View Article Online / Journal Homepage / Table of Contents for this issue

TECHNICAL NOTE www.rsc.org/loc | Lab on a Chip

Rapid, cost-efficient fabrication of microfluidic reactors in thermoplastic

polymers by combining photolithography and hot embossing†
Jesse Greener,a Wei Li,‡a Judy Ren,a Dan Voicu,a Viktoriya Pakharenko,a Tian Tangb
and Eugenia Kumacheva*a
Received 11th September 2009, Accepted 11th November 2009
First published as an Advance Article on the web 1st December 2009
DOI: 10.1039/b918834g
Published on 01 December 2009. Downloaded by Reprints Desk on 7/24/2018 7:41:13 AM.

We report a cost-efficient and easy to implement process for fabricating microfluidic reactors in
thermoplastic materials. The method includes (i) the fabrication of an imprint template (master), which
consists of a photoresist deposited on a metal plate; (ii) the thermoembossing of the reactor features
into polymer sheets; (iii) the activation of the embossed and planar thermoplastic surfaces; and (iv) the
low-temperature bonding of these surfaces. The generality of the method is established by fabricating
microfluidic reactors with a complex geometry in a range of thermoplastic polymers, including
cycloolefin, polycarbonate, and UV-transparent acrylic polymers and by the multiple, high-fidelity use
of the master.

Introduction transition temperatures, e.g., poly(methyl methacrylate). These

Chemical reactions performed in microfluidic reactors are char-
acterised by efficient mixing, enhanced heat and mass transfer,
protection from contamination, and the ability to perform multi-
step reactions in a continuous mode.1–5 There is a growing
interest in rapid and cost-efficient fabrication of microreactors
with the purpose of high-throughput screening and optimization
of formulations or increased productivity of microfluidic
synthesis.
Hot embossing (HE) is a promising technique for the fabri-
cation of microfluidic reactors with channel dimensions on the
order of tens to hundreds of micrometres and high aspect ratio
features.6–12 This method requires relatively low heating
(in comparison with e.g., injection molding), thereby reducing
residual thermal stress in the fabricated device.11
Commercially available imprint templates (i.e., stamps or
masters) for HE are usually fabricated from metals, e.g., from
nickel, by using mechanical machining,13 laser ablation,14 or
electroforming.15 These methods suffer from high cost and long
production times. Stamps based on patterned silicon wafers are
fragile and prone to breaking, especially with repeated use.16
Similarly, masters produced from silicon wafers coated with
a patterned photoresist are not robust, and they frequently suffer
from poor adhesion of the photoresist to the wafer in the de-
embossing step.17–20 Imprint templates based on epoxy and
polyester resins21,22 have been generated in a multi-step process
by replicating a primary mold (usually, fabricated in poly-
(dimethylsiloxane)) and, to the best of our knowledge, they have
only been used to pattern materials with relatively low glass
drawbacks limit the implementation of HE for the production of
microfluidic reactors.
We overcame the limitations listed above by using photoli-
thography to directly generate inexpensive and robust masters,
which consisted of the two-layer patterned photoresist film
applied to a metal plate. The imprint template was sufficiently
robust to be used to imprint the features of a microfluidic reactor
for the synthesis of polymer particles at least 40 times. No
noticeable degradation or change in pattern fidelity was
observed, important in the scaled up synthesis of polymer
microbeads with narrow size distribution.23 Following the HE
step, a microfluidic reactor was produced by bonding the
patterned thermoplastic sheet with a planar thermoplastic sheet.
The method was utilized to fabricate microfluidic reactors in
a range of thermoplastic materials with the glass transition
temperature, Tg, in the range 113  C # Tg # 149  C using
a standard temperature-controlled hydraulic press.
We selected the following thermoplastic materials for the
fabrication of microreactors by HE: a cycloolefin polymer
(COP), a UV-transparent acrylic polymer, and polycarbonate
(PC). The use of a COP had advantages over the two other
polymers owing to the combination of high transmission in the
UV-Vis, relatively low viscosity at elevated temperatures, low
water absorption, low fluorescence background, and its ability to
activate its surface by O2 plasma treatment.24–27 Therefore,
whereas both the acrylic polymer and PC were successfully used
for the fabrication of microreactors, most of the results reported
in the current paper are presented for the fabrication of micro-
reactors in COP.

a
Fabrication of the imprint template
Department of Chemistry, University of Toronto, 80 St. George Street,
Toronto, Ontario, M5S 3H6, Canada. E-mail: ekumache@chem. Fig. 1 shows the schematic of the fabrication of the master. A
utoronto.ca; Fax: +1 416-978-3576; Tel: +1 416-978-3576
b
seed layer of SU-8 3050 (MicroChem Corp. USA) with a thick-
EV Group, Inc., 7700 S. River Parkway, Tempe, Arizona, 85284, USA
† Electronic supplementary information (ESI) available: Experimental ness of approximately 40 mm was spin-coated on a square copper
details. See DOI: 10.1039/b918834g base sheet with dimensions of 76  76  1 mm. The photoresist
‡ Co-first author was subsequently cross-linked using UV irradiation (MA6 mask

522 | Lab Chip, 2010, 10, 522–524 This journal is ª The Royal Society of Chemistry 2010

Published on 01 December 2009. Downloaded by Reprints Desk on 7/24/2018 7:41:13 AM.
Fig. 1 Fabrication of a Cu/SU-8 HE imprint template (see text).
aligner, SUSS Microtech) and baked at 90  C (Fig. 1a). A second
layer of SU-8 3050 with a thickness of 75–110 mm was spin-
coated on top of the primary layer and soft baked at 95  C for
40 min (Fig. 1b). The hardened SU-8 was then selectively cross-
linked by placing a mask (resolution of 5080 dpi) with micro-
reactor features on top of the second photoresist layer and
exposing the system to UV irradiation (Fig. 1c). After removing
the uncross-linked photoresist with SU-8 Developer (Microchem
Corp.), the remaining photoresist was hardened at 175  C for 2 h
(Fig. 1d) and the imprint template was ready for use.
In the present work, the thermoembossed microfluidic reactor
had the topography of a planar flow-focusing droplet gener-
ator,29 followed by the polymerization compartment (the wavy
downstream channel). Fig. 2 shows the imprint template and the
sealed microfluidic reactor. The height of the channels was
106 mm and the total area of the embossed features was 230 mm2.
Fig. 2 (a) A photograph of the imprint template (master) comprising
a two-layer SU-8 film applied to a copper base plate and the fabricated
microfluidic reactor for the synthesis of polymer particles (b). In (b) the
emulsification compartment (i) and reaction compartment (ii) are high-
lighted with dashed boxes. The scale bar represents 1 cm. To enhance the
visibility of the features of the master in (a) they have been coated with
ink.28 The reactor channels in (b) were filled with a green food dye.
This journal is ª The Royal Society of Chemistry 2010
View Article Online
The details of the embossing and de-embossing processes as well
as the low-temperature bonding of the imprinted and patterned
surfaces are given in ESI†.
Results and discussion
To assess the quality of the imprinting, we focused on the highest
aspect ratio element of the reactor: the orifice region of the
droplet generator.29 Fig. 3a and b show optical microscopy
images (top view) of the master and of the patterned COP sheet,
respectively. The embossed features in the polymer were the
exact inverse of the features of the master used for HE. In Fig. 3a
defocused vertical microgrooves (the result of polishing of the
copper sheet) are seen through the transparent photoresist layer.
The fidelity of the imprinting process was examined by
comparing the dimensions of the orifice on the mask, the master,
and the imprinted pattern in COP. Fig. 3c shows a typical profile
(acquired using optical profilometry) of the cross-section of the
orifice in a master and the COP sheet embossed at Te ¼ 166  C and
de-embossed at Td ¼ 125  C. The side walls in both fabrication steps
had the wall angle of 3.9  0.5 from the normal. The imprinted
orifice had a 3 mm smaller width compared to the master, which led
to the 6% decrease in its volume. We note that the fidelity in the HE
process benefited from the closer match of SU-8 thermal expansion
coefficient (a ¼ 52 ppm K1) to that of the thermoplastic polymers
used (see ESI†), compared to templates made from silicon (a ¼
3.3 ppm K1)30 or Ni (a ¼ 14.4 ppm K1).31
The HE process conducted with the SU-8/Cu master had high
fidelity of imprinting over the course of many cycles. Using
Fig. 3 Optical microscopy images (top view) of the orifice region on the
two-layer SU-8/Cu master (a) and the corresponding patterned COP
sheet (b). The scale bar represents 200 mm. (c) The profile of the cross-
section of the orifice acquired from the master (solid line) and the
patterned COP sheet (dotted line). The profile of the cross-section of the
master is presented upside down, in order to compare it with the channel
cross-section. Vertical dashed lines in (a) and (b) show the locations
where measurements were taken. (d) The profile of the orifice cross-
section on the imprint template before (solid line) and after (dotted line)
40 HE cycles.
Lab Chip, 2010, 10, 522–524 | 523

a single SU-8/Cu master, we imprinted 40 microfluidic reactors in
COP with the average orifice width at the bottom and the top of
44.8 mm  1.0 mm and 58.5 mm  2.1 mm, respectively. The
tapering angle was in the range of 2.5–5.4 , with an average of
3.7  1.3 . The tapering angle depended on the embossing
temperature and pressure and it was minimised for Te $ 155  C
and Pe ¼ 600 kPa. The SU-8/Cu imprint template was durable:
after 40 HE cycles, the features of the template did not noticeably
change (Fig. 3d). We attribute the durability of the imprint
template to four factors minimising stresses in the de-embossing
step: (i) matched thermal expansion coefficients (a) of the master
Published on 01 December 2009. Downloaded by Reprints Desk on 7/24/2018 7:41:13 AM.
and thermoplastic polymers, (ii) smooth template walls, (iii) the
existence of a seed protoresist layer and (iv) strong adhesion of
the photoresist to the copper base. The selection of copper (a ¼
18.0 ppm K1)32 as the base material also made it superior to e.g.,
silicon, owing to the closer match of thermoexpansion properties
of copper and SU-8. Low surface roughness of the walls of the
master minimised friction between the template and the ther-
moplastic polymer during de-embossing (see ESI†). The seed
photoresist layer increased the adhesion of the patterned SU-8
layer to the master base and prevented its peeling during de-
embossing. Furthermore, based on shear force tests, adhesion of
SU-8 3050 to copper is the strongest compared to any other
substrate, including silicon.33
In the next step, we bonded the patterned and planar sheets of
the thermoplastic polymer by exposing them to either vapours of
methylcyclohexane (MCH), or to O2 plasma followed by sealing
via the application of pressure and temperature lower than Tg
(see ESI†).26,27,34 In both cases bonding was strong (e.g., when
MCH activation was used, we were able to achieve flow rates in
excess of 160 mL h1) and the channel dimensions were not
noticeably altered. Finally, we utilized the fabricated devices for
the emulsification of polar and non-polar liquids, namely, an
aqueous solution of N-isopropyl acrylamide, a photoinitiator,
and a crosslinking agent and a mixture of tripropylene glycol
diacrylate (ESI†).
Summary and outlook
We used photolithographically generated imprint templates
(masters) for the HE-based fabrication of microfluidic reactors in
a range of thermoplastic polymers. Our work has two important
implications: fast fabrication of robust, low-cost SU-8/Cu
masters (4 stamps in 6–8 h) and the subsequent high-fidelity
fabrication of thermoplastic microreactors (ca. 2 h per reactor).
The imprint template is durable in multiple imprinting steps.
These features are beneficial in rapid prototyping of microfluidic
reactors. Further improvements of the fabrication process
include automated temperature and pressure control and
vacuum capabilities.
Acknowledgements
The authors acknowledge financial support from NSERC Can-
ada and Ontario Ministry of Research and Innovation. The
524 | Lab Chip, 2010, 10, 522–524
View Article Online
authors thank Professor H. Naguib (University of Toronto) for
the use of a temperature-controlled hydraulic press.
Notes and references
1 H. R. Sahoo, J. G. Kralj and K. F. Jensen, Angew. Chem., Int. Ed.,
2007, 46, 5704–5708.
2 D. M. Ratner, E. R. Murphy, M. Jhunjhunwala, D. A. Snyder,
K. F. Jensen and P. H. Seeberger, Chem. Commun., 2005,
578–580.
3 A. J. deMello, Nature, 2006, 442, 394–402.
4 Microreactors: New Technology for Modern Chemistry, ed. W.
Ehrfeld, V. Hessel and H. L€ owe, Wiley-VCH Verlag Gmbh,
Weinheim, Germany 2000, 1–288.
5 A. Gunther and K. F. Jensen, Lab Chip, 2006, 6, 1487–1503.
6 N. S. Cameron, H. Roberge, T. Veres, S. C. Jakeway and
H. J. Crabtree, Lab Chip, 2006, 6, 936–941.
7 P. W. Leech, J. Micromech. Microeng., 2009, 19, 1–5.
8 Y. Chen and A. Pepin, Electrophoresis, 2001, 22, 187–207.
9 B. D. Gates, Q. B. Xu, J. C. Love, D. B. Wolfe and G. M. Whitesides,
Annu. Rev. Mater. Res., 2004, 34, 339–372.
10 H. Becker and C. Gartner, Electrophoresis, 2000, 21, 12–26.
11 D. G. Yao, P. Nagarajan, L. Li and A. Y. Yi, Polym. Eng. Sci., 2007,
47, 530–539.
12 M. Heckele and W. K. Schomburg, J. Micromech. Microeng., 2004,
14, R1–R14.
13 T. Schaller, L. Bohn, J. Mayer and K. Schubert, Precis. Eng., 1999,
23, 229–235.
14 P. P. Shiu, G. K. Knopf, M. Ostojic and S. Nikumb, J. Micromech.
Microeng., 2008, 18, 025012.
15 T. Shibata, Y. Takahashi, T. Kawashima, T. Kubota, M. Mita,
T. Mineta and E. Makino, Microsyst. Technol., 2008, 14,
1359–1365.
16 T. Nielsen, D. Nilsson, F. Bundgaard, P. Shi, P. Szabo, O. Geschke
and A. Kristensen, J. Vac. Sci. Technol., B, 2004, 22, 1770–1775.
17 J. Zhang, K. L. Tan, G. D. Hong, L. J. Yang and H. Q. Gong,
J. Micromech. Microeng., 2001, 11, 20–26.
18 A. Bhattacharyya and C. M. Klapperich, Anal. Chem., 2006, 78,
788–792.
19 M. B. Esch, S. Kapur, G. Irizarry and V. Genova, Lab Chip, 2003, 3,
121–127.
20 A. Mathur, S. S. Roy, M. Tweedie, S. Mukhopadhyay, S. K. Mitra
and J. A. McLaughlin, Curr. Appl. Phys., 2009, 9, 1199–1202.
21 T. Koerner, L. Brown, R. X. Xie and R. D. Oleschuk, Sens. Actuators,
B, 2005, 107, 632–639.
22 G. S. Fiorini, G. D. M. Jeffries, D. S. W. Lim, C. L. Kuyper and
D. T. Chiu, Lab Chip, 2003, 3, 158–163.
23 W. Li, J. Greener, D. Voicu and E. Kumacheva, Lab Chip, 2009, 9,
2715–2721.
24 M. Yamazaki, J. Mol. Catal. A: Chem., 2004, 213, 81–87.
25 C. G. Koh, W. Tan, M. Q. Zhao, A. J. Ricco and Z. H. Fan, Anal.
Chem., 2003, 75, 4591–4598.
26 C. W. Tsao and D. L. DeVoe, Microfluid. Nanofluid., 2009, 6,
1–16.
27 Hartmut Steffen, K. Schr€ oder, B. Busse, A. Ohl and K. D. Weltmann,
Plasma Processes Polym., 2007, 4, S392–S396.
28 J. W. Jung, Y. J. Lee, M. S. Shin, H. R. Kim and J. H. Kim, Mol.
Cryst. Liq. Cryst., 2007, 470, 183–190.
29 S. L. Anna, N. Bontoux and H. A. Stone, Appl. Phys. Lett., 2003, 82,
364–366.
30 H. Watanabe, N. Yamada and M. Okaji, Int. J. Thermophys., 2004,
25, 221–236.
31 T. G. Kollie, Phys. Rev. B: Solid State, 1977, 16, 4872.
32 T. A. Hahn, J. Appl. Phys., 1970, 41, 5096–5101.
33 http://www.microchem.com/products/pdf/SU-8-Adhesion-Results.pdf.
34 P. Kettner, R. L. Pelzer, T. Glinsner, S. Farrens and D. Lee, J. Phys.
Conf. Ser., 2006, 34, 65–71.
This journal is ª The Royal Society of Chemistry 2010