Bob Hawsey, director of ORNL's
Superconductivity Technology Center, discusses the bright future
he sees for ORNL's work with industry to develop a
commercializable superconducting wire. On the computer screen is the
World Wide Web site that contains information on
the development of ORNL's superconducting wire
(http://www.ornl.gov/HTSC/htsc.html). Photograph
by Tom Cerniglio.
Electrons, like people, like to go with the flow. As they travel the electrical highway, they try to avoid electromagnetic roadblocks and get in the fast lane. Easing the flow of electron traffic has been a goal at ORNL since 1986 when high-temperature superconductivity was discovered. But, like electrons, ORNL researchers have encountered roadblocks. They have found it exceedingly difficult to develop a high-temperature superconducting wire. After much persistence, secrecy, team work, and just plain hard work over the past seven years, ORNL scientists have found a promising new route to forming a superconducting wire.
On April 10, 1996, at the spring meeting of the Materials Research Society in San Francisco, ORNL researchers rolled out a new superconductora tape 3 millimeters wide and 15 millimeters long, and it became a hot item. They announced the development of a process for making the backbone of superconducting wire using a pair of rollers, heat, and thin crystalline films. The rolling-assisted biaxial textured substrates (RABiTS) process generated excitement among researchers in the electrical industry because it represented a leap forward in the race to develop fabricable superconducting wire. RABiTS is faster and probably cheaper than the ion beamassisted deposition (IBAD) process perfected in 1995 by DOE's Los Alamos National Laboratory (LANL).
Considered a cornerstone technology in high-temperature superconducting wire, the ORNL process conditions the substrate, or template, upon which a superconductor can be formed. The substrate, which is made of nickel covered with thin buffer layers, provides the underlying foundation for the superconducting wire. The superconducting material, called YBCO for short (for yttrium-barium-copper oxide, or YBa2Cu3O7-x), is made of crystalline grains. Each grain is a homogeneous region in which atoms are lined up in fixed directions. In materials available before the RABiTS or IBAD processes were developed, the grains pointed in different directions, like cars at a very busy intersection. RABiTS gives the YBCO grains a consistent orientation, like cars whizzing by on an interstate highway, unlocking electronic gridlock. A high degree of grain alignment in all directions is the key to more efficient electrical current flow through the superconductor.
Efficient current flow is the Holy Grail of the electrical industry. Today electricity is delivered by copper wire. But atomic bumps in the unpaved copper road slow the flow, and electrons pay tolls in the form of energy loss. The copper's resistance causes the wire to heat up and dissipate energy. Superconductors avoid this wasteful process. Because superconductors have no resistance to the flow of electricity, electrons get a free ride in the interstate's fast lane.
| Our substrate process will pave the way for the future manufacture of the next generation of high-temperature superconducting wires. |
"Our substrate process will pave the way for the future manufacture of the next generation of high-temperature superconducting wires," says Robert Hawsey, director of ORNL's Superconductivity Technology Center. "These wires will be the first to conduct electricity efficiently in the presence of high magnetic fields at liquid nitrogen temperatures.
"If these wires can be made long enough and strong enough, they can be used in transmission cables, transformers, current limiters, and motors and generators—anywhere large amounts of electricity are produced, transmitted, or distributed to customers. They can be used in devices that contain magnets, such as particle accelerators for research and medical diagnostic machines. They may also be used in large motors used in heavy industry, such as paper or steel mills."
Hawsey knows all about the struggle to make a YBCO-based wire (the name YBCO inspires the question, "Why be so out-of-line?). "Before I became the center's director in 1991, many groups had tried in vain to put YBCO coatings that were superconducting on standard silver wire. We lost interest in YBCO after it was shown that bismuth-based wires could carry a lot of current in low magnetic fields."
A compound containing some of the YBCO ceramic was first produced in January 1987 by Paul Chu of the University of Houston and M. K. Wu at the University of Alabama in Huntsville. They had substituted yttrium for the lanthanum in lanthanum-barium-copper oxide, the first high-temperature superconductor, discovered in January 1986 by J.Georg Bednorz and K. Alex Müller of the IBM Zurich Research Laboratory in Switzerland.
If it's kept cold enough and if its atoms are properly aligned, YBCO can conduct electricity without resistance, unlike copper metal used in conventional conductors. As a superconductor, YBCO has two important properties. First, it can conduct electricity in the presence of a relatively strong magnetic field, making it potentially useful for motors, generators, and other devices that have magnets, such as accelerators and medical diagnostic machines. Second, it superconducts if chilled to 77 K by liquid nitrogen, making it a high-temperature superconductor. Alloys of niobium and titanium (NbTi) and of niobium and tin (Nb3Sn) conduct electricity with zero resistance when chilled to 4.2 K by liquid helium, making them low-temperature superconductors. High-temperature superconducting materials like YBCO are preferred because they are far more economical: liquid nitrogen is 50 times less expensive than helium (10 cents a liter instead of $5 a liter).
| So the big challenge has been to fabricate wires and tapes coated with properly aligned YBCO. After such coated conductors were made finally at ORNL, the next hurdle was to improve the YBCO coating's ability to carry large amounts of current. |
The problem with YBCO has been the difficulty of aligning its grains and incorporating it into a long flexible wire. It is possible to make small-area (<1 in.), high-quality YBCO films on single crystals, but they are too short for use as wire and easily break when bent like wire. It is also possible to make thick YBCO conductors, but these "bulk conductors" superconduct as well as films only if chilled to 20 K by liquid helium. So the big challenge has been to fabricate wires and tapes coated with properly aligned YBCO. After such coated conductors were made finally at ORNL, the next big hurdle on the road to a practical superconducting wire was to improve the YBCO coating's ability to carry large amounts of current (critical current density).
"Normally, YBCO's crystalline grains are randomly oriented, so we tried various ways to align them to get good current flow," Hawsey says. "In 1989 our researchers, working with Georgia Tech scientists, tried coating flat silver tapes and round silver wires with YBCO using a metal organic chemical vapor deposition process. We were not successful in making a good superconductor this way. At that time, a bismuth-based material called BSCCO looked more promising than YBCO for making superconducting cables."
BSCCO (pronounced "bisco") is shorthand for bismuth-strontium- calcium-copper oxide. Several domestic and foreign companies are now fabricating BSCCO tapes and wires in lengths of kilometers. They pack BSCCO in silver tubes and draw the tubes into wires or flatten them into tapes. "BSCCO performs well in underground transmission cables and transformers at temperatures of 65 to 77 K and in higher magnetic fields at 20 to 30 K," says Hawsey. "BSCCO's grains are platelike, so when the material is rolled, they tend to line up in one direction more easily than YBCO grains. But unlike YBCO at 77 K, BSCCO's performance drops like a rock in the presence of a magnetic field."
In 1990, ORNL researchers took a completely different approach to aligning grains in YBCO films. They first examined YBCO films deposited onto metal tapes by pulsed laser ablation. In initial work using silver sheets as substrates, they noticed that even metal sheets obtained from the stockroom possessed a certain useful degree of grain alignment. Metallurgists have known for many years that grain alignment, called "texture," is often introduced into metals when they are rolled to form thin sheets or foils. The ORNL researchers realized that, if they could intentionally roll highly textured metal foils and then transfer this alignment to a deposited YBCO film, they should be able to make high-current wires that could operate in high magnetic fields at 77 K.
The many steps needed to roll metal tapes and to deposit high-quality buffer layers and YBCO turned out to be much more challenging than researchers first envisioned. However, five years after they first began working on this alternative approach, YBCO was no longer on the back burner. Some 15 ORNL researchers from three different divisions, along with five students and postdoctoral researchers, had teamed to solve a series of problems. They found a way to roll nickel to make a flexible textured tape coated with YBCO and three buffer layers—palladium, cerium oxide, and yttrium-stabilized zirconia. The pattern on the nickel template forced the alignment of the crystalline grains of the buffer layers and the YBCO on top.
"In the summer of 1995," Hawsey says, "we had low critical current densities. So we figured out how to improve the architecture of the buffer layers after Christmas to raise the critical current density to 300,000 amperes per square centimeter from January through March 1996. By late April, we had simplified the architecture by eliminating the palladium and raised the current density to 710,000 amperes per square centimeter in zero applied field. On April 10 we announced our development of a process to make a practical superconducting wire at a Materials Research Society meeting and in a news release. We caught the Japanese by surprise with our announcement of RABiTS. That was pretty darn exciting."
| High-temperature superconducting wires, says Hawsey, offer the possibility of fabricating new electric power devices that are more compact, cost less to operate, and use less energy. |
Hawsey said that the development of a YBCO superconducting wire earlier at LANL was encouraging to the ORNL group. "They used a similar architecture and similar materials like cerium oxide, yttrium-stabilized zirconia, and nickel, except they use Hastelloy nickel alloy and we use rolled pure nickel. These material similarities gave us the confidence our materials would work."
How do the two processes differ? ORNL relies on a high degree of alignment of atoms in the substrate, which propagate through the buffer layers to the YBCO. LANL starts with Hastelloy nickel alloy that is not well aligned and deposits an yttrium-stabilized zirconia (YSZ) layer on it as a buffer. LANL uses IBAD to align the YSZ layer under the YBCO film. The ion beam selectively removes the YSZ atoms not aligned at certain orientations biaxially. This is the limiting step that slows up the process and makes it potentially more costly than the RABiTS process for roll-texturing nickel.
Los Alamos has made impressive improvements over the IBAD process invented in the late 1980s in Japan. So far, Los Alamos has achieved 1 million amps per square centimeter in zero applied field at 75 K. The Japanese achieved a half-million amps per square centimeter using IBAD. Los Alamos broke the million mark by careful tuning of the process and careful attention to detail for depositing buffer layers.
Why must critical current density be so high in superconductors? Says Hawsey: "A lot of supercurrent must flow in the superconducting YBCO film to make up for the fact that the substrate, which is just a building block, is not carrying any current. The YBCO superconductor in thickness represents only a couple percent of the actual wire. Unless the current density is extremely high in the YBCO, the tape or wire won't have the overall current density of 50,000 to 80,000 amps per square centimeter required in a conductor. Standard household wires typically carry less than 1000 amperes per square centimeter. The higher the current density, the smaller the wire can be. That's important because engineers want wires used in motors, transformer coils, and transmission lines to be as small as possible."
High-temperature superconducting wires, says Hawsey, offer the possibility of fabricating new electric power devices that are more compact, cost less to operate, and use less energy. U.S. electricity demand is expected to double by the year 2030, so energy-efficient systems are needed to reduce requirements for costly new power plants.
"What Oak Ridge has done," Hawsey says, "is to come up with an industrially scalable way to get the alignment needed to ensure efficient electrical current flow. However, a few technical issues must be addressed before this technique is commercialized. Industrial partners must help us address the technical challenges of developing an industrial process of reproducing a reliable long-length superconducting wire."
One challenge is to determine the best way of chilling the wires to make them superconducting. Should the wire be wrapped around a hollow tube containing liquid nitrogen? Or, should it be conductively cooled by putting a cryocooler in contact with part of the wire?
Can the length of the superconducting wire be extended from a few centimeters to a kilometer? "The longest textured nickel tape we have produced is about a meter in length," Hawsey says. "The question is, can a uniform texture be reproduced along a wire hundreds of meters long? Industrial partners such as the Westinghouse Science and Technology Center will help us address this technical issue."
Hawsey also expressed concerns about buffer layer continuity and mechanical integrity of the substrate. "Will the buffer layers consistently have the desirable mechanical properties when the wire is wrapped around a tube or made into a coil?" he wonders. "Will superconducting properties hold up under the strain experienced by film wrapped around a tube? What happens to the superconducting properties if the wire is chilled to cryogenic temperatures 100 times? If a RABiTS superconductor is coated with a passivation layer, can it sit on a shelf in East Tennessee humidity all summer and still function when built into a device?"
The fact that nickel is magnetic may also be a problem. "The electrical energy in the superconductor could dissipate in the presence of an alternating magnetic field, causing unacceptable alternating-current losses," Hawsey says. "Also, if you wind a magnet with nickel wire, the magnetic nickel may distort the magnetic field of the magnet by sucking in the flux lines. Such an effect could make it difficult to predict the shape of the magnetic field, which may be essential in magnetic resonance imaging (MRI) and accelerator magnets. But, if this turns out to be a problem, it may be overcome by good engineering."
There are also economic issues. Can a substrate be produced that is economically attractive to industry? How will the cost of making a RABiTS wire compare with the cost of fabricating a wire using the Los Alamos IBAD process or the current BSCCO processes?
One of ORNL's industrial partners is Midwest Superconductivity of Lawrence, Kansas, the first company to license the RABiTS technology. It uses metal organic chemical vapor deposition to grow YBCO on single-crystal ceramics. "Their process is scalable," Hawsey says, "but we want to know if it will work on the nickel substrate. They wanted to work with us on this question under a cooperative research and development agreement (CRADA). If their process is scalable for making commercial wires from nickel and YBCO, that may bring costs down. They've already duplicated ORNL's results on short samples."
Midwest Superconductivity has teamed with Westinghouse Science and Technology Center, and the two partners are collaborating with ORNL. "We will work with both partners to get them in position to produce buffered textured metals in long lengths," Hawsey says. He notes that Westinghouse is a potential purchaser of wire for use for applications in their new generation of motors, transformers, generators, and energy storage systems.
"Westinghouse will tell us what the wire will have to do to work in new applications," Hawsey says. "We will try to solve problems to make sure the wire performs as needed."
Southwire, another CRADA partner of ORNL, will work as an adviser to Midwest Superconductivity on the use of RABiTS wire in transmission cables. On April 3, 1997, ORNL signed a four-way CRADA with 3M (a Fortune 100 company), Southwire Company, and LANL to further develop the RABiTS and IBAD substrates. 3M will use its expertise in electron beam deposition, laser control, and process modeling. Southwire will apply its expertise in thermomechanical processing of metals and underground cable product definition.
"We will continue to do research to develop a more economical process," Hawsey says. "One way to save money is to avoid using expensive laser deposition systems because they require high-power lasers and a vacuum, which are costly to produce and maintain. We are looking at solution-based techniques like sol-gel processes. We have experience in ceramic sol gel processes that we have patented. A number of people have demonstrated that sol-gel processes can deposit a superconductor that grows epitaxially on single-crystal substrates such as magnesium oxide or lanthanum aluminate. We are proposing to develop a way to use sol-gel processes to deposit YBCO and buffer layers, avoiding the cost of vacuum techniques and lowering the manufacturing cost.
"I'm not ready to call a winner between ORNL or LANL but I hope one or both of these two DOE-funded processes will work for industry," Hawsey adds. "A nickel-YBCO wire could finally kick off some of these applications that have never been possible in the electric industry because of the high cost of cooling metallic superconductors such as niobium-tin and niobium-titanium. To make BSCCO work for a motor or generator, it must be cooled to 20 K by expensive helium or a helium-gas cryocooler.
"If this YBCO wire works when chilled by much cheaper liquid nitrogen, all of a sudden we're talking about cost-effective electric power applications that hadn't been possible before. Underground cables that carry twice the power as conventional cables will be available within the next five years, and superconducting transformers that can be installed inside office or apartment buildings without concerns about fire or oil leakage may be developed. Other, non-utility applications are on the horizon.
"For example, superconducting magnets in MRI machines in hospitals are cooled by liquid helium. If we can replace them with magnets that can be cooled by liquid nitrogen, then these machines' cooling systems will be simplified. One result may be lowered health care costs. In addition, high-temperature superconducting wires would enable less expensive operation of superconducting accelerator magnets and fusion magnets at DOE research laboratories."
Hawsey says that the development of the RABiTS wire has several interesting aspects. "We kept quiet for five years and published no papers. Our effort was highly multidisciplinary. The development required staff from three ORNL divisions and funding from two DOE sources--the Office of Energy Efficiency and Renewable Energy and the Office of Energy Research."
The DOE laboratories should be "big players" in superconducting research, he says. "The fact that DOE laboratories have gone this far is testimony to the value of our partnerships with industry."
In the quest to pave the way for the efficient flow of electrons, ORNL researchers have taken a long, tortuous route that has proved also to be an exciting intellectual journey.
[ Next article | Search | Mail | Contents | Review Home Page | ORNL Home Page ]