micromachines

Article
Micromachining Microchannels on Cyclic Olefin
Copolymer (COC) Substrates with the Taguchi Method
Pin-Chuan Chen 1, *, Ren-Hao Zhang 1 , Yingyot Aue-u-lan 2 and Guo-En Chang 3
1 Department of Mechanical Engineering, National Taiwan University of Science and Technology,
Taipei 10607, Taiwan; [email protected]
2 Materials and Production Engineering Program, The Sirindhorn International Thai-German Graduate
School of Engineering (TGGS), King Mongkut’s University of Technology North Bangkok, Bangkok 10800,
Thailand; [email protected]
3 Department of Mechanical Engineering, and Advanced Institute of Manufacturing with High-Tech
Innovations (AIM-HI), National Chung Cheng University, Chia-Yi 62102, Taiwan; [email protected]
* Correspondence: [email protected]; Tel.: +886-2-27376456; Fax: +886-2-27376460

Received: 19 July 2017; Accepted: 23 August 2017; Published: 29 August 2017

Abstract: Micromilling is a straightforward approach to the manufacture of polymer microfluidic

devices for applications in chemistry and biology. This fabrication process reduces costs, provides a
relatively simple user interface, and enables the fabrication of complex structures, which makes it
ideal for the development of prototypes. In this study, we investigated the influence of micromilling
parameters on the surface roughness of a cyclic olefin copolymer (COC) substrate. We then employed
factor analysis to determine the optimal cutting conditions. The parameters used in all experiments
were the spindle speed, the feed rate, and the depth of cut. Roughness was measured using a stylus
profilometer. The lowest roughness was 0.173 µm at a spindle speed of 20,000 rpm, feed rate of
300 mm/min, and cut depth of 20 µm. Factor analysis revealed that the feed rate has the greatest
impact on surface quality, whereas the depth of cut has the least impact.

Keywords: cyclic olefin copolymer (COC); micromilling; rapid prototyping

1. Introduction
Microfluidics have been developing since the advent of micro gas chromatography in 1979 [1].
These developments have made it possible to conduct bio/chemical reactions on a small platform to
lower reagent demand, accelerate reaction rates, minimize labor, reduce contamination, and enable
integration with other functional components. These so-called micro total analysis systems (µTAS) are
used in genetic analysis, clinical testing, drug discovery, food control, and environmental monitoring.
Several low-cost fabrication methods have been reported to fabricate microfluidic platforms for
bio-applications. For example, Pinto et al. [2] reported a method to fabricate a polydimethylsiloxane
(PDMS) microfluidic chip, which used a cutting plotter to create patterns on adhesive papers as
molds, followed by a PDMS casting and sealing process, to realize microfluidic chips for experiments.
Liu et al. [3] reported a method for PDMS microfluidic chips, which used a laser to write patterns
directly on photoresist as molds for a subsequent PDMS casting process. This article not only discussed
this fabrication method, but also compared the cost, advantages, and drawbacks of different fabrication
methods. Pinto et al. [4] used equipment for printed circuit board (PCB) industries instead of a
cleanroom facility to create well-defined SU-8 microstructures with a minimum resolution of 10 µm
and an aspect ratio of 20. Jang et al. [5] reported a method to fabricate glass microfluidic device for blood
plasma separation, in which they used multiple replication processes to create glass microfluidic chips.
Compared to other methods used for fabricating glass microfluidic chips, this method is relatively

Micromachines 2017, 8, 264; doi:10.3390/mi8090264 www.mdpi.com/journal/micromachines

Micromachines 2017, 8, 264 2 of 9

low-cost. To better understand the advantages and drawbacks of the reported low-cost fabrication
methods for microfluidic chips, an article published by Faustino et al. [6] is highly recommended.
Micromachining is a time-efficient, low-cost approach to the manufacture of polymer microfluidic
devices [7]. Unlike etching [8], lithography, electroplating, molding (Lithographie, Galvanoformung,
Abformung, LIGA) [9,10], and PDMS casting [11], micromilling can be used on a wide range of
materials to fabricate complex multi-level microstructures. Micromilling involves the mechanical
removal of substrate material; therefore, the operating parameters, such as spindle speed, depth of cut,
feed rate, and working environments, can greatly affect the surface quality of the resulting micromilled
substrate. Growing interest in polymer microfluidic devices is driving the need for new methods used
to prototype polymer microfluidic chips. Polymers are preferred for disposable microfluidic devices,
due to their low cost, wide range of materials, and the maturity of manufacturing methods, such as
injection molding.
Two methods have been proposed for the prototyping of polymer microfluidic devices using a
micromilling machine: (1) milling a mold insert on metal substrates followed by hot embossing on the
polymeric material, and (2) the direct milling of microchannels on polymer substrates. The productivity
of the second approach tends to be somewhat lower; however, it is convenient for concept validation
during the early stages of development. Direct milling involves only four steps to manufacture a
device for testing: (1) design in computer-aided design (CAD), (2) conversion of the CAD file into
G-code for the micromilling controller, (3) micromilling, and (4) bonding. The entire process generally
takes less than 4 h to complete a ready-to-use chip for testing.
Computer numerical controlled (CNC) machines, such as lathes and mills, are commonly used in
the manufacture of polymer microfluidic devices for chemical applications [12,13]. Researchers have
sought to overcome the low accuracy, high surface roughness, and round corner features [14,15] by
improving the manufacturing process [13] or using micromachining [16,17]. Many researchers have
focused on micromilling a metal mold insert with a smooth surface quality for injection molding or
hot embossing [18,19]. A critical concern of surface roughness in microfluidics is that a high surface
roughness would influence the streamline in a small microchannel or affect the microfluidic device’s
performance, especially in those cases which require surface force such as electrophoresis. For example,
Hupert et al. compared the DNA separation efficiency on two types of microfluidic chip; one that
was fabricated with the LIGA process and another that was fabricated with the micromilling process.
They concluded that a smooth micromilled microchannel would not significantly affect the DNA
separation performance [16].
Cyclic olefin copolymer (COC) is widely used in the manufacture of compact-disks and glasses.
COC has been used in the manufacture of microfluidic devices for a variety of applications, such as
polymerase chain reaction (PCR) [20], OLEDs [21], and biological microelectromechanical systems
(BioMEMS) [22]. Spindle speed, feed rate, and depth of cut are the primary cutting parameters
associated with micromilling; however, the final surface quality ultimately depends on the final cut.
In this study, we examined the depth of cut at 10 µm, 15 µm, and 20 µm. We disregarded the effects of
temperature and tool wear due to the softness of the polymer and the shallow cut depth. During the
cutting process, we used compressed air to cool and clean the substrate. Our aim was to elucidate the
effects of each cutting parameter on the surface roughness of a micromilled COC substrate. We then
employed factor analysis to determine the optimal cutting parameters [23].

2. Experiment Design and Procedure

In this study, we used the micromilling machine in Figure 1. This device comprises five major
components, including a micromilling spindle (E3000c, Nakanishi, Kanuma-Shi, Japan), a laser
non-contact tool setting system (NC4, Renishaw, Wotton-under-Edge, UK), a numerical controller
(M515i, LNC Technology Co. Ltd., Taichung, Taiwan), a compressed air/oil coolant system, and a
bit holder for tool exchange. To minimize uncertainty, the COC substrates used in this study were
self-manufactured by Professor Chang of National Chung Cheng University. The micromilling bit used
Micromachines
Micromachines2017,
2017,8,8,264
x FOR PEER REVIEW 33of
of99
Micromachines 2017, 8, 264 3 of 9
in
inthe
theexperiment
experimentwas wasaadouble-flute
double-fluteendmill endmillwith withaadiameter
diameterof of200
200μm μm(Taiwan
(TaiwanMicrodrill
MicrodrillCo. Co.Ltd.,
Ltd.,
Taipei, Taiwan).
Taipei, Taiwan).
Micromachines 2017, 8, 264 3 of 9
in the
Theexperiment was a double-flute endmill with a diameter ofmeasured
200 µm (Taiwan Microdrill Co. Ltd.,
The roughness
roughness of of the
the micromilled
micromilled COC COC substrates
substrates was was measured at at four
four points
points inside
inside the the
in Taiwan).
Taipei, the experiment was a double-flute endmill with a diameter of 200 μm (Taiwan Microdrill Co. Ltd.,
reservoir
reservoir using
using a stylus profilometer (Hommel Werke T400 & P2000
a stylus profilometer (Hommel Werke T400 & P2000 Pick-up TKL 300, Jenoptik, Jena, Pick-up TKL 300, Jenoptik, Jena,
Taipei,
The Taiwan).
roughness of the micromilled COC substrates was measured at four points inside the reservoir
Germany)
Germany)The with
with aaresolution
resolutionof theofmicromilled
of 0.01
0.01μm μm(Figure
(Figure
COC 2).
2). Figure
Figure33was shows
shows the
the18 reservoirs
18atreservoirs micromilled
micromilled
inside the on a
on a
roughness substrates measured four points
using
single a stylus profilometer (Hommel Werke T400 & P2000 Pick-up TKL 300, Jenoptik, Jena, Germany)
substrate
singlereservoir
substrateusingusing
using various
various
a stylus cutting
profilometer parameters.
cutting(Hommel
parameters.Werke The
The reservoirs
T400reservoirs are
are 88 mm
& P2000 Pick-up TKLin
mm length
in300,
length and
Jenoptik, 33 mm
andJena, mm in in
with a and
width, resolution of 0.01 µm (Figure 2). Figure 3 shows the 18 final
reservoirs micromilled on a single
width,Germany)the depth
and thewith depth is the
is the sum
a resolution sum of
of 0.01 0.1
of 0.1 mm
μmmm plus
(Figure the
plus2).the depth
depth
Figure of
3 showsthe
of thethefinal cut (10
18 reservoirs μm,
cut (10 μm, 15
15 μm, oron20aμm).
μm,
micromilled or 20 μm).
substrate
We singleusing various cutting parameters. The reservoirs are 8aremm inwall
length andand 3 mm induring
width,
We designed
designed aa wide
substrate reservoir
wideusingreservoir
various to avoid
avoid damage
tocutting damage to
parameters. the
the stylus
to The stylus from
reservoirs from the the
8 mm in of
wall of the
the reservoir
length reservoir
3 mm during
in
and the depth
measurement. is the sum of 0.1 mm plus the depth of the final cut (10 µm, 15 µm, or 20 µm). We designed
measurement.
width, and the depth is the sum of 0.1 mm plus the depth of the final cut (10 μm, 15 μm, or 20 μm).
a wideWe reservoir
designed to wide
avoidreservoir
damage totoavoid
thevarious
stylus from thestylus
wall of theto reservoir
wall of during measurement.
As shown
As shown ininaTable
Table 1, 1, wewe adopted
adopted damage
various to the
parameter
parameter from
levels
levels the
to elucidate
elucidate thethe
the reservoir
impactduring
impact of spindle
of spindle
speed As shown
measurement. in Table 1, we adopted various parameter levels to elucidate the impact of spindle
speed (N),
(N), depth
depth of of cut
cut (DOC),
(DOC), and and feedfeed rate
rate (F)
(F) onon the
the micromilled
micromilled surface. surface. The The three
three parameters
parameters
speed (N), Asdepth
shownof incutTable 1, weand
(DOC), adopted various
rate (F)parameter
feedsurface levels to elucidate
on the micromilled surface. theTheimpact of spindle
three parameters
studiedherein
studied hereinare aredirectly
directlyrelated
relatedto tothe
the surfaceroughness
roughnessbecause becausethey theycontrolled
controlledthe theremoval
removalrate rate
speed
studied (N),
herein depth of
are directly cut (DOC), and feed rate (F) on the micromilled surface. The three parameters
ofmaterial
of material per
per flute.
flute. Table
Table 2related
2 lists
lists theto experiment
the
the surface roughness
experiment conditions
conditions because
and
and the
the they controlled results
corresponding
corresponding the removal
results based
based rate
on
on
of studiedper
material herein are directly
flute. Table 2 related
lists to the
the surface roughness
experiment conditions becauseand they
the controlled
corresponding the removal
results ratebased
the various
the various levels
levels
of material per of
of three
three
flute. factors.
2 lists In
factors.
Table theall
In all experiments,
experiments,
experiment the step-over
the
conditions
step-over between each
between
and the corresponding
each milling
milling
results based path
path
on was
was
on the various levels of three factors. In allbit. experiments, the step-over between eachamilling path
20% of
20% of the
the the diameter
diameter
various levels of ofthree
of the factors.
the micromilling
micromilling bit. Each
Each the
In all experiments, reservoir
reservoir
step-over was
was inspected
inspected
between using path
using
each milling a measuring
measuring
was
was 20% of to theensure
diameter of the micromilling bit. Each reservoir wasmilled
inspected using aaa measuring
microscope
20% of to
microscope theensure
diameter thatofno
that nothe significant
significant
micromilling scratching
scratching occurred
bit. Eachoccurred
reservoiron on
was the
the milled surface
inspected surface
using aas as result of
result
measuring of aa
microscope to ensure that no significant scratching occurred on the milled surface as a result of a
broken
broken micromilling
micromilling
microscope bit. that no significant scratching occurred on the milled surface as a result of a
bit.
to ensure
broken micromilling
broken micromilling bit.bit.
Micromilling Spindle and Compressed Air

Laser Tool Setting System Manufacturing Platform

Figure
Figure 1. Micromillingsystem
1. Micromilling systemused
used in
in this
thisstudy:
study:five major
five components.
major components.
Figure
Figure1.1.Micromilling
Micromillingsystem
systemused
usedin
inthis
thisstudy:
study:five
fivemajor
majorcomponents.
components.

Figure 2. Stylus profilometer used in this study: (a) stylus profilometer; (b) measurement platform.

Figure
Figure2.2.
2.Stylus
Stylusprofilometer used
usedin
inthis
thisstudy:
study:(a)
(a)stylus
stylusprofilometer;
profilometer;(b)
(b)measurement
measurementplatform.
Figure Stylus profilometer
profilometer used in this study: (a) stylus profilometer; (b) measurement platform.
platform.

(a)
Figure 3. Dimensions of test sample in experiments: (a) top view; (b) side view.
(b)

Figure
Figure3.
Figure 3.Dimensions
3. Dimensionsof
Dimensions oftest
of testsample
test samplein
sample inexperiments:
in experiments:(a)
experiments: (a)top
(a) topview;
top view;(b)
view; (b)side
(b) sideview.
side view.
view.
Micromachines 2017, 8, 264 4 of 9

Table 1. Three levels of three factors examined in this study: spindle speed, feed rate, and depth of cut.

Factor Level 1 Level 2 Level 3

Spindle speed (N) (mm/min) 10,000 15,000 20,000
Feed rate (F) (mm/min) 300 600 800
Depth of cut (DOC) (µm) 10 15 20

Table 2. Experimental results based on the factors and levels listed in Table 1 (N: Spindle speed; F:
Feed rate; DOC: Depth of cut; Average Ra : Average roughness; S: Standard deviation; S/N ratio:
Signal-to-noise ratio).

No. N F DOC Average Ra S S/N Ratio

1 10,000 300 10 0.317 0.021 9.975
2 10,000 600 15 0.337 0.021 9.445
3 10,000 800 20 0.353 0.021 9.026
4 15,000 300 10 0.247 0.006 12.156
5 15,000 600 15 0.267 0.006 11.479
6 15,000 800 20 0.323 0.012 9.803
7 20,000 300 10 0.223 0.006 13.019
8 20,000 600 15 0.297 0.015 10.547
9 20,000 800 20 0.310 0.000 10.173
10 10,000 300 15 0.300 0.026 10.435
11 10,000 600 20 0.343 0.006 9.285
12 10,000 800 10 0.357 0.006 8.954
13 15,000 300 15 0.227 0.006 12.890
14 15,000 600 20 0.290 0.035 10.711
15 15,000 800 10 0.287 0.006 10.851
16 20,000 300 15 0.223 0.006 13.019
17 20,000 600 20 0.287 0.023 10.834
18 20,000 800 10 0.300 0.017 10.448
19 10,000 300 20 0.240 0.010 12.391
20 10,000 600 10 0.307 0.006 10.266
21 10,000 800 15 0.337 0.023 9.442
22 15,000 300 20 0.217 0.006 13.282
23 15,000 600 10 0.273 0.006 11.265
24 15,000 800 15 0.307 0.006 10.266
25 20,000 300 20 0.173 0.006 15.219
26 20,000 600 10 0.223 0.012 13.013
27 20,000 800 15 0.250 0.010 12.037

3. Results and Discussion

3.1. Experimental Result Analysis

Figure 4 shows the measured roughness values based on the cutting parameters listed in Table 2.
All roughness values fell between 0.173 µm and 0.357 µm. Reducing the feed per flute (by increasing
the spindle speed, reducing the feed rate, or reducing the depth of cut) was expected to reduce
roughness; however, this did not prove to be true in the experimental results. This can be explained by
the fact that the DOC in this experiment was on the same order as the edge radius of the micromilling
bit (5 µm), which may have led to the incomplete removal of the COC substrate during the cutting
process [2]. The results of the microscope analysis also revealed chips stuck to the micromilling bit,
which may have affected the cutting process.
Micromachines 2017, 8, 264 5 of 9

Micromachines 2017, 8, 264 5 of 9

Figure 4. Roughness measurements obtained under various cutting conditions: Depth of cut (10 μm,
Figure 4. Roughness measurements obtained under various cutting conditions: Depth of cut (10 µm,
15 μm, 20 μm), feed rate (300 mm/min, 600 mm/min, 800 mm/min), and spindle speed (10,000 rpm,
15 µm, 20 µm), feed rate (300 mm/min, 600 mm/min, 800 mm/min), and spindle speed (10,000 rpm,
15,000 rpm, 20,000 rpm): (a) measured surface roughness at spindle speed of 10,000 rpm; (b) measured
15,000 rpm, 20,000 rpm): (a) measured surface roughness at spindle speed of 10,000 rpm; (b) measured
surface roughness at spindle speed of 15,000 rpm; (c) measured surface roughness at spindle speed of
surface roughness
20,000 rpm. at spindle speed of 15,000 rpm; (c) measured surface roughness at spindle speed of
20,000 rpm.
3.2. Factor Analysis
3.2. FactorFactor
Analysis
analysis was used to identify the key cutting parameters involved in micromilling a
microchannel
Factor analysisdirectly
wason a COC
used to substrate. Table
identify the key1 lists the three
cutting controllinginvolved
parameters factors with
in three levels
micromilling a
each, which resulted in 27 combinations of cutting parameters. Table 2 lists the standard deviation
microchannel directly on a COC substrate. Table 1 lists the three controlling factors with three levels
(S) and signal-to-noise ratio (S/N) at four measurement points. Equation (1) [14] was used to calculate
each, which resulted in 27 combinations of cutting parameters. Table 2 lists the standard deviation (S)
the signal-to-noise ratio (S/N), where n = 4 for four-time roughness measurements and yi2 is the sum
and signal-to-noise ratio (S/N) at four measurement points. Equation (1) [14] was used
of the four-time roughness measurements. Table 3 lists the average S/N ratio at each level to elucidate to calculate the
signal-to-noise ratio (S/N), where n = 4 for four-time roughness measurements and y 2 is the sum of
the impact of each factor (at different levels) on the surface roughness. Figure 5 shows the i average
the four-time
S/N ratiosroughness
in Table 3 asmeasurements.
well as the averageTableroughness
3 lists the(Raverage
a) values S/N ratio at each
corresponding level
to the S/Ntoratios.
elucidate
the impact of each
These results factor
show that(at
thedifferent levels) onisthe
average roughness surface
inversely roughness.toFigure
proportional 5 shows
the average the average
S/N ratio. In
S/N Table
ratios3,inthe range
Table of well
3 as each as
factor
the is definedroughness
average as the difference between
(Ra ) values the highest and
corresponding the lowest
to the S/N ratios.
Theseaverage
resultsS/N ratios.
show A the
that larger range indicates
average roughness thatisthe corresponding
inversely factor has
proportional to athe
more pronounced
average S/N ratio.
effect on the surface quality of the micromilled COC substrate. Factor B presents
In Table 3, the range of each factor is defined as the difference between the highest and the the largest rangelowest
average S/N ratios. A larger range indicates that the corresponding factor has a more pronounced
effect on the surface quality of the micromilled COC substrate. Factor B presents the largest range
Micromachines 2017, 8, 264 6 of 9

(2.376), indicating that the feed rate has the greatest impact on surface quality. Factor C presents the
smallest range (0.129),
Micromachines indicating that the DOC has the least impact on surface quality.
2017, 8, 264 6 of 9
 
1
(2.376), indicating that the feed rate has the greatest impact
S/N = −10 log (∑ yon 2 surface quality. Factor C presents the
i ) (1)
smallest range (0.129), indicating that the DOC has thenleast impact on surface quality.
Figure 5a–c show measured roughness in terms 1 of feed rate and depth of cut based on three
S/N 10log (1)
spindle speeds. Based on the criteria of low roughness and higher S/N ratio, Figure 5 can be used
to identifyFigure 5a–ccutting
the best show measured
parameters roughness in termsthe
to minimize of roughness
feed rate and
ofdepth of cut based
a micromilled COCon three
substrate.
spindle speeds. Based on the criteria of low roughness and higher S/N ratio, Figure
The combination that resulted in the lowest roughness values (with a roughness of 0.173 µm and 5 can be used to an
identify the best cutting parameters to minimize the roughness of a micromilled COC
S/N ratio of 15.219) was No 27 in the Table 2, which corresponds to a spindle speed of 20,000 rpm,substrate. The
a feedcombination
rate of 300that resulted in the lowest roughness values (with a roughness of 0.173 μm and an S/N
mm/min, and a depth of cut of 20 µm.
ratio of 15.219) was No 27 in the Table 2, which corresponds to a spindle speed of 20,000 rpm, a feed
rate of 300 mm/min, and a depth of cut of 20 μm.
Table 3. Factor analysis results identifying key cutting parameters.
Table 3. Factor analysis results identifying key cutting parameters.
A B C
Levels and Response
(Spindle A
Speed) (FeedBRate) C (Depth of Cut)
Levels and Response
(Spindle Speed) (Feed Rate) (Depth of Cut)
Level 1 9.913 12.487 11.105
Level 1 9.913 12.487 11.105
Level 2 11.412 10.760 11.062
Level 2 11.412 10.760 11.062
Level 3 12.034 10.111 11.192
Range Level 3 2.121 12.034 10.111
2.376 11.192 0.129
Rank Range 2 2.121 2.376
1 0.129 3
Rank 2 1 3

Figure 5. Calculated S/N ratio and measured roughness values associated with three major cutting
Figure 5. Calculated S/N ratio and measured roughness values associated with three major cutting
parameters at three levels: (a) spindle speed, (b) feed rate, (c) depth of cut.
parameters at three levels: (a) spindle speed, (b) feed rate, (c) depth of cut.
Micromachines 2017, 8, 264 7 of 9

3.3. Confirmation Run

Experimental results revealed that the lowest roughness values were obtained from a spindle
speed of 20,000 rpm, a feed rate of 300 mm/min, and a depth of cut of 20 µm. However, it is likely
that these parameters are not directly applicable to other micromilling machines. Furthermore, several
factors that are generally disregarded in conventional machining (substrate grain size and tool edge
geometry), may actually play a dominant role in determining the surface quality when machining
at the micro-scale [2]. To confirm our analysis, we micromilled 10 reservoirs on COC substrates
using the optimal parameter settings identified in this study, the results of which are listed in Table 4.
We employed the same measurement method to obtain values at four randomly selected locations in
each reservoir for measurement using a stylus profilometer. The original roughness values obtained
using the optimal cutting parameters was 0.173 µm; however, the average roughness in Table 4 was
0.183 µm, indicating a standard deviation of 0.02 µm.

Table 4. Measurement results of 10 samples with the best cutting parameters as a confirmation run.

No. 1 2 3 Average (Ra ) µm

1 0.14 0.14 0.15 0.143
2 0.18 0.19 0.18 0.183
3 0.20 0.20 0.20 0.200
4 0.15 0.14 0.15 0.147
5 0.18 0.20 0.19 0.190
6 0.21 0.20 0.20 0.203
7 0.17 0.17 0.18 0.173
8 0.19 0.19 0.18 0.187
9 0.19 0.20 0.20 0.197
10 0.20 0.22 0.21 0.210

4. Conclusions
Micromilling is a useful tool for the rapid prototyping of polymer microfluidic devices, particularly
during the initial stages of development. This approach entails far lower costs and far less time than
micromilling a metal mold insert followed by hot embossing a polymer microfluidic device. In this
study, we sought to identify the optimal parameters for micromilling a COC substrate, with the aim of
minimizing surface roughness. We focused on three parameters: spindle speed, feed rate, and depth
of cut. Using 27 parameter combinations, measured roughness values fell between 0.173 µm and
0.357 µm. Factor analysis revealed that the feed rate has the greatest impact on surface roughness of
a micromilled COC substrate, whereas the depth of cut has the least impact. The lowest roughness
values in this study (0.173 µm) were obtained using a feed rate of 300 mm/min, a spindle speed of
20,000 rpm, and a depth of cut of 20 µm. The difference in roughness values between these two sets of
cutting parameters falls within the resolution of the profilometer.

Acknowledgments: This work was funded by the Ministry of Science and Technology (MOST 106-2218-E-011-010)
and the Mechanical Engineering Department of National Taiwan University of Science and Technology (NTUST).
Author Contributions: This research is discussed between Pin-Chuan Chen and Yingyot Aue-u-lan, the purpose
of which is to understand the surface roughness of COC substrate after a micromilling process. Ren-Hao Zhang
conducted all experiments. To ensure the identical material properties of all COC substrates, Guo-En Chang’s lab
used an injection molding machine to fabricate COC films for the experiments.
Conflicts of Interest: The authors declare no conflict of interest.
Micromachines 2017, 8, 264 8 of 9

References
1. Terry, S.C.; Jerman, J.H.; Angell, J.B. A gas chromatographic air analyzer fabricated on a silicon wafer.
IEEE Trans. Electron Devices 1979, 26, 1880–1886. [CrossRef]
2. Pinto, E.; Faustino, V.; Rodrigues, R.O.; Pinho, D.; Garcia, V.; Miranda, J.M.; Lima, R. A Rapid and Low-Cost
Nonlithographic Method to Fabricate. Micromachines 2015, 6, 121–135. [CrossRef]
3. Liu, Z.H.; Xu, W.C.; Hou, Z.N.; Wu, Z.G. A Rapid Prototyping Technique for Microfluidics with High
Robustness and Flexibility. Micromachines 2016, 7, 201. [CrossRef]
4. Pinto, V.C.; Sousa, P.J.; Cardoso, V.F.; Minas, G. Optimized SU-8 Processing for Low-Cost Microstructures.
Micromachines 2014, 5, 738–755. [CrossRef]
5. Jang, H.; Haq, M.R.; Ju, J.; Kim, Y.; Kim, S.M.; Lim, J. Fabrication of All Glass Bifurcation Microfluidic Chip.
Micromachines 2017, 8, 67. [CrossRef]
6. Faustino, V.; Catarino, S.O.; Lima, R.; Minas, G. Biomedical microfluidic devices by using low-cost fabrication
techniques: A review. J. Biomech. 2016, 49, 2280–2292. [CrossRef] [PubMed]
7. Dornfeld, D.; Min, S.; Takeuchi, Y. Recent advances in mechanical micromachining. CIRP Ann.-Manuf. Technol.
2006, 55, 745–768. [CrossRef]
8. Xu, B.Y.; Yan, X.N.; Zhang, J.D.; Xu, J.J.; Chen, H.Y. Glass etching to bridge micro- and nanofluidics. Lab Chip
2012, 12, 381–386. [CrossRef] [PubMed]
9. Park, D.S.; Chen, P.C.; You, B.H.; Kim, N.; Park, T.; Lee, T.Y.; Datta, P.; Desta, Y.; Soper, S.A.; Nikitopoulos, D.E.;
et al. Titer-plate formatted continuous flow thermal reactors for highthroughput applications: Fabrication
and test. J. Micromech. Microeng. 2010, 20, 055003. [CrossRef]
10. Chen, P.C.; Park, D.S.; You, B.H.; Kim, N.; Park, T.; Soper, S.A.; Nikitopoulos, D.E.; Murphy, M.C.
Titer-plate formatted continuous flow thermal reactors: Design and performance of a nanoliter reactor.
Sens. Actuators B-Chem. 2010, 149, 291–300. [CrossRef] [PubMed]
11. Kim, J.A.; Lee, J.Y.; Seong, S.; Cha, S.H.; Lee, S.H.; Kim, J.J.; Park, T.H. Fabrication and characterization of a
PDMS-glass hybrid continuous flow PCR chip. Biochem. Eng. J. 2006, 29, 91–97. [CrossRef]
12. Mecomber, J.S.; Hurd, D.; Limbach, P.A. Enhanced machining of micron-scale features in microchip molding
masters by CNC milling. Int. J. Mach. Tools Manuf. 2005, 45, 1542–1550. [CrossRef]
13. Mecomber, J.S.; Stalcup, A.M.; Hurd, D.; Halsall, H.B.; Heineman, W.R.; Seliskar, C.J.; Wehmeyer, K.R.;
Limbach, P.A. Analytical performance of polymer-based microfluidic devices fabricated by computer
numerical controlled machining. Anal. Chem. 2006, 78, 936–941. [CrossRef] [PubMed]
14. Becker, H.; Heim, U. Hot embossing as a method for the fabrication of polymer high aspect ratio structures.
Sens. Actuators A Phys. 2000, 83, 30–135. [CrossRef]
15. Becker, H.; Gartner, C. Polymer based micro-reactors. Rev. Mol. Biotechnol. 2001, 82, 89–99. [CrossRef]
16. Hupert, M.L.; Guy, W.J.; Llopis, S.D.; Shadpour, H.; Rani, S.; Nikitopoulos, D.E.; Soper, S.A. Valuation of
micromilled metal mold masters for the replication of microchip electrophoresis devices. Microfluid. Nanofluid.
2007, 3, 1–11. [CrossRef]
17. Park, C.H.; Song, C.K.; Hwang, J.; Kim, B.S. Development of an ultra precision machine tool for
micromachining on large surfaces. Int. J. Precis. Eng. Manuf. 2009, 10, 85–91. [CrossRef]
18. Vazquez, E.; Rodriguez, C.A.; Elias-Zuniga, A.; Ciurana, J. An experimental analysis of process parameters
to manufacture metallic micro-channels by micro-milling. Int. J. Adv. Manuf. Technol. 2010, 51, 945–955.
[CrossRef]
19. Zhang, J.Z.; Chen, J.C.; Kirby, E.D. Surface roughness optimization in an end-milling operation using the
Taguchi design method. J. Mater. Process. Technol. 2007, 184, 233–239. [CrossRef]
20. Chung, Y.C.; Lai, L.W.; Yang, L.J.; Liao, W.J. Comparison of Different Metal Film Thicknesses of
COC-Substrate Polymerase Chain Reaction Chips With Single-Side and Double-Side Heaters. In Proceedings
of the ASME 2009 International Mechanical Engineering Congress and Exposition, Lake Buena Vista, FL,
USA, 13–19 November 2009; Volume 149, pp. 435–437.
21. Yu, H.H.; Hwang, S.J.; Hwang, K.C. Preparation and characterization of a novel flexible substrate for OLED.
Opt. Commun. 2005, 248, 51–57. [CrossRef]
Micromachines 2017, 8, 264 9 of 9

22. Ma, K.S.; Reza, F.; Saaem, I.; Tian, J. Versatile surface functionalization of cyclic olefin copolymer (COC) with
sputtered SiO2 thin film for potential BioMEMS applications. J. Mater. Chem. 2009, 19, 7914–7920. [CrossRef]
23. Yoon, H.S.; Wu, R.; Lee, T.M.; Ahn, S.H. Geometric optimization of micro drills using Taguchi methods and
response surface methodology. Int. J. Precis. Eng. Manuf. 2011, 12, 871–875. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).