Wednesday, October 28, 2009

Punching bag

Unnoticed in the recent pile-ons, our friends from outer space have been taking it in the chin. Rasmus @ Real Climate wondered how an idea can be knocked down so often and still get off the mat.

Things Break has posts about two (one, two) new papers that falsify two of the most repeated claims of the guys from interstellar space.

Overholt et al. examine the evidence in a paper entitled Testing the link between terrestrial climate change and galactic spiral arm transit and surprise surprise, find it wanting
and another titled
Atmospheric data over a solar cycle: no connection between galactic cosmic rays and new particle formation
which says it all.

And over at delayed oscillator, we have a new participant, transient eddy who looks at a new claim by Dengel, et al, who find a correlation between galactic cosmic rays and tree growth in a Scottish grove. Eddy, finds it wanting on statistical and methodological grounds, in particular because Dengel and friends studied from a young managed forest, where the trees were pampered and not stressed. Not the sort of place you would expect to find a strong climate signal.

In following the link to Atmospheric Chemistry and Physics Discussions, one of the open review journals, Eli happened to find another:

Results from the CERN pilot CLOUD experiment by J. Duplissy and a lot of alls including the Svensmarks.

If Eli were paying taxes in the EU he would be throwing tea bags at the entrance to CERN. The paper describes some test runs from like 2006, which according to the paper show that

a) The designers of the experiment were not very aware of wall effects and
b) The aims of the experiments were not thought out as in, at the end of the day what was going to be proved

Although in the introduction very positive, the referee's comments make it clear why wall effects are important and that could be a show stopper.
To paraphrase, the experiment is designed to study how aerosols can form around ions created by cosmic rays which then grow to cloud condensation nucleii size (larger aerosols) over periods of hours, if not days. To do that, they have to have excellent control of the loss of ions, aerosols and the CCNs on the surface of the chamber.

Wall loss is an obsession in reactive chemical kinetics and is usually dealt with by coating the walls with various substances. For free radical kinetics, phosphoric acid (an early favorite), lots of strange kinds of waxes, teflon and secret sauce (Eli kids you not, he knows one aerosol guy who swears by anodized aluminum about the roughest surface you could think of on the micro/nano scale). It's not clear that any of these would work for CLOUD, because of the long design times for holding the aerosols in the chamber which means that wall loss has to be minimized so most of the aerosols are unaffected and remain long enough to be able to measure slow growth.

Worse, what the CLOUD team found was that they were getting outgassing off the walls and that the walls had a memory of what they were previously exposed to. This showed up when the temperature went up by a little bit. They say that they can clean the walls by flowing clean gas through the system for a long time based on a decrease in the bumping with time exposed to gas flow. In the original proposal it said that the walls would be cleaned by heating in vacuum (a standard procedure), but they only remain clean as long as the vacuum is maintained below 10^-9 atm. Since the cell is going to have significant amounts of air, water and SO2 in it during the experiment, the walls are always going to be saturated with water, SO2, sulfuric acid and similar during the experiment. Eli's first guess is that wall chemistry is going to be, well, complicated and confusing.



Marion Delgado said...

I remember reading an astronomer, a woman, on the spiral arm/climate thing, and thinking she was pushing a speculation as if it were highly evidenced, but not presenting that evidence very much.

A lot of "it's simply ..."

bigcitylib said...

You know I'm sure I saw a link to the whole of the Overholt paper when it was 1st mentioned at Slashdot. But now I can't find it.

John said...

Eli, at a pressure of 10-6 torr, you get a new monolayer in about a second. At 10-9 torr, it'll take a thousand seconds, about 17 minutes.
At that point, the surface is no longer pristine (if it was once). So if you are waiting for hours or days for condensation nuclei to grow, you better not neglect the effects of the wall. True, I have discussed the effect of the gas on the surface of the wall, not the effect of the wall on the gas phase reactions. But still...don't assume it's negligible.

bigcitylib said...

And I was right.

Overton's pre-print here:

bigcitylib said...

Overholt. Sorry. TeeHee.

EliRabett said...

Eli is with John, but since they are using elastomer O-rings 10^-9 atm is about the best they can get. As Eli said, this is not going to turn out well.

jyyh said...

Eli explaining one of the reasons for the simplified stories in my blog... this rabbit thinks building an additional cage for the experiment with upper atmosphere conditions might get the experiment to stabilize over time but at the same time admits this will take a long time, probably longer than it takes to build and send a wide range spectrometer/GCR detector to space. This rabbit also thinks that pions/kaons (whatever it was GCRs produced) likely need to hit atm gases that are in sufficient consentrations to have the effect some are claiming they have, and wouldn't be surprised if this sort of point would be claimed to constitute an additional variable to the QM equation that could be argued (but not with me). Maybe it's time to go political (conservation vs. ecological adaptation) for me.