See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/332071601

Label-free biosensing using a microring resonator integrated with poly-

(dimethylsiloxane) microfluidic channels

Article  in  The Review of scientific instruments · March 2019

DOI: 10.1063/1.5074134

CITATIONS READS
18 153

5 authors, including:

Shangquan Wu
University of Science and Technology of China
48 PUBLICATIONS   520 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Microcantilever sensor View project

Pixelated-polarization-camera-based polarimetry system View project

All content following this page was uploaded by Shangquan Wu on 05 January 2020.

The user has requested enhancement of the downloaded file.

Label-free biosensing using a microring
resonator integrated with poly-
(dimethylsiloxane) microfluidic channels
Cite as: Rev. Sci. Instrum. 90, 035004 (2019); https://doi.org/10.1063/1.5074134
Submitted: 22 October 2018 . Accepted: 08 March 2019 . Published Online: 28 March 2019

Shangquan Wu, Yingying Guo, Wanjun Wang, Jie Zhou, and Qingchuan Zhang

Rev. Sci. Instrum. 90, 035004 (2019); https://doi.org/10.1063/1.5074134 90, 035004

© 2019 Author(s).
 Review of ARTICLE scitation.org/journal/rsi
Scientific Instruments

Label-free biosensing using a microring resonator

integrated with poly-(dimethylsiloxane)
microfluidic channels
Cite as: Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134
Submitted: 22 October 2018 • Accepted: 8 March 2019 •
Published Online: 28 March 2019

Shangquan Wu,1 Yingying Guo,1 Wanjun Wang,2 Jie Zhou,2 and Qingchuan Zhang1,a)

AFFILIATIONS
1
CAS Key Laboratory of Mechanical Behavior and Design of Material, Department of Modern Mechanics,
University of Science and Technology of China, Hefei 230027, China
2
China Electronics Technology Group Corporation No. 38 Research Institute, Anhui 230001, China

a)
Author to whom correspondence should be addressed: zhangqc@ustc.edu.cn. Telephone: +86-551-63607613.
Fax: +86-551-63601248.

ABSTRACT
Microring resonators have shown promising potential for highly sensitive, label-free, real-time detection of biomolecules. Accurate quan-
titative detection of target molecules through use of photonic integrated circuits has been demonstrated for environmental monitoring
and medical diagnostics. Here, we described the design, fabrication, and characterization of a highly sensitive, label-free microring opti-
cal resonator integrated with poly-(dimethylsiloxane) microfluidic channels, which consumes only 30 µl of sample solution. The resonance
wavelength shifts resulting from the change in the effective refraction index can be measured in situ, and thus the binding events on the res-
onator surface, including antibody immobilization, blocking of the resonator surface, and the specific binding of antibody and antigen, can be
recorded throughout the entire experimental process in real time. We measured the binding events for the detection of human immunoglob-
ulin G. The system had a detection limit of 0.5 µg/ml, a value substantially (14 times) lower than that of a previously reported microring
resonator. To verify the usefulness and adaptability of this technique, human epidermal growth factor receptor 2 was used for the detection.
The microring optical resonator was able to monitor reactions between biological molecules in real time and thus can be used in quantitative
detection and biological sensing with little sample consumption.
Published under license by AIP Publishing. https://doi.org/10.1063/1.5074134

I. INTRODUCTION semiconductor (CMOS) fabrication technology,15,16 as well as the

Biochemical sensors are powerful detection and analysis tools
with diverse applications in biomedical research, environmental
monitoring, and health care.1,2 According to the different output
physical signals, biochemical sensors can be grouped into func-
tional types such as optical biosensors (including microring biosen-
sors3–5 ), electrochemical biosensors,6,7 and mechanical biosensors
(microcantilever sensors8–10 ). Among these sensors, the silicon
microring resonator, a highly sensitive, label-free, real time, rapid
response, and efficient optical sensor,11–14 has become one of the
most promising photonic integration platforms. Its promise can be
mainly attributed to the combination of a high contrast in refrac-
tive index (RI) and the availability of complementary metal oxide
substantial experimental cost reduction associated with the use of
electronics fabrication facilities.17–19 Silicon microring resonators
can be integrated with microfluidic technology, thus providing a
relatively stable liquid environment. Using advanced fluid handling
techniques for photonic devices at or below the micrometer scale has
the potential to provide compact, effective sensors for lab-on-a-chip
tools.
Microring resonators take advantage of the highly sensitive
evanescent field generated near silicon-on-insulator (SOI) photonic
waveguides, owing to a contrast in RI between core silicon and
silicon dioxide.20–23 The evanescent field is a type of electromag-
netic wave produced at the boundary of two different media. Its
amplitude decays exponentially with increasing vertical dimension

Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134 90, 035004-1
Published under license by AIP Publishing
 Review of
Scientific Instruments
to the boundary. Biorecognition molecules, such as antibodies and
aptamers, are immobilized on the resonator surface. When a solu-
tion containing target analytes, such as specific antigens, flows over
the resonator surface, the binding of antigens to antibodies causes
the effective RI to change, thus shifting the resonance wavelength.
The effective RI is related to the RI of the microring and the
microring’s surroundings. Resonance occurs when the change in
the optical path length of the resonator is exactly a whole num-
ber. Microring resonators can sense the effective RI changes induced
by analyte binding on the surface. By examining the shift in res-
onance wavelength, the binding of biomolecules to target-specific
capture agents can be monitored. The resonance wavelength is
defined as
λ = 2πRneff /m, (1)
where m is an integer (m = 1, 2, 3 . . .), λ is the wavelength of light, R
is the radius of the microring, and neff is the effective RI.
Human immunoglobulin G (IgG), the most abundant
immunoglobulin in serum, plays an important role in the immune
response.24 Many analytical methods are currently available to
detect IgG, including enzyme-linked immunosorbent assay (ELISA),
and protein chip and fluorescein labeling methods. Fluores-
cein labeling and ELISA are sensitive and selective methods for
detection.25 However, either target molecules or biorecognition
molecules must be fluorescently tagged, a process that is often dif-
ficult, typically requiring extensive sample pretreatment or lengthy
procedures, and may increase experimental cost. In contrast,
microring resonators, allow for label-free detection of molecules in
their natural forms on the chip surface without the added complex-
ity of the forms on labels or enzymatic tags. Microring resonators
have the notable advantages of being label-free, being able to detect
multiple analytes in real time to provide in situ monitoring, and hav-
ing lower analytical costs. Thus, this method is relatively easy and
inexpensive to perform.
Here, we demonstrate the design, fabrication, and characteriza-
tion of a highly sensitive, label-free microring optical resonator, inte-
grated with poly-(dimethylsiloxane) (PDMS) microfluidic channels,
which consumes only 30 µl of the sample solution. Using different
concentrations of ethanol solutions with known refractive indices,
we determined the RI detection sensitivity and the detection limit
(DL). We demonstrated the label-free, quantitative detection of IgG
by using a microring resonator and measured the resonance wave-
length shift resulting from the effective RI changes caused by the
specific binding between antibodies and antigens. Furthermore, the
binding events on the resonator surface, including antibody immo-
bilization, blocking of the resonator surface, and specific binding of
antibody and antigen, were recorded during the entire experiment
in real time. To demonstrate the usefulness and adaptability of the
presented technique, we also demonstrated the label-free detection
of human epidermal growth factor receptor 2 (Her2).
II. MATERIAL AND METHODS
A. Materials and apparatus
IgG and goat anti-human IgG were purchased from San-
gon Biotech (Shanghai) Co., Ltd. Human Her2 protein and anti-
Her2 monoclonal antibody were purchased from Sino Biological,
Inc. Bovine serum albumin (BSA), 3-aminopropyltriethoxysilane
Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134
Published under license by AIP Publishing
ARTICLE scitation.org/journal/rsi
(APTES), glutaraldehyde (GAD) solution [25% (w/v) in water], and
phosphate-buffered saline (PBS; 0.1M phosphate buffer containing
0.9% sodium chloride, pH 7.5) were obtained from Sigma-Aldrich
(St. Louis, MO, USA). Other chemicals were of analytical reagent
grade and were used as received.
SOI resonator chips and microfluidic channels (30 µl), both
made in house, were used. A temperature control system was
purchased from Shimax (MAC50, JP), and 24-well polystyrene
microplates were purchased from Costar (Corning, NY, USA).
B. Resonator design, fabrication,
and characterization
Figure 1 shows an SEM micrograph of the home-built SOI
microring resonator with a radius of 20 µm used in the experiment.
The circular section is the microring waveguide, and the structure
beside the ring is the straight waveguide. In this work, microring
resonators were fabricated on a SiO2 substrate (RI: 1.44) with a Si
core layer (RI: 3.45). Owing to the large RI difference between Si and
SiO2 , most optical fields were restricted in the higher index Si waveg-
uide. In order to achieve the high sensitivity, the most key design rule
is to improve the overlap between the optical mode and the analyte.
Among the single microring resonator, one of the key parameter
is the thickness of the silicon waveguide; with thinner waveguide,
more optical modes will overlap with the analyte. If we use thinner
waveguide, we can get higher sensitivity. Based on sensitivity and
micro-nano processing technology, we have chosen a microring of
20-µm radius, a silicon waveguide of 220 nm × 450 nm, and a gap of
160 nm between the straight waveguide and the ring.
The microring resonators were patterned by using standard
optical photolithography and formed by using reactive ion etching
(RIE).26 Our chips were fabricated at the Interuniversity Micro-
electronics Centre (IMEC), where 150 nm line or gap can be pro-
cessed. The fabrication was started by spin-coating approximately
FIG. 1. (a) A schematic illustration of the microfluidic channels system, exposing
the four layers of the microfluidic channel system (from bottom to top): the bottom
base, the optical chip, the fluidic layer, and the top base. Illustration: SEM image
of a microring resonator and linear waveguide in the corner. (b) Photograph of the
PDMS microfluidic channel system.
90, 035004-2
 Review of
Scientific Instruments
280–320 nm of photoresist on an SOI wafer, then drying for 90 s
at 90 ○ C (microfabrication process sequence in Fig. S1). After con-
ventional photolithography, the pattern of the photoresist was trans-
ferred to the SOI wafer through inductively-coupled plasma RIE.
The photoresist was then removed, thus allowing the analyte to con-
tact the sensing waveguide surface. Finally, the resonator chip was
cleaned in a solution (1:1 H2 SO4 /30% H2 O2 ) for 10 min to remove
the protective photoresist coating and any residual organic contam-
inants introduced from fabrication and then washed with copious
amounts of deionized water.
C. Fabrication of the microfluidic channel system
In the process of detection, steady control of sample solutions
flowing through the chip surface is crucial. Microfluidic channels,
one of the most commonly used methods for the detection of tar-
get molecules with little sample consumption, not only ensure the
activity of target and biorecognition molecules but also make the
modification of the chip more efficient.
Owing to the high sensitivity of the microring, the output spec-
trum is easily affected by external factors, such as vibration, temper-
ature, and dust. Many materials are available with a broad range of
optical, mechanical, and chemical properties. Polymers have been
mostly used as substrate materials for microfluidic devices in recent
years.27,28 We designed the system of microfluidic channels inte-
grated with a microring resonator, which prevented salt coales-
cence on the chip surface, owing to liquid volatilization in an open
environment. The microfluidic channels consumed only 30 µl of
sample solution, thus potentially allowing for substantially lower
experiment costs.
We used two acrylic polymers for easy processing, poly-
(methylmethacrylate) (PMMA) and poly-(dimethylsiloxane)
(PDMS) to fabricate the microfluidic channels. PDMS is a type of
macromolecule organic silicon compound, which has good biocom-
patibility and is widely used in biological microelectromechanical
systems, as a gap filling agent or lubricant, and in contact lenses.29,30
First, we placed the microring chip in the PMMA base. Then,
the precise PDMS microfluidic channels were aligned, and another
PMMA base was deposited to produce a seal. Simultaneously, inlet
and outlet holes above the base were made to form solution chan-
nels. As shown in Fig. 1, the base encapsulated the surface of the
chip, thus not only protecting the chip but also forming the solution
channels.
D. Measurement setup
The resonator consisted of two parts: the integrated optical
waveguide structure and the microfluidic channel system. Solutions
were controlled by a peristaltic pump, which supplied a continu-
ous flow of solution through the microfluidic channels and output
pipes, then to the waste solution bottle. We use the broadband light
source and optical spectrum analyzer to monitor the wavelength
shift of the microring resonator sensor. Light from the broad band
light source was coupled into the waveguide of the resonator through
polarization-maintaining lens fibers. After a series of coupling oscil-
lations, the output light was collected by a spectrometer through the
output optical fiber. For the optical measurement system, a three-
dimensional adjustable high precision optical test platform was used
to accurately adjust the input and output optical fibers and chip.
Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134
Published under license by AIP Publishing
ARTICLE scitation.org/journal/rsi
In addition, a high precision temperature controller was applied to
adjust the experiment temperature. The control of the entire mea-
surement setup was automated by a computer (measurement setup
in Fig. S2). The polarization used is TM mode.
E. Preparation of the microring
The surface of the sensor was modified with APTES and glu-
taraldehyde before being integrated with the microfluidic channels.
The chip was first cleaned in piranha dip (3:1 H2 SO4 /30% H2 O2 ) for
approximately 10 min (with stirring) to remove the protective pho-
toresist coating and any residual organic contaminants introduced
from fabrication, and then it was washed with copious amounts of
deionized water and dried under nitrogen gas. To promote anti-
body adhesion to the silicon surface, the chip surface was immersed
in a solution of 2.5% APTES (1:100 deionized water/ethanol) for
2 h at 4 ○ C, then rinsed with ethanol and deionized water, and
dried under nitrogen gas. The resonator chip was incubated with
5% glutaraldehyde (GAD) in deionized water for 1 h and rinsed
with deionized water three times. Next, we integrated the chip with
microfluidic channels as described in Sec. II C. The temperature con-
trol of the microfluidic channels was 27 ± 0.01 ○ C, and the room
temperature was controlled to 25 ± 0.1 ○ C. The functionalization
procedures of the chip were performed in the microfluidic chan-
nels. PBS containing 100 µg/ml of anti-IgG monoclonal antibody
was flowed through the microfluidic channels and over the resonator
surface at a rate of 4.2 ml/h for 10 min, and then the chip was incu-
bated for 1.5 h to allow functionalization. A rinse with PBS was
performed to remove any loosely bound antibody. The microflu-
idic channel was then filled with 5% solution (w/v) of BSA in PBS
for 40 min for blocking. After these procedures, the functionalized
microring was washed with PBS and was then ready to use, as shown
in Fig. S3.
III. RESULTS AND DISCUSSION
A. Stability of resonance wavelength shift
The stability of the signal when no biomolecular interaction
takes place on the chip surface is crucial for a low detection limit.
To determine whether the microfluidic channels system was able to
measure wavelength shifts with high signal stability, we monitored
the real time resonance wavelength shift in an open liquid environ-
ment and microfluidic channels by using the same chip [Fig. 2(a)]. In
an open liquid environment, the chip is placed on a glass slide or in
a petri dish, and then 2 ml of liquid is added to cover the chip. In an
open liquid environment, the resonance wavelength shifted 110 pm
in the short wavelength direction after 10 min in the buffer solution
[black line in Fig. 2(a)]. After the experiment, there was abundant
salt on the chip surface. One reason for this deviation may have
been salt precipitation during the process of water loss, thus chang-
ing the effective RI constantly and causing a continued resonance
wavelength shift. In the microfluidic channel system, the continuous
liquid flow and small volatile area provided the chip with a stable
buffer solution environment, and thus the resonance wavelength
was in the steady state [red line in Fig. 2(a)]. These findings indi-
cated the stability of the microfluidic channels system, suggesting
that the microfluidic channels system was essential for the accuracy
of detection.
90, 035004-3
 Review of ARTICLE scitation.org/journal/rsi
Scientific Instruments

FIG. 2. Real time wavelength shifts of the microring caused by environmental factors. (a) Comparison between wavelength shifts obtained in PBS with or without microfluidic
channels. (b) Comparison between wavelength shifts obtained in ethanol with or without a temperature control device.

Temperature is another major contributor to unwanted
changes in effective RI in biochemical detection. We placed the chip
in ethanol with or without a temperature control device. The reso-
nance wavelength changed with changes in room temperature and
shifted 30 pm in the long wavelength direction after 30 min with-
out a temperature control device [Fig. 2(b)]. The wavelength was
essentially in a stable state at a constant temperature of 25 ○ C. These
results suggested that temperature stability was also very important
for the accuracy of detection.
B. Refractive index sensing
To demonstrate the microring sensing ability and to character-
ize the RI sensitivity of the resonator, the resonator was calibrated
by using different concentrations (0, 2, 4, 6, and 8M) of ethanol. Its
resonant wavelength was measured when the ethanol solution was
flowed across the microring surface. The results showed a linear shift
in the resonant wavelength with increasing ethanol concentration
(Fig. S4). Higher ethanol concentrations with a larger RI caused a
larger shift in the resonant wavelength. From the slope of the curve,
we obtained an RI sensitivity of 140.15 nm/RI unit (RIU). There are
many kinds of the silicon ring resonators and we summarize them
in Table I. As we known, the most key performance is the sensitiv-
ity. In order to achieve the high sensitivity, the most key design rule
is to improve the overlap between the optical mode and the ana-
lyte. The strip waveguide based on TE or TM is robust and reliable
for the fabrication and can achieve the 135 nm/RIU in a previous
reported study. In this paper, the sensitivity is improved by the using
of the TM mode. As we know, the TM mode has more overlaps
with the analyte than the TE mode, and so maybe that is why we
can achieve the higher sensitivity than the convention single micror-
ing sensor based on TE mode. Slot waveguide and sub wavelength
grating is also a very competitive solution and can push the sensi-
tivity to almost twice/three as compared to the strip waveguide. But
they suffer from the problem that the minimum linewidth of slot
or subwavelength grating is challenge for the fabrication process.
They need high resolution Lithography technology. It will intro-
duce extra cost in the fabrication. The detection limit (DL) was
defined as31
σ
DL = , (2)
RI sensitivity
where σ is the system noise which can be obtained from the sys-
tem filled with a blank solution. The DL reports the smallest RI

TABLE I. Comparison of the sensitivity of different microring resonator.

Sensitivity (nm/RIU) Reference

MRR based on strip waveguide working at TE 70 32

MRR based on ultra-thin strip waveguide working at TE 100 33
MRR based on strip waveguide working at TM 135 34
MRR based on the slot waveguide 298 35
MRR based on the sub wavelength grating 490 36
MRR based on strip waveguide working at TM 140 This work

Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134 90, 035004-4
Published under license by AIP Publishing
 Review of ARTICLE scitation.org/journal/rsi
Scientific Instruments

FIG. 3. Real time resonance wavelength shifts of the microring resonator for antibody immobilization, blocking of the sensor surface, and the specific binding of antibody and
antigen. After each procedure, the functionalized microring was washed with PBS. 1 PBS, 2 anti-IgG, 3 BSA, 4 1 µg/ml IgG, or 5 5 µg/ml IgG.

change that can be measured. The resolution of the spectrometer net shift (10 pm) was much less than that of the microring with spe-
used in the experiment was 4 pm. From Fig. 2(b), we can get the cific binding on the surface (90 pm) and slightly greater than that of
system noise which is 8 pm. Thus, we determined the DL to be the microring in PBS (4 pm), as shown in Figs. 3 and 4. The results
5.708 × 10−5 RIU. again suggested that the measured wavelength shift of the micror-
ing was generated by the specific binding interaction between the
C. Real-time detection of human IgG anti-IgG antibody and IgG on the microring surface.
The density of biorecognition molecules on the chip surface
The resonance wavelength shifts resulting from effective RI
may affect the wavelength shift; thus, we compared the wavelength
changes can be measured in situ, and thus, the binding events on the
shifts on different chip surfaces for antibody immobilization and
resonator surface, including antibody immobilization, blocking of
blocking of the resonator surface under the same modified condi-
the resonator surface, and specific binding of antibody and antigen,
tions. The wavelength shifts were measured when the chip surface
can be recorded during an entire experiment in real time (Fig. 3). We
was successively covered with anti-IgG monoclonal antibody and
recorded the resonance wavelength every 30 s. The resonance wave-
length remained essentially unchanged in PBS and shifted 408 pm
after 1.5 h in the anti-IgG solution and then reached the steady
state. The spectra did not significantly change after a rinse with PBS,
thus suggesting that the antibody had been steadily immobilized on
the microring surface. The BSA coating buffer caused a net shift
of 132 pm after 40 min. After the baseline resonance wavelength
was stabilized in the PBS solution again, the human IgG solution
with a concentration of 1 µg/ml in PBS was flowed over the sur-
face. After the chip was rinsed with PBS, 5 µg/ml IgG solution was
circulated in the microfluidic channels. For both concentrations of
IgG, the effective RI changed when the antigen-antibody complex
formed on the chip surface. The shift signals for human IgG solu-
tions were 44 and 90 pm for 1 and 5 µg/ml IgG solution, respectively.
The curve did not show an obvious shift during a continual rinse
with PBS, thus indicating that the resonance wavelength shift was
caused by an RI change on the resonator surface due to the spe-
cific binding of antigen rather than to chemical molecules in the
liquid.
To further determine whether the measured wavelength shift FIG. 4. Real-time resonance wavelength shifts of the microring resonator for spe-
was induced by the specific binding on the microring surface, the cific binding of antigen and nonspecific binding of BSA. The pink area is the real
microring was flowed with the BSA solution at a high concentration time shift in resonance wave-length during specific binding. The blue area is the
real time shift in resonance wavelength during nonspecific binding. 1 PBS, 3
(5 mg/ml) after 5 µg/ml IgG (Fig. 4). Although the BSA concentra-
BSA, 5 5 µg/ml IgG.
tion was much higher than the antigen concentration, the measured

Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134 90, 035004-5
Published under license by AIP Publishing
 Review of
Scientific Instruments
BSA solutions. As seen in Fig. S5, the wavelength shift in antibody
immobilization was approximately twice that in BSA blocking. One
reason for this deviation might be that anti-IgG flowed over the
chip earlier than BSA, thus affording the anti-IgG monoclonal anti-
body more opportunities to occupy positions on the chip surface.
Another reason may be that BSA has a smaller molecular weight
than anti-IgG. The wavelength shifts on different chip surfaces in the
same modified condition were nearly identical. These results thus
established a foundation for quantitative detection.
D. Quantitative detection of human IgG
To explore the quantitative detection capability of microrings,
solutions of human IgG with various concentrations were circulated
in the microfluidic channels after the baseline resonance wavelength
was stabilized in PBS. The resonance wavelength shifts were moni-
tored in situ for approximately 60 min for each concentration. The
real time monitoring of the resonance wavelength in response to var-
ious concentrations of human IgG is shown in Fig. 5. Different new
sensors were used in each experiment. The measured values showed
jumps attributed to the resolution of the spectrometer used in the
experiment (4 pm). The shift signals were 12, 52, 64, and 88 pm
for 0.5, 2.5, 5, and 10 µg/ml human IgG solutions, respectively.
These results indicated that the specific binding between the anti-
IgG antibody and IgG appeared to cause the resonance wavelength
shift. In addition, with increasing of antigen concentration, the shift
increased. The results demonstrated that the microring based on
the evanescent field was sensitive enough to detect IgG (7.23–16.85
mg/ml in human serum) and to quantify the IgG. The developed sys-
tem had a detection limit of 0.5 µg/ml, 14 times lower than the detec-
tion limit (7.1 µg/ml) of a previously reported microring resonator
for IgG detection.37
To verify that this technique was applicable to other ana-
lytes, a microring resonator for human epidermal growth factor
FIG. 5. Real time wavelength shifts in the microring resonator after injection of
different concentrations of IgG solutions.
Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134
Published under license by AIP Publishing
ARTICLE scitation.org/journal/rsi
FIG. 6. Representative binding curve for Her2 and PBS to an anti-Her2 modified
microring resonator. The solid lines indicate the functions fit to the initial binding
response.
receptor 2 (Her2), an important breast cancer biomarker, was devel-
oped through the same procedure described above for IgG. Over-
expression of Her2 protein occurs in approximately 25% of human
breast cancers and leads to a particularly aggressive disease.31 The
shift signals were 96 and 125 pm for 2.5 and 5 µg/ml human Her2
solutions, respectively (Fig. 6). We verified that the use of our devel-
oped system was able to sensitively detect Her2 at a concentra-
tion as low as 2.5 µg/ml, a value comparable to the detection limit
(5.0 µg/ml) of a previously reported microring resonator for Her2
detection.31 The high sensitivity of the sensing system for both IgG
and Her2 may be mainly attributed to microfluidic channels, which
provide a relatively stable liquid environment, and also to optical
waveguide structure of the sensor. As we know, the TM mode have
more overlap with the analytes than the TE mode, so maybe that
is why we can achieve higher sensitivity than the convention sin-
gle microring sensor based on TE mode. For the single microring
resonator, one of the key parameter is the thickness of the silicon
waveguide so that the thinner the waveguide, the more optical mode
will overlap with the analytes. We get a higher sensitivity based
the 220 nm × 450 nm waveguide. On the basis of these findings,
our developed system composed of a microring optical resonator
and PDMS microfluidic channels can be used for target molecule
detection, providing the advantage of high sensitivity, label-free,
quantitative, rapid, and real time detection.
IV. CONCLUSION
We demonstrated the design, fabrication, and characteriza-
tion of a highly sensitive and label-free microring optical resonator,
integrated with PDMS microfluidic channels, which avoided salt
coalescence on the chip surface resulting from volatilization in
an open environment, and required only 30 µl of sample solu-
tion. The microfluidic channels significantly decreased the exper-
imental cost. The RI detection sensitivity was determined to be
140.15 nm/RIU, and the RI detection limit was determined to be
90, 035004-6
 Review of ARTICLE scitation.org/journal/rsi
Scientific Instruments

10
5.708 × 10−5 RIU. We demonstrated label-free, real time, quan- S. Q. Wu, Z. G. Zhang, X. R. Zhou, H. Liu, C. G. Xue, G. Zhao, Y. X. Cao,
titative detection of IgG by measuring the resonance wavelength Q. C. Zhang, and X. P. Wu, Nanoscale 9, 18258–18267 (2017).
11
shift resulting from effective RI changes caused by the specific bind- Y. Shin, A. P. Perera, and M. K. Park, Sens. Actuators, B 178, 200–206 (2013).
12
ing between antibodies and antigens. The detection limit of IgG E. Kim, M. D. Baaske, and F. Vollmer, Lab Chip 17, 1190–1205 (2017).
13
was as low as 0.5 µg/ml, a value 14 times lower than that detected J. H. Wade, A. T. Alsop, N. R. Vertin, H. Yang, M. D. Johnson, and R. C. Bailey,
ACS Central Sci. 1, 374–382 (2015).
by a previously reported microring resonator. To verify the use- 14
I. A. Grimaldi, G. Testa, and R. Bernini, RSC Adv. 5, 70156–70162 (2015).
fulness and adaptability of this technique, Her2 was used for the 15
W. Bogaerts, R. Baets, P. Dumon, V. Wiaux, S. Beckx, D. Taillaert, B. Luyssaert,
detection. The advantages of the low sample volume required, the J. Van Campenhout, P. Bienstman, and D. Van Thourhout, J. Lightwave Technol.
convenience, and the capability of real-time, label-free, quantita- 23, 401–412 (2005).
tive detection suggest that the microring optical resonator may hold 16
W. Shi, X. Wang, W. Zhang, H. Yun, C. Lin, L. Chrostowski, and N. A. F. Jaeger,
promise for testing in a point-of-care environment. With these Appl. Phys. Lett. 100, 121118 (2012).
17
capabilities, microring resonators may open up new perspectives W. M. J. Green, M. J. Rooks, L. Sekaric, and Y. A. Vlasov, Opt. Express 15,
in drug discovery, clinical medicine, healthcare, and environmental 17106–17113 (2007).
18
monitoring. J. M. Fedeli, E. Augendre, J. M. Hartmann, L. Vivien, P. Grosse, V. Mazzocchi,
W. Bogaerts, D. Van Thourhout, and F. Schrank, “Photonics and electronics inte-
gration in the HELIOS project,” in 7th IEEE International Conference on Group IV
SUPPLEMENTARY MATERIAL Photonics (IEEE, 2010), pp. 356–358.
19
C. Gunn, Photon. Spectra 40, 62–68 (2006).
See supplementary material for a detailed microfabrication pro- 20
D. Erickson, S. Mandal, A. H. J. Yang, and B. Cordovez, Microfluidics Nanoflu-
cess sequence of microring resonator, the entire measurement setup idics 4, 33–52 (2008).
of the sensing system, and the schematic diagram of antibody immo- 21
S. Rosenblum, Y. Lovsky, L. Arazi, F. Vollmer, and B. Dayan, Nat. Commun. 6,
bilization. Furthermore, the supplementary material contains wave- 6788 (2015).
22
length shifts caused ethanol solutions and corresponding variance in A. Bozzola, S. Perotto, and F. De Angelis, Analyst 142, 883–898 (2017).
23
the shift curve when the chip was coated with antibody and blocked J. Witzens and M. Hochberg, Opt. Express 19, 7034–7061 (2011).
24
with the BSA solution. A. Ramachandran, S. Wang, J. Clarke, S. J. Ja, D. Goad, L. Wald, E. M. Flood,
E. Knobbe, J. V. Hryniewicz, S. T. Chu, D. Gill, W. Chen, O. King, and B. E. Little,
Biosens. Bioelectro. 23, 939–944 (2008).
25
ACKNOWLEDGMENTS W. E. Moerner, Proc. Natl. Acad. Sci. U. S. A. 104, 12596–12602 (2007).
26
L. H. Wang, J. Ren, X. Y. Han, T. Claes, X. G. Jian, P. Bienstman, R. Baets,
This work was supported by the National Natural Sci- M. S. Zhao, and G. Morthier, IEEE Photonics J. 4, 920–930 (2012).
ence Foundation of China (Nos. 11872355 and 11627803), the 27
H. Becker and C. Gartner, Anal. Bioanal. Chem. 390, 89–111 (2008).
Fundamental Research Funds for the Central Universities (No. 28
H. Becker and L. E. Locascio, Talanta 56, 267–287 (2002).
WK2480000002), and the International S&T Cooperation Program 29
C. F. Carlborg, K. B. Gylfason, A. Kazmierczak, F. Dortu, M. J. Banuls
of China (ISTCP) (No. 2013DFE13090). Polo, A. Maquieira Catala, G. M. Kresbach, H. Sohlstrom, T. Moh, L. Vivien,
J. Popplewell, G. Ronan, C. A. Barrios, G. Stemme, and W. van der Wijngaart,
REFERENCES Lab Chip 10, 281–290 (2010).
30
I. M. White and X. D. Fan, Opt. Express 16, 1020–1028 (2008).
1
S. Zhang, R. Geryak, J. Geldmeier, S. Kim, and V. V. Tsukruk, Chem. Rev. 117, 31
K. W. Kim, J. F. Song, J. S. Kee, Q. Liu, G. Q. Lo, and M. K. Park, Biosens.
12942–13038 (2017). Bioelectron. 46, 15–21 (2013).
2
S. Zhan, Y. Wu, L. Wang, X. Zhan, and P. Zhou, Biosens. Bioelectron. 86, 32
K. De Vos, I. Bartolozzi, E. Schacht, P. Bienstman, and R. Baets, Opt. Express
353–368 (2016). 15, 7610–7615 (2007).
3
M. S. Luchansky and R. C. Bailey, Anal. Chem. 84, 793–821 (2012). 33
S. T. Fard, V. Donzella, S. A. Schmidt, J. Flueckiger, S. M. Grist, P. TalebiFard,
4
Q. Liu, X. G. Tu, K. W. Kim, J. S. Kee, Y. Shin, K. Han, Y. J. Yoon, G. Q. Lo, and Y. C. Wu, R. J. Bojko, E. Kwok, N. A. F. Jaeger, D. M. Ratner, and L. Chrostowski,
M. K. Park, Sens. Actuators, B 188, 681–688 (2013). Opt. Express 22, 14166–14179 (2014).
5 34
M. K. Park, J. S. Kee, J. Y. Quah, V. Netto, J. F. Song, Q. Fang, E. M. La Fosse, and D. X. Xu, M. Vachon, A. Densmore, R. Ma, S. Janz, A. Delage, J. Lapointe,
G. Q. Lo, Sens. Actuators, B 176, 552–559 (2013). P. Cheben, J. H. Schmid, E. Post, S. Messaoudene, and J. M. Fedeli, Opt. Express
6
K. Abnous, N. M. Danesh, M. Alibolandi, M. Ramezani, A. Sarreshtehdar 18, 22867–22879 (2010).
Emrani, R. Zolfaghari, and S. M. Taghdisi, Biosens. Bioelectron. 94, 374–379 35
K. B. Gylfason, C. F. Carlborg, A. Kazmierczak, F. Dortu, H. Sohlstrom,
(2017). L. Vivien, C. A. Barrios, W. van der Wijngaart, and G. Stemme, Opt. Express 18,
7
Z. H. Yang, Y. Zhuo, R. Yuan, and Y. Q. Chai, Anal. Chem. 88, 5189–5196 (2016). 3226–3237 (2010).
8 36
X. Zhou, S. Wu, H. Liu, X. Wu, and Q. Zhang, Sens. Actuators, B 226, 24–29 J. Flueckiger, S. Schmidt, V. Donzella, A. Sherwali, D. M. Ratner, L.
(2016). Chrostowski, and K. C. Cheung, Opt. Express 24, 15672–15686 (2016).
9 37
S. Wu, H. Liu, X. M. Liang, X. Wu, B. Wang, and Q. Zhang, Anal. Chem. 86, Y. Chen, F. Yu, C. Yang, J. Song, L. Tang, M. Li, and J.-J. He, Opt. Commun.
4271–4277 (2014). 344, 129–133 (2015).

Rev. Sci. Instrum. 90, 035004 (2019); doi: 10.1063/1.5074134 90, 035004-7
Published under license by AIP Publishing

View publication stats