[Compaq]

Hi!

I have an oral exam in a course called Inorganic Instrumental Analysis tomorrow, and there are some questions that I just cannot find the answer to. Please do not believe I am merely "doing it the easy way" by asking here, I have been thinking a lot about it and been trying to find some literature.

1) Why does ICP-MS have lower detection limits than ICP-OES? Is this question true for all elements, or is it more like "in general, or on average, ICP-MS will give you lower detection limits"?

Firstly, if ICP-MS has LOD, this tells me that the MS instruments handle background noise better than OES instruments; LOD is after all a multi-plum, usually 3, of the blanks' standard deviation.

The two measures I can recall for controlling background in ICP-MS, are these: Ion lenses right after the cones ensure that only positive ions are allowed into the quadruple or the reaction chamber. And the electron multiplier is positioned off-axis to reduce continuous background. Anything else I should know? (there probably is!)

In OES, however, elements can have numerous emission lines. Background correction is more "manual" than for ICP-MS. One- or two-point background corrections are sort of "user-dependent". Reflections in the mirrors can cause background noise. Chemometrics (for example Perkin Elmer's Multicomponent Spectral Fitting, MSF) can give good results with a complex background.

So, am I onto something here? Will this make sense at the exam?

I appreciate any *delete me*
Anders

[Corribus]

You're right that it all has to do with background. I'm not as familiar with the physics of detectors for MS instrument. But the photon detectors in spectrometers all have what's called a dark current that provide a natural nonzero background level. There are also effects like photon bleeding and stuff with array detectors which can raise the background levels. Fluorescence is a very low background technique, but apparently not as low background as MS.

[Compaq]

Thank you for replying. I will look up that "dark current" - it may be useful.

Another thing I thought of is this. In the ICP we get polyatomic compounds. ICP-MS efficiently handles these, but there is no controlling for that in ICP-OES. Molecules's electrons (some, there are some "illegal" transitions, yes?) are also excited by heat, and emit light when de-exciting. This light may cause spectral overlap, or it may contribute to the continuous background. Maybe this is worth mentioning as well?

[Corribus]

Molecular interferences can be a problem for both ICP-MS and ICP-OES.  Actually I believe this is more of an issue for ICP-MS because fluorescence is usually still restricted to excited atoms. As far as I know, most of the interferences from ICP-OES arise due to the multiple spectral lines for each atom - there is a high probability of a spectral line for another atom overlapping with the analyte line. In ICP-MS, interferences are due to molecular fragments (oxides, say) which coincidentally have the same or nearly the same mass as the analyte atom. 

Note that there's a distinction to be made between interferences and detection limits. These are two different things. Detection limits for ICP-MS are undoubtedly better. I believe ICP-OES actually has less interferences overall than ICP-MS because spectral lines are everywhere and you can usually find a line free of interferences for your atom of interest. In ICP-MS your only options are to select a different isotope - and there are a limited number of isotopes with sufficient abundance - or use some kind of collision/reaction cell to remove the interference before it gets to the detector.  This is one of the reasons that ICP-MS is a more difficult technique overall, even though it does offer orders of magnitude sensitivity. (Samping issues are another complication.)

[Compaq]

Would not the fact that an atom has many, many more emission lines than it has isotopes be a two-edged knife? On the one side, it seems we can only switch to another emission line free of interferences. On the other side, since there are so many emission lines throughout the wavelength range used, some lines will overlap with analyte lines?

In our lab exercise using ICP-OES (Perkin Elmer), we used a large database of emission lines.  Nowhere in there were lines from molecules. Is molecular emission really a problem in optical emission spectrometry?

[Corribus]

Would not the fact that an atom has many, many more emission lines than it has isotopes be a two-edged knife? On the one side, it seems we can only switch to another emission line free of interferences. On the other side, since there are so many emission lines throughout the wavelength range used, some lines will overlap with analyte lines?

You will almost always have overlapping lines, you will almost always have at least one line for your analyte element - and usually many - that is free of interferences. I guess you could get unlucky, but there are usually at least 5-10 strong lines for most elements so it seems this should rarely be an issue. I don't have a broad experience, though, so maybe there are some problematic elements that I'm not aware of.

In our lab exercise using ICP-OES (Perkin Elmer), we used a large database of emission lines.  Nowhere in there were lines from molecules. Is molecular emission really a problem in optical emission spectrometry?

Based on my little bit of practical ICP experience and more extensive background in fluorescence spectroscopy, I would think this is unlikely. There are exceptions of course, but for various reasons most molecular fluorescence occurs from molecules with very specific structural requirements. Still, I would not contest it if an ICP expert came here and said I was wrong.

[Compaq]

The exam went great. I got a strong B for my presentation, which is more of an A minus in American terms. Due to some poor lab reports, my overall european grade is a B.

Thank you for taking the time to discuss and help me, Corribus!

Have a nice Christmas!
Anders