Improved Piezoelectric Ceramic-Polymer Composites

ELSEVIER

Synthetic

IMPROVED

Metals 85 (1997)

1391-1392

PIEZOELECTRIC CERAMIC/POLYMER HYDROPHONE APPLICATIONS

CdMPOSITES

FOR

C. Cui a, R.H. Baughman a: Z. Iqbal a, T.R. Kazmar b, and D.K. Dahlstromb aAlliedSignal, Research and Technology, Morristown, New Jersey 07962-1021, USA bAlliedSignal Ocean Systems, 15825 Roxford St.? Sylmar, CA 91342, USA

Abstract Composites of piezoelectric ceramic powders and organic polymers are described that provide highly desirable properties for sonar sensors: (1) high figures of merit, g,d,= 50(?5)x10-‘3 m2/N and g,d,/tanS = 3.3~10-‘~ m2M: (2) pressure-independent sensitivity up to high pressures (14 MPa, which is 2OOOpsi), (3) thermal stability of the poled state for months at lOO”C, (4) high dielectric constants (6070), (5) convenient melt processibility, and (6) robust mechanical properties. Key words: Applications,

piezoelectricity,

pressure sensor, dielectric constant, percolation

1. Introduction Piezoelectric materials have been used for a host of sensor and actuator applications. For example, lead zirconate titanate (PZT), polyvinylidene fluoride and vihylidene (PVW, fluoride/trifluoroethylene copolymers (PVDF-TrFE) are widely used sensor materials [1,2]. However, major performance liabilities exist for the application of such homogeneous composition piezoelectrics as sensors [3], which has motivated interest in ceramic/polymer composites. In order to guide the design of such composites for sensor applications, Newnham, Skinner, and Cross developed the concept of phase connectivity [3]. While there are many possible phase connectivity patterns, the least expensive composites consist of piezoelectric ceramic particles in a continuous, three-dimensionally connected matrix. These are referred to as either O-3, 1-3, 2-3, or 3-3 composites, depending upon whether or not the ceramic particles are percolated in zero, one, two, or three dimensions. It is typically difficult to assess the nature of ceramic particle percolation. Consequently, we will refer to ceramic particle composites in a three-dimensionally connected polymer host as O-3 composites independent of the degree of percolation of the ceramic particles. The present paper will describe a new O-3 piezoelectric composite that has convenient processibility, a high figure of a low dielectric loss, pressure-independent merit Wd, performance, and high thermal stability. This combination of properties suggests important applications opportunities, especially for hydrostatic sonar sensors. 2. Hydrophone

material

figure of merit

The hydrophone figure of merit (FOM) for piezoelectric materials is defined as sg,2, where g, is the piezoelectric voltage 0379-6779/97/%17.00 0 1997 Elsevier PII SO379-6779(96)04405-0

Science S.A. All rights reserved

coefficient and E is the dielectric permittivity [4]. It can be easily shown that the FOM is the volumetric electrical energy density produced by a pressure fluctuation of unity. This conclusion follows since a piezoelectric sensor is a capacitor in which a pressure change p generates a voltage V = tg,p. If the area and thickness of a piezoelectric in a plate-shaped sensor are A and t; respectively, then the sensor capacitance is C = A&/t. Hence, the electrical energy stored as a result of the pressure change p is E = CV’/2= (Aait)(tg,p)2/2 = Atagh2p2/2. If the sensor volume (v) is approximately equal to the volume of the piezoelectric: then the FOM can be expressed in terms of this energy as: FOM = agh2 = 2Elvp2

(1)

This equation implies that the FOM is twice the volumetric electrical energy density produced by unity applied pressure. If the FOM is divided by the average sensor density p, we have a new figure of merit WFOM = FOM/p = 2Elpvp2 = 2E/Wp”,

(2)

where W is the sensor weight. Consequently, the densitynormalized figure of merit is the gravimetric electrical energy density produced by a unity change in pressure. 3. The optimization

of FOM

3. I. Dielectric constants The present high dielectric and chemically particles. The

properties optimization resulted from our choice of constant, low dielectric loss, high melting point, inert polymers for the host matrix for the ceramic field (E,) acting during poling on unpercolated

1392

C. Cui et al. /&‘ynthehcddetals

ceramic particle in a polymer matrix is approximately related to the applied field (E,) in the short-voltage-pulse limit by E,=3E0sp/(2sp+ E,),

(3)

where sp and E, are the dielectric permittivity of the polymer and ceramic, respectively. Because the dielectric constant (E/E,) of most polymers is lower than 7 and the dielectric constant of most ceramics is above 200, E, is generally a small fraction of E,. Consequently, the choice of a high dielectric polymer matrix can be important for effectively poling the embedded ceramic particles. Also, this choice of a high dielectric constant is important because the remanent polarization (P,) of the composite is approximately related to the remanent polarization (P,) of the ceramic particles by [6] P,= 3v,P,sd(2~r +

EC);

(4)

where v, is the volume fraction of the ceramic powder in the composite. In order to obtain sufficient remanent polarization, the dielectric constant of the polymer component should be as large a fraction of the dielectric constant of the piezoelectric ceramics as possible. The above arguments led to our choice of a non-piezoelectric phase of PVDF (E/E, = 10) as the host matrix and Ca-modified lead titanate (Ca-PT) (E/E, = 270) as the ceramic component. The polyvinylidene fluoride (from Polyscience? Inc.) had a weightaverage molecular weight of 35O:OOO and a melting point of 166170°C. The ceramic powders used were prepared by fractionating (according to particle size) the ED0 EC-97 ceramic obtained from ED0 Ceramics, which has a reported composition of Ph &~,sJ((CO,~~W~~ 04Ti0.9~P3. 3.2. Ceramic particle size and distribution We find that it is important to use ferroelectric ceramic particles that are nearly monodispersed in size. When a ceramic powder/polymer composite is formed under high pressures from many differently sized ceramic particles, the smaller particles can reside in cages formed by larger ceramic particles. Stresses applied externally to such a composite will not be efficiently transferred to these smaller particles that reside in the cages. As a result, the smaller ceramic particles will not make significant contributions to the piezoelectricity of the composite. We call this the cage effect. When the particle size is larger than 30 ym, ceramic powders can be conveniently obtained that are free-flowing. This aspect facilitates composite processing to form void-free composites. For much smaller particles, agglomeration hinders uniform dispersion of the ceramic powder in the polymer matrix. For these reasons, the Ca-PT ceramic powders used for the present investigations were 70-90 pm in diameter. 3.3 Ceramic particle percolation In order to obtain adequate inter-particle contacts, we and others (71 believe that the ceramic powder loading level should be above about 65% by volume. This is the loading level used for the presently described composites. The degree of percolation in the present samples has not been directly determined, and

85 (i997)

1391-1392

Equations (3) and (4) are not applicable if the composite is fully percolated. Nevertheless, our experiments support the conclusion derived from these equations that sensor performance is optimized by maximizing EWE,. We find that the use of high pressures (typically about 100 MPa) during the melt fabrication of composites is desirable in order to eliminate voids from the ceramic/polymer composites. While the presence of voids can increase device sensitivity and FOM, such voids typically introduce an undesirable pressure dependence of hydrostatic sensitivity for the pressures that are of interest for sonar sensors. 4. Results The above analysis, and our work on other compositions, led to an optimized PVDFKa-PT composite containing 65% by volume of the ceramic, whose properties are here described. Pellet formation was performed by the compaction of thoroughly mixed ceramic and polymer powders at 100 MPa and 190°C. The composite pellets were poled at 110°C and 125 kV/cm for 20-30 minutes. The resulting dielectric and piezoelectric properties for the poled composite pellets are listed in Table 1. Table 1. Dielectric composites d,,(pCiN)

and piezoelectric

d,(pCM)

58

56

g,,(mV/N) 0.094

properties

(EJE$

70

of the present

FOM(lK’3mZ/N) 54

aDielectric constant in the poling direction. The above measured figure of merit for our composite (FOM = 54) is substantially higher than for conventional ceramics (19.3 for lead titanate) and for available polymers (17 for a commercially available vinylidene fluorideitrifluoroethylene copolymer). Sonar hydrophones based on this composite provided a negligible sensitivity change with pressure up to 14 MPa. These results, combined with the low loss and frequencyindependent performance we obtained for this composite in hydrophones, suggest important applications opportunities. 5. Acknowledgment We thank Professor L.E. Cross, Professor A. Safari, and their colleagues for useful discussions and assistance with measurements. References [l] B. Jaffe, W.R. Cook, and H. Jaffe, Piezoelectric Ceramics, Academic Press, New York, 197 1, [2] H.S. Nalwa, Rev. Mucromol. Chem. Phys., Cl3 (1992) 341. [3] R.E. Newnham, D.P. Skinner, and L.E. Cross, Mat. Res. Bull. 13 (1978) 525. [4] A. Safari, J. Phys. III, Appl. Phys., Mater. Ski., Fluids, Plasma, Instrum., 4 (1994) 1129, and references therein. [5] A.R. von Hippel, Dielectrics and IJ’aves, Wiley, New York, 1954. [6] Y. Wada and R. Hayakawa, Ferroelectrics, 22 (198 1) 115. [7] J.A. Hossack and B.A. Auld, Ferroeiectrics, 156 (1994) 13.