Fingerprinting Conflict Minerals

When it comes to “conflict minerals,” diamonds get the most attention. But diamonds aren’t the only minerals being sold to finance wars. In the Democratic Republic of the Congo, minerals such as cassiterite—a tin oxide—and columbite-tantalite, also known as coltan—a source of niobium and tantalum—are mined and sold to underwrite militias.

To prevent companies from inadvertently funding wars through the purchase of conflict minerals, the U.S. Congress passed legislation in 2010 as part of the Dodd-Frank Act that requires firms to verify that their products contain no cassiterite, columbite-tantalite, wolframite, or gold obtained from the Democratic Republic of the Congo and its neighbors. The problem is that legitimate mining operations can be located near militia-controlled ones.

To distinguish between minerals obtained from conflict regions and those from approved areas, Richard R. Hark, a chemistry professor at Juniata College, in Huntingdon, Pa., is developing a method based on laser-induced breakdown spectroscopy (LIBS). The method provides a geochemical “fingerprint” for such minerals. Two of his students, Ian K. Potter and Benjamin M. Tansi, described the use of LIBS to discriminate among ore samples at the American Chemical Society national meeting in San Diego last month.

LIBS generates those fingerprints with a laser-induced plasma that breaks down samples into atoms and excites them to emit light at specific wavelengths. The resulting spectrum can contain thousands of spectral lines. Pattern recognition and statistical methods, typically partial least squares discriminant analysis, can tease out subtle differences between samples. The potential to do real-time analysis in the field makes this an attractive approach.

Geologically, the ores formed from solutions associated with the passage of molten rock through Earth’s crust. The ores contain trace elements—especially rare-earth elements—that reflect this origin. Although any element can be excited during LIBS, the rare-earth elements account for most of the differences between signatures.

Immediately after mining, the ore is processed into so-called ore concentrate. It is this ore concentrate that Hark wants to be able to fingerprint. “It ends up looking like a powder or something more akin to Grape-Nuts cereal, small granular bits,” Hark says. “At that point, the signature of the crust is still going to be present.” But after smelting and especially when ores from different mines are mixed, identification becomes more complicated. And after the metal has been extracted from the ore, there may be no way to determine provenance. “We’re not sure how much of the fingerprint is still going to be there” at that point, Hark says.

Hark is working with Applied Spectra, a company based in Fremont, Calif., that builds commercial LIBS systems, to develop a field-portable instrument. “A portable LIBS prototype has already demonstrated significant capability,” Hark says. Now, more work is needed to make the instrument rugged enough for field use.

Frank Melcher and his colleagues at the Federal Institute for Geosciences & Natural Resources, in Hannover, Germany, are also developing techniques for differentiating conflict minerals.

Melcher sees both advantages and disadvantages in the LIBS technique relative to his group’s method, which incorporates scanning electron microscopy and laser ablation inductively coupled plasma mass spectrometry. The advantages include higher throughput and the ability to take the measurement into the field. However, the disadvantages of LIBS, Melcher says, include poor detection limits of trace elements such as rare earths, complex spectra, and the inability to get age information because of the inability to measure lead isotopes accurately.

Melcher, who provided some of the ore concentrate samples that Hark’s group analyzed, remains unconvinced that the current LIBS technique provides a suitable fingerprint.

“Ore concentrates are complex mixtures of various minerals of variable grain size in various proportions,” he says. “Just lasering some of the grains does not help because you will most probably miss the right grains. You need an additional method to identify the target grains before lasering.” For example, he says, “if you want to certify the tantalum minerals in a mixed Ta–Sn concentrate, you have to make sure you only hit the Ta grains and even within those grains that you only laser ones with the same structure.”

Nancy J. McMillan, a geologist at New Mexico State University who has worked with LIBS for about a decade, is more optimistic about LIBS’s potential for geochemical fingerprinting.

“The technology is extremely promising for many geologic applications,” she says. “The LIBS spectrum contains an enormous amount of information and provides a detailed geochemical fingerprint of the material analyzed.”

But for companies to use geochemical fingerprinting to determine the source of their raw materials, thousands of samples of known origin from around the world will be needed to construct a database. “If one is to determine the provenance of a sample, the database used for comparison must fully capture the chemical variability of each ore deposit,” McMillan says. “My work and that of Materialytics”—a company in Killeen, Texas, developing LIBS for provenance determination—“has shown that even 30 samples from each deposit may not completely characterize the deposit. LIBS is definitely a case of the more, the merrier.”

Hark is confident that, with a large enough database, LIBS will be able to pinpoint the origins of conflict minerals. “We’ve been doing this same sort of analysis on obsidian—volcanic glass—with fairly large data sets,” he says. His success in using LIBS to trace the source of obsidian-containing Neolithic artifacts gives him hope for similar success with conflict minerals.