Imagine a young couple that has just “seen” their infant still in the womb for the first time after a routine ultrasound scan. What joy! This news is worth sharing, so they click an image of the monitor on their smartphones and send it to grandparents, family, friends, and their entire social media networks.
Prenatal ultrasound Credit: iStock / SanyaSM
Meanwhile, across the medical campus, a recent retiree undergoes a cardiac procedure guided with real-time ultrasound scans to repair a blockage discovered, also with ultrasound. What relief! This patient just found out he will be a grandfather and now confidently awaits that day with anticipation.
These routine activities of modern life are possible because of piezoelectric ultrasound transducers, many thousands of ferroelectric multilayer ceramic capacitors, and a host of other electronics, sensors, actuators, and devices based on ferroelectric materials.
Ferroelectrics have intrigued materials scientists since their first report in 1920 and first publication the following year.1 The label “ferroelectric” can be slightly misleading, as only a small fraction of ferroelectric compositions contain iron. Rather, the nomenclature was adopted in recognition of the parallels between the newly discovered phenomenon of ferroelectricity and the already-familiar ferromagnetism, which originally was thought to be tied exclusively to iron.
The defining characteristic of a ferroelectric material is the existence of spontaneous dipoles within its crystal structure whose direction can be reoriented by the application of an electric field. In other words, all materials belonging to a polar point group are potentially ferroelectric, but it takes a demonstrated polarization reversal for the material to earn the title.2 (See sidebar, “Genesis of hysteresis…” ). This strict definition highlights an important conundrum: a variety of phenomena can produce polarization-vs-field measurements that suggest ferroelectricity, but they are instead artifacts masquerading as polarization reversal.3 At the same time, only a small fraction of the applications enabled by ferroelectric materials directly take advantage of the switchable polarization. Instead, as summarized here, the tremendous utility of ferroelectric materials in modern life typically arises from phenomena associated with, but not necessarily directly leveraging, the full reorientation of this spontaneous dipole.
Discovery and early developments
In 1919, Joseph Valasek began his Ph.D. work at the University of Minnesota under professor W.F.G. Swan, who suggested that Valasek investigate the curious crystals of Rochelle salt (KNaC4H4O6∙4H2O). These crystals were relatively straightforward to grow and were already known to be piezoelectric, pyroelectric, and optically active. Annoyingly, they were also very sensitive to humidity, and nearly all of their interesting properties seemed to depend on everything, including temperature, electric field, and previous history.
Valasek’s first task was to develop sensitive measurements that could finally pin down the properties of this mischievous Rochelle salt. Valasek’s thesis and the associated publications are a veritable treatise on crystal physics and careful measurements. It is interesting to note that of Valasek’s five papers in Physical Review on Rochelle salt, he is the sole author on four of them, and the seminal paper1 reporting the first ever ferroelectric hysteresis loop received just over 200 citations in the ensuing century. The scientific enterprise certainly has changed!
Reading Valasek’s papers, it is clear he imagined this new phenomenon would lead to new functionality and new devices, but he could scarcely have predicted how his careful measurements of nonlinear dielectric response of finicky Rochelle salt crystals would end up playing such a fundamental role in so many technologies a century later. Figure 1 presents an approximate timeline of the development of ferroelectric materials from prediction in 1912 to discovery in 1920 to the first ceramic transducer in 1945 through to the present time’s search for eco-friendly compositions. This history, juxtaposed against global advances at the time, gives interesting context to the significance of ferroelectric devices to the evolution of digital technology.
Figure 1. Timeline presenting development of ferroelectric materials alongside global advances at the time.
One thing that has not changed significantly since Valasek’s time is that much of the funding devoted to development of new materials is driven by potential military applications. In Valasek’s case, the interest was in detection of submarines. To this day, needs for improved materials for sonar remain a strong driver of ferroelectric and piezoelectric materials development. In fact, World War II spurred development of BaTiO3, Pb(Zr,Ti)O3, and Bi4Ti3O12 based ferroelectrics roughly simultaneously in isolated groups in the United States, Japan, and Russia. This development, discussed in more detail below, is an excellent example of how the phenomenon of ferroelectricity enables otherwise unachievable performance even when the switchable polarization itself is only indirectly related to the application.
In the century since Valasek first discovered the ferroelectric effect, enormous research effort went into understanding these extraordinary materials and how to make them better, controlling their properties precisely, shrinking the size of components, and deploying them in new and novel applications. Much of that research was—and continues to be—reported in ACerS publications. The sidebar “Historically significant ferroelectrics papers” highlights key papers over the years that advanced the science and art that led to the implementation of ferroelectric components in devices we enjoy today.
It is worth noting that ferroelectric polymers, most notably polyvinylidene fluoride, do exist and find significant application. However, in this article we focus on inorganic, nonmetallic (i.e., ceramic) ferroelectrics, which are both more numerous and more widely used than their squishier counterparts.
Fundamentals of ferroelectric ceramics
One of the most significant characteristics of ferroelectric ceramics is their ability to exhibit single crystal-like behavior even when fabricated as polycrystalline ceramics. In addition to typically being significantly less costly and easier to produce than single crystals, fabrication of polycrystalline ceramics also introduces opportunities for microstructure engineering and formation of complex shapes. Use of polycrystalline ceramics as piezoelectrics depends upon the ability to break full isotropic symmetry through the application of a poling field and thus relies on extrinsic effects (domain wall motion) in ferroelectrics. The sidebar “What does symmetry have to do with it?” reviews the importance of crystallographic symmetry to piezoelectricity and ferroelectricity.
Intrinsic piezoelectricity is linear under small fields, so applying a +5 V field in one direction will produce equal and opposite strain to applying –5 V in the same direction. In a typical sintered polycrystalline ceramic, even if each of the billions of individual grains is piezoelectric, they all collectively cancel out their neighbors, and the net result is zero macroscopic piezoelectric response. This fact is why quartz piezoelectrics, such as those used for timekeeping in watches, must be single crystals.
Therefore, to use polycrystalline materials as piezoelectrics, the macroscopic random symmetry of the grains must be broken in some way. One option is to force the grains to all (or at least mostly) grow in a coordinated direction. This approach is used for AlN thin films in MEMS resonators and clever ceramic processing approaches such as templated grain growth.4 Another option is to start with a crystallographically random polycrystalline ceramic sample and break the macroscopic symmetry in some way, such as with the application of a large electric field. The applied field aligns crystallographic dipoles, and this “poling” process is enabled by the field-induced alignment of spontaneous dipoles, in other words, ferroelectricity (Fig. 2).
Figure 2. Of the crystalline or polycrystalline materials that exhibit piezoelectric behavior, some are pyroelectric, and of those, some are ferroelectric. Materials are classified into these categories based on the symmetry of their unit cell, which in turn dictates how they interact with thermal, electrical, and mechanical energies. Credit: Sherbondy
The outstanding properties of ferroelectric materials for piezoelectric and charge storage applications arise from a combination of intrinsic and extrinsic contributions. Ceramics based on barium titanate (BaTiO3) dominate the capacitor market while lead zirconate titanate (Pb(Zr,Ti)O3, PZT) based ceramics are the most widely used piezoelectric ceramic for more than six decades, finding applications in medical ultrasound transducers, sonar, micropositioners, and more. These materials are cubic and thus nonferroelectric at temperatures above TC (the Curie temperature) and transform to a lower symmetry polar state below this temperature. Figure 3 shows how the ferroelectric phase transition occurs in the prototypic ferroelectric, BaTiO3 (Fig. 3).
Figure 3. During the phase transition that results upon cooling through the Curie temperature, the BaTiO3 unit cell expands by roughly 1% in the polar direction while shrinking accordingly in the two perpendicular directions. Displacement of the cation species (the Ba2+ and Ti4+) and anion species (O2-) in opposing directions results in a polar (tetragonal) crystal structure and the presence of a small electrical dipole, a spontaneous polarization, within each unit cell. While the dimensional changes associated with this phase transition are relatively small (about 0.1 Å), the lower symmetry is the key to enabling ferroelectricity and its derivative properties. Credit: Geoff Brennecka
The change in shape and the resulting spontaneous polarization along the (by definition) c-axis occurs in every unit cell in the sample, and of course each unit cell is influenced by those around it. Having all of the unit cells transition in the same direction would be both a tremendous coincidence and a higher energy state than having some disorder. Thermodynamic free energy minimization drives the formation of domains to reduce both elastic (strain) and electrostatic (charge) energies. Domain sizes and distributions depend on boundary conditions and defects, but the ability to configure and reconfigure these domains by applying electric fields is both the defining property of ferroelectrics and the source of extrinsic contributions that often dominate the properties.
It is worth noting that in recent decades, researchers have found that electrically driving single crystal piezoelectrics along nonpolar directions can produce very large extrinsic contributions to electromechanical strain, taking advantage of similar mechanisms to those described here for polycrystalline ceramics.
Some of the same factors that enable large piezoelectric responses in ferroelectrics also contribute to large permittivities, which is of great importance for charge storage in capacitors. Relative permittivity (also referred to as dielectric constant, though it is anything but constant in the materials discussed here) is proportional to the change in polarization with an applied electric field. When a small electric field is applied to a ferroelectric, the intrinsic response discussed above results in a change in polarization that is typically much larger than in linear dielectrics, such as silica or alumina. This change in polarization leads to a large relative permittivity and a high volumetric capacitance. Additional extrinsic effects, small motions of domain walls, lead to even larger changes in polarization and even higher relative permittivities. These attributes make ferroelectrics well-suited for many capacitor applications. For example, the relative permittivity of BaTiO3-based dielectrics in the multilayer ceramic capacitors (MLCC, Fig. 4) that are present in our cell phones, computers, and other electronics, is greater (often much greater) than 1,000. For comparison, silica has a relative permittivity value of just 3.9.
Figure 4. Schematic structure of a multilayer ceramic capacitor. Layers are (1) dielectric ceramic, (2) outer ceramic layer, (3) electrode, and (4) contact surface. Credit: Wikimedia, Jens Both Elcap, CC BY-SA 3.0
Abundant applications
Sophisticated ceramic processing and microstructure engineering enabled scaling of ferroelectric-containing MLCCs, such that today our electronics operate with dielectric layer thicknesses far less than one micron packaged into capacitors with more than 100 active layers. Packaged MLCC devices today can be many times smaller than a grain of salt. In fact, MLCC scaling over the past 50 years rivaled—and often outpaced—the aggressive transistor scaling of the semiconductor industry known as Moore’s Law.
As discussed in the sidebar “What does symmetry…”, polar crystals exhibit the pyroelectric effect, or a change in polarization magnitude with a change in temperature. While all polar materials possess this property, ferroelectrics may have pyroelectric coefficients that are much larger than their nonferroelectric counterparts, often in the proximity of a ferroelectric phase transition. This property makes ferroelectric pyroelectrics particularly well-suited for thermal sensors. A wide variety of ferroelectrics are used for pyroelectric sensors, but some of the most common are single crystal LiNbO3 or LiTaO3 owing to their combination of high pyroelectric coefficient, low losses, and low relative permittivity, which increases the measurable voltage for a given amount of charge produced.
LiNbO3 and LiTaO3 single crystals also find tremendous use in the optics sector, often in the form of periodically poled lithium niobate (PPLN) and periodically poled lithium tantalate (PPLT), structures for waveguides, phase matching, difference frequency generation, and many others. This periodic poling takes advantage of the nearly strain-free 180-degree domain walls in these crystals as well as their large coercive fields. In fact, the coercive fields of as-grown LiNbO3 and LiTaO3 crystals at room temperature are often very close to the breakdown strengths of the crystals, so the pure materials are often poled as they are pulled from a melt or at temperatures approaching their Curie temperature where the coercive fields are lower. Doping with magnesium reduces the coercive field and increases the laser damage threshold of LiNbO3 and LiTaO3 by compensating for the lithium deficiency inherent in melt-grown LiNbO3 and LiTaO3 crystals. Creating an array of inverted domains with periodicity related to the wavelength of an incident laser beam enables a variety of clever manipulations of the beam as it passes through the PPLN or PPLT crystal, and the high coercive fields ensure that once the material is poled, it will stay that way during operation, even under large optical powers and occasionally large applied electric fields.
One technology that utilizes the full switching of polarization, the very hallmark of ferroelectricity, is ferroelectric random access memory, FeRAM. In this technology, the polarization of the ferroelectric represents a binary bit of information (i.e., 1 or 0). When coupled in series with a transistor, the amount of current flowing through the channel when the transistor is activated allows differentiation of the positive or negative polarization (memory) state of the ferroelectric capacitor. A voltage pulse is then used to set the polarization of the ferroelectric for the next stored bit. The result is a nonvolatile memory that can maintain its state for timescales up to 45 years.5 FeRAM has advantages over other memory technologies in terms of the number of read/write cycles possible and the energy required for each switching event, though its physical scaling into the tens of nanometer range remains a significant challenge.
While ferroelectric and related materials dominate the high strain piezoelectric technologies of sonar, ultrasound, and micro- and nano-positioners, these technologies rely on bulk ceramics or single crystals fabricated into specific geometries for best performance. To affect technologies at smaller size scales, such as millimeter-scale robotics, RF switches, and actuators for inkjet printers, the device sizes must also decrease. This requirement has driven significant efforts into development of ferroelectric piezoelectric thin films and micro-electro-mechanical systems (MEMS) devices that can perform at reduced dimensions—a field known as piezoMEMS. Figure 5 shows an example piezoMEMS device designed to imitate dragonfly flight.6 The most broadly studied material for piezoMEMS is PZT. In approximately 30 years of piezoMEMS study, the piezoelectric performance of PZT thin films substantially improved through advances in processing and understanding of the roles of mechanical boundary conditions.
Figure 5. U.S. Army Research Laboratory’s prototype PZT piezoMEMS dragonfly with stroke and pitch actuated wing design. The image on the right shows a platform suspended on the dragonfly between individually controllable wings.7Credit: U.S. Army Research Laboratory
While PZT and other lead-based ferroelectric piezoelectrics demonstrate the most utility for electro-mechanical applications, the looming European Union Restriction on Hazardous Substances (ROHS) requirements drive the search for reduced lead in materials and has also led to the study of lead-free piezoelectric ceramics and films. The piezoelectric coefficients of the lead-free materials, such as (K,Nb)NbO3 and Bi0.5Na0.5TiO3-Bi0.5K0.5TiO3 to date lag behind those of lead-based systems and represent a large area of current and future study and growth.
Frontiers of ferroelectricity
In the first 90 years of ferroelectrics research, attention focused on a few primary crystal structures and materials populating those structures at the forefront for commercial applications. These include the perovskite ferroelectrics LiNbO3 and LiTaO3 described previously; layered structures, such as SrBi2Ta2O9; and dihydrogen phosphates, such as KH2PO4. However, in the past 10 years, two new structural classes have joined the ferroelectric family and bring with them some potential technological superiority to traditional materials on the basis of chemical compatibility with dominant semiconductor technologies. These include fluorite-structured HfO2 and wurzite-structured (Al,Sc)N.
In 2011, Boscke et al. first reported ferroelectricity in fluorite-structured silicon-doped HfO2 thin films.7 The observation of a switchable polarization in this material surprised the community and generated a great deal of excitement. This excitement was driven largely by the inherent silicon compatibility of hafnia and the fact that the first observation of ferroelectric response in HfO2 occurred in films that were only 10 nm thick just shortly after HfO2-based gate dielectrics emerged in commercial integrated circuits. HfO2-based ferroelectrics are poised to enable scaling of existing ferroelectrics-based technologies, such as FeRAM, to even smaller dimensions. In addition to demonstrated and commercialized ferroelectric thin film technologies, the integrated circuit process compatibility of HfO2 positions it to enable other new devices, such as negative differential capacitance field effect transistors (NC-FETs) and ferroelectric-FETs, which may offer continued performance increases to silicon based integrated circuits.
All of the commercially relevant ferroelectric ceramics discussed so far are oxides, but recent efforts have identified promising nitride ferroelectrics. Primary among them is (Al,Sc)N.8 MEMS resonators based on AlN thin films dominate the wireless communications market since the early 2000s, and in recent years, such resonators showed improved piezoelectric response via alloying with ScN or other transition metal nitrides. Increasing the concentration of scandium in (Al,Sc)N reduces the c/a ratio of the polar AlN wurtzite structure, gradually pushing the unit cell closer to the nonpolar hexagonal BN-type structure in which the cation sits in the same plane as the anions. At sufficiently high scandium contents, the amount of electric field required to switch the polarity of this structure can be less than the breakdown strength of the sample, and the material is ferroelectric. In large part due to the size of the MEMS industry already established around AlN thin films for radio frequency communications, ferroelectric (Al,Sc)N films and their derivatives have attracted tremendous attention for a number of integrated devices, but the enormous driving force for phase separation makes reliable fabrication rather challenging.
Recent predictions suggest ferroelectric behavior in nitride perovskites, including LaWN3.9 Calculations point to a polar structure with a sufficiently low energy barrier between anti-aligned polarities to make it a strong candidate for ferroelectricity,10 but these await experimental confirmation. Similarly, ferroelectricity reported in several of hybrid halide perovskites took the photovoltaic world by storm, though the veracity of the genuine ferroelectric nature and the potential role(s) of the spontaneous polarization on the properties of these materials remains controversial.
We will not pretend to know exactly what lies ahead for the second century of ferroelectricity, but it is a safe bet that ferroelectrics will continue to play the role of enabler, quietly operating behind the scenes, facilitating critical functions and systems.
Read more: “Genesis of hysteresis in ferroelectric materials“
Read more: “Historically significant ferroelectrics papers“
Read more: “What does symmetry have to do with it?“
Capsule summary
A salty start
University of Minnesota graduate student Joseph Valasek gave the first presentation on ferroelectricity in Rochelle salt in 1920. He could scarcely have predicted how his careful measurements would end up playing a fundamental role in many of today’s technologies.
Versatile uses
The switchable spontaneous polarization that defines ferroelectricity directly enables applications such as nonvolatile ferroelectric random access memory, but many other applications, such as multilayer ceramic capacitors, are enabled by ferroelectric materials even without directly using the switchable polarization.
Future ferro
In the past 10 years, two new structural classes—including fluorite-structured HfO2 and wurzite-structured (Al,Sc)N—joined the ferroelectric family and bring with them some potential technological superiority to traditional materials.
Related Articles
Research Insights
Review of “Encyclopedia of Glass Science, Technology, History, and Culture”
The two-volume Encyclopedia of Glass Science, Technology, History, and Culture, edited by Pascal Richet, is a rich compendium of information spanning a vast number of aspects of the use of glass as a material. The two volumes cover essential information on the structure of glass and its industrial applications and…
Market Insights
Lessons learned from the rollout of 5G technology help guide 6G preparations
While it may seem like the 5G communications and data network has just been rolled out—and is, in fact, still in the process of being deployed—communications providers, governments, and researchers are already preparing for the sixth generation of wireless technology, or 6G. Compared to 5G, which aimed to expand the…
Bulletin Features
The synthetic data revolution: How AI is redefining quality control in ceramics manufacturing
The ceramics industry stands at the brink of a fundamental shift in how quality control is conceived, implemented, and scaled. For decades, the promise of artificial intelligence-powered inspection systems was held back by a persistent obstacle: the scarcity of high-quality data. This bottleneck made it difficult for ceramics manufacturers to…