PVF2 Transducers for NDT

We have investigated the spatial dependence of the longitudinal piezoelectric stress constant e of PVF . Experiments were performed on a series of no.minally identical brass-backed PVF longitudinaF wave t~ansducers in water using commercial PVF7 film. Computer programs were designed t6 predict the performance of the transducers as a function of e (Z) , the thickness mode piezoelectric stress constant as a function of position Z through the thick~~ss of the film. Our experiments indicate that the coupling coefficient is uniform across the film thickness. These computer programs were also used to model the insertion loss and bandwidth performance of the transducers. High voltage pulses were applied to PVF2 transducers to determine the region of linearity and to find the maximum nondestructive voltage that can be used. Transducers of this type were also made using PVF films fabricated in the Stanford Center for Materials Research using commercial resin, and found to pertorm as well as transducers using commercial film. Transducers of the above kind have also been used as bulk wave sources in wedge transducer assemblies for the production of surface acoustic waves on ceramic plates, and initial observations of reflections from surface cracks have been made. It has been reported the PVF film may have a spatially non-uniform piezoelectr~c coupling coefficient.l The claim is that the active piezoelectric region is concentrated near the positive side of the film, the positive side being the side that was charged positively during the poling process. The PVF?. transducer, which consists of a piece of electroaed PVF bonded to a brass backing,2 may have either th~ positive or negative side of the film exposed. The two possible configurations are shown below in Fig. 1. If the piezoelectric coupling constant is non-uniform across the thickness of the film, the insertion loss characteristics of a PVF2 transducer should depend on the configuration of the transducer. This statement was verified using a computer matrix model that calculated the insertion loss of a single multi-layer PVF2 transducer transmitting into water.3 Fig. 1. Schematic of the PVF2 brass-backed transducer configurations used. Results obtained using this program are shown in Fig. 2.. The PVF2 was modeled as ten separate 3.1 lJ thick layers, with e values of (.1, .05, .025, .0125, .00625, .00312~~ 0, 0, 0, 0) in order across the 10 layers. The top curve is the predicted insertion loss when the layer with the maximum e value is located next to the loading materialztwater). This configuration results in a resonance valley around w = .6 w due to the layer of nonpiezoelectric PVF?. trapped Between the backing material (brass) ana the active PVF . The bottom curve is the predicted insertion lo~s when, the layer with ezz = .1 is located next to the brass. Figure 2 shows that large insertion loss dif153 ferences should be apparent if ezz is strongly non-homogenous. ~L-----~.2~-----.~4------~.,~-----.·~----~1.0 Fig. 2. Calculated insertion loss curves for PVF2 with non-homogenous value of ezz Six brass backed PVF transducers were constructed using 25 lJ thick2film obtained from Kureha Corp. The aluminum electrodes originally on the PVF7 were removed and replaced by a fl~sh of chromium and a thin layer of gold (1000 ~). deposited using an e-gun evaporation system. The polymer was then repoled with a voltage of 2500 volts (~ 106 V/cm) placed across the PVF for two hours. The positive electrode of the p6ling apparatus was connected to the side of the PVF which had been poled positively by Kureha. The poting wgs performed in an air oven, at a temperature of 80 C. Using this repoled material the six transducers mentioned were constructed.3 Three of the transducers were bonded so that the side of PVF2 which was charged positive during poling was exposed to the air. (These were labeled the +1, +2, and +3 transducers.) The other three transducers were bonded so that the positive side during poling was epoxied to the brass. (These were labeled the -1, -2, and -3 transducers.) Two-way reflection mode insertion loss measurements were then performed on all transducers. Insertion loss is defined on the basis of a 50 n source generator driving the transmitting transduce~ and a 50 n load connected across the receiving transducer. The results of this experiment are shown in Fig. 3. The graph reveals no significant 70 l'lASURED INSUHOH LOSS 6 TRNI:SDOC£R$