View Article Online / Journal Homepage / Table of Contents for this issue
`
`Integrated flow-cells for novel adjustable sheath flows
`
`J. H. Nieuwenhuis,a J. Bastemeijer,b P. M. Sarroc and M. J. Vellekoopa
`
`a Industrial Sensor Systems, Institute IEMW, Vienna University of Technology, Austria
`b Electronic Instrumentation Lab. – DIMES, Delft University of Technology, The Netherlands
`c Laboratory of ECTM – DIMES, Delft University of Technology, The Netherlands
`
`Received 30th October 2002, Accepted 26th February 2003
`First published as an Advance Article on the web 13th March 2003
`
`In this paper two integrated flow-cells are presented that can generate novel sheath flows. The flow-cells allow for
`dynamic orthogonal control of the sample flow dimensions. In addition to this, the sample flow can be freely
`positioned inside the channel. The flow-cells are attractive, because they are very simple to fabricate and are
`compatible with the integration of sensors. Experiments have been carried out demonstrating that the sample flow
`dimensions can be controlled over a wide range; also the results show good agreement with finite element
`simulation results.
`
`Introduction
`
`In this paper two integrated flow-cells are presented that can
`generate novel sheath flows. The flow-cells presented here
`allow dynamic orthogonal control of the sample flow dimen-
`sions. Also the sample flow can be freely positioned inside the
`flow-channel so that the sample flow touches or nearly touches
`one side of the channel; this assures a good contact with any
`integrated sensor interface. Finally, another attractive feature of
`the flow-cells presented here is that they are very simple to
`fabricate, consisting of a two layer structure and requiring only
`2 etch-steps to fabricate.
`Due to the limited interaction between the sample flow and
`sheath flow the sample flow can be considered a virtual flow
`channel with adaptable dimensions (e.g. diameter 10% of the
`physical flow-channel). As a result the presented sheath flows
`combine the advantages of a large diameter channel with those
`of a small diameter channel: due to the small dimensions of the
`sample flow the low detection limits and modest sample
`consumption of a small channel device can be achieved; at the
`same time the large physical dimensions of the channel alleviate
`many problems such as clogging, air bubbles and strict
`fabrication tolerances. Most of these advantages are obvious,
`but the final advantage might need some further explanation.
`The imperfections in the physical flow channel are smoothed
`out by the sheath liquid, which is in direct contact with the wall.
`The sample flow, which is screened from the wall by the sheath
`liquid, will not be disturbed as much. Therefore, the imperfec-
`tions of the channel are not as critical as without the application
`of a sample flow.
`Because of the advantages mentioned above, a number of
`realisations of sheath flow on a chip have been presented in the
`literature. Most of them are limited to a semi-sheath flow,
`consisting of a layered-flow configuration. Here the sample
`liquid is confined in one dimension between the sheath liquid
`and the wall of the physical channel1,2 or between two layers of
`sheath liquid.3,4,5 In the other dimension the sample flow is
`fixed by the physical channel. Therefore these are not sheath
`flows in the classical sense. However, a few flow-cells have
`been presented where a classical, coaxial sheath flow was
`realised on chip (in a coaxial sheath flow the sheath liquid
`surrounds the sample flow completely). One of the first
`realisations consists of a miniaturised cytometer configuration,
`which resulted in a rather complex 5-layer device.6 More
`recently less complicated devices were demonstrated where a
`special ‘U-shaped’ sample inlet7 or a circular ‘chimney-shaped’
`
`56
`
`Lab Chip, 2003, 3, 56–61
`
`sample inlet8 was used to lift the sample liquid from the bottom
`of the channel. In these flow-cells the shape of the sample flow
`is fixed by the geometry of the flow-cell and therefore these
`flow-cells offer little control over the sample flow dimensions.
`The flow-cells described above use a sheath liquid to confine
`the sample liquid, but also a rare example is known where a gas
`was used to confine the sample flow.9 The main problem with
`this device is that special measures are required, such as
`controlling the surface properties of the channel walls, to create
`a stable flow.
`The flow-cells presented in this paper distinguish themselves
`from the flow-cells known in the literature by their versatility.
`Both layered sheath flows and coaxial sheath flows can be
`realised within the same device. What is even more important,
`due to a number of orthogonal control mechanisms both the
`dimensions and the position of the sample flow can accurately
`be controlled. As a consequence the flow-cells can create an
`optimal sample flow for each specific application. And because
`they can be fabricated by adding a few post-processing steps to
`a standard IC-process, many known sensors can be integrated
`into the device.
`Some examples of applications are: an integrated Coulter
`counter,10 sensors for particle shape analysis11 and sensors for
`cell analysis and (bio)chemical analysis.
`
`Sheath flows for microfluidics systems
`
`Traditionally, sheath flows have been applied in cytometers12 to
`confine the sample liquid into a very narrow sample stream for
`optical analysis, mainly by fluorescence and light scattering
`techniques. By integrating the flow-channel on a chip it has
`become possible to include the flow-channel and the sensors in
`the same device. The flexibility of the flow-cells presented here
`allows optimisation of the sample flow dimensions and sample
`flow position to the integrated sensor. For example, for an
`optical evanescent wave sensor an optimal sample flow would
`have a small height and would be positioned on the same side as
`the sensor interface,2 whereas for an integrated Coulter counter
`a sample flow with equal horizontal and vertical dimensions
`will be optimal.10
`Due to the small channel dimensions of microfluidics
`systems the Reynolds number in these devices is usually very
`low. As a consequence the flow behaviour in these systems is
`typically completely laminar and no turbulence occurs. There-
`
`DOI: 10.1039/b210724d
`
`This journal is © The Royal Society of Chemistry 2003
`ABS Global, Inc. and Genus plc – Ex. 1014, p. 56
`
`
`
`Published on 13 March 2003. Downloaded by University of California - Los Angeles on 04/10/2015 20:18:06.
`
`

`

`View Article Online
`
`the pictures from top to bottom the vertical sample flow control
`is shown.
`The horizontal dimensions of the sample flow are controlled
`by two horizontal control ports that are located on the sides of
`the flow-channel, downstream of the sample inlet (see Fig. 1).
`By adding or removing sheath liquid through these control ports
`at an equal rate the already present sheath flow is horizontally
`compressed or expanded respectively which leads to a narrow
`or wider sample flow. In Fig. 2 from left to right this effect is
`shown. Notice that the height of the sample flow is not affected
`by this control mechanism.
`
`Flow-cell 2: additional control of sample flow
`position
`
`Flow-cell 2 looks quite similar to flow-cell 1, but there is one
`main difference. Flow-cell 2 has an additional inlet, located in
`between the focussing section and the sample inlet (see Fig. 3).
`This additional inlet gives flow-cell 2 the added functionality to
`freely position the sample flow anywhere inside the flow-
`channel. Two control mechanisms are required to achieve
`this.
`The vertical position of the sample flow is controlled by the
`additional vertical control inlet. When sheath liquid is added
`through this inlet the entire sheath flow in the channel is lifted
`up from the channel bottom. As a result a coaxial sheath flow is
`formed that no longer has any contact with the channel bottom.
`The more liquid is added through this inlet the higher the sample
`flow is positioned. This vertical position control inlet has a
`narrow shape (625 μm 3 50 μm) to create a flow profile through
`this inlet that is as uniform as possible. This means that the
`shape of the sample flow is hardly influenced by the vertical
`position control, except for some vertical compression. Vertical
`position control of the sample flow is demonstrated in Fig. 4 in
`the pictures from top to bottom.
`The same horizontal control ports that are used to control the
`width of the sample flow can also be used to control its position.
`In this type of operation the direction of flow through both inlets
`is now opposite. By adding sheath liquid through one of the
`control ports and removing it from the other inlet at the same
`flow-rate the sample flow is shifted in the horizontal plane. The
`results can be seen in Fig. 4 in the pictures from left to right.
`Apart from the additional inlet the configuration of the inlets
`of flow-cell 2 is similar to that of flow-cell 1, therefore the
`sample dimension control mechanisms described for that flow-
`cell work for flow-cell 2 as well. The coaxial flow that can be
`achieved with this flow-cell is very suitable for sensors that do
`not require any contact with the sample liquid and especially
`those sensors in which the interface might get polluted such as
`optical sensors.
`
`fore, the main interaction between liquids in a microfluidics
`system is by diffusion,13 which is a relative slow process. The
`laminar flow behaviour and the limited interaction between
`different liquids make microfluidics systems very suitable to be
`used with sheath flow. In the following sections two flow-cells
`using sheath flows are described that have been designed using
`finite element modelling.
`
`Flow-cell 1: control of sample flow dimensions
`
`Flow-cell 1 has been developed to allow dynamic control of the
`dimensions of the sample flow (see Fig. 1). In this flow-cell a
`non-coaxial sheath flow is formed by vertically injecting a
`sample liquid into a channel through which sheath liquid is
`flowing. By hydrodynamic focussing a smooth flow of sample
`liquid is formed that still touches the bottom of the channel. A
`focussing section brings the channel width down from 625 μm
`to a width of 160 μm. The application of such a focussing
`section allows the use of fairly large inlets which are convenient
`in handling the liquid connections to the chip. The less critical
`alignment of the inlets and the lower pressure drop over the
`wider sections are additional advantages. The non-coaxial type
`of sheath flow that is formed with this flow-cell is suitable for
`sensors that require contact with the sample liquid such as
`impedance sensors. The dimensions of the sample flow can
`dynamically be adapted by 2 orthogonal control mechanisms.
`The vertical dimensions of the sample flow are controlled by
`the relative flow-rate at which the sample liquid is injected in
`relation to the flow-rate of the sheath flow. At higher relative
`flow-rates of the sample liquid, it penetrates further into the
`sheath liquid thereby increasing the sample flow height.
`Lowering the relative flow-rate of the sample liquid will result
`in a sample flow with less height. In Fig. 2 simulation results are
`shown, that were obtained using the model shown in Fig. 1. In
`
`Fig. 1 Model of flow-cell 1, this flow-cell creates a non-coaxial sheath
`flow and allows control of the sample flow dimensions; here the control
`ports are used to widen the sample flow.
`
`Published on 13 March 2003. Downloaded by University of California - Los Angeles on 04/10/2015 20:18:06.
`
`Fig. 2 Overview of the control of the sample flow dimensions in a cross-
`section of the channel.
`
`Fig. 3 Model of flow-cell 2, the additional vertical position inlet of this
`flow-cell also allows positioning of the sample flow in the channel; here the
`control ports are used to position the sample flow vertically in the centre of
`the channel with a horizontal shift.
`
`Lab Chip, 2003, 3, 56–61
`ABS Global, Inc. and Genus plc – Ex. 1014, p. 57
`
`57
`
`

`

`View Article Online
`
`Due to the planar nature of the device it is only possible to
`analyse it from the top. Performing measurements on the width
`of the sample flow is straightforward in this configuration, but
`analysing the vertical dimensions of the sample flow is more
`complicated. However by analysing the intensity of the dye,
`also quantitative information about the sample height can be
`obtained. Before each series of experiments photographs were
`taken of the channel without any dye present and photographs
`were taken when the channel was completely filled with the dye.
`After the experiment the height of the sample flow was
`calculated by regression between these two references. To
`obtain a reliable result each measurement point was taken from
`the average of three photographs. The noise was reduced by
`averaging the intensity of the dye over 100 adjacent image lines.
`The area of the photographs that was used in the experiments is
`marked with the dotted white squares in Fig. 7, 9 and 11.
`Unfortunately analysing the vertical position of the sample flow
`from the top is not possible with this method; therefore it was
`decided to base the experiments only on flow-cell 1.
`The finite element simulations were performed with the
`Netflow-module of the finite element package Coventorware.
`Detailed models of up to 100 000 elements with a well designed
`grid were required to obtain a high-enough accuracy. Besides
`the geometry of the model and the flow-rates also the diffusion
`constant of the dye was included in the simulations. From the
`simulation results the height of the sample flow could easily be
`obtained by just summing the concentrations of the sample
`distribution along the vertical axis for all data points in the
`measurement area. Finally the experimental results and the
`simulation results were compared by plotting them in the same
`graph.
`
`Fabrication of the flow-cells
`
`Both types of flow-cells can be fabricated using the same,
`relatively simple process. In a glass wafer by isotropic etching
`the channel is defined with a depth of 100 μm and a minimum
`width of 160 μm. In a silicon wafer by isotropic etching
`through-holes are defined that form the liquid inlets of the
`device. The glass wafer and the silicon wafer are then anodically
`bonded together to form the complete devices. A photograph of
`a flow-cell chip is depicted in Fig. 5.
`There are a number of reasons to make the devices like this.
`Firstly, the dimensions of the sample-inlet need to be accurate
`and etching of silicon can be easier controlled than glass
`etching. Secondly, with the liquid inlets in the silicon part the
`glass side of the device is still available for optical inspection
`during operation. Finally, since the channel was etched in the
`glass wafer the surface of the silicon wafer is still smooth and
`suitable for future integration of sensors to form a complete
`integrated analysis system.
`
`Experimental verification
`
`A series of experiments was carried out to verify the simulation
`results. The measurement setup is shown in Fig. 6. The chip
`with the flow-cell is placed in a custom made aluminium holder
`which is used to make the connections to the tubing. The liquid
`connections to the chip are sealed by O-rings. Two syringe
`pumps (kdScientific model 200 series) are used to control the
`flow-rates at the inlets. The flow-cell is positioned underneath a
`microscope (Zeiss Stemi SV11) and a diluted red dye (standard
`plotter dye) is used as a sample liquid. The sheath liquid is
`purified water. A digital still camera (Sony Cybershot DSC-75)
`captures images of the flow-cell for quantitative analysis and
`comparison with the simulation results.
`
`Published on 13 March 2003. Downloaded by University of California - Los Angeles on 04/10/2015 20:18:06.
`
`Fig. 4 Overview of the position control of the sample flow in a cross-
`section of the channel.
`
`Fig. 6 Overview of the experiment set-up.
`
`Fig. 5 The flow-cell chip with dimensions of 2 cm by 1 5 cm (the
`electrical contacts are not used in the experiments).
`
`Fig. 7 Illustrative results from the diffusion experiment.
`
`58
`
`Lab Chip, 2003, 3, 56–61
`
`ABS Global, Inc. and Genus plc – Ex. 1014, p. 58
`
`

`

`View Article Online
`
`In Fig. 9 it can be seen that the width of the sample is constant
`for relative flow-rates of the sample flow up to 10%, at the
`highest flow-rate the sample flow becomes slightly wider. For
`the height control of the sample flow there is a good match
`between the finite element simulations and the experimental
`results. This means that the model can be used to reliably predict
`the vertical control of the sample flow. The slight mismatch at
`the highest vertical dimension is due to saturation of the camera
`for the red colour.
`
`Horizontal control of sample flow dimensions
`
`In the second control experiment the horizontal dimensions of
`the sample flow were controlled. During this experiment the
`sheath liquid and sample liquid flow-rates were kept constant at
`10 μl min21 and 1 μl min21 respectively. The flow-rate through
`each control port was varied from 23 to +25 μl min21. The dye
`was measured at the same location as in the previous
`experiments; here it was even more important to have a visible
`area that is maximally wide. Some illustrative results are
`depicted in Fig. 11 and the quantitative results are shown in Fig.
`12.
`The results in Fig. 12 show that the width of the sample flow
`can be controlled over a wide range: for the flow-rates used in
`the experiment the width of the sample flow varies from 40 to
`
`Fig. 9 Illustrative results from the sample height control experiment.
`
`Sample diffusion
`
`First a reference experiment was carried out to determine the
`influence of diffusion and to see how well it can be modelled. In
`this experiment the ratios of the flow-rate of the sample liquid,
`the flow-rate through the control ports and the flow-rate of the
`sheath-liquid were kept constant at 1+5+10. During the
`experiment the total flow-rate was varied from 1 to 50 μl min21.
`The intensity of the dye was measured in the wider section
`following the narrow section (see Fig. 7), since in this location
`the visible area is not blocked so much by the rounded corners
`of the isotropic etching of the channel. A detailed finite element
`model similar to the one in Fig. 1 was made, only the model
`used here has a wide section following the narrow section so
`that the quantitative data from the simulations can be obtained
`in the same location as in the experiment. In Fig. 7 some
`illustrative results of the experiments are shown.
`In Fig. 8 the quantitative results of both the simulation and the
`experiment are shown. They show that there is a very good
`agreement between experimental and simulation results. The
`entire shapes of the dye distributions match very well. This
`means that the finite element model represents the real flow-cell
`very well and that this model can be used for the further
`experiments. The influence of diffusion is realistically taken
`into account in the experiments and for a flow-rate higher than
`10 μl min21 the influence of diffusion is small, so measure-
`ments can be carried out at this flow-rate.
`
`Vertical control of sample flow dimensions
`
`In the first control experiment the vertical dimensions of the
`sample flow were controlled. Since for this measurement the
`width control of the sample was not necessary, a device without
`horizontal control ports was used similar to flow-cell 1. During
`the experiment the flow-rate of the sheath liquid was kept
`constant at 10 μl min21 and six different flow-rates for the
`sample liquid were applied in a range from 0.05 to 2 μl min21,
`which correspond to relative sample flow-rates of 0.5% to 20%
`of the sheath flow-rate. Again the intensity of the dye was
`measured in the wide section of the channel, downstream of the
`narrow section. Some illustrative results are depicted in Fig. 9
`and the quantitative results are shown in Fig. 10.
`The results in Fig. 10 show that the height of the sample can
`be controlled over a wide range. For the relative flow-rates used
`in the experiment the height of the sample flow can be
`controlled from 3 to 60 μm in a channel with a depth of 100 μm.
`
`Published on 13 March 2003. Downloaded by University of California - Los Angeles on 04/10/2015 20:18:06.
`
`Fig. 8 Quantitative comparison of the simulation (blue) and the experimental (red) results for the diffusion experiment, for sheath flow-rates of 1, 2.5, 5,
`10, 25 and 50 μl min21.
`
`Lab Chip, 2003, 3, 56–61
`ABS Global, Inc. and Genus plc – Ex. 1014, p. 59
`
`59
`
`

`

`View Article Online
`
`292 μm in the wide part of the channel, which has a width of 625
`μm. As a result the width of the sample flow in the narrow
`section where the sensor should be positioned (width 160 μm)
`is varied in the range from 10 to 73 μm. The width of the sample
`flow is defined here as the width of the sample at 50% of the
`maximum dye intensity. So also for the horizontal control of the
`sample flow dimension the results show a good match between
`simulation and experimental results.
`
`Sample flow splitting and switching
`
`The flow-cell used in the experiment turned out to be a very
`flexible device. Besides the horizontal and vertical control of
`the sample also other useful functions can be realised. The first
`example is sample splitting. For this application only one
`control port is used. Through this inlet 50% of the total flow of
`liquid is removed. This results in an equal splitting of the sample
`(see Fig. 13). By removing more or less liquid through the
`control port an unequal splitting of the sample is also possible of
`course.
`A second example of the flexibility of the flow-cell is sample
`switching. Here again only one control port needs to be active.
`
`This time the liquid is removed through the control port at a
`flow-rate that equals the flow-rate of the total flow of liquid. As
`a result the complete sheath flow including the sample liquid is
`switched (see Fig. 14). When a coaxial sheath flow is used such
`as can be generated with flow-cell 2 this is even possible
`without the sample liquid touching any wall of the channel.
`
`Conclusions
`
`Two new flow-cells have been developed that are very versatile.
`The flow-cells can generate both non-coaxial and coaxial sheath
`flows. Using an orthogonal control mechanism the flow-cells
`allow dynamic control of the sample flow dimensions. With one
`additional vertical position inlet the sample flow can also be
`freely positioned inside the channel. The processing of the flow-
`cells is very simple, and can be easily combined with the
`integration of sensors to form a complete integrated analysis
`system.
`Experiments were carried out with one of the flow-cells and
`a comparison of experimental results and simulation results
`shows that the complicated flow behaviour of the device can be
`modelled very accurately. Also the influence of diffusion is
`
`Fig. 10 Comparison of the simulation (Finite Element Modelling, FEM)
`results and the experimental (Dye) results for the sample height control
`experiment.
`
`Fig. 12 Comparison of the simulation (Finite Element Modelling, FEM)
`results and the experimental (Dye) results for the sample width control
`experiment.
`
`Published on 13 March 2003. Downloaded by University of California - Los Angeles on 04/10/2015 20:18:06.
`
`Fig. 11 Illustrative results from the sample width control experiment.
`
`60
`
`Lab Chip, 2003, 3, 56–61
`
`ABS Global, Inc. and Genus plc – Ex. 1014, p. 60
`
`

`

`View Article Online
`
`The authors also would like to thank Dr. A. Bossche and Mr. P.
`Turmezei of the Electronic Instrumentation Laboratory, Delft
`University of Technology for fruitful cooperation and useful
`discussions. Finally, the authors thank the Hochschuljubiläums-
`stiftung der Stadt Wien for financial support.
`
`References
`
`1 M. Tidare, H. Roos and U. Lindberg, Hydrodynamic Addressing; A
`Flow Cell for Improved Signal Quality in Bioanalytical Systems, , in
`Proc. of Transducers ’01, Munich, Germany, June 10–14, 2001, pp.
`1154–1157.
`2 O. Hofmann, P. Niedermann and A. Manz, Modular Approach to
`Fabrication of Three-Dimensional Microchannel Systems in PDMS –
`Application to Sheath Flow Microchips, Lab Chip, 2001, 1(2),
`108–114.
`3 U. D. Larsen, G. Blankenstein and J. Branebjerg, Microchip Coulter
`particle counter, , in Proc. of Transducers ’97, Chicago, June 16–19,
`1997, pp. 1319–1322.
`4 K. Roberts, M. Parameswaran, M. Moore and R. S. Muller, A Silicon
`Microfabricated Aperture for Counting Cells using the Aperture
`Impedance Technique, , Proc. of 1999 IEEE Canadian Conference on
`Electrical and Computer Engineering, Edmonton, Alberta, Canada,
`May 9–12, 1999, pp. 1668–1673.
`5 G. Lee, B. Hwei and G. Huang, Microfluidic Chips with MxN
`Continuous Sample Introduction Function Using Hydrodynamic
`Flow Switching, , in Proc. of Transducers ’01, Munich, Germany,
`June 10–14, 2001, pp. 1158–1161.
`6 R. Miyake, H. Ohki, I. Yamazaki and R. Yabe, A Development of
`Micro Sheath Flow Chamber, , in Proc. of MEMS ’91, 1991, pp.
`265–270.
`7 K. Tashiro, T. Sekiguchi, S. Shoji, T. Funatsu, W. Masumoto and H.
`Sato, Design and Simulation of Particles and Biomolecules Handling
`Micro Flow Cells with Three Dimensional Sheath Flow, , in Proc. of
`μTAS 2000, Enschede, The Netherlands, May 14–18, 2000, pp.
`209–212.
`8 A. Wolff, U. D. Larsen, P. Friis and P. Telleman, Chip-Integrated
`Microfluidic System for Cell Sorting and Cell Culturing, , in Proc. of
`Eurosensors XIV, Copenhagen, Denmark, August 27–30, 2000, pp.
`235–238.
`9 D. Huh, Y. Tung, H. Wei, J. B. Grotberg, S. J. Skerlos, K.
`Kurabayashi and S. Takayama, Use of Air-Liquid Two-Phase Flow in
`Hydrophobic Microfluidic Channels for Disposable Flow Cyt-
`ometers, Biomed. Microdev., 2002, 4(2), 121–149.
`10 J. H. Nieuwenhuis, J. Bastemeijer, P. M. Sarro and M. J. Vellekoop,
`Integrated Coulter counter with non-coaxial sheath-flow and dy-
`namic aperture control, , in Proc. of Eurosensors XVI, Prague, Czech
`Republic, September 15–18, 2002, pp. 1194–1197.
`11 J. H. Nieuwenhuis, J. Bastemeijer, A. Bossche and M. J. Vellekoop,
`Dynamic Particle-Shape Measurements Using a Near-Field Optical
`Sensor, , in Proc. of 1st IEEE Sensors Conference, June 12–14, 2002,
`Orlando, Florida, USA.
`12 J. V. Watson, Introduction to flow cytometry, 1st edn., Cambridge
`University Press, Cambridge 1991, pp. 2–4.
`13 M. Koch, A. Evans and A. Brunnschweiler, Microfluidic Technology
`and Applications, 1st edn., Research Studies Press Ltd., Hertfordshire,
`2000, pp. 227–228.
`
`Fig. 13 The flow-cell applied for sample splitting.
`
`Fig. 14 The flow-cell applied for flow-switching.
`
`realistically taken into account. The results on the horizontal
`and vertical control of the sample flow dimensions show that
`both mechanisms work well and that they can be predicted from
`simulation results. The control ports make the flow-cells very
`versatile so that they can also be used for other applications such
`as sample splitting and flow switching.
`The flow-cells can be used in a broad range of applications
`such as an integrated Coulter counter,10 sensors for particle
`shape analysis11 and sensors for cell analysis and (bio)chemical
`analysis.
`
`Acknowledgement
`
`The authors would like to thank the DIMES Technology Centre
`of the Delft University of Technology for the fabrication of the
`device, especially Mr. C. R. de Boer and Mr. W. van der Vlist.
`
`Published on 13 March 2003. Downloaded by University of California - Los Angeles on 04/10/2015 20:18:06.
`
`Lab Chip, 2003, 3, 56–61
`ABS Global, Inc. and Genus plc – Ex. 1014, p. 61
`
`61
`
`