Treated buckyballs can remove potentially toxic metal particles from water and other liquids while recovering valuable particles for future use, according to scientists at Rice University.

The Rice lab of chemist Andrew Barron has discovered that carbon-60 fullerenes (aka buckyballs) that have gone through the chemical process known as hydroxylation aggregate into pearl-like strings as they bind to and separate metals from solutions.

Potential uses of the process include environmentally friendly removal of metals from acid mining drainage fluids, a waste product of the coal industry, and from fluids used for hydraulic fracturing (fraking) in oil and gas production.

The study led by Rice undergraduate Jessica Heimann appeared in the Royal Society of Chemistry journal Dalton Transactions.

Previous research in Barron’s lab had shown that hydroxylated fullerenes (known as fullerenols) combined with iron ions to form an insoluble polymer. Heimann and colleagues conducted a series of experiments to explore the relative binding ability of fullerenols to a range of metals.

“For more complex fluids, the problem is to take out the ones you actually want,” Barron said. “Acid mining waste, for example, has large amounts of iron and aluminum and small amounts of nickel and zinc and copper, the ones you want. To be frank, iron and aluminum are not the worst metals to have in your water, because they’re in natural water, anyway.

“So our goal was to see if there is a preference between different types of metal, and we found one. Then the question was: Why?”

The answer was in the ions. An atom or molecule with more or fewer electrons than protons is an ion, with a positive or negative charge. All the metals the Rice lab tested were positive, with either 2-plus or 3-plus charges.

“Normally, the bigger the metal, the better it separates,” Barron said, but experiments proved otherwise. Two-plus metals with a smaller ionic radius bound better than larger ones. (Of those, zinc bound most tightly.) But for 3-plus ions, large worked better than small.

“That’s really weird,” Barron said. “The fact that there are diametrically opposite trends for metals with a 2-plus charge and metals with a 3-plus charge makes this interesting. The result is we should be able to preferentially separate out the metals we want.”

The experiments found that fullerenols combined with a dozen metals, turning them into solid cross-linked polymers. In order of effectiveness and starting with the best, the metals were zinc, cobalt, manganese, nickel, lanthanum, neodymium, cadmium, copper, silver, calcium, iron and aluminum.

The “pearl” reference isn’t far from literal, as one inspiration for the paper was the fact that metal ions are cross-linking agents for proteins that give certain marine mussels an amazing ability to adhere to wet rocks.

Heimann, a senior, started on the project before spending a semester at Rice’s sister institution in Germany, Jacobs University. “I initially worked with carbon nanotubes, oxidizing them to see how they would bind metals, and then I went abroad,” she said. By the time she came back, Barron was ready to try C-60. “Coming from Rice and its history with buckyballs, I thought that would be really cool,” Heimann said.

“I liked being able to see the end goal of making a filter that could be used to address contaminated water,” she said.

Barron said fullerenols act as chelate agents, which determine how ions and molecules bind with metal ions. Experiments with various metals showed the fullerenols bound with them in less than a minute, after which the combined solids could be filtered out.

Barron said the choices of aluminum, zinc and nickel for testing were due to their co-presence with iron in acid mining drainage water. Similarly, cadmium was tested for its association with fertilizer and sewage sludge and copper with mining discharge. Nickel, lanthanum and neodymium are used in batteries and drive motors in hybrid vehicles.

Barron said the research shows the versatility of the buckyball, discovered at Rice in 1985 by Nobel Prize winners Rick Smalley, Robert Curl and Harold Kroto. It also points the way forward. “The understanding we now have is allowing us to find alternatives to C-60s to design ways in which we can separate out metals more efficiently,” he said.

Co-authors of the paper are Rice graduate student Lauren Morrow and alumnus Robin Anderson. Barron is the Charles W. Duncan Jr.-Welch Professor of Chemistry and a professor of materials science and nanoengineering.

The Robert A. Welch Foundation and the Welch Government Sêr Cymru Programme supported the research.

---

Abstract of Understanding the relative binding ability of hydroxyfullerene to divalent and trivalent metals

Metal contamination of water is a serious challenge faced by environmental chemists, but there is also economic value in the removal of metals for recycling or extraction. Our prior observation that hydroxyfullerenes [C₆₀O_x(OH)_y]ⁿ⁻ act as chelate agents to Fe³⁺ suggests that these material, or derivatives, may be used for co-precipitation. We report the removal of main group (Al³⁺, Ag⁺, Ca²⁺, and Zn²⁺) as well transition metal (Fe³⁺, Co²⁺, Cu²⁺, Mn²⁺, and Ni²⁺) and lanthanide (La³⁺and Nd³⁺) ions from solution. The resulting complexes have been characterized by XPS, SEM, TEM, and DLS. The competitive binding with Fe³⁺ shows that the binding affinity with hydroxyfullerenes follows the order: Zn²⁺ > Co²⁺ > Mn²⁺ > Ni²⁺ > La³⁺ > Nd³⁺ > Cd²⁺ > Cu²⁺ > Ag⁺ > Ca²⁺ > Fe³⁺ > Al³⁺. The relative binding ability is dependent on ionic radius; however, divergent trends are observed for M²⁺ and M³⁺, i.e., for divalent ions the binding is stronger for smaller ions, while the opposite trend is observed for trivalent ions. Previously the coordination of hydroxyfullerene to metals was assumed to be analogous to a 1,2-diol or catechol. However, while ab initio calculations using [M(catecholate)_n]ⁿ⁻ (n = 2, 3) provide an explanation of the observed trend for M²⁺, the use of cis–cis-1,3,5-cyclohexanetriol and cis-1,3-cyclohexanediol as model ligands provides insight into the relative binding efficiency for M³⁺.