LIQUID HYDROGEN Fuel-cell-powered cars could get a boost if there were better ways to store H2 on board. A global competition last year showcased environmentally friendly vehicles such as DaimlerChrysler's prototype Necar 4.
	PHOTO BY JOE WILSSENS/WIECK MEDIA SERVICES

Among those dreams is the controversial one that carbon nanotubes might provide the best medium for storing hydrogen. Hydrogen is viewed by many as an ideal fuel for the future because it is lightweight and plentiful and its oxidation product (water) is environmentally benign. When hydrogen reacts with oxygen in a fuel cell, for example, it produces electricity. If hydrogen fuel cells were used to power automobiles and other vehicles, air pollution would be reduced, and so would our dependence on imported oil. But such fuel cells would be practical only if the fuel--in this case, hydrogen--could be stored on board the vehicle in a safe, efficient, compact, and economical manner.

And that's the big stumbling block: After having explored the possibility of using compressed or liquefied hydrogen or existing hydrogen-storage materials such as metal hydrides and activated carbons, many scientists and engineers have concluded that none of these is adequate for mobile applications.

When carbon nanotubes first became available, they drew interest because these tubes are typically produced in bundles that are lightweight and have a high density of small, uniform, cylindrical pores (the individual nanotubes). Scientists found that they could fill these pores with many substances--so why not hydrogen?

Under the right conditions, there's no reason why nanotubes would not welcome hydrogen molecules into their interior space or into the channels between the tubes. But the crucial question is this: Can nanotubes store and release practical amounts of hydrogen under reasonable conditions of temperature and pressure?

THE ANSWER to that question depends on whom you ask. Michael J. Heben, a materials scientist at the National Renewable Energy Laboratory (NREL) in Golden, Colo., holds one of the more optimistic views. Studies by his group indicate that single-walled carbon nanotubes (SWNTs) can store up to 8% by weight of hydrogen at room temperature and moderate pressure. That figure is promising, given that the Department of Energy has set 6.5 wt% as the target capacity for a practical hydrogen-storage material for use in vehicles.

Heben's results are disputed by Michael Hirscher, a physicist at Max Planck Institute for Metal Research in Stuttgart, Germany. Hirscher has tried to reproduce Heben's findings, but in his hands the SWNTs adsorb less than 1 wt% hydrogen at room temperature and ambient pressure. Hirscher suspects that the H₂ uptake observed by Heben may be due to a hydrogen-adsorbing metal alloy contaminant in the nanotube samples.

Other scientists have weighed in as well. For example, physics professor Peter C. Eklund and postdoctoral fellow Bhabendra K. Pradhan at Pennsylvania State University have observed hydrogen uptake of 6 wt% in nanotubes--but at 77 K, the temperature of liquid nitrogen. At room temperature, they tell C&EN, nanotubes adsorb only "negligible" amounts of hydrogen. The 6 wt% uptake at 77 K is not of practical consequence for vehicles, they believe, because of the costs that would be involved in maintaining the storage system at that temperature.

J. Karl Johnson, an associate professor of chemical and petroleum engineering at the University of Pittsburgh, says that a number of theoretical modeling studies, including his own, are consistent with Eklund's experimental observations. If you assume that the intact hydrogen molecules physically adsorb to the nanotubes--a process known as physisorption--and you make reasonable assumptions for the molecular interaction potentials, Johnson says, the modeling indicates that "nanotubes are not capable of storing large amounts of hydrogen at room temperature and moderate pressures." The simulation results agree fairly well that only about 1 to 2 wt% of hydrogen is stored in nanotubes at room temperature and reasonable pressures, he adds.

But the modeling results are only as good as the assumptions used in the modeling, theorists point out. If something more than physisorption is occurring in Heben's experiments--as Heben has suggested--then the results of the studies modeling physisorption of hydrogen may be misleading.

"I don't think my simulations prove that Mike Heben's results are wrong," Johnson says. "I think they indicate that his results cannot be explained through simple physisorption." Heben's results are so intriguing, Johnson remarks, that they "make me think there's something else going on. I think you have to keep an open mind."

Certitude also eludes theorist Milton W. Cole, a physics professor at Penn State. "There are all kinds of experimental results out there, and I don't know how seriously to take them, to tell you the truth." The field is in disarray, grappling with "an open, controversial problem," he says, and "I don't know who's right and who's wrong."

MEETING OF MINDS NREL's Heben (left) and Penn State's Pradhan (right foreground) discuss a point during a break in a nanotube session at the November 2001 Materials Research Society meeting in Boston.
PHOTOS BY RON DAGANI

IT'S NO SURPRISE that Cole and other scientists are confused. The research literature addressing hydrogen storage in nanotubes and related carbon materials has become muddled because of conflicting and sometimes dubious reports. In 1999, for instance, Jianyi Lin and coworkers in the physics department at the National University of Singapore reported remarkable hydrogen uptake by alkali-metal-doped multiwalled nanotubes formed by the catalytic decomposition of methane [Science, 285, 91 (1999)]. The H₂ uptake was claimed to be 20 wt% for lithium-doped nanotubes at 380 °C and 14 wt% for potassium-doped nanotubes at room temperature. Subsequent studies in other labs cast doubt on these results, attributing them to the presence of water impurities.

Even more startling were the claims of chemist Nelly M. Rodriguez and coworkers at Northeastern University. They reported that certain graphite nanofibers can store hydrogen at levels exceeding 50 wt% at room temperature [J. Phys. Chem. B, 102, 4253 (1998); C&EN, May 25, 1998, page 6]. The results were seen as incredible when one considers that methane, which has four hydrogens per carbon atom, has a hydrogen content of 25 wt%. Attempts at other labs to reproduce the Northeastern findings were unsuccessful.

Both the Singapore and the Northeastern claims have been dismissed as red herrings. Now the field's attention has shifted back to Heben, who is the most vocal proponent of the view that carbon nanotubes--under the right conditions and after proper processing--can meet or exceed DOE's target for hydrogen storage.

Despite considerable heat from his critics, Heben has stuck to his guns on this point ever since he and his coworkers published the first paper on hydrogen storage in SWNTs almost five years ago [Nature, 386, 377 (1997)]. This paper was caught in the peer review process for almost two years and was revised several times before Nature agreed to publish it, according to Heben. But since its appearance, it has driven a lot of the subsequent work in this field, he says.

At the time the work reported in the Nature paper was begun, the art of nanotube synthesis was still in the Stone Age. The NREL team was obliged to work with highly impure samples--SWNT-containing soots that were prepared by Heben's IBM collaborators by coevaporating cobalt and graphite in an electric arc. Milligram quantities of these soots (containing at most about 0.2 wt% SWNTs) were exposed to H₂ at specific temperatures and pressures. Using temperature-programmed desorption (TPD) spectroscopy, the researchers monitored the amount of hydrogen that was released as the sample was gradually heated to high temperatures. From their data, Heben and coworkers estimated that the SWNTs in their samples were adsorbing about 5 to 10 wt% H₂ under ambient conditions.

The NREL team later began producing single-walled nanotubes in much higher yield and purity using a laser vaporization process pioneered by Richard E. Smalley of Rice University and his coworkers. Heben's group developed a chemical process to purify these laser-generated tubes. The tubes were found to be very long and were assumed to be capped with fullerene hemispheres, which explained why no hydrogen uptake could be observed--the H₂ molecules couldn't enter the tubes. And because of their near-perfection, the tubes were very stable, resisting attempts to cut them open.

The NREL scientists finally resorted to "very aggressive" measures, Heben says: By subjecting a suspension of the nanotubes in 5 M nitric acid to high-power ultrasonic waves for 16 hours, they succeeded in cutting them. Some samples of cut tubes were found to store as much as about 8 wt% H₂.

	Hirscher

TO CARRY OUT the sonication, Heben's group employed a commonly used ultrasonic probe whose tip, or horn, is composed of an alloy that is 90% titanium, 6% aluminum, and 4% vanadium. During the sonication, they later found, the tip breaks down and introduces alloy particles into the cut nanotube sample. Titanium is known to adsorb hydrogen, so this was a potentially worrisome discovery. But some samples were found to be taking up more hydrogen than would be expected for the alloy itself, Heben says.

Meanwhile, Max Planck Institute's Hirscher, who had studied hydrogen diffusion in metals and alloys but hadn't previously worked with nanotubes, took an interest in the hydrogen-storage problem. He says that his group carefully tried to reproduce Heben's work, including the sonication procedure, using SWNTs prepared by arc discharge and laser vaporization. They, too, found that the sonication treatment contaminates the nanotube samples with Ti-Al-V particles. And they showed that these particles do take up hydrogen.

In Hirscher's lab, the nanotube samples were found by TPD spectroscopy to adsorb a maximum of 1.5 wt% hydrogen. All of this, he says, can be explained by assuming that the hydrogen is stored only in the titanium-alloy particles [Appl. Phys. A, 72, 129 (2001)].

Hirscher and coworkers have shown that the amount of Ti alloy in any given sample increases with sonication time, and the amount of hydrogen storage by that sample increases proportionately. Furthermore, when they use a stainless steel horn instead of a Ti-Al-V horn, they observe no hydrogen storage in the samples.

Hirscher described these published results last November in an invited talk at the Materials Research Society (MRS) meeting in Boston. He also reported that additional experiments using a different technique to measure hydrogen adsorption confirm his group's definitive conclusion that SWNTs store less than 1 wt% hydrogen at room temperature.

Heben's turn to speak came next, and he started out in defensive mode, noting that the "accusations that have been raised about the quality of our results need to be addressed." Heben set out to respond to his critics, who he says believe his group has been "misled by inaccurate equipment, poor calibration, and incomplete methods." In his talk, he went to great lengths to describe the nanotube samples and how they are processed, the laboratory equipment used, and the extensive calibrations for the TPD experiments.

These calibrations indicate that the Ti-Al-V alloy in his samples adsorb hydrogen in the 2.5 to 3.4 wt% range. When this is taken into account, a few of the data points he showed at the MRS meeting still suggest that the SWNT fraction adsorbed around 8 wt% hydrogen in those experiments. The problem is that these results cannot be reproduced in every experiment. Most of the data points indicate less than 3 wt% hydrogen uptake by SWNTs. "More than half the time you don't see any hydrogen adsorption by the samples because the process is so poorly controlled," Heben points out. Even samples having similar purity and similar tube sizes and chiralities do not always exhibit the same H₂ capacity after the tubes are sonicated, he says.

Heben suggests that nanotubes in some of the samples are being "activated" in some way by the alloy particles so that they adsorb more hydrogen than they would in the absence of the metal. The alloy, he believes, may assist hydrogen uptake "by a catalytic effect in which the titanium dissociates the H₂" or "via thermal effects--the reaction of hydrogen with the titanium is exothermic, and this could locally create thermal disturbances to the tube that might aid in the addition of the hydrogen."

The exact mechanism, though, is far from clear, Heben admits. But because much of the stored hydrogen desorbs at higher temperatures than is typically found for physisorption, he favors a mechanism that is midway between physisorption and chemisorption, the process in which covalent carbon-hydrogen bonds would be formed. The mechanism, he thinks, could involve fractional electronic charge transfer between the H₂ molecule and the nanotube, with the H₂ molecule remaining intact.

Heben believes there is some experimental and theoretical support for such a mechanism. For instance, he cites the work of chemists Hansong Cheng, Guido P. Pez, and Alan C. Cooper at Air Products & Chemicals. Their quantum mechanical molecular dynamics simulations suggest that when H₂ molecules interact with the nanotube walls, the walls flex inward and the gas molecules orient themselves so that they can accept charge (electron density) into a * (antibonding) orbital from the highly distorted areas of the tube [J. Am. Chem. Soc., 123, 5845 (2001)]. These enhanced electron-transfer interactions make adsorption more likely, the chemists believe.

	FLEXING TUBES Quantum mechanical molecular dynamics simulations by Cheng and coworkers suggest that the walls of nanotubes in a bundle fluctuate in position, giving rise to distortions that create the opportunity for charge-transfer interactions between the walls and hydrogen molecules. This illustration includes a greater number of hydrogen molecules than were involved in the initial simulation.
	IMAGE BY HANSONG CHENG

One major difficulty that other researchers have with Heben's results is that the TPD curves--the plots of how much H₂ desorbs from the sample at various temperatures--are similar in shape for all samples containing the titanium alloy, whether the samples are 100% alloy or contain, in addition, either SWNTs or graphite. As Eklund puts it: "The desorption data from samples without nanotubes look the same as the desorption data from samples with nanotubes. So what would a reasonable person conclude?"

HEBEN COUNTERS that the TPD curves "are similar in shape but are clearly different in magnitude" for nanotube samples exhibiting high H₂ uptake. Their similarity in shape, he suggests, may reflect the alloy's role in mediating the uptake and release of hydrogen by the nanotubes.

While Heben has spent a lot of time responding to criticisms of his work, he also has questioned some aspects of the Max Planck group's work. For example, he thinks that their findings of essentially no H₂ uptake by nanotubes could be explained if the tubes don't survive the sonication treatment or remain capped or if their electronic properties are different from those of the NREL tubes. Hirscher admits that he and his coworkers don't even know what fraction of their tubes is being opened by the sonication treatment.

Scientists who have remained on the sidelines in this debate are perplexed by the fact that a range of different hydrogen-uptake results for nanotubes has been reported--from Heben's to Hirscher's and beyond. But perhaps it's not too surprising, Penn State's Cole suggests, because "different laboratories have different nanotube samples and different techniques of preparation, so it's like comparing apples and oranges."

"My feeling is that the only way to solve this controversy is to share samples," says Patrick Bernier, a physicist at the University of Montpellier II, in France, who collaborates with Hirscher and has provided arc-produced nanotube samples for Hirscher's hydrogen-uptake experiments. "We can work on their samples, and they can work on ours." Bernier shared some of his samples with Heben but was told by the NREL scientist that those samples were not of high enough quality to get a meaningful result. Bernier would like a sample of Heben's material to scrutinize in his own lab, and Heben says he intends to send him one.

Heben says he has given some samples to other researchers, but they didn't handle the samples correctly. He says his team is now working closely with a Canadian group and with researchers at Honda R&D Americas, which has been funding his work, to gain "external validation."

The real bottleneck in this research, explains John E. Fischer, a professor of materials science and engineering at the University of Pennsylvania, is that Heben hasn't figured out how to optimize hydrogen uptake by his nanotubes and so cannot make enough of the optimized tubes for the necessary testing. In the TPD experiments, Heben is using a milligram or so of material. For nuclear magnetic resonance spectroscopy, which would shed light on the mechanism, Fischer says, "you need 50 mg." For neutron scattering experiments, which could help pinpoint the sites where H₂ molecules are adsorbed, "you need 100 mg." The NREL scientists need to understand the activation process better so that they can produce more of the optimized tubes, Fischer says. These tubes would need to be analyzed in detail to find out how they differ from nanotube samples being studied in other labs.

Bernier says his lab has already started collecting nanotube and related carbon samples from different labs so that their H₂ uptake can be determined under identical conditions.

ANOTHER STRATEGY that might help crack the hydrogen-storage mystery, Heben suggests, is to "decouple" the nanotube-cutting process from the alloy-introduction step. With this in mind, his group and others are searching for ways to cut nanotubes without sonication.

During his MRS talk, Heben remarked that being forced to address the criticisms against his work "has been good for us because we've really tightened our belts" and gained a better understanding of what's going on in this system. But Heben was not above hurling a couple of accusations against Hirscher. And he couldn't help but mention that he was "upset" that Hirscher never acknowledged in his preceding talk and in his 2001 Applied Physics A paper that he had been informed in 2000 that the NREL team knew the sonication process introduces alloy impurity into nanotube samples. By not acknowledging this, Heben asserted, Hirscher made it appear as if the Max Planck group had discovered the alloy impurity while the NREL team was clueless about it.

In an interview with C&EN after the meeting, Heben expressed regret at venting his feelings during the talk, calling it "somewhat unprofessional."

As the putative underdog in this contest, he's feeling increasing pressure to prove that his results are not wrong.

While some scientists have deserted Heben, others remain supportive. For example, Air Products' Cooper says that Heben is "a very careful researcher" and his experimental work is "pretty solid."

And Fischer remarks: "A lot of people beat up on Michael Heben. He asks for it sometimes because he's a pugnacious fellow. But I have to give him the benefit of the doubt."

When asked if he thinks nanotubes have a future in hydrogen storage, Fischer says, "I'm reluctant to say it'll never go anywhere. I'm still holding out hope."

[Previous Story] [Next Story]

Chemical & Engineering News
Copyright © 2001 American Chemical Society