Actuation and In-situ Assembly of Colloidal Devices via Magnetic Fields
D.W.M. Marr
Chemical Engineering Department, Colorado School of Mines, Golden, CO 80401
For the practical realization of microscale and nanoscale pumps that mimic the operation
of their macroscale counterparts, two significant hurdles must be overcome. The first,
fabrication, is currently and in general being addressed through either improvements in
the resolution of macroscale techniques (“top down”) or through synthetic chemical
approaches to create increasingly complex molecular structures (“bottom up”). Though
both approaches have achieved some success, the creation of working devices within
their final functioning environments remains elusive. The second significant hurdle
involves the delivery of energy across length scales to effectively power devices in a
desired and controlled fashion once they have been created. In this proposal we present a
bulk-field based in-situ technique that addresses both issues by not only driving device
assembly from simple colloidal building blocks but also by subsequently powering the
devices in their microenvironments once fabricated. By using this approach we can both
scale down to the microscale and “scale up” to the assembly and control of massively
parallel device networks.
Intellectual Merit
As we shift to systems of smaller and smaller size scales, the need for new methods of
directing microscale transport will become acute. Taking advantage of the ability to
manipulate colloidal particles with external fields, we will both fabricate and operate
large-scale parallel device networks within microfluidic systems. Our preliminary
investigations have shown the clear interplay between confining geometry, selfassembled structure, and bulk field. The proposed investigations will seek to characterize,
understand and more directly take advantage of this coupling between applied field and
assembly function for improved microdevice operation in novel integrated microsystems.
Broader Impacts
Microfluidic-based technology has gained relevance through its promise to miniaturize,
consolidate and automate the time and sample-consuming steps inherent in a great
number of research disciplines. To date, however, microfluidic research has focused
upon fluid control schemes for the purpose of realizing complex micro total analysis
systems (TAS). The work proposed here will not only advance the TAS goal but, in
the process, develop new methods of device self-assembly for massively parallel
fabrication and actuation. Also, in the work described here the materials used are
biocompatible and of low toxicity, and the fabrication techniques relatively easy to learn.
These factors combine to lend this research extremely well to both graduate and
undergraduate level investigation. The proposed studies will provide the continued
support necessary to encourage these kinds of mentoring opportunities.
Actuation and In-situ Assembly of Colloidal Devices via Magnetic Fields
D.W.M. Marr
Chemical Engineering Department, Colorado School of Mines, Golden, CO 80401
Results of Prior NSF Support
NSF Award #CTS9734136 entitled “CAREER: Optical Manipulation of Colloidal
Particles”, award period 6/1/98-5/31/02. This award focused on the development of
techniques for directly manipulating colloidal particles using light to control formation of
large colloidal structures. NSF Award #CTS0097841 entitled “Sensing, Actuation, and
Flow Control with Colloidal Devices”, award period 4/15/01-3/31/04. This award focuses
on the use of optical trapping techniques for the development and actuation of colloidbased sensors, micropumps, and valves. NSF Award #CTS0304158 entitled “NER: 3D
Nano-Colloidal Crystallization via Electrokinetic Flows”, award period 8/15/03-7/31/04
focused on the use of electric fields to direct colloidal crystallization. The 14 publications
resulting from support under these NSF awards are given in the bibliography as [1-14].
Also, many of these results concern the use of optical manipulation techniques for the
simultaneous control of multiple colloidal particles and their assembly into small-scale
structures and are provided here as background information.
Introduction
For the practical realization of microscale and nanoscale pumps that mimic the operation
of their macroscale counterparts, two significant hurdles must be overcome. The first,
fabrication, is currently and in general being addressed through either improvements in
the resolution of macroscale techniques (“top down”) or through synthetic chemical
approaches to create increasingly complex molecular structures (“bottom up”). Though
both approaches have achieved some success, the creation of working devices within
their final functioning environments remains elusive. The second significant hurdle
involves the delivery of energy across length scales to effectively power microdevices in
a desired and controlled fashion once they have been created. In this proposal we present
a bulk-field based in-situ technique that addresses both issues by not only driving device
assembly from simple colloidal building blocks but also by subsequently powering the
devices in their microenvironments once fabricated. By applying global fields we will
avoid the need for individual manipulation allowing for both scale down to the nanoscale
and “scale up” to the assembly and control of massively parallel device networks.
Because of the small relevant length scales, microfluidics holds great promise as the basis
for future sensing and analysis systems. It offers both rapid transport and the ability to
manipulate smaller sample volumes than previously possible, advantages that are ideal
for the investigation of rapid kinetic processes in very expensive protein solutions. In
addition to laboratory studies however, these factors allow minute samples to be split to
perform multiple diagnoses simultaneously upon a single chip, moving analytical
processes out of the laboratory to point of care situations. As such, portable, handheld
sensors capable of collecting samples from dynamic environments and providing
immediate feedback may now be considered. One goal is the micro total analysis system
(TAS) where a single sample drop (blood, saliva, etc.) could be split, mixed with
varying solutions, and analyzed all in a single small device (the “lab on a chip”).
1
a)
b)
Figure 1: Pump design with 30° rotation steps to illustrate lobe movement (the top pair rotate
clockwise, the bottom counterclockwise). 3 m colloidal silica undergoing rotation at 2 Hz within a 6
m channel are also shown. Frames are separated by 2 cycles to show movement of the 1.5 m
colloidal silica tracer particles [11]. b) detailed simulations of these colloidal pumps show oscillatory
pump flow rate [15].
To achieve this goal will require the integration of a number of vital technologies,
including the ability to manipulate and transport, not only fluids and mass, but light and
information as well. Recently, our group has spent significant effort on the enhanced
manipulation and control of colloidal systems within microfluidic geometries. Using
optical trapping based techniques we have actuated individual colloidal-based
microdevices, making pumps, valves, and cell/particle separators. Because of the limited
capabilities of the specific optical manipulation tools employed however, the true goal of
an integrated design, one which simultaneously allows control of combined pumping,
valving, and separating architectures has not been achievable.
One approach to overcoming these difficulties is to move to better, more capable, optical
scanning techniques. These have been used to create extended arrays of trapping sites by
means of interfering laser beams [16-19], allowing the study of fundamental aspects of
phase transitions on patterned substrates, for example. In addition, scanned optical
tweezers can be used to create e.g. boundary boxes to two-dimensional systems which
allows to control particle densities very precisely [20] [21]. The same technique has been
also used to study the diffusion of particles along one-dimensional lines, a situation
usually referred to as single-file diffusion [22].
Such use of fast scanning techniques using acousto-optical deflectors (AOD’s) to create
more complex light patterns seems to be promising because it allows quasi-simultaneous
manipulation of up to a thousand particles under well-controlled conditions. However,
further scale up will be exceedingly difficult due to not only geometric concerns but also
due to the power requirements associated with simultaneously optically trapping a great
number of colloids. This need to both fabricate the colloidal devices and subsequently run
them will become a significant future limitation. In addition, as we create smaller feature
sizes, especially below optical wavelengths, the optical manipulation of colloidal systems
for such nanodevices becomes problematic. Focusing of the light and creation of the
2
necessary optical gradients for nanocolloidal assemblies becomes limiting and alternative
field manipulation approaches must be found.
The use of magnetic fields may address many of these issues that currently limit the
massive parallelization and nanoscale operation of colloidal devices. Because these
fields are external to the device and are not local, a system capable of operating a single
device will be equally capable of running as many devices as can be fabricated within the
available field geometry. In fact, because these magnetic fields can be used to modify
the interaction between individual colloidal device elements, they can be used not
only for actuation but in-situ device assembly. In addition, scale down to smaller sizes
no longer becomes limiting as the manipulation mechanism no longer requires local field
control.
Background
Colloidal Devices
Our group has developed functional devices out of
colloidal systems at length scales significantly smaller
than previously achieved by other techniques. He has
successfully created gear pumps, peristaltic pumps,
and two-way valves, all at sizes that approximate those
of a human red blood cell. The primary advantage of
doing microfluidics at such small scales is that vastly
smaller quantities are required than is needed for
current technologies; however, additional advantages
rely on the unique, viscosity-dominated, nature of the
fluid dynamics at these sizes. In these, we have
manipulated colloids by optical trapping, a noncontact, non-invasive technique that eliminates the
need to physically interface to the macroscopic world,
thus circumventing a traditional obstacle to
microfluidic device miniaturization. The optical
trapping principle is based upon a focused laser beam
encountering a colloid of refractive index different
than its surrounding solvent, causing the particle to
reflect and refract the focused beam. Such photon
redirection must be balanced by a change in colloid
momentum, the net result of which is the trapping of
small micron-sized objects in the focal point of a
converging laser beam [23]. In order to manipulate
complex asymmetric objects or multiple objects at
once, as is required for the actuation of a microfluidic
pump, a great number of optical traps are
simultaneously required. To accomplish this, we have
employed a scanning approach in which a
piezoelectric mirror is translated to rapidly reflect a
laser beam in a desired pattern. If the piezoelectric
a)
b)
Figure 2: 3 m colloidal silica used
as a peristaltic pump operating at 2
Hz inducing flow from right to left
within a 6 m channel. Frames are
separated by 2 cycles to show
movement of the 1.5 m tracer
particle [11]. b) simulation results of
resulting pumping action [15].
3
mirror is scanned over the desired pattern at a frequency greater than that associated with
Brownian time scales, a time-averaged trapping pattern is created. The details of this
approach, called scanning laser optical trapping (SLOT), can be found elsewhere [6, 7].
Under microfluidic conditions where viscous effects dominate, the fluid dynamics are
unique. The Reynolds number, defining the ratio of inertial to viscous forces, is very
small reducing the equations of motion to a simple time reversible differential form
known as the Stokes equation. These microfluidic flows are completely dominated by
viscous effects and are therefore laminar in nature, time reversible and turbulent free. The
physical nature of these microfluidic flows determines the approaches one can use in
designing both microscale pumps as well as microscale mixers which must rely heavily
on diffusion.
The first design is a two-lobe gear pump in which small, trapped pockets of fluid are
directed through a specially-designed cavity fabricated in a microchannel by rotating two
colloidal dumbbells or “lobes” in opposite directions. Over repeated and rapid rotations,
the accumulated effect of displacing these fluid pockets is sufficient to induce a net flow.
This motion is illustrated in Figure 1, where clockwise rotation of the top lobe combined
with counterclockwise rotation of the bottom lobe induces flow from left to right. In the
experiments also shown in Figure 1, each of the lobes consisted of two, independent 3
m silica spheres. To create these structures, the colloids were first maneuvered using the
optical trap to a 3 m deep section of channel designed with a region of wider gap to
accommodate lobe rotation. Once the particles were properly positioned, the laser was
scanned in a manner such that a time-averaged pattern of four independent optical traps
was created, one for each microsphere comprising the two-lobe pump. By rotating the
two traps in the upper part of the channel and the two traps in the lower part of the
channel in opposite directions and offset by 90°, the overall pump and the corresponding
fluid movement was achieved. Flow direction was easily and quickly reversed by
changing the rotation direction of both top and bottom lobes.
The gear pump design illustrates the success of positive displacement pumping through
the use of colloidal microspheres; however, its design may prove particularly harsh to
certain solutions. Though individual cells can be pumped using the gear pump,
concentrated cellular suspensions may be damaged by the aggressive motion of the
meshing “gears” of the pump. A second approach has been developed that incorporates a
peristaltic design also based upon the concept of positive fluid displacement, effectively a
pseudo two-dimensional analog of a three-dimensional, macroscopic screw pump. If
Figure 3: Because of the shape of the beam at the
optical trap focus (left) in “classic” 3D optical trapping
the relatively weak optical gradient force in the zdirection Fz must overcome gravity, strictly
constraining the optics. (right) In confined microfluidic
geometries the stronger lateral optical forces, Fx and Fy
dominate, allowing for greatly simplified optical
designs.
4
instead of rotating the particles as in the gear pump, they are translated back and forth
across the channel in a cooperative manner, fluid propagation can be achieved. One main
advantage of this peristaltic design lies in the simplified, reciprocal motion of the
microspheres, which may allow actuation by other methods such as electrophoresis.
The colloidal movement required to direct flow via the peristalsis approach is illustrated
in Figure 2. The optical trap moves the colloids in a propagating sine wave within which
a plug of fluid is encased. Direction of the flow can be reversed by changing the
direction of colloidal wave movement. Once again, these experiments were performed
with independent, 3 m silica spheres; however, more colloids were used in the
experiments of Figure 2 to represent a complete wavelength. Fabrication of these pumps
required first maneuvering the colloids into the channel section. Once in place, the optical
trap was scanned such that multiple independent traps were created, one for each colloid
compromising the peristaltic pump.
What truly enables this approach is that optical trapping in our systems occurs within
thin, confining geometries. Traditionally, and in non-confining 3D systems, much effort
has been expended in the optical design of laser traps that are very tightly focusing (e.g.,
the use high numerical aperture objectives). This has been driven by the need to create
strong optical gradients in the z-direction as illustrated in Figure 3. This is generally
required to overcome gravity and optical scattering
forces that typically come into play. With a confining
geometry, however, these requirements are greatly
diminished and one requires only optical intensity
gradients in the lateral dimensions, a condition that
greatly minimizes the optical design requirements. We
have taken full advantage of this by employing low
power objectives that allow us to simultaneously access
relatively large areas within the device at low
magnification.
In addition to pumps, simple valves can be created using Figure 4: a) An actuated, threea similar technique. These are shown in Figure 4, where way colloidal valve. [10].
the valves consist of a 3 m silica sphere
photopolymerized to several 0.64 m silica spheres forming a linear structure. For
passive application, the device was maneuvered into a straight channel and the 3 m
sphere held next to the wall allowing the arm to rotate freely in the microchannel. As the
flow direction is changed (Figure 4a) the valve selectively restricted the flow of large
particles in one direction while allowing passage of all particles in the other. To actively
direct particulates to one of two exit channels, the passive valve was maneuvered into a
confining T geometry. As the valve structure was rotated about its swivel point using the
optical trap, the top or bottom channel was sealed, directing flow of particulates toward
the open channel (Figure 4).
5
What We Propose
Our studies of the optical-based assembly and control of colloidal systems have
demonstrated that these devices can function in microfluidic systems. As discussed above
however, their utility can be limited in systems in which high degrees of parallelization
are desired; the power requirements needed to simultaneously fabricate and direct
systems of many devices becomes highly problematic. To overcome this limitation we
propose to investigate the use of bulk magnetic fields to both direct the assembly and
operation of microscale colloidal devices and do so in a highly parallel fashion.
Our preliminary investigations have demonstrated that, in combination with paramagnetic
colloidal particles, magnetic fields can be used to create analogous microdevices. In
addition however, we have observed the importance of confining microstructure in
determining assembled colloidal device function and it is exactly this interaction that will
be exploited in creating locally powered microstructures. Here, we will probe this
interplay between confining microgeometry, applied field, and assembled colloidal
morphology in determining device function. Our goal is to better understand
approaches for directing microscale transport that use simple fields and inexpensive
building blocks for the creation of easy to use microfluidic technologies.
…it (microfluidics) must become successful commercially, rather than remain a field
based on proof-of-concept demonstrations and academic papers. The impact of
microfluidic systems…will only become apparent when everyone is using them.
Microfluidics must be able to solve problems for users who are not experts in fluid
physics or nanolithography, such as clinicians, cell biologists, police officers or public
health officials.
-George M. Whitesides[24]
(Nature July, 2006)
Over the past decade there has been a tremendous growth
in the use of microfluidic systems for a variety of
proposed applications. A good review of the current state
of the art, the most pressing needs, as well as some of the
more promising applications can be found in a recent issue
of Nature [24-27]. Because these are relatively simple to
generate and provide the possibility of transferring energy
across length scales and without direct contact, magnetic
fields may solve some of the issues preventing wide-scale
adoption of microfluidic technologies. As such, previous
microfluidic studies have employed magnetic fields for
separations of cells [28, 29], separations using
superparamagnetic particles such as those we employ here
[30-32] as well as pumps [33, 34]. Though our approach
differs greatly from these previous investigations in that
we are creating very small-scale devices of distinct local
function, our goal is similar - to develop approaches to the
Figure 5: A vertically aligned
magnetic field can be used to
induce repulsions between
individual particles in a confined
microfluidic channel (top). With
fields instead applied in the
microfluidic plane, dipoles attract
leading to assembled
morphologies (bottom).
6
operation of microfluidic devices that are simple yet capable and practical to aid in the
push of microfluidic technologies into public use.
Preliminary Results
Magnetic Devices
As discussed previously we intend to develop complementary magnetic field
manipulation techniques to aid the assembly and operation of our fluidic-based colloidal
microdevices. For our preliminary investigations, Dynabeads® (www.dynalbiotech.com)
of diameter either 2.7 m or 4.5 m were used. Developed for bioassaying applications,
these readily-available particles are superparamagnetic due to the presence of Fe2O3 and
therefore exhibit magnetic properties only in the presence of a magnetic field. They are
available at low polydispersities making them a convenient system for our studies.
Our microfluidic systems are planar in nature and have been fabricated such that channel
height is typically little more than the particle diameter. This confinement plane provides
the reference point for the application of our external magnetic fields. Here coils are
placed in this same plane and external to the entire microfluidic device. Upon application
of current through these coils a magnetic field is created that induces an effective
attraction between Dynabeads. As the polarization of the magnetic field is rapidly rotated
using the three distinct coils, a torque is induced that can be used to rotate these colloidal
assemblies. In this setup, an optical trap has been included for ease of particle
manipulation.
Figure 6A-D demonstrate the assembly of seven 4.5 m particles into a compact rotating
cluster in the presence of a rotating magnetic field [35]. It has been shown previously [36,
37] that slow field rotation frequencies lead to the formation of chain-like structures
which rotate around their center of mass. We observe however that when the clusters are
located inside channels, compact structures independent of the rotation frequency are
always favored. This process also allows formation of independent clusters of different
sizes as shown in Figure 6E-H. In this, three particles are each forming a particle cluster
in the presence of a rotating field. Afterwards, an additional vertical field is turned on
which leads to a repulsion of the clusters and thus to the formation of two independent
Figure 6: Self-assembly of A-D) seven 4.5 m magnetic
particles in a channel structure into a compact micropump
in the presence of an external rotating magnetic field. In this
specific case, cluster formation is irreversible due to vander-Waals forces. Accordingly, the micropumps do not
disassemble in the subsequent presence of repulsive particle
interactions. E-H) The size of the resulting micropumps can
be adjusted by applying an additional static vertical field
inducing a repulsive particle interaction. Here, after the
formation of two three-particle clusters, the vertical field is
turned on. In these studies, super paramagnetic colloidal
particles of 4.5 or 2.7 m diameter were employed.
7
pumps. Note that cluster formation in these systems can be either reversible or
irreversible depending on specific colloid surface chemistry and strength of the applied
magnetic field.
Application of a rotating magnetic field to a compact particle cluster leads to a cluster
rotation rate dependent on a balance between viscous drag and the magnetic forces.
Figure 7A-D are frames taken from a video sequence showing a rotating cluster
composed of seven particles in a microchannel structure filled with water. The external
field rotates at a frequency of 100 Hz in the plane of the particles in a counter-clockwise
direction. It induces a torque on the cluster due to the interaction with the magnetic
dipoles and leads to a cluster rotation. In the case of the 7 particle cluster shown here this
leads to a maximum cluster rotation frequency of approximately 20 Hz. Due to the length
scales of our microfluidic channels and pumps, flow is laminar and a rotating particle
cluster can only induce a net flow if the channel symmetry is broken. We therefore
fabricated the channel walls with depressions on one side. When applying a rotating field,
the cluster aligns itself close to the curved side of the channel. This becomes apparent
when considering the flow created by the pump as shown by observing the motion of
tracers sketched in Figure 7E. Here, the pump is situated in a structured channel, with a
maximal width of approximately 16 m and height 6 m. Flow was visualized by the
motion of non-magnetic polystyrene tracer particles and it can be seen that pumping
increases with the strength of the rotating field (Figure 7F). This is measured by taking
the time the tracer needs to pass the bypass (the distance indicated by the dotted line in
Figure 7E) for different field strengths. In fact, because the pumps can rotate very
rapidly, we have been unable to quantify the exact rotation speed in the preliminary
setup. Though this is an issue we can address, the data of Figure 7F shows the speed of a
non magnetic colloidal tracer in the bypass channel as a function of the applied coil
voltage instead of pump rotation rate. Nonetheless, it can be easily seen that pumping
rate is strongly influenced by the speed of particle pump rotation.
Figure 7: Pumping mechanism. A-D), seven
super paramagnetic particles of 4.5 m diameter
form a cluster driven counter clockwise by an
external magnetic field that rotates at 100 Hz in
the sample plane. The maximum width of the
channels is 16 m, the height approximately 6
m. Liquid flow is visualized with a 3 m
polystyrene tracer particle. E), illustration of the
liquid flow originating from the rotating particle
cluster. F), velocity of tracer particles as a
function of the current through the magnetic
coils. The flow rate is measured in the upper,
approximately 5 m wide bypass channel.
8
In these studies we created pumps of two, three and seven particles in similar geometries
of varying channel width as well as pumps connected in series. We found that the pump
efficiency increases with increasing pump diameter. More particles have a bigger
collective magnetic moment, an effect that leads to faster rotation for a given applied
magnetic field. These larger clusters also have more surface leading to a stronger
hydrodynamic interaction with the surrounding fluid. In addition, pumps connected in
series lead to larger flows than single pumps. In this, the external application of the
magnetic field, leads to reversible aggregation of smaller numbers of paramagnetic
colloidal spheres. Seven such particles in a confined, two-dimensional geometry such as
we have here typically leads to a flower-like cluster. In this image, two such clusters have
been formed and, as the magnetic field is rotated, these cluster rotate as well. Though
certainly better seen in movie clips not available here, the cluster rotation leads to fluid
flow; this is verified by changing the rotation direction of the clusters where the tracer
now moves in the opposite direction. Using larger 4.5 m Dynabeads we have achieved
rapid rotation rates of at least 5 Hz, significantly increasing fluid flow velocities.
In our studies, the channels were designed to capture the pump in the asymmetric part
and prevent translation along the wall because of the strong interaction between walls and
cluster. This interaction can induce a small circular movement of the pump center of
mass, which has no observable influence on the pump efficiency. In addition however,
the microchannel design plays an important role in device function. As these devices are
powered using an external source, their rotation is driven in the same direction, a feature
that, at first glance, may limit function. However, pumping direction depends both on the
cluster rotation direction and the channel geometry. As illustrated in Figure 8 for pumps
rotating in identical directions, net flow is determined by the location of the channel
asymmetry. This is a vital point – though pump assembly and powering are driven by
the external field, pumping direction is dictated by the physical geometry in which
the pump is fabricated. We will take advantage of this in future studies for the creation
and integration of pumping networks.
Though colloid-based pumping has been demonstrated by us previously through the
using of scanning optical trapping techniques, our direct manipulation of the colloids had
Figure 8: Illustration of A) the flow lines induced by the
rotating cluster. Nearer points indicate faster liquid flow.
The asymmetry in the pump geometry is essential for two
reasons: It is required for a net flow in one direction and for
pump confinement, without which the cluster would
translate along the walls due to hydrodynamic interactions.
On average the cluster is localized in the concavity of the
asymmetry due to the Bernoulli effect. It is important to
realize that the liquid flow can be reversed either by
changing the rotation direction of the cluster or by moving
the cluster to a channel concavity with opposite orientation
B), C).
9
drawbacks. The most obvious is that of creating multiple pumps simultaneously where
splitting of the trapping laser optically is likely the most promising approach. The
significant advantage of the magnetic technique is that,
because the fabrication and actuation field is applied
externally, simultaneous creation and control of
multiple pumps is straightforward. This is
demonstrated in Figure 9 where two pumps are run in
parallel. Note that, because device rotation rate is fixed
externally, different flow rates can be achieved through
the use of multiple pumps in series.
The approach discussed here allows the simultaneous
creation of large numbers of micropumps inside
microfluidic devices. This is demonstrated in Figure 9
where six three-particle pumps composed of 2.7 m
Figure 9: A) Multiple colloidal
pumps run simultaneously. Here, particles rotate in the same direction with roughly the
because rotation rate is externally same speed. Though certainly more dense configurations
fixed, different pumping rates are are possible, this image corresponds to a pump density
achieved through the use of
of approximately 30,000 pumps/cm2. Note that the
multiple assemblies.
energy required to drive all of these individual devices
simultaneously is provided by a single external source.
Despite the large number of available pumps and the ability to direct pumping with static
channel designs, more dynamic control is of interest for some applications. In our
approach, a global field is used to power all of the individual devices simultaneously [38];
however, local modifications to the field or application of a separate, supplementary field,
can alter local function. For example, as illustrated in Figure 8, because the pumping
direction depends on the pump position within the channel geometry, a simple translation
of a micropump inside the channel structure will reverse flow if two neighboring channel
concavities with opposite orientations are present.
Research Design and Methods
Fabrication Methods
These microfluidic systems are assembled using a methodology coined “rapid
prototyping”. In this, and using standard photolithography techniques, a pattern is
produced on silicon or silicon dioxide substrates in thick SU-8 photoresist. Following the
photolithography step, the pattern is then used directly as a “master” to produce positive
relief replicas in PDMS, an optically transparent elastomer, a process that has come to be
known as “soft lithography” [39].
Specifically, templates of microchannels (Chs) and microfluidic networks (FNs) are
created lithographically with ultraviolet (UV) light by transposing the pattern of a shadow
mask to a UV sensitive negative photoresist. The patterns are subsequently developed in
an appropriate solution, leaving only the negative relief of the desired pattern, which may
be used directly as a PDMS master or etched to produce a permanent master. If used
directly to create PDMS replicas, photoresist films may be prepared with thickness from
25 nm to 250 m, thus providing a wide range of accessible sizes and aspect ratios.
10
Figure 10: Square wheels and other regular polygons translate well along roads of
specific geometry.
Except for situations in which extremely thin films are required, SU-8 series negative
photoresist (MicroChem Corp., Newton, MA) is employed, which is capable of
producing rugged patterns with high aspect ratios that can be directly cast into PDMS
replicas and reused many times.
The PDMS replicas are created using a commercially available two-component kit
(Sylgard 184 Kit, Dow Corning). A mixture of elastomer and curing agent are poured
over the silicon master and cured under vacuum to degas the elastomer solution. PDMS
makes an ideal candidate for FN production because it can be cured quite rapidly,
patterns are faithfully reproduced, even on the nanoscale and the process can be
conducted in a non-clean room environment [40, 41]. Once cured, PDMS replicas are
peeled from the master, leaving a clean, reusable template. The replica is finally placed
in conformal contact with either a glass slide or PDMS flat forming a tight, reversible
seal and enclosing channels capable of conveying fluids. PDMS is natively hydrophobic,
but can be easily modified to create a hydrophilic surface through brief exposure to an
oxygen plasma. Replica films as thin as 1 m may also be created by spin coating PDMS
onto a silicon master. Such films may be patterned and used as soft components such as
micro gaskets, seals and spacers for multilevel functional devices. Thicker films (> 40
m) may be removed from the substrate and used as shadow masks for the deposition of
metal features, such as electrodes, onto other replicas [42] or a wet etching mask for the
patterning of conducting tin oxides. Other microfluidic network concepts that are capable
of accessing the z-dimension through the stacking of multiple thin PDMS films have also
been employed [43].
Geometric Coupling of Assembly Rotation and Translation
The preliminary results above demonstrate the strong interplay between local geometry
and colloidal assembly and function. In this work we will take this one step further by
designing locally varying geometries that allow a coupling of field-induced rotation and
colloidal assembly translation. Such translations can be used to greatly improve the
pumping efficiencies demonstrated in our preliminary studies of magnetically assembled
structures. The general idea is shown in Figure 10 where a rotation of a square “wheel”
(here on a bicycle) leads to a smooth translation along a “road” of specific design. In fact,
for regular polygons the general solution is a series of truncated catenaries where y = 11
coshx and the angle at which these catenaries meet corresponds to the interior angle of
the polygon wheel [44]. Our intention here is to investigate this geometric coupling and
exploit it for the development of colloidal devices of varying function, including both
mixing and pumping devices driven by bulk magnetic field rotation.
Under the bulk field conditions we will employ colloidal particles are attracted within the
microfluidic plane leading to close-packed selfassembled structures (Figure 11). For small
numbers of particles this results in differing
external geometries, some of which are excellent
approximates to the regular polygons shown in
Figure 10 and we will initially focus our efforts on
the 3-mer, 6-mer, and 7-mer. It should be noted
Figure 12: Colloidal check valve
that mathematical solutions for the “roads”
corresponding to wheels of arbitrary structure such
as those constructed out of assembled spheres can also be generated; however, their
structure is not significantly different than the simpler solutions represented here. Also,
these structures have sharper profiles, features that replicate poorly during the soft
lithography process at the necessary resolutions corresponding to the colloidal particles
we employ here.
Task Plan
The combination of microfluidics and magnetic-field manipulation outlined in this
proposal provides the potential for massive parallelization on very small length scales.
Here, the task plan has been developed with the following goals in mind:
1) The construction of an experimental system capable of both magnetic field and
optical trapping based manipulation.
2) The demonstration of in-situ colloidal assembly through check valve construction.
3) The study of the coupling of colloidal assembly rotation and translation along
appropriate geometries for mixing in microfluidics.
4) Integration of 1-3 for demonstration of field-induced microfluidic pumping.
Tasks
Task 1 – Experiment Construction
The preliminary data generated for this proposal was obtained
during a von Humboldt fellowship at the Universität Stuttgart.
Here the experimental setup was constructed and basic optical
trapping integrated with magnetic field application and
rotation capabilities. Our goal in the first task is to create a
more capable version of this feasibility setup in our own
laboratories, including more advanced optical scanning (for
isolation of specific colloidal numbers) and magnetic field
generation geometries more amenable for microfluidic device Figure 11: Assembled
testing.
colloidal close-packed
Task 2 – In-situ Assembly
structures.
12
Check Valve: One of the significant advantages associated with the proposed approach is
that structures larger than the width of the fluid inlet and outlet streams can be
constructed in-situ using smaller colloidal building blocks. One example of a device that
requires such fabrication is the check valve as illustrated in Figure 12. Though
conceptually simple, fabrication of a check valve requires that the key element be
assembled within the device geometry (ie. the microscopic version of the macroscopic
construction of the ship in a bottle). Because individual colloids can be brought into the
microfabricated structure and then assembled upon application of the magnetic field, this
fabrication limitation is readily overcome.
Assembly Translation: Illustrated in Figure 13a, we will probe the coupling between
field-induced colloidal assembly rotation and confining channel geometry for directed
translation within microgeometries. In this we will use the soft-lithography techniques to
create appropriate concatenated catenaries of correct dimension, a process well within the
our resolution limits given chrome masks and micron-scale paramagnetic colloids.
Illustrated here is that the aggregate translation direction can be fixed for a given field
rotation direction simply by choosing the side upon which the “roads” are located. Note
also that because of the close registry of the colloidal assembly and the channel width, the
effective influence on fluid flow will be dramatic; the effective impact on fluid motion is
no longer simply a result of anisotropic decays in the surrounding laminar flow fields as
seen in our preliminary results.
Mixers: It is well known and a significant area of research in microfluidics that mixing in
microscale geometries is difficult due to the laminar nature of the fluid flows, the
associated lack of turbulence, and the resulting reliance on diffusion (see for example the
review [45]). One design that has been employed to mix in relatively large microfluidic
devices is to use multi-layer soft lithography to create pumps and valves and mix fluids
by circulating them within a circular channel [46], an approach that has been used, for
example, in screening for protein crystallization [47]. This technique however relies on
the flexibility of the polymer matrix
(preventing practical size reduction to smaller
length scales) and the supporting pressurization
equipment associated with operating many of
these pumps simultaneously limits very high
level parallelization. As depicted in Figure
a)
13b, we will employ a similar geometry in the
investigation of mixing using the coupling of
assembly rotation and translation. Here a
colloidal assembly will be created and used to
translate fluid upon application of the field
enhancing microscale mixing.
b)
Task 3 Integration of Multiple Assemblies
Figure 13: a) Assembly rotation coupled to
translation via wall geometry b) Magnetic
field rotation used to enhance mixing.
13
The previous tasks are employed to investigate issues regarding in-situ colloidal
assembly and geometric coupling in these magnetically-driven colloidal systems. Here
we will link individual devices to study their interaction for the creation of more
sophisticated devices. Such an example device is illustrated in Figure 14, where repeated,
alternating rotations of the magnetic field can be used to induce net fluid flow. In this,
assembly translations push fluid in a rectified fashion when employed with the check
valves developed in Task 1.
Timeline
Our expected timeline follows the outline of the task plan closely. Initial efforts will be
in the construction of the required experimental setup, an effort that will require some
time but should be straightforward because of our experience in similar setup
construction. This will lead into colloidal assembly, geometric coupling and finally into
individual device integration over the proposed two-year funding period.
Broader Impacts
Though this proposal is aimed at the creation and actuation of small micro devices, it is
clear that many of the applications will be those that are focused on microfluidics. In this,
the ability to rapidly, accurately, and simultaneously screen many samples will prove
instrumental in the development of sensors for applications in a host of fields. Remote
biosensing for
environmental monitoring,
the protection of armed
forces and civilian
populations from pathogens,
and even interplanetary
exploration are all needs
that may be satisfied
through the use of capable
microfluidic networks.
Environmental Health
Monitoring and
Bioprospecting: Often fieldFigure 14: Pumping using a translating colloidal assembly.
based research demands
simple, functional results
such as whether a particular sample is toxic or not, toxic to a particular organism, or toxic
to a cultured organ cell of an organism. Classified under this rubric are two seemingly
distant endeavors: environmental monitoring and bioprospecting. In each case, where
only a fatality event or persistent viability must be delineated, broad arrays of various
cells may be created and maintained for field screening of samples from ground water for
mercury to the extraction of a microbial endosymbiant for potential anti-tumor properties.
Clinical Health Monitoring: Similar to the detection of pathogens, particle and cell-based
sensor arrays may be customized to perform multiple immunoassays in parallel, and
therefore screen for a host of potential indicators of disease simultaneously. Beyond
14
being far more comprehensive than current techniques, the devices themselves will be
inexpensive and portable, thus enabling point-of-care diagnostics and home care. Given
our use of Dynabeads® such bioassaying applications appear a natural extension of much
of the work we intend to do. To eliminate the need for additional mixing steps, one can
even envision incorporating simple mechanical cell lysing via rotating, meshing colloids
and other such capabilities.
Microfluidic Separations: Though beyond the scope of this proposal, it is apparent that
magnetic fields can be employed to effect particle separation in ways similar to those we
have already demonstrated with optical trapping. In addition, Dynabeads have been
designed with specific functionalities for DNA, organelle, or specific protein isolation for
example. Taking advantage of these capabilities in future applications could add
significant capability for highly-specific microfluidic separations.
Education
Studies, such as those proposed here, provide a nice context in which both undergraduate
and graduate students learn to do research. The PI has been a very strong supporter of
these future scientists within his laboratory. Though not required in the Chemical
Engineering curriculum at CSM, the PI has had 30 undergraduates (of whom ten were
women) working on individual research projects. These currently include Regina
Hutchinson, senior; Justin Chichester, senior; and Jennifer Cho, senior.
Also, within the past two years, a high percentage of the undergraduates who have
worked in the research group have continued on to graduate school, indicating the
intellectual interest our laboratory investigations have helped foster. These include:
Rachel Fallon, University of Arizona – Chemical Engineering
Michael Gower, University of California, Davis - Bioengineering
Curt Schneider – Massachusetts Institute of Technology – Chemical Engineering
Matt Brown – Colorado School of Mines (MS) – Chemical Engineering
Tor Vestad – Colorado School of Mines (MS) – Chemical Engineering
These students have played a vital and important role in the research carried out within
the research group, as evidenced by publications including one in Science. In addition,
the materials required for these investigations consist primarily of PDMS (a relatively
safe material) and the techniques required build upon the tools used in microlithography.
The students therefore who gain experience within the laboratory and do not go on to
graduate school are well prepared for employment in the microelectronics industry, one
of the primary employers of BS level chemical engineers.
15
References
1.
Gong, T. and D.W.M. Marr, Electrically Switchable Colloidal Ordering in
Confined Geometries. Langmuir, 2001. 17: p. 2301-2304.
2.
Gong, T. and D.W.M. Marr, Photon-Directed Colloidal Crystallization. Applied
Physics Letters, 2004. 85: p. 3760.
3.
Gong, T., D.T. Wu, and D.W.M. Marr, Two-Dimensional
Electrohydrodynamically-Induced Colloidal Phases. Langmuir, 2002. 18: p.
10064.
4.
Gong, T., D.T. Wu, and D.W.M. Marr, Electric Field Reversible ThreeDimensional Colloidal Crystals (cover article). Langmuir, 2003. 19: p. 5967.
5.
Mio, C., et al., Morphological Control of Mesoscale Models. Fluid Phase
Equilibria, 2001. 185: p. 157.
6.
Mio, C., et al., Design of a Scanning Laser Optical Trap for Multiparticle
Manipulation. Review of Scientific Instruments, 2000. 71: p. 2196.
7.
Mio, C. and D.W.M. Marr, Tailored Surfaces Using Optically Manipulated
Colloidal Particles. Langmuir, 1999. 15(25): p. 8565.
8.
Mio, C. and D.W.M. Marr, Optical Trapping for the Manipulation of Colloidal
Particles. Advanced Materials, 2000. 12: p. 917.
9.
Oakey, J., J. Allely, and D.W.M. Marr, Laminar Flow-Based Separations at the
Microscale. Biotechnology Progress, 2002. 81: p. 1555.
10.
Terray, A., J. Oakey, and D.W.M. Marr, Fabrication of Linear Colloidal
Structures for Microfluidic Applications. Applied Physics Letters, 2002. 81: p.
1555.
11.
Terray, A., J. Oakey, and D.W.M. Marr, Microfluidic Control Using Colloidal
Devices. Science, 2002. 296: p. 1841.
12.
Vestad, T., D.W.M. Marr, and J. Oakey, Flow Control for Capillary-Pumped
Microfluidic Systems. Journal of Micromechanics and Microengineering, 2004.
14: p. 1503.
13.
Vestad, T., T. Munakata, and D.W.M. Marr, Flow Resistance for Microfluidic
Logic Operations (cover article). Applied Physics Letters, 2004. 84: p. 50745075.
14.
Viravathana, P. and D.W.M. Marr, Synthesis of Colloidal Aluminosilicate for
Light Scattering Investigations. Journal of Colloid and Interface Science, 2003.
265: p. 15.
15.
Liu, D., M. Maxey, and G.E. Karniadakis, Modeling and Optimization of
Colloidal Micro-Pumps. J. Micromech. Microeng., 1994. 14: p. 567-575.
16.
Wei, Q.-H., et al., Experimental study of laser induced melting in two dimensional
colloids. Phys. Rev. Lett., 1998. 81: p. 2606-2609.
17.
Bechinger, C., M. Brunner, and P. Leiderer, Phase behavior of two-dimensional
colloidal systems in the presence of periodic light fields. Phys. Rev. Lett., 2001.
86: p. 930-933.
18.
Brunner, M. and C. Bechinger, Phase behavior of colloidal molecular crystals on
triangular light-lattices. Phys. Rev. Lett., 2002. 88: p. 248302.
19.
Baumgartl, J., M. Brunner, and C. Bechinger, Locked-floating-solid to lockedsmectic transition in colloidal systems. Phys. Rev. Lett., 2004: p. in press.
16
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
Brunner, M., et al., Density-dependent pair-interactions in 2D colloidal
suspensions. Europhys. Lett., 2002. 58: p. 926-932.
Brunner, M., et al., Measuring the equation of state of a hard-disc fluid.
Europhys. Lett., 2003. 63: p. 791.
Lutz, C., M. Kollmann, and C. Bechinger, Single-file diffusion of colloids in onedimensional channels. Phys. Rev. Lett., 2004. 93: p. 026001.
Ashkin, A., Acceleration and trapping of particles by radiation pressure. Phys.
Rev. Lett., 1970. 24: p. 156-159.
Whitesides, G.M., The Origins and the Future of Microfluidics. Nature, 2006.
442: p. 368-373.
El-Ali, J., P.K. Sorger, and K.F. Jensen, Cells on Chips. Nature, 2006. 442: p.
403-411.
Psaltis, D., S.R. Quake, and C. Yang, Developing Optofluidic Technology through
the Fusion of Microfluidics and Optics. Nature, 2006. 442: p. 381-386.
Yager, P., et al., Microfluidic Diagnostic Technologies for Global Public Health.
Nature, 2006. 442: p. 412-418.
Inglis, D.W., et al., Continuous Microfluidic Immunomagnetic Cell Separation.
Applied Physics Letters, 2004. 85: p. 5093-5095.
Pamme, N. and C. Wilhelm, Continuous Sorting of Magnetic Cells via On-chip
Free-flow Magnetophoresis. Lab on a Chip, 2006. 6: p. 974-980.
Jiang, G. and D.J. Harrison, mRNA Isolation in a Microfluidic Device for
Eventual Integration of cDNA Library Construction. The Analyst, 2000. 125: p.
2176-2179.
Kim, K.S. and J.K. Park, Magnetic Force-based Multiplexed Immunoassay using
Superparamagnetic Nanoparticles in Microfluidic Channel. Lab on a Chip, 2005.
5: p. 657-664.
Tondra, M., et al., Design of Integrated Microfluidic Device for Sorting Magnetic
Beads in Biological Assays. IEEE Transactions on Magnetics, 2001. 37: p. 26212623.
Atencia, J. and D.J. Beebe, Magnetically-driven Biomimetic Micro Pumping using
Vortices. Lab on a Chip, 2004. 4: p. 598-602.
Hatch, A., et al., A Ferrofluidic Magnetic Micropump. J. MEMS, 2001. 10(2): p.
215-221.
Bleil, S., D.W.M. Marr, and C. Bechinger, Field Mediated Self Assembly and
Actuation of Highly Parallel Microfluidic Devices. Applied Physics Letters, 2006.
88: p. 263515.
Sonia Melle, O.G.C., Miguel A. Rubio, Gerald G. Fuller, Microstructure
evolution in magnetorheological suspensions governed by Mason number.
Physical Review E, 2003. 68: p. 041503.
Yukata Naogaoka, H.M., Toru Maekawa, Dynamics of disklike clusters formed in
a magnetorheological fluid under a rotational magnetic field. Physical Review E,
2005. 71: p. 032502.
Jung, C.W. and P. Jacobs, Physical and chemical properties of
superparamagnetic iron oxide MR contrast agents: ferumoxides, ferumoxtran,
ferumoxsil. Magnetic Resonance Imaging, 1995. 13(5): p. 661-74.
17
39.
40.
41.
42.
43.
44.
45.
46.
47.
Xia, Y. and G.M. Whitesides, Soft Lithography. Annu. Rev. Mater. Sci., 1998. 28:
p. 153-184.
Duffy, D.C., et al., Rapid prototyping of Microfluidic Systems in
Poly(dimethylsiloxane). Anal. Chem., 1998. 70(23): p. 4974-4984.
Folch, A., et al., Molding of deep polydimethylsiloxane microstructures for
microfluidics and biological applications. Transactions of the ASME, 1999. 121:
p. 28-34.
Duffy, D.C., et al., Rapid prototyping of microfluidic switches in poly(dimethyl
siloxane) and their actuation by electro-osmotic flow. J. Micromech. Microeng.,
1999. 9: p. 211-217.
Schneider, C., et al., Laser Fabrication of Three Dimensional Microfluidic
Networks for Droplet Formation. submitted to Advanced Materials.
Hall, L. and S. Wagon, Roads and Wheels. Mathematics Magazine, 1992. 65(5):
p. 283-301.
Ottino, J.M. and S. Wiggins, Introduction: Mixing in Microfluidics. Phil. Trans.
R. Soc. Lond. A, 2004. 362: p. 923-935.
Chou, H.P., M.A. Unger, and S.R. Quake, A Microfabricated Rotary Pump.
Biomedical Microdevices, 2001. 3(4): p. 323-330.
Hansen, C.L., M.O.A. Sommer, and S.R. Quake, Systematic Investigation of
Protein Phase Behavior with a Microfluidic Formulator. PNAS, 2004. 101: p.
14431-14436.
18