Content uploaded by Govind Kaigala
Author content
All content in this area was uploaded by Govind Kaigala on Aug 12, 2015
Content may be subject to copyright.
The floating microfluidic probe: Distance control between probe and sample using
hydrodynamic levitation
Martina Hitzbleck, Govind V. Kaigala, Emmanuel Delamarche, and Robert D. Lovchik
Citation: Applied Physics Letters 104, 263501 (2014); doi: 10.1063/1.4886117
View online: http://dx.doi.org/10.1063/1.4886117
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/104/26?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Highly selective biomechanical separation of cancer cells from leukocytes using microfluidic ratchets and
hydrodynamic concentrator
Biomicrofluidics 7, 034114 (2013); 10.1063/1.4812688
Two- and three-dimensional modeling and optimization applied to the design of a fast hydrodynamic focusing
microfluidic mixer for protein folding
Phys. Fluids 25, 032001 (2013); 10.1063/1.4793612
An integrated, multiparametric flow cytometry chip using “microfluidic drifting” based three-dimensional
hydrodynamic focusing
Biomicrofluidics 6, 024113 (2012); 10.1063/1.3701566
Three-dimensional hydrodynamic focusing in a microfluidic Coulter counter
Rev. Sci. Instrum. 79, 046104 (2008); 10.1063/1.2900010
The generation of highly monodisperse droplets through the breakup of hydrodynamically focused microthread in
a microfluidic device
Appl. Phys. Lett. 85, 3726 (2004); 10.1063/1.1812380
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 178.83.19.16
On: Wed, 02 Jul 2014 04:58:02
The floating microfluidic probe: Distance control between probe and sample
using hydrodynamic levitation
Martina Hitzbleck, Govind V. Kaigala, Emmanuel Delamarche, and Robert D. Lovchik
a)
IBM Research-Zurich, S€
aumerstrasse 4, CH-8803 R€
uschlikon, Switzerland
(Received 28 May 2014; accepted 17 June 2014; published online 30 June 2014)
Microfluidic probes (MFPs) are an emerging class of non-contact scanning devices used to perform
local chemical reactions on surfaces covered with liquid. Typically, the probe is scanned at a dis-
tance between 10 lm and 50 lm over the surface. For proper functioning, the distance between the
probe and the surface needs to be kept stable. Here, we present a self-regulating distance control
for a microfluidic probe based on hydrodynamic levitation, and we call the device the “floating
MFP.” By injecting a liquid between the probe head and the surface (flow rates: 5–500 ll min
1
),
we were able to achieve levitation heights up to 15 lm without perturbation of the probe function.
We provide an analytical solution describing the levitation, which fits well with the experimental
data. This work helps in the design and implementation of distance control in MFPs for a broad
range of applications. V
C2014 AIP Publishing LLC.[http://dx.doi.org/10.1063/1.4886117]
There is a class of microfluidic devices that are based on
the concept of having a probe dispensing liquids locally to
biological samples, tissues, or on surfaces.
1
These devices
are well-suited for localized, low-volume, and precision
processing of biological samples typically placed on stand-
ard glass slides or Petri dishes. We are interested in the
microfluidic probe (MFP), which is highly versatile in its
operation and its range of applications. The MFP has been
used for several applications, such as microscale staining of
tissue sections,
2
biochemical patterning of surfaces,
3
and
single-cell electroporation.
4
Central to the MFP is a micro-
fluidic chip (MFP head) comprising at least two channels for
liquids that terminate at an apex. The apex is aligned parallel
to the sample surface, and a processing liquid is simultane-
ously injected and aspirated resulting in a hydrodynamically
confined volume of liquid (flow confinement) on the surface,
(Fig. 1). MFPs operate at distances between 10 lm and
50 lm from the surface within a liquid environment, making
the distance regulation challenging for the following reasons:
(1) the liquid can vary in its composition, e.g., pH, ion con-
tent, and turbidity, (2) compatibility with standard biological
substrates is required, and (3) direct contact with the sample
should be avoided. MFP heads used to be leveled manually
with a reference slide and the distance was set using a motor-
ized stage. This worked well for applications involving flat
and transparent microscope slides, for example, creating an
array of proteins.
3
However, for surfaces with topographical
variations, a manual and continuous intervention was
required during operation that limited the technology to be
used with transparent substrates.
Controlling the probe-to-sample distance is critical not
only in the MFP
5,6
but also in other scanning technologies
such as nanopipettes
7
and atomic force microscopy-based
approaches.
8
In order to sense distance at the sub-millimeter
scale, there are numerous feedback mechanisms based on
force,
9
current,
7
voltage, or frequency. As precise and fast
these sensing approaches may be, they are not suitable for
measuring the probe-to-sample distance during operation of
MFPs.
10
A MFP that allows processing of surfaces with to-
pographical variations and also on opaque substrates, with-
out the need for manual intervention, will significantly
broaden the range of applications.
Inspired by granite sphere fountains,
11
we hypothesized
that a “floating” mechanism could allow for probe-to-sample
distance control in a stable and self-sustained manner at the
micrometer length scale. We therefore implemented the
hydrodynamic levitation on the MFP and established key pa-
rameters for design considerations, which support the imple-
mentation of such a distance control. We call this particular
implementation of hydrodynamic levitation the “floating
MFP” (fMFP).
In contrast to the vertical MFP, wherein the head is
mounted to a motorized Z-stage, the fMFP is not constrained
along the Z-direction. Our approach was to clamp the head
onto a tone arm of a standard record player mounted to an
inverted light microscope ensuring the fMFP head is aligned
to the optics (Fig. 1(a)). The vertical movement during oper-
ation generally is between 10 lm and 50 lm, and the tilt vari-
ation caused by the 200 mm long tone arm was therefore
neglected (0.01for 50 lm height difference). The weight
of the head was by design adjusted using the counter weight
of the tone arm. A precise balance, embedded in the micro-
scope stage, allowed the weight of the head to be set to 0.1 g.
As shown in Figs. 1(b) and 1(c), the fMFP head comprises 4
microchannels. The two inner processing channels were used
to sustain a hydrodynamic flow confinement (Fig. 1(d)), and
the two outer floating channels were used to inject a floating
liquid that generated a lifting force (F
lift
). This force depends
on factors such as the gap height h, the geometry of the head,
and the weight of the head. In addition, the capillary force
F
capillary
, the viscosity lof the floating liquid, as well as the
momentum of the injected liquid are to be taken into account
to characterize the system. The key principle of the fMFP is
to use a stream of incompressible fluid flowing in a slit-like
confinement inducing a pressure due to the hydrodynamic re-
sistance exerted on the sidewalls of the system. The resulting
a)
yrl@zurich.ibm.com
0003-6951/2014/104(26)/263501/4/$30.00 V
C2014 AIP Publishing LLC104, 263501-1
APPLIED PHYSICS LETTERS 104, 263501 (2014)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 178.83.19.16
On: Wed, 02 Jul 2014 04:58:02
F
lift
counteracts with the above mentioned forces and reaches
a steady state at height h, which in practice is important for
functioning of the fMFP.
First, we looked at the hydrodynamic levitation of a
fMFP head having a circular apex with a radius Rand a sin-
gle circular floating channel of radius r0in the center. This
head has a cylindrical symmetry in which the partial differ-
ential equations of the Navier stokes equation decouple and
an analytical solution under approximation of creeping flow
can be established. At steady state and on account of the
cylindrical symmetry, the radial component of the velocity
vector is independent from the angle hand the only compo-
nent that is not equal to zero ðu¼uðr;zÞ~
erÞ. We assume a
parabolic flow profile in the Z-direction (no-slip condition)
and describe the velocity vector as ~
uðr;zÞ¼AðrÞzðhzÞ~
er.
The continuity equation for incompressible fluids
ðr ~
u¼0Þrepresents the mass balance of the system and
requires that the flow rate of floating liquid entering the head
(Q
in
) equals the mass flow through the outlet, or any other cyl-
inderofradiusr0rRaround the inlet
Qin ¼ðh
0
dz ð2p
0
rdhAðrÞzðhzÞ;8rR;(1)
~
ur;z
ðÞ
¼3Qin
ph3rzhz
ðÞ
~
er:(2)
For a velocity field of Eq. (2), the left hand side of the
Navier Stokes equation (3) is identically zero, and in the ab-
sence of external forces acting on the head, the equation can
be simplified to rP¼lD~
u(creeping flow approximation)
q@~
u
@tþ~
ur
ðÞ
~
u
¼rPþlD~
uþ~
f:(3)
Furthermore, D~
u¼
r~
u¼0r ðr ~
uÞ)@P
@r¼l@2ur
@2z2
applies for the vector field in Eq. (2). As a boundary condition,
we set the pressure at the edge of the apex equal to 0, with
respect to the ambient pressure, or P(R)¼0. Integration
results in
Pr
ðÞ
¼6lQin
ph3ln r
R
:(4)
The pressure below the apex integrated over its
surface area generates a force acting on the head, which
we name lift force (F
lift
). Flif t ¼ÐR
rinlet dr Ð2p
0rdhPðrÞ
¼3lQin
2ph3ðR2r2
inletÞþ2r2
inletln rinlet
R
hi
:
Generally, the radius of the inlet is small with respect to
the radius of the apex ðrinlet RÞ
Flif t ¼3lQin
2ph3R2:(5)
There are mainly four forces acting on a levitating fMFP
head. These are: F
lift
,F
weight
,F
capillary
(from the pinning
forces of the liquid meniscus at the three-phase contact line
of the immersed head), and F
momentum
of the floating liquid
injected through the channels and redirected by the sample
surface (see Fig. 1).
We were interested to study the interplay of the above
forces at flow rates between 0 and 1000 llmin
1
with head
weights of 0.1–2 g to achieve levitation heights up to 20 lm.
Practical considerations suggested the above conditions to be
suitable for operating the fMFP. The range of F
weight
was
limited by the mechanical setup in Fig. 1(a).
Fig. 2shows the magnitudes of forces in relation to the
flow rate of the floating liquid for a circular fMFP. The capil-
lary action can be assumed to be constant since the length of
the three-phase contact line does not change during opera-
tion. For the highlighted flow rates and head weights shown
in Fig. 2, the momentum and the capillary force are small
compared to the weight and the lift force and therefore were
neglected in the analysis. Rearrangement of Eq. (5) results in
h¼3lQin
mnetgR2
1=3
:(6)
During levitation, the lift force counteracts the weight
and the fMFP head reaches equilibrium at height h.
We verified the analytical solution for h(6) by meas-
uring the lift heights of a circular fMFP at different flow
rates and with various head weights. In addition, we tested
the influence of the viscosity of the floating liquid on h. The
FIG. 1. Concept of a fMFP. (a) A modified tone arm was used to position a
fMFP head atop a glass slide placed on the stage of an inverted microscope.
The counter weight allowed adjusting F
weight
below 1 mN. (b) Injecting liquid
through floating channels generates a lifting force F
lift
resulting in hydrody-
namic levitation of the fMFP head at height habove the substrate. Processing
channels ending at the apex of the head are used to hydrodynamically confine
a liquid of interest on the substrate. (c) Photograph of a fMFP head compris-
ing microfabricated channels and vias for liquid interfacing. (d) Micrograph
showing the apex of a fMFP head levitating above a glass substrate. A flow
confinement was generated using a solution of fluorescein, and fluorescent
beads were added to the injected floating liquid as flow tracers.
263501-2 Hitzbleck et al. Appl. Phys. Lett. 104, 263501 (2014)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 178.83.19.16
On: Wed, 02 Jul 2014 04:58:02
vertical position of the head was determined with a resolu-
tion of 130 nm using an optical measuring device (STIL,
CHR 150-N, France). Figure 3shows the lifting measure-
ments with head weights of 125, 250, and 1000 mg plotted
together with the corresponding analytical solution. We find
that for flow rates higher than 100 llmin
1
the experimental
data are in good agreement with the analytical solution. For
lower flow rates, the analytical solution slightly deviates
from the experimental measurements, which is also observed
for hlower than 10 lm. Under these conditions, where the
resulting lift forces are low, the setup is more easily per-
turbed by the experimental conditions (e.g., mechanical re-
sistance, bending of capillary tubing, precision of pumps).
This is not the case if a solution of higher viscosity is used,
such as 30% glycerol in water, which results in higher levita-
tion heights already at flow rates below 10 llmin
1
.
Vertical MFP heads typically have a rectangular apex;
we therefore also studied the lifting behavior of such
geometries. The fabrication of vertical MFP heads is per-
formed by etching channels into a silicon wafer, sealing it
with a glass layer, and singulation of the head by dicing.
10
Typical apex dimensions are 1 2mm
2
. Two channels at
the locations (x
0
, 0), (x
0
, 0) are for injection of the floating
liquid. The other two channels for generating the flow con-
finement are disregarded for this analysis. We modified the
equation for a circular fMFP (4) to approximate the pressure
distribution of a vertical fMFP with rectangular apex.
In Eq. (7),f
x,y
is the minimal distance to the border of
the head in x and y direction, respectively, P
max
is the maxi-
mum pressure, i.e., the pressure at the borders of the inlet,
and K(a) a factor relating to the aspect ratio of the head.
Px;y
ðÞ
¼
0;8x6x0
ðÞ
2þy2>fx;y;
Pmax;8x6x0
ðÞ
2þy2winlet
2
2
;
6lQin
2ph3Ka
ðÞ
ln x6x0
fx
2
þy
fy
2
!
:
8
>
>
>
>
>
>
<
>
>
>
>
>
>
:
(7)
In addition to studying the interplay between flow rate,
head weight, and apex geometry (footprint and placement of
channels), we were also interested in establishing levitating
conditions that do not perturb the probe function.
Figures 4(a) and 4(b) show the results of a finite element
simulation and the analytical solution (7) of the pressure
FIG. 3. Lifting of a circular fMFP at different flow rates, various F
weight
and
viscosity l. Data points with error bars indicate experimental measurements,
and the curves represent the analytical solution based on Eq. (6). The inset
shows a stream of floating liquid spiked with fluorescent beads injected
through the central apertures of a circular fMFP.
FIG. 4. Hydrodynamic levitation of a vertical fMFP. (a) Simulated pressure
distribution within the floating liquid at a gap height of 10 lm and a flow
rate of 150 ll min
1
(left). Comparison between the simulated pressure dis-
tribution (dots) and the modified analytical solution (Eq. 7) along the two
indicated cross sectional views (right). (b) Experimental results (connected
black dots) showing the lifting behavior of an operational fMFP head having
a net weight of approximately 0.125 g. The color band visualizes simulated
lifting heights for different head weights (purple, m ¼0.3 g; dark red
m¼0.1 g). Stable flow confinements using red fluorescent processing liquid
were possible with floating liquid flow rates of more than 500 ll min
1
(flow
confinements shown in insets).
FIG. 2. Magnitudes of forces acting on a circular fMFP head. The area of
the circular apex is p(3 mm)
2
, the three-phase contact line p1.6 mm and the
radius of the central inlet channel 60 lm.
263501-3 Hitzbleck et al. Appl. Phys. Lett. 104, 263501 (2014)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 178.83.19.16
On: Wed, 02 Jul 2014 04:58:02
distribution in the gap between the fMFP head and the sur-
face. The simulation and the analytical solution of pressure
distribution corroborate. These solutions provide guidelines
for the placement of the processing channels relative to the
floating channels and the apex geometry.
Using the optical measuring approach, as employed
with the circular MFP, the lifting of a vertical fMFP at
125 mg 620 mg was monitored for different flow rates and
compared with the simulation results, Fig. 4(b). For each
flow rate, we performed a simulation to estimate the lifting
heights for head weights between 0.1 g (purple) and 0.3 g
(dark red). The influence of the head weight is more pro-
nounced at high flow rates, hence enabling better control of
h. This characteristic may further be leveraged by using a
higher viscous floating liquid. For the head weight of about
125 mg, the measured lifting behaviour is in agreement with
the simulation results.
While increased flow rates of the floating liquid allow
for precise distance control, a high flow rate, such as 1000 ll
min
1
, perturbs the flow confinement of a vertical fMFP (see
inset of Fig. 4(b)). From a practical standpoint, we operate at
hof 10–15 lm, which can readily be achieved at flow rates
lower than 500 llmin
1
where no distortion of the flow con-
finement was observed.
In conclusion, distance control using hydrodynamic lev-
itation allows for easy, rapid, and self-sustaining operation
of MFPs on surfaces covered with liquids. This approach is
applicable to both static and scanning mode of operation of
the MFP and does not need any additional active peripheral
devices other than the ones inherent to the MFP, while ensur-
ing compatibility to a standard inverted microscope. By
altering the flow rates and the head weights, the desired levi-
tation was readily achieved. Further control over the
levitation is possible using liquids with varying viscosities,
changing the apex geometry and surface properties of the
head. We believe that hydrodynamic levitation can more
generally be applicable to other probe-based devices, ena-
bling their use in liquid environments without complex
means of distance control.
We acknowledge financial support by the European
Research Council (ERC) Starting Grant, under the 7th
Framework Program (Project No. 311122, BioProbe). We
thank Folkert Horst for his help with the optical distance
measurements and Marcel Buerge and Urs Kloter for the
support in building the experimental setup. Viola Vogel
(ETH Zurich), Urs Duerig, Bruno Michel, and Walter Riess
are acknowledged for their continuous support.
1
G. V. Kaigala, R. D. Lovchik, and E. Delamarche, Angew. Chem. Int. Ed.
51, 11224 (2012).
2
R. D. Lovchik, G. V. Kaigala, M. Georgiadis, and E. Delamarche, Lab
Chip 12, 1040 (2012).
3
D. Juncker, H. Schmid, and E. Delamarche, Nature Mater. 4, 622 (2005).
4
A. Ainla, S. Xu, N. Sanchez, G. D. M. Jeffries, and A. Jesorka, Lab Chip
12, 4605 (2012).
5
M. A. Qasaimeh, T. Gervais, and D. Juncker, Nat. Commun. 2, 464
(2011).
6
K. V. Christ and K. T. Turner, Lab Chip 11, 1491 (2011).
7
B. Babakinejad, P. Jnsson, A. Lpez Crdoba, P. Actis, P. Novak, Y.
Takahashi, A. Shevchuk, U. Anand, P. Anand, A. Drews, A. Ferrer-
Montiel, D. Klenerman, and Y. E. Korchev, Anal. Chem. 85, 9333 (2013).
8
A. Meister, M. Gabi, P. Behr, P. Studer, J. Vrs, P. Niedermann, J. Bitterli,
J. Polesel-Maris, M. Liley, H. Heinzelmann, and T. Zambelli, Nano Lett.
9, 2501 (2009).
9
K.-H. Kim, N. Moldovan, and H. Espinosa, Small 1, 632 (2005).
10
G. V. Kaigala, R. D. Lovchik, U. Drechsler, and E. Delamarche, Langmuir
27, 5686 (2011).
11
J. H. Snoeijer and K. van der Weele, “Physics of the granite sphere
fountain,” Am. J. Phys. (submitted).
263501-4 Hitzbleck et al. Appl. Phys. Lett. 104, 263501 (2014)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 178.83.19.16
On: Wed, 02 Jul 2014 04:58:02
