Advanced Sonar Projector Materials
Specific requirements for acoustic transducers are driven by the application. Sometimes they are a combined reciprocal device; i.e., the same element is used to both transmit and receive with the appropriate switching. More often, separate projectors and hydrophones are used. This allows the designer more freedom to choose materials optimized for either transmitting or receiving and to design the physical configuration for the desired acoustic response. In this article, we will focus on underwater transmit, i.e., projector, applications.
Most underwater projectors today use piezoelectric materials; i.e., materials that change shape in response to an electric field. The type of piezoelectric material used is a key selection criterion when trying to achieve high-power undersea acoustic performance. Quartz crystals and Rochelle salts were some of the first piezoelectric materials used for undersea sonar applications during World War I. These materials were chosen, in part, because the preferred single-crystal form is easy to produce. However, neither quartz nor Rochelle salts possess the acoustic coupling factor of more modern material, and they therefore provide limited acoustic power output.
Late in World War II, improved materials such as permanently polarized barium titanate ceramics were discovered; and in the early 1950s, polarized lead zirconate titanate (PZT) ceramics were introduced. These piezoelectric ceramics initiated the modern era of acoustic transducers. At that time, single PZT crystals of any significant size could not be grown. However, these new ceramics quickly replaced the earlier single quartz crystal materials because of the higher amount of piezoelectric activity or the coupling factor that could be achieved through ceramics. Sixty years later, PZT ceramics are still being used in most underwater transducers.1
Over the past decade, developers have demonstrated the ability to grow large PZT single crystals with performance that is significantly better than that of conventional PZT ceramic.2 This single-crystal material has gained acceptance in specific medical ultrasound and high-strain actuator applications in which only small quantities of high-power material are required. However, due to their high cost and process limitations, these PZT single-crystal structures have not been widely used for underwater sonar applications.
Raytheon has been collaborating with the Penn State University Materials Research Laboratory to develop and evaluate new transducer materials for sonar applications. A new “textured” ceramic material promises to achieve near single-crystal performance at a fraction of the cost. Similar to single crystal, the low sound speed of textured ceramic results in a smaller projector for a given frequency. In addition, the higher electro-acoustic activity of textured ceramic provides greater acoustic output and bandwidth, even from the smaller device. Applications range from shipboard sonar (frequencies less than 10 KHz) to medical ultrasonics (frequencies greater than 1 MHz). To better understand the advantages of this newer textured ceramic material, it is helpful to understand how ceramic piezoelectric materials are made.
Traditional Ceramic. The manufacturing of traditional ceramics begins with the mixing of powders of the various constituent elements in specific, usually proprietary, proportions. The resulting powder mix is then combined with a binder and heated in a furnace. The temperature profile of the furnace is set so the particles adhere to one another but do not melt. This process is called sintering. As the ceramic cools, a dense polycrystalline structure is created and adjoining crystals form domains. Although these individual domains have a net dipole moment, the bulk material does not because the domains are randomly oriented relative to each other. This inert ceramic is transformed into a useful transducer material during the subsequent application of a high-DC electric field within the poling process. During poling, the domains that are most closely aligned with the electric field lengthen and become permanently oriented with the electric field. When the high-DC electric field is removed, these lengthened domains produce a voltage when tension or compression is applied. It is important to note, however, that not all domains become aligned. Some become partially aligned, and some are not aligned at all.
Single Crystal. One can see that the performance of the piezoelectric material could be improved if a larger number of domains were aligned. To accomplish this, new manufacturing methods and ceramic formulations were devised. In the Bridgman method, a newer ceramic formulation — e.g., lead magnesium niobate-lead titanate (PMN-PT) — is grown in a platinum crucible. A seed crystal is located in the bottom (cooler end) of the crucible and partially melted. Crucibles vary in size, but a typical laboratory crucible is approximately 15 cm in length with diameters ranging from 1 to 2.5 cm. A temperature gradient is then applied to melt the ceramic and grow the crystal from the seed. Unidirectional heating and subsequent solidification are accomplished by translation of the crucible through the temperature gradient. Translation rates of 0.8 mm per hour are typical, and the temperature gradient maximum is about 1,400 degrees Celsius. After growth, crystal is slowly cooled to room temperature to prevent cracking. The total time to cool is generally 100 hours. A crucible cannot be reused because it is cut away to remove the grown crystal.3
When a poling voltage is applied to the preferred crystallographic direction of a single crystal, the number of domains that are aligned with the electric field is greatly increased. The resulting single-crystal ceramic has a much higher coupling factor. The primary disadvantage of this single-crystal structure is the inability to grow large crystals at a reasonable cost. As discussed above, growing crystals is a slow process. Additionally, single-crystal ceramic vendors are still learning how to grow large crystals with uniform properties, and this lack of uniformity seriously affects the yield and the resulting cost.
Tape Casting and Templated Grain Growth. Tape casting is a method to create uniform sheets of material. It has been in use since the 1950s to create dielectric tape for the capacitor industry. A slurry of the raw ceramic and special additives is placed in a chamber and extruded through a small gap onto a conveyor belt (Figure 1). The gap and resulting tape thickness are controlled using a “doctor blade.” The conveyor belt can be covered with a non-stick polymer or sacrificial paper that is burned in the sintering process. Tape casting has been used to produce conventional thick-film piezoceramics in the past. It has recently been combined with the process called templated grain growth (TGG) to produce textured PMN-PT ceramics. This breakthrough, known as “textured material” (TM), was accomplished at the Penn State University Materials Research Laboratory.
The key to achieving improved properties is to mechanically align (texture) the jumbled crystal domains rather than align them by crystal growth. Texturing polycrystalline ceramic causes its properties to become more anisotropic, yielding performance gains in certain directions. A small fraction of template particles or grains is added to a slurry containing a much finer equiaxed matrix powder. The slurry is then shear-formed using the tape casting method described above to align the anisotropic particles in the direction of least resistance to flow. An electron microscope photograph of the textured result is shown in Figure 2.
TGG is a cost-effective alternative to Bridgman-grown single crystals. Textured PMN-PT obtained through this process not only shows better piezoelectric properties than random ceramic PMN-PT, but it can approach Bridgman single-crystal performance.4 The tape casting process, combined with TGG, can be used to make flat plates and other shapes. Wrapping the tape around a mandrel produces a ring-shaped ceramic with domains aligned primarily in the radial direction. A photo of a Raytheon-fabricated cylindrical transducer is shown in Figure 3.
Traditional ceramic, single crystal, and template grain growth material may be compared by way of their electromechanical coupling coefficients k, a unit-less parameter between 0 and 1 (higher is better). Performance and cost comparisons of three modern materials, as used in 33-mode, are shown in Table 1. The 33 subscript refers to the mode where the dominant output strain is along the same axis as the applied electric field. (The 33-mode parameters are typically used as a power comparison because 33 is the most effective transmitting mode.) The 31-mode excitation is relevant to cylindrical transducer geometries. The “28” in the “Materials” column description indicates that the lead titanate (PT) is 28 percent of the mixture, by volume.
The cost comparison in the table, while showing nearly an order of magnitude lower cost, is approximate because the TM material has not yet been produced in a factory setting (only in laboratories). Nevertheless, the TM material exhibits performance comparable to that of a single crystal.
The scaling of laboratory processes to repeatable commercial production has yet to be accomplished. However, early indications are that combining the well known commercial tape casting process with TGG enables the production of high-performance flat plates and cylindrical shapes at a fraction of the cost of a single crystal. Textured material promises to be a significant part of the unfolding piezoelectric ceramic technology revolution.
1Sherman, Charles and Butler, Jack: Transducers and Arrays for Underwater Sound, 2007, Springer Science + Business Media, LLC.
2Oakley, Clyde and Zipparo, Michel: Single Crystal Piezoelectrics: A Revolutionary Development for Transducers, Proceedings of the IEEE Ultrasonic Symposium, October 2000, Pages 1157–1167.
3Zawilski, Kevin T. et al, Segregation during the vertical Bridgman growth of lead magnesium niobate–lead titanate single crystals, Journal of Crystal Growth, Volume 258, Issues 3–4, November 2003, Pages 353–367.
4Kristen H. Brosnan, Stephen F. Poterala, Richard J. Meyer, Scott Misture, and Gary L. Messing, Templated Grain Growth of <001> Textured PMN-28PT Using SrTiO3 Templates, J. Am. Ceram. Soc., 92 [S1] S133–S139 (2009).
Michael Janik, William Marshall