Peak from plastics?

[Mattias]

Hi,

I was asked to run LC/MS on a new unknown peak in one of our products. The peak shows up clearly in UV and TIC (positive). The UV spectrum shows one UVmax at 200 nm.

But the MS-spectrum puzzles me...
From about 300 Da and up I have a peak for every second mass (+2 all the time). The intensity of the peaks goes slowly up and down over the mass scale with a frequency of 57.92 Da (Q-TOF with lockmass is used). Examples of masses with highest intensity: 490.331 - 548.252 - 606.173.

Has anyone of you seen this before, and knows what it is? I guess it is a polymer of some kind, and that the +2 phenomenon comes from Cl or Br?

[lmh]

58 is usually NaCl as clusters [nNaCl + Na](+), and this has copious +2 peaks because each cluster has lots of +2 peaks for the chlorine isotope. For the interested, the intensity of each cluster varies with the stability of the structure. NaCl is one of those things that forms its proper cubic crystal structure even at a very small scale (not everything does!), so a 3*3*3 lattice containing 14 Na and 13 Cl (or vice versa in negative mode) is often particularly intense. There should be peak-free gaps between the clusters though, as with 10Cl there can only be peaks up to +20...

[Mattias]

Thanks! I have never heard of NaCl clusters before, but they seem to explain what I am seeing above.

I have some low molecular masses for this unknown peak as well, and I think that I should focus on them at the moment.

I got a more concentrated and pure sample of the impurity, and now I can see that I get signals at 190.96, 192.97, 172.96 and 174.96 Da (ESI, positive mode). Cl/Br must be present?

I can also detect a UVspectrum with UVmax at 210, 239 and 285 nm.

Has anyone of you seen these masses before? I have a strong feeling that the impurity must come from the plastic (it is packaged in a very soft and sticky type of plastic). Massbank.jp did not give any answers this time.

[Mattias]

I performed a prediction of the elemental composition, and the top hit turned out to be the same formula as dichlorobenzoic acid. It sounds reasoble as this molecule would easily release water to give the -18 peak that I am seeing. But can dichlorobenzoic acid really be protonated?

Anyhow, I have ordered the most common isomers (i.e. the cheapest). I will cross my fingers.

[krickos]

Hi

This might be related to your issue:

The experiment showed that 2,4-dichlorobenzoic acid was released from silicone tubes placed in contact with natural soil and rainwater. The compound is probably a transformation product from bis (2,4-dichlorobenzoyl)peroxide used in production of silicone rubber.

Taken from:
Release of 2,4 - dichlorobenzoic acid from silicone tubing
Environmental Technology
Volume 11, Issue 9, 1990, Pages 863 - 866
Authors: Per Nmberga; Elsa Voldbjerg Sflrensenb

[Mattias]

Thank you very much!

I downloaded this article, and 2,4-dichlorobenzoic acid has exactly the same UV-spectrum as my unknown peak! The product is packaged in PVC, but silicone tubing is used in the manufacturing process.

I will get the standards on Monday. I will let you know how it goes, but it feels like a 90% chance right now.

Does anyone know if 2,4-dichlorobenzoic acid is toxic?

[Mattias]

For information: I got a perfect match for 2,4-dichlorobenzoic acid!

I still don't understand how I can add a proton to this molecule, but apparantly it was possible.

[krickos]

Hi

Well at least chemically speaking, think it was confirmed by NMR already in the 60s that benzoic acid could be protonated.

Also mentioned at NIST website: http://webbook.nist.gov/cgi/cbook.cgi?I ... 83&Mask=40

But I am not "good enough" to say that it relates to LC/MS.

Nice that you got a good match.

[aldehyde]

Wow lot of great help in this thread, good job guys.

[Camisotro]

For information: I got a perfect match for 2,4-dichlorobenzoic acid!

I still don't understand how I can add a proton to this molecule, but apparantly it was possible.

Can't it be protonated at the carbonyl? Assuming that your ionization method is ESI-MS, this is not highly unexpected. The positive ions formed in highly charged droplets can be protonated in otherwise less favourable places.

I found that a few carboxylic acids would produce strong [M + Na]+ adducts in methanol/water mixtures; but with low concentrations of ammonium acetate or formic acid as an additive, the [M + H]+ peak became much stronger. I don't know the mechanistic details of why this is the case.

[Mattias]

I guess it has to be the carbonyl that is being protonated, because I assume that benzene does not produce any [M+H] (?)

I am running ESI-MS, and it is not always easy to predict how the signal will look. It has happened a couple of times that I have been able to increase the signal a lot by adding a small amount of Na+ to the mobile phase (not this case though).

This case with 2,4-dichlorobenzoic acid was a huge success for me. There have been a quite large team working with this peak for weeks, but they were all stuck in the strange cluster pattern, that everyone thought was a polymer. By using QTOF, exact mass, this forum and Internet solved the identification in a couple of days...

[Camisotro]

[quote="Mattias"
I am running ESI-MS, and it is not always easy to predict how the signal will look. It has happened a couple of times that I have been able to increase the signal a lot by adding a small amount of Na+ to the mobile phase (not this case though).[/quote]

The problem with [M + Na]+ adducts is, the trace levels of Na+ in the mobile phase can vary greatly and change over time due to glassware. Add to it the fact that non-volatile additives are detrimental to ESI-MS, and it becomes a recipe for inconsistent results. So it's always been my impression that a better strategy is to use additives which boost the [M + H]+ signal or which create additive-based adducts such as [M + NH4]+.

Anyway, I'm glad you got your answer