Consult for help with bad profile and column degradation

[JosueR29UDG]

Hello everyone. I have been reading and consulting this forum for some months now but this is my first time posting. I need help with a particullary challenging methodolgy that has been giving us problems for a couple weeks now.

We are working on a methodology for separation and quantitation of API and Specified Impurity. API is a pyrazolone family salt. Separation itself is no problem. Sample is injected at high concentration so we can also see and report unidentified impurities. We know API peak shape can not be the best because of this but our problem goes above that

Our main problem can be seen here:



Black being sample, red being diluent. First big peak is obviously API. Second peak is our expected, specificified and identified degradation product. This "baseline" that does not finish to drop happens in all systems, all methodologies and operation conditions we have tried. Its very critical as we are unable to properly quantitate the impurity because of it. Injection volume itself does not help, unless we "hide" this "tail" by lowering sample concentration so much that unidentified impurities get lost.

Other problems range from fronting of API peak, tailing, split peak, unretained API peak (eluting before death volume time and diluent peak), among others.

And beside this, we have been experiencing very, very fast degradation and death of our columns. After around 50 to 80 injections, API peak shape starts looking bad, and just after a little while, chromatographic separation profile is completely lost. We suppose stationary phase is being attacked but as we have tried very different aquous mobile phases, only possibility is the API itself.

We have tried many methdologies/aplication notes/methodologies from publications, but after a few injections our problems remain. Ill try to briefly describe main operation conditions because they all differ from each other

Columns are C18 endcapped and certified for working at high pHs. Dimensions and brands differ. Tried 8 different columns to date.
Mobile phase has been different proportions of methanol (<50%) and different aquous buffers, ranging from pH 7.0 up to 11.0. Most of them are on the 7-8 pH region. Some have been phosphate buffers, others contain triethilamine, others require sodium hidroxide for pH adjustment.
Diluent for everyhing on the system is Methanol, because of said sample degradation (hydrolisis) with pure water.
Tried both HPLC and UPLC Systems.
Temperature is below 35°C on all systems
UV is 254 nm
API does not require sample preparation. Sample contains no contaminants, excipients, reagents, etc. Just API and diluent.

Sorry if my english was bad. I am not a native speaker.

Many thanks in advance for your help and suggestions. Regards!

[vmu]

Our main problem can be seen here:

The problem is not clear. The black trace only shows a large peak with apparently moderate tailing at 4.5 min, a large peak at 10.5 min, and a small peak at 3.5 min.

[DR]

Baseline issue is one of integration parameters. Depending on your CDS, either a smaller peak width setting or a force valley event should help.

As to column life, even "high pH tolerant columns" have their limitations.
Keep shopping around, keep setting your conditions to allow for at least some retention...

[vmu]

API is a pyrazolone family salt...
This "baseline" that does not finish to drop...
Diluent for everyhing on the system is Methanol, because of said sample degradation (hydrolisis) with pure water...
Mobile phase has been different proportions of methanol (<50%) and different aquous buffers.

The "bridge" that connects the two peaks is due to conversion of one compound to the other during their passage through the column.
https://en.wikipedia.org/wiki/Pyrazolone

[LuccasLN]

Hi Josue,

If your chromatograms start out ok and then look bad after 50~80 injections it means there is no "sistematic" issue, in my opinion. i would eliminate all sorts of issues that look like this including tautomerism. in this very example, if tautomerism was the problem it would be observable since injection #1. is this what you see sinnce injection 1? by the looks of it it starts ok-ish, gets bad like this and proceeds to look terrible after 50~80 injections.

The problem appears after continoues work, therefore it is something somewhere that send tiny contributions per cycle. myy best bet is that there is something wrong with the sample: either something physical (like debris), chemical (like excipients that get highly adsorbed) or, what i think is the highest probability, chemical through a precipitation standpoint: you ate injecting 100% MeOH on a water:MeOH mixture. there is also buffering involved, so if the API or any matrix component is less hydrophobic in the pH you are working... the higher the chance of having it crashed out.

I am not going to dismiss that some of your 8 columns might have been damaged by something else like the pH but that is not your main case. what was said by other users is true. there are columns and columns regarding extreme pH resistance!

now for the good part: solutions! hey now that's a pickle. there is no good solution since your molecule is fragile agaist what is supposed to be your diluent (rule #1: always have your diluent the exact same as your MP!). so in order to do this you would have to go to another chrom mode like NP, HILIC or God knows what... i wouldn't go that way but it is something to keep in mind. Besides this we can head to the most obvious answers: either inject less sample to have lower clogging contribution per injection or have a guard column protecting the analytical column.

If the peak looks kind of bad from the start than there are other points to hit, feel free to contact me, we can discuss this as well this could be a lot of stuff including poorly endcapped phases, pH choice, the actual tautomerism, etc.

i hope this helps.

best regards,