This article is one of five papers to be presented exclusively on the web as part of the October 2000 JOM-e the electronic supplement to JOM.
	The following article appears as part of JOM-e, 52 (10) (2000), http://www.tms.org/pubs/journals/JOM/0010/Kumar/Kumar-0010.html JOM is a publication of The Minerals, Metals & Materials Society

Biosensors Based on Piezoelectric Crystal Detectors: Theory and Application

TABLE OF CONTENTS
INTRODUCTION PIEZOELECTRIC BIOSENSORS: BACKGROUND EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION CONCLUSIONS ACKNOWLEDGEMENTS References
Figure 1. Schematic diagram of a biosensor device.

Biosensors are chemical sensors that take advantage of the high selectivity and sensitivity of a biologically active material. It is well known that the resonant frequency of an oscillating piezoelectric crystal can be affected by a change in mass at the crystal surface. Piezoelectric immunosensors are able to measure a small change in mass. This paper describes the construction of an antibody-based piezoelectric sensor capable of detecting mycobacterial antigen in diluted cultures of attenuated M. tuberculosis.

INTRODUCTION

A biosensor is an analytical tool consisting of biologically active material used in close conjunction with a device that will convert a biochemical signal into a quantifiable electrical signal. Biosensors have many advantages, such as simple and low-cost instrumentation, fast response times, minimum sample pretreatment, and high sample throughput. Although biosensors are beginning to move toward field testing and commercialization in the United States, Europe, and Japan, relatively few have been commercialized. Increased research in this area demands the development of novel materials, new and better analytical techniques, and new and improved biosensors.^7-11 Some potential applications of biosensors are agricultural, horticultural and veterinary analysis; pollution, water and microbial contamination analysis; clinical diagnosis and biomedical applications; fermentation analysis and control; industrial gases and liquids; mining and toxic gases; explosives and military arena; and flavors, essences and pheromones.^1-6

A biosensor has two components: a receptor and a detector. The receptor is responsible for the selectivity of the sensor. Examples include enzymes, antibodies, and lipid layers. The detector, which plays the role of the transducer, translates the physical or chemical change by recognizing the analyte and relaying it through an electrical signal. The detector is not selective. For example, it can be a pH-electrode, an oxygen electrode or a piezoelectric crystal. Figure 1 describes a typical biosensor configuration that allows measurement of the target analyte without using reagents. The device incorporates a biological-sensing element with a traditional transducer. The biological-sensing element selectively recognizes a particular biological molecule through a reaction, specific adsorption, or other physical or chemical process, and the transducer converts the result of this recognition into a usable signal, which can be quantified. Common transduction systems are optical, electro-optical, or electrochemical; this variety offers many opportunities to tailor biosensors for specific applications.^1-6 For example, the glucose concentration in a blood sample can be measured directly by a biosensor (which is made specifically for glucose measurement) by simply dipping the sensor into the sample.

The objective of the research described here is to use biosensor technology to develop a rapid method for the diagnosis of tuberculosis and other infections caused by mycobacteria. The work encompassed here describes the construction of antibody-based piezoelectric crystals capable of detecting mycobacterial antigens in diluted cultures of attenuated M. tuberculosis in an immunologically specific manner. The antigen were detected in either liquid or vapor phase.

Table I. Some Common Biosensor Materials
Analytes	Examples
Respiratory Gases	O2, CO2
Anesthetic Gases	N2O, Halothane
Toxic Gases	H2S, Cl2, CO, NH3
Flammable Gases	CH4
Ions	H+, Li+, K+, Na+, Ca+, Phosphates, Heavy Metal Ions
Metabolites	Glucose, Urea
Trace Metabolites	Hormones, Steroids, Drugs
Toxic Vapors	Benzene, Toluene
Proteins and Nucleic Acids	DNA, RNA
Antigens and Antibodies	Human Ig, Anti-human Ig
Microorganisms	Viruses, Bacteria, Parasites

PIEZOELECTRIC BIOSENSORS: BACKGROUND

Two classes of bio-recognition processes-bio-affinity recognition-and bio-metabolic recognition, offer different methods of detection. Both processes involve the binding of a chemical species with another, which has a complementary structure. This is referred to as shape-specific binding. In bio-affinity recognition, the binding is very strong, and the transducer detects the presence of the bound receptor-analyte pair. The most common types of processes are receptor-ligand and antibody-antigen binding. In bio-metabolic recognition, the analyte and other co-reactants are chemically altered to form the product molecules. The biomaterials that can be recognized by the bio-recognition elements are as varied as the different reactions that occur in biological systems. Table I lists a number of common analytes that could prove attractive for developing biosensors of appropriate specificity and sensitivity. Almost all types of biological reactions, (chemical or affinity), can be exploited for biosensors. The concept of shape-specific recognition is commonly used to explain the high sensitivity and selectivity of biological molecules, especially antigen-antibody systems. The analyte molecule has a complementary structure to the antibody, and the bound pair is in a lower energy state than the two separate molecules. This binding is very difficult to break. Table II summarizes a variety of biosystem-transducer combinations in terms of transducer, measurement mode and potential application.

Table II. Biosensor Components
Transducer System	Measurement Mode	Typical Applications
Ion-Selective Electrode	Potentiometric	Ions in biological media, enzyme electrodes
Gas-Sensing Electrodes	Potentiometric	Gases, enzyme, organelle, cell or tissue electrodes
Field-Effect Transistors	Potentiometric	Ions, gases, enzyme substrates immunological analytes
Optoelectronic and Fiber-Optic Devices	Optical	pH; enzymes; immunological analytes
Thermistors	Calorimetric	Enzyme, organelle, gases, pollutants, antibiotics, vitamins
Enzyme Electrodes	Amperometric	Enzymes, immunological systems
Conductimeter	Conductance	Enzyme substrates
Piezoelectric Crystals	Acoustic (mass)	Volatile gases and vapors, antibodies

The interaction of antibodies with their corresponding antigens is an attractive reason for attempting to develop antibody-based chemical biosensors, i.e. immunosensors. Theoretically, if an antibody can be raised against a particular analyte, an immunosensor could be developed to recognize it. Despite the high specificity and affinity of antibodies towards complementary ligand molecules, most antibody-antigen interactions do not cause an electronically measurable change. However, the remarkable selectivity of antibodies has fueled much research to overcome this intrinsic problem. The piezoelectric effect in various crystalline substances is a useful property that leads to the detection of analytes. Figure 2 shows a schematic diagram of an immunosensor device.^12-21

Figure 2. Schematic diagram of an immunosensor device.

The piezoelectric immunosensor is thought to be one of the most sensitive analytical instruments developed to date, being capable of detecting antigens in the picogram range. Moreover, this type of device is believed to have the potential to detect antigens in the gas phase as well as in the liquid phase.

Almost all current methods of diagnosing tuberculosis (TB) have drawbacks. They tend to be either nonspecific or too time-consuming. In most cases of pulmonary and extrapulmonary TB, diagnosis depends upon culturing the mycobacterial organism, a process requiring 4-8 weeks. Significant attention has been devoted to developing more rapid diagnostic methods for TB, but some of them do not have the high specificity or sensitivity required for proper diagnosis.^22-29

A piezoelectric sensor that could reliably detect the mycobacterial antigen in biological fluids would be of enormous use. For instance, detection of the antigen in saliva could constitute a noninvasive method of screening high-risk populations. One tested piezoelectric crystal sensor gives results within a couple of hours after exposing the electrode to a liquid containing the antigen. The apparatus would be quite portable, so the immunological tests could be performed virtually anywhere, and the results could be obtained very quickly. The feasibility of using piezoelectric immunosensors to diagnose TB based upon the detection of mycobacterial antigens in liquid depends upon the degree of sensitivity and specificity that can be achieved and upon overcoming any problems caused by potentially interfering substances in biological fluids. The feasibility of gas- or vapor-phase detection of antigen depends upon these same factors, plus any difficulties that may be unique to gas-phase antigen capture by antibodies.

Theoretical Principals

The basic equations describing the relationship between the resonant frequency of an oscillating piezoelectric crystal and the mass deposited on the crystal surface have been derived by Sauerbrey,³⁰ Stockridge,³¹ and Lostis.³² Each followed a different path, but their final equations are similar, the Sauerbrey equation being the most widely accepted. In 1959, Sauerbrey developed an empirical equation for AT-cut quartz crystals vibrating in the thickness shear mode that describes the relationship between the mass of thin metal films deposited on quartz crystals and the corresponding change in resonant frequency of the crystal:

(1)

where, DF = frequency change in oscillating crystal in Hz, F = frequency of piezoelectric quartz crystal in MHz, DM = mass of deposited film in g, and A = area of electrode surface in cm².

These relationships not only apply to film deposition but also to particulate deposition. When vibrating in the thickness-shear mode, the oscillating frequency of an AT-cut quartz crystal is given by:

(2)

where F is the frequency of the crystal, N is the material constant (N = 1.66 MHz-mm for AT-cut quartz crystal), and a is the thickness of the crystal plate. Thus,

(3)

for finite amount of change, we may write

(4)

Dividing Equation 4 by Equation 2 gives:

(5)

If the thickness is defined as

(6)

where M is the mass of the electrically driven portion of the crystal, A is the area of the electrically driven portion of the crystal, and r is the density of the crystal.

Assuming constant density, we have:

(7)

If the changes are finite,

(8)

Using Equation 6 yields

(9)

Thus,

(10)

Since F and M are constants, we may say:

DF = -k DM

(11)

The oscillating frequency of the crystal changes linearly with the change in mass on the crystal. The mass change occurs due to deposition of materials on the surface of the crystal. This relationship is only valid for small mass changes. For larger mass changes, it would be invalid since the density would change.

Equation 10 may also be stated as:

(12)

This shows that the term

DF  DM

would increase if the base oscillating frequency of the crystal is increased. The term itself is the sensitivity of the crystal sensor, and Equation 12 shows that sensitivity is directly proportional to the base frequency of the crystal.

If a gas stream is sent flowing over the surface of the piezoelectric crystal and if it contains an analyte with concentration C, then

(13)

where C is the concentration of the analyte in the gas stream, DM is the mass of analyte in the gas stream, and V is the volume of the gas in the stream. The volume of the gas stream is related to the sampling time by the following relationship:

V = q · t

(14)

where q is the flow rate of the gas stream, and t is the sampling time. Equation 13 can be rewritten as:

(15)

where E is the collection efficiency of the coating material. If E is assumed to be equal to one and substituting Equation 15 into Equation 12, we may say

(16)

or

(17)

Thus, the change in the oscillating frequency of the crystal is related to both the sampling time and the concentration of the particles in the carrier gas. If we keep the sampling time, career gas flow rate and the mass of the analyte in the gas stream constant, we may restate Equation 17 as

DF = K C

(18)

Equation 18 shows that the change in frequency of the crystal is directly proportional to the concentration of the analyte in the gas stream flowing over it.

The piezoelectric crystal detector can be a very powerful analytical tool because of the relationship shown for the change in frequency to the analyte concentration with high sensitivity. Conversely, the above explanation shows that the crystal detector indiscriminately changes frequency due to the deposition of mass of any material on its surface. Thus, it is the task of the researcher to choose a coating that will undergo a highly selective chemical or physical binding with the substance to be detected. Only then can a highly selective sensor be constructed that will be sensitive to the subject to be detected.

EXPERIMENTAL PROCEDURES

Figure 3. Quartz crystal and holder.

Figure 4. Experimental apparatus for a piezoelectric sensor.

Figure 5. Schematic diagram of the antibody-antigen binding.

Electrode Fabrication Process

The most frequently used detector crystal is alpha quartz. These crystals are most suitable for piezoelectric application because they are insoluble in water and resistant to high temperatures. Alpha quartz crystals can be resistant to temperatures up to 579°C with no loss of piezoelectric properties. The resonant frequency of quartz crystal depends on the physical dimensions of the quartz plate and the thickness of the electrode deposited. AT and BT-cut crystals are most useful as piezoelectric detectors. These cuts refer to the orientation of the plate with respect to the crystal structure. The AT-cut crystal is the most stable, with a temperature coefficient of 1 ppm per degree centigrade over a temperature range of 10°C to 50°C. The crystals usually take the form of discs, squares, and rectangles.

All crystals in this investigation were general-purpose 10 MHz AT-cut quartz crystals with an electrode coating deposited on each side using sputtering method. The crystal was mounted on a holder with stainless steel with leads. A silver composite was used to connect the electrode to wire. The crystals were 14 mm in diameter, and the electrodes on both sides of the crystal were 8 mm in diameter. The crystals were mounted on size HC6/u holders. Figure 3 shows the schematic diagram of the fabricated crystal attached to the base.

Figure 4 is a block diagram of the apparatus used for the biosensor experiment. The piezoelectric quartz crystal was driven by a low-frequency transistor oscillator, powered by a 1-30 V d.c. regulator power supply and set at 9 V d.c. The frequency of the vibrating crystal was monitored by a Protek multifunction frequency counter. The crystal mounted on its holder was connected to the oscillator circuit and the frequency counter was connected to the oscillator device. After each step in the coating process-first with the various metal depositions and then with the biomolecular analytes-the frequency reading was recorded.³³

Immunosensing

The crystal electrodes were first modified with a 5 ml coating of protein A for better adhesion of the antibodies to the surface of the transducer. Protein A is a polypeptide isolated from Staphylococcus aureus that binds specifically to the immunoglobulin molecules, especially IgG antibodies, without interacting at the antigen site. This property permits the formation of tertiary complexes consisting of protein A, antibody, and antigen. Prior to modification, the electrodes were anodically oxidized at constant current in 0.5 M NaOH. They were then cleaned in 0.5 M HCl and 0.5 M HCrO₂. They were dried in an incubator for one hour, and the antibody (IgG) coating was then applied to the protein A coating. Using a pipette, 10 ml of antibody was applied on both sides of the crystal. After another hour of drying, 10 ml of antigens were coated onto the crystal, and they were methodically dried again. After each step, the frequency of the crystal was recorded, and the crystal was washed as a precaution against non-specific binding.

Control crystals and experimental crystals were coated with antibodies. The former were coated with an irrelevant antibody (HBV-honey bee antibody), one specific for an antigen not present in the solution containing the analyte. The experimental sample was coated with antibody (M. tuberculosis) specific for binding to the antigen. Both were exposed to the solution containing the analyte. Then the difference in frequency change between the control and experimental crystals were compared, reflecting the immunologically specific binding of analyte. This procedure was carried out first with crystal with gold substrate, and then using crystals coated with magnetic materials. A magnetic field was induced during the investigation of the latter. The nature of the binding of antigen to antibody to the surface of the transducer is shown in Figure 5. The protein helps the antibody to bind to the electrode and the antigen in gaseous state binds to specific antibody.³³

RESULTS AND DISCUSSION

Figure 6 shows the x-ray diffraction patterns of the gold films deposited on quartz. The gold film is polycrystalline in nature. The theoretical values at varying gold thicknesses, compared with experimental values at different thicknesses, are shown in Figure 7, which clearly shows that the experimental and theoretical values of change in frequency are almost equal for different thicknesses of electrode coating. The deposited coating changes the frequency approximately in agreement with the projections made by Sauerbrey.

Figure 6. X-ray diffraction pattern of gold electrode coating on a quartz substrate.		Figure 7. Frequency change (experimental and theoretical) vs. thickness for the deposited gold coating.

Although platinum is a more noble (non-reactive) material compared to the gold, the adherence of the platinum films was very poor on quartz substrates and the platinum reacted with different buffer solutions, (e.g. HCl and NaOH) during the specimen preparation for antigen-antibody binding. Thus, gold electrodes are preferred due to superior adhesion and non-reactive properties. The optimum thickness of the gold electrode layer was estimated at 1,000 Angstroms. Although the binding between antigen and antibody did show a change in frequency, the results were not always reproducible. The antibody binding to the protein layer was critical to achieve desirable results. The use of magnetic materials underneath the gold coating helped make the antigen detectable.^33-35

CONCLUSIONS

These results are preliminary. More analysis is needed to shed more light on the optimum binding characteristics of the antigen and antibody to the piezoelectric transducer.

ACKNOWLEDGEMENTS

This research was supported by a National Science Foundation grant. The author thanks S. Perlaky, I. Hussain, and A. Mangiaracina for their contributions in doing research at the University of South Alabama.

References

1. A.S. Dewa and W.H. Ko, "Biosensors," Semiconductor Sensors, ed. S.M. Sze (New York: Wiley Interscience, 1994), p. 415.
2. Elizabeth A.H. Hall, ed., Biosensors (New York: Prentice Hall, 1991).
3. D. Diamond, ed., Principles of Chemical and Biological Sensors, vol. 150 (New York: John Wiley & Sons, 1998).
4. C.R. Lowe, "An Introduction to the Concepts and Technology of Biosensors," Biosensors, 1 (1985), p. 4.
5. K.R. Rogers, "Biosensors for Environmental Applications," Biosensors Bioelectronics, 10 (1995), pp. 533-541.
6. A.P.F. Turner, "Current Trends in Biosensor Research and Development," Sensors Actuators, 17 (1989), pp. 433-450.
7. E.C. Hahn, "Piezoelectric Crystal Detectors and Their Applications," Ph.D. dissertation, University of New Orleans, 1988.
8. W.P. Mason, Piezoelectric Crystals and Their Application to Ultrasonics (Princeton, NJ: Van Nostrand, 1950).
9. G.G. Guilbault and J.M. Jordan, "Analytical Uses of Piezoelectric Crystals: A Review," CRC Crit. Rev. Anal. Chem., 19 (1) (1988), pp. 1-28.
10. R.A. Heising, Quartz Crystal for Electrical Circuits (New York: Van Nostrand, 1946), p. 24.
11. M. Ho, Applications of Piezoelectric Quartz Crystal Microbalances, ed. C. Lu and A.W. Czanderna (Amsterdam: Elsevier/North Holland, 1984).
12. J.H.T. Luong and G.G. Guilbault, "Analytical Applications of Piezoelectric Crystal Biosensors," Biosensor Principles and Applications, ed. L.J. Blum and P.R. Coulet (New York: Marcel Dekker, 1991), pp. 107-138.
13. G.G. Guilbault, "Detection of Formaldehyde with an Enzyme-Coated Piezoelectric Crystal Detector," Anal. Chem., 55 (1983), pp. 1682-1684.
14. J. Ngeh-Ngwainbi et al., "Parathion Antibodies on Piezoelectric Crystals," J. Am. Chem. Soc., 108 (1986), pp. 5444-5447.
15. G.G. Guilbault, B. Hock, and R. Schmid, "A Piezoelectric Immunobiosensor for Atrazine in Drinking Water," Biosensors Bioelectronics, 7 (1992), pp. 411-420.
16. M. Minunni, P. Skladal, and M. Mascini, "A Piezoelectric Quartz Crystal Biosensor for Atrazine," Life Chemistry Reports, 11 (1994), p. 391.
17. M. Minunni, P. Skladal, and M. Mascini, "A Piezoelectric Quartz Crystal Biosensor as a Direct Affinity Sensor," Anal. Lett., 27 (1994), p. 1475.
18. K. Nakanishi et al., "Detection of the Red Tide-Causing Plankton Alexandrium Affine by a Piezoelectric Immunosensor Using a Novel Method of Immobilizing Antibodies," Anal. Lett., 29 (1996), pp. 1247-1258.
19. M. Minunni, P. Skladal, and M. Mascini, "A Piezoelectric Quartz Crystal Biosensor as a Direct Affinity Sensor," Anal. Lett., 27 (1994), pp. 1475-1487.
20. K.R. Rogers and E.N. Koglin, "Biosensors for Environmental Monitoring: An EPA Perspective," Biosensors for Direct Monitoring of Environmental Pollutants in Field, ed. D.P. Nikolelis et al. (Norwell, MA: Kluwer Academic Publishers, 1997), pp. 335-349.
21. H. Muramatsu et al., "Piezoelectric Crystal Biosensor Modified with Protein A for Determination of Immunoglobulins," Anal. Chem., 59 (1987), pp. 2760-2763.
22. C.J.L. Murray, K. Styblo, and A. Rouillon, Disease Control Priorities in Developing Countries, ed. D.T. Jamison and W.H. Mosley (New York: Oxford University Press, 1990), p. 50.
23. B.R. Broom and C.J.L. Murray, "Tuberculosis: Commentary on a Reemergent Killer," Science, 257 (1992), pp. 1055-1064.
24. M.F. Goldsmith, "Medical Exorcism Required as Revitalized Revenant of Tuberculosis Haunts and Harries the Land," JAMA, 268 (2) (1992), pp. 174-175.
25. P.F. Barnes et al., "Tuberculosis in Patients with Human Immunodeficiency Virus Infection," N. Engl. J. Med., 324 (1991), pp. 1644-1650.
26. A.D. Harries, "Tuberculosis and Human Immunodeficiency Virus Infection in Developing Countries," Lancet, 335 (1990), pp. 387-390.
27. J.C. Weissler, "Southwestern Internal Medicine Conference: Tuberculosis-Immunopathogenesis and Therapy," Amer. J. Med. Sci., 305 (1993), pp. 52-65.
28. D.E. Sada, L.E. Ferguson, and T.M. Daniel, "An ELISA for the Serodiagnosis of Tuberculodid Using a 30,000-Da Native Antigen of Mycobacterium Tuberculosis," J. Infect. Dis., 162 (1990), pp. 928-931.
29. T.C. Huang et al., "Analysis of Cobalt Doped Iron Oxide Thin Films by Synchrotron Radiation," Thin Solid Films, 154 (1987), pp. 439-445.
30. G. Sauerbrey, Z. Phys., 155 (1959), p. 206.
31. C.D. Stockbridge, Vac. Microbalance Tech., 5 (1996), p. 193.
32. M. Lostis, Ph.D. dissertation, Faculty of Science, University of Paris, 1958.
33. I. Hussain, "Development and Applications of Piezoelectric Biosensors," M.S. thesis, University of South Alabama, 1998.
34. I. Hussain et al., "Fabrication of Piezoelectric Sensors for Biomedical Applications," MRS Symp. Proc. Materials for Smart System, 459 (1997), pp. 501-506.
35. Ashok Kumar et al., "Design and Implementation of a Piezoelectric Biosensors for Multifunctional Applications," Integrated Design and Process Technology, IDPT-Vol . 1 (1998), pp. 35-41.

Ashok Kumar is with the Department of Mechanical Engineering and Center for Microelectronic Research at the University of South Florida.

For more information, contact Ashok Kumar, University of South Florida, Department of Mechanical Engineering and Center for Microelectronics Research, Tampa, Florida 33620.