In IRG1, a strong team of theoretical and experimental physicists and materials scientists are exploring the properties of a class of materials called complex oxides that have the potential to spawn new technologies in sensing, computing, and superconductivity. Complex oxides are an abundant group of materials that are composed of at least one oxygen atom and one other element. Their interesting properties can include high temperature superconductivity, ferroelectricity, and ferromagnetism. The goal of IRG1 is to create new science by coupling ferromagnetism, ferroelectricity, and ferroelasticity in a single material at room temperature.
The most familiar of the ferroic properties is ferromagnetism, the basis of the magnetic compass, which was studied by the Greek philosopher Thales as early as the 6th century B.C. Magnetism is the result of the quantum spin of electrons. In a ferromagnetic material, the magnetic field can point in only one of two directions, called spin up or spin down. When all of the spins are pointing in the same direction, the result is a strong magnetic field. The ability to flip the spin direction of a ferromagnetic material using an external magnetic field is the basis for the memory in a computer hard drive. Early in the 20th century, materials with a spontaneous electrical polarization were discovered. These ferroelectrics are crystalline materials with positive and negative poles aligned in a single direction. Like ferromagnetism, the direction of the alignment can be flipped using an external electric field. Ferroelectrics are used in nonvolatile random access memory, capacitors, sensors, night vision devices, and precision actuators for electron microscopy, to name a few applications.
Magnetism and electrical polarization are rarely found in a single material. On the occasions when they are found together, both are typically weak. However, theory predicted that strong ferroelectric ferromagnetic materials could exist and recently such materials have been created in the lab. The past decade has seen an explosion of the synthesis and characterization of complex oxide multiferroics combining two or more ferroic properties in a single material system.
Penn State has been a leader in the science of ferroic materials for more than fifty years. The Center for Dielectric Studies has been funded continuously by the National Science Foundation and industry since 1983, and noted experts such as Robert Newnham and Eric Cross made pioneering advances in sonar and in ultrasound for medical purposes based on ferroic materials in the 1950s and ‘60s. Multiferroics research at Penn State builds on their achievements.
Penn State’s work on multiferroics originated about a decade ago when the National Science Foundation started a program called Nanoscale Interdisciplinary Research Teams (NIRTs), IRG1 leader Venkatraman Gopalan explains. His colleague, Darrell Schlom, was an expert in molecular beam epitaxy. The high vacuum chamber of Schlom’s device allowed him to lay down vapors of atomic elements one layer at a time. "He called it atomic spray painting," says Gopalan. Schlom gathered a group of experts that included the optics expert Gopalan, a theorist, Long-Qing Chen, a physicist, Xiaoxing Xi, a ferroelectrics expert, Susan Trolier-McKinstry, and a number of experts from other universities across the country. This core group formed the NIRT team, which was funded for two four-year cycles by the NSF, during which time they made a number of fundamental discoveries.
"One of the great discoveries we made was an idea that Darrell had about inducing strain in some well-known oxide materials to see if it would induce new properties," Gopalan recalls. Strain was known to affect the mobility of electrons in semiconductors, and strain was being used in making transistors. Schlom thought something similar might occur in oxides.
However, oxides are brittle materials and compressing or stretching them causes cracks. To overcome this obstacle, Schlom developed a series of single crystal substrates made of rare earth materials that produced fine textured templates. By precisely depositing a thin-film layer of oxide with a near matching template, he could coax the thin film to stretch or compress the small amount needed to exactly fit the substrate lattice structure without cracking. One of the materials he used to coat his substrate is a well-known oxide material, strontium titanate. In the world of ferroics, strontium titanate is a bland material, with no magnetism and only a small amount of ferroelectricity at low temperatures. But when he grew it on a substrate with a little strain – 1or 2 percent – he found that the material developed a strong polarization. "He found that a property that does not exist in this material – ferroelectricity – suddenly pops up," Gopalan says.
In another much-studied oxide, barium titanate, the ferroelectric properties that were already present were enhanced three-fold by strain. Strain was shown to be an important factor in inducing the new properties, properties that could subsequently be used in new devices. This work resulted in papers in both Nature and Science that received a great deal of attention, Gopalan says. As the NIRT funding came to an end, the Penn State MRSEC was undergoing its own renewal process for a third funding cycle. In preparation, the MRSEC put out a proposal for a new IRG to expand the Center for Nanoscale Science. Gopalan suggested a theme to his colleagues, which was to go beyond ferroelectricity and see if they could develop oxides with strong ferroelectricity and ferromagnetism that could actually couple the two effects, in short get the two fields to talk to one another and flip each other’s switches.
At this point, interest in multiferroics was beginning to heat up. Nicola Spaldin, a theorist then at UC Santa Barbara had reignited the field by predicting a number of new materials with both properties. Spaldin joined the team, which had now grown to around a dozen researchers, about equally divided between Penn State and half a dozen other universities.
Gopalan thought they could use the strain techniques Schlom had developed to induce ferromagnetism and ferroelectricity. The Penn State MRSEC provided the team with a year’s worth of seed funding to hold them together while they waited to see if the National Science Foundation would fund the new IRG. "At that time we didn’t have a clear idea if we could make these materials. Those two properties, ferroelectricity and ferromagnetism, don’t easily coexist," says Gopalan. "What we did have were all these oxides, these accomplishments, and the theory predictions by our collaborators." In 2008, the funding from NSF came through and the newest IRG was officially in existence under the title Strain-enabled Multiferroics.
Their history of collaboration and their wide array of skills gave the IRG1 team a head start. From the beginning the multitalented team embodied the National Science Foundation’s purpose in the creation of MRSECs, which is to solve problems that no individual investigator would be able to tackle on his or her own. One group provided theory, another group grew thin films, and other members measured the resultant properties and looked at their structure using various microscopy techniques.
Gopalan explains why this IRG is successful: "The idea of these NSF teams has conventionally been that people come from one campus, maybe from different departments, but with the core always from one institution. But from the beginning in the NIRT days, we have had half from Penn State and half from other universities. We didn’t want to change that composition because we’ve worked together so well." In the four years of their existence to-date, IRG1 has published around 130 research papers, the majority of them coauthored or with multiple IRG authors, an indication of both their productivity and their collaborative approach. When Schlom moved from Penn State to Cornell, he remained an integral part of the team, as did Xiaoxing Xi, now at Temple University. Roman Engel-Herbert, a recent faculty hire in materials science and engineering, brought new molecular beam epitaxy capabilities to Penn State.
"There are maybe four or five different mechanisms where you can couple magnetism to ferroelectricity," Gopalan mentions. "Out of those five mechanisms, four have come out of our IRG. It’s a new field, but that shows the impact of this IRG."
Although the team excelled in making and measuring atomically thin oxide materials, they were equally good in theoretical predictions. Three of the IRG members – Karen Rabe at Rutgers, Nicola Spaldin, now at ETH Zurich, and Craig Fennie at Cornell, specialize in first-principles calculations, basically building up models of materials an atom at a time. Long-Qing Chen’s group at Penn State uses phase-field techniques to model energy expressions on a different length scale.
Fennie, who had been a student in Karen Rabe’s group during the NIRT research, brought some exciting new ideas to the IRG. One of these was the prediction that there could be a coupling between electron spins and phonons, which are bundles of vibration energy in the lattice structure of crystals. The phonons, he believed, vibrate in such a way that they polarize the material. The vibrations are also coupling to the spins and making them line up to make a ferromagnet. He predicted that at its lowest energy state – the ground state – if the material were strained with either compression or tension, the phonons would get stuck and stop vibrating. This would freeze the material into a multiferroic state with a strong polarization and a strong magnetization. Gopalan compares it to a spring that loses tension.
They went to work to see if they could prove Fennie’s theory with experiment. The material they chose to work with is called europium titanate, a bland oxide that is both paraelectric – not electric – and antiferromagnetic – the spins line up in opposite directions, cancelling out magnetism. In other words, it had neither of the properties the group was looking for.
The material was not easy to grow in the exact state they were looking for, with the europium oxidation state of 2 plus. Finding the right crystal with the right strain on the right substrate without too many defects took Schlom and company three and a half years of effort. Finally their efforts bore fruit. Their new strained europium titanate showed exactly the properties Fennie had predicted. As they cooled the material down, it first became ferroelectric and then as it was cooled below 4 kelvin, strongly ferromagnetic. The europium titanate thin film outperformed the best ferroelectric ferromagnetic material by a factor of 1,000. Strain was the key for turning on two phenomena in a material for the first time.
Not many laptop users would be comfortable working on a machine that only functioned at four degrees above absolute zero. The goal, which no one has reached yet, is to build a material that can be harnessed to flip electron spins and magnetic moments to create low-power magnetic memory devices that operate at room temperature.
The theorists have come up with three other mechanisms - they call them knobs - that can be used to turn on polarization and magnetism. The first of these knobs involves laying down layers of oxides with a particular arrangement of atoms in what is called a perovskite structure, of which barium titanate is probably the best known example. This layer could be followed by a layer of rock salt and then another layer of perovskite material, such as strontium oxide. Because they can deposit the layers with atomic precision, there are multitudes of possible combinations and thicknesses.
"Layering of structures is one of the ways we are moving forward," Gopalan says. "And they are not just theoretical. We are studying them right now in my lab." Learning what these composite materials can do could lead to new discoveries in materials science that translate into advances in technology.
Another knob relates to oxygen cages. The cages are crystals with oxygen atoms that sit in the center of the six faces of an octahedral crystal. It is possible, Gopalan explains, to literally stuff other atoms inside the cages. Depending on what size atoms are stuffed into the cages, and the nearby atoms on the outside, the crystal will twist and turn to try to seek its lowest energy level. The twisting and turning of the crystal is another knob to couple electricity to magnetism.
Yet another knob is similar to the earlier layering mechanism, but with layers so thin that two different materials are so interconnected at an atomic scale that they become a new material. It is as if the layers have so little material that they do not know their own identity. This knob might make it possible to tune the ferroic properties at a much higher temperature.
On final knob on their tuner is called gradients. Apply strain on one end of a material but not the other, for example, and the resulting gradient of strain leads to new phenomena. Very little work has been done in this area- Eric Cross’s pioneering research at Penn State being the rare exception. But gradients appear to be a promising route to controlling properties. Additional gradient effects include electric field gradients, magnetic field gradients, and compositional gradients.
All of these knobs for fine tuning the properties of complex oxides are available with the precise control of atomic deposition. IRG1’s goal of creating room temperature ferroelectric ferromagnets might or might not one day spin-out new technologies, but it has already added to our store of fundamental knowledge.