Carbon Cages: LBL Scientists Study Fullerenes

Summer 1993

LBL scientists are probing the properties of fullerenes--a family of molecules with immense industrial potential. See also the related article fullerenes-study-techniques.

By Lynn Yarris,

They look like soccerballs--pentagons and hexagons bound together into round hollow molecular cages along the same architectural lines as the geodesic domes created by the noted architect and philosopher R. Buckminster Fuller. About a billionth of a meter in diameter, they are incredibly stable-- slammed against a steel surface at 17,000 miles per hour, they bounce off undisturbed. They join diamond and graphite as the third known form of pure carbon. They are the 60-atom clusters called buckminsterfullerene (C60). Fullerene for short. Buckyballs to the general public. By any name, they and their 70-atom cousins (C70) are the most talked-about discovery in materials since high-temperature superconductors.

Fullerenes were discovered in 1985 by Robert Curl and Richard Smalley of Rice University while doing experiments with carbon clusters in supersonic beams. However, the evidence for the existence of these molecules remained indirect until 1990, when researchers at the Max Planck Institute in Germany used a carbon-arc plasma to produce the first directly observable (visible to the naked eye) quantities.

Within months, physicists, chemists, and materials scientists all over the world were making their own buckyballs and in the process spawning an entire family of fullerenes, from "buckybabies" with 32 carbon atoms (C32) to giant fullerenes with 960 carbon atoms (C960). In 1991, Science magazine named the buckyball its choice for "molecule of the year," calling it "the discovery most likely to shape the course of scientific research in the years ahead."

At LBL's Materials Sciences Division (MSD), experimentalist Alex Zettl and theorist Marvin Cohen have been working together for the past two years to find answers to the many questions posed by fullerenes--answers that will help researchers tap into the enormous potential of this most amazing new class of molecules.

Experimentalists and theorists have shown or suggested that solids based on buckyballs can be insulators, conductors, semiconductors, or even superconductors when doped with other atoms or molecules. Pure buckyball solids form crystal structures, like graphite or diamond, that are insulators or semiconductors. However, when doped with an alkali metal, such as potassium or rubidium, these solids can become electricity-conducting metals. Buckyballs doped with an organic reducing agent exhibit ferromagnetic properties. In the absence of metals, this is a phenomenon without precedence.

"Buckyballs are like a playground for chemists and materials scientists," says Zettl. "They are easily modified, using organic chemistry and other synthesis techniques, into other molecular structures, which makes them remarkably versatile. This versatility gives them enormous practical applications."

Even if there weren't any practical applications, fullerenes would probably be worthy of scientific study because of their unique shape. C60 is the roundest molecule known. Like all closed geodesic structures, it has the 12 pentagons that are required to transform a network into a spheroid, as demonstrated by the 18th-century Swiss mathematician Leonhard Euler. The remaining carbon atoms are configured as 20 hexagons to form the molecule's soccerball shape.

As hexagons are added or removed, the molecule begins to loose its roundness. C70, which has 25 hexagons, is shaped more like a rugby ball. Giant fullerenes take on a pentagonal shape. Smaller fullerenes look like asteroids. With the loss of roundness comes a loss of stability.

The versatility of fullerenes stems in part from the belief that they also can be doped either by attaching atoms to the outer bonds of the ball or by inserting one or more atoms inside the cage. To learn more about the properties of fullerenes and how they can best be used, scientists eventually have to produce the materials in a crystalline form large enough for electronic transport and mechanical measurements.

Zettl is used to being at the forefront of making exotic materials. He was among the first to make the yttrium-barium-copper-oxide compound that became superconducting at what was then a record high "critical temperature" of 90 Kelvin--well above the 77 K boiling point of liquid nitrogen. His experiences with the so-called high-temperature superconductors and the facilities he developed in his laboratory to make and study them served Zettl well when he began making his own fullerenes.

"My facility is flexible so I can quickly jump in and start making and exploring new classes of materials," Zettl says.

Zettl has been working with an "excellent team of researchers," starting with his postdoctoral associates, Xiao-Dong Xiang, Jian Guo Hou, and Li Lu, and including graduate students Gabriel Briceno, Brian Burk, Nasreen Chopra, Michael Fuhrer, and William Vareka. Together, Zettl and his group have been able to successfully synthesize and characterize the highest quality of sizable single crystals of both pure and doped buckyballs.

Until this breakthrough, characterizations of fullerene properties had been based on granular thin films and pressed powders containing an abundance of fullerene clusters. Often such forms of a material are desirable, but in some cases, Zettl says, reliable measurements on single crystals are essential for establishing intrinsic properties.

"Earlier electronic transport studies had shown anomalous fluctuations in thin films of fullerenes doped with potassium just above the critical temperature for superconductivity," says Zettl. "Based on our studies with single crystals, we conclude that these fluctuations are not intrinsic properties but are due to imperfections such as grain boundaries."

Zettl's experimental results appear to support the theoretical work of Cohen, who maintains that almost all of the measurements of the electronic nature of fullerenes that have been made so far are consistent with the "standard model of electronic structure."

Explains Cohen, "The standard model assumes that valence electrons move almost freely among the positively charged ions that form the crystal lattice."

Using this standard model, Cohen and Vincent Crespi, a graduate student, have been able to explain the data obtained by Zettl's group.

Says Cohen, "These successes also imply that the conventional theory of superconductivity is likely to be a correct description for fullerene solids."

The conventional theory of superconductivity was first advanced in 1957 by physicists John Bardeen, Leon Cooper, and Robert Schrieffer. Ever since, it has been known as the "BCS theory." It explains superconductivity as the result of electron-pairings through their mutual attraction to vibrations in the crystal's lattice. When a crystal's electrons are paired, every motion of one electron is precisely countered by its partner so that there is no overall resistance to the flow of an electrical current through the crystal.

Speculation about the need for an alternative theory to explain the superconductivity of fullerene solids was fueled by the discovery that crystals of doped buckyballs have an unusually high critical temperature. For two decades, thin films of niobium-germanium had held the record for standard BCS-type superconductors at 23 K. Buckyball crystals doped with alkali metals have gone superconducting at temperatures as high as 33 K. While far from the 130 K record of the latest and best of the high-temperature superconductors (a class of materials that may not fit BCS theory), the critical temperature of metallic buckyballs is achieved with a relatively simple chemical system that could prove to have its own advantages.

Says Cohen, "At this point, a model of phonons interacting with itinerant electrons may be a good first approximation for explaining the properties of the metallic fullerenes. When fullerenes are doped with alkali metals, they appear to conform to this model."

For example, Cohen points out that, to date, a state of superconductivity has been observed in C60 but not in C70, nor in fullerene tubes. The standard model predicts this because the greater amount of curvature in the perfectly round buckyballs increases the opportunity for interactions between electrons and phonons, the quanta associated with lattice vibrations. It also helps explain Zettl's observations that "the phonons relevant to the superconductivity may be coming from the buckyballs themselves as opposed to the entire crystal."

An indication that BCS theory is relevant for fullerenes, Cohen says, is the observation in experiments of the "isotope effect." An isotope is a form of the same element with a different atomic mass--meaning it has the same number of protons but a different number of neutrons in its nucleus. The discovery that in conventional superconductors different isotopes of the same element have different critical temperatures was one of the major clues leading to BCS theory because it showed that superconductivity can involve electron-phonon interactions.

Says Cohen, "Experimentalists have found big changes in the critical temperatures of fullerenes when carbon-13 is substituted for carbon-12. This means that the crystal's vibrational frequencies are changing--and that sounds like phonons!"

A number of novel mechanisms have been proposed to explain the relatively high critical temperatures of doped fullerenes, including one that links their superconductivity to interactions between their electrons. According to Cohen, however, the observation of the isotope effect in the materials tends to discredit some of these alternative theories.

"The standard model may be viewed as dull by some researchers," he says, "but if it is correct, it does unify a lot of concepts about solids and demonstrates that spectacular properties are possible within this model when the relevant parameters are allowed to vary over a wider range."

Cohen believes that theorists are now at the stage where they can explain experimental data and begin making predictions "based on trends rather than calculations." Theory can be used to tell expermentalists what properties to measure in both pure and doped fullerenes and give them, says Cohen, "new targets to shoot at." Already, Zettl and his collaborators, working with Cohen, have measured structural, thermodynamic (specific heat), and mechanical (elasticity) as well electronic properties of pure and doped fullerene crystals.

Zettl's fullerene crystals are much in demand by researchers all over the world. At the time of this writing, no one else has been able to duplicate his success for synthesizing large, high- quality doped crystals, even though his basic recipe for producing fullerenes is well known.

Zettl starts by creating an arc between two carbon rods in the presence of helium gas. The arc's heat vaporizes the rods into carbon atoms that can coalesce into sheets. It is believed that the inert helium holds the carbon sheets near the arc long enough for them to close in on themselves, forming millions and millions of fullerene spheres--most of which are C60 and C70. For the creation of pure crystals, the C60 is separated from the C70 through chromatography and condensed into powder. This powder of buckyballs is then vaporized at a high temperature inside a quartz tube.

"If the combination of temperature, gradient, and the amount of gas inside the tube is exactly right," he says, "fullerene crystals will grow on the walls of the tube. We are continually changing the details of our technique in order to optimize it."

It takes about 10 minutes to make fullerene powders but about 10 days to grow crystals. Once Zettl has produced the pure crystals, he can subsequently dope them. Atoms, such as potassium or rubidium, are inserted into the hollow spaces between individual buckyballs.

Zettl and his group have gotten so good at making single crystals of fullerenes that they are now constructing crystals of reduced geometric dimensions. Perhaps the most intriguing of these experiments involves the production of fullerene tubes-- essentially one-dimensional cylindrical crystals which were first discovered in Japan.

"Fullerene tubes have perhaps an even wider range of potential applications than buckyballs," says Zettl. "For example, you could make the world's strongest wire or fiber that could be either an insulator, a conductor or a semiconductor, depending on how you closed up the tubes."

The versatility of tubes stems from being able to attach other atoms to their ends which are highly reactive because of their curvature. It is believed that reactions could also be used to temporarily open tube ends for the insertion of atoms or molecules to produce a bonanza of exciting new materials. The idea of placing atoms inside fullerene cages rather than attaching them to the outside bonds is intriguing. For years, scientists have been working to design molecular cages that can encapsulate other atoms for purposes ranging from industrial catalysis, to the safe removal of toxic substances from the body. With the arrival of fullerenes, they now have access to an entire family of naturally hollow molecules that are so stable they could withstand the most demanding manufacturing processes and might even be used to safely contain and transport radioactive materials.

Zettl is currently exploring several ways of putting atoms inside fullerenes, including growing a carbon cage around another atom and using a scanning tunneling microscope to forcibly inject the atom through the carbon walls. He is also participating in a series of experiments on the ultrafast dynamics of fullerenes that will literally shed more light on the optical properties of these materials.

This latter undertaking involves a collaboration between Zettl's group and the femtosecond laser research group of LBL Director Charles V. Shank. Fullerene samples will be flashed with a succession of visible light pulses only 12 femtoseconds long (a femtosecond is a millionth of a billionth of second). Spectroscopic analysis, led by MSD's Robert Schoenlein, will subsequently allow the researchers to follow the photoexcitation process step by step.

Another effort underway at LBL, led by MSD physicist Yuen-Ron Shen, will explore the nonlinear optical properties of fullerenes (see sidebar).

Zettl and Cohen would like to establish a central fullerene research program at LBL.

"The program will take advantage of the initial lead Berkeley presently has in many areas of fullerene-derivative synthesis, single crystal growth, and superconductivity," says Zettl.

In addition to their staff positions at LBL, Zettl, Cohen are both professors at UC Berkeley. Their proposed fullerene program would thus be able to bring together several strong physics and chemistry research groups from both institutions.