Access to this full-text is provided by Wiley.
Content available from Advanced Science
This content is subject to copyright. Terms and conditions apply.
FULL PAPER
www.advancedscience.com
Controllable Cell Deformation Using Acoustic Streaming for
Membrane Permeability Modulation
Xinyi Guo, Mengjie Sun, Yang Yang, Huihui Xu, Ji Liu, Shan He, Yanyan Wang, Linyan Xu,
Wei Pang, and Xuexin Duan*
Hydrodynamic force loading platforms for controllable cell mechanical
deformation play an essential role in modern cell technologies. Current
systems require assistance from specific microstructures thus limiting the
controllability and flexibility in cell shape modulation, and studies on real-time
3D cell morphology analysis are still absent. This article presents a novel
platform based on acoustic streaming generated from a gigahertz device for
cell shape control and real-time cell deformation analysis. Details in cell
deformation and the restoration process are thoroughly studied on the
platform, and cell behavior control at the microscale is successfully achieved
by tuning the treating time, intensity, and wave form of the streaming. The
application of this platform in cell membrane permeability modulation and
analysis is also exploited. Based on the membrane reorganization during cell
deformation, the effects of deformation extent and deformation patterns on
membrane permeability to micro- and macromolecules are revealed. This
technology has shown its unique superiorities in cell mechanical
manipulation such as high flexibility, high accuracy, and pure fluid force
operation, indicating its promising prospect as a reliable tool for cell property
study and drug therapy development.
1. Introduction
Living cells are always undergoing continuous internal and ex-
ternal mechanical forces. Response to these forces can trig-
ger the deformation of cells, which will further influence
X. Guo, M. Sun, Y. Yang, H. Xu, J. Liu, S. He, Dr. Y. Wang, Prof. X. Duan
State Key Laboratory of Precision Measuring Technology & Instruments
Tianjin University
Tianjin 300072, China
E-mail: xduan@tju.edu.cn
Dr.L.Xu,Prof.W.Pang
College of Precision Instrument and Opto-electronics Engineering
Tianjin University
Tianjin 300072, China
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/advs.202002489
© 2020 The Authors. Advanced Science published by Wiley-VCH GmbH.
This is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in
any medium, provided the original work is properly cited.
DOI: 10.1002/advs.202002489
the functions and biological proper-
ties of cells.[1–4 ] Exogenous mechanical
stimulation can regulate the growth,[5]
differentiation,[6] intercellular signaling
of cells,[7] providing important guidance
for tissue engineering and regenerative
medicine.[8,9 ] Besides, controllable force ex-
ertion and cell extrusion are demonstrated
to have profound effects on transmembrane
transportation by producing transient dis-
ruptions on the membrane,[10,11 ] which
can be used for drug targeting[12] and gene
engineering.[13] However, details on the
mechanism of cell deformation process
and the relationship between deformation
and membrane permeability are still less
studied. Thus, exploiting microscaled force
manipulation systems, which enables me-
chanical operations at cellular level with
high controllability and real-time analysis
become essential.
Recent advances in micro- and nanotech-
nology have extended the research tools for
cell force loading and mechanical deforma-
tion studies. These tools can be generally
classified into contact and noncontact approaches, according to
their way of applying forces to cells. A majority of these ap-
proaches, such as micropipette aspiration,[14,15 ] atomic force mi-
croscopy (AFM),[16] microchannel cell squeezing,[17,18] magnetic
manipulators,[19,20 ] and surface microstructures[5,8,21,22 ] require
direct contact between cell membrane and external apparatus.
Each of these techniques has its own strengths and its most ap-
propriate applications. However, due to the direct contact in these
methods, contamination of the surfaces and irreversible dam-
ages to cells are typical issues. Noncontact approaches such as
optical stretchers,[23,24 ] acoustic tweezers[25] and dielectrophore-
sis (DEP)[26,27 ] taking advantage of field forces make a step for-
ward in controlling cell deformation. Among these approaches,
hydrodynamic method, using the flow force generated from bub-
ble oscillation[28,29 ] or shear and extensional flow,[30–33] has been
successfully applied in cell deformation studies. Nevertheless,
the assistance of bubbles or specific microfluidic structures is
required in these hydrodynamic systems. Bubbles may increase
the difficulty to achieve highly accurate control due to the un-
certainty in bubble generation and positioning, and specific mi-
crofluidic structures in cells’ culturing environments will limit
the flexibility in tuning the force pattern and the in situ cell
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (1 of 11)
www.advancedsciencenews.com www.advancedscience.com
culture and analysis. Besides, cell shapes analyzed from micro-
graphs or 2D slices limit the perceptions of real cell morphol-
ogy and the possibility for multidirectional analysis. Thus, a 3D
real-time mechanical deformation system which can achieve pre-
cise control with high flexibility, and have no restrictions on ex-
ogenous materials and operating environments, is still high in
demand.
Here, we present a noncontact platform for the controllable
and tunable cell deformation based on the hydrodynamic force
generated from acoustic streaming under gigahertz (GHz) res-
onator excitation. Previous studies from our group have revealed
the ability of such microfabricated resonator to generate local-
ized and high-speed fluid motion due to the rapid attenuation
of acoustic energy in liquid environment, and its biomedical ap-
plications for microscale fluidic mixing,[34] tuning biomolecule–
surface interactions[35,36 ] and particle manipulations.[37] We hav e
reported a drug delivery method based on gigahertz resonator,[38]
in which the delivery mechanism and the acoustic streaming gen-
eration were not fully explored. Following studies have tried to
model the cell response to gigahertz acoustic field using unil-
amellar vesicles,[39] in which material exchange over the syn-
thetic membrane was observed. In this article, we integrated
the miniaturized acoustic resonator with the confocal micro-
scope for the real time 3D analysis of the cell mechanical re-
sponses under acoustic stimulation. The platform can realize
gradual and restorable cell deformation with tunable deforma-
tion extent by power adjustment. We also realized reciprocat-
ing and pulsating cell shape change under square-waved fluid
excitation. To our knowledge, this is the first real-time 3D ob-
servation of tunable cell deformation under complex hydrody-
namic force patterns. The application of this system in mod-
ulating cell membrane permeability were further studied. Re-
organization of the cell membrane during deformation pro-
cess was observed, and relationship between the cell deforma-
tion character and the cell membrane permeability was thor-
oughly studied. A novel phenomenon is revealed by our sys-
tem that fast and periodic cell deformation under hydrodynamic
stimulation will facilitate the membrane permeability for in-
tracellular transportation applications. Comparing with acous-
tics at lower frequencies, this platform has shown its partic-
ular advantage on providing large body force for fluid driven
due to the fast-attenuation of gigahertz vibration, which facili-
tates the generation of localized high-speed stream flow at ap-
proximately m s−1level and enables the microscale force con-
trol over a wide range of several tens of micro Newton. No as-
sistance from external structures, such as sharp edges, needles,
or bubbles is needed to couple with a transducer for acoustic en-
ergy transfer and fluid flow magnification,[40–42 ] thus increases
the system integration and stability and decreases the possibil-
ity to induce cell damage due to the sharp structures. The effec-
tive range can be adjusted by tuning the distance or resonator
size which enables both localized cell manipulation or large-
scaled cell batch processing. There are also no limitations on
specific cell growth environments, which perfectly fits the sub-
sequent in situ cell cultivation and conventional cell detection
techniques. In summary, this platform is a promising tool for
the development of new cell manipulation and cell therapeutic
methods.
2. Results and Discussion
2.1. Simulation and Force Characterization of the Acoustic
Streaming under GHz Device Excitation
The working mechanism of gigahertz resonator and acoustic
streaming generation is illustrated in Figure 1a–c. The resonator
is mainly composed of a piezoelectric layer (aluminum nitride,
AlN) sandwiched between two electrodes. When applying an ex-
ternal vertical electric field through the electrode, geometric de-
formation is triggered within the piezoelectric material due to
the inverse piezoelectric effect, thus generates mechanical vibra-
tion. The resonator was designed working at thickness exten-
sional (TE) mode, in which symmetric displacement occurs on
the upper and lower boundary of the piezoelectric layer to gen-
erate vertically propagating longitudinal wave. The resonant fre-
quency is mainly governed by the thickness of the piezoelectric
layer and can be estimated by the following equation:
f=1
2d√c
𝜌(1)
where dis the thickness of the piezoelectric material, cand 𝜌are
the material stiffness coefficient and density respectively.[43] The
thickness of the AlN thin film can be precisely controlled from
nanometers to micrometers during deposition process, thus en-
ables the device to achieve an ultrahigh vibration frequency over
gigahertz. The working area of the device was pattern into a pen-
tagon shape to minimize unwanted lateral-wave resonance and
achieve higher energy utilization.
When the resonator works in liquid, acoustic wave propagates
into liquid and attenuation occurs due to the fluid viscosity.[44]
The acoustic propagation can be described as follow:
v=v0e−𝛽z(2)
where v0is the initial velocity amplitude, and z is the direction of
propagation. 𝛽is the attenuation coefficient which is defined by:
𝛽=𝜔2
2𝜌cl3(4
3𝜇+𝜇B)(3)
In which 𝜔is the angular frequency, clis the sound speed in liq-
uid, 𝜌,𝜇,and𝜇Bdenote the density, dynamic viscosity and bulk
viscosity of the liquid. The equation indicates that the acoustic at-
tenuation coefficient is proportional to 𝜔2, and acoustic wave of
a higher frequency will attenuate much more rapidly. The decay
length of an acoustic wave is defined by the distance where the
amplitude decreased to 1/eof its original state, which is 1/ 𝛽.We
can calculate that the decay length of a 1.64GHz acoustic wave
is only 17.4 µm in water. In this paper, a minimum distance of
100 µm was used for cell manipulation where the acoustic wave
already attenuated to 0.3% of its initial amplitude. Therefore, the
influence of the acoustic pressure on cells can be ignored in our
study.
The acoustic energy dissipation into liquid will cause a body
force (F) for fluid driven, which is expressed as follow:
F=2𝛽𝜌v02e−2𝛽z(4)
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (2 of 11)
www.advancedsciencenews.com www.advancedscience.com
Figure 1. Acoustic streaming characterization and cell behavior simulation. a) Vertical structure of the gigahertz resonator. b) Resonator vibration. The
resonator works at thickness extensional (TE) mode in which longitudinal expansion and contraction occurs to generate vertically propagated acoustic
wave. c) The acoustic wave transmits into liquid, and acoustic streaming is generated due to the acoustic attenuation. d) Schematic of the AFM force
detection of the acoustic streaming. The yellow pentagon indicates the resonator, and the red line shows the detected position. e) Detected force
distribution of the resonator vibration (black) and the fluid motion (red) under 0.1 mW power supply. f) Relationship between the power applied to the
resonator and the fluid force measured by the ultrasensitive force sensor (n=3). g) Simulation of cell deformation under acoustic streaming stimulation.
The cell is represented by an elastic semicircle surrounded by water and is placed 100 µm away from the resonator surface. White arrows indicate flow
directions, and color bars indicate streaming velocity distribution and cell deformation extent.
where v0is the initial velocity amplitude. In conclusion, acoustic
waves with higher frequencies have a shorter travelling distance,
meanwhile generate a larger local force for liquid actuation. For
example, compared with acoustic wave of 1.5 MHz frequency, the
initial body force generated by a 1.5 GHz wave at the solid–liquid
interface can be 106larger, indicating this gigahertz resonator
to be a perfect tool for generating micro-scaled high-speed fluid
motion. Besides frequency, the resonator size also influences the
generated acoustic streaming. A simulation study is given in Fig-
ure S1 in the Supporting Information for detailed discussion.
To understand the cell deformation under such acoustic
streaming, we first characterized the force distribution in such
small-scale liquid flow with atomic force microscope (AFM) and
ultrasensitive force sensor, followed by a simulation of the cell
behavior under the hydrodynamic force field. Schematic and the
result for AFM force mapping of the acoustic steaming is given in
Figure 1d,e. AFM tip was scanning along a 90 µm line above the
resonator. Before fluid force measurement, resonator vibration
was detected to confirm the relative position of the AFM tip with
the active resonant region. The up-lifting fluid motion above the
resonant area under 0.1 mW power supply was tested by placing
the AFM tip 100 nm away from the solid surface. When the tip
was located outside the pentagon resonant region, the detected
force almost equals to zero. By gradually sweeping the tip toward
the resonator active area, fluid force increased sharply. In the pen-
tagon center, the force shows a relatively flat distribution, and a
vertical force of about 4 nN was recorded. The distribution of the
fluid force is well matched with the shape of the device, which
confirms the generation of the localized fluid motion based on
gigahertz nanoscale oscillation. The AFM results provide a clear
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (3 of 11)
www.advancedsciencenews.com www.advancedscience.com
Figure 2. Controllable cell deformation platform. a) Schematic of the controllable cell deformation and real-time 3D observation system based on the
acoustic streaming generated from gigahertz ultrasonic device. b) Optical image of the gigahertz resonator. c) Cell deformation under the stimulation
of the acoustic streaming.
observation of the fluid force distribution, however due to the lim-
its of the AFM system it cannot precisely measure larger forces.
To further characterize the streaming at larger acoustic energy,
we then detected the streaming force under different applied
power using a commercialized miniaturized force sensor (Au-
rora Scientific, 405A). The result in Figure 1f indicates that the
strength of the force is linearly proportional to the power ap-
plied from a few to a few tens of µN, demonstrating a good con-
trollability of this force loading platform over a wide adjustable
force range. This is the first detailed force characterization on
the acoustic streaming generated from gigahertz ultrasonic de-
vice. It provides a new perspective and a deeper insight for the
mechanism studies of the interactions between acoustic stream-
ing and biomaterials. We also provided a streaming velocity study
by recording the microparticle movement in liquid, and the video
and data analysis are given in Figure S2 and Video S1 in the Sup-
porting Information.
To further understand the effects of the streaming force on
the cell deformation, simulation of the interaction between the
acoustic streaming and a single cell is carried out. As shown in
Figure 1g, a cell model was placed 100 µm away from the GHz
resonator. When the device works in liquid environment, acous-
tic streaming is generated and hydrodynamic force is applied
along the cell surface. Under such flow field excitation, the cell
is squeezed from the semicircular (white area) to a compressed
condition. The most significant displacement occurs at the top of
the cell, while the bottom of the cell remains at its original status.
This displacement difference results in the oblique deformation
of the cell. The dynamic simulation result is given in Video S2
(Supporting Information), from which the transient response of
the cell under acoustic streaming can be obtained. To explore the
influence of the horizontal relative position between the cell and
the resonator on cell deformation, simulation results under dif-
ferent cell locations were compared as well (Figure S3, Support-
ing Information), and obvious cell deformation can be obtained
when cells are located within 100 µm from the center of the res-
onator. Since a 2D simulation cannot fully represent the real 3D
space, we also performed a 3D simulation under the same con-
dition as Figure 1g (Figure S4 in Supporting Information). Com-
paring the results of 2D and 3D simulations, similar cell defor-
mation feature and deformation extent can be seen. Thus, in the
following analysis, 2D simulation will be used for modeling sim-
plification.
2.2. Controllable and Restorable Cell Deformation
After simulation of the cell behaviors under the acoustic stream-
ing, we experimentally recorded the real-time cell deformation
process by integrating the GHz resonator with a confocal micro-
scope to provide a 3D dynamic characterization of the cell. The
schematic of the GHz hydrodynamic cell deformation and 3D cell
morphology recording system is shown in Figure 2. A bulk acous-
tic resonator with frequency of 1.64 GHz was utilized to excite
acoustic streaming. The resonator was inserted into the solution
with its resonant area facing down toward the cells seeded Patri
dish, and the vertical distance was controlled to 100 µm according
to our simulation in Figure 1g. Under this condition, the size of
the major streaming vortex area (Figure S3, Supporting Informa-
tion) is around 150 µm in diameter, and will only deform the cells
near and within the resonator working area. Thus, this distance
was used to maximize force utilization for localized cell deforma-
tion. Z-stacking function in confocal microscope was applied to
obtain the 3D cell morphology (Video S3 and Figure S5, Support-
ing Information). By continuously recording the complete vol-
umes of the cells, real-time analysis of the cells before, during
and after the streaming stimulation was achieved.
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (4 of 11)
www.advancedsciencenews.com www.advancedscience.com
Figure 3. Cell deformation process and cell shape characterization. a) Real-time cell outline deformation before, during and after the treatment of acoustic
streaming (400 mW applied). The cell is squeezed to the left by the acoustic streaming, and gradually restored to its original shape after stopping the
stimuli. White thin dashed line indicates the original cell central position. Scale bar is 5 µm. b) Calculation method of cell deformation coefficient. c) Cell
deformation coefficient change with the treating time (n=3).
A typical recorded cell deformation dynamic process is given
in Video S4 (Supporting Information). The video provides an in-
tuitive understanding of the cell deformation in the 3D space, and
enables the cell shape characterization from different directions.
To maximize the effect of the acoustic streaming, here we extract
the vertical sections of the treated cell along the maximum cell
deformation direction for the following analysis, as shown in Fig-
ure 3a. At the initial stage, the cell maintains original hemispher-
ical shape. When applying a power of 400 mW to the resonator,
the stimulated acoustic flow flushes the substrate and forms a
lateral streaming from right upon the cell. Since the bottom of
the cell adheres to the substrate firmly, the fluid force pushes the
cell deflecting to the left. Such deformation is happened in a very
short time (≈5 s). Due to the press, the right half of the cell is
compressed, meanwhile the left half is ballooned, which agrees
with the aforementioned simulation results. With the increase of
the treating time, the deflection angle gets larger and remains
almost stable after 15 s. In order to analyze the long-term influ-
ence of the streaming on the cell, we also recorded the cell shape
change after stopping the stimulation. A quick recovery from the
cell stretch status happens within 5 s after turning off the power,
and the cell shape continues to change and recover to its original
condition after 20 s relaxation.
To quantify the deformation feature of cells, we introduce an
indicator based on the angle of cell deflection, as shown in Fig-
ure 3b. Longitudinal section of a cell can be simplified as an arc
(arc-bc) with its two end points (point b and c) on the substrate.
We define the mid of the cell outline to be the point (point a)
which divide the arc into two parts with the same length (arc-ab
=arc-ac), and the mid of the substrate (point O) to be the cen-
ter of the area on the petri dish occupied by the cell (segment
bc), that angle 𝜃(∠aOc) represents the cell deflection. When cell
is pressed by the streaming flow, midpoint of cell outline moves
along the direction of acoustic streaming due to the cell deflec-
tion, thus making angle 𝜃increases (from 𝜃0to 𝜃1). The deflection
coefficient (DC) is defined by the ratio of these two angles:
Deflection coefficient (DC)=
𝜃1
𝜃0
(5)
which can be used to quantify the cell deformation. DC
should be equal to or larger than 1 ( =1 indicates cell
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (5 of 11)
www.advancedsciencenews.com www.advancedscience.com
Figure 4. Cell deformation under different powers. a) The cell was treated with a time period of 5 s. Scale bar is 5 µm. b) Relationship between deformation
coefficient and the power applied (n=3).
original shape), and a larger DC indicates a more severe
deformation.
The cell DC is extracted and plotted with different time in Fig-
ure 3c. The coefficient gradually increased to about 1.15 when
maximum cell deformation occurs at 35 s. The most rapid shape
change happens within 5 s after the power is applied, and then
the changing rate slowed down until the cell shape reaches the
equilibrium position. When the stimulation is stopped, a quick
decrease of the coefficient is also happened in the first 5 s, which
is consistent with the observation from Figure 3a. The DC keeps
decreasing in the following recordings and the final value reaches
to almost 1, indicating that the streaming force under this power
intensity has negligible long-term influence on the integrity of
the cell and the shape can finally recover to its original state. The
results also confirm that the cell deformation extent produced by
our platform is time controllable.
To further investigate the influence of applied power on the cell
deformation, different powers were applied to the same cell, and
the cell deformations after treated for the same periods are given
in Figure 4a (also see Video S5 in the Supporting Information).
It is observed that the deformation extent increased by increas-
ing the power intensity. When applying 300 mW power, the cell
shows slight deflection. When the power is increased to 500 mW,
the cell is severely compressed to an oblate shape, and the defor-
mation coefficient increased to more than 1.25 (Figure 4b). The
result reveals that the cell deformation extent can be delicately
controlled in our platform by tuning time duration or power of
the stimulation, proving it to be a powerful tool for cell shape
manipulation. Besides cell deformation, we also discussed the in-
fluence of fluid on cell detachment. We have found that cell peel-
ing may occur when treating a separated single cell of a small
attaching area on the substrate (Figure S6 and Video S6 in the
Supporting Information) or using larger acoustic powers (larger
than 500 mW), and a detailed study is given in S7 (Supporting
Information).
2.3. Complex Cell Shape Manipulation under Periodic Streaming
Excitation
The previous data has proved that the cell deformation or the re-
covery process closely follows the period of the acoustic stream-
ing. Thus, we assume that if applying periodically acoustic
streaming with tunable duration, the treated cells would be ex-
cited by cyclic hydrodynamic force, and exhibiting an oscillating
shape change. To verify our analysis, we performed cell defor-
mation experiments using different periodic excitation methods,
as shown in Figure 5. The period of the acoustic streaming is
controlled by applying and tuning a square-waved power to the
GHz resonator. Two kinds of pulse durations, 5 s/5 s (indicat-
ing a square wave with 5 s power on and 5 s power off in a pe-
riod of 10 s) and 15 s/15 s were used. The real-time measure-
ment of the streaming force under these two pulse excitations
is given in Figure S8 (Supporting Information). The force result
shows the same variation as the applied electric signal, indicat-
ing the acoustic streaming can be well controlled under periodic
gigahertz excitation. Video S8 (Supporting Information) shows
the cell status under these two stimulation methods, the maxi-
mum and minimum deformation within two cycles are plotted
in Figure 5a. Repeating cell extrusion and recovery process are
observed, which are closely following the variation of the applied
power. Extracted cell deformation coefficient (DC) is given in Fig-
ure 5b, from which we can see that the feature and the extent of
the cell deformation in each cycle is consistent, indicating a good
repeatability. To describe the feature of cell deformation under
cyclic acoustic streaming, another parameter, deformation veloc-
ity, is introduced here which is defined by:
Deformation velocity =Deformation coefficient −1
pulse duration (6)
The comparation of maximum DC and deformation velocity
under these two excitations is given in Figure 5c. Resulted by the
longer treating time in each cycle, cells under 15 s/15 s excitation
show higher deformation extent. While the deformation velocity
of 5 s/5 s excitation is significantly larger than that of 15 s/15 s
excitation, which is due to the higher frequency of the stimulation
used in 5 s/5 s excitation method.
These results demonstrated that this hydrodynamic force load-
ing platform could realize controllable cell deformation following
complex patterns. Benefited from the easy and flexible regulation
of the electric signal and the instant conversion from the electric
energy to the fluid motion, we can further assume that by apply-
ing a code-controlled multiparameter signal with tunable power,
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (6 of 11)
www.advancedsciencenews.com www.advancedscience.com
Figure 5. Cyclic cell deformation under periodic streaming excitations. a) Cell shape change in two cycles under 5 s/5 s excitation and 15 s/15 s excitation.
Scale bar is 5 µm. b) Deformation coefficient variation with time under two excitation methods. Blue area indicates the period when the power is applied.
c) Comparation of cell deformation extent and average deformation velocity under different excitation methods (n=3).
pulse duration and duty cycle, a desired final cell shape or cell
shape regulating trajectory can be automatically achieved, and a
programmable cell behavior regulation platform can be extended.
2.4. Application of the Controllable Cell Stretching for Membrane
Permeability Modulation
We have observed an interesting phenomenon on our platform
that during the cell deformation process, the membrane fluores-
cence intensity on one side of the cell which is facing the acous-
tic streaming tend to weaken, meanwhile the fluorescence on the
other side tend to become stronger (Figure 6a). Considering the
fluidity of the plasma membrane, we assume that this variation in
fluorescence is due to the change of membrane molecular den-
sity and construction. When the cell is pressed by the stream-
ing, the flow along the cell surface will provide shear force to the
membrane. This force may overcome the intermolecular forces
in the membrane, compel the phospholipids to flow to the back
side of the cell, and provide chances for crack formation on the
membrane.
This unique feature indicates that the platform could be ap-
plied for membrane permeability study and manipulation. In or-
der to confirm our assumption, penetration of DOX molecules
through the cell membrane under streaming stimulation was
evaluated to reveal the change in cell membrane permeability. To
achieve larger area of treatment, distance between the cell sub-
strate and the resonator was increased to 1 mm. As simulated
in Figure S9 (Supporting Information), the vortex area is signifi-
cantly increased under this condition. A relatively uniform force
distribution on cells within 1 mm diameter area can be seen, thus
enables simultaneous treatment on multiple cells. DOX fluores-
cence inside the Hela cells without and with streaming excita-
tion under different power was observed, respectively. As shown
in Figure 6b, low fluorescence intensity of the cells with no hy-
drodynamic force treatment indicates a low permeability of the
membrane. The increased DOX fluorescence is observed after
gigahertz stimulation, and it is power dependent. This result is
consistent with the deformation result in Figure 4, where a larger
power gives rise to a higher shape deformation efficiency. Thus,
we can confirm that cell deformation can lead to crack formation
on the cell membrane, and the permeability to molecules can be
controlled by tuning the cell deformation extent. This permeabil-
ity change was proved to be reversible, and the membrane per-
meability reversibility and cell viability under acoustic streaming
excitation are given in Figures S10 and S11 in the Supporting
Information. Besides, we found that cell aggregation status also
influences cell deformation and membrane permeability, which
is thoroughly discussed in Figures S12 and S13 in the Supporting
Information.
We further performed the cellular uptake experiment with
larger molecules (FITC-dextran, Mw 40K), as shown in Figure 7.
The control group and experiment group with 500 mW contin-
uous stimulation shows obvious difference in the dextran pene-
tration efficiency, which proves that the continuous cell deforma-
tion process can also alter the membrane permeability for larger
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (7 of 11)
www.advancedsciencenews.com www.advancedscience.com
Figure 6. Cell membrane reconstruction and permeability change under streaming excitation. a) Cell membrane fluorescence before and during acoustic
streaming treatment. White and yellow dashed circles indicate the pressed side and the back side of the cell, respectively. Scale bar is 5 µm. b) DOX
uptake with or without acoustic streaming stimulation. Cells were treated under different power for 10 min. Scale bar is 100 µm.
Figure 7. Membrane permeability for dextran under different excitations. a) Penetration of 40k FITC-dextran molecules into HeLa cells using different
excitation methods. Each group was treated in 2 mg mL−1dextran solution under 500 mW power for 10 min. Scale bar equals to 50 µm. Quantitative
analysis of the dextran penetration efficiency (indicating the percentage of the dextran-delivered cells in view) and the mean fluorescence intensity in
delivered cells are given in b,c) (n=3).
biomolecules entering, however the penetration efficiency only
reaches about 10%. In order to further optimize the membrane
permeability for macromolecules, two cyclic hydrodynamic exci-
tation methods which have been studied in Figure 5, 5 s/5 s and
15 s/15 s square-waved excitations, were applied in this exper-
iment. Compared with the continuous excitation, 15 s/15 s ex-
citation shows no obvious improvement in intracellular fluores-
cence. However, under 5 s/5 s treatment, the dextran penetration
efficiency is significantly increased, and the average fluorescent
intensity in the dextran-delivered cells is also much higher than
that of continuous excitation.
The result indicates that fast and repeating periodic cell
squeezing can induce remarkable effect on increasing the mem-
brane permeability to macromolecules. A possible explanation
for this phenomenon is that when the cell is treated with a
stable force, the cell deformation and membrane phospholipid
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (8 of 11)
www.advancedsciencenews.com www.advancedscience.com
distribution will reach an equilibrium status. In contrast, the cell
shape will be in a continuously changing state by alternatively
apply and withdraw forces; the cell membrane experiences an
extruding-relaxing process, and there will be a mechanical accu-
mulative effect, which facilitates the membrane molecules break
their force balance and in favor of the macromolecule membrane
transportation. Such cell membrane dynamic fatigue behaviors
have been reported before for red blood cells treatment under
cyclic stretching loads.[45,46 ] An increase in the membrane shear
modulus of red blood cells was found after several tens of cycles
of cell stretching and extensional recovery, indicating a decrease
in elasticity and an increase in membrane brittleness in response
to cyclic stresses. Besides, theoretical study which combines the
cell resonance and cell fatigue theories was reported as well.[47]
They explain the effect of cell resonance using the cumulative ef-
fect under ultrasonically driven strains with a smaller magnitude
but enormous number of loading cycles. Thus, we assume that
the enhanced dextran delivery under periodic cell deformation is
likely from a fatigue-like behavior on cell membrane, which fa-
cilitates the macromolecule transportation.
In summary, this GHz resonator based hydrodynamic force
loading and cell deformation platform can serve as a novel tool
for cell membrane permeability modulation. Our results have
revealed the change of membrane permeability under different
force environments. Most prominently, we have observed a novel
regulation approach based on this platform that cells tend to pro-
duce larger cracks for macromolecules transportation under the
fast and periodic oscillation. Considering the significance of the
membrane permeability, this discovery may help with the in-
depth mechanism study and provide a novel way for the optimiza-
tion of the cellular drug uptake efficiency. The future steps will
be focused on improving the current system for real-time simul-
taneous observation of cell deformation and drug penetration to
deepen our understanding on fluid–membrane and molecule–
membrane interactions.
3. Conclusion
In this article, we developed a novel hydrodynamic cell defor-
mation platform based on acoustic streaming generated from a
microfabricated gigahertz resonator. The direct conversion from
acoustic energy to localized fluid motion of the gigahertz device
enables the noncontact and pure fluid force manipulation on this
platform, which avoids the need of bubble or structure assistance
for energy transformation in traditional technologies, improves
the system flexibility and reduces the difficulty to manipulate at
high accuracy. By integrating the deformation platform with con-
focal microscope, real-time 3D observation of cell deformation is
realized, making it possible for the analysis of complex cell mor-
phological response under different stimulation patterns. Our
DOX and dextran uptake experiments further investigated the
effect of hydrodynamically controlled cell deformation on mem-
brane molecule construction. The results revealed a promising
application area of this platform for cell membrane permeabil-
ity modulation, strongly proved the potential value of this plat-
form in developing new cell manipulation and cell therapeutic
methods.
4. Experimental Section
Device Fabrication:Gigahertz acoustic resonator was fabricated as de-
scribed in previous publication.[35] Briefly, Bragg mirror structure consist-
ing of alternating layers of silicon dioxide (SiO2) and aluminum nitride
(AlN) was deposited on Si wafer for acoustic wave reflection. 600 nm
molybdenum (Mo), 1000 nm AlN, 60 nm chromium (Cr), and 300 nm
gold (Au) were then deposited and patterned to form a sandwich struc-
ture for acoustic vibration. The size of the resonator working area is 0.01
mm2.
Cells Preparation:HeLa cells were grown in Dulbecco’s modified Ea-
gle medium (DMEM) supplemented with 10% fetal bovine serum and 1%
penicillin–streptomycin. The cell line was maintained in T-25 cell culture
flasks in an incubator at 37 °C and 5% CO2level. For laser confocal obser-
vation, cells were cultured on glass bottom culture dishes. The bottom of
the culture dish was precoated with Fibronectin from human plasma (0.2
mg mL−1, Solarbio, China) for cell adhesion. Cell suspension (0.5 ×105
cells mL−1, 2 mL) was loaded and cultured for 2 d before experiment. For
DOX (doxorubicin hydrochloride) and dextran delivery, cells were cultured
on glass slides with same conditions.
System Setup:The acoustic resonator was controlled by a sinusoidal
signal (1.64 GHz), which was generated by a signal generator (Agilent,
N5171B) and preamplified by a power amplifier (Mini-Circuits, ZHL-5W-
422+). The resonator was wire-bonded to evaluation boards for signal
transmission. For confocal experiments, the resonator was held by a po-
sition control platform and was placed facing the bottom of the culture
dish. In drug delivery experiments, a PDMS chamber with a thickness of
1 mm was sealed on the resonator to form a drug container and fulfilled
with DOX or dextran solution (100 µL). Glass slides cultured with Hela
cells were then covered on the chamber and in contact with the drug so-
lution.
Acoustic Streaming Characterization:The resonator vibration and fluid
force distribution along the resonator surface were characterized by AFM
(Dimension Icon, Bruker Corp., Bremen, Germany), which were performed
following the procedures described in previous publication.[48] Briefly, the
gigahertz resonator was actuated by an amplitude-modulated signal dur-
ing AFM detection. The carrier frequency was equal to the resonance fre-
quency, and the modulation frequency was equal to the second resonance
frequency (≈110 kHz) of the AFM cantilever. Thus, the AFM tip follows
the envelope of the resonator or fluid vibration, and its response at the
modulation frequency is proportional to the detected vibration amplitude.
Considering the AFM detection range (90 µm in length), resonator with a
smaller working area (0.005 mm2) was used in this experiment. AFM tip
was placed at the surface of the resonator when mapping the device vi-
bration, and was lifted for 100 nm (tip offset =100 nm) when testing the
fluid force to make sure that the tip was not influence by the solid surface.
The final force value (F) was calculated by multiplying the spring constant
(k,0.35Nm
−1), dthe sensitivity (S,70nmV
−1) of the cantilever and the
response of electric voltage (V, which was giver by the AFM software):
F=k∗S∗V(7)
Force transducer (Aurora Scientific, 405A) with a tip of 60 µm in diameter
was used to characterize acoustic streaming under different power excita-
tion, and the value was extracted and recorded with a Source Measure Unit
(Keithley 2400, USA). Fluid velocity was recorded using a self-established
particle tracing system. The resonator was immersed in DI water filled with
fluorescent microparticles (15 µm diameter, FluoSpheres, Life technolo-
gies, OR, USA). Green laser (532 nm, 100 mW, LD-WL206, Changchun
New Industries Optoelectronics Tech. Co., China) and a cylindrical lens
were used to generate a plane fluorescent illumination which passes
through the center of the resonator. Particle movement was recorded us-
ing a high-speed camera (Phantom v7.2, Vision Research, NJ, USA).
Finite Element Simulation:Simulation of acoustic streaming field and
cell deformation process was given using the fluid-structure interaction
model in COMSOL Multiphysics 5.5. Gigahertz vibration induced acous-
tic streaming is governed by a decaying body force generated from acous-
tic attenuation, which is calculated following Equations (1–4). In 2D
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (9 of 11)
www.advancedsciencenews.com www.advancedscience.com
simulation, body force area was set to 100 µm (width) * 50 µm (height).
Liquid material is set to water, and the velocity field is described by the
incompressible Navier-Stokes equation which is given in COMSOL soft-
ware. Cell was modeled using a semicircular-shaped linear elastic mate-
rial with a diameter of 20 µm with its straight side fixed to the fluid field
boundary, and the cell elastic properties were used as published in previ-
ous publications.[49] In 3D simulation, body force was given in an area with
a pentagon-shaped bottom (side length 100 µm) and 50 µm height. Cell
was represented with a semisphere of 20 µm in diameter. Other settings
are the same as the 2D simulation.
Cell Deformation Recording and Analysis:Confocal microscope (Leica,
SP8) was used for cell deformation recording. Before experiments, cells
were washed with 1×PBS for three times and stained the membrane with
wheat germ agglutinin (Alexa Fluor 647 conjugate, Life Technologies, USA)
(10 µg mL−1) for 20 min. After that, cells were rinsed with 1×PBS for an-
other three times, and fresh PBS solution was added to the culture dish
for the experiment. Culture dish was then placed on the stage of the confo-
cal microscope, and repeated acquisition of complete volumes (xyzt scan
mode) was used at an excitation wavelength of 633 nm to trace the cell
shape before, during and after acoustic streaming treatment. The device
was held by a manually controlled three-axis linear translational stage with
a 10 µm position accuracy for position control and was inserted into the
PBS solution. The relative position of the cell and the resonator is tuned
with the following procedures. Position of the device was firstly confirmed
under a low power objective lens (20×), and the pentagon area was moved
to the center of the view by tuning xand yaxis of the stage. Then, the zpo-
sition of the device was lowered down carefully until the device surface
and the cells on the substrate are all clear in the same focal plane. This
position was defined to be z=0 position for the device, which means the
distance between the device and the cells is 0. After that, the device was
again carefully lifted for 100 µm, and the microscope was switched to a
high-power lens (60×) for cell scanning. The objective stage of the micro-
scope together with the petri dish was moved to select cells for scanning,
and the selected cells were placed at the center of the view to make sure
that the cell is in the resonator streaming area.
Intracellular Delivery Recording:Doxorubicin hydrochloride (DOX, Al-
addin, China) (2 µg mL−1) and fluorescein isothiocyanate labeled dextran
(FITC-Dextran, Mw 40 000, Sigma-Aldrich, USA) (2 mg mL−1)wereusedto
characterize the cell membrane permeability and drug delivery efficiency.
For each experiment, cells on the glass slides were treated with acous-
tic streaming for 10 min, then were washed with 1×PBS for three times.
Fluorescence microscope (Olympus BX53) with a CCD camera (Olympus
DP73) was utilized to observe the intracellular fluorescence. The distance
between the resonator and the cells was controlled to 1 mm by sealing
a 1 mm thickness round PDMS chamber in between. The resonator was
fixed at the center of the round PDMS area, and only the cells at the cen-
ter of the substrate (indicating cells opposite to the resonator area) were
selected for fluorescence observation.
Cell Viability Evaluation:Cell viability change under acoustic stream-
ing stimulation was evaluated using a MTT assay. Cells were seeded into
a 96-well plate at a density of 1.0 ×104cells per well in 200 µL culture
medium and were grown for 1 d. Thereafter, the cells were stimulated with
or without gigahertz resonator for 10 min, followed by another 48 h incu-
bation. After that, MTT solution (5 mg mL−1, 10 µL) was added into each
well and the cells were incubated for 4 h. Then the media was completely
removed and the cells were washed with 1×PBS for three times. 150 µL of
dimethylsulfoxide (DMSO) was added to each well, and the formazan was
dissolved for 10 min on an orbital shaker at a shaking speed of 300 rpm.
The absorbance was measured with a microplate reader at the wavelength
of 490 nm.
Statistical Analysis:The data are presented as mean ±SD in fluid force
measurements (n=3), cell shape analysis (n=3), intracellular fluores-
cence analysis (n=3), and cell viability analysis (n=5). Particle tracking
for fluid velocity analysis was processed with TrackMate plugin in ImageJ
software and plotted with 3D scatter function in MATLAB (MathWorks,
USA). Cell cross-sections were obtained with Leica Application Suite X.
Extraction of cell outlines and calculation of intracellular fluorescence in-
tensity were made in ImageJ software.
Supporting Information
Supporting Information is available from the Wiley Online Library or from
the author.
Acknowledgements
X.G. and M.S. contributed equally to this work. The authors gratefully ac-
knowledge financial support from the National Natural Science Founda-
tion of China (NSFC No. 61674114, 91743110, 21861132001), National
Key R&D Program of China (2017YFF0204604, 2018YFE0118700), Tianjin
Applied Basic Research and Advanced Technology (17JCJQJC43600), the
111 Project (B07014), and the Foundation for Talent Scientists of Nan-
chang Institute for Micro-technology of Tianjin University.
Conflict of Interest
The authors declare no conflict of interest.
Keywords
acoustic streaming, cell deformation, drug delivery, hydrodynamic force,
membrane permeability
Received: July 1, 2020
Revised: October 3, 2020
Published online: December 21, 2020
[1] W. Li, Z. Yan, J. Ren, X. Qu, Chem.Soc.Rev.2018,47, 8639.
[2] M. Merkel, M. L. Manning, Semin. Cell Dev. Biol. 2017,67, 161.
[3] H. Chen, J. Cornwell, H. Zhang, T. Lim, R. Resurreccion, T. Port, G.
Rosengarten, R. E. Nordon, Lab Chip 2013,13, 2999.
[4] J.D.Humphrey,E.R.Dufresne,M.A.Schwartz,Nat. Rev. Mol. Cell
Biol. 2014,15, 802.
[5] W. Jiang, D. Niu, L. Wei, G. Ye, L. Wang, H. Liu, P. Chen, F. Luo, B. Lu,
Carbon 2018,139, 1048.
[6] K. H. Vining, D. J. Mooney, Nat. Rev. Mol. Cell Biol. 2017,18, 728.
[7] C. Guilluy, L. D. Osborne, L. Van Landeghem, L. Sharek, R. Superfine,
R. Garcia-Mata, K. Burridge, Nat. Cell Biol. 2014,16, 376.
[8] M. Werner, S. B. G. Blanquer, S. P. Haimi, G. Korus, J. W. C. Dunlop,
G. N. Duda, D. W. Grijpma, A. Petersen, Adv. Sci. 2017,4, 1600347.
[9] N. Huebsch, E. Lippens, K. Lee, M. Mehta, S. T. Koshy, M. C. Darnell,
R. M. Desai, C. M. Madl, M. Xu, X. Zhao, O. Chaudhuri, C. Verbeke, W.
S. Kim, K. Alim, A. Mammoto, D. E. Ingber, G. N. Duda, D. J. Mooney,
Nat. Mater. 2015,14, 1269.
[10] A. Sharei, J. Zoldan, A. Adamo, W. Y. Sim, N. Cho, E. Jackson, S. Mao,
S. Schneider, M. - J. Han, A. Lytton-Jean, P. A. Basto, S. Jhunjhunwala,
J. Lee, D. A. Heller, J. W. Kang, G. C. Hartoularos, K. - S. Kim, D. G.
Anderson, R. Langer, K. F. Jensen, Proc. Natl. Acad. Sci. USA 2013,
110, 2082.
[11] M. Sun, X. Duan, Nanotechnol. Precis. Eng. 2020,3, 18.
[12] J. Li, B.u Wang, B. M. Juba, M. Vazquez, S. W. Kortum, B. S. Pierce,
M. Pacheco, L. Roberts, J. W. Strohbach, L. H. Jones, E. Hett, A. Tho-
rarensen, J. -. B. Telliez, A. Sharei, M. Bunnage, J. B. Gilbert, ACS
Chem. Biol. 2017,12, 2970.
[13] X. Han, Z. Liu, M. C. Jo, K. Zhang, Y. Li, Z. Zeng, N. Li, Y. Zu, L. Qin,
Sci. Adv. 2015,1, e1500454.
[14] A. Sawicka, A. Babataheri, S. Dogniaux, A. I. Barakat, D. Gonzalez-
Rodriguez, C. Hivroz, J. Husson, Mol. Biol. Cell 2017,28, 3229.
[15] E. A. Evans, R. M. Hochmuth, Biophys. J. 1976,16, 13.
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (10 of 11)
www.advancedsciencenews.com www.advancedscience.com
[16] M. Lekka, D. Gil, K. Pogoda, J. Duli´
nska-Litewka, R. Jach, J. Gostek, O.
Klymenko, S. Prauzner-Bechcicki, Z. Stachura, J. Wiltowska-Zuber, K.
Oko´
n, P. Laidler, Arch. Biochem. Biophys. 2012,518, 151.
[17] Q. Guo, S. P. Duffy, K. Matthews, A. T. Santoso, M. D. Scott, H. Ma,
J. Biomech. 2014,47, 1767.
[18] J. R. Lange, C. Metzner, S. Richter, W. Schneider, M. Spermann, T.
Kolb, G. Whyte, B. Fabry, Biophys. J. 2017,112, 1472.
[19] Y.Zhang,F.Wei,Y.-C.Poh,Q.Jia,J.Chen,J.Chen,J.Luo,W.Yao,W.
Zhou, W. Huang, F. Yang, Y. Zhang, N. Wang, Nat. Protoc. 2017,12,
1437.
[20] S. Hu, L. Eberhard, J. Chen, J. C. Love, J. P. Butler, J. J. Fredberg,
G. M. Whitesides, N. Wang, Am. J. Physiol.: Cell Physiol. 2004,287,
C1184.
[21] X. Liu, R. Liu, Y. Gu, J. Ding, ACS Appl. Mater. Interfaces 2017,9,
18521.
[22] A. Sutton, T. Shirman, J. V. I. Timonen, G. T. England,P. Kim, M. Kolle,
T. Ferrante, L. D. Zarzar, E. Strong, J. Aizenberg, Nat. Commun. 2017,
8, 14700.
[23] C.-W. Lai, S.-K. Hsiung, C.-L. Yeh, A. Chiou, G.-B. Lee, Sens. Actuators,
B2008,135, 388.
[24] J. Czerwinska, S. M. Wolf, H. Mohammadi, S. Jeney, Cell. Mol. Bioeng.
2015,8, 258.
[25] J. Y. Hwang, J. Kim, J. M. Park, C. Lee, H. Jung, J. Lee, K. K. Shung,
Sci. Rep. 2016,6, 27238.
[26] L. A. Macqueen, M. D. Buschmann, M. R. Wertheimer, J. Micromech.
Microeng. 2010,20, 065007.
[27] I. Doh, W. C. Lee, Y. - H. Cho, A. P. Pisano, F. A. Kuypers, Appl. Phys.
Lett. 2012,100, 173702.
[28] Y. Xie, N. Nama, P. Li, Z. Mao, P.o-H. Huang, C. Zhao, F. Costanzo,
T. J. Hua n g , Small 2016,12, 902.
[29] F. Yuan, C. Yang, P. Zhong, Proc. Natl. Acad. Sci. USA 2015,112,
E7039.
[30] A. M. Forsyth, J. Wan, W. D. Ristenpart, H. A. Stone, Microvasc. Res.
2010,80, 37.
[31] D. R. Gossett, H. T. K. Tse, S. A. Lee, Y. Ying, A. G. Lindgren, O. O.
Yang,J.Rao,A.T.Clark,D.DiCarlo,Proc. Natl. Acad. Sci. USA 2012,
109, 7630.
[32] S. S. Lee, Y. Yim, K. H. Ahn, S. J. Lee, Biomed. Microdevices 2009,11,
1021.
[33] A. Sohrabi Kashani, M. Packirisamy, AIMS Biophys. 2017,4,400.
[34] H. Qu, Y. Yang, Y.e Chang, Z. Tang, W. Pang, Y. Wang, H. Zhang, X.
Duan, Sens. Actuators, B 2017,248, 280.
[35] X. Guo, H. Zhang, Y. Wang,W. Pang, X. Duan, J. Nanobiotechnol. 2019,
17, 86.
[36] S. Pan, H. Zhang, W. Liu, Y. Wang, W. Pang, X. Duan, ACS Sens. 2017,
2, 1175.
[37] W. Cui, M. He, Y. Yang, H. Zhang, W. Pang, X. Duan, Part. Part. Syst.
Charact. 2018,35, 1800068.
[38] Z. Zhang, Y. Wang, H. Zhang, Z. Tang, W. Liu, Y. Lu, Z. Wang, H. Yang,
W. Pang, H. Zhang, D. Zhang, X. Duan, Small 2017, 13, 1602962.
[39] Y. Lu, W. C. De Vries, N. J. Overeem, X. Duan, H. Zhang, H. Zhang,
W. Pang, B. J. Ravoo, J. Huskens, Angew. Chem., Int. Ed. Engl. 2019,
58, 159.
[40] M. Ovchinnikov, J. Zhou, S. Yalamanchili, J. Acoust. Soc. Am. 2014,
136, 22.
[41] N. Li, J. Hu, H. Li, S. Bhuyan, Y. Zhou, Appl. Phys. Lett. 2012,101,
093113.
[42] P.o-H. Huang, Y. Xie, D. Ahmed, J. Rufo, N. Nama, Y. Chen, C. Y.u
Chan, T. J. Huang, Lab Chip 2013,13, 3847.
[43] W. Pang, H. Zhao, E. S. Kim, H. Zhang, H. Yu, X. Hu, Lab Chip 2012,
12, 29.
[44] W. Cui, W. Pang, Y. Yang,T. Li, X. Duan, Nanotechnol. Precis. Eng. 2019,
2, 15.
[45] Y. Qiang, J. Liu, E. Du, Acta Biomater. 2017,57, 352.
[46] S. Sakuma, K. Kuroda, C.-.H. D. Tsai, W. Fukui, F. Arai, M. Kaneko,
Lab Chip 2014,14, 1135.
[47] M. Or, E. Kimmel, Ultrasound Med. Biol. 2009,35, 1015.
[48] F. Xu, X. Guo, L. Xu, X. Duan, H. Zhang, W.Pang, X. Fu, Micromachines
2017,8, 244.
[49] W. Kim, A. Han, in The 14th Int. Conf. on Miniaturized Systems for
Chemistry and Life Sciences (MicroTAS 2010), (Eds: S. Verpoorte, H.
Andersson-Svahn, J. Emnéus, N. Pamme) Vol. 1, Chemical and Bi-
ological Microsystems Society (CBMS), San Diego, California 2010,
253.
Adv. Sci. 2021,8, 2002489 © 2020 The Authors. Advanced Science published by Wiley-VCH GmbH
2002489 (11 of 11)
Available via license: CC BY
Content may be subject to copyright.
