nose in the air
Professor Allen Goldstein's mechanical nose analyzes compounds in the atmosphere's soup
By identifying unique sets of compounds in the air, Allen Goldstein and his team can discover what types of human actitivities are contributing to local air pollution - everything from cooking meat to brewing coffee to driving cars. The technology has the potential to play a vital role in understanding sources of pollution - and how to combat them.
When graduate students working with biogeochemistry professor Allen Goldstein first took the readings from the machine they had built, they weren't prepared for what they saw. They had hoped it would pick up minute chemicals in the air-they just hadn't expected to see such a vast array of compounds.
Goldstein's crew had designed a mechanical nose: a machine that could pick compounds out of the air with enough precision to tell researchers where they were coming from-to differentiate not just between particles and ozone, but to separate particles coming from car exhaust and diesel emissions. Scientists already had the ability to make these fine distinctions, but the typical process of taking an air sample in the field and later analyzing it in the lab was very slow and expensive. This nose would breathe: sucking in the air and analyzing compounds on the fly.
They had been tinkering with it in a room in Hilgard Hall and it seemed to be working, so they decided to try another location. The team collected a few samples in West Berkeley. The neighborhood they chose is a bit scruffy-a mix of old Victorian cottages and modern industrial buildings just behind Aquatic Park. They hadn't picked that location for any particular reason; it just happened to be where Aerosol Dynamics, Inc., a company collaborating on the project, is located. When they ran the samples, the results were remarkable. The machine took smells and made them visible. "We could see the traffic from the morning commute," Goldstein says, "and we could see caffeine in the air when the coffee shops were roasting."
But what was most amazing was that the system was picking up far more compounds than they had expected to see-from tree emissions to nicotine smoke to a variety of illegal drugs.
The point all along had been to turn the ebb and flow of invisible compounds into a legible printout. The analysis displayed a whole world of hidden acts in crisp focus.
The machine is known as TAG-an acronym so complex it contains another acronym. It stands for thermal desorption aerosol GC/MS-FID (that is, gas chromatography/mass spectrometry-flame ionization detection). TAG rolls off the tongue a little easier. Goldstein fondly refers to his crew as "the TAG team."
The TAG system was never meant to detect drugs. Goldstein and his team developed it because they had larger questions to answer, and didn't have the tools to do it. So they created the tools themselves.
TAG stands for Thermal desorption Aerosol Gas chromatography/ mass spectrometry–flame ionization detection.
Goldstein, chair of the Department of Environmental Science, Policy, and Management, is a clean-cut man who has the look of someone for whom function routinely trumps form. As an undergraduate at UC Santa Cruz in the 1980s, Goldstein majored in both chemistry and politics. "From the beginning there," Goldstein said, "I was interested in working on fundamental science that could give us the information needed to make sound policy decisions."
This theme runs through Goldstein's work, which is otherwise wide-ranging. He's worked with the National Oceanic and Atmospheric Administration (NOAA) to track global pollution flow, recorded forests breathing to understand their carbon uptake, updated the science on emissions from cattle (it turns out they produce fewer ozone-forming pollutants than previously thought), and analyzed the particles blowing into Riverside from Los Angeles to find out what they are and where they're coming from. He's also likely to play an important role in Berkeley's new Energy Biosciences Institute, looking at the potential effects that new biofuels could have on the atmosphere. It's all work that can affect the course of politics and policy.
Goldstein is particularly interested in organic compounds in the atmosphere -that is, compounds built around carbon atoms. Organic compounds, though they make up far less than one percent of the atmosphere, are important. Some are toxic and cause cancer if inhaled in high concentrations. Some react with other gases to form ozone and particles affecting both human health and the earth's climate. "You can't understand air pollution," Goldstein says, "until you understand what's naturally in the atmosphere and how that is changed by human activities."
We know a lot about the atmosphere, he says, but "when it comes to organic compounds, there's much more that we don't know." In a recent paper, Goldstein points out that if you look at all the organic compounds going into the atmosphere (aside from methane), 90 percent are emitted directly by natural plants, and we only know what happens to half of them. The other half must go somewhere-we just don't know exactly where. In the same paper, he notes that the last comprehensive summary of organic compounds in the atmosphere listed a total of 2,857. Since then, scientists have identified some 10,000 others and, as Goldstein writes, "That may be only a small fraction of the number actually present." The air, it turns out, is a great soup of compounds constantly jostling and reacting with each other.
The TAG machine works by collecting a sample for 30 minutes, then compressing and slowly heating it from 85 to 660 degrees Fahrenheit. One by one, the compounds vaporize, get separated according to their chemical properties, and fly into a chamber where they are bombarded with electrons. The pieces from this blast form a distinct pattern —a sort of fingerprint— which the scientists use to determine what is in the air.
This soup is a tricky thing to study. First of all, it's big-too big to fit in a lab. The second problem is, it's small-made up of components so tiny they are invisible. And finally, it's fast-these compounds move around constantly. They are emitted en masse then float away, diluting into the soup. They transform in midair. To really be able to "see" the air you'd need a device that could move from place to place, that could tell the difference between thousands of tiny compounds and could do so quickly.
Atmospheric chemists already had machines that could analyze gases on the fly, but nothing that could detail what organic particles are made of as they pass by. Then, about five years ago, a colleague in the private sector, Susanne Hering of Aerosol Dynamics, mentioned that her business had developed a tool to analyze the nitrogen and sulfur in particles, but could not differentiate the organic compounds. Goldstein approached her, and together they set out to build a system to combine their technologies-a portable machine that could quickly analyze particles, every hour, on location, and identify and quantify the organics present. The machine they developed together sucks air in, traps the particles, and analyzes the organics. Instead of taking samples back to the lab for analysis, the researchers take home data.
To test the machine, the team decided to give it a specific set of compounds that can only be created by burning wood-and see how it did. Dabrina Dutcher, who was working at Aerosol Dynamics at the time, created a fire and started collecting samples. When she ran the sample, she was elated-the machine gave a perfect reading for wood smoke. "It was like a miracle," Dutcher said. "It just seemed too good to be true."
With further tests and new ideas, the TAG team is continuously tweaking and improving their electronic nose, creating a system powerful enough to sniff out the sources of air pollution. And while those initial results from West Berkeley are more interesting for their novelty than their scientific value, they were indicative of the TAG system's potential. Almost anything that pollutes-a car running, a log burning, or someone cooking meat-will release a unique organic compound that the machine can identify.
When Goldstein and his team took TAG to Nova Scotia as part of a project to analyze global pollution flow, they were able to identify compounds from fires burning in Alaska. When one of Goldstein's graduate students took measurements in Riverside, he found that a whole class of compounds in the particles blowing in from Los Angeles were coming from the cooking of meat- a source of particle pollution that most people never even consider.
This is the kind of information that policy makers need if they have any hope of making meaningful change. By sniffing out the link between events on the ground and particles in the air, TAG could help regulators track down polluters and determine how to counteract the harmful effects of our emissions. "It's very invigorating as a scientist to see what we've learned affecting public policy," Goldstein says.