Hybrid microfluidics combined with active and passive approaches for continuous cell separation

Abstract

Microfluidics, which is classified as either active or passive, is capable of separating cells of interest from a complex and heterogeneous sample. Active methods utilise external fields such as electric, magnetic, acoustic, and optical to drive cells for separation, while passive methods utilise channel structures, intrinsic hydrodynamic forces, and steric hindrances to manipulate cells. However, when processing complex biological samples such as whole blood with rare cells, separation with a single module microfluidic device is difficult. Hybrid microfluidics is an emerging technique which utilises active and passive methods whilst fulfilling higher requirements for stable performance, versatility, and convenience, including (i) the ability to process multi-target cells, (ii) enhanced ability for multiplexed separation, (iii) higher sensitivity, and (iv) tunability for a wider operational range. This review introduces the fundamental physics and typical formats for subclasses of hybrid microfluidic devices based on their different physical fields; presents current examples of cell sorting to highlight the advantage and usefulness of hybrid microfluidics on biomedicine, and then discusses the challenges and perspective of future development and the promising direction of research in this field. This article is protected by copyright. All rights reserved.

Full text

Electrophoresis 2016,00,1–12 1
Sheng Yan1
Jun Zhang1,2
Dan Yuan1
Weihua Li1
1School of Mechanical, Materials
and Mechatronic Engineering,
University of Wollongong,
Wollongong, Australia
2School of Mechanical
Engineering, Nanjing University
of Science and Technology,
Nanjing, P. R. China
Received August 24, 2016
Revised September 29, 2016
Accepted September 29, 2016
Review
Hybrid microfluidics combined with active
and passive approaches for continuous cell
separation
Microfluidics, which is classified as either active or passive, is capable of separating cells
of interest from a complex and heterogeneous sample. Active methods utilise external
fields such as electric, magnetic, acoustic, and optical to drive cells for separation, while
passive methods utilise channel structures, intrinsic hydrodynamic forces, and steric hin-
drances to manipulate cells. However, when processing complex biological samples such
as whole blood with rare cells, separation with a single module microfluidic device is
difficult. Hybrid microfluidics is an emerging technique, which utilises active and pas-
sive methods whilst fulfilling higher requirements for stable performance, versatility, and
convenience, including (i) the ability to process multi-target cells, (ii) enhanced ability for
multiplexed separation, (iii) higher sensitivity, and (iv) tunability for a wider operational
range. This review introduces the fundamental physics and typical formats for subclasses
of hybrid microfluidic devices based on their different physical fields; presents current
examples of cell sorting to highlight the advantage and usefulness of hybrid microfluidics
on biomedicine, and then discusses the challenges and perspective of future development
and the promising direction of research in this field.
Keywords:
Cell separation / Continuous / Hybrid techniques / Microfluidics / Review
DOI 10.1002/elps.201600386
1 Introduction
The isolation and separation of cells from complex, hetero-
geneous mixtures is an upstream step for many biomedical
applications, such as clinical diagnosis [1, 2] and drug dis-
covery [3]. Due to fast sorting rates, improved accuracies and
portability, simple operating procedures, reduced cost and
reagent volume, microfluidic lab-on-a-chip devices have dis-
tinct advantages over macro-scale components, such as cen-
trifuge and flow cytometry [4]. Biomedical samples typically
consist of cells, protein, debris, viruses, and other organisms
which make microfluidic separation very challenging. For ex-
ample, since approximately 1–100 circulating tumour cells
(CTCs) are found in 1 mL of peripheral blood from a cancer
Correspondence: Professor Weihua Li, School of Mechanical, Ma-
terial and Mechatronic Engineering, University of Wollongong,
NSW 2522, Australia
E-mail: weihuali@uow.edu.au
Fax:+61-2-4221-3238
Abbreviations: AP, acoustophoresis; CTC, Circulating tumour
cells; DLD, deterministic lateral displacement; IDTs, interdig-
ital transducers; JM, human lymphocyte cell line; MCF-7,
human breast cancer cells; MOFF, multi-orifice flow frac-
tionation; MP, magnetophoresis; OEPFF, optically enhanced
pinched flow fractionation; OP, optophoresis; PC3-9, human
prostate cancer cells; PFF, pinched flow fractionation; SKBR3,
human breast cancer cells; WBC, white blood cell
patient, the rarity of CTCs is the biggest obstacle for CTC iso-
lation [5]. Another tremendous challenge of isolating CTCs
is that some are smaller than peripheral blood leukocytes,
which limits the recovery yield of microfluidic separation plat-
forms based on cell size. Although the recovery yield may be
increased by altering parameters such as the Reynolds num-
ber and external physical field, poor purity inevitably follows,
which makes it difficult to achieve a high recovery yield and
high purity simultaneously [6].
To bridge this gap, hybrid microfluidics has been emerg-
ing over the past decade. The hybrid technique is a com-
bination of active and passive approaches for processing
samples whilst fulfilling higher requirements for perfor-
mance, versatility, and convenience, including (i) the ability to
process multi-target cells, (ii) enhanced ability for multiplexed
separation, (iii) higher sensitivity, and (iv) tunability for a
wide operational range. Active technologies involve external
physical fields such as electric field [7], magnetic field [8],
acoustic streaming [9], and optical tweezer [10], whereas pas-
sive means utilising the channel geometry or intrinsic hy-
drodynamic phenomena such as pinched flow fractionation
(PFF) [11], deterministic lateral displacement (DLD) [12, 13],
hydrophoresis [14–16] and inertial microfluidics [17–19]. Ac-
tive techniques can precisely control cells of interest and be
adjusted in real time, but their flow rate is slow and so too is
Colour Online: See the article online to view Figs. 1–4 in colour.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

2S. Yan et al. Electrophoresis 2016, 00,1–12
their corresponding throughput because a long residual time
is needed to expose cells to a physical field for efficient sorting.
A passive microfluidic platform, on the other hand, can easily
and reliably manipulate cells in a high-throughput manner.
Although passive devices have simple structures and can eas-
ily be parallelised, the fixed geometry and restricted design
of passive micro-channels limit their range of operation for
different samples [20]. This is why hybrid techniques, which
combine active and passive methods, have been proposed to
overcome these shortcomings and benefit from the advan-
tages of active and passive methods. Although other types
of hybrid techniques such as active-active [21, 22] or passive-
passive [23,24] formats have been proposed, they may not take
both advantages of active and passive approaches so they are
not discussed in this review, whereas the hybrid techniques
presented here contain at least one active and one passive
manipulating technology.
The superior features of hybrid approaches mean they
can separate cells based on their dielectric property, mag-
netic property, refractive index and compressibility and also
on their volume, size, and deformability. A hybrid device can
also be used in a visual feedback system where the results of
separation are monitored under a microscope and the run-
ning parameters are adjusted in real time to obtain an opti-
mal outcome [25]. Apart from live feedback, the tunability of
a hybrid microfluidic platform relaxes the strict prerequisite
of designing and fabricating the micro-channels for specific
applications.
This review focuses on continuous flow-based cell separa-
tion using hybrid microfluidics which integrates passive and
active components in the same chip. Hybrid microfluidics
is divided into four categories according to the active meth-
ods they use: (i) dielectrophoresis (DEP)-assisted, (ii) magne-
tophoresis (MP)-assisted, (iii) acoustophoresis (AP)-assisted,
and (iv) optophoresis (OP)-assisted sorting. Within each cat-
egory, the fundamental physics behind the applied physical
field and typical cases of subclasses of hybrid microfluidics
will be presented. A variety of applications for the technology
will then be summarised to prove its uniqueness and useful-
ness. Finally, current challenges and future perspectives of
hybrid microfluidics will be discussed. We hope this review
might inspire readers who are interested in hybrid microflu-
idics to devise novel micro-chips and expand the applications
of hybrid techniques.
2 Categories of hybrid microfluidics based
on operating strategies
Based on the physical principles that govern the separation
process, hybrid microfluidics is divided into four categories:
(i) DEP-assisted, (ii) MP-assisted, (iii) AP-assisted, and (iv)
OP-assisted techniques (Fig. 1). Within each category the
theory behind each physical field, the approaches used to
tune particle motion, and advanced hybrid technologies are
presented.
2.1 Dielectrophoresis-assisted hybrid techniques
DEP refers to the movement of polarised (neutral) cells in a
non-uniform electric field where once exposed, the cells mi-
grate towards or away from the strongest part of the electric
field according to the electric permeability and conductivity
of the cell and the fluid. In a non-uniform electric field, the
polarisable cells suspended in an aquatic medium will expe-
rience a time average dielectrophoretic force. The net force
FDEP is given by [26]:
FDEP =2␲εmr3Re[K(␻)]∇ERMS 2(1)
Acve Passive
Electric
Magnec Opcal
Acousc
Microfluidic separaon
etc
DLD
Hydrophoresis
Ineral
PFF
Hybrid
DEP
N
S
-assisted MP-assisted AP-assisted OP-assisted
Figure 1. Manipulating techniques for microfluidic cellular separation. Hybrid microfluidics combines the advantages of both active and
passive methods, which is classified into DEP-assisted, MP-assisted, AP-assisted and OP-assisted technique.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2016, 00,1–12 General 3
where εmis the absolute permittivity of the suspending
medium, rdenotes the radius of the cell, ERMS2is the gradi-
ent of the square of the applied field ERMS,andK(␻) refers to
the Clausius–Mossotti (CM) factor. Re indicates the real part
of the factor, K(␻) depends on the complex permittivities of
the cell and the suspending medium, and the frequency of
the external electric field, via,
K(␻)=(ε∗
cell −ε∗
medium)/(ε∗
cell +2ε∗
medium)(2)
where ε*=ε–i␴/␻(i2=–1) is the complex permittivity, ␴
is the electrical conductivity, and ␻is the frequency of the
electric field. Since cells have a complex structure of cyto-
plasm, membrane, and nuclear envelope, a spherical multi-
shell model is utilised to measure the Re[K(␻)] [27]. The K(␻)
factor is a dominating role in DEP force that represents the
dielectric properties of cells and suspending medium under
different frequencies of the electric field applied. If the per-
mittivity of a cell is greater than the suspending medium
(K(␻)⬎0), a positive DEP (p-DEP) is generated in this mode,
where the cell migrates to the maxima of the electric field.
However, if K(␻)⬍0, the cell is repelled from the maxima
of the electric field which is referred to as a negative DEP
(n-DEP).
According to Eq. (1), the DEP force is proportional to the
square of the applied field, which means it can be tuned by the
amplitude of the external signal. To control the movement of
particles, altering either the frequency of the electric field or
fluid permittivity would be an alternative, although label-free
DEP-based devices can separate cells based on their intrin-
sic characteristics, including volume and dielectric properties
(i.e., polarisability).
DLD is a passive method proposed by Huang et al. [12] to
continuously separate beads with a high resolution down to
20 nm and also separate blood cells and isolate plasma from
whole blood [28]. Even though DLD can spatially manipulate
cells in a deterministic manner, the new devices lack the
flexibility needed when samples are varied. To bridge this
gap, DLD was combined with DEP to improve separation
efficiency. Chang and Cho [29] replaced the mechanical pillar
array with spot electrodes which could generate a virtual DLD
array with an n-DEP force generated by an AC electrical field.
The range of separable particle sizes was tuned using the
frequency and amplitude of the electric field. Beech et al.
[25] described a method which combines DEP with DLD,
thus taking advantage of the two approaches offered. Here,
platinum wires inserted into the inlets and outlets of the
DLD devices to serve as electrodes that generate an n-DEP
force around the micro-posts. When a DEP force is applied it
pushes the beads to the neighbouring lamina while shifting
from “zigzag” mode to “displacement” mode. This means the
critical size for separating micro-beads was tuned from 6 to
2␮m in a single device by altering the external AC signals
(Fig. 2A).
Hydrophoresis is a newly emerging hydrodynamic ap-
proach that utilises the steric effects between particles and
grooves [15]. These hydrophoretic devices have been used for
biological applications such as blood cell separation [14, 30],
cell cycle synchronisation [31], focusing of mammalian cell
lines [15], and cell removal [32]. However, the threshold for
hydrophoretic cell separation depends almost entirely on the
height of the channels, and that requires accurate fabrication
for a specific application. Recently, our group [33] proposed
a DEP-assisted hydrophoresis device with interdigitated elec-
trodes and hydrophoretic channels, where once the particles
experience the n-DEP force generated by the electrodes the
will be levitated towards the top wall of the channel where
the intensive particle-groove interactions allow the particles
to form a hydrophoretic ordering. Although the diameter of
bio-particles or cells was less than the critical diameter, they
were still focused in the channel due to the DEP force. Since
the lateral positions of particles can be tuned by AC signals,
particles with distinct sizes were separated based on their dif-
ferent focusing positions in the lateral direction (Fig. 2B) [34].
Inertial force can lead to the cells migrating across the
streamlines in laminar flows. Due to its merits of high
throughput, simplicity, robustness, and ease of parallelisa-
tion and fabrication, inertial microfluidics has also been in-
vestigated for bioengineering and clinical diagnosis, includ-
ing but not limited to the extraction of blood plasma [35, 36],
separation of particles and cells [37–41], solution exchange
[42–44], cell enrichment [45], isolating circulating tumour
cells (CTCs) [46–52], detecting the malaria pathogen [53, 54],
cell cycle synchronisation [55], cell encapsulation [56], and hy-
drodynamic stretching of single cells [57, 58]. Unfortunately,
the only parameter that can be adjusted is the Reynolds num-
ber for a specific inertial microfluidic device, so most inertial
microfluidic devices can only work for limited flow condi-
tions. Moreover, separation resolution is not as accurate as
active methods, and is highly dependent on experimental tri-
als [6]. Therefore, a combination of inertial microfluidics and
DEP may be an ideal solution. Moon et al. [59] proposed a de-
vice consisting of a multi-orifice flow fractionation (MOFF)
channel and a DEP channel for isolating the CTCs from blood
(Fig. 2C). The blood cells were filtrated in a massive and
high-throughput manner by inertial separation in an MOFF
channel, while the serially connected DEP sorter acted like
a precise post-processor to further enhance the separation
efficiency and purity. Apart from the hybrid device whose ac-
tive and passive components work independently, our group
proposed a DEP-inertial microfluidic platform which coupled
the inertial lift force with the DEP force [60]. This hybrid tech-
nique can modify the inertial focusing patterns in a serpen-
tine channel by a vertical n-DEP force generated by interdigi-
tated electrodes patterned on the bottom of the microchannel.
With the help from DEP force, particles were levitated along
the height of the channel, and the three-dimensional focusing
pattern of particles can be adjusted in real time.
Dielectrophoretic field-flow-fractionation (DEP-FFF) is
another approach which integrates DEP and FFF for par-
ticle or cell separation. Cells are balanced by DEP force,
sedimentation, and hydrodynamic lift forces at the equilib-
rium positions. Since the flow field is parabolic along the
height direction, cells with a faster speed at their equilibrium
positions will reach the outlet in a shorter time, and therefore
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

4S. Yan et al. Electrophoresis 2016, 00,1–12
Figure 2. (A) Schematic diagram of DLD with DEP; at a low voltage of 80 V/cm the trajectories of the 3 and 5 ␮m particles are overlaid and
there is no obvious separation, but at 265 V/cm, the critical diameter is tuned and the 5 ␮m beads are in displacement mode and then
separated from 3 ␮m beads. Reproduced from [25] with permission. (B) The layout of the DEP-assisted hydrophoretic sorter. Both large
and small particles can be focused onto the sidewall under a certain external electric field in the prefocusing region, whereas particles
of different sizes are separated in the sorting region due to distinctions in the lateral positions. Reproduced from [34]. (C) An illustration
of a microfluidic device for high-throughput separation of MCF-7 from blood cells using MOFF and DEP. The relatively larger MCF-7 cells
and a small portion of blood cells enter the centre of the channel, while most blood cells are filtered in the MOFF. Finally, MCF-7 cells are
selectively isolated via DEP. Reproduced from [59] with permission.
cells with different densities, and dielectric and mechanical
properties are separated [61,62]. However, DEP-FFF operates
at a batch-mode, which limits its throughput. To improve this
throughput, Shim et al. [63] developed a continuous flow-
based DEP-FFF for cell separation where target cells expe-
riencing a positive DEP were attracted to the bottom of the
channel, while non-target cells are pushed to a higher position
in the channel due to the negative DEP force. The target cells
are isolated by skimming the non-target cells. Since the cells
are separated in a vertical direction, the width of the channel
has almost no effect on separation efficiency. Therefore, in-
creasing the width of the channel can significantly improve
throughput.
The passive methods utilise intrinsic hydrodynamic phe-
nomena to manipulate particles, and they are highly relying
on microchannel structures, which can generate hydrody-
namic forces or physical collisions. The particles with vari-
ous sizes have different migration paths in the flow fields
induced by the specially designed microchannel, leading to a
size-based separation. The DEP-assisted approach combines
passive methods, which can remove the majority of non-target
objects from the raw samples, and DEP followed by in series,
which provide a finer process. Therefore, the high separation
efficiency and purity can be secured simultaneously. Apart
from improving separation performance, the DEP-assisted
method operating in parallel with coupled physics can extend
the operational range and enable the tunability of passive
techniques.
2.2 Magnetophoresis-assisted hybrid techniques
The magnetic force (Fmag) exerting on a particle (Eq. (3)) relies
on the difference between the magnetic susceptibility of the
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2016, 00,1–12 General 5
particle ␹pand the base fluid ␹m, as well as the volume of
the particle Vp, the magnetic flux density and gradient of the
magnetic field (B·)B,and␮0, the permeability of free space
(4␲×10−7H/m) [64]:
Fmag =(␹p−␹m)Vp
␮0
(B·∇)B.(3)
For a magnetic particle (␹p⬎0) dispersed in an aqueous
diamagnetic medium (␹m⬍0), a positive magnetophoresis
is produced in this mode, where the particle migrates to the
maxima of a magnetic field. However, if the particle is dia-
magnetic (␹p⬍0) and the medium is paramagnetic (␹m⬎
0) then the difference between the values becomes negative
and the particle is pushed towards an area of field minima.
The above equation indicates that a gradient in the mag-
netic field and a difference in susceptibility are required to
generate a magnetic force on a particle. Therefore, magnetic
cell separation is achieved by differences in the magnetic
forces, which mainly depend on the size, the difference in
magnetic susceptibility, and the field gradient. A magnetic
field is typically generated by permanent magnets or electro-
magnets, and with assistance from the ferromagnetic struc-
ture, the magnetic field gradient is enhanced to increase the
magnetic force [65–67]. The susceptibility of a mismatch is
another significant factor that affects the separation efficiency
or throughput of MP-based devices. Since biological cells con-
sist of non-magnetic particles, they can be specifically labelled
with particles of high magnetic susceptibility to distinguish
them from the sample mixtures [68,69]. Alternatively, the use
of ferrofluids and paramagnetic ions improves the magnetic
force by modifying the susceptibility of the medium [70, 71].
Seo et al. [72] proposed a hybrid cell sorter that com-
bined hydrodynamics and MP to improve the efficiency of
inertial-based separation. Their device consists of an align-
ment component with a paramagnetic line and an inertial-
based separation component. In the presence of a magnetic
field, paramagnetic particles were repelled towards the side-
walls of the channel, and diamagnetic particles were attracted
into the centre of the channel, but in the separation segment,
the particles were further separated and collected by their
sizes. Kirby et al. [73] devised a centrifuge-magnetophoretic
system to separate magnetic particles of different sizes, as
well as magnetic and diamagnetic particles of the same size.
Mizuno et al. [74] introduced a simple microfluidic platform
that utilised hydrodynamic filtration and MP to sort cells
based on their size and magnetic properties (Fig. 3A). Ini-
tially, immunomagnetic bead-conjugated cells were focused
onto the sidewall of the channel by the hydrodynamic effect,
and then sorted into individual separation lanes based on
their size difference. In the second stage, cells were driven
laterally by the magnetic force and then recovered through
multiple outlet branches on the basis of their magnetic char-
acteristics.
Researchers from the Toner’s lab [75] recently described
an inertial focusing-enhanced microfluidic CTC isolation
platform that combined the advantages of passive microflu-
idics for rare cell manipulation while integrating the mer-
its of magnetophoretic cell sorting. The proposed platform
consists of three modules within a single device: (i) separa-
tion of nucleated cells, including white blood cells (WBCs)
and CTCs, from RBCs and platelets via DLD, (ii) position-
ing cells in an almost single line using inertial focusing, and
(iii) deflection of magnetically tagged CTCs into a collection
channel using positive MP (Fig. 3B).
Giudice et al. [76] described the deterministic separation
of particles by combining MP and viscoelasticity. Their mi-
crofluidic device contained two processes; in the first module
the magnetic and diamagnetic particles were focused by ex-
ploiting fluid viscoelasticity in a straight rectangular channel,
and in the second, when the channel was exposed to a mag-
netic field, the magnetic beads were deflected from the origi-
nal stream and then separated from the diamagnetic beads.
Our group proposed a novel micro-device that combines
negative magnetophoresis and hydrophoresis for rapid par-
ticle ordering [77]. The non-magnetic particles in the diluted
ferrofluid experiencing negative magnetic forces were pushed
close to the grooves where particle-groove interaction enabled
hydrophoretic ordering of the particles. By changing the con-
centration of ferrofluid and external magnetic field, the par-
ticle focusing patterns can be tuned.
Currently, most of MP-assisted devices require the conju-
gation of cells with magnetic beads to achieve precise control.
The passive component was arranged in the front to align
the particles into a single stream. Then, a magnetic field
was applied to displace the immunomagnetic bead-labelled
cells from the main stream. Although they showed impres-
sive results, their efficacy was demonstrated with laborious
pre-treatment of samples, including media exchange and pre-
incubation of cells and magnetic beads. Instead, negative MP
is a label-free method for separation of diamagnetic cells us-
ing the difference between the magnetic susceptibility of cells
and the media. However, the negative MP-based devices typi-
cally run at a low flow rate in order to take sufficient exposure
time to the magnetic field. Integrating with other passive
components may overcome this issue. Unfortunately, very
few publications have been reported such hybrid formats.
2.3 Acoustophoresis-assisted hybrid techniques
Acoustic streaming is generated when a microfluidic chan-
nel is excited by ultrasound to a resonance mode. Here a
suspended particle with a radius r, will experience a one-
dimensional primary acoustic radiation force which is given
by the following equation [78]:
Fa=4␲r3Eacksin(2kz)␸(4)
where Eac represents the acoustic energy density, zdenotes
the distance from the pressure antinode in the wave prop-
agation axis, kis the wavenumber (2␲f/c0), where fis the
frequency of the wave, and c0is the speed of sound in the
medium, and ␾is the acoustic contrast factor defined as:
␸=␳p+2
3(␳p−␳0)
2␳p+␳0
−␳0−c2
0
3␳0c2
p
(5)
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

6S. Yan et al. Electrophoresis 2016, 00,1–12
where cpis the speed of sound in particle material and ␳pand
␳0is the density of the particle and the medium, respectively.
The direction of the acoustic radiation force depends on the
sign of the acoustic contrast factor ␾.If ␾⬎0, cells will mi-
grate towards the node of an acoustic standing wave, whereas
a negative contrast factor means the cells move to an antin-
ode. Typically, as the density of cells is slightly higher than
the physiological buffers, most cells show a positive contrast
factor under an acoustic field [79].
From Eq. (4), the acoustic radiation force is proportional
to the cube of the cell radius, so large cells experience larger
forces than small ones. Moreover, cell size is not the only
property that can be used for acoustophoretic sorting; cells
with a different density ␳por speed of sound cp, can be sep-
arated continuously according to Eq. (5). Alternatively, cells
in a mixture with different contrast factors can be sorted,
that is, cells with a positive contrasting factor are driven to
the centre of the channel (pressure node) while cells with a
Figure 3. (A) A continuous-flow microfluidic channel integrating hydrodynamic filtration and MP for cell sorting. The cells are first sorted
into corresponding channels based on their size. In the downstream, the cells conjugated with magnetic beads are separated under a
magnetic field. Reproduced from [74] with permission. (B) Schematic showing a hybrid microfluidic device which integrates DLD, inertial
focusing, and MP to isolate rare CTCs. In the DLD region, magnetically tagged CTCs and white blood cells are washed and separated from
the blood, while red blood cells and platelets are removed from the chip. In the second part, CTCs and white blood cells are well focused
via inertial focusing in an asymmetrical serpentine channel. Finally, CTCs are effectively separated in a magnetic field. Reproduced
from [75] with permission.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2016, 00,1–12 General 7
negative contrasting factor migrate towards the sidewalls of
the channel (pressure antinode) in a half-wavelength res-
onator.
Collins et al. [80] presented a virtual DLD to sort tun-
able particles by a combination of acoustophoresis where the
virtual DLD system integrated with interdigital transducers
(IDTs) which produced acoustic radiation forces at an angle
to the direction of flow. Principally, the radiation force on a
particle whose diameter is above the critical diameter over-
came the drag force and drove particles laterally along the
angle of IDTs. Smaller particles were not influenced enough
by the force field and thus maintained the original direction
of fluid flow without any disturbance. Tunability was enabled
by modifying the applied voltages, from which ⬎97% sepa-
ration of 5.0 and 6.6 ␮m particles and 87% separation of
300–500 nm particles were achieved.
Compared with DEP that requires in-chip electrodes and
has a special preference on working buffer, AP is a non-
contact and biocompatible method for cell separation. In ad-
dition, ultrasound under specific intensities has been proven
safe to biological samples [81] and has been widely used in
biomedical applications [82, 83]. Since the common issue of
these AP-based devices is low throughput, the combination
of AP with other passive techniques can bring more merits
to AP-assisted methods. AP, in turn, will allow the passive
parts to be more flexible. However, only one publication has
been found about the AP-assisted hybrid technique, which is
basically a blank area, waiting for further exploration.
2.4 Optophoresis-assisted hybrid techniques
As well as forces from an electric field, magnetic field, and
acoustic field, optical forces generated by a highly focused
optical beam of light have been rapidly developed for particle
sorting over the past decades [84]. Optical methods are con-
sidered promising due to the preservation of cell function,
precise spatial control in three dimensions, and of the ability
to manipulate small targets such as cells and molecules [85].
Typically, a focused laser beam has a Gaussian intensity pro-
file and can manipulate cells using a mismatch in the re-
fractive index between the cells and the medium to produce
optical scattering and gradient forces. While scattering forces
tend to push cells away from the centre of the beam, gra-
dient forces attract cells towards the beam maxima. Similar
to p-DEP, objects with a high refractive index travel to the
point of highest intensity and are optically trapped; this be-
came known as “optical tweezers” [86]. Moreover, the laser
wavelength, power, and trap geometry can be adjusted to sort
objects ranging from 100 nm to 100 ␮m [87].
Optical forces can be used to sort particles in active and
passive ways [88]. Cells may be sorted actively by activating
lasers when a rare event occurs to push the cell of interest
out of the original trajectory into a sorted fraction [89]. Passive
sorting has been proposed by MacDonald et al. [90]. They used
an extended, interlinked, dynamically reconfigurable, three-
dimensional optical lattice, similar to DLD, to sort protein
microcapsules by size and colloidal particles by the refractive
Figure 4. (A) Microfluidic sorting in an
optical lattice. The virtual DLD is gen-
erated by an interlinked, dynamically
reconfigurable optical lattice for opti-
cal fractionation. Reproduced from [90]
with permission. (B) Schematic dia-
gram of a microfluidic device based on
optically enhanced PFF. In the pinched re-
gion, the equilibrium positions of differ-
ent particles are modulated by an optical
scattering force. The differences in parti-
cle positions are further amplified in the
broadened region along the streamlines.
Reproduced from [91] with permission.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

8S. Yan et al. Electrophoresis 2016, 00,1–12
index. When a mixture passes through the lattice, the target
particles are deflected from their original positions, while
others remain at the same position (Fig. 4A). The efficiency
of this method of separation depends largely on how particles
respond to their optical potential, because unlike the physical
micro-posts in DLD which could suffer from clogging, the
virtual posts generated by the optical lattice led to a 45ºangular
deflection.
Pinched flow fractionation (PFF), reported by Yamada
et al. [11] was a simple and passive technique for continu-
ously separating particles in a microchannel by utilising the
spreading flow. Due to its simple structure, the PFF are eas-
ily integrated with other active techniques. To modify the
PFF, optically enhanced pinched flow fractionation (OEPFF)
has been developed by Lee et al. [91], whereby after pinch-
ing the particles to one side of the channel by the sheath
flow, they then travelled through the channel after being ex-
posed to the laser beam. This optical force moved the parti-
cles laterally in the channel, as shown in Fig. 4B, and since
this lateral displacement is associated with particle volume
and laser parameters, the equilibrium positions of different
particles at the pinched region will be enhanced. Following
this spreading flow at the broadened region, the particles be-
came separated by larger distances but then recovered with
greater efficiency. This method could also separate two kinds
of leukaemia cells [92].
Apart from an optical tweezer, an optoelectronic tweezer
is another technique that requires a photoconductive sub-
strate and a programmable display device that can project
the images onto a photoconductive surface to form tran-
sient electrodes which can precisely manipulate micro- and
nano-particles [93]. Optoelectronic tweezers thus provide a
low power approach for cell capture, transportation, and sep-
aration without labels; moreover, they not only provide high-
resolution manipulation of single particles but also require
100 000 times less optical intensity than optical tweezers [94].
Even though optoelectronic tweezers are simple, flexible, and
programmable, integration with other passive components
can enable high throughput and complicated processes [95].
Since optoelectrofluidic platforms do not need extra fluidic
components such as tubing and pumps, they are difficult
to couple with other passive techniques for complex sample
processing such as the injection of multiple flows, change of
medium buffer, and continuous sample processing. To the
best of our knowledge, no publication has reported on the
hybrid scheme based on an optoelectronic tweezer.
3 Applications in cell separation based
on hybrid microfluidics
The fundamentals of hybrid microfluidics and the various ex-
amples used to achieve tunable and continuous cell/particle
separation are described above; this section presents applica-
tions of hybrid microfluidics in cell sorting to highlight its
enormous potential for biological studies.
3.1 Blood
Mammalian blood consists of white blood cells, red blood
cells, platelets and plasma [28]. WBCs play a significant role
in the immune system, while platelets are responsible for
formatting blood clots [30]. Red blood cells are the most
common, accounting for 98% of all blood cells [96], whilst
the remainder consists of plasma, a straw coloured aque-
ous medium. Sorting or removing cells from blood has been
found to be very beneficial for diagnostic purposes such as
malaria [97], HIV [98], stroke [99], etc. This section presents
separation of blood cells and blood plasma using hybrid tech-
niques.
A cell separator using an n-DEP virtual pillar array was
used to separate red blood cells (5.4 ±1.3 ␮m-diameter)
from white blood cells (8.1 ±1.5 ␮m-diameter) [29]. Despite
their large differences in size, red blood cells and WBCs were
successfully separated with a purity of more than 99%, at a
voltage of 500 kHz, and with a 10 Vrms sinusoidal wave and a
flow rate of 0.11 ␮L/min.
Our group [100] devised a DEP-active hydrophoretic de-
vice to separate plasma from whole blood; plasma being the
liquid composition of blood that may contain proteins and
circulating nucleic acids and viruses that can be used for clin-
ical diagnostics. In our work the large cells (red blood cells
and WBCs) and small cells (platelets) were focused simulta-
neously at an appropriate flow rate of 10 ␮l/min and a voltage
of1MHz,20V
p-p AC signal, which enabled all the blood cells
to separate from the blood and plasma to be extracted with
high purity (94.2%).
Seo et al. [72] presented a hybrid method to separate red
and white blood cells using MP and inertial microfluidics.
The efficiency of separating red blood cells at the main outlet
increased from 75.2 to 86.8% when a magnetic field was
applied, whilst the classification efficiency of WBCs dropped
from 83.8 to 70.9% with the hybrid scheme.
3.2 Circulating tumour cells
Cancer metastasis is responsible for approximately 90% of
cancer-related deaths caused by CTCs [101]. CTCs and ma-
lignant cells originate from primary or second tumour sites
and enter the peripheral blood [102], but because extracting
blood from cancer patients is clinically less invasive, CTCs
potentially act like a liquid spy for prognosis, evaluation of
treatment efficacy, and studying molecular alterations under
therapy [103,104]. Since the rarity of CTCs and complexity of
blood, it is extremely difficult to isolate CTCs with both high
purity and high yield in the conventional microfluidics. How-
ever, the hybrid techniques have shown its outstanding capa-
bility to isolate CTCs using multi-physics and multi-steps.
Hybrid techniques for CTC isolation can be classified
as either positive and negative methods [105] where positive
methods typically capture the CTCs and elute haematologi-
cal cells using the expression of cell surface antigens. Glynn
et al. [106] proposed a centrifuge-magnetophoretic system for
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2016, 00,1–12 General 9
isolating HL60 cells from whole blood. The cells expressing
the HIV/AIDS relevant epitode (CD4) were bound with su-
perparamagnetic microparticles, so all the cells were loaded
into a chamber under rotation and bead-conjugated target
cells were deflected to a designated reservoir under a mag-
netic field, and almost 92% of CTCs were separated from the
whole blood. Ozkumur et al. [75] proposed an innovative CTC-
iChip to separate magnetically tagged CTCs from peripheral
blood for clinical diagnosis. It should be mentioned that the
novel chip could capture CTCs using strategies that are either
dependent on or independent of tumour surface makers, and
therefore applicable for a wide range of cancers. This chip also
benefitted from inertial focusing because it could also isolate
CTCs from whole blood in a high-throughput manner (107
cell s−1). In this work, 98.6% of SKBR3 human breast cancer
cells were recovered from the whole blood and almost 89.7%
of human prostate PC3-9 cancer cells were also captured.
Unlike the positive isolation of tumour cells, negative en-
richment is the preferred approach because it does not rely on
the biomarker expression on the tumour cells and can keep
the cells intact. Moon et al. [59] successfully separated human
breast cancer cells (MCF-7) from a sample of spiked blood by
combining MOFF and DEP techniques. The MOFF compo-
nent consisted of an alternating series of contraction channels
and expansion chambers. Once the blood samples were in-
jected through the inlet, most of the blood cells were focused
on the sides of the channel and then extracted through outlet
I, while the MCF-7 cells with unseparated blood cells headed
toward the DEP separator. The DEP separator commenced
with an expansion channel where the flow velocity decreased
dramatically (×1/200) to better fit the DEP separation. At the
first set of DEP electrodes, all the cells were driven to the
sidewalls due to the absence of sheath flow, the MCF-7 cells
migrated to the centre of the channel after passing through
the second slanted electrodes, while blood cells remained at
the same position, while the cancer cells exited through out-
let II and the residual blood cells exited through outlet III
(Fig. 2C). Up to 162-fold enrichment of the MCF-7 cells was
achieved and red and white blood cells were removed with
separation efficiencies of 99.24 and 94.23%, respectively.
Lau et al. [92] devised an integrated optofluidic plat-
form that combines the laser tweezers Roman spectroscopy
and PFF for separating leukaemia cells. Here the targeted
leukaemia cells were trapped by the laser tweezer and then
moved to a neighbouring side channel, and then the cells
of interest were sorted from the cell mixture. Sajay et al.
[107] developed a microfluidic system for negatively isolating
CTCs using magnetophoresis and membranes. In the first
step, the CD45-positive WBCs were magnetically immune
depleted from the whole blood, and step two consisted of
a microfabricated filter membrane implemented RBC de-
pletion and target cell isolation. The RBCs and platelets
freely passed through the micro slit membrane, whereas the
CTCs could not, which meant that 90% CTCs were recov-
ered from the blood sample. Another negative isolation of
CTCs is based on the dielectric signature [63]. This technique
combines DEP and FFF to separate tumour cells with high
throughput (10 mL/h). The tumour cells and peripheral
blood mononuclear cells (PBMCs) with different dielectric
properties were brought to different heights due to a balance
between the DEP force, sedimentation, and hydrodynamic lift
force, and eventually collected from different outlets. Various
CTC types spiked into PBMCs were recovered at an average
rate of 75%.
3.3 Other cell types
Apart from separation of blood cells and CTCs, hybrid mi-
crofluidics is also used to sort and isolate a variety of other
cells. However, due to differences in membrane potentials
under a non-uniform electric field, viable and non-viable cells
were separated using a modified DEP-active hydrophoretic
device [34] integrated with the focusing and sorting parts,
where cells were concentrated in the focusing section and
then separated in the sorting region. Based on the distinct
dielectric property of viable and non-viable Chinese Hamster
Ovary cells at a medium conductivity of 0.03 S/m, the live cells
exerting a larger DEP force focused well and were then sep-
arated from the sample cell with a purity of 99.6%. Another
approach is to tag target cells with magnetic beads so they can
be manipulated with a magnetic field. A magnetophoresis-
integrated hydrodynamic filtration system for cell sorting has
been reported by Mizuno et al. [74]. The JM (human lym-
phocyte cell line) cells conjugated with anti-CD 4 immuno-
magnetic beads and unlabelled HeLa cells were separated at
a purification ratio higher than 90%.
4 Challenges and perspectives
This review highlights the recent development of continuous
cell sorting in hybrid microfluidic devices. Since making its
appearance, hybrid microfluidics has attracted a substantial
amount of press and interest, but there are big challenges
that must be addressed to make it more practical. First, com-
mercialising hybrid technologies into industry and hospitals
will keep this field attractive and dynamic, but most of the mi-
crofluidic technologies reviewed here are still in the prototype
or proof-of-concept stage; only a few have the capability for
clinical usage. Commercial investment from industrial com-
panies or medical centres is one way to satisfy their needs,
but in the interim, researchers should fight to determine what
can be done for biomedicine using hybrid techniques.
Second, while the marriage of active and passive methods
combines their strengths, it also includes their weaknesses
because positive MP requires pre-incubation to conjugate
the cells with magnetic particles, which is time-consuming,
whereas negative MP is a label-free method for manipulat-
ing diamagnetic cells. Here, the magnetic gradient in the
medium can form easily by controlling the concentration of
paramagnetic nano-particles or ions such as ferrofluid [108]
or paramagnetic salt [67]. However, the biocompatibility of
the paramagnetic medium is a huge challenge because while
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

10 S. Yan et al. Electrophoresis 2016, 00,1–12
E. coli and yeast cells are sorted using a commercial ferrofluid
with a continuous-flow fashion, [108] few have demonstrated
the capability of processing mammalian cells. Shen et al. [67]
successfully achieved a label-free separation of U937 cells
from red blood cells using a paramagnetic salt called Gd-
DTPA, but higher concentrations of Gd-DTPA might cause
cell apoptosis due to the effect of osmotic pressure. Zhao
et al. [109] provided a custom-made ferrofluid for label-free
and continuous-flow separation of HeLa cells and blood cells,
but further work is needed to develop a biocompatible para-
magnetic solution for mammalian cells otherwise MP will be
pushed to its limit and become unattractive. The appropriate
paramagnetic solution should have such feature: (i) the PH
value must be maintained around 7, which is the best situ-
ation for cell culture, and (ii) the salt concentration, toxicity,
and surfactant should cater for the physiological conditions
to avoid cell death.
Like the drawbacks existing in MP, a DEP buffer will have
a similar issue because DEP-based devices do not work very
well with physiological media (conductivities ⬎1 S/m), where
cells become less polarisable than the medium [110, 111],
which is why they usually operate in a low conductive
medium. Here, the sucrose solution typically acts as the base
of a DEP buffer whose conductivity is adjusted by adding
phosphate buffered saline to achieve a better separation per-
formance [112]. However, it is difficult to balance the osmotic
pressure of a solution which is suitable for cell viability with
its conductivity, and therefore an alternative biocompatible
DEP buffer should be chosen to meet the requirements of
cell polarisation and viability.
In contrast, AP and OP have non-special requirements
on the cell medium where typically, cell samples are pre-
pared in a physiological media such as PBS, where cells re-
main viable due to isosmotic pressure and the cells sorted
in PBS can be used for cell fixation, enumeration, and stain-
ing [113]. Moreover, the viable cells being processed by AP
or OP are valuable for further cell culture [114] and drug
screening [115, 116], but unfortunately AP-assisted and OP-
assisted microfluidics is still a blank field that needs further
exploitation.
The active and passive components in hybrid microflu-
idics currently operate in series with independent physics
[59,75]; or in parallel with coupled physics [25,33]. The main
concern is the incompatible throughput between each field
of physics. While passive microfluidics can only work at high
speed, like inertial microfluidics, the flow rate in active mi-
crofluidics is relatively low. The hybrid technique with in-
dependent physics has simpler mechanism because each
section has a weak connection, but the interface between
sections, such as the flow rate or pressure should be bal-
anced carefully, otherwise the downstream section cannot
work properly. In contrast, coupled physics is more compli-
cated but versatile for cell manipulation; for instance, hydro-
dynamic filtration, hydrophoresis and DLD are easy to couple
with the active forces because their throughputs are compa-
rable to their active peers and therefore more combinations
are possible in this format.
As a newly developed technology, hybrid microfluidics
has shown enormous potential at cell separation, but there
is still a vast and raw territory urgently waiting to be ex-
plored. The practical applications of hybrid microfluidics
should be demonstrated further because this will attract more
researchers to this field and more industrial partners to in-
vest in this technology. We believe that hybrid microfluidics
remains at the forefront of the next generation of microflu-
idic devices because it has the capacity to accurately separate
complex samples and provide real-time feedback.
This work is supported by the University of Wollongong-
China Scholarship Council joint scholarships.
The authors have declared no conflict of interest.
5 References
[1] Chin, C. D., Laksanasopin, T., Cheung, Y. K., Steinmiller,
D., Linder, V., Parsa, H., Wang, J., Moore, H., Rouse,
R., Umviligihozo, G., Karita, E., Mwambarangwe, L.,
Braunstein, S. L., van de Wijgert, J., Sahabo, R., Just-
man, J. E., El-Sadr, W., Sia, S. K., Nat. Med. 2011, 17,
1015–1019.
[2] Schulte, T. H., Bardell, R. L., Weigl, B. H., Clin. Chim.
Acta. 2002, 321, 1–10.
[3] Dittrich, P. S., Manz, A., Nat. Rev. Drug Discovery. 2006,
5, 210–218.
[4] Bhagat, A. A. S., Bow, H., Hou, H. W., Tan, S. J., Han, J.,
Lim, C. T., Med. Biol. Eng. Comput. 2010, 48, 999–1014.
[5] Hyun, K.-A., Jung, H.-I., Lab Chip. 2014, 14, 45–56.
[6] Zhang, J., Yan, S., Yuan, D., Alici, G., Nguyen, N.-T.,
Warkiani, M. E., Li, W., Lab Chip. 2016, 16, 10–34.
[7] C¸etin, B.,Li, D.,Electrophoresis. 2011, 32, 2410–2427.
[8] Forbes, T. P., Forry, S. P., Lab Chip. 2012, 12, 1471–1479.
[9] Wang, Z., Zhe, J., Lab Chip. 2011, 11, 1280–1285.
[10] Grier, D. G., Nature. 2003, 424, 810–816.
[11] Yamada, M., Nakashima, M., Seki, M., Anal. Chem.
2004, 76, 5465–5471.
[12] Huang, L. R., Cox, E. C., Austin, R. H., Sturm, J. C.,
Science. 2004, 304, 987–990.
[13] McGrath, J., Jimenez, M., Bridle, H., Lab Chip. 2014, 14,
4139–4158.
[14] Choi, S., Song, S., Choi, C., Park, J.-K., Lab Chip. 2007,
7, 1532–1538.
[15] Choi, S., Song, S., Choi, C., Park, J. K., Small. 2008, 4,
634–641.
[16] Song, S., Choi, S., J. Chromatogr. A. 2013, 1302,
191–196.
[17] Di Carlo, D., Lab Chip. 2009, 9, 3038–3046.
[18] Zhang, J., Li, W., Li, M., Alici, G., Nguyen, N. T., Mi-
crofluid. Nanofluid. 2014, 17, 305–316.
[19] Zhang, J., Yan, S., Sluyter, R., Li, W., Alici, G., Nguyen,
N.-T., Sci. Rep. 2014, 4, 4527.
[20] Sheng, Y., Jun, Z., Chao, P., Dan, Y., Gursel, A., Haiping,
D., Yonggang, Z., Weihua, L., J. Micromech.Microeng.
2015, 25, 084010.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2016, 00,1–12 General 11
[21] Kim, U., Soh, H. T., Lab Chip. 2009, 9, 2313–2318.
[22] Adams, J. D., Th´
evoz, P., Bruus, H., Soh, H. T., Appl.
Phys. Lett. 2009, 95, 254103.
[23] Beech, J. P., Tegenfeldt, J. O., Lab Chip 2008, 8, 657–659.
[24] Bhagat, A. A. S., Hou, H. W., Li, L. D., Lim, C. T., Han, J.,
Lab Chip. 2011, 11, 1870–1878.
[25] Beech, J. P., Jonsson, P., Tegenfeldt, J. O., Lab Chip.
2009, 9, 2698–2706.
[26] Jones, T. B., Electromechanics of Particles. Cambridge
Unviersity Press, Cambridge, 1995.
[27] Tang, S. Y., Zhang, W., Baratchi, S., Nasabi, M.,
Kalantar-Zadeh, K., Khoshmanesh, K., Anal. Chem.
2013, 85, 6364–6371.
[28] Davis, J. A., Inglis, D. W., Morton, K. J., Lawrence,
D. A., Huang, L. R., Chou, S. Y., Sturm, J. C., Austin,
R. H., Proc.Natl.Acad.Sci.U.S.A.2006, 103,
14779–14784.
[29] Chang, S., Cho, Y.-H., Lab Chip. 2008, 8, 1930–1936.
[30] Choi, S., Ku, T., Song, S., Choi, C., Park, J.-K., Lab Chip.
2011, 11, 413–418.
[31] Choi, S., Song, S., Choi, C., Park, J.-K., Anal. Chem.
2009, 81, 1964–1968.
[32] Song, S., Kim, M. S., Choi, S., Small. 2014, 10,
4123–4129.
[33] Yan, S., Zhang, J., Li, M., Alici, G., Du, H., Sluyter, R.,
Li, W., Sci. Rep. 2014, 4, 5060.
[34] Yan, S., Zhang, J., Yuan, Y., Lovrecz, G., Alici, G., Du, H.,
Zhu, Y., Li, W., Electrophoresis. 2015, 36, 284–291.
[35] Lee, M. G., Choi, S., Kim, H. J., Lim, H. K., Kim, J. H.,
Huh, N., Park, J. K., Appl. Phys. Lett. 2011, 98, 253702.
[36] Mach, A. J., Di Carlo, D., Biotechnol. Bioeng. 2010, 107,
302–311.
[37] Warkiani, M. E., Tay, A. K. P., Guan, G., Han, J., Sci. Rep.
2015, 5, Art. no. 11018.
[38] Zhang, J., Yan, S., Sluyter, R., Li, W., Alici, G., Nguyen,
N.-T., Sci. Rep. 2014, 4, Art. no. 4527.
[39] Lee,M.G.,Choi,S.,Park,J.K.,J. Chromatogr. A. 2011,
1218, 4138–4143.
[40] Kuntaegowdanahalli, S. S., Bhagat, A. A. S., Kumar, G.,
Papautsky, I., Lab Chip. 2009, 9, 2973–2980.
[41] Wu, L., Guan, G., Hou, H. W., Bhagat, A. A. S., Han, J.,
Anal. Chem. 2012, 84, 9324–9331.
[42] Sollier, E., Amini, H., Go, D., Sandoz, P., Owsley, K., Di
Carlo, D., Microfluid. Nanofluid. 2015, 19, 53–65.
[43] Dudani, J. S., Gossett, D. R., Henry, T., Lamm, R. J.,
Kulkarni, R. P., Di Carlo, D., Biomicrofluid. 2015, 9,
014112.
[44] Gossett, D. R., Tse, H. T. K., Dudani, J. S., Goda, K.,
Woods, T. A., Graves, S. W., Di Carlo, D., Small. 2012,
8, 2757–2764.
[45] Martel, J. M., Smith, K. C., Dlamini, M., Pletcher, K.,
Yang, J., Karabacak, M., Haber, D. A., Kapur, R., Toner,
M., Sci. Rep. 2015, 5, Art. no. 11300.
[46] Sollier, E., Go, D. E., Che, J., Gossett, D. R., O’Byrne, S.,
Weaver, W. M., Kummer, N., Rettig, M., Goldman, J.,
Nickols, N., Lab Chip. 2014, 14, 63–77.
[47] Che, J., Mach, A. J., Go, D. E., Talati, I., Ying, Y., Rao, J.,
Kulkarni, R. P., Di Carlo, D., PLoS One. 2013, 8, e78194.
[48] Mach,A.J.,Kim,J.H.,Arshi,A.,Hur,S.C.,DiCarlo,D.,
Lab Chip. 2011, 11, 2827–2834.
[49] Lee, M. G., Shin, J. H., Bae, C. Y., Choi, S., Park, J.-K.,
Anal. Chem. 2013, 85, 6213–6218.
[50] Sun, J., Li, M., Liu, C., Zhang, Y., Liu, D., Liu, W., Hu, G.,
Jiang, X., Lab Chip. 2012, 12, 3952–3960.
[51] Hou, H. W., Warkiani, M. E., Khoo, B. L., Li, Z. R., Soo,
R.A.,Tan,D.S.-W.,Lim,W.-T.,Han,J.,Bhagat,A.A.S.,
Lim, C. T., Sci. Rep. 2013, 3, Art. no. 1259.
[52] Warkiani, M. E., Guan, G., Luan, K. B., Lee, W. C., Bha-
gat, A. A. S., Chaudhuri, P. K., Tan, D. S.-W., Lim, W. T.,
Lee, S. C., Chen, P. C., Lab Chip. 2014, 14, 128–137.
[53] Warkiani, M. E., Tay, A. K. P., Khoo, B. L., Xiaofeng, X.,
Han, J., Lim, C. T., Lab Chip. 2015, 15, 1101–1109.
[54] Birch, C. M., Hou, H. W., Han, J., Niles, J. C., Sci. Rep.
2015, 5, 11347.
[55] Lee, W. C., Bhagat, A. A. S., Huang, S., Van Vliet, K. J.,
Han, J., Lim, C. T., Lab Chip. 2011, 11, 1359–1367.
[56] Kemna, E. W., Schoeman, R. M., Wolbers, F., Vermes,
I., Weitz, D. A., van den Berg, A., Lab Chip. 2012, 12,
2881–2887.
[57] Gossett, D. R., Henry, T., Lee, S. A., Ying, Y., Lindgren,
A. G., Yang, O. O., Rao, J., Clark, A. T., Di Carlo, D., Proc.
Natl. Acad. Sci. U. S. A. 2012, 109, 7630–7635.
[58] Dudani, J. S., Gossett, D. R., Henry, T., Di Carlo, D., Lab
Chip. 2013, 13, 3728–3734.
[59] Moon, H. S., Kwon, K., Kim, S. I., Han, H., Sohn, J., Lee,
S., Jung, H. I., Lab Chip. 2011, 11, 1118–1125.
[60] Zhang, J., Yan, S., Alici, G., Nguyen, N.-T., Di Carlo, D.,
Li, W., RSC Advances. 2014, 4, 62076–62085.
[61] Gascoyne, P. R. C., Noshari, J., Anderson, T. J., Becker,
F. F. , Electrophoresis. 2009, 30, 1388–1398.
[62] Gascoyne, P. R. C., Anal. Chem. 2009, 81, 8878–8885.
[63] Shim, S., Stemke-Hale, K., Tsimberidou, A. M., Noshari,
J., Anderson, T. E., Gascoyne, P. R., Biomicrofluidics.
2013, 7, 011807.
[64] Peyman, S. A., Kwan, E. Y., Margarson, O., Iles, A.,
Pamme, N., J. Chromatogr. A 2009, 1216, 9055–9062.
[65] Adams, J. D., Kim, U., Soh, H. T., Proc. Natl. Acad. Sci.
U. S. A. 2008, 105, 18165–18170.
[66] Lee, H., Jung, J., Han, S.-I., Han, K.-H., Lab Chip. 2010,
10, 2764–2770.
[67] Shen, F., Hwang, H., Hahn, Y. K., Park, J.-K., Anal. Chem.
2012, 84, 3075–3081.
[68] Lou, X., Qian, J., Xiao, Y., Viel, L., Gerdon, A. E., Lagally,
E. T., Atzberger, P., Tarasow, T. M., Heeger, A. J., Soh, H.
T. , Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 2989–2994.
[69] Pamme, N., Wilhelm, C., Lab Chip. 2006, 6, 974–980.
[70] Winkleman, A., Perez-Castillejos, R., Gudiksen, K. L.,
Phillips, S. T., Prentiss, M., Whitesides, G. M., Anal.
Chem. 2007, 79, 6542–6550.
[71] Zhu, T., Cheng, R., Liu, Y., He, J., Mao, L., Microfluid.
Nanofluid. 2014, 17, 973–982.
[72] Hye-Kyoung, S., Yong-Ho, K., Hyun-Ok, K., Yong-Jun,
K., J. Micromech.Microeng. 2010, 20, 095019.
[73] Kirby, D., Siegrist, J., Kijanka, G., Zavattoni, L., Sheils,
O., O’Leary, J., Burger, R., Ducr´
ee, J., Microfluid.
Nanofluid. 2012, 13, 899–908.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

12 S. Yan et al. Electrophoresis 2016, 00,1–12
[74] Mizuno, M., Yamada, M., Mitamura, R., Ike, K., Toyama,
K., Seki, M., Anal. Chem. 2013, 85, 7666–7673.
[75] Ozkumur, E., Shah, A. M., Ciciliano, J. C., Emmink, B.
L., Miyamoto, D. T., Brachtel, E., Yu, M., Chen, P.-i.,
Morgan, B., Trautwein, J., Sci. Transl. Med. 2013, 5,
179ra147–179ra147.
[76] Del Giudice, F., Madadi, H., Villone, M. M., D’Avino, G.,
Cusano, A. M., Vecchione, R., Ventre, M., Maffettone, P.
L., Netti, P. A., Lab Chip. 2015, 15, 1912–1922.
[77] Yan, S., Zhang, J., Chen, H., Yuan, D., Alici, G., Du, H.,
Zhu, Y., Li, W., Biomed. Microdevices. 2016, 18,1–9.
[78] Lenshof, A., Magnusson, C., Laurell, T., Lab Chip 2012,
12, 1210–1223.
[79] Augustsson, P., Laurell, T., Lab Chip. 2012, 12,
1742–1752.
[80] Collins, D. J., Alan, T., Neild, A., Lab Chip. 2014, 14,
1595–1603.
[81] Lin, S.-C. S., Mao, X., Huang, T. J., Lab Chip. 2012, 12,
2766–2770.
[82] Zalis, M. C., Reyes, J. F., Augustsson, P., Holmqvist, S.,
Roybon, L., Laurell, T., Deierborg, T., Integr. Biol. 2016,
8, 332–340.
[83] Chen, Y., Wu, M., Ren, L., Liu, J., Whitley, P. H., Wang,
L., Huang, T. J., Lab Chip. 2016, 16, 3466–3472.
[84] Lenshof, A., Laurell, T., Chem.Soc.Rev.2010, 39,
1203–1217.
[85] Neuman, K. C., Nagy, A., Nat. Meth. 2008, 5, 491–505.
[86] Ashkin, A., Proc. Natl. Acad. Sci. U. S. A. 1997, 94,
4853–4860.
[87] Sajeesh, P., Sen, A. K., Microfluid. Nanofluid. 2013,
1–52.
[88] Jon ´
aˇ
s, A., Zem ´
anek, P., Electrophoresis. 2008, 29,
4813–4851.
[89] Perroud, T. D., Kaiser, J. N., Sy, J. C., Lane, T. W., Branda,
C. S., Singh, A. K., Patel, K. D., Anal. Chem. 2008, 80,
6365–6372.
[90] MacDonald, M. P., Spalding, G. C., Dholakia, K., Nature.
2003, 426, 421–424.
[91] Lee, K. H., Kim, S. B., Lee, K. S., Sung, H. J., Lab Chip.
2011, 11, 354–357.
[92] Lau, A. Y., Lee, L. P., Chan, J. W., Lab Chip. 2008, 8,
1116–1120.
[93] Wu, M. C., Nat. Photon. 2011, 5, 322–324.
[94] Chiou, P. Y., Ohta, A. T., Wu, M. C., Nature. 2005, 436,
370–372.
[95] Hwang, H., Park, J.-K., Lab Chip. 2011, 11, 33–47.
[96] Kersaudy-Kerhoas, M., Sollier, E., Lab Chip. 2013, 13,
3323–3346.
[97] Birch, C. M., Hou, H. W., Han, J., Niles, J. C., Sci. Rep.
2015, 5, 11347.
[98] Kim, Y.-G., Moon, S., Kuritzkes, D. R., Demirci, U.,
Biosens. Bioelectron. 2009, 25, 253–258.
[99] Rainer, T. H., Wong, L. K., Lam, W., Yuen, E., Lam, N. Y.,
Metreweli, C., Lo, Y. D., Clin. Chem. 2003, 49, 562–569.
[100] Yan, S., Zhang, J., Alici, G., Du, H., Zhu, Y., Li, W., Lab
Chip. 2014, 14, 2993–3003.
[101] Gupta, G. P., Massagu ´
e, J., Cell. 2006, 127, 679–695.
[102] Aggarwal, C., Meropol, N. J., Punt, C. J., Iannotti, N.,
Saidman, B. H., Sabbath, K. D., Gabrail, N. Y., Picus, J.,
Morse, M. A., Mitchell, E., Miller, M. C., Cohen, S. J.,
Ann. Oncol. 2013, 24, 420–428.
[103] Lianidou, E. S., Strati, A., Markou, A., Crit. Rev. Clin.
Lab. Sci. 2014, 51, 160–171.
[104] Qin, X., Park, S., Duffy, S. P., Matthews, K., Ang,
R. R., Todenhofer, T., Abdi, H., Azad, A., Bazov, J.,
Chi, K. N., Black, P. C., Ma, H., Lab Chip. 2015, 15,
2278–2286.
[105] Jin, C., McFaul, S. M., Duffy, S. P., Deng, X., Tavassoli,
P.,Black,P.C.,Ma,H.,Lab Chip. 2014, 14, 32–44.
[106] Brotherton, C. M., Sun, A. C. and Davis, R. H., Mi-
crofluid. Nanofluid., 2008, 5, 43–53.
[107] Sajay, B., Chang, C.-P., Ahmad, H., Khuntontong, P.,
Wong, C., Wang, Z., Puiu, P., Soo, R., Rahman, A.,
Biomed. Microdevices. 2014, 16, 537–548.
[108] Zhu, T., Cheng, R., Lee, S. A., Rajaraman, E., Eiteman,
M. A., Querec, T. D., Unger, E. R., Mao, L., Microfluid.
Nanofluid. 2012, 13, 645–654.
[109] Zhao, W., Zhu, T., Cheng, R., Liu, Y., He, J., Qiu, H.,
Wang, L., Nagy, T., Querec, T. D., Unger, E. R., Adv.
Funct. Mater. 2015, 26, 3990–3998.
[110] Park, S., Zhang, Y., Wang, T.-H., Yang, S., Lab Chip. 2011,
11, 2893–2900.
[111] Yan, S., Zhang, J., Chen, H., Alici, G., Du, H., Zhu, Y., Li,
W., Biomicrofluidics. 2014, 8, 064115.
[112] Cheng, I. F., Huang, W.-L., Chen, T.-Y., Liu, C.-W., Lin,
Y.-D., Su, W.-C., Lab Chip. 2015, 15, 2950–2959.
[113] Hou, H. W., Warkiani, M. E., Khoo, B. L., Li, Z. R., Soo,
R.A.,Tan,D.S.-W.,Lim,W.-T.,Han,J.,Bhagat,A.A.S.,
Lim, C. T., Sci. Rep. 2013, 3, 1259.
[114] Chen, H., Sun, J., Wolvetang, E., Cooper-White, J., Lab
Chip. 2015, 15, 1072–1083.
[115] Chan, C. Y., Huang, P.-H., Guo, F., Ding, X., Kapur, V.,
Mai, J. D., Yuen, P. K., Huang, T. J., Lab Chip. 2013, 13,
4697–4710.
[116] Zhao, H., Zhou, L., Zhang, Q., Zhou, X., Zhang, Y., Chen,
H., Du, Y., Lab Chip. 2015, 15, 3481–3494.
C2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com