By Devin Powell
MINNEAPOLIS — To the naked eye, a diamond dug up in peaceful Australia is often indistinguishable from a diamond dug up to finance civil war in the Democratic Republic of the Congo. But a new means of interrogation with lasers reveals fingerprints that can trace the gems’ origins.
In a small-scale pilot study, a laser probe identified which of eight countries a diamond came from with 95 percent certainty. It did an even better job with emeralds, rubies and other gems, as well as with gold and some kinds of ore.
“With enough data, we could identify which country, which mining region, even the individual mine a mineral comes from,” says Catherine McManus, director of research at Materialytics Inc. in Killeen, Texas, who presented her team’s research October 10 at the Geological Society of America’s annual meeting.
The technique isn’t ready for commercialization, but it may one day help companies ensure their mineral purchases aren’t helping to finance bloodshed.
“This is a fundamentally new tool that could provide a better fingerprint of a material from a particular locality,” says Barbara Dutrow, a mineralogist at Louisiana State University in Baton Rouge, who was not involved in the research.
McManus and her collaborators use laser-induced breakdown spectroscopy, a technique that is fast and nondestructive. A focused laser vaporizes a minuscule bit of the material, creating plasma. This hot plasma gives off a pattern of light with bright spots at different wavelengths that can be used to figure out what the material is made of. The researchers then compare this data against their library of samples, which for diamonds includes more than 200 samples collected in eight countries.
At first, the researchers tried to use these bright spots to compare the ratios of a handful of elements, hoping this would reveal a sample’s origins. Other researchers have tried similar approaches in the past — measuring elements, impurities or imperfections in search of a physical signature that separates one place from another. But differences in element ratios proved too small for a definitive diagnosis.
“It was like trying to tell someone what Niagara Falls looks like by examining a tiny bit of water from the waterfall,” says McManus.
So McManus’ team decided to stick with raw data itself, the flood of 40,000 data points in each splash of light released by the plasma. A group of mathematicians wrote a computer program that finds and compares patterns in each sample, without a care for how the data or its patterns relate to the material itself. Similar Bayesian algorithms, which approach the problem probabilistically, have been used to analyze the rise and fall of stock prices and changes in greenhouse gases.
The more data these algorithms are fed, the better they work. Ultimately, though, the researchers will need even more samples, including from the very mines that fund violence, for the laser method to be practical for identifying conflict minerals. But acquiring such samples may be a problem beyond the reach of science.
