A piezoelectric film transducer for dental occlusal analysis

Clinical

A4aieriais

10 (1992)

e P1e

145-151

lectric Film Transducer

for Den

Analysis R. L. Sakaguchi,” B. S. Wenande,a R. DeLong,” G. C. Andersonb a Minnesota Research Center for Biomaterials and Biomechanics, Department of Oral Science, ‘Department of Diagnostic and Surgical Sciences, University of Minnesota School of Dentistry, Minneapolis, Minnesota 55455, USA

(Received I June

1991; accepted 25 August 1991)

Abstract: Occlusal load, contact sequence and location are necessary parameters for the study of dental biomechanics, simulation and clinical treatment planning. A piezoelectric film transducer for dental occlusal analysis was developed and in vitro evaluations of the transducer were conducted in a servohydraulically driven artificial mouth. The transducer is designed to provide occlusal contact force information. The transducer thickness (9 pm) minimizes interference with normal mandibular closure. Voltage outputs from the sensor as a result of occlusal contacts generated in the artificial mouth were sampled through a computer controlled data acquisition system. The transducer output recorded by the data acquisition system was calibrated to the applied load in the artificial mouth. The output of the piezoelectric film was evaluated during varying loads, cycling frequencies, surface areas of contact, and transducer surface areas. The preliminary studies indicate that the piezoelectric film holds out considerable promise and with further development may be well suited as a diagnostic too! in dental occlusal analysis.

INTRODUCTION Dental occlusion involves the study of ‘the morphological and functional features of the contacting surfaces of opposing teeth’.l These relationships impact directly on jaw function, including mastication, swallowing and speech; and are of significance to all the specialty and subspecialty disciplines of clinical dental science.’ Three occlusal parameters: contact location, load and sequence, are required for the study of masticatory biomechanics and simulation. Accurate determination of these parameters would also be invaluable in planning and executing occlusal treatments such as occlusal adjustment, restorative treatment, orthodontics and orthognathic surgery. In presently accepted procedures, occlusal contacts are located clinically with articulating paper or Clinical

Materials

0267-6605/92/$05.00

0

mylar shimstock. 3 However, these techniques are poorly adapted to the concurrent determination of the parameters of contact location, sequence and load.3s a Several contemporary approaches attempt to quantify one or more contact parameters. These include photo-occlusion,5 sonic transmission6 and electric resistive elements.’ However, none of these techniques provide all of the necessary parameters simultaneously. One electronic device, T-Scan (Tekscan Corporation, Boston: MA, USA)’ attempts to integrate the occlusal parameters by using a sensor consisting of a large array of electrically conducting elements. One drawback of this system is the thickness of the sensor which impedes normal mandibular closure. Piezoelectric materials demonstrate properties which may be useful in determining occlusal parameters. In piezoelectric materials, transient

145 1992 Elsevier Science Publishers Ltd, England

146

R. L. Sakaguchi,

B. S. Wenande,

R. DeLong,

electric charge is produced by mechanical strain. The amount of charge produced is directly proportional to the amount of strain.g Polymers which normally do not exhibit piezoelectric properties, can be made to do so through the poling process. The polymer is heated to a temperature at which the molecules are free to move. It is then subjected to a strong electric field which aligns the molecules (electric dipoles) along the axis of the applied field. The polymer is then cooled under a constant electric field which prevents the dipoles from returning to their original random state. The level of the resultant piezoelectric activity depends upon the heating temperature, electric field strength and poling time. Piezoelectric film (PEF) is commercially available in thicknesses similar to those of conventional dental films used for analysis of occlusion. When an external force is applied to the film, the film develops an electrical charge which is proportional to changes in mechanical strain. This charge diminishes over time. A cycling load results in a corresponding alternating electrical signal output. Piezoelectric film transducers have been used in gait analysis for measurement of animal foot contacts and 1oads.l’ Generally, piezoelectric films can be used as force measurement devices because the charge produced varies linearly with load and is independent of loading frequency. PEF shows high elastic compliance due to the film’s flexible substrate and thin metal surface coating. The thin cross-section and high flexibility are desirable to minimize interference with normal occlusion.” The piezoelectric film also exhibits excellent piezoelectric properties and high mechanical strength. These properties are important design criteria for an accurate occlusal contact sensor. If PEF can be configured for dental use, it may contribute greatly to the knowledge base of dental biomechanics in which there is limited information concerning occlusal contact parameters, as well as provide important data for clinicians performing occlusal treatments. Piezofilm consists of a vinylidene fluoride polymer which is coated with silver ink. Because of the transient application of the film in a clinical setting, the composition of the film should not pose any health hazards. The electrical output of the film (2-3 V) is expected to be similar to the difference in potential between dissimilar alloy restorations commonly used in clinical dentistry.ll Shielding and isolation of the film should minimize any risks. The water absorption of the film is 0.02% and the operating temperature range is -40 to 100 “C.

G. C. Anderson,

W. H. Douglas

This paper describes the development and configuration of a piezoelectric film transducer for dental occlusal analysis and preliminary results concerning the feasibility for clinical applications. METHODS

AND MATERIALS

Kynar Piezo Film (Pennwalt Corporation, Valley Forge, PA, USA) is a semi-crystalline long-chain polymer of repeating vinylidene fluoride units (-CH,=CF,-). The film is available with two types of metallization: thin conducting metal layers deposited by vacuum metallization and thicker coatings obtained with conductive silver ink. Silver inked film was used in this project because of its enhanced durability and ease of etching for production of custom sensor geometries. The evaluations were conducted on a single piezo film strip (12 mm x 60 mm). Film and load cell outputs were collected by a Masscomp Computer MC500 supermicrocomputer (Concurrent Computer-formerly Masscomp Computer, MA, USA). A 16-channel, l-MHz analog-to-digital converter sampled the output at 100 Hz. The data were transferred in ASCII format to an Apple Macintosh IIci microcomputer (Apple Computer, Cupertino, CA, USA) for analysis. A battery of evaluations were performed on the film transducer prototype (Fig. 1) in an artificial mouth. The controller of the servohydraulically driven artificial mouth12 can be programmed to

Load cell spherical loading element

Fig. 1. Experimental configuration. (a) Amplifier circuit design for piezoelectric film transducer. (b) Schematic for piezoelectric film transducer in an artificial mouth with data acquisition system.

147

Piezoelectric jilm transducer Table 1. Amplifier

circuit evaluation

(output

in mV)

AnqdiJier type

1 Hz

5Hz

HP7090A FET Charge amp” Charge ampb

3.3 252.8 Results 880

35.5 16.9 1003.3 1166.7 too variable for analysis 920 880 860

O1Initial charge amplifier design. b Charge amplifier design with modified resistive elements.

10 Hz

capacitive

20 Hz

and

produce sine wave loading, square wave loading, ramp loading, etc. at various frequencies and forces. A cyclic ramp loading function was used to simulate the dynamics of tooth contact. The load was applied by a stainless steel spherical loading element at a constant rate up to the selected maximum load then unloaded at a constant rate to a point just short of lifting off the film.

provided consistent results, however, output signals were significantly smaller than those obtained with the FE7 and were also highly frequency dependent. The results obtained with the initial charge amplifier circuits were highly variable. Commercially available charge amplifiers were found to have insufficient input impedances. Modifications were made to the charge amplifier circuit through addition of a 0.29-nF capacitor between the film and ground which was designed to stabilize the circuit dynamics. A LOOresistor was installed to minimize output variability and frequency dependence. Using the modified circuit design, output voltages were significantly higher and the outputs were more consistent over the range of contact frequencies. The modified charge amplifier circuit design was used for all subsequent evaluations.

Calibration of data acquisition sy$t~ Development

of amplifier circuitry

Three amplifier circuits were tested and compared (Table 1). The three configurations included : (1) an HP709OA data acquisition instrument and plotter with memory buffer (Hewlett-Packard, Palo Alto, CA, USA) ; (2> an FET (field effect transistor) amplifier; and (3) a charge amplifier for signal amplification and hold. The charge and FET amplifiers were constructed using component parameters provided by Pennwalt Company. Since the charge produced by the film drains rapidly after initial contact, it is necessary to use an amplifier which samples and records the peak output signal at the time of load application. Initial circuit performance was evaluated using an electronic signal generator which provided known and consistent signal amplitudes, frequencies and wave forms. Since the film acts as a capacitor, circuit testing was performed by substituting an equivalent capacitor for the film into the circuit. The equivalent capacitor was determined by using eqn (1) : Capacitance,

C= $

(1)

where E, a permittivity constant = 106 x 10-l’ F/m, A’is the area of the film and t is the thickness of the film. Although the FET circuit provided the largest output signal it was highly frequency dependent at low cycling frequencies (Table 1). The HP 7090A

1

The data acquisition system was calibrated to the load cell by relating the output of the load cell on the artificial mouth to the DAS values for selected loads. DAS values representing film output were converted to volts by applying a constant load to the film and reading the amplifier outputs with a voltmeter.

Sensitivity to load and loading fre

The charge, Q, generated by an. external force stressing the film is proportional to the force, ~5’:~~ Q = d,o,A, = dttF

(2)

where d, is the piezoelectric strain constant, cZ is the stress applied perpendicular to the plane of the film, and A, is the contact surface area. The potential difference across the metallized surfaces or voltage, V equals the charge, Q, divided by the capacitance, C:

vy Q Substituting eqns (1) and (2) in (3); we get v-

d,Ft &A’

(4)

Thus, in theory the output voltage is directly proportional to the load applied to the film and the film thickness for a given strain constant.

R. L. Sakaguchi,

148

B. S. Wenande, R. DeLong,

The response of the film to loading frequency is a function of the amplifier circuit design rather than the film. The frequency response of the film ranges from near 0 to 10 GHz and has a flat frequency response over a wide frequency range.13 The response of the film to load and loading frequency was evaluated by the following experiment. Using a 6-mm spherical striker to simulate a cusp tip, five loads (4.4, 88,22,44, 88 N) were applied to the film. Five loading frequencies were used (0.1, 1, 5, 10, 20 Hz) for each of the loads listed above. These frequencies cover the range of normal physiologic chewing patterns6 Mean outputs were calculated from multiple cycles (- 25). Sensitivity

G. C. Anderson,

For a 28-pm film thickness, V, = - 6.048 x lo-%,

RESULTS

Based on eqn (4) above, the output voltage is theoretically independent of occlusal contact area. This was evaluated using two sizes of spherical strikers, 6 mm and 3 mm diameter, to apply a 4*4N load to the surface of the film. Sensitivity to compressive vs. shear loads

For Pennwalt’s piezofilm described earlier, the polarization axis is the thickness axis.13 The piezoelectric stress constant, g, is the negative of the electric field developed under open circuit conditions by the film relative to the stress applied along a specific axis.

Sensitivity to load and frequency (Table 2, Figs 2

and 3) For each loading frequency, the relationship between generated potential of the film and load was relatively linear (Y’= 0.99) for loads of 44 N or less. At 88 N some deviation from linearity was noted. The output voltage of the film was relatively Table 2. Contact mW Cycling frequency (Hz) 0.5 1 5 10 20

electric field developed along z (normal)-axis g,, = stress applied along z-axis x 10-37v/m

N/m Induced voltage, V = -got

e9>

The threshold of sensitivity to compressive loading was evaluated by placing an unrestrained strip of film into a cavity which simulated the occlusal surface of a molar. Compressive loads were applied with a metal spherical loading element. As a result of occlusal anatomy, the intra-oral PEF sensor is expected to bend when conforming to the tooth surface prior to contact. For use in clinical applications, the film output resulting from the compressive load needs to be distinguishable from the output resulting from bending and shearing of the film.

to contact area

= -339

W. H. Douglas

load

and frequency

44N

128 132 152 116 104

evaluation

(output

22N

44N

88 N

120 880 920 880 860

1360 1840 1800 1680 1360

2600 3300 3350 3000 Inadequate control in artificial mouth

(5)

where g is the applied stress and t is the film thickness. For the thickness or z-axis then, loading

(6) For a 2%pm film thickness, V, = 9.492 x IO-%,

(7)

For stresses in the plane of the film (x-axis), v, = --g&Z& electric field developed along z-axis g,, = stress applied along x-axis v/m = 216 x 10-3----, N/m

(8)

0

10

20

30 40 50 60 load (Newtons)

Fig. 2. Film output

70

80

vs. contact

90

load.

in

149

Piezoelectric film transducer

less than 35 mV will be indisti~~u~s~abie from the bending mode. This corresponded to a compressive force of 2.2 N.

load (Newtons)

r-l -o-

4.4

-e-

22

I+44 j t

i

i loading

Fig. 4.

Table 3. Contact

lb li

1’4 16 (Hz)

frequency

Film output

element

1’8

vs. contact

surface

88

I J

2b

frequency.

area evaluation

44 N load: 5 Hz cycling frequency 3 mm loading element 6 mm loading element

186 mV (+4%) 200 mV ( f 3.5)

For n = 26, t = 12.286, P < 0.001.

/

50 mV film output

i

2.2N (0.5 lb) load waveform

output resulting from bending of film Fig.

4. Compressive

vs. bending

load

constant between loading frequencies of 1 and 5 Hz. At 4~4Iv: the film output was relatively constant over the range of frequencies from 0*5 to 20 Hz. As the load increased, the film output decreased more severely from frequencies of O-5Hz and lo-20 Hz. sensitivity

to contact area (Table 3)

A statistically significant difference (P < O*OOl)in film output was noted between the two striker diameters. The larger striker resulted in a greater film 0utpu.t (200 2 3.5 mV) than the smaller striker (186k4.8 mV). Sensitivity to corn ressive vs. shear loads (Fig. 4)

Evaluation of the sensitivity of the film to compressive load versus bending or shearing indicates that a compressive force producing an output signal

DISCUSSIO The optimal thickness of an inter-occlusal sensor appears to be approximately 8 pm or less which is the thickness of Mylar shimstock used clinically for detection of occlusal contacts.3 The T-Scan (Tekscan Corporation, Boston, MA, USA)’ utilizes a sensor consisting of a large array of electrically conducting elements which describe discrete points in a plane. As the sensor is deformed, changes in electrical resistance within the array are detected and sampled over time. Although the T-Scan system is theoretically able to measure many of the required parameters of occlusion, the stiffness and lhickness (60 pm) of the intra-oral bite sensor interferes with normal occlusion. Piezoelectric film is approximately the same thickness as Mylar shimstock (available in 9 and 28 pm thicknesses) and exhibits similar compliance while having the additional benefit of piezoelectric properties. It is, therefore, an excellent candidate for a clinical dental occlusion analysis transducer. Equation (2) implies that the charge on the film resulting from an applied force, 4”, is directly proportional to that force. Q = d,F

(2)

From eqn (4), the generated potential of the film is directly proportional to the force applied to the film, F and film thickness, t. J,l

-

d@ &A’

44)

The voltage generated increases with decreasing metallized area as long as the force is concentrated within the metallized area,13 thus smaller sensor geometries should exhibit higher film output than larger geometries. This will be beneficial when multiple channel sensor arrays of small metallized islands are developed in further studies. Equation (4) also indicates that output voltage is directly related to the applied force. However, there is a cutoff frequency which is a characteristic of the amplifier design below which the output voltage is no longer linearly related to applied force. The large resistive element (100 MR) was installed in the charge amplifier circuitry in an attempt to drive the

150

R. L. Sakaguchi, B. S. Wenande, R. DeLong, G. C. Anderson, W. H. Douglas

cut-off frequency below the physiologic minimum contact frequency. 6 It appears that the cut-off frequency for the current circuit design is approximately 1 Hz. Further modifications of the circuit are necessary to decrease the cut-off frequency further. Theoretically, output voltage is independent of a contact frequency. However, experimentally, decay in voltage output was noted for frequencies above 5 Hz. This was most evident at the higher loads of 44 and 88 N and not as evident in the 4.4 and 22 N load applications. At higher loads and loading frequencies, there might be some deterioration of the film which decreases the output. Since the film is relatively flexible, application of high loads tends to plastically deform the film. When tests were conducted on sputter coated films, the films deteriorated rapidly. This was exhibited by a loss of conductance across the surface of the film. Silver inked film has a thicker coating and has less tendency for open circuiting. However, the effective thickness of the film may change as it is compressed by the loading element which will alter the output voltage as indicated by eqn (4). The independence of output voltage to contact area was evaluated using two sizes of loading elements. Loading the film at 4.4 N in the artificial mouth, the 6-mm and 3-mm strikers created 160-psi and 645-psi contacts respectively. Although no difference in output voltage was expected between the two striker diameters, a significant difference (P < 0.001) was noted in a measurement over 26 cycles. The larger diameter striker produced a larger output signal (200& 3.5 mV) than the smaller diameter striker (186 k 4.8 mV). This was not expected because eqn (4) indicates that the film is sensitive to load, not pressure. In eqn (4) the area parameter, A’, pertains to the metallized surface area of the film, not the surface area of contact. Factors such as bending and shearing of the film contribute to the response of the film to load application which masks the independence of output to contact surface area. If the bending and shearing responses can be filtered to provide a true measurement of force, the piezoelectric film transducer will have significant clinical applications. Variations in surface area of occlusal contact will not have an effect on occlusal load determination. This will be a strong point in favor of the use of PEF in occlusal measurements of force. When the film transducer is utilized in a clinical situation, it is anticipated that the film will bend and deform according to the occlusal anatomy. The film

is sensitive to any stress placed on it. Therefore, it is expected that there will be output resulting from bending and shearing forces on the film as it conforms to the occlusal surface. The voltage outputs can be estimated from eqns (7) and (9). The film is very sensitive to loads applied in the plane of the film since the load is applied over small crosssectional area. This, of course, assumes that film is anchored at two points. In the clinical setting, it will be unusual for the film to be held by two simultaneously occurring occlusal contacts. The film will be expected to bend to conform to the occlusal anatomy, but the shearing forces will be limited because of the nature of the occlusal contacts. In the preliminary studies, it appears that a compressive load of 22 N results in film output which is distinguishable from bending. It is also possible that further discrimination between bending and compressive contacts will be achieved through the incorporation of filters in the hardware and software. Based on the results of the preliminary studies, it appears that the piezoelectric film transducer is a reasonable instrument for characterizing occlusal loads. The output signal appears to be relatively linear for applied forces of 44 N or less. There is some deviation from linearity for larger forces. Generated potential of the film is reasonably constant for frequencies of 1 Hz and above for loads in the physiologic range (44-22 N). The output signal due to compressive loading is distinguishable from that due to pure bending for loads of 2.2 N or greater. Further investigations are continuing to enhance the amplifier circuitry and to develop and evaluate a multiple strip sensor for measurement of occlusal contact locations and sequences in addition to loads.

ACKNOWLEDGEMENTS This project was supported by research funds from the Graduate School of the University of Minnesota and National Institutes of Health/National Institute of Dental Research Grant No. IP30 DE09737. REFERENCES 1. Jablonski, S., Illustrated Dictionary of Dentistry. W. B. Saunders Company, Philadelphia, 1982. 2. Mohl, N., Zarb, G. A., Carlsson, G. E. & Rugh, J. D., (eds), A Text of Occlusion. Quintessence, Chicago, 1988, pp. 15-17.

Piezoelectric jiim transducer 3. Halperin,

4. 5.

6.

I.

G., Halperin, A. & Norling, B., Thickness, strength and plastic deformation of occlusal registration strips, J. Puosthet. Dent., 48 (1982) 575-8. Cazit, E. & Fitzig, S.: Reproducibility of occlusal marking techniques. J. Prosthet. Dent., 55 (1986) 505-9. Dawson, P. E. & Arcan, M., Attaining harmonic occlusion through visualized strain analysis. J. Prosthet. Dent., 46 (1981) 61522. Gibbs, C. H.; Mahan, P. E., Lundeen, H. C., et al., Occlusal forces during chewing and swallowing as measured by sound transmission. J. Pvosthet. Dent., 46 (1981) 443-9. Mannes, W. & Benjamin, M., Computerized occlusal analysis: a new technology. Quint. Int., 18 (1987) 287-92.

151

8. Rhee, E. & Maryniuk, G. A., Diagnostic efficacy of an electronic occlusal analysis instrument. J. Dent. Res., 68 (1989) abst. 437. McGraw-Hill, New York, 9. Cady, W. G., Piezoelectricity. 1949. 10. Koal, J. G., Record of Invention. 11. Craig, R. G. (ed.), Restorative Dental Materials, C. V. Mosby, St. Louis, 1989, p. 56. 12. DeLong, R. & Douglas, W. H., Development of an artificial oral environment for the testing of dental restoratives : bi-axial force and movement control. J. Dent. Res., 62 (1983) 32-6. 13. Pennwalt Corporation, Kynar Piezo Film technical manual, 1987.