LAB-ON-A-CHIP WITH FLUID ACOUSTIC MICROAGITATION

Piezoelectric Polymer β-PVDF used as Ultrassonic Transducer

V. F. Cardoso, J. G. Rocha, F. O. Soares, G. Minas

Dept. Industrial Electronics,University of Minho, Campus de Azurem, 4800-058 Guimaraes, Portugal

S. Lanceros-Mendez

Dept. Physics,University of Minho, Campus de Gualtar, 4710-057 Braga, Portugal

Keywords: Lab-on-a-chip, PVDF, acoustic microagitation, SU-8.

Abstract: The main objective of this article is to describe the development of a fully-integrated disposable lab-on-a-

chip for point of care testing and monitoring of biochemical parameters in biological fluids. The lab-on-a-

chip is composed mainly by two dies: the fluid and the detection. The fluid die, fabricated in SU-8,

comprises three microfluidic cuvettes, containing the fluids into analysis, and a β-PVDF microagitation

system. The detection die is fabricated in a CMOS standard process and contains the photodetectors and the

electronics for signal actuation and detection. The main innovation of this lab-on-a-chip is the application of

an acoustic microagitation technique by the deposition of the β-PVDF piezoelectric polymer underneath the

microfluidic structures, with automatic electronic control. This piezoelectric polymer produces mechanical

vibrations, which allow the enhancement of the mixing and the reaction.

1 INTRODUCTION

Microfluidic technology has become an important

tool for analytical biochemistry applications. It

enables the fabrication of precise and small

structures in glass, quartz, silicon or polymeric

materials: the lab-on-a-chip concept. The great

interest in that technology stems from the inherent

performance gains: reduced sample size, higher

degree of integration and thus enhanced potential for

automation and control fluids of submicroliter

volumes, shortened response time, potential for

improved analytical performance, reduced storage of

chemicals and hence laboratory safety, and reduced

costs (Kopp et. al, 1997).

There is a large demand in the healthcare system

to develop lab-on-a-chips for rapid and reliable point

of care (POC) testing and monitoring. Such lab-on-

a-chips would significantly enhance the quality of a

diagnostic by offering immediate measurement of

several clinically relevant parameters that can be

used to assess the health of the patient. The

biochemical analysis of the patient’s biological

fluids is a good start. Most diseases leave a

molecular fingerprint in those fluids and by

measuring that fingerprint in the right way, the

precision of the diagnostic can be improved

(Connolly, 1995). However, the physician or the

patient has no routine in performing the advanced

biochemical analysis on-chip. Therefore, to develop

point of care disposable lab-on-a-chips, avoiding

cross-contamination of samples and measurements

errors, it is necessary to have a reliable and highly

automated microfluidic control system. This system

should be fully integrated with the control and the

detection electronics implemented on a low-cost

substrate and performed by a low-cost fabrication

process. The use of MEMS (Micro Electro

Mechanical Systems) based devices, such as

microvalves and micropumps, increases the cost of

the system, needs complex control systems and are

difficult to integrate (Reyes et. al, 2002). Mixing

only by diffusion avoids these drawbacks. However,

long transit times and consequently long

microchannels are necessary, if large molecules with

small diffusivities must react. This illustrates how

dramatic the diffusion limitation is.

To overcome the long transit times due to

diffusion and to enable high-efficient reactions it is

262

F. Cardoso V., G. Rocha J., O. Soares F., Minas G. and Lanceros-Mendez S. (2008).

LAB-ON-A-CHIP WITH FLUID ACOUSTIC MICROAGITATION - Piezoelectric Polymer β-PVDF used as Ultrassonic Transducer.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 262-267

DOI: 10.5220/0001050202620267

 SciTePress

necessary to induce the microfluidic die by a

mechanism that accelerates the mixing and the

reaction, preferably with no moving parts. In this

paper, such mechanism is presented.

2 BACKGROUND

The need for rapid and in-situ measurements with

low sample volumes has led to the development of

miniaturized analyses devices with the fluidic,

detection and readout systems integrated in a single

chip. The vision of those devices was presented in

1990 (Manz et. al., 1990) and has been since then

the inspiration for an intense research effort pursuing

miniaturization of macroscopic biochemical analy-

ses methods. Within the field of clinical diagnostics,

several lab-on-a-chip approaches have been

published (Auroux et. al, 2002, Reyes et. al., 2002).

Our group has developed a lab-on-a-chip for

measuring the concentration of some biomolecules

in urine samples by optical absorption technique

(Minas, et. al., 2005). It comprises three parts in a

multi-chip-module: (1) a microfluidic system die

containing the microchannels fabricated using SU-8

techniques (Ribeiro et. al., 2005); (2) an optical

filtering system based on highly selective Fabry-

Perot optical resonators using a stack of CMOS

process compatible thin-film layers (Minas et. al.,

2006); (3) a detection and readout system fabricated

in a CMOS microelectronic process (Minas et. al.,

2004). The optical filtering system enables the

measurement using white light illumination, thus

avoiding the use of a wavelength dependent light

source. This characteristic makes the lab-on-a-chip

portable and ensures that the analysis can be

performed at any location with instantaneous results.

However, in that lab-on-a-chip, mixing the samples

with the reagents was performed by diffusion, which

leads to long transit times, especially when large

molecules with small diffusion coefficients must be

analysed. Therefore, to be valuable for point of care

testing and monitoring, the microfluidic die of the

lab-on-a-chip requires a microagitation mechanism.

It is desirable that this mechanism does not require

any external apparatus, internal moving parts or

valves.

The flows used in microfluidic systems are very

small, of the order of 100 μm, with velocities of the

order of 1 mm/s, which leads to Reynolds number

less than 1. For this small Reynolds numbers the

flow is dominated by the viscous effects and

turbulent mixing is impossible. Therefore, mixing in

a straight channel rely on diffusion, but since

diffusion coefficients of some biological species

(e.g. enzymes and other proteins) are very small,

mixing may take several minutes or even hours. This

may be undesirable for some applications and

alternative mixing methods must be developed.

Acoustic waves are an interesting solution for

this problem. They have been used both to promote

mixing (Bengtsson et. al., 2004) and to pump fluids

(Rife et. al., 2000). Sound waves that propagate in

the fluid generate pressure differences that induce

the so called acoustic streaming. Secondary flows

can be created in the plane perpendicular to the main

flow in the channel. These secondary flows promote

mixing by convection between otherwise parallel

currents. Therefore, acoustic microagitation could be

a simple actuation source for mixing and promoting

fluids reaction.

In order to produce the ultrasounds, the channels

and the reaction chambers of the microfluidic die

must have an ultrasound transducer, which can be

fabricated from a piezoelectric polymer, such as the

PVDF - Poly(Vinylidene Fluoride) in its beta-phase.

PVDF is a polymer with interesting piezoelectric

properties, which allows important electro-optical,

electromechanical and biomedical applications. This

polymer shows at least four crystalline phases. The

one with the best piezo- and pyroelectric properties,

after poling, is the beta-phase (Sessler, 1987). Until

recently, this phase was exclusively obtained by

mechanical stretching of films originally in the non-

polar alfa-phase, which is the most stable one from a

thermodynamic point of view and directly obtained

from the melt. This process results in films mostly in

the beta-phase, but with a small percentage of alfa-

phase material (Sencadas et. al., 2004). Unoriented

films exclusively in the beta-phase were obtained

from the crystallization of PVDF from solution with

N,N-Dimethyl Formamide or Dimethyl Acetamide

at temperatures below 70ºC. The electromechanical

properties of this film were improved by a patented

process by our group (Lanceros-Mendez et. al.,

2006).

3 DEVICE DESCRIPTION

This paper describes the concept, operation and

implementation of a portable, low-cost, plastic

based, disposable and highly automated lab-on-a-

chip for biochemical analyses of biological fluids. It

uses optical absorption spectrophotometry as

measurement analytical technique and acoustic

microagitation for mixing the fluids. Its application

is for point-of-care systems.

LAB-ON-A-CHIP WITH FLUID ACOUSTIC MICROAGITATION - Piezoelectric Polymer ß-PVDF used as Ultrassonic

Transducer

263

3.1 Measurement Analytical Technique

The spectrophotometric analysis, the study of the

interaction of electromagnetic radiation with

(bio)chemical compounds, is a very convenient and

often used analytical technique in clinical

laboratories for routine tests analyses, especially the

ones based on colorimetric detection (Sigma, 2006).

However, many of the analytes of interest for

clinical analysis do not have chromophores that

absorb light in a useful part of the visible range.

Specific chemical reagents are available to transform

these analytes into colored products that do have

adequate absorbance. Therefore, in that detection

method, the sample (ex: urine) is mixed with a

proper reagent for the biochemical parameter in

analysis (ex: uric acid reagent) and from that

mixture, a visible colour is produced. The intensity

of that colour is proportional to the biochemical

parameter concentration and can be quantified by

measuring the optical absorption of the mixture,

once the mixture has an absorption maximum at a

specific wavelength. As an example, for determining

the uric acid concentration in a urine sample, the

mixture has an absorption maximum at 495 nm

(Figure 1). For an accurate measurement a complete

and homogeneous mixing must be achieved.

495

Uric

acid

Rise of

biochemical

parameter

concentration

Magnesium

Wavelength (nm)

Absobance (arb. units)

Calcium

550

650

600

Albumin

Figure 1: Absorption spectra for some biochemical

parameters with different concentrations in urine.

3.2 Lab-on-a-chip Operation

The lab-on-a-chip is composed by two dies: the

fluidic die and the detection die (Figure 2). For the

measurement of the concentration of the

biochemical parameters, a light beam is directed to

the three fluidic cuvettes. The intensity of the light

beam transmitted through the mixture is measured

using the three photodetectors, placed underneath

the microfluidic cuvettes, forming three optical

channels. The photodetectors signal is converted, by

the readout electronics, into a digital signal that

allows simpler computer interfacing. The accurate

sample and reagent mixing is performed by the

control electronics that actuates an electroactive

polymer (β-PVDF), deposited underneath the

microfluidic cuvettes. This polymer vibrates and

produces the proper acoustic microagitation of the

mixture, e. g., the produced waves are coupled to the

slide and propagate into the microfluidic cuvette.

Chip

Microfluidic cuvettes

β−

PVDF

Photodetectors

Electronics

Figure 2: Schematic drawing of the lab-on-a-chip

structure.

3.3 Microfluidic Die

The microfluidic die includes the microfluidic

cuvettes containing the fluids to be analysed

(Figure 2). Three cuvettes are needed for each

analysis: one for the chemical reagent, in order to

obtain the baseline reference; other for the mixture

of the sample plus the reagent, to perform the

analysis of the coloured mixed solution; and a third

one with a standard sample with a well-known

concentration of the biochemical parameter that is

being analyzed, for the calibration of the

biochemical parameter concentration.

Under the microfluidic cuvettes a piezoelectric

polymer, β-PVDF, is deposited (Figure 2). This

material will be responsible for the acoustic

microagitation of the fluid. β-PVDF is a transparent

piezoelectric polymer that will produce the

necessary vibration, when an electrical alternating

voltage is applied to its contacts. These vibrations

result in the acoustic microagitation necessary to

mix the sample with the reagent inside the cuvettes.

Moreover, it accelerates the required time for the

complete and homogeneous mixing, improving the

efficiency of the reaction.

The microfluidic cuvettes are fabricated using a

photoplastic material: the photoresist SU-8. This

photoresist is an epoxy-based material that offers

relevant properties, such as high mechanical

strength, good adhesion on many different substrate

materials and biocompatibility. The SU-8 based

fabrication is a low-cost process, UV (from 350 nm

to 400 nm) lithography semiconductor compatible

and does not require expensive masks. It can be

processed by using a spincoating and an UV

maskaligner. In addition, the microfluidic system

can be a disposable die, which minimize the costs

associated with cleaning of the microfluidic cuvettes

and avoids the contamination between analyses.

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

264

Moreover, SU-8 based processing enables the

fabrication of deep structures with very low sidewall

roughness which is suitable for optical absorption

measurements (IBM, 1989).

The negative mask to be used for patterning the

structure of the microfluidic cuvettes is fabricated

from a regular transparency foil, such as the ones

used in printed circuit boards. The chosen SU-8

photoresist is the SU-8-2150, which has a high

viscosity and is the most appropriate for the required

depth (≥ 500 μm). The following paragraphs

describe the microfluidic cuvettes processing steps,

including the SU-8 and the β-PVDF depositions.

After cleaning and drying the glass substrate, a

thin-film of Itrium-Tin-Oxide (ITO) of approxima-

tely 40 nm is deposited by PVD, in order to form the

bottom contact of the piezoelectric β-PVDF. Then, a

600 nm β-PVDF layer is spun of on top of the ITO.

The coated β-PVDF is activated by applying an

electrical field of tens of megavolts by meter, which

is known as the electrical poling of the polymeric

material. After that, the top electrical contact of the

β-PVDF is deposited using the same procedure used

for the bottom contact. These steps complete the

fabrication of the microagitation mechanism.

Next, the fabrication of the microfluidic cuvettes

starts with the spin of 1.25 ml/cm

of SU-8 on top of

the β-PVDF top ITO contact. This process requires

two steps: first, ramping up the spin rotation to

500 rpm during 5 s and keeping that velocity for

10 s; second, ramping up again the spin rotation up

to 1000 rpm during 3 s and keeping it for 30 s. Then,

a soft bake process is needed for evaporation of the

solvent and for the hardening of the SU-8. The soft

bake is performed at 65 ºC during 420 s and at 95 ºC

during 3600 s. The initial temperature allows a

slower evaporation, which results in a better coating

and, mainly and very important, without grooves.

Then, the cuvettes are exposed to UV light, using a

negative mask, for patterning the structure. Figure 3a

shows a cross-section of the device after the

aforementioned steps.

An excessive dose of light with wavelength

lower than 350 nm results in an excessive absorption

at the surface of the film. The effect can be a

pyramidal shape of the microfluidic cuvettes. To

avoid this effect, a commercially available glass

optical filter was used to cut of the wavelengths

lower than 350 nm. The used exposition energy was

700 mJ/cm

. The post-exposure bake is performed at

65 ºC during 60 s, followed by 900 s treatment at

95 ºC. The fabrication ends with the development,

during 2400 s, with the SU-8 developer: an agitation

bath of 1 methoxy-2-propanol acetate. In this way,

the unexposed resist is dissolved and the

microfluidic cuvettes are formed. The structure is

then cleaned with isopropanol and deionised water

and dried with nitrogen. Figure 3b shows a cross

section of the patterned structure. The structure of

the microfluidic cuvettes is shown in Figure 4. Each

cuvette is 1 mm wide, long and height, having a 1 μl

sample volume.

(a)

Glass substrate

ITO

-PVDF

SU-8 photoresist

Mask

-PVDF top

contact

-PVDF bottom

contact

(b)

Glass substrate

ITO

-PVDF

SU-8

Cuvettes

-PVDF top

contact

-PVDF bottom

contact

SU-8

Patterned SU-8

after development

Figure 3: Fabrication sequence of the microfluidic

cuvettes of the prototype: (a) deposition, spin coating, soft

bake, UV exposure; (b) development of the SU-8.

Figure 4: Picture of the microfluidic cuvettes.

3.4 Detection Die

The detection die includes the detectors and the

electronics for signal actuation and detection, all

fabricated in CMOS technology. Specifically, it

comprises the photodetectors, its readout electronics

and the electronics that control the voltages and

frequencies applied to the β-PVDF. Figure 5 shows a

block diagram of the photodiode and its readout

electronics.

The readout electronics consists basically in a

current amplifier and in a sigma-delta analog to

digital converter. The circuit is repeated for each

microfluidic cuvette. An additional circuit for the

photodiode dark current compensation is also

implemented.

LAB-ON-A-CHIP WITH FLUID ACOUSTIC MICROAGITATION - Piezoelectric Polymer ß-PVDF used as Ultrassonic

Transducer

265

After the light reaches the photodiodes, the four

analog to digital converters start the conversion and

their output signal is placed in four separated lines.

Further computer processing perform additional

calculations of these four output signals to achieve

the concentrations of the biochemical parameters

into analysis. The oversample frequency of the

sigma-delta converters is determined by the desired

number of output bits (signal to noise ratio). For this

particular application, an oversampling ratio of 64

and a first order one-bit sigma-delta analog to digital

converter is adequate since the input signal has no

time variations, allowing a high oversample ratio

without the need for a high clock frequency.

Vdd

Reset Clk

Out

Amplifier

Current

Photodiode

Figure 5: Block diagram of the readout electronics for

each microfluidic cuvette.

For these sigma-delta converters, the gain of the

integrator is very high, so, it is finite and larger than

the oversampling ratio. At these conditions, the

noise in the signal band increases only 0.15 dB

(Candy et. al., 1992). Finally, the integrator must be

initialized with a known voltage level at the

beginning of each conversion. This initialization

allows an improvement in the sigma-delta modulator

of 3 dB in the signal to noise ratio (Netravali, 1977).

It was used a digital filter with constant weights

(simple average) and the Pearson coefficient

obtained for the response curve is larger than 0.99,

which demonstrates a good linearity of the device.

This result can even be improved by increasing the

oversampling ratio and by using an optimized digital

filter. Figure 6 shows a picture of the detection die.

Figure 6: Picture of the detection die.

4 EXPERIMENTAL RESULTS

The evaluation of the mixing process was carried out

using the Sigma Diagnostic kit (Infinity

Uric Acid

Reagent) and standards of urine with 30 mg/dl of

uric acid concentration (Sigma, 2006). The reagent

reacts with a sample of urine containing uric acid in

a 50:1 ratio, and produces a pink colour with an

absorption maximum at 495 nm. Its manual

procedure states that a gentle inversion of the

cuvette that contains the mixture is enough for the

complete and homogeneous mixing. Performing this

manual agitation, and after approximately 5 minutes

at room temperature, the mixing is complete.

Without any agitation and due to the high diffusion

coefficients of uric acid in the Infinity

reagent, the

complete mixing is achieved but it takes

approximately 15 minutes at room temperature. In

clinical laboratories, the macroscopic equipments

have mechanical agitation of the cuvettes for

improving mixing and reducing the reaction time.

The experimental results for evaluating the

mixing process are presented in Figure 7. The

microagitation was performed by powering the

β-PVDF though the electrical contacts, with a

sinusoidal signal of 5 V amplitude at the frequencies

shown in the Figure 7. The system was calibrated for

an absorbance of 0 a. u. filling the cuvette with

deionised water, and for an absorbance of 1 a. u.

filling the cuvette with the perfectly mixed solution

(perfectly mixing was guaranteed by external

agitation of the mixture using a slow-rotation

vortex).

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 100 200 300 400 500 600 700 800 900

Time (s)

Absorbance (a.u.)

Without oscilation

Oscilation - 100 Hz

Oscilation - 300 Hz

Oscilation - 1 kHz

Oscilation - 10 kHz

(a)

0.80

0.85

0.90

0.95

1.00

200 250 300 350 400

Time (s)

Absorbance (a.u.)

Without oscilation

Oscilation - 100 Hz

Oscilation - 300 Hz

Oscilation - 1 kHz

Oscilation - 10 kHz

(b)

Figure 7: (a) Measured absorbance at 495 nm for 30 mg/dl

of uric acid concentration in urine as a function of time for

different driving frequencies. (b) Zoom of dashed-square.

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

266

The analysis cuvette is filled with the reagent. Its

measured absorbance is 0.24 a. u.. Afterwards, the

urine standard with 30 mg/dl of uric acid is

dispensed on the reagent. The absorbance increases

as the sample is being mixed with the reagent. From

Figure 7 it can be seen that with the application of

oscillations, the mixing occurs in a faster way, being

the time necessary to obtain the complete mixing

(absorbance of 1 a. u.) at 1 kHz only one third

(300 s) of the complete mixing time without

oscillation (900 s).

From those results it can be concluded that the

application of acoustic microagitation by the

β-PVDF piezoelectric polymer improves the mixing

time. Moreover, uric acid has high diffusion

coefficients, which is a good characteristic for

mixing by diffusion only: after 15 minutes the

mixture will be completed, even without agitation.

However, when other biochemical parameters

present in biological fluids have to be analysed, such

as enzymes and some macromolecules, the mixing

by diffusion can take hours or can even not occur.

An example is a DNA analysis when PCR

(Polymerase Chain Reaction) must be performed.

5 CONCLUSIONS

A lab-on-a-chip device with fluidic acoustic

microagitation that reduces the mixing time of the

analytes with the reagents was reported. The device

consists of two dies: a microfluidic die, composed

by three cuvettes and a β-PVDF acoustic

microagitator, fabricated on a glass substrate; and an

electronic detection die, composed by the readout

circuits and the microagitation control electronics.

The main innovative concept is the application of a

β-PVDF film that produces acoustic microagitation,

increasing the mixing velocity. Experimental results

show that at 1 kHz microagitation, the mixing time

is reduced to one third of the time needed without

microagitation. As a conclusion, it can be stated that,

for decreasing device sizes, acoustic microagitation

becomes a preferred technology for effective

mixing.

ACKNOWLEDGEMENTS

Work supported by the Portuguese Science

Foundation (grants PTDC/BIO/70017/2006 and

POCI/CTM/59425/2004).

REFERENCES

Kopp, M. U., Crabtree, H. J. Manz, A., 1997,

Developments in technology and applications of

microsystems, Current Opinion in Chemical Biology,

1, p. 410-419.

Connolly, P., 1995, Clinical diagnostics opportunities for

biosensors and bioelectronics, Biosensors &

Bioelectronics, 10, p. 1-6.

Reyes, D. R., Iossifidis, D., Auroux, P., Manz, A., 2002,

Micro Total Analysis Systems. 1.Introduction, Theory,

and Technology, Anal. Chem., 74, p. 2623-2636.

Manz, A., Graber, N., Widmer, H. M., 1990, Miniaturized

total chemical systems: a novel concept for chemical

sensing. Sensors and Actuators B, 1, p. 244-248.

Auroux, P., Iossifidis, D., Reyes, D. R., Manz, A., 2002,

Micro Total Analysis Systems. 2.Analytical Standard

Operations and Applications, Anal. Chem., 74,

p. 2637-2652.

Minas, G., Wolffenbuttel, R. F., Correia, J. H., 2005, A

Lab-on-a-Chip for Spectrophotometric Analysis of

Biological Fluids, Lab-on-a-Chip, 5, p. 1303-1309.

Ribeiro, J. C., Minas, G., Turmezei, P., Wolffenbuttel, R.

F., Correia, J. H., 2005, A SU-8 Fluidic Microsystem

for Biological Fluids Analysis. Sensors and Actuators

A, 123-124 p. 77-81.

Minas, G., Wolffenbuttel, R. F., Correia, J. H., 2006, An

array of highly selective Fabry-Perot optical-channels

for biological fluids analysis by optical absorption

using white light source for illumination. Journal of

Optics A: pure and applied optics, 8, p. 272-278.

Minas, G., Martins, J. S., Ribeiro, J. C., Wolffenbuttel, R.

F., Correia, J. H., 2004, Biological microsystem for

measuring uric acid in biological fluids, Sensors and

Actuators A, 110, p. 33-38.

Bengtsson, M., Laurell, T., 2004, Ultrasonic agitation in

microchannels, Anal Bioanal Chem, 378-7, p. 1716-

1721.

Rife, J.C. et. al., 2000, Miniature valveless ultrasonic

pumps and mixers, Sensors and Actuators B, 86,

p. 135–140.

Sessler, G.M, 1987, Topics in Applied Physics: Electrets,

2nd Ed., Springer, Berlin Heidelberg New York.

Sencadas, V., Lanceros-Mendez, S., Mano, J.F., 2004,

Thermochimica Acta, 424, p. 201.

Lanceros-Mendez, S., Sencadas, V., Gregorio Filho, R.,

2006, Portuguese patent n.º 103318.

Biochemistry and Organic Reagents: for bioscience

investigation. Sigma-Aldrich Diagnostics®, 2006.

IBM, Photoresist composition and printed circuit boards

and packages made therewith. J. D. Gelorme, R. J.

Cox, S. A. R. Gutierrez, US Patent 4882245, 1989.

Candy, J. C., Temes G. C:, 1992, editors, Oversampled

Delta-Sigma Data Converters, IEEE Press, New York.

Netravali, A. N., 1997, Optimum digital filters for

interpolative A/D converters, Bell Syst. Tech. J., 56,

p. 1629.

LAB-ON-A-CHIP WITH FLUID ACOUSTIC MICROAGITATION - Piezoelectric Polymer ß-PVDF used as Ultrassonic

Transducer

267