Endohedral fullerenes--carbon cages encapsulating atomic or molecular species--have been objects of fascination and frustration since their discovery in 1985. They have captivated scientists with their tantalizing properties and possibilities for application. But they also have stymied research efforts because they have proven to be challenging to prepare, purify, and understand. It's not surprising, then, that practical applications for these exotic materials remain elusive.

At the 201st meeting of the Electrochemical Society (ECS), held last month in Philadelphia, attendees were treated to a smorgasbord of new findings about endohedral fullerenes that is likely to feed their fascination. Speakers at the conference reported intriguing preliminary results on upping the synthetic yield of "endohedrals," encapsulating a carbene inside a carbon cage, testing a new metallofullerene derivative with potential for medical imaging, and measuring for the first time the nuclear magnetic resonance (NMR) spectrum of a xenon atom trapped inside a buckyball, among other advances.

In his ECS talk, chemist Lothar Dunsch of Leibniz Institute for Solid State & Materials Research, in Dresden, Germany, pointed out some of the advantages and problems associated with endohedrals. For example, he noted that these fullerenes can be used to stabilize reactive species inside the cage, as in N@C₆₀ or Sc₂C₂@C₈₄ (C&EN, Jan. 22, 2001, page 16). Endohedrals can also serve as nondissociating salts in electrochemistry--for example, with a negatively charged cage encapsulating a positively charged ion. In addition, endohedral structures offer exciting electronic and magnetic properties and might be applied to quantum computing or biomedical applications.

But obstacles stand in the way of continued progress, Dunsch noted. Typically, endohedrals are prepared in an electric arc by evaporating carbon electrodes in the presence of other atomic or molecular species that sometimes end up inside the carbon cages. The endohedrals, though, are produced in very low yield, typically less than 1%. And scientists still don't understand how these molecules form in the arc burning process. Furthermore, Dunsch said, there are no rules for predicting which elements or clusters will be incorporated into a particular cage structure or how the encapsulated cluster will affect the cage. Also, some endohedrals have limited stability, even in the dark and at room temperature.

Dunsch and his coworkers produce endohedrals on a larger scale by using an arc burning apparatus that can accommodate several pairs of carbon rods. To make trimetallic-nitride-containing fullerenes such as Ho₃N@C₈₀, which are being studied in a number of labs around the world, the Dresden group uses holmium-containing graphite rods and adds nitrogen compounds, even using reactive gases in the atmosphere of the arc burning chamber to fine-tune the synthetic process.

Endohedrals can stabilize reactive species inside the cage, can serve as nondissociating salts in electrochemistry, and offer other exciting properties.

Because of patent considerations, Dunsch avoided divulging details of the process. But he did claim that by adjusting the reaction conditions in the arc chamber, it's possible to get an endohedral structure like Ho₃N@C₈₀ "as a main product." This is the first time, he suggested in his talk, that anyone has obtained an endohedral structure as a main product in an arc burning synthesis. The purification of the compound, using high-performance liquid chromatography, is simple, he said.

Dunsch and his collaborators, including chemistry professor Hisanori Shinohara's group at Nagoya University in Japan, compared the vibrational spectra of three M₃N@C₈₀ fullerenes, where M = Sc, Y, or Ho, and drew some preliminary conclusions about the compounds' structural similarities and differences.

But the more pulse-quickening part of his lecture came when he discussed a new endohedral structure that appeared in a mass spectrum at a mass-to-charge ratio of 854. At first, the researchers thought it might be N@C₇₀, but the electron spin resonance spectrum clearly appeared to rule that out. Now they believe it is CH₂@C₇₀--the simplest of all carbenes, and a highly reactive one at that, trapped inside a fullerene. They ruled out the cage-functionalized isomer C₇₀CH₂--a known cyclopropanated molecule--by reacting C₇₀ with diazomethane and comparing the product fullerene's infrared spectrum with that of the mystery compound. The two spectra are not identical, Dunsch declared.

The group members' evidence for an encaged carbene is still short of conclusive, so they are working on further characterization, including accumulating enough of the compound to determine its NMR spectrum. But at the meeting, Dunsch was confident that they had produced yet another example of a highly reactive species that is stabilized inside a fullerene cage. He told C&EN that the carbene-endohedral can be purified to greater than 90% purity using HPLC and that it is stable for months.

The issue of how this compound was prepared was left vague. Dunsch would say only that it can be produced by adding an organic compound to the graphite rod or--better yet--by adding a reactive gas to the atmosphere inside the arc chamber.

CONNECTIONS Crystal structure of the first organic derivative of Sc₃N@C₈₀ (Yellow = Sc; blue = N; red = O; white = H). This work, published by Alan L. Balch and coworkers, demonstrates that additions to C₈₀ occur at the junctions of a five-membered ring and a six-membered ring rather than at the more usual site, the junction of two six-membered rings [J. Am. Chem. Soc., 124, 3494 (2002)].
HON MAN LEE AND ALAN L. BALCH

THE ABILITY OF fullerenes to sequester one or more metal atoms inside the cage has led to a research effort aimed at exploring their potential as contrast-enhancing agents for magnetic resonance imaging. Contrast agents enhance the quality of MRI images, aiding in the detection and diagnosis of injuries or abnormalities in the human body. The leading commercial MRI contrast agents are gadolinium(III) chelates such as gadolinium-diethylenetriaminepentaacetic acid, known by its brand name Magnevist. Gd³⁺ works so well because of its unique electronic structure--it is the only ion with seven unpaired electrons. Thus, once injected into the body, it can magnetically "tickle" water protons present in tissues, accelerating their relaxation between radio-frequency pulses. Faster relaxation leads to higher signal intensity and therefore greater contrast in the MRI images.

Organic chelates of gadolinium seem to do a good job of keeping this toxic heavy metal locked up and away from biomolecules as the chelate wends its way through the body. But encapsulating the gadolinium inside a fullerene might be even safer, and such endohedrals could offer additional advantages. For example, the trimetallic-nitride-containing endohedral fullerenes, first reported by chemistry professor Harry C. Dorn's group at Virginia Polytechnic Institute & State University, Blacksburg, can accommodate three metal atoms inside each cage, potentially offering a more potent agent (C&EN, Sept. 20, 1999, page 54). These compounds are being developed for medical applications by Luna Innovations, a company based in Blacksburg.

Before such endohedrals can be tested in vivo, though, they must be made water-soluble. One way to do this is to attach hydroxyl groups to the outer surface of the cage. Shinohara and coworkers prepared polyhydroxylated Gd@C₈₂ and found that, compared with Magnevist, it provides almost 20 times better signal enhancement for water protons at much lower gadolinium concentration [Bioconjugate Chem., 12, 510 (2001)].

Although this result is promising, there are problems with C₈₂ endohedral fullerenes that could knock them out of contention, according to Robert D. Bolskar, a chemist at TDA Research, a contract R&D firm in Wheat Ridge, Colo. In his presentation at the ECS meeting, Bolskar noted that most studies of metallofullerenes have centered on some C₈₂ isomers primarily because their solubility allows them to be more easily separated from empty fullerenes and purified using HPLC. But Shinohara's study of polyhydroxylated Gd@C₈₂ in rats revealed significant uptake of the material by the reticular endothelial system, such as the lung, liver, and spleen.

In a related study, Bolskar noted, a U.S. collaborative team observed similar uptake of a polyhydroxylated derivative of Ho@C₈₂ in these tissues, as well as in bone. These results are not particularly encouraging for pharmaceutical applications, he said, because you want the metal-containing agent to be excreted without long-term retention in tissues.

 

---

CAGE MASTER Bolskar stands in front of TDA's custom arc-discharge apparatus, which is used to generate endohedral metallofullerenes.
TDA RESEARCH

ANOTHER DRAWBACK of C₈₂ endohedral fullerenes is that they are difficult to make in large quantities in high purity, which would be necessary for a pharmaceutical, he added.

Bolskar and his colleagues, including chemists J. Michael Alford at TDA and Lon J. Wilson at Rice University, decided to strike out on a different path and explore the potential of metallofullerenes in the C₆₀ family, such as Gd@C₆₀. According to the TDA researchers, this metallofullerene class has been overlooked for pharmaceutical applications because its members are generally insoluble and air-sensitive. On the plus side, though, M@C₆₀ compounds can be produced in a carbon arc in yields up to 10 times higher than soluble M@C₈₂ species. With this potential production advantage, it seemed worthwhile to explore ways to improve the solubility and stability of C₆₀ metallofullerenes, Bolskar suggested.

The TDA chemists first developed procedures for separating the fullerene products from the nonfullerene soot, then for separating the metallofullerenes from the empty fullerenes. Eventually, they got a fullerene fraction that is largely Gd@C₆₀.

Next they derivatized the metallofullerene using a variation of the well-known Bingel cyclopropanation reaction. The final product they obtained is Gd@C₆₀[C(COONa)₂]₁₀, a carbon cage in which 10 double bonds have been transformed into cyclopropane rings substituted with two carboxyl groups. The sodium salt is water soluble and fully air-stable, Bolskar announced. And he added that the synthetic procedure has been scaled up so that they can produce hundreds of milligrams of this derivative in a single run.

Bolskar was not able to show his audience an X-ray structure or any conclusive structural evidence for this compound, and he admitted that the material may contain several different isomers.

Nevertheless, his medical collaborators at the M. D. Anderson Cancer Center in Houston have shown that the compound functions as a safe and effective MRI contrast agent in the rat. The rat survived the test "without any complications," Bolskar noted, and the compound provided contrast enhancement in MRI images comparable to that of commercial gadolinium chelates.

The MRI images of the rat's organs revealed where the compound goes, showing that it does not localize in the liver, bone, or spleen, but is rapidly excreted by the kidneys. "So it's quite promising to see that the biodistribution of this carboxymetallofullerene is quite distinct" from that of the polyhydroxylated fullerenes, Bolskar said.

Although polycarboxylated metallofullerenes like the Gd@C₆₀ derivative may have some advantages over polyhydroxylated C₈₂ metallofullerenes, it's too early to know if either will find their way to the marketplace. Even if they cannot compete directly with commercial MRI contrast agents, it's possible "they could fit into niche markets where commercial organic chelates are not optimal," Bolskar said.

Besides metal atoms, molecular clusters, and reactive species, chemists also have been keen to encapsulate noble-gas atoms inside fullerene cages and study the interactions between the host and guest. At Yale University, chemistry professors R. James Cross Jr. and Martin Saunders have developed a cottage industry of sorts: They insert helium-3 into C₆₀ and send the endohedral to various collaborators who chemically modify the outside of the cage in different ways and then send the fullerene products back to Yale for NMR analysis. Since every ³He-labeled fullerene has a distinctive helium chemical shift, this shift can be used to pin down the structure of the derivative, as well as monitor the molecule's subsequent chemical transformations. ³He NMR spectroscopy thus has become one of the most powerful tools for following fullerene chemistry.

In addition to helium, the Yale researchers also have put neon, argon, krypton, and xenon into fullerenes, making unusual and highly stable noble-gas compounds in which no formal bond exists between the noble gas and the surrounding carbons. These compounds typically are made by heating the fullerene in the presence of the gas at 650 °C and 3,000 atm. Under these conditions, though, no more than one in 1,000 fullerene cages ends up with a noble-gas atom inside.

Aside from ³He, ¹²⁹Xe is the only other noble-gas isotope having a spin of one-half, which makes the nucleus easily observable using NMR spectroscopy. As Cross explained in his ECS talk, xenon NMR is notable for one characteristic: The chemical shift of ¹²⁹Xe changes dramatically from one solvent to another, often by many parts per million. For ³He, by contrast, the effect is very much smaller. "So we would expect all sorts of differences" between the xenon shifts and the helium shifts of fullerenes, he said.

"We've been trying to do xenon NMR for years," Cross continued. "We showed we could put xenon [inside C₆₀] many years ago." But several difficulties have thwarted their efforts. First, ordinary xenon is only about 26% ¹²⁹Xe. Second, when they try to force xenon into C₆₀, they get three to five times less xenon inside than helium, probably because xenon is so much larger.

Finally, the gyromagnetic ratio of ¹²⁹Xe is much smaller than that of ³He, which makes xenon's NMR sensitivity much lower than helium's.

After many unsuccessful attempts, Saunders, Cross, and postdoc M. S. Syamala finally found a way around these problems late last year. First, they purchased xenon enriched with ¹²⁹Xe (86%). Second, they ground the fullerene together with potassium cyanide before subjecting it to the high-pressure gas reaction. This had the effect of boosting the amount of xenon incorporation by about an order of magnitude, although it's not clear how KCN improves the yield. Finally, they used HPLC to separate the xenon-stuffed buckyballs from the empty C₆₀, obtaining NMR samples that were enriched in the endohedral fullerenes.

	OBSCURED The xenon atom inside this buckyball is largely hidden by the carbon cage. IMAGE BY R. JAMES CROSS JR.

THIS LAST STEP was particularly laborious, Cross explained. The Xe@C₆₀ peak in the chromatogram isn't discernible--it's lost in the noise underneath the tail of the peak for the empty C₆₀. "You know it's got to be there somewhere," Cross told his listeners. "So we collected aliquots coming off the HPLC, evaporated the solvent, and then analyzed for the presence of ¹²⁹Xe using the mass spectrometer." They repeated this at 15-second intervals during the elution and found where the Xe@C₆₀ was coming off. Then they injected multiple samples--"one after another after another after another"--collecting the Xe@C₆₀ aliquot that comes off at the tail. By repeatedly pooling and reinjecting these aliquots, they were finally able to get a sufficiently enriched sample of ¹²⁹Xe@C₆₀.

The ¹²⁹Xe NMR measurements were performed on an 800-MHz (for proton) instrument. The resonance for ¹²⁹Xe@C₆₀ in benzene was found to be 28.89 ppm (upfield) from the resonance of ¹²⁹Xe dis-solved in benzene and +179.24 ppm from the resonance of ¹²⁹Xe in the gas phase. Such measurements are intriguing, Cross told C&EN, because, thus far, theory has not provided accurate predictions for ¹²⁹Xe chemical shifts like it has for ³He chemical shifts.

The Yale researchers also prepared the dimethylanthracene adduct of both ³He@C₆₀ and ¹²⁹Xe@C₆₀ and measured the respective helium or xenon shifts. The ³He adduct's resonance appeared at 23.36 ppm relative to ³He@C₆₀, while the ¹²⁹Xe adduct's resonance appeared at +11.31 ppm relative to ¹²⁹Xe@C₆₀.

These results--"a great surprise to us"--illustrate the differences between helium and xenon inside C₆₀, Cross noted. The helium atom, being much smaller than the cavity, "rattles around inside," he said. The helium electrons are tightly bound to the nucleus, and there is little interaction between these electrons and the p electrons of the fullerene. The ³He chemical shift simply gives a measure of the magnetic field inside the cavity that is being shielded by the p electrons.

"Xenon, on the other hand, is a tight fit inside the cage," Cross pointed out. Xenon's 5p electrons are much closer to and interact much more strongly with the fullerene's p electrons. When you alter the cage environment in any way, such as by making a fullerene adduct, "you can imagine that the cage may pucker slightly, the dimensions may change." The modified cage or the new group on the outside "will interact very strongly with the xenon" in ways that aren't as easy to describe, he said.

These results, some of which were published recently in the Journal of the American Chemical Society [124, 6216 (2002)], show that xenon inside C₆₀ is quite different from helium inside C₆₀. Nevertheless, Cross doesn't expect the insertion of xenon into fullerenes to become another cottage industry for him and Saunders. After all, he noted in his talk, achieving the first ¹²⁹Xe NMR spectrum of xenon inside a fullerene required laborious work and 48 hours on the NMR spectrometer--all to get data with a "really lousy-looking" signal-to-noise ratio. So this is not something the Yale team will be doing routinely, he thinks.

Still, with advances continually being made in fullerene research, who can predict what will be possible in a few years?

CAGEY CHEMISTRY

Chemical & Engineering News
Copyright © 2002 American Chemical Society