From Powder to Parts: An Introduction to Manufacturing Piezoelectric PZT Ceramic Components
Piezoelectric materials are ideal for medical, defense and industrial applications.
Piezoelectric materials have the unique ability to transform mechanical energy into electrical energy, and vice versa. If an electric charge is applied to a piezoelectric material, it will deform. The magnitude and orientation of the deformation is determined by the polarity, magnitude, and orientation of the charge relative to the polarization direction. If an alternating charge is applied, the material will deform cyclically. At the appropriate frequency, an audible sound (< 20 kHz) or ultrasound (> 20 kHz) can be generated. Harnessed effectively through transducer or sensor design, this unique property makes piezoelectric materials ideal for medical, defense and industrial applications.
The properties of a finished lead zirconate titanate (PZT) ceramic component are determined by the starting PZT powder composition and the physical geometry of the component being manufactured. Common geometries include disks, rings, plates and cylinders. In addition to geometry, electrical properties are also targeted during manufacturing to ensure compatibility among the component, the housing or package, and the electronics within the system for which it is designed.
One of the most common electrical requirements is resonant frequency (Fr). The resonant frequency is the frequency at which the component exhibits the lowest electrical impedance. It is therefore important because parts driven at this frequency have high mechanical efficiency and require less power to function; however, a component can have more than one resonant frequency.
Typically, each physical dimension exhibits its own resonance. For example, a disk will have a diameter resonance and a thickness resonance. A plate will have a length mode resonance, a width mode resonance and a thickness resonance. This is due to the dependence of the resonant frequency on the speed at which sound travels through the material (frequency constant N) and the physical dimension in any orientation. A typical frequency constant for PZT material is 2,050 Hz/m, but this can shift with compositional changes in the material.
Grain structure and domain alignment can also affect the frequency constant, resulting in a different value for each direction in which the sound is travelling through the material (i.e., NThickness and NRadial can be different). The relationship between frequency constant and physical size for the radial mode is shown in the following equation:
NRadial = ƒRadialD
where NRadial = frequency constant for radial mode, ƒRadial = component frequency, and D = diameter of component.
Following the resonant frequency is the anti-resonant frequency (Fa). This is higher in the frequency range, and is the frequency at which the component’s electrical impedance is at its maximum. For each resonant frequency, there is an anti-resonant frequency. The difference between the resonant and anti-resonant frequency is referred to as the delta frequency (DF), and the ratio DF/Fr is one figure of merit indicating the component’s efficiency at converting physical to electrical energy and vice versa. A large ratio indicates a more efficient component. Typical acceptable values for the DF/Fr ratio can range from 0.10 to 0.30, depending on the composition of the material and the mode in which the component is being measured.
Figure 1 shows a log-impedance vs. frequency curve for a PZT disk with an aspect ratio of approximately 10:1 (diameter:thickness), measured using an HP 4194 impedance analyzer. Note the resonant peak of lowest impedance for the radial mode, the corresponding anti-resonant peak, and the resulting DF and DF/Fr ratio.
It is also critical that the capacitance and the d33 of the ceramic be considered when manufacturing components for existing electronic systems. The capacitance is a measurement of the material’s ability to store electrical energy, and is controlled using the physical geometry of the component and its material composition:
where C = capacitance, KT = relative dielectric constant of the material, A = active (electrode) area of the component, and T = thickness of the component.
The d33 is an additional indicator of the component’s ability to convert mechanical to electrical energy, is related to the DF, and is easily evaluated using a d33 meter. The relationship between d33, applied voltage, and the resulting dimensional change of the component are shown in the following equation:
where Dh = change in thickness, V = voltage applied, and d33 = piezo charge constant.
Prior to manufacturing the components, it is important to understand the end use, the existing electronic system, and the specific requirements (e.g., duty cycle, power requirements, etc.), since all are used when selecting the appropriate PZT material. Many other properties can be considered when designing components, but using these equations and others it is possible to design a basic component for the desired end use with an appropriate shape (to fit inside existing packaging, or lend to a frequency), frequency (to maximize efficiency in existing or new systems using physical dimensions), and capacitance (to ensure compatibility with existing electronics using physical dimensions, electrode dimensions and material selection). Once a component’s approximate shape and size have been determined, the manufacturing process can begin.
Forming and Firing
The first step is to form green bodies from PZT powder. The starting powder for this process is pre-reacted, spray-dried PZT powder containing a polyvinyl alcohol-based binder system, ready for dry pressing (read more about this process in “PZT Powder Manufacturing,” Ceramic Industry, May 2010). The purpose of spray drying the powders is to yield a free-flowing material in the form of hollow spherical agglomerates in the 100-200 um range, comprised of 0.3-0.5 um PZT crystallites. Figure 2 illustrates binder containing spray-dried hollow spheres comprised of individual PZT crystallites.
The free-flowing powder of the appropriate composition is dry pressed at a compaction ratio of 2:1 at approximately 10,000 PSI to a near shape 15-20% larger than the desired finished shape. This is to account for additional densification by sintering in later firing stages (parts shrink through the sintering process, resulting in densification). The compacted PZT parts are considered “green” in this state, and are approximately 4.90 g/cc (~ 60% dense).
Basic and complex shapes, including disks, rings, plates, lenses and large bricks (for the manufacture of very complex shapes), can be pressed for additional processing at a later stage. It is common to add a lubricant to the powder prior to pressing to reduce particle-on-particle friction, and particle-on-die wall friction to improve density distribution and reduce die wall friction when pressing and ejecting.
The presence of organic binder in the green ceramic requires the parts to be fired in the presence of oxygen (if oxygen is not available, a reducing atmosphere will occur); sintering completely in air allows the lead to evaporate from the ceramic, shifting the stoichiometry and affecting the desired phase. To circumvent this issue, the green ceramics can be fired in air to approximately 600°C. This allows the organic binders to burn away while retaining the lead in the part (volatile at ~ 800°C) and building some strength in the ceramic. The parts can then be enclosed in high-purity alumina (Al2O3) crucibles (limiting potential lead loss) and moved to higher temperature furnaces, where solid state sintering between 1,200 and 1,350°C occurs, resulting in densities of 7.6-7.8 g/cc (~ 95% dense) and grain growth to the desired 3-5 um range for good mechanical strength and optimum electrical properties. Figure 3 shows an SEM micrograph of a sintered PZT fracture surface with 3-5 um grains.
Shaping and Machining
Since the matching of the ceramic to the electrical driving circuits and end-use requirements depends on the frequency, capacitance, etc. of the device, the parts are precisely machined in all dimensions. This is accomplished using a variety of machining techniques. The hardness and degree of brittleness of the sintered PZT ceramic requires the use of diamond or silicon carbide abrasives for machining. Material removal is based on grinding, not cutting. Coolant must be used to cool the grinding interface, lubricate the tool, and carry away the by-products (swarf). Machining operations include lapping, polishing, surface or center-less grinding, dicing, slicing, and multi-axis milling for complex shapes.
When removing material abrasively, several critical factors need to be considered, including surface finish (Ra) of the PZT, tool wear rate, surface speed of the abrasive, and feed rate of the abrasive into the piece. Successful machining is accomplished through a compromise among surface finish, cutting rate, tool wear, and desired precision (tool tracking through the piece). Ra < 0.7 um is a typical surface finish target for machined PZT surfaces. An abrasive used to accomplish this would typically be diamond in the 320 grit (~ 44 um) range, metal bonded for durability, and operating at surface speeds anywhere from 1,200-1,800 meters per minute (MPM) with feed rates ranging from 25-200 mm/min. Figure 4 shows a common setup for dicing PZT plates to their final size using appropriate abrasives and coolant.
The precision of the completed piece (e.g., flatness, perpendicularity, tolerance level, etc.) is achieved using good work-holding techniques and fixture design, while still considering the inherent compromises of machining listed previously.
Once the parts are completely machined to the appropriate physical dimensions, they are cleaned and conducting electrodes are applied using silk screening, air brushing, dipping, rolling, sponging, hand brushing or chemical deposition processes. Other techniques available to the industry include sputtering or evaporative processes, but these typically require vacuum and therefore larger, more complex and more expensive equipment.
Silk screening with a silver-based electrode is the most commonly used technique in this industry. The deposition thickness and resolution of the print are controlled primarily by the physical attributes of the screen, including open area, mesh count and wire diameter. The electrode paste and its properties are equally critical to the silk screening process. Paste formulations generally contain ~ 70% metal, 25% organic vehicles and 5% glass. The vehicle system and particle size of the solids in the paste determine its rheology, a thixotropic paste being preferred for print consistency and ease of deposition.
After printing, the electrode is dried of its volatiles and fired in a flowing air atmosphere where softening of the glass and subsequent bonding of the electrode to the ceramic take place. The temperature at which this occurs is dependent on the composition of the glass in the paste, but is typically in the 500-850°C range. The quality of the bond can be easily determined by soldering a wire to the surface and performing a pull test. It is not uncommon to expect the electrode/ceramic bond to support an 8+ lb tensile force applied to the solder joint. If component geometry does not support a solder pull test, an adhesive tape test can be performed as an alternative (details of these techniques are available in MIL-STD1376B). Figure 5 shows general work-holding for printing electrodes on small PZT disks.
Internal to (within the structure of) the ceramic, there are multiple domains that have individual dipole orientations. The orientations of these dipoles are such that they cancel each other out, leaving the component without a net polarization. The purpose of the poling process is to re-orient and realign these dipoles within the part structure to yield a net remnant polarization.
This re-alignment occurs when the part is exposed to a high DC electric field (1,600-3,200 volts per mm of thickness) for a fixed period of time at an elevated temperature. The process is generally performed in a fluid with a dielectric strength greater than that of air. The exact voltage, time, and temperature are determined by the desired final properties and the material composition. The resulting remnant polarization is reflected by the d33 value, and its magnitude is an additional indicator of the component’s ability to convert physical to electrical energy, and vice versa.
Testing and Inspection
After polarization, the piezoelectric components undergo final testing and inspection. Final electrical testing is performed to the initial component design inputs (e.g., capacitance, d33, dielectric loss, frequency, etc.) using equipment such as impedance analyzers, d33 meters, LCR meters, etc., as dictated by the specific requirements. The inspection process may include physical (for dimensions) and visual (for chips, cracks or flaws in the electrode material). These final tests and inspections are performed on 100% of the components produced to ensure product quality and validate the manufacturing process for the components being tested.