Cellphone-Enabled Microwell-Based Microbead Aggregation Assay
for Portable Biomarker Detection
and Xuexin Duan*
State Key Laboratory of Precision Measuring Technology & Instruments, College of Precision Instrument and Optoelectronics
Engineering, Tianjin University, Tianjin 300072, China
Department of Electrical Engineering and Yale University, New Haven, Connecticut 06520, United States
ABSTRACT: Quantitative biomarker detection methods featured
with rapidity, high accuracy, and label-free are demonstrated for the
development of point-of-care (POC) technologies or “beside”
diagnostics. Microbead aggregation via protein-speciﬁc linkage
provides an eﬀective approach for selective capture of biomarkers
from the samples, and can directly readout the presence and amount
of the targets. However, sensors or microﬂuidic analyzers that can
accurately quantify the microbead aggregation are scared. In this work,
we demonstrate a microwell-based microbeads analyzing system, by
which online manipulations of microbeads including trapping,
arraying, and rotations can be realized, providing a series of
microﬂuidic approaches to layout the aggregated microbeads for
further convenient characterizations. Prostate speciﬁc antigen is
detected using the proposed system, demonstrating the limit of
detection as low as 0.125 ng/mL (3.67 pM). A two-step reaction kinetics model is proposed for the ﬁrst time to explain the
dynamic process of microbeads aggregation. The developed microbeads aggregation analysis system has the advantages of label-
free detection, high throughput, and low cost, showing great potential for portable biomarker detection.
KEYWORDS: microbead aggregation, microﬂuidics, cellphone-enabled detection, PSA detection, two-step reaction, POC
Since biomarkers are objectively measured and evaluated as
indicators of normal biological processes, pathogenesis, or a
pharmacological response to therapeutic intervention,
accurate detection and quantiﬁcation of biomarkers in a
convenient way is a central goal of modern biotechnology.
For clinical applications, sandwich immunoassays are the most
widely applied assay formats for target biomarker detections.
Generally, these methods require some type of labeling
(enzymatic or ﬂuorescent) for selective and sensitive report
of target analytes.
Specialized instruments (e.g., plate
reader) and numerous washing steps are required as well.
Recently, advances in quantitative biomarker detection
methods featured with rapidity, high accuracy, portability, and
label-free readout have enabled the concept of point-of-care
(POC) technologies or “bedside”diagnostics.
them, aﬃnity biosensors such as impedance spectroscopy,
surface-enhanced Raman scatter-
and gravimetric sensor
developed to directly recognize and quantify the target
biomarkers by immobilizing speciﬁc acceptors on the trans-
ducers, which do not require any type of labeling. Another
promising label-free detection method is based on biofunction-
alized micro/nanoparticles, which have been demonstrated and
widely applied as molecule probes or carriers for direct capture
of target protein or DNA from complex samples.
Colorimetric method or particle size analyzer has been applied
for biomarker detection through analyzing the aggregation
status of these particles.
Aggregations of nanoparticles are
induced by the speciﬁc protein interactions (e.g., antibody−
antigen) between the receptor immobilized on the particle
surface and the target analyte in solution. Generally, more and
larger aggregated clusters will be generated with higher
biomarker concentrations. However, the accurate relationship
between nanoparticle aggregations and the biomarker concen-
tration has not been thoroughly understood because of the lack
of compatible methods or tools to precisely quantify the
aggregation status, including information regarding the number
of the nanoparticles within each cluster. In addition to
nanoparticles, aggregation of antibody functionalized microbe-
ads for antigen detection has been recently demonstrated using
an impedance sensor to count the number of the aggregated
Though an impedance sensor can directly read out
the number of clusters, it suﬀers problems such as easy
Received: November 22, 2017
Accepted: January 19, 2018
Published: January 19, 2018
Cite This: ACS Sens. 2018, 3, 432−440
© 2018 American Chemical Society 432 DOI: 10.1021/acssensors.7b00866
ACS Sens. 2018, 3, 432−440
congestion, low throughput, and limited resolution for diﬀerent
sized beads. Besides, the developed impedance analyzing
system requires dedicated ﬂuid delivery setup and expensive
data acquisition system for fast and low noise electrical signal
processing which hindered the development of such systems
into portable assays. Microﬂuidic analyzers with the capability
of processing the samples in a high throughput and well-
controllable manner are attractive for microbeads character-
izations. Especially, arraying microbeads into a small space
contributes to more eﬃcient observation and characterization
of their aggregations.
In this work, a cellphone-enabled microwell-based POC
system is developed for biomarker detections by analyzing the
aggregation status of the biomarker functionalized microbeads.
The microﬂuidic chip is made of a (PDMS) pattern using a
reusable silicon mold, with microsized features deﬁned by
standard optical lithography. Using the developed system, we
demonstrated the successful detection of prostate speciﬁc
antigen (PSA), a critical biomarker in early stage detection of
prostate cancer using anti-PSA functionalized polystyrene (PS)
beads. The system provides a dedicated tool for microbead
trapping and arraying via ﬂuidic approaches. Aggregation status,
deﬁned as the percentage of microbeads dimers, was acquired
by a cellphone enabled portable imaging system. The limit of
detection (LOD) of this system is demonstrated to be as low as
3.67 pM, which is beyond the requirement of clinical PSA
detections. Based on the experimental results, the kinetics of
microbeads aggregation is carefully studied, and a two-step
reaction model is proposed to explain the aggregation process.
These results demonstrate that this platform provides a simple,
label-free, low-cost, and point-of-care system for biomarker
Materials. Carboxyl-terminated PS beads (108/mL) with diameter
of 5 μm were purchased from Sigma; polyclonal PSA (prostate speciﬁc
antigen) and anti-PSA (1 mg/mL) were purchased from Sigma-
Aldrich. Poly-L-lysine grafted with oligoethylene glycol (PLL-OEG,
MW = 15−30 kDa) is synthesized in our lab according to procedures
described in previous publications.
perazine-N-ethane-sulfonicacid, 10 mM, pH = 7) containing 0.2% v/v
Tween (Sigma) as the surfactant was used as the solvent. Full serum
(F2442, MFCD00132239) was purchased from Sigma. The portable
microscope was purchased from Shenying Optics (SAGA002, Suzhou,
China), and the cellphone can be any commercial cellphone with a
camera of 12 Megapixels (more details about the cellphone enabled
portable microscope system are presented in the Supporting
Information, Figure S1).
Surface Modiﬁcations of PS Microbeads. Anti-PSA with a
concentration of 8 μg/mL was applied to modify PS beads using
EDC/NHS (N-(3-(dimethylamino)propyl)-N′-ethylcarbodiimide hy-
drochloride and N-hydroxysuccinimide) activation approach.
Microbeads were washed twice with HEPES buﬀer (10 mM, pH =
7.4), and then activated by EDC (N-3-dimethylanimopropyl-N-
ethylcarbodimide) and sulfo-NHS (N-hydroxysulfosuccinimide) at a
typical concentration of 2 mg/mL (EDC) and 0.5 mg/mL (NHS) in a
shaker (800 rpm) at room temperature. After activation for 30 min,
the microbeads were separated by centrifugation and washed twice
again with HEPES. Antibodies (anti-PSA, 8 μg/mL) diluted in HEPES
were then quickly added to the activated particles. After incubation at
room temperature for 40 min, the supernatant was removed after
centrifugation. Ethanolamine aqueous solutions (20 mM, pH = 7.4)
were then added and reacted with microbeads for another 40 min to
deactivate the unbound carboxyl groups on the surface, followed by
washing twice with HEPES buﬀer. Finally, the antibody conjugated
microbeads were resuspended in HEPES and stored at 4 °C.
Chip Design and Fabrication. The MicroWell-based Microﬂuidic
Chip (MW-MFC) consists of a microchamber for ﬂuid delivery and
microwell arrays on the bottom, both of which are made of PDMS.
The microchamber is designed to be 30 mm in length, 5 mm in width,
and 20 μm in height supported by a 4 ×25 pillars array to prevent
channel collapse. Two input channels and one exit channel are
connected with to the chamber, and the width of each is 600 μm.
Microwells on the bottom of the chamber are 10 μm in depth and 20
μm in diameter. Silicon mold was fabricated by reactive ion etch
process, and a releasing agent perﬂuorodecyltriethoxysilane (PFDTS)
was vacuum deposited overnight and then cured in 85 °C for 30 min.
Two-part Sylgard 184 was mixed with a mass ratio of 10:1 and then
degassed for 30 min. The mixture was poured over the silicon mold
and cured at 85 °C for 90 min, after which the cured PDMS was
carefully peeled oﬀand holes were punched to serve as the inlet and
outlet ports. Each PDMS chip was cleaned with ethanol and deionized
water (1:1) solution and gently dried with nitrogen, followed by
treatment with O2plasma cleaner under 120 W for 30 s. The PDMS
was aligned and irreversibly bound together after 30 min curing under
85 °C. The inner surface of PDMS chip was incubated with the PLL-
OEG solution (100 mM, dissolved in HEPES buﬀer, pH = 7.4) for 30
min to achieve a hydrophilic coating.
Principles of Microbeads Manipulations and Quantiﬁca-
tions. The sensing mechanism of our system is based on the counting
of the number of the aggregated beads. Anti-PSA modiﬁed PS beads
are mixed with diﬀerent concentrations of PSA samples. After
incubation at room temperature, the PS beads will aggregate together
through the speciﬁc interactions between PSA and anti-PSA. In
principle, when the number of the beads is designed to be in excess of
the target biomarkers, dimers (two PS beads linked together as Figure
1a shows) are the most likely aggregation status in the ﬁnal
The reacted microbeads were then processed by the
microwell analyzing system to obtain the percentage of the dimers.
Figure 1b presents the arrayed microbeads within the MW-MFC,
which is imaged with a cellphone equipped with a portable optical
microscope. The acquired images were processed with a home
developed MATLAB program and the percentage of the dimer, i.e.,
aggregation ratio, was calculated (as Figure 1c shows) to represent the
aggregation status by eq 1
ggregation ratio number of dimers
number of arrayed microbeads 100% (1)
Figure 1. (a) Diagram of the microbeads aggregation process: mixing
PSA into anti-PSA modiﬁed polystyrene microbeads suspension forms
bead aggregations induced by the antibody−antigen interactions. (b)
Image of the arrayed microbeads within microwells. (c) Statistics of
the aggregated microbeads acquired by processing the photos. Scale
bar in (b) is 40 μm.
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
As Figure 2c shows, the microbeads suspended within the MW-MFC
would be dragged by the ﬂuidic drag force
where K= 81.2, Ris the radius of the microbead, νis the kinematic
viscosity, and γis the ﬂuidic shear rate induced by velocity gradient
near the microbead. Vand Vprepresent the ﬂow velocity and the
microbead moving velocity, respectively. When the ﬂow rate is very
low (below the ﬂow rate Q1labeled in Figure 2c, which represents the
case of a balance between gravity and drag force inside the microwell),
most of the microbeads is deposited onto the substrate by gravity, G.
While the microbeads can be lifted from the substrate and the
microwells when the ﬂow rate is increased to be higher than Q2
(corresponding to the ﬂow rate that induces a lift force equaling to the
gravity). By tuning the ﬂow rate (between Q1and Q2), the
microvortices induced by the microwells
would contribute to
trapping microbeads into the microwells. Finally, HEPES buﬀer was
injected into the chip with a ﬂow rate of 400 nL/min (referring to Q2),
which is high enough to generate a shift force to release the
microbeads outside the microwells and wash them away from the
chamber, while the trapped beads inside the microwells are well
protected from lifting out. When increasing the ﬂow rate up to 500
nL/min (higher than Q2), the trapped microbeads will also be released
by the lift force induced by the ﬂow rotation,
following the eq 3
Here, ωis the rotation angular velocity of microbeads.
Image Acquisition and Microbeads Detection. PSA of
diﬀerent concentrations was respectively mixed with anti-PSA
modiﬁed PS beads suspension in a 500 μL centrifugal tube and then
incubated for 30 min at room temperature. The ﬁnal concentration of
microbeads in the incubated mixture was 2 ×106/mL. After
incubation, the suspension was diluted four times with HEPES buﬀer.
Typically, 20 μL of aggregated microbeads suspension was introduced
into the MW-MFC. By controlling the ﬂuidic ﬂow as outlined above,
the suspended microbeads are well trapped and arrayed for imaging
analysis. The manipulations of microbeads trapping and arraying
requires 10 min, and the picture acquisition of the arrayed microbeads
in the region of interest (ROI) takes 5 min. For all the experiments, a
total population of approximately 3000 microbeads was trapped and
arrayed corresponding to trap ratio of 3%. Typically, 70 images were
taken to include a suﬃcient number of beads. To eﬀectively acquire
the beads aggregation information from these images, a MATLAB
image-processing program was developed to directly count the
number of the dimers. Microbeads with diameter of 5 μm can be
easily identiﬁed with the Hough-transform algorithm. Furthermore, a
cellphone APP has been developed, enabling the whole assay job,
including the image capture, analysis, and data processing all operated
within a single cellphone (the source code of the APP can be
downloaded at the GitHub page https://github.com/linzuzeng/
Microsphere/releases, and detailed information can be found in the
Supporting Information, Figures S2−S4).
Chip Fabrication and System Design. Figure 2a,b shows
the schematic of the MW-MFC and the fabrication process.
The chamber is designed as 20 μm in height supported by
micropillars. Two connecting channels are used to introduce
microbeads suspensions and HEPES buﬀer via syringe pumps.
The depth of the microwell is set as 10 μm for high eﬃciency
trapping of 5 μm microbeads in combination with the
microchamber. Hydrophilic modiﬁcation of the inner surface
of the chip is required to reduce the nonspeciﬁc adsorption of
suspended microbeads. This is achieved through coating the
chips with PLL-OEG.
Figure 2c presents the principles of
the microbeads trap and release with MW-MFC, which forms
the basis of the microbeads array. Since the size of the particle
and microwell is rather large, a cellphone camera connected to
a standard portable microscope is used to acquire the images of
the arrayed microbeads by directly attaching the camera to the
objective lens of the microscope (Figure 2d). Figure 2e
presents a comparison of the MW-MFC images obtained via a
professional microscope (top image) and the cellphone
(bottom image). The aggregation status of the microbeads
can be easily identiﬁed from both images, proving the capability
of the developed portable system for image-based microbeads
Figure 2. (a) Schematic and (b) fabrication process of the MicroWell-based Microﬂuidic Chip. (c) Principles of the microbeads trap and release,
wherein the dotted arrow presents the integration force on the microbead. (d) Setup of the cellphone-enabled image acquiring system. (e)
Comparison of the images taken by a professional microscope (top) and by the cellphone (bottom). The labeled rectangle in (b) represents the ROI
which is located at the center of the chip. Images of the particles were taken within the ROI to avoid the errors induced by the higher trapping
eﬃciency near the entrance, side wall, and exit of the chip.
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
analysis. It is noted that we used a portable microscope to
demonstrate the portability of the system. It is also possible to
use other commercially available cellphone attachments to
achieve even better portability.
The arrayed microbeads within the MW-MFC are presented
in Figure 2e. Generally, there are four cases of the trapped
microbeads within one microwell as shown in Figure 3a: (1) no
microbeads; (2) single; (3) dimer; (4) two separated
microbeads. There exists one case that two microbeads may
physically sink together into one microwell and contact each
other without the real linkage through the antigens which
would form a false dimer and hinder the detection accuracy. To
solve this issue, rotation or swing of the trapped microbeads is
introduced by changing the ﬂow rate which could stretch the
physically linked beads within the microwells. The advantage of
using a ﬂuidic method is not only due to the capability to move
microbeads into and out of the microwells by tuning the ﬂow
rate in real time, but the ﬂuidic manipulation helps to reduce
the false dimers. As Figure 3b shows, the trapped microbeads
rotate or swing under the ﬂow forces. The rotating motions of
the microbead will divide the false dimers into two single beads
as Figure 3c shows.
In addition, the ratio of trapped microbeads within the MW-
MFC, μ(deﬁned as the ratio of number of trapped microbeads
to the number of microwells), is an important factor. The
probability (P) of the trapped microbeads number (x) within
one microwell obeys to the Poisson distribution
When μ= 0.03, the probability of a false dimer in a single well
is 0.044%. To reduce the false aggregated microbeads, the value
of μis kept below 0.03 in the experiments.
PSA Detection. Figure 4a presents the size distribution of
the applied microbeads with average diameter of 5.1 μm. The
commercial counter instrument (Beckman Coulter Multisizer
3) calculates the equivalent diameter of the counted particle by
assuming it with spherical shape by ﬁtting the electrical signal.
To test the variation of the dimer number, PSA with
concentration of 12.5 ng/mL, 0.125 ng/mL, and control
group solution were, respectively, introduced into the anti-PSA
modiﬁed microbeads suspensions and incubated for 30 min.
Figure 4b shows the microbeads size distribution changes
because of the microbeads aggregation. It is easy to count the
bead number, but the size is diﬃcult to precisely distinguish
between aggregations and single beads via the diameter
distribution curve. This is due to low-eﬃcient of microbead
aggregations via protein speciﬁc interactions. Moreover, the
equivalent size of a dimer is comparable with single ones.
Considering the wide distribution range of the single bead size
shown in Figure 4a, the diameter distribution changes will be
diﬃcult to evaluate. Besides, the instrument is expensive and
sample-consuming. While, the developed MW-MFC system
can directly recognize the single and aggregated beads via
microﬂuidic chip-assisted image methods, enabling an accurate
description of the microbeads size distribution.
To demonstrate the capability of using microbeads
aggregation for protein detections, PSA of diﬀerent concen-
trations (250, 125, 12.5, 1.25, and 0.125 ng/mL, corresponding
to 7.35 nM, 3.67 nM, 0.367 nM, 0.0367 nM, and 3.67 pM)
were, respectively, detected using the developed system. Figure
4c presents the experimental results of PSA detection, showing
a negative correlation between dimer percentage and PSA
concentrations. The decrease of dimer percentage with
increased biomarker concentration can be explained by the
fact that the high concentrations of PSA will block the anti-PSA
sites on the PS beads, thus preventing particle aggregations.
Additional evidence is that the dimer percentage even dropped
to a value close to the negative control for the highest
concentration used in the experiment. It means that high
concentrations of protein would block the particle aggregation
process and increase the ratio of individual microbeads. As
lower PSA concentration leads to more aggregations in Figure
4c, it is possible to reduce the LOD of PSA detection. To
further study the LOD of this method, PSA with concentrations
of 367 and 36.7 fM was respectively tested following the same
process. The results are plotted in Figure 4d, which shows a
non-monotonous response curve of the relationship between
dimer percentage and PSA concentration. The result indicates
that the microbeads aggregation system can achieve a rather
low LOD. In practical applications, the clinical concentration of
PSA is about 10 ng/mL (corresponding to 0.3 nM),
is shown as the shaded range in Figure 4c. For real diagnosis
applications, the target sample and a diluted solution of the
target sample can both be measured, and the diﬀerence of
aggregation ratio between the two experimental groups can be
used to determine the relevant concentration region of the
In order to test the practicality of this method for clinical
applications, experiments in serum have also been performed to
detect PSA with concentrations of 0, 2, 5, 25, and 100 ng/mL
(corresponding to 0, 0.0588, 0.147, 0.735, and 2.94 nM). The
experiments were conducted following the detection procedure
as buﬀer samples. For each case, PSA serum solution was added
into and uniformly mixed with the anti-PSA functionalized
microbeads suspension. Figure 5 presents the results of the
Figure 3. (a) Four cases of the trapped microbeads within one
microwell. (b) Local rotation of the microbeads. (c) Principle of
rotation to identify the false dimer and the imaging algorithm to
distinguish and count the single bead and dimers arrayed within the
MW-MFC system. The diameter of the microwells in (a)−(c) is 20
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
serum test, which reveals a relationship consistent with that in
buﬀer solution as shown in Figure 4. The microbeads
aggregation due to the beads interactions and nonspeciﬁc
serum protein adsorption is calculated to be 4.41%, which is a
little higher than that in buﬀer solutions. The linear relationship
between dimer percentage and PSA concentration demon-
strates the great potential of the microbeads aggregation
strategy for biomarker quantiﬁcations in clinical diagnostics.
To understand the special behaviors of the formation of the
dimers, the contact probability and the aggregation kinetics of
the microbeads were further studied. For microbeads,
aggregation is less eﬃcient compared with nanoparticles
because of their diﬀerent diﬀusion rate and size. As Brownian
diﬀusion for the number of micro/nanoparticles scales with 1/
R2, the diﬀusion of microbeads is much slower compared with
which prevents the aggregation of the microbe-
As Figure 6a shows, PSA in the solution is primarily binding
with anti-PSA and adsorbed onto the surface of the microbeads
at the original period
(this reaction step is named as process
I) which follows eq 5:
where A1and Rare molecule capture sites and target
biomarker, respectively. In process I, anti-PSA on the
microbeads surface works as the capture molecule (A1)to
interact with PSA (R) in the bulk solution. Kon and Koff are the
association and dissociation rate of PSA and anti-PSA
interaction, and they deﬁne the equilibrium dissociation
Figure 4. (a) Microbeads size distribution and (b) microbeads aggregations measured via a commercial counter. Experimental results of PSA
detection using MW-MFC: (c) negative corrective relationship between dimer percentage and PSA concentration from 3.67 pM to 7.35 nM,
meanwhile the concerned concentration in diagnostic application is in the range from 30 to 300 pM. (d) Response curve of PSA detection using
microbeads aggregation method.
Figure 5. Experimental results of PSA serum detections. PSA serum
with concentrations in the region that is relevant in the clinical
diagnostics is detected, showing a negative corrective relationship.
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
where Nsis the minimum number of molecules to be captured
for detection, Dis the diﬀusion coeﬃcient of the molecules, and
ris the radius of micro/nanoscale matters. For a simpliﬁed
discussion, the molecule and microbeads are all considered as
spheres in the following analysis. The number of captured PSA
(N) by anti-PSA modiﬁed microbeads
can be calculated
with eq 7:
on 00off (7)
where N0is the real density of the receptor molecules (anti-
PSA on the microbead surface) and ρ0is the bulk concentration
of analyte molecule (PSA). The diﬀusion of analyte molecules
obeys the following eq 8:
As Figure 6b shows, PSA captured on a microbead can react
with residual anti-PSAs immobilized on an adjacent microbead
to generate a dimer (process II). For microbeads aggregation,
the interaction of PSA coated microbeads (A1R) and anti-PSA
modiﬁed microbeads (A1) can be expressed as eq 9:
HIoooAR A A R
where A1Ris the PSA coated microbeads formed in the process
Iand A12Ris the aggregated microbeads formed in the process
II. (A1R) and (A12R) represent the number of anti-PSA
modiﬁed microbeads and PSA coated microbeads, respectively.
The aggregation ratio (β) of the microbeads in the above
process can be described with Heidelberger−Kendall (H−K)
=bSARAf5.2 4 ( )( )
where bhis the steric hindrance coeﬃcient for available binding
sites on one PS bead bound to another bead (0 < bh< 1), and
Shis the corresponding steric hindrance coeﬃcient for available
binding sites on one PSA molecule bound to the anti-PSA
modiﬁed PS beads (0 < Sh< 1). fis the surface fraction
occupied by one binding site on the particle given by
where bis the radius of a circular binding site of the receptor
(PSA-binding site) and Ris the radius of the particle.
For three beads aggregation, the forming ratio is β2. And for
three more microbeads aggregation, the ratio should be higher
order (β3,β4, ...) of the dimers because they are formed on the
basis of dimers. As revealed by the experimental results in
Figure 5, the aggregation ratio is low, generally lower than 11%.
Therefore, the generation of three or three more aggregations is
more diﬃcult. In fact, there is no three or three more
aggregations observed or detected in the MW-MFC system.
Combining eqs 10 and 11, the relationship between
processes I and II can be further understood to elucidate the
diﬀerence between microbeads aggregation and nanoparticle or
molecular interactions. In the experiment of PSA detection,
when PSA is mixed with anti-PSA modiﬁed PS beads, processes
I and II would take place. The value of bhand Shis assumed to
be constant. The aggregation ratio of the two processes will be
diﬀerent as it is determined by the radius of the bounded
particles or molecules from eqs 10 and 11.Thus,the
relationship between processes I and II can be described as
in Figure 6c, in which particle size is a critical factor, revealing
the special features of microbeads aggregation strategy. In
addition, the sum value of (A1R) and (A1) equals the number of
coated anti-PSA sites on the microbeads surface; thus, the
maximum aggregation ratio obtained by eq 10 would occur
when the number of A1equals the A1R, corresponding to the
concentration locating at the peak curve in Figure 4b.
As shown in Figure 6c, the area of the sharing region, C,is
used to represent the relationship between processes I and II.
For nanoparticle aggregations, the area of Cis comparable to
the size of the biomolecules. With increase in particle size, C
decreases as the aggregation ratio scales with R−3.For
microbeads with a diameter of 5 μm, the area of Cwould be
very small and even close to zero, indicating that the reaction
process will be divided into two steps as described above. In
process II, aggregation of the microbeads takes place through
the interactions between the adsorbed PSA on microbeads and
the residual anti-PSA coated on other microbeads. At low PSA
concentration, there exist many residual anti-PSA sites on the
microbeads, while less PSA is bound to the particle for the
aggregation reaction. On the contrary, most of the anti- PSA
sites would be blocked in cases of high PSA concentrations, and
few anti-PSA binding sites remain that can allow for further
formation of the aggregates which mainly occur in process II.
This case is similar to the low PSA concentration, and
aggregation process is highly blocked as the number of PSA
sites and anti-PSA sites deviate far from maximum concen-
tration predicted by eq 11.
Equation 10 reveals that improving the concentration of
microbeads (corresponding to (A1R) and (A12R)ineq 10
would contribute to forming more aggregated dimers in process
II. Because of the limited diﬀusion of microbeads, higher beads
concentration in conﬁned space leads to higher contact
probability, which further contributes to generating more
aggregated microbeads. Microbeads of diﬀerent concentrations
diluted from the original microbeads suspension by 20, 100,
200, and 500 times have been tested for PSA detections
Figure 6. Two-step reaction model of microbeads aggregations: (a)
process I indicates the adsorption of PSA onto anti-PSA modiﬁed
microbeads, and (b) process II presents the process of microbeads
aggregations. (c) Relationship between aggregation dynamics and
particle size, which reveals that the aggregation process is dependent
on the micro/nanoparticle size.
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
following the same procedure. In the experiment, the ﬁnal
concentration of PSA is set as 0.25 μg/mL. The results are
shown in Figure 7, which indicates that higher microbeads
concentration leads to greater aggregation rate which agrees
with our theoretical analysis. This conclusion is diﬀerent from
the results of nanoparticles,
in which nanoparticle concen-
tration seems to have no inﬂuence on the aggregation result,
showing the diﬀerent feature of the microbeads. Besides, the
portable MW-MFC system developed here is suitable for
microbeads aggregation experiments using higher beads
concentration due to its inherent advantages of high
throughput and hydrodynamic control.
In this work, we have demonstrated a cellphone-enabled
portable particle analysis system for biomarker detection based
on the microbeads aggregation strategy. With the MW-MFC,
aggregated microbeads are trapped and arrayed in a high
throughput and hydrodynamically controllable way. The bead
aggregation status can be simply analyzed by processing the
images with a custom-built MATLAB program or using the
developed cellphone APP directly. The detection limit of PSA
is demonstrated to be as low as 0.125 ng/mL (3.67 pM) in
buﬀer, satisfying the clinic diagnostics demands. Moreover, the
clinically relevant range for PSA was measured in serum with a
linear relationship. The developed new platform does not
require any bulky and expensive optical or complex electrical
setup, thereby making immunoassay even cheaper and suitable
for point-of-care applications. Signiﬁcantly, the behaviors of
protein induced microbeads aggregation have been carefully
studied, and a two-step kinetics model has been developed to
explain the response curve of PSA detection, with which the
kinetics of microbeads aggregation process can be ﬁgured out.
The proposed microbeads aggregation strategy and measure-
ment system shows great potential for a wide range of
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acssen-
Structure of the portable microscope system; picture
processing with the cellphone-enabled application; the
algorithm used in the cellphone application (PDF)
Luye Mu: 0000-0002-6810-7598
Xuexin Duan: 0000-0002-7550-3951
W.C. and M.H. contributed equally.
The authors declare no competing ﬁnancial interest.
The authors gratefully acknowledge ﬁnancial support from the
Natural Science Foundation of China (NSFC No. 61176106,
91743110), and the 111 Project (B07014).
(1) Rifai, N.; Gillette, M. A.; Carr, S. A. Protein biomarker discovery
and validation: the long and uncertain path to clinical utility. Nat.
Biotechnol. 2006,24, 971−983.
(2) Andre, F.; McShane, L. M.; Michiels, S.; Ransohoff, D. F.;
Altman, D. G.; Reis-Filho, J. S.; Hayes, D. F.; Pusztai, L. Biomarker
studies: a call for a comprehensive biomarker study registry. Nat. Rev.
Clin. Oncol. 2011,8, 171−176.
(3) Poste, G. Bring on the biomarkers. Nature 2011,469, 156−157.
(4) Xu, S.; Liu, Y.; Wang, T.; Li, J. Positive potential operation of a
cathodic electrogenerated chemiluminescence immunosensor based
on luminol and graphene for cancer biomarker detection. Anal. Chem.
(5) Zhang, H.; Liu, L.; Fu, X.; Zhu, Z. Microfluidic beads-based
immunosensor for sensitive detection of cancer biomarker proteins
using multienzyme-nanoparticle amplification and quantum dotslabels.
Biosens. Bioelectron. 2013,42,23−30.
(6) Xi, Q.; Zhou, D. M.; Kan, Y. Y.; Ge, J.; Wu, Z. K.; Yu, R. Q.;
Jiang, J. H. Highly sensitive and selective strategy for microRNA
detection based on WS2 nanosheet mediated fluorescence quenching
and duplex-specific nuclease signal amplification. Anal. Chem. 2014,
(7) Choi, J. W.; Kim, G. J.; Lee, S.; Kim, J.; deMello, A.; Chang, S.-I.
A droplet-based fluorescence polarization immunoassay (dFPIA)
platform for rapid and quantitative analysis of biomarkers. Biosens.
Bioelectron. 2015,67, 497−502.
(8) Kaushik, A.; Vasudev, A.; Arya, S. K.; Pasha, S. K.; Bhansali, S.
Recent advances in cortisol sensing technologies for point-of-care
application. Biosens. Bioelectron. 2014,53, 499−512.
(9) Wang, Q.; Liu, F.; Yang, X.; Wang, K.; Wang, H.; Deng, X.
Sensitive point-of-care monitoring of cardiac biomarker myoglobin
using aptamer and ubiquitous personal glucose meter. Biosens.
Bioelectron. 2015,64, 161−164.
(10) Chen, Y.; Cheng, N.; Xu, Y.; Huang, K.; Luo, Y.; Xu, W. Point-
of-care and visual detection of P. aeruginosa and its toxin genes by
multiple LAMP and lateral flow nucleic acid biosensor. Biosens.
Bioelectron. 2016,81, 317−323.
(11) Modak, S. S.; Barber, C. A.; Geva, E.; Abrams, W. R.; Malamud,
D.; Ongagna, Y. S. Y. Rapid point-of-care isothermal amplification
Figure 7. Results of 125 ng/mL PSA detection using microbeads of
diﬀerent concentrations, showing higher microbeads concentration
leads to the more eﬃcient generation of aggregated microbeads. The
relationship between microbeads concentration and dimer percentage
is linearly ﬁtted as shown in the inset.
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
assay for the detection of malaria without nucleic acid purification.
Infect. Dis.: Res. Treat. 2016,9,1−9.
(12) Romeo, A.; Leung, T. S.; Sanchez, S. Smart biosensors for
multiplexed and fully integrated point-of-care diagnostics. Lab Chip
(13) Choi, G.; Song, D.; Shrestha, S.; Miao, J.; Cui, L.; Guan, W. A
field-deployable mobile molecular diagnostic system for malaria at the
point of need. Lab Chip 2016,16, 4341−4349.
(14) Daniels, J. S.; Pourmand, N. Label-free impedance biosensors:
Opportunities and challenges. Electroanalysis 2007,19, 1239−1257.
(15) Varshney, M.; Li, Y. Interdigitated array microelectrode based
impedance biosensor coupled with magnetic nanoparticle−antibody
conjugates for detection of Escherichia coli O157: H7 in food samples.
Biosens. Bioelectron. 2007,22, 2408−2414.
(16) Yang, T.; Wang, S.; Jin, H.; Bao, W.; Huang, S.; Wang, J. An
electrochemical impedance sensor for the label-free ultrasensitive
detection of interleukin-6 antigen. Sens. Actuators, B 2013,178, 310−
(17) Lu, Y.; Yao, Y.; Zhang, Q.; Zhang, D.; Zhuang, S.; Li, H.; Liu, Q.
Olfactory biosensor for insect semiochemicals analysis by impedance
sensing of odorant-binding proteins on interdigitated electrodes.
Biosens. Bioelectron. 2015,67, 662−669.
(18) Miyahara, Y.; Moriizumi, T.; Ichimura, K. Integrated enzyme
FETs for simultaneous detections of urea and glucose. Sens. Actuators
(19) Sethi, R. S. Transducer aspects of biosensors. Biosens. Bioelectron.
(20) Gao, X. P.; Zheng, G.; Lieber, C. M. Subthreshold regime has
the optimal sensitivity for nanowire FET biosensors. Nano Lett. 2010,
(21) Duan, X.; Li, Y.; Rajan, N. K.; Routenberg, D. A.; Modis, Y.;
Reed, M. A. Quantification of the affinities and kinetics of protein
interactions using silicon nanowire biosensors. Nat. Nanotechnol. 2012,
(22) Yan, F.; Vo-Dinh, T. Surface-enhanced Raman scattering
detection of chemical and biological agents using a portable Raman
integrated tunable sensor. Sens. Actuators, B 2007,121,61−66.
(23) Henry, A. I.; Sharma, B.; Cardinal, M. F.; Kurouski, D.; Van
Duyne, R. P. Surface-enhanced raman spectroscopy biosensing: In vivo
diagnostics and multimodal imaging. Anal. Chem. 2016,88, 6638−
(24) Zhang, M.; Huang, J.; Cui, W.; Pang, W.; Zhang, H.; Zhang, D.;
Duan, X. Kinetic studies of microfabricated biosensors using local
adsorption strategy. Biosens. Bioelectron. 2015,74,8−15.
(25) Sonato, A.; Agostini, M.; Ruffato, G.; Gazzola, E.; Liuni, D.;
Greco, G.; Travagliati, M.; Cecchini, M.; Romanato, F. A surface
acoustic wave (SAW)-enhanced grating-coupling phase-interrogation
surface plasmon resonance (SPR) microfluidic biosensor. Lab Chip
Romaschin, A.; Thompson, M. Acoustic wave biosensor for the
detection of the breast and prostate cancer metastasis biomarker
protein PTHrP. Biosens. Bioelectron. 2016,78,92−99.
(27) Liu, W.; Zhang, H.; Zhao, H.; Tang, Z.; Wang, Y.; Sun, C.; Pang,
W.; Duan, X. Comparative analysis of static and non-static assays for
biochemical sensing using on-chip integrated field effect transistors
and solidly mounted resonators. Sens. Actuators, B 2017,243, 775−
(28) Huang, L.; Harvie, G.; Feitelson, J. S.; Gramatikoff, K.; Herold,
D. A.; Allen, D. L.; Amunngama, R.; Hagler, R. A.; Pisano, M. R.;
Zhang, W. W.; Fang, X. Immunoaffinity separation of plasma proteins
by IgY microbeads: meeting the needs of proteomic sample
preparation and analysis. Proteomics 2005,5, 3314−3328.
(29) Grandhi, T. S. P.; Mallik, A.; Lin, N.; Miryala, B.; Potta, T.; Tian,
Y.; Rege, K. Aminoglycoside antibiotic-derived anion-exchange
microbeads for plasmid DNA binding and in situ DNA capture.
ACS Appl. Mater. Interfaces 2014,6, 18577−18589.
(30) Park, B. H.; Oh, S. J.; Jung, J. H.; Choi, G.; Seo, J. H.; Kim, D.
H.; Lee, E. Y.; Seo, T. S. An integrated rotary microfluidic system with
DNA extraction, loop-mediated isothermal amplification, and lateral
flow strip based detection for point-of-care pathogen diagnostics.
Biosens. Bioelectron. 2017,91, 334−340.
(31) Jans, H.; Huo, Q. Gold nanoparticle-enabled biological and
chemical detection and analysis. Chem. Soc. Rev. 2012,41, 2849−2866.
(32) Xie, X.; Xu, W.; Liu, X. Improving colorimetric assays through
protein enzyme-assisted gold nanoparticle amplification. Acc. Chem.
Res. 2012,45, 1511−1520.
(33) Leslie, D. C.; Li, J.; Strachan, B. C.; Begley, M. R.; Finkler, D.;
Bazydlo, L. A.; Barker, N. S.; Haverstick, D. M.; Utz, M.; Landers, J. P.
New detection modality for label-free quantification of DNA in
biological samples via superparamagnetic bead aggregation. J. Am.
Chem. Soc. 2012,134, 5689−5696.
(34) DuVall, J. A.; Cabaniss, S. T.; Angotti, M. L.; Moore, J. H.;
Abhyankar, M.; Shukla, N.; Mills, D. L.; Kessel, B. G.; Garner, G. T.;
Swami, N. S.; Landers, J. P. Rapid detection of Clostridium difficile via
magnetic bead aggregation in cost-effective polyester microdevices
with cell phone image analysis. Analyst 2016,141, 5637−5645.
(35) Rodriguez-Trujillo, R.; Ajine, M. A.; Orzan, A.; Mar, M. D.;
Larsen, F.; Clausen, C. H.; Svendsen, W. E. Label-free protein
detection using a microfluidic Coulter-counter device. Sens. Actuators,
(36) Rajan, N. K.; Rajauria, S.; Ray, T.; Pennathur, S.; Cleland, A. N.
A simple microfluidic aggregation analyzer for the specific, sensitive
and multiplexed quantification of proteins in a serum environment.
Biosens. Bioelectron. 2016,77, 1062−1069.
(37) Duan, X.; Mu, L.; Sawtelle, S. D.; Rajan, N. K.; Han, Z.; Wang,
Y.; Qu, H.; Reed, M. A. Functionalized Polyelectrolytes Assembling on
Nano-BioFETs for Biosensing Applications. Adv. Funct. Mater. 2015,
(38) Han, Z.; Wang, Y.; Duan, X. Biofunctional polyelectrolytes
assembling on biosensors−A versatile surface coating method for
protein detections. Anal. Chim. Acta 2017,964, 170−177.
(39) Qu, Z.; Xu, H.; Xu, P.; Chen, K.; Mu, R.; Fu, J.; Gu, H.
Ultrasensitive ELISA using enzyme-loaded nanospherical brushes as
labels. Anal. Chem. 2014,86, 9367−9371.
(40) Saffman, P. G. T. The lift on a small sphere in a slow shear flow.
J. Fluid Mech. 1965,22, 385−400.
(41) Lutz, B. R.; Chen, J.; Schwartz, D. T. Hydrodynamic tweezers: 1.
Noncontact trapping of single cells using steady streaming micro-
eddies. Anal. Chem. 2006,78, 5429−5435.
(42) Lieu, V. H.; House, T. A.; Schwartz, D. T. Hydrodynamic
tweezers: Impact of design geometry on flow and microparticle
trapping. Anal. Chem. 2012,84, 1963−1968.
(43) Tongmanee, T.; Pimpin, A.; Srituravanich, W.; Jeamsaksiri, W.;
Sripumkhai, W.; Ketpun, D.; Sailasuta, A.; Piyaviriyakul, P. November.
Development of a triangular micro well for single cell trapping
Computational study. Biomedical Engineering International Conference
(BMEiCON), 2016 8th 2015, 15758584.
(44) Barkla, H. M.; Auchterlonie, L. J. The Magnus or Robins effect
on rotating spheres. J. Fluid Mech. 1971,47, 437−447.
(45) Zhang, D.; Liu, Q. Biosensors and bioelectronics on smartphone
for portable biochemical detection. Biosens. Bioelectron. 2016,75, 273−
(46) Quesada-González, D.; Merkoçi, A. Mobile phone-based
biosensing: An emerging “diagnostic and communication”technology.
Biosens. Bioelectron. 2017,92, 549−562.
(47) Guan, W.; Chen, L.; Rane, T. D.; Wang, T. H. Droplet digital
enzyme-linked oligonucleotide hybridization assay for absolute RNA
quantification. Sci. Rep. 2015,5, 13795.
(48) Kwiatkowski, M. K.; Recker, F.; Piironen, T.; Pettersson, K.;
Otto, T.; Wernli, M.; Tscholl, R. In prostatism patients the ratio of
human glandular kallikrein to free PSA improves the discrimination
between prostate cancer and benign hyperplasia within the diagnostic
“gray zone”of total PSA 4 to 10 ng/mL. Urology 1998,52, 360−365.
(49) Haese, A.; Graefen, M.; Steuber, T.; Becker, C.; Pettersson, K.;
Piironen, T.; Noldus, J.; Huland, H.; Lilja, H. Human glandular
kallikrein 2 levels in serum for discrimination of pathologically organ-
ACS Sensors Article
ACS Sens. 2018, 3, 432−440
confined from locally-advanced prostate cancer in total PSA-levels
below 10 ng/mL. Prostate 2001,49, 101−109.
(50) Rajan, N. K.; Duan, X.; Reed, M. A. Performance limitations for
nanowire/nanoribbon biosensors. Wiley Interdisciplinary Reviews:
Nanomedicine and Nanobiotechnology 2013,5, 629−645.
(51) Nair, P. R.; Alam, M. A. Performance limits of nanobiosensors.
Appl. Phys. Lett. 2006,88, 233120.
(52) Wiklund, M.; Nord, O.; Gothäll, R.; Chernyshev, A. V.; Nygren,
P.Å.; Hertz, H. M. Fluorescence-microscopy-based image analysis for
analyte-dependent particle doublet detection in a single-step
immunoagglutination assay. Anal. Biochem. 2005,338,90−101.
ACS Sensors Article
ACS Sens. 2018, 3, 432−440