Multiple peaks from one substance

[kata]

Dear friends,
I am studying metabolites of popular pharmaceuticals in the context of its presence in wastewater.
I am at the first stage of elaborating GC-MS method (Thermo Scientific TRACE 1300 GC, ITQ 900 Ion Trap Mass Spectrometer, Rxi-5Sil MS Cap. Column).

Now I work with metabolites of diclofenac, i.e. 4' hydroxydiclofenac and 5-hydroxydiclofenac (http://www.sigmaaldrich.com/catalog/pro ... &region=PL, http://www.sigmaaldrich.com/catalog/pro ... &region=PL).

Before analysis, analytes are derivatised with MSTFA (http://www.sigmaaldrich.com/catalog/pro ... &region=PL).

And the problem is I've got two peaks from 4'-OH diclofenac as well as 5-OH diclofenac (and actually diclofenac also).
4'OHD spectrum is: RT.15.28 - 190, 437, 402; RT. 16.21 - 302, 330, 455
5OHD spectrum is: RT. 15.62 - 437, 278, 402, 338; RT. 16.10 365, 455, 302
Diclofenac spectrum is: RT.13.45 - 190, 349, 314, RT. 242, 214, 367 (this second peak is for sure from diclofenac).
Ions are presented in descending intensities.

It is permanent situation (same results were observed from months). It is not about the column. Blank sample does not show any similar peaks.

As you can see the second peak of 4’OHD and 5OHD has similar spectrum but the intensities of ions are different. Is it possible for the second peaks to be the same compound while the intensities of ions are slightly different? And could it be some degradation product of both 4’OHD and 5OHD (I suppose they are not stable and reactive)?

In the spectrum of original diclofenac there is 190 ion most abundant in the first peak, the same as in 4OHD spectrum, in the first peak also. Is it possible for 4’OHD to be present in the spectrum of diclofenac, as a degradation product? But only the 190 ion is coincident...

Maybe some expert could give me a hint how I should interpret this results? I am planning to do stability tests for each compound, maybe it could give me some answer which peak correspond to the analyte and which is degradation product (??)... And of course I have to study fragmentation patterns of these compounds, however it is not so easy for me..
Please give me some links where I could extend my knowledge on the topic. I am fully aware, I have to learn a lot, but I have no chemical background so I ask for the help.

Thanks for your time!

[Rndirk]

Is LC-MS(/MS) an option? It would make your project much easier: not having to derivatize and seeing the molecular ion.

Because your compounds have more than 1 functional group (alcohol, carboxylic acid, amine(?)*) that can react with MSTFA, derivatization of 1 molecule can give rise to 2, 3,... different molecules.

(*I don't know if amines react with MSTFA.)

Depending on the reaction conditions, your analyte could also partially "change", for example oxidize, reduce, intramolecular rearrangement,.......

Bottom line: If you insist on using GC-MS I suggest to either

- Reconsider the derivatization procedure. If you took it from a literature source, read their observations/discussion carefully and if possible contact the authors to discuss your problem.

- Try a different derivatization procedure (if any).

EDIT: It is also possible that the derivatization went totally fine, but the compounds degrade in the liner or column. Having more information about the GC method would be helpful.

[kata]

Thank you for the answer, very much apreciated!

I wish I had an LC-MS option, but I have only old HPLC-UV-Vis in stock..

In theory, MSTFA can react with amines, however this group is hidden in the structure and in the case of diclofenac there is no evidence for derivatisation of this group.
OK, so should I try to make analysis skipping derivatisation process? I am mot sure if I get some peaks, but I will try.
But yes, double peaks are observed only in hydroxy compounds (I can see it also for 2 hydroxyibuprofen, and don't see it for example for carboxyibuprofen and desmethylnaproxen which I try also).

Unfortunatelly I did not found reference for GCMS method with sililation. LC MS are almost exclusively used. I have one paper on GC Ms with metylation but I don't remember the authors desribed any problem.

Sililation is the most popular for pharmaceuticals. Presently, I run analysis with another sililation reagent, MTBSTFA, but I suppose there will be the same situation.

My GC method is 90C to 300C in 17 minutes. Inlet temp is set to 250C, transfer line 280C and ion trap 250C.

[Rndirk]

Derivatization is probably, and unfortunately, a must to get a decent signal if it has to pass trough a gass chromatograph. A carboxylic acid function typically means a no-go for GC.

Are you sure you can't find alternative derivatization/ GCMS methods? I googled 'diclofenac GCMS' and came across a couple of recent articles.

Also this: http://www.sigmaaldrich.com/technical-d ... gents.html

Where a diagonal look learns me that diclofenac forms a byproduct during derivatization by ring-closure...

[kata]

I agree there is a need of derivatisation in order to get good signals. But maybe it would be good to check the signal and ions after adding no reagent (?).

Yes, there are lots of derivatisation methods, but most of them are based on silylation reaction. I checked 3 silylation reagents and the trend is the same (two peaks). I have got also one reagent for acylation (trifluoroacetamide), but it damage the column I use.
Thank you for the last hint. I definitely need to go through this topic..