SIGNAL IMPROVEMENT BY DIELECTRIC FOCUSING IN

MICROFLUIDIC IMPEDANCE CYTOMETERS M. Evander

*, B. Dura, A.J. Ricco, G.T.A. Kovacs and L. Giovangrandi

Department of Electrical Engineering, Stanford University, USA

ABSTRACT We report the design and the electrical and microfluidic characterization of a device that improves signal strength and

increases the detection levels in microfluidic impedance cytometry. Low-conductivity aqueous and two-phase dielectric

sheathing allows the use of wider, more easily fabricated channels, increases the precision of cell- and particle-size

measurements, and decreases the smallest detectable object size. For aqueous dielectric sheathing with a fixed excitation

voltage of 400 mVrms, the highest relative signal strength measured was 1.4 ± 0.6% for a 27-µm core. The use of a two-

phase system and a core width of 33 µm resulted in a signal strength of 5.1 ± 0.5%. This two-phase system increases

SNR by 23 dB compared to the aqueous system.

KEYWORDS: Impedance cytometry, Microfluidics, Dielectric focusing, Signal improvement

INTRODUCTION Impedance cytometers are increasingly used within the biomedical community for counting and measuring the sizes

and/or dielectric properties of particles and cells, and single and populational cell analysis is increasingly important for

applications from drug discovery to diagnostics [1]. Microfluidic technologies have enabled the use of miniaturized

Coulter counters and micro-impedance cytometers for label-free cell analysis [2, 3]. The use of microfluidics can

improve performance relative to traditional cytometers by implementing a low-conductivity aqueous dielectric sheath [4-

6], but a systematic approach to the design of a device utilizing this technique is lacking. Using two-phase flow can

further improve performance by eliminating diffusion, a limiting factor in all-aqueous systems. It will also thus widen

the range of flow rates that can be used and make smaller core widths possible. Use of oil as sheath fluid was recently

demonstrated, but implications for detection were not reported [7].

EXPERIMENTAL

Our microfluidic impedance cytometer, Figures 1 and 2, uses a fluidic layer fabricated from 50-µm-thick, double-

sided adhesive film (Adhesives Research, Glen Rock, PA) using a CO2 laser engraving and cutting system (Universal

Laser Systems, Scottsdale, AZ). Two sets of 20-µm-wide platinum electrodes separated by 100 µm, patterned on

Borofloat® glass using standard photolithography and sputtering, were used for differential measurements. The fluidic

layer was sandwiched between two glass chips and bonded using a hot plate at 70° C and a hand roller. Fluidic

connections were fabricated from polydimethylsiloxane (PDMS) and bonded to the top glass chip with the aid of air

plasma.

Figure 1: (A) Illustration of the components of the microfluidic impedance cytometer. A laser-cut, 50-µm-thick,

polyester-supported adhesive layer is sandwiched between two glass chips with 20-µm-wide Pt microelectrodes

separated by 100 µm. (B) Photograph of the actual device in an acrylic fixture providing fluidic and electric connections

and a U.S. quarter dollar for scale. The pogo® pins used for electric connections are not attached to the fixture in this

image.

978-0-9798064-3-8/µTAS 2010/$20©2010 CBMS 1277 14th International Conference onMiniaturized Systems for Chemistry and Life Sciences

3 - 7 October 2010, Groningen, The Netherlands

Figure 2: The fluidics layout of the microfluidic impedance cytometer allows for hydrodynamic focusing of the sample,

dielectrophoretic focusing/centering (5 MHz, 1.6 Vrms), and impedance detection of single cells/particles.

The effects of dielectric focusing were characterized using both aqueous and two-phase systems. Deionized water

(D.I.), εr ≈ 80, was used as the aqueous dielectric sheath whereas heavy mineral oil (EMD Chemicals Inc., Darmstadt,

Germany), εr ≈ 2.3, was used for two-phase experiments. For both, a conductive core of phosphate-buffered saline (PBS)

with 10-µm polystyrene particles (61946, Fluka Analytical) was hydrodynamically focused at various flow ratios to

achieve and characterize varying core widths. Gravity flow was used for all aqueous experiments due to the high

sensitivity of the system to small pressure fluctuations created by conventional syringe pumps. For the two-phase

experiments, a syringe pump (EW-74900-20, Cole-Parmer, Vernon Hills, IL) was used due to the higher pressures

needed to drive these flows. The particle velocities were kept at 6.3 mm/s for the aqueous experiments and at 43 mm/s

for the two-phase experiments.

As varying particle positions within the channel are a common cause of signal variation in impedance cytometry [8],

negative dielectrophoresis was applied to center the particles in the channel, homogenizing their velocities. A signal of

5 MHz and 1.6 Vrms was applied to focusing electrodes in the hydrodynamic focusing zone, producing a negative

dielectrophoretic force centering the particles.

Figure 3 shows a block diagram of the detection system. A lock-in amplifier (SR830 DSP, Stanford Research

Systems, Sunnyvale, CA) was used to detect particles passing the microelectrodes. A frequency of 65 kHz and a fixed

excitation voltage of 400 mVrms (below the water hydrolysis window) was used for all measurements. The in-phase

component was high-pass filtered with a 1 Hz cut-off filter (SR 650, Stanford Research Systems, Sunnyvale, CA) to

remove any DC offset. The filtered signal was then acquired though a PCI-6024E DAQ-card (National Instruments,

Austin, TX) and analyzed using MATLAB® (Mathworks, Natick, MA). As the base current, Ibase, passing through the

channel is different for each core width, the signal strength (relative change in current for a particle passing the

electrodes, ∆I/Ibase), as well as the signal-to-noise ratio (SNR, ∆I/Inoise), were calculated for each core width. Core widths

were measured from channel micrographs.

Figure 3: Block diagram of the detection instrumentation used for the microfluidic impedance cytometer.

RESULTS AND DISCUSSION

Particles were analyzed for core widths ranging from 145 µm to 27 µm in the aqueous system and for a single core

width of 33 µm for the two-phase system. Results (Figure 4) show the relative detected signal from the passage of a

Lock-in

amplifier1 V / µA

Filters DAQ

ParticleFluid flow

+ -

Drive signal

Shunt resistor Shunt resistor

Dielectrophoretic

focusing

Detection electrodes

Detection electrodes

Differential

AmplifierGain: 50.4

15 kΩ 15 kΩ

1278

single particle increasing with decreasing core width for aqueous sheaths, culminating at 1.4 ± 0.6% for the smallest

(27 µm) core. For the two-phase flow, a signal of 5.1 ± 0.4% was measured for a 33 µm core. The SNR shows no

significant increase with decreasing core size for the aqueous setup; however for the two-phase flow an improvement of

23 dB is measured.

Figure 4: (A) Average detection signal, ∆I/Ibase, for the passage of single 10-µm polystyrene beads in PBS with a range of

core sizes (all aqueous) and for a two-phase flow (oil sheath, 33 µm core). For a fixed excitation voltage of 400 mVrms, a

detection signal of 1.4 ± 0.6 % is measured for a 27-µm aqueous core, whereas the two-phase system has a signal of 5.1

± 0.5 %. (B) The signal-to-noise ratio for the various core sizes, showing no significant dependence on core size for the

aqueous cores. For the two-phase system, a 23 dB improvement compared to the smallest aqueous core is observed.

Error bars correspond to ± 1 standard deviation for averages of 21 measurements.

CONCLUSION There are many advantages of using dielectric sheathing in a microfluidic impedance spectrometer. Hydrodynamic

focusing enables the use of larger channels that are simpler to fabricate, less prone to clogging, and need lower pressure

to drive the fluid. Utilizing a dielectric sheath flow improves both the detected signal strength and the SNR for single-

particle measurements without increased excitation voltages and currents, enabling the analysis of particles and cells with

smaller sizes or lower dielectric contrast. Using a two-phase system that is immune to diffusion effects at the interface

between the core and sheath also results in decreased signal variability, which will translate to higher precision when

performing size measurements and comparing the signal response to that of a known particle size. For the low frequency

used in this study, the detected signal is dependent only on the size of the object and the microparticles used are expected

to result in very similar signals to cells [9]. It’s anticipated that the use of higher frequencies will further improve the

signal which should improve the size resolution and make this approach more suitable for cells.

ACKNOWLEDGEMENTS This work was supported by the National Institutes of Health through grant 1R21-EB007390-01A1.

REFERENCES [1] O. E. Beske and S. Goldbard, High-throughput cell analysis using multiplexed array technologies, Drug Discov.

Today, Vol 7, pp. S131-S135, 2002.

[2] K. Cheung, S. Gawad, and P. Renaud, Impedance spectroscopy flow cytometry: On-chip label-free cell

differentiation, Cytom. Part A, vol. 65A, pp. 124-132, 2005.

[3] D. Holmes, D. Pettigrew, C.H. Reccius, J.D. Gwyer, C.v. Berkel, J. Holloway, D.E. Davies, and H. Morgan,

Leukocyte analysis and differentiation using high speed microfluidic single cell impedance cytometry, Lab Chip,

vol. 9, pp. 2881-2889, 2009.

[4] U.D. Larsen, G. Blankenstein, and J. Branebjerg, Microchip Coulter particle counter,” in Proc. 1997 International

conference on solid-state sensors and actuators, 1997, pp. 1319-1322.

[5] S. Chung, S.J. Park, J.K. Kim, C. Chung, D.C. Han, and J. K. Chang, Plastic microchip flow cytometer based on 2-

and 3-dimensional hydrodynamic flow focusing, Microsyst. Technol., vol. 9, pp. 525-533, 2003.

[6] J.H. Niewenhuis, F. Kohl, J. Basemeijer, P.M. Sarro, and M.J. Vellekoop, Integrated Coulter counter based on 2-

dimensional liquid aperture control, Sens. Actuator B-Chem, vol. 102, pp. 44-50, 2004.

[7] C. Bernabini, D. Holmes, and H. Morgan, Micro-impedance spectroscopy for detection and discrimination of

bacteria and micro-particles, in Proc. 13th international conference on miniaturized systems for chemistry and life

sciences, 2009, pp. 1641-1643.

[8] L. Spielman and S. L. Goren, Improving resolution in a Coulter Counting by hydrodynamic focusing, J. Colloid

Interface Sci., vol 26, pp. 175-182, 1968

[9] D. Holmes and H. Morgan, Single Cell Impedance Cytometry for Identification and Counting of CD4 T-cells in

Human Blood Using Impedance Labels, Anal. Chem., vol 82, pp. 1455-1461, 2010.

CONTACT

M. Evander, Tel: +1-650-723 5646; evander@stanford.edu

1279