How often do you see radical formation in ESI fragmentation?

[mrmiller]

I've found a few unknown identification papers and the occasional mechanism contains a radical. Some people I've discussed fragmentation with say you shouldn't see radical formation at all in ESI MS/MS. Trying to form my own opinion but could use more information.

Links to articles or advice from personal experience appreciated.

[Rndirk]

Recently I was working on a method that includes beta-carotene on LC-ESI-MS/MS. The parent ion in ESI+ is the same mass as the molecular mass of beta-carotene (537). After googling a bit, I assumed this was the [M]°+ radical ion. I've seen a similar effect for retinol and retinyl acetate, which are structurally related to beta-carotene. I found that the latter ones, however, show a much higher signal for a parent ion of [retinol-H2O]+.

This article seems to be about the effect for carotenoids in general, but I haven't read it in detail.

[Ian_R]

Are you only asking about fragments, or are you interested in molecular ions as well?

Have been running some diimides recently that have an extended aromatic system. They give M- ions (radicals) in negative mode. (Agilent 6530 QTof)

[lmh]

In precursor ions it is rare, but it can happen (and, as in previous replies, it depends on the class of compound). In fragments it's surprisingly common. In both cases it only happens where the free electron is highly stabilised. This usually means a lot of conjugated double bonds (and aromatic systems are good). It is essential that the free electron begins its life somewhere where it can move around the conjugated bonds freely (must be possible to draw curly-arrow diagrams that don't make pentavalent carbons or mess up the aromatic rings; for example, if you have a methoxy group on a phenolic ring, if there's enough conjugation, it's possible that you'll see loss of a methyl radical, mass 15, because a radical electron left on the oxygen can wander off round the phenyl ring; it, and an electron taken from the ring, can form a double bond to the oxygen, leaving another electron as the radical on the ring. You will never see loss of a methoxy radical in this situation, because an electron left on the ring itself can't actually go anywhere, because there's nowhere for it to go. In curly arrow terms, you will have written the ring with a double bond to one side, a single bond to the other, and a dot on the carbon, but if you try to do anything with the dot, you'll end up with a carbon in the ring with double bonds on both sides, which is geometrically not possible).
My attitude to radicals in ESI is that I prefer not to believe in them (if someone tries to convince me that they've got a radical I will first look to see whether one of their oxygens is actually a nitrogen or vice-versa; an ester bond differs from an amide bond by just one mass unit), but if there's a good chemical mechanism for their existence, and the data look good, then Yes, they can happen.

[sav]

I've found a few unknown identification papers and the occasional mechanism contains a radical.

Yes, it happens by redox mechanism of ESI (formation of radical molecular ion). I've had experience with that with carotenoids, PAHs and other redox sensitive molecules.

Links to articles or advice from personal experience appreciated.

I'd recommend this (detail explanation and lot of examples):

Electrospray and MALDI Mass Spectrometry (ed. by Richard B. Cole)
Ch.3. Electrochemistry of the Electrospray Ion Source (Gary J. Van Berkel and Vilmos Kertesz)

[sav]

Some people I've discussed fragmentation with say you shouldn't see radical formation at all in ESI MS/MS.

Yes, its called "even-electron rule", but... Although it's not common for formation of ion-radical fragments (odd-electron ions) from pseudo-molecular ions (even-electron ions) at CID, but it happens.

[Rndirk]

Are there any analytical implications when you happen to select a radical parent and/or daughter ion for a certain compound, provided there is an alternative?

[lmh]

I see no implication whatsoever for radical daughter ions. Automated transition-finding software such as Waters' Intellistart will simply find the most abundant fragment with complete disregard of how it happened, and I'm sure there are plenty of methods out there where people have used such software. Also, the daughter ion is formed in a nice, clean, reproducible environment.
I suppose it could be argued that a radical parent ion might be more susceptible to cosuppression effects than a non-radical because it's more reactive and likely to react with anything else that happens to be in the spray chamber at the same time - but I have no evidence for this.
In any case, radical parents are rarer than radical fragments.