Thursday, July 03, 2008

Gas background suppression during ICPMS analysis

This is a brief technical note for any ICPMS users out there who are trying to analyse elements with carrier gas interferences. Although I am not in the private sector, I still interact socially with academic researchers from time to time, and a post doc recently approached me with a analytical problem that may be related.
In laser ICPMS analysis, it is commonly assumed that any background that exists in the carrier gas prior to sample acquisition will remain constant when the laser comes on and ablation introduces sample into the system. However, as a result of this recent query I dug up some 2 year old data where I investigated this effect, and I noticed that for gas interferences with high first ionization potentials, sample introduction can reduce the background. Thus if you subtract the unreduced background from the sample signal, you will oversubtract, and underestimate your abundance.
The following table shows a) the mean background counts for one Xe and two Hg isotopes (which are both contaminants in the Ar gas); b) The mean counts during signal acquisition, and c) the ratio of signal/background.

Note that when ablating the doped Na-Ca-Al-Si NIST 612 glass, the Xe signal is reduced to 93% of the background, while the Hg is reduced to 98%. Interestingly, this effect seems to be matrix-specific; ablation of rutile only reduced the Xe to 98% while leaving Hg unchanged.
I have no idea if this sort of thing has been published- its easier to mine the data then slog through a library which I can’t access. And since I’m not an academic, I don’t particularly care. But chances are this sort of effect is machine specific, so anyone trying to measure isotopes such as 130Te, 82Se, or 204Pb would be wise to check their own instruments before trying to quantify unknowns.


Dr Mike at Otago said...

Very interesting (and somewhat scary) information. I don't try to subtract 204Hg backgrounds from 204Pb, but this might explain some funny business at m/e=48 for NIST glass vs quartz. Thanks.

Chuck said...

NIST at mass 48 will also have Ca issues.