Download Optical detection of electrokinetically manipulated single molecules

Optical detection of electrokinetically manipulated
single molecules in a nanofluidic chip
Gea O.F. Parikesit*a, Vladimir G. Kutchoukovb,
Andre Bosscheb, Ian T. Younga and Yuval Garinia
Quantitative Imaging groupa, Electronic Instrumentation groupb,
Delft University of Technology, Lorentzweg 1, 2628 CJ Delft, The Netherlands.
ABSTRACT
We report on the progress of a novel nanofluidic device for detecting and manipulating single molecules in solution. This
paper discusses the development of an earlier proposed molecule separation method, where electrokinetic forces separate
different molecules based on their masses and charges. Optical imaging using confocal microscopy is applied to perform
the detection of the single molecules. Potential applications of this device will be assessed. This research aims for the
high spatial and spectral resolutions, both in manipulation and detection, which can lead to molecular identification.
Keywords: optical detection, nanofluidics, single molecules, lab on a chip, molecular identification.
1. INTRODUCTION
In conventional measurements, where only averaged properties of the measured ensemble are sensed, the properties of
single molecules remain hidden. By being able to detect the behavior and properties of those individual molecules, the
limit of the measurement technology will be achieved. Therefore, the research on single molecules has been progressing
rapidly in the recent years1,2.
The study of single molecules can be performed using various techniques, among others by employing small-scale
fluidic channels. The molecules are put in the liquid solution, which in turn flows through the fluidic channels. The
dimensions of the channels confine the position of the molecules. In our approach, the channels’ height are in the range
of 50-200nm, while the lateral dimensions are in the micron scale. Thus, in a sense, our solution of molecules flows in a
2D-like manner.
Our fluidic channels are built inside some chips, using fabrication technologies commonly used for silicon-based
integrated-electronic chips. This approach is in conjunction with the research trend on laboratory miniaturization,
commonly known as LOC (lab on a chip) or PTAS (micro total analysis system). Manipulation as well as detection of
the molecules can be performed on the chips. Due to the nanometer-scale of the channels’ height, we call our chips the
nanofluidic chips.
Electrokinetic forces are used to manipulate the solution of molecules. We avoid any mechanical parts that can move
inside our chips, because they have inherently low robustness and durability. Electrokinetic forces are also preferred
because they result in a plug-like velocity profile of the flow, providing much less molecule dispersion than conventional
pressure-driven flow’s velocity profile3,4.
Detection of single molecules using optical methods is chosen in this research, because they provide high sensitivities.
Moreover, they are also capable of detecting multiple wavelengths simultaneously. Another main advantage is the
abundance of optical methods that can be used on the same chip without modifying it.
*
[email protected]; phone: +31 15 278 6054; fax +31 15 278 6740
Microfluidics, BioMEMS, and Medical Microsystems III, edited by Ian Papautsky,
Isabelle Chartier, Proceedings of SPIE Vol. 5718 (SPIE, Bellingham, WA, 2005)
0277-786X/05/$15 · doi: 10.1117/12.583566
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
133
The ultimate goal of this study is to achieve molecular identification and classification. To perform those things, high
spatial and spectral resolutions are required. Therefore the resolutions, in electrokinetic manipulation as well as in optical
detection, must be defined and improved.
In our research, we develop some nanofluidic devices, each designed with different objectives. Fabrication of all devices
was done using silicon-based microchip technology, particularly with the methods we developed recently6. Using one of
those devices, we proposed a novel molecule separation method5. This paper reports on its development. The molecule
separation method is reviewed in section 2. Section 3 provides simulation of the forces inside the device. Our
experimental set-up is described in section 4, while the results and discussions are given in section 5. Definition of the
resolutions in our set-up is the subject of section 6. Section 7 foresee the potential applications of the particular device.
Section 8 concludes this paper and provides the direction for further research.
2. MOLECULE SEPARATION METHOD
Three different electrokinetic forces will be employed in this research, namely the electroosmosis, electrophoresis, and
dielectrophoresis. In electroosmosis, as liquid solution fills a channel, an electric double layer will exist in the interface
between liquid and solid. This layer consists of the negative ions of the walls and the positive ions of the solution,
attracted to each other. If the inlet and the outlet of the channel are connected to a positive and a negative electrode,
respectively, than the solution’s positive ions in the electric double layer will be attracted to the outlet. Due to the
viscosity of the fluid, those positive ions will drag the whole bulk fluid along. Eventually the whole solution will move
with a velocity of
v EO
H 0H l ]
E
K
,
(1)
where İ0 is the vacuum permittivity, İl is the relative permittivity of the liquid, ȗ is the electric potential in the double
layer, Ș is the fluid viscosity, and E is the electric field vector. Therefore the electroosmotic force3,4 is given as
FEO
ml
dv EO
,
dt
(2)
where ml is the mass of the bulk liquid.
When charged molecules or particles are added into the solution described above, then they will be attracted either to the
positive or negative electrode. Therefore positively charged molecules and particles will move to the outlet of the
channel, while negatively charged counterparts will move to the inlet. This attraction forces between opposite sign of
charges is the electrophoresis force3,4, defined as
(3)
FEP q p E ,
with qp as the charge of the particle.
Meanwhile, in the case of dielectrophoresis, we have neutral (i.e. with zero net charge) molecules or particles in the
solution. When applied to an electric field, they will be polarized and form dipoles. Should the electric field is nonuniform, the two types of poles will feel attraction forces to different directions with unequal magnitudes. Because the
resultant force is independent of the orientation of the electric field, then AC as well as DC electric fields can be used.
Normally, due to the resultant force, those molecules or particles will move toward the highest electric field; this is called
the positive dielectrophoresis. However, if the solution medium is more polarizable than the molecules or the particles,
the aligning of the dipoles to the electric field will experience lagging. Electric field will be distorted, the dipole
orientation will be inverted, and the molecules or the particles will move toward the lowest electric field; this is called
the negative dielectrophoresis. The dielectrophoresis force3,4 is expressed as
FDEP
1
Re ^[4SH l rp 3 K (Z )E] [’E* ]` ,
2
(4)
where rp is the particles’ radius, Ȧ is the electric field frequency, and K(Ȧ) is the Claussius-Moratti factor described by
K (Z )
134
H cp H lc
,
H cp 2H lc
(5)
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
with ȑp and ȑl the complex permittivities of the particles and the liquid, respectively. When Re{K(Ȧ)} is positive, the
particles experience positive dielectrophoresis, while a negative Re{K(Ȧ)} leads to negative dielectrophoresis. Therefore,
different Ȧ could lead to different types of dielectrophoresis acting on a certain particle.
The electrokinetic forces described above can be used to manipulate the solution of molecules or particles inside
nanofluidic channels. One of our designed nanofluidic devices is depicted in Figure 1 in top-view, in which those forces
will be employed to perform separation. This device’s inlet and outlet, together with the flow’s direction, can be seen in
Figure 1(a). The channels’ height is 150 nm. Because this height is much less than the channels’ width or length, the
two-dimensional flow assumption can be applied.
The solution will pass a straight channel with the width of 100 µm, before entering a semi-circular chamber with a radius
of 1 mm. After the chamber, twelve 100µm-wide sub-outlets are available. In this preliminary design, the sub-outlets are
united with a main outlet with the width of 273 µm. Nevertheless, modification in future designs could also results in
those sub-outlets leading to different new chambers or channels. At the right-upper and left-upper corners of Figure 1(a),
two separated electrodes were made, encircling the semi-circular chamber. A microphotograph of the fabricated device is
given in Figure 1(b), where these two electrodes are more clearly visible.
(a)
(b)
(c)
(d)
Figure 1. The nanofluidic device which is designed for separation of molecules or particles.
(a) The schematic design, with arrows showing the incoming and outgoing direction of flow; (b) A microphotograph of the fabricated
structure. A and B show the two microelectrodes used for dielectrophoresis; (c) AC electric field between the dielectrophoresis
electrodes; (d) Traveling path of charged particles (dashed line), neutral particles with positive dielectrophoresis (straight line), and
neutral particles with negative dielectrophoresis (dotted line).
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
135
Separation will be performed based on the different physical properties of the particles or the molecules. Positive and
negative voltages will be applied to the inlet and the outlet, respectively. Generally the solution will move from the inlet
to the outlet, because of electroosmosis. In this design, the magnitude of electrophoresis is significantly smaller than
electroosmosis. Therefore, charged particles will also move from the inlet to the outlet. Nevertheless, their velocity
depends on the electrophoresis force applied on them. Positive molecules will travel faster than the solution, while
negative molecules will experience the opposite.
Moreover, AC voltages will be applied through the two dielectrophoresis microelectrodes. The induced AC electric field
is illustrated in Figure 1(c). AC voltages are preferred than their DC counterparts, so that no unwanted electrophoresis
will be induced by the two dielectrophoresis microelectrodes. Along the wall of the semi-circular chamber, magnitude of
the electric field is the highest. Meanwhile, the lowest electric field will exist in the center of the chamber. Therefore, the
neutral molecules or particles that experience positive dielectrophoresis will be attracted to the wall, thus traveling close
to it with a longer path. They will enter the sub-outlets further away from the inlet. On the other hand, the neutral
molecules or particles with negative dielectrophoresis will be attracted to the center of the chamber. Hence they will
travel with a shorter path, and exit through the sub-outlets closer to the inlet.
As the result, the molecules or particles with different properties will travel in different paths and velocities inside the
structure, as illustrated in Table 1 and Figure 1(d).
Table 1: Different charge and polarizability will lead to different paths and velocities inside the nanofluidic structure.
Neutral
Charged
Particle properties
Easily polarized (Pos. DEP)
Not easily polarized (Neg. DEP)
Positively charged
Negatively charged
Path length
Long
Short
Normal
Normal
Velocity
Normal
Normal
Fast
Slow
3. SIMULATION OF FORCES
Simulation was performed using analytical calculation on the field of all electrokinetic forces. The goal is to predict the
behavior of molecules or particles inside the semi-circular chamber. Based on the simulation result, the feasibility of the
proposed separation method can be assessed. Some results are shown in the figures 2 and 3. The horizontal and vertical
axis in the figures represents an arbitrary coordinate system, with the center of the chamber as the origin.
Figure 2(a) shows the force field for the case of electroosmosis and electrophoresis. The field of those two forces are
similar, because they both depend on the same electric field, induced by the electrodes in the ends of the nanofluidic
channel. In either case, the magnitude and the direction of the forces depends on the polarity of those electrodes. Figure
2(a) particularly illustrates the case of positive electrode in the inlet and negative electrode in the outlet. Meanwhile,
Figure 2(b) represents the force field of negative dielectrophoresis. For the case of positive dielectrophoresis, the force
field will look similar except for the reversed direction. Both types of dielectrophoresis depend on the same electric field,
exerted by the electrodes in the semi-circular chamber’s wall.
When charged molecules or particles flow in the device, they feel the combination of electroosmosis and electrophoresis
forces. Because the fields of those two forces are similar, the resultant force field depends solely on the ratio of the
forces. As described in Section 2, the molecules’ charge determine whether the molecules move faster or slower than the
fluid medium. In the meantime, the combination between electroosmosis and dielectrophoresis determine the behavior of
neutral molecules or particles. The modeling results are shown in Figure 3. Different ratio are applied on the
electroosmosis and dielectrophoresis forces. Figure 3(a) and (b) represents the 1:1 and 10:1 ratio between electroosmosis
and negative dielectrophoresis. The same combination, but where the positive dielectrophoresis replaces the negative
dielectrophoresis, are illustrated in Figure 3(c) and (d). From Figure 3, we can predict the flow of the neutral molecules
as a function of the combined forces.
Simulation results in Figure 2 and 3 generally confirm the feasibility of the molecule separation method described in
section 2. However, more simulation will be done using numerical methods. The numerical model should provide the
complete flow properties of the molecules.
136
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
(a)
(b)
Figure 2. The analytic calculation of force fields, in the case of (a) electroosmosis and electrophoresis, and (b) dielectrophoresis.
(a)
(b)
(c)
(d)
Figure 3. The analytical calculation of force fields, for the case of combined electroosmosis and dielectrophoresis.
(a) Electroosmosis : Negative Dielectrophoresis = 1:1, (b) Electroosmosis : Negative Dielectrophoresis = 10:1,
(c) Electroosmosis : Positive Dielectrophoresis = 1:1, (d) Electroosmosis : Positive Dielectrophoresis = 10:1
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
137
4. EXPERIMENTAL SET-UP
The setup built for this research comprises several main parts: the nanofluidic chip itself, an interface, a Leica confocal
epi-fluorescence microscope, and a Hamamatsu camera. The nanofluidic chips, with the devices built in them, have been
described in the previous sections. To provide a stable fluidic and electronic connection between the chip’s features
(channels’ inlet/outlet, the micro-electrodes) and the outer-world instruments (fluid dispensers, fluid reservoirs, voltage
sources, voltmeters), we built an interface. Developed from Perspex plastic, this interface is equipped with 8 fluidic ports
and 8 electrode ports.
The fluidic ports connect the channels’ inlet/outlet to fluid dispenser/reservoirs. These ports also have gold wires to
enable electrokinetic manipulation directly to the solution. Meanwhile, using some pogo-pins and gold wires, the
electronic ports connect the microelectrodes to voltage sources/sensors, allowing electrokinetic manipulation through the
channels’ walls. A window exists in the middle part of the interface, allowing transmission and reflection microscopy
through the chips. The interface holds the chip under the confocal epi-fluorescence microscope, which in turn is
connected to the camera. The confocal ability of the microscope gives lateral resolutions7, while the nm-scale height of
the channels in the chip provides the axial resolution (and also optical sectioning) for the microscope. The epifluorescence setup enables specific fluorescence excitation and emission wavelengths applied on the solution of
molecules inside the chip. The camera captures the signal from the microscope, and a software interface connects it to a
computer.
5. RESULTS AND DISCUSSIONS
An experiment was done by diluting fluoresence beads (FluoSpheres from Molecular Probes, with 20nm in diameter, and
the emission wavelength of 515 nm) with distilled water. A voltage difference of 150 Volts was applied between the inlet
and the outlet. The result is shown in Figure 4.
(b)
Trace length (in micron)
8
(a)
6
4
2
0
0
5
10
15
(c)
Figure 4. The electroosmotic flow of the solution of fluorescence beads in water.
(a) The trajectories of beads observed in the experiment;
(b) The schematic of the device, where the white rectangle depicts the region of interest;
(c) The measured trace length, based on the brightest trajectory in figure (a), for the first 15 traces; black dotted line is
measured data, and red straight line is the trend of the data.
138
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
Due to the electroosmosis flow, the solution of beads enter the chamber. The trajectories of the beads, as illustrated in
Figure 4(a), are observed by analyzing the recorded flow sequence. The region of interest in the device is shown by the
white rectangle in Figure 4(b). The unequal brightness of the trajectories can be explained by the various sizes of the
beads. Even though the specification states the diameter as 20nm, the real dimension might vary. This can be avoided by
using a filter.
As seen in Figure 4(a), the trajectories are shown as dashed lines with unequal dash lengths. Those dashes represent the
‘trace lengths’ of the beads during each exposure time of the camera. The ‘trace lengths’ of the brightest trajectory was
measured for the first 15 traces. The measured ‘trace lengths’ (in micron) are shown as the black line in Figure 4(c), with
the trend given as the red line. The trend shows that the ‘trace lengths’ get shorter when the bead goes into the chamber.
In other words, the beads became slower. This agrees with the assumed laminar flow, where the velocity will be reduced
as the flow goes from a smaller cross-section area (i.e. the inlet) into a bigger one (i.e. the semi-circular chamber).
In the experiment described above, it took about 30 minutes for the water-based solution to enter and fill the nanofluidic
device. Therefore we tried to use another solution. Ethanol-based solution was chosen, because ethanol’s surface tension
is smaller than water. We diluted Rhodamine 6G with ethanol, and fill another device with it. The result was satisfactory,
where now only about 10 minutes were required. Snapshots of the filling of the solution are given in Figure 5(a)-(e). The
regions of interest of those snapshots are depicted in Figure 5(f). The solution is seen as start filling from the walls, due
to the capillary forces.
(a)
(b)
(c)
(d)
(e)
(f)
Figure 5. (a)-(e) The snapshots of the filling of the device using an ethanol-based solution, and
(f) and the location of the regions of interest.
6. DEFINITION OF RESOLUTIONS
To be able to identify and investigate single molecules properly, high spatial and spectral resolutions are required. The
common strategy to achieve that is by improving the resolutions limits of the optical detection scheme. However,
improvement can also be done on the manipulation techniques. Resolutions limits of the manipulation scheme determine
the highest accuracy of control of the molecules inside nanofluidic devices.
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
139
Definitions of the resolution limits should be made. Based on them, actual limitations can later be measured
quantitatively, and then improvements can be performed. Those definitions are summarized in Table 2. The spatial
resolutions are related to the properties in space, while the spectral resolutions are considered as the properties dependant
on frequencies or wavelengths. For the resolutions in detection, the limitations are determined by the microscope, the
camera, and the fluorescence signals8. As for the resolutions in manipulation, the limits are mainly due to the
electrokinetic forces, the interaction between the solution and the channels, the properties of the molecules, and the
geometry of the devices3,4.
Table 2: Definition of the resolutions limits of the system.
Spatial
Spectral
Manipulation
The highest controllable
accuracy of size, motion, and
position difference of molecules.
The smallest controllable
difference of electrokinetic
frequency response.
Detection
The smallest observable
size, motion, and position
difference of molecules.
The smallest observable
color difference, emitted by
molecules.
7. POTENTIAL APPLICATIONS
The nanofluidic device described in section 2 will be developed further, and its capabilities will be explored. Several
potential applications for the device are already foreseen. Definitely, the first potential application is the separation of
molecules. A proposed separation method has been described in section 2. Molecules or particles could be classified, by
observing their trajectory or by capturing them in different sub-outlets, as a function of their physical properties. The
second potential application is the mixing of two or more flowing solutions. A modulation on the dielectrophoresis
frequency would induce a modulated cycle between positive and negative dielectrophoresis, as shown from equations (4)
and (5). Hence the total force field, due to the combination between electroosmosis and modulated dielectrophoresis,
would look like a periodic cycle consisting the fields shown in Figure 3(a) to (d). In that way, should several solutions of
molecules enter the chamber, they would be stirred by the alternating total force field.
The third potential application is to perform single molecule studies. For instance, interaction between biomolecules
could be explored. If the ends of DNA strands are bound to polarizable neutral tagging particles, then interacting
complimentary strands would reduce the distance between the particles. The two particles being close to each other could
behave differently under the combinatory force field, than other single particles. Thus by monitoring the behavior of the
tagging particles, interaction between the biomolecules could be revealed.
Ultimately, all the potential applications of this device should be able to integrate with each other. The integration,
formed in a series of device, would provide another alternative in the field of ‘lab on a chip’.
8. CONCLUSION
The latest development of a novel nanofluidic device has been reported. An earlier proposed molecule separation method
was reviewed. The feasibility of the method was investigated, by performing simulation on the field of the electrokinetic
forces. Further simulation will be focused on numerical methods, to provide the complete flow properties inside the
nanofluidic device.
Experimental results show that the trajectories of the beads agree with the laminar flow assumption. Measurement on one
of the trajectories proves that the velocity decreases as the cross-section area increases. Faster filling of the device was
desired, and replacing water-based solution with ethanol-based solution solved the problem.
Resolution limits of the system, both in manipulation and in detection, are defined. Therefore the actual limits can be
measured quantitatively later and be improved. Meanwhile, foreseen potential applications of the novel nanofluidic
device are given. Development of the device will aim to fulfill those potentials, and to enable molecular identification
and investigation.
140
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
ACKNOWLEDGEMENTS
The Dutch Foundation for Fundamental Research on Matter (FOM) financially support this research.
REFERENCES
1. Ch. Zander, J. Enderlein, R.A. Keller (eds) Single Molecule Detection in Solution – Methods and Applications, Wiley
VCH, 2002.
2. Y. Ishii, T. Yanagida, Single Molecule Detection in Life Science, Single Molecules, vol. 1, pp. 5-16, 2000.
3. M.P. Hughes, Nanoelectromechanics in Engineering and Biology, CRC Press, 2003.
4. N. Nguyen, S. Wereley, Fundamentals and Applications of Microfluidics, Artech House, 2002.
5. G.O.F. Parikesit, V.G. Kutchoukov, W. van Oel, G. Liqui Lung, A. Bossche, I.T. Young, and Y. Garini, Optical
detection of single molecules in nanofluidic chips, Proceedings of SPIE Annual Meeting, 2004, accepted.
6. V.G. Kutchoukov, F. Laugere, W. van der Vlist, L. Pakula, Y. Garini, and A. Bossche, Fabrication of nanofluidic
devices using glass-to-glass anodic bonding, Sensors and Actuators A, vol. 114, no. 2-3 (Selected papers from
Transducers 03), pp. 521-527, 2004.
7. J.B. Pawley (eds.), Handbook of Biological Confocal Microscopy, Kluwer Academic Publisher, 1995.
8. I.T. Young, Quantitative microscopy, IEEE Engineering in Medicine and Biology Magazine, vol. 15, no. 1, 59-66,
1996.
Proc. of SPIE Vol. 5718
Downloaded from SPIE Digital Library on 21 May 2010 to 131.180.130.114. Terms of Use: http://spiedl.org/terms
141