Advanced Materials, Processing, Synthesis, and Characterization

Laser-Deposited Films Improved with Aid of Plume Pictures
New Approach To Thermoelectric Refrigeration
ORNL Technology Leading to Better Transistor for Computer Memory Chips
Better Crystalline Substrates Open New Markets

David Geohegan (left) and Alex Puretzky capture digitized snapshots of the vaporized plume of carbon species created during laser ablation of pyrolytic graphite to synthesize new materials. An intensified charge-coupled detector array camera is used in fundamental studies of the target vaporization, vapor transport, and deposition processes that occur during only a few microseconds. The studies are used to optimize vacuum growth of transparent films of amorphous carbon with the properties of diamond, or the growth of clusters in background gases to useful nanoparticles (e.g., buckyballs). Photograph by Tom Cerniglio

To produce energy and use it efficiently while protecting the environment, it’s important to have the right materials. ORNL’s multidisciplinary program in materials research and development (R&D) is one of the nation’s largest, as well as one of the most respected internationally. By integrating basic, applied, and industrial R&D, we’ve achieved many successes in synthesizing, processing, and characterizing materials through development of new techniques, modeling, and performance evaluation. These materials include ceramics and composites, metals and alloys, polymers, superconductors, and thin films. Specialized equipment and the special expertise of ORNL and outside researchers have come together in our unsurpassed materials characterization and processing user facilities supported by the Department of Energy. Users seek to link a material’s microstructure with its properties at the High Temperature Materials Laboratory and the Shared Research Equipment Program facility. Scientists use neutron scattering to probe the structure of materials at the Neutron Scattering Research Facilities at the High Flux Isotope Reactor. Other users at the reactor look for internal stresses in welds and other materials at the Residual Stress User Center. Experiments with implanting ions in materials to improve their properties are carried out at our Surface Modification and Characterization Research Center. And through our Metals Processing Laboratory User Center, we are working with representatives from metals processing and other U.S. industries to help them use less energy, produce less pollution in their processes, and, hence, save money and jobs while becoming more competitive in the world marketplace. Through all these user facilities, we are establishing broad partnerships with industry.

Laser-Deposited Films Improved with Aid of Plume Pictures

Stop-action photographs reveal how a plasma plume composed of laser-vaporized carbon ions, atoms, and clusters moves away from a graphite surface into vacuum (bottom) and into a low-pressure background gas (70-mTorr argon, top). The digital images of the plasma light are taken at different times after the vaporizing laser pulse and are assigned false colors to reveal different regions of emission for spectroscopic identification. Expanding into vacuum (bottom sequence), the bright ball of fast carbon ions responsible for high-quality amorphous diamond films is the dominant plume component, while slower atoms and clusters are barely visible. However, the top image sequence shows how a low-pressure gas scatters and slows these ions, confining the plume so that carbon-carbon collisions can form clusters, such as buckyballs, near the target surface for fabrication of novel materials in the gas phase.

By imaging the plume in pulsed laser
deposition, researchers can influence
the quality of the thin-film product.

Just as solar rays turn comet ice into a blazing, cloudlike tail, pulsed beams of ultraviolet laser light can rapidly heat and vaporize a solid target in a vacuum chamber. But in pulsed laser deposition, the vaporized material travels in a luminescent plume toward a heated substrate where the material is deposited. The plume is a dense plasma of energetic electrons and neutral atoms, ions, and molecules in both excited and ground states. The plume’s contents are laid down in the same composition as the starting polycrystalline target material, but in the form of a film whose crystalline structure is locked into place by the substrate’s atomic template. Although the first pulsed laser deposition experiment was conducted in 1965, the technique has been employed widely only since 1987, when it was discovered that it can be used to grow high-temperature superconducting films.

ORNL researchers have taken high-tech snapshots of many a plume on its trip from target to substrate. They have used an assortment of sophisticated tools to get the picture. These include a fast-intensified charge-coupled detector camera to acquire images of the fluorescent plasma; optical emission and absorption spectroscopies that differentiate among the plume’s atomic and molecular species and states; and an ion probe, which measures ion energy and plasma density. From these tools, scientists can learn how fast the plume is moving and how much material in different atomic forms is present in the plume’s different parts.

By monitoring the plume, ORNL researchers can predict whether the film being produced will be uniformly thick. They can determine which conditions—laser beam energy and wavelength, gas pressure, substrate temperature—have changed and how much any change should be corrected to optimize film quality. From the diagnostics, they have found that most of the plume blasts from target to substrate in 20 microseconds and that the fast plume material moves from 1 to 5 centimeters per microsecond.

How fast the material moves, where it goes, how much rebounds from the substrate, and what type of product is formed depend partly on whether the plume is in a vacuum or background gas. For example, if the target is graphite in a background gas like argon, all-carbon buckyballs can be produced. If graphite is zapped in a vacuum with a laser beam, the product could be an amorphous diamond film that emits electrons when subjected to a strong electric field. Because this film is a superhard, transparent field emitter which can excite phosphor coatings, one company is using laser ablation to make amorphous diamond films for flat panel displays; these devices are needed to make much less bulky television sets and computer monitors, similar to the flat screens of notebook computers. ORNL researchers are looking at ways to deposit amorphous diamond films that could emit electron beams in lieu of light to pattern silicon chips. In this way, more circuits can be packed on each chip, making a chip both smaller and smarter.

ORNL’s plume diagnostics have revealed just how different background gases and varying gas pressures slow down the plume, causing the material to cluster together. As a rule, the gas absorbs energy from collisions between the expanding plasma and gas molecules, slowing down many plasma species. As the gas pressure is increased, some of the rapidly expanding plasma escapes with few collisions, but most gets caught in a snowplowed pile of gas which acts like a shock wave. As a result, the plume splits into two or three components. ORNL researchers developed the first accurate model of plume splitting and were among the first groups to understand it.

Monitoring of many types of ablation plumes showed that clusters of bound atoms ejected from the target can sometimes ruin an amorphous diamond film, but that clusters created by collisions with a background gas—the “third component” of a split plume—can be useful in other ways. For example, in experiments involving doping zinc telluride films to learn how to make cheaper blue-light-emitting diodes, the nitrogen gas itself was incorporated in the film. When film growth was conducted using each of the three components of material observed with the plume pictures, researchers produced films that possess quite different electronic properties. The best films were obtained when the nitrogen pressure was high enough to kill the damaging superfast “first” component, but the nitrogen content was low enough to avoid formation of the clunky nanoparticles of the “third” plume component. ORNL researchers now are intentionally making larger clusters and nanoparticles at higher pressures, from supertough nanotubes to nanocrystallites (which emit light in electric fields), all of interest for flat panel displays.

In respect to laser deposition plumes, ORNL researchers don’t use film to get good pictures, but instead use pictures to get good films.

The research is supported by DOE’s Office of Energy Research, under the Division of Materials Sciences in the Office of Basic Energy Sciences.

New Approach To Thermoelectric Refrigeration

The most efficient refrigerators—the compressor-based kinds found in our kitchens—are noisy because they have moving parts. They are not, therefore, the best choice for chilling foods and beverages in nuclear submarines. Because subs must operate as quietly as possible, the needed cooling is provided by refrigerators using thermoelectric materials such as bismuth-telluride alloys. These thermoelectric materials are much less efficient than compressor-based refrigerators—they provide less cooling per watt of electricity consumed. But, although less economical, they are quiet, more reliable, and convenient. Some can reach temperatures as low as 160 K, or ­113°C. Thus, they are also used for beer coolers, water coolers, and spot cooling of electronic circuits on chips.

Thermoelectric materials conduct electricity well, but they conduct heat poorly. If these materials could be designed to be much more efficient, they could be used to make car air conditioners and home refrigerators, avoiding the need for refrigerants that threaten the ozone layer or strengthen the greenhouse effect. It might even be possible to design a thermoelectric material that can chill high-temperature superconducting cables to 77 K, replacing liquid nitrogen, the cheapest coolant that enables such wires to conduct electricity without resistance.

ORNL researchers have demonstrated
a novel approach to designing
thermoelectric materials that
could provide the required
cooling at much improved

ORNL researchers have demonstrated a novel approach to designing thermoelectric materials that could provide the required cooling at much improved efficiencies. This approach has been described in articles in Science (May 31, 1996) and Physics Today (March 1997).

Thermoelectric materials can provide electricity when heated, or they can provide cooling when conducting an electric current. In 1823 Thomas Seebeck observed that if two different metals are joined in a closed circuit and if the junction for the two materials is heated, an electric current will be produced—the basis for thermocouples and radioisotope thermoelectric generators in deep space probes. In 1838 Heinrich Lenz observed that, if electrical current is passed in one direction through the junction between dissimilar materials, cooling will result.

ORNL researchers have designed an intriguing thermoelectric material. Because it can be used in a temperature range of 800 to 900 K (500 to 600°C), this material shows promise for generating electricity from waste heat in power plants and chemical production facilities. The material is more efficient than currently used silicon-germanium alloys.

Because of the potential of the ORNL approach for developing novel materials for thermoelectric refrigeration, our researchers are working with Marlow Industries in a cooperative research and development agreement. Marlow manufactures thermoelectric cooling modules in which the two ends of the device are linked by two legs (made of semiconducting material) whose electrons (or positively charged electron holes—the vacant positions in a crystal left by the absence of electrons) carry heat. The negative terminal of a battery is connected to the end to be chilled, and the positive terminal is connected to the device’s other end. Because like charges repel, the battery’s electrons force the electrons in the legs away from the “cold” end, driving them and the heat they carry to the positively charged end. The cold end becomes colder, and refrigeration results.

In this filled skutterudite crystal structure, iron or cobalt atoms are represented by red spheres, antimony atoms by blue spheres, and rare-earth atoms (lanthanum or cerium) by yellow spheres. The “rattling,” or thermal vibration of the rare-earth atoms, is thought to be responsible for the low thermal conductivity of these compounds.

Using an idea suggested by a researcher now at Rensselaer Polytechnic Institute, ORNL researchers developed a new crystalline solid with the “filled skutterudite” structure. The word skutterudite is derived from the name of a Norwegian location that has minerals with this same cubic structure. When iron, cobalt, and antimony are melted together, the iron and cobalt atoms form cubic “cages,” many of which are filled with antimony atoms in groups of four. But the researchers changed the skutterudite structure by also filling the empty cages with rare-earth atoms (cerium or lanthanum), which are too small for the cages. As a result, while iron, cobalt, and antimony atoms conduct electricity, each rare-earth atom rattles around in its cage, scattering heat waves and making the material a very poor thermal conductor. This novel approach of trapping small atoms in large cages could make possible the design of more efficient thermoelectric materials for refrigeration, creating cooling modules whose only moving parts are the vibrating atoms inside.

The research is sponsored by DOE’s Office of Energy Research, Basic Energy Sciences, Division of Materials Sciences.

ORNL Technology Leading to Better Transistor for Computer Memory Chips

(AO)n(A´BO3)m : The above figure shows an enlarged panel of the initial interface between crystalline oxides and silicon. The enlarged panel (right) is an overlay depicting positions of silicon atoms, the repeated alkaline oxide (AO) planes and the first layers of the perovskite, A´BO3. The lattice image shows our series of oxides in which 4 atomic layers of BaSrO are at the interface between silicon and the perovskite, CaxSr1-xTiO3. This composite interface structure is commensurate to silicon and provides the basis for ferroelectric transistor development.

A faster transistor that doesn’t forget has been shown to be feasible, thanks to an ORNL materials technology. This next-generation ferroelectric, or ferro-gated, transistor will enable the electronic industry to pack more information in smart cards and memory circuits in future processor chips for portable computers—and speed up retrieval of this information.

A new transistor being developed will allow
computer memory chips to hold more
information and enable users to
read and write it faster, making
the device ideal for smart
cards and laptop

A smart card is a pocket-sized computer processor that could replace your bulky wallet. This invention for consumer convenience can be used as electronic cash, driver’s license, credit card, information source, ticket for transportation services, and personal identifier. A smart card consists of a receiver (to accept signals from a card reader requesting information such as a credit card number), power supply, microprocessor, memory chip, and antenna (to send out requested information). Ferro-gated transistors are expected to be used in future smart cards’ memory chips because, compared with today’s smart card electronics, they cram in more information and require much less power to get information in and out.

The ferro-gated transistor is a kind of “smart transistor” because of its retentive memory. It will be especially useful for dynamic random access memory (DRAM) in computers because it is nonvolatile—the information stored as a bit on such a transistor during operation won’t vanish when the computer’s power is turned off. In today’s computers, DRAM is volatile, so any application not saved to a hard disk drive or floppy disk is lost when the computer is shut down. A chip containing millions of the new transistors could act as the hard disk drive of a laptop computer and extend the lifetimes of laptop batteries. Memory chips made of these new transistors will be able to retain up to a million times more information than chips of the same size today.

The ferro-gated transistor is a metal-on-oxide field-effect transistor, a very small semiconducting device that consists of three metal electrodes and a silicon base. When a conventional field-effect transistor is turned on, electrons injected by a source electrode flow as a current through the silicon base for collection at a drain electrode. To turn the transistor off, a gate electrode between the other electrodes applies an electrical voltage to a dielectric film which pinches off the current by raising resistance in the silicon base. In this way, a transistor can function as an on-and-off switch or as a repository of a bit of information coded as a combination of 1’s and 0’s (the “on” transistors are each storing a 1 and the “off” transistors are each storing a 0).

The new ferro-gated transistor is different from the conventional field- effect transistor in one way. The dielectric film between the gate electrode and the silicon base in the new transistor is formed from barium titanate, rather than silicon oxide. ORNL researchers perfected a technique for depositing barium titanate as a crystalline film on silicon, made metal oxide silicon capacitors, and proved that the properties of barium titanate meet requirements of a ferro-gated transistor.

The crystal structure of barium titanate gives it desirable ferroelectric properties—in certain regions, positive and negative ions separate, setting up an internal electric field. Alternatively, such an electric field is maintained in silicon oxide by using external power. The internal electric field of barium titanate remains unchanged and requires no external power unless it is flipped. Depending on whether it’s up or down, the field either pulls up or pushes away electrical charges in the silicon substrate, facilitating or resisting the flow of electrical current.

The semipermanence of the barium titanate’s internal field gives the new transistor two advantages. First, information held by millions of “on” and “off” ferro-gated transistors is retained when the power is shut off. Second, chips containing such transistors require less electricity because power is needed only when it is necessary to flip the internal electric field of any transistor’s barium titanate film.

Because its dielectric constant is 250 times higher than that of silicon oxide, the barium titanate layer exerts a stronger influence on the transistor’s conductivity; thus, less area is required for the gate electrode, making it possible to put the source and drain electrodes closer together. As a result, the transistor can be made smaller, and because electrons would travel a shorter distance between the source and drain electrodes, the device is faster.

In today’s computers, two transistors and two power-hungry capacitors are required to store one bit. If ferro-gated transistors are used, only one transistor is needed per bit—a change in the “logic state” that greatly cuts power requirements. Memory chips made of these new transistors will be able to retain a thousand to a million times more information. For example, a one-centimeter-square chip today (about the size of a small shirt button) would hold 64,000 to 256,000 bytes, whereas a chip the same size containing ferro-gated transistors would hold one billion bytes (gigabyte) of information.

To make computers even faster in the future, memory circuits will be incorporated into the processor chip. The spinning hard drive will be a thing of the past. Today it takes hundreds of milliseconds to retrieve information from a gigabyte hard drive disk and to write information to it. If ferro-gated transistors are used for gigabyte memory circuits of processor chips, it will take only fractions of a nanosecond to read the information and a few hundred nanoseconds to write information for storage in future computers.

Rodney McKee received a NOVA Award for Technical Excellence from Lockheed Martin Corporation for this development. It was funded by ORNL’s Laboratory Directed Research and Development Program and by DOE’s Office of Energy Research, Basic Energy Sciences, Division of Materials Science.

Better Crystalline Substrates Open New Markets

An oriented substrate cut from a potassium tantalate crystal. Thin films can be grown on substrates like this, forming the heart of devices that depend on ferroelectric, waveguide, high-temperature superconducting, or superlattice materials.

They don’t peer into a crystal ball to predict the successful materials of tomorrow. But ORNL researchers have helped transform crystals into new technologically important materials such as substrates for the growth of epitaxial electro-optic and superconducting thin films.

Electronic and optical devices can be made of thin crystalline films deposited on and supported by a durable substrate. A substrate is an underlying template that lines up the crystals of a thin film grown on it so that it conducts electrons or light. The substrate should not react with the film and, to avoid film breakage, it should expand at the same rate as the film when both are heated.

ORNL and a Florida company learned how
to grow larger crystals for use as
substrates in advanced
electronic devices.

Few crystals can meet these criteria, but single-crystal magnesium oxide (MgO) is an excellent substrate for thin-film devices. However, for some devices, the typical MgO crystal has previously been too small. Through a cooperative research and development agreement, ORNL and Commercial Crystal Laboratories (Naples, Florida) developed new insights into the MgO crystal growth process, leading to the formulation of a new concept for the nucleation and growth of this material. Drawing on this concept, researchers can grow MgO crystals whose diameters are double or triple those previously producedjust the right size for customers needing larger substrates for developing new electronic and optical devices such as switches and modulators for light-based communication networks and all-optical computers. This new MgO crystal-growth technology will help U.S. high-technology firms avoid total reliance on foreign suppliers of MgO substrates.

Crystals of potassium tantalate/niobate grown at ORNL have also been shown to make excellent substrates because they are mechanically and chemically stable. By adding calcium and barium to potassium tantalate crystals, ORNL researcher Lynn Boatner found a way to make them either insulating or conducting. By adding niobium, he turned the crystal into a ferroelectric material whose separated positive and negative electrical charges can be switched when an electric field is applied. These new substrates are important because high-temperature superconducting films can be grown on them for applications such as magnetic sensors for medical, geological, and industrial applications; microwave components for radar and communication technologies; and ultrafast switches.

Potassium tantalate substrates can be produced as large wafers, reducing the fabrication cost of thin-film devices that are grown on the substrate material. Commercial Crystal Laboratories has developed improved ways to prepare epitaxial-quality surface finishes on these substrates and is currently marketing the material for commercial and research applications.

The development of these new substrates received an R&D 100 Award in 1996, and the technology for producing these crystals has been patented and licensed to Commercial Crystal Laboratories. The ORNL researchers, Boatner and Ron Feenstra, also received a 1997 Federal Laboratory Consortium Award for Excellence in Technology Transfer.

The commercial prospects for these ORNL-developed materials are not known, but it is clear that the collaborators have completed a gem of a project.

This research was funded by DOE’s Office of Energy Research, Basic Energy Sciences, and the Laboratory Technology Research Program.

Next article