UVA Professor Jon Ihlefeld combines materials and electrical engineering, fueling innovation in both fields.

Materials scientists and electrical engineers have divergent yet complementary viewpoints. Electrical engineers tend to see a material as a given, as a substance with a specific set of properties they can harness to make a more efficient, more powerful semiconductor device. They are perpetually on the lookout for materials that expand their options for innovation.

Materials scientists, on the other hand, see a material as something that is potentially malleable. By uncovering the links between processing methods and material structure and between material structure and properties, they gain the insights needed to design advanced materials with a specific set of characteristics. In effect, innovation in materials science fuels innovation in electrical engineering.

Associate Professor Jon Ihlefeld exemplifies the relationship that materials scientists have with electrical engineers. At UVA, he holds a primary appointment in the Materials Science and Engineering Department with a joint appointment in the Charles L. Brown Department of Electrical and Computer Engineering, and he is an integral part of UVA Engineering’s Multifunctional Materials Integration initiative, helping colleagues in electrical and computer engineering achieve breakthroughs in the performance, functionality and application of engineered systems by combining disparate materials and technologies.

This is a role that Ihlefeld is comfortable with. He was recruited from Sandia National Laboratories in New Mexico where, “My job was to be the materials scientist who interfaced with electrical engineers and physicists to develop new devices.”

Ihlefeld is a specialist in oxides, materials containing oxygen in their chemical makeup. When his colleagues at Sandia talked about oxides, he noted they instinctively thought of silicon dioxide, which had been universally used as an insulator in microelectronics. “My role was to point out that they had the whole periodic table from which to choose oxides, each with its own functionalities and properties, and that I could make them.”

Proving Himself Wrong

At UVA Engineering, Ihlefeld is continuing his exploration, begun at Sandia, of oxides with properties that are both new and useful. His leading candidate is hafnium oxide, a material that is used as an insulator on computer chips. In 2011, a group of researchers in Germany showed that thin hafnium oxide films prepared under specific conditions seemed to demonstrate ferroelectric properties. Ihlefeld was initially skeptical.

“The early data was not entirely convincing,” he said. “I just didn’t buy it.”

The Germans’ announcement, however, was intriguing because ferroelectric materials can be used for computer memory. In a ferroelectric material, one surface has a positive charge, and the other a negative charge. If sufficient voltage or an electric field is applied, these charges reverse. This characteristic makes them an ideal medium for digital electronics, where information and instructions are expressed in binary code consisting of ones and zeros. In fact, Fujitsu and Texas Instruments produce ferroelectric memory for subway smart cards and other devices using a ferroelectric material called lead zirconate titanate.

Given the technological potential for the material, Ihlefeld overcame his doubts and gave the German claims a careful examination. His research, along with that of several other groups, confirmed that hafnium oxide, when made extremely thin and heat-treated under specific conditions, possesses outstanding ferroelectric properties.

“We made it and measured it,” he said. “It is real.”

Extending Moore’s Law for Another Generation

These results are significant because lead zirconate titanate possesses a number of drawbacks, foremost among them is that it contains lead. The European Union is pushing to remove lead from consumer products—and the current exemption lead zirconate titanate enjoys may not be extended indefinitely. Equally important, it is not directly compatible with silicon, the platform for nearly all electronics. Engineers must go to elaborate lengths to integrate it on a chip, limiting its applicability.

Hafnium oxide, on the other hand, has already been used with silicon, making it potentially possible to build memory directly on a processing chip. It also possesses ferroelectric properties at thicknesses much smaller than lead zirconate titanate, a critical characteristic in a field driven by miniaturization.

Another property of interest is that when switched from positive to negative polarity, it produces an effect known as negative differential capacitance. One of the impediments to continuation of Moore’s Law, the prediction that computer power effectively doubles approximately every two years, is the lower limit on voltage required to operate chips – a property related to “Boltzmann’s Tyranny,” a menacing name describing a fundamental lower limit of transistor switching power. Materials that produce negative differential capacitance act as a voltage amplifier. In an ultra-low-power chip, hafnium oxide could be used to increase the voltage at the transistor where it is needed and extend Moore’s Law for another generation.

Ihlefeld sees that his next challenge is to make ferroelectric hafnium oxide a viable commercial technology. A major challenge is that thin hafnium oxide films are not uniformly ferroelectric. This is the kind of challenge, however, that is perfect for a materials scientist.

“We are working on a strategy to control the microstructure,” Ihlefeld said. “We believe that this will enable us to produce the properties we want and that every device will work exactly the same across an entire microchip.”