drug discovery. Nature reviews. Drug discovery,
5(3):210–218.
Xu, B., Zhang, Y., Xia, H., Dong, W., Ding, H., Sun, H.,
2013. Fabrication and multifunction integration of
microfluidic chips by femtosecond laser direct writing.
Lab on a chip, 13:1677–1690.
Liu, K., Yang, Q., He, S., Chen, F., Zhao, Y., Fan, X., Li,
L., Shan, C., Bian, H., 2013. A high-efficiency three-
dimensional helical micromixer in fused silica.
Microsystem Technologies, 19(7):1033–1040.
Hwang, D., Kim, M., Hiromatsu, K., Jeon, H.,
Grigoropoulos, C., 2009. Three-dimensional opto-
fluidic devices fabricated by ultrashort laser pulses for
high throughput single cell detection and processing.
Applied Physics A: Materials Science and Processing,
96(2):385–390.
Vos, K., Bartolozzi, I., Schacht, E., Bienstman, P., Baets,
R., 2007. Silicon-on-Insulator microring resonator for
sensitive and label-free biosensing. Optics Express,
15(12):7610–7615.
Wu, H., Wheeler, A., Zare, R., 2004. Chemical cytometry
on a picoliterscale integrated microfluidic chip.
Proceedings of the National Academy of Sciences of
the United States of America, 101(35).
Osellame, R., Cerullo, G., Ramponi, R., 2012.
Femtosecond Laser Micromachining. Springer, 1
st
edition.
Sugioka, K., Cheng, Y., 2014. Femtosecond Laser 3D
Micromachining for Microfluidic and Optofluidic
Applications, Springer, 1
st
edition.
Corbari, C., Champion, A., Gecevicius, M., Beresna, M.,
Bellouard, Y., Kazansky, P., 2013. Femtosecond
versus picosecond laser machining of nano-gratings
and micro-channels in silica glass. Optics Express,
21(4).
Zeil, P., Voigtländer, C., Thomas, J., Richter, D., Nolte,
S., 2013. Femtosecond laser-induced apodized Bragg
grating waveguides. Optics Letters, 38(13):2354–6.
Davis, K., Miura, K., Sugimoto, N., Hirao, K., 1996.
Writing waveguides in glass with a femtosecond laser.
Optics Letters, 21(21):1729.
Bellini, N., Vishnubhatla, K., Bragheri, F., Ferrara, L.,
Minzioni, P., Ramponi, R., Cristiani, I., Osellame, R.,
2010. Femtosecond laser fabricated monolithic chip
for optical trapping and stretching of single cells.
Optics Express, 18(5):4679–4688.
Applegate, R., Squier, J., Vestad, T., Oakey, J., Marr, D.,
Bado, P., Dugan, M., Said, A., 2006. Microfluidic
sorting system based on optical waveguide integration
and diode laser bar trapping. Lab on a chip, 6(3):422–
426.
Stone, H., Stroock, A., Ajdari, A., 2004. Engineering
Flows in Small Devices: Microfluidics Toward a Lab-
on-a-Chip. Annual Review of Fluid Mechanics,
36(1):381–411.
Said, A., Dugan, M., Bado, P., Bellouard, Y., Scott, A.,
Mabesa, J., 2004. Manufacturing by laser direct-write
PHOTOPTICS 2017 - 5th International Conference on Photonics, Optics and Laser Technology
112
of three-dimensional devices containing optical and
microfluidic networks. Proceedings of SPIE,
5339:194–204.
Taylor, R., Hnatovsky, C., Simova, E., 2008. Applications
of femtosecond laser induced self-organized planar
nanocracks inside fused silica glass. Laser and
Photonics Reviews, 2(1-2):26–46.
Champion, A., Beresna, M., Kazansky, P., Bellouard, Y.,
2013. Stress distribution around femtosecond laser
affected zones: effect of nanogratings orientation.
Optics express, 21(21):24942–51.
Agarwal, A., Tomozawa, M., 1997. Correlation of silica
glass properties with the infrared spectra. Journal of
Non-Crystalline Solids, 209(1-2):166–174.
Kiyama, S., Matsuo, S., Hashimoto, S., Morihira, Y.,
2009. Examination of etching agent and etching
mechanism on femotosecond laser microfabrication of
channels inside vitreous silica substrates. Journal of
Physical Chemistry C, 113(27):11560–11566.
Vishnubhatla, K., Bellini. N., Ramponi, R., Cerullo, G.,
Osellame, R., 2009. Shape control of microchannels
fabricated in fused silica by femtosecond laser
irradiation and chemical etching. Optics Express,
17(10):8685–8695.
Ho, S., 2013. Femtosecond Laser Microfabrication of
Optofluidic Lab-on-a-Chip with Selective Chemical
Etching. PhD thesis, University of Toronto.
Fabrication of Microfluidic Channels by Femtosecond Laser Micromachining and Application in Optofluidics
113