Sunday, March 19, 2006

Lithium drives me batty

There are all sorts of ways to categorize the 83 long-lived, naturally occurring elements. There are the physical- solids, liquids, and gasses at the STP of your favorite planetary surface, for example. There are the chemical- based on electron orbitals, valences, electronegativity, and all that other stuff that I really ought to understand better than I do. There are the behavior-based methods- lithophiles, chalcophiles, and siderophiles. And then there’s the Sergio Leone method, to which I subscribe.

Under this system, all elements fall into three well defined groups: The Good, the Bad, and the Ugly. And under this strict categorical system, the defining characteristic under which all elements are grouped is the ease with which I can set my mass spectrometer up to analyze them at 8:30 on a hungover Friday morning.

Lithium is one of the Ugly. But unlike many of the ugly, like sodium, iodine, and tellurium, which are basically set in their ways, lithium really, really wants to be good. This is fine, because the analytical community wants it to be good as well- we can’t wait to bring it into the fold. We even give it help. When lithium starts acting out or getting ornery, we try to coax it back into the light.

The heartbreaking thing about lithium is that just when you think it is finally starting to behave in a respectable and informative manner, it acts out. Call it light element rebellion, call it three-proton obstinacy, call it what you will; the fact of the matter is that just when you think the lithium problem has finally been licked, it goes out and does something really bizarre.

In ICP-MS mass spectrometry, the main problem with Lithium is that, like most alkalis, it dissolves into the skimmer cone, and then gets reionized off the back of the cone. I don’t know if this is due to cone-plasma interactions, or if it is some sort of accidental TIMS- I honestly haven’t dug into the specifics of this phenomena. I’m not paid to do that. I’m paid to make it go away.

Specifically, my aim it to get- and keep- the lithium background down to below 1000 counts per second. That may sound like a lot, but under standard operating conditions, our system runs at about 5-10 thousand cps/ppm with a 70 micron laser spot. So if I can keep the background down, we can measure tenths of a ppm concentrations, which is roughly what you can expect in a mantle olivine, for example.

The main villain in the Lithium background drama is Lithium Borate glass. Lithium borate glasses (aka “flux”) are popular for whole rock analysis for a couple of reasons. Firstly, the correct composition melts at a reasonably low temperature, so you don’t volatilize all your non-refactories when you remelt your rock. Secondly, Lithium and Boron are fairly X-ray transparent (owing to their lack of deep electron shells), so they are great for doing XRF to get the major and minor elements. This allows scientists to measure the same samples by XRF and laser ablation ICP-MS, so that they can then pull their hair out when the two methods give contradictory results.

Trouble is, if you’re ablating something that is mostly lithium oxide, it doesn’t take very long for the Li to soak into the cones to the point where the background is in the tens of millions counts per second.

So, two months ago, I set aside a special, dedicated set of “LiBO only” cones. All LiBO work is done on them, and everything else is done on “normal” cones. Since then, the Li background has been steadily declining. Last week, in fact, we were regularly under 1000 cps, and the problem appeared to be solved. We were even able to detect Li in amphiboles and quartz. Everyone was happy. Until Friday, that is.

On Friday, the background jumped by a factor of 20. So, I checked with the Thursday user, no make sure he didn’t sneak some fluxed whole rocks into the machine while I was at seminar just to piss me off. He did not; all he analyzed was natural olivine, which generally has less Li than our tuning and calibration standards. I checked the 6/7 ratio, just to make sure I wasn’t looking at a doubly charged nitrogen problem, but the Li looked real. And I have no idea where it came from.

Oh well. I have better things to worry about over the weekend, so we’ll see what it wants to do on Tuesday.

No comments: