Thursday, November 19, 2009

The plot thickens; the ozone does not

Hello, ozone. (Image from NASA).
Here's a scientific mystery for you: where has all the ozone gone?

In 1985, researchers discovered a gaping hole in the layer of ozone-rich cloud over Antarctica. Ozone, a molecule of three oxygen atoms, may seem unassuming, but it absorbs light in the UV-B range, considerably reducing how much of it reaches the surface of the earth. When it comes to sunlight, this is the nasty stuff—skin cancer-causing, molecule-destroying, plant-killing wavelengths of ultraviolet. Ozone, happily, absorbs the lion's share before it starts frying our eyes.

So what was causing the hole? Another chemical was the culprit: chlorine. A chlorine atom will tear an oxygen atom away from O3, producing chlorine monoxide (ClO) and molecular oxygen (O2). Then things get worse: the molecular oxygen grabs the oxygen back from the chlorine atom, not to form ozone again, but to form two atoms of molecular oxygen. This leaves the chlorine atom free to do its dastardly business all over again—a vicious circle if I've ever seen one.

Where does all the chlorine come from? Some of it comes from chlorine-bearing molecules naturally present in clouds, but a lot of it, at least in the 1980s, was coming from us. In the 1930s, manufacturers began making refrigerator coolants, aerosol propellants, and cleaning solvents from chlorofluorocarbons, this wonderful a chlorine-bearing compound that, to their delight, couldn't react with anything, catch on fire, or poison anybody—except the ozone, that is.

Not long after the discovery that ozone in Antarctica was suffering thanks to the millions of safe refrigerators humming away in kitchens across America, the Montreal Protocol phased out production of so-called CFCs, although the CFCs already released will continue to do damage for years. There's just one hitch in this story: bound up in a compound like CFC, chlorine is pretty much harmless. So what was jimmying the molecular locks, letting chlorine loose on defenseless ozone molecules?

Scientists went back to UV light. This radiation could break down the CFCs at high altitudes, after which the fragments circulated to lower reaches. In colder months, the freezing clouds in the stratosphere provide a sort of lab-table for these fragments to engage in ozone-depleting reactions.

One scientist, however, wasn't buying it. In 2001, Qing-Bin Lu, then at the University of Sherbrooke in Canada, published a paper linking ozone depletion to cosmic rays, the high-energy nuclei, electrons, and protons produced in galaxies far, far away. Examining ozone and cosmic-ray data from satellites, balloons, and ground stations for the years 1997—1992, Lu found a statistical correlation between cosmic ray intensity, related to the 11-year solar cycle, and global ozone levels. He also showed in the lab that when CFCs bound in a "cloud" (water vapor condensed on a metal rod at below-freezing temperatures) were bombarded with low-energy electrons, they were a million times more likely to let loose active chlorine. (Cosmic rays rarely make it through the atmosphere intact, but instead produce showers of other particles, such as low-energy electrons, when they collide with atmospheric molecules.)

In April 2009, Lu came out with even stronger support for his idea, or so it seemed. This time his study spanned the years 1980-2007. Again, he showed a correlation between cosmic ray intensity and mean global ozone levels, and between cosmic ray intensity and fluctuations in Antarctic ozone.

"These correlations mean that nearly 100 percent of the ozone loss over Antarcitca must be driven by cosmic rays," he told Physics World, implying that UV was just not the culprit here. But other climate experts were very skeptical of Lu's attempt to explain ozone depletion by a completely new mechanism.

Maybe they shouldn't have been. An unexpected measurement presented at a conference 2007 seems to make ample living room for Lu's cosmic ray explanation. Markus Rex, an atmospheric scientist at the Alfred Wegener Institute of Polar and Marine Research in Potsdam, Germany, got a shock when he measured how quickly dichlorine peroxide (Cl2O2) broke down into chlorine and oxygen when exposed to wavelengths of light available in the stratosphere, where the vicious cycle of ozone-depleting chemistry takes place. Dichlorine peroxide is a crucial intermediate step in the cycle's second half, where chlorine oxide and molecular oxygen recombine to form two molecules of oxygen and free chlorine.

Rex found that the UV-induced re-liberation of chlorine occurred an order of magnitude more lowly than previously thought. This seemed to pull the rug out from under the accepted mechanism for ozone depletion—if UV light was really the culprit for breaking down these molecules, then we'd have much more ozone today. So now we need a catalyst. Cosmic rays seem like they might do the trick. Has Lu solved the mystery?

Perhaps—but now the plot has thickened—Physical Review Letters, the journal that published Lu's two previous cosmic ray papers, has recently accepted a paper by Rolf Müller and Jens-Uwe Strooß that picks apart Lu's work. (Incidentally, Müller called Lu's correlation potentially "spurious" in a New Scientist article about the April 2009 paper.) In his abstract, he announces, "measurements of total ozone in Antarctica do not show a compact and significant correlation with cosmic ray activity."

Müller proceeds to go through Lu's research with a fine-toothed comb; sure enough, it hitches on subtle snarls in argument and approach. He says that the connection between the solar cycle and global ozone levels is well known, and that a correlation has been observed in years (1960-1980) when "no substantial Antarctic O3 loss has been observed." So an argument for cosmic ray-induced depletion of polar ozone can't rest on correlations between the solar cycle and global ozone.

He then looks at Lu's second piece of evidence, the percentage variation of cosmic ray intensity and mean total ozone in the polar region from one October to the next in the period 1990 to 2007. Again, Müller thinks Lu's looking at the wrong indicator. Rather than looking at mean total ozone, which has confounding influences, he should have looked at the minimum of daily average total ozone in the region. When you do, Müller says, there's no correlation with cosmic ray intensity.

Finally, Müller looks at data from 1991, when Mount Pinatubo erupted, throwing up large amounts of sulfur that formed sulfate in the atmosphere. The surfaces of these particles are ideal for ozone-depleting chemical reactions to take place. Consequently, the data shows low ozone levels, but it also shows high cosmic ray intensity. So what was it, Mount Pinatubo or the cosmic rays?

I'm sort of fascinated that two completely contradictory pieces of research that examine the same dataset and reach entirely different conclusions could be published in the same journal within months of each other. Reading Müller's paper, I get a sense for how disturbingly subtle science can be. When it seems like a piece of research has all the answers, a few well-placed holes deflate the whole argument.

Whether Lu's explanation is right or wrong, chlorine is still a big problem for the ozone, so we won't go back on the Montreal Protocol. But if cosmic rays don't catalyze these crucial reactions, what does? The mystery persists.


  1. "I'm sort of fascinated that two completely contradictory pieces of research that examine the same dataset and reach entirely different conclusions could be published in the same journal within months of each other."

    This is the rule of progress from the process of "Conjectures and refutations" (See the book with this title by Karl R. Popper).

  2. Usually this happens when the journal decides that the first piece was a piece of crap. I think you were referring to Pope et al (Stan Sanders lab @ jpl) on the ClO dimer cross-section. That's been pretty well, let us say, dismembered. We are back to the old values but with much better precision.

  3. Interesting!

    In his new publication [Physics Reports 487, 141-167(2010)], Lu has now showed that there exist not only a clear correlation between cosmic ray intensity and total ozone in the polar hole but also a nice correlation between cosmic ray intensity and stratospheric cooling induced by polar ozone loss.

    It seems now clear whose explanation is right.