Screening methods for LC

[Mattias]

I have been asked to look for any cross-contamination of other API in copies (legal and illegal) of our product. This means I have no clue what to look for, just that I need to be very broad in my search.

I have access to an LC/MS, but I need some screening LC methods first. The first obvious choice would be to run a long gradient between 0 - 100% ACN on a RP-column, with full PDA and MS detection.

What would you do next? Some drugs have no retention on RP-columns, so I guess some kind of ion-exchange gradients is the next step. How would you set up screening gradients for SAX and SCX? Any other good ideas in this complex but very interesting task.

[oscarBAL]

Dear Mattias; I think that ion exanchege could be a good option;' other you can do is to do a trial(This is just an idea); using MeOH and MeCN run each solvent at two diferent pH (3 and 9) runing gradients from 95 o 30% using RP and Ion Exchance; that would be a very exaustive screening, (wath out you using a column aqble to hadle these pH.

I do not know if you have acces, but the test is much straightforward, using LC/MS TOf.

Regard.

Oscar

[Uwe Neue]

I agree with your basic approach: gradient from 0 to 100% acetonitrile, with 0.1% formic acid, PDA and MS. Column: Atlantis T3 (good retention in 100% water). If you see something that comes out at V0, I suggest to do HILIC with same setup: 95% to 50% acetonitrile with 0.1% formic acid on a HILIC silica column such as Atlantis HILIC Silica.

The latter is only needed, if indeed you see some very polar constituents of your sample. I do not recommend ion-exchange, since you would need to do a lot of tap-dancing (choice of buffer, post column splitting) with this method if you want to use MS.

[SIELC_Tech]

Go with mixed-mode approach, you can retain much wider range of compounds when you employ two mechanisms on one column (ion-exchange and reverse phase).
Here are relative comparison of modern HPLC columns (including Atlantis T3) and SIELC’s columns for various compounds (Competing in HPLC Olympics - Comparison of Modern LC Columns)
http://www.sielc.com/pdf/SIELC_August_2008.pdf

You can do simple gradient of ACN or double or triple gradient (any combination of ACN, buffer concentration and buffer pH). Buffer pH and concentration will affect ionization states of your compounds and ion-exchange mechanism. ACN will affect reverse phase properties.
Nothing can compare to mixed-mode when you have compounds with various properties. You are basically applying 2-D separation performed on one column:
http://www.sielc.com/Technology_2D_Properties.html

Here is application for 8 different compounds:
http://www.sielc.com/application_130.html ((8 compounds on 50 mm column, all with different properties)

You need to look at properties of your molecules which define separation/mechanism. Ask yourself what is logP and LogD of my compounds? You don't need to pay attention to exact structures of the compounds. Just use properties of compounds affecting your retention/separation.

P.S. Mattias, I am sure that you have couple of our columns in drawers, so you don't need to buy a new one

[Hollow]

I would have had proposed doing some 0-100% gradient with ACN or MeOH @ 2 pH (2-3 and 9-10) on a C18 column which can tolerate this (e.g. XBridge C18), or pH of 2 and 8 on the Atlantis T3.
If compounds elutes at V0 in both cases (acidic/basic), trying HILIC mode.
(Pay attention to minimize the content of organic solvent in your sample when injecting to 100% water mobile phase)

Maybe you can also do some HPTLC and apply some selective color-derivatization.

[Mattias]

Thank you for all suggestions! I have had a lot of trouble to log in to the forum, thus the delay.

I have already started with the reversed-phase method, using an Acquity column and a long acetonitrile gradient (UPLC+QTOF). You can be amazed how much you can find on the Internet - and this first small generic company (in a distant part of the world) had a home page with all compounds listed. First injection showed clear responses for one painkiller, one antibiotic, one sedative and of course our molecule... (QTOF response within 10 mDa).

I am quite fond of my Primesep columns, and might give them a try as a next step. One problem is that I will probably not be able to "classic" ion-exchange with increasing salt concentrations. I doubt that the QTOF will be too happy with 100-200 mM of NH4OAc. So I am thinking of using pH gradients instead:

Primesep D: from pH 5 to pH 2 ?
Primesep 100: from pH 2 to pH 5? (with constant acetonitrile conc - 50%?). Would that be a good approach?

[SIELC_Tech]

Mattias,

You don't need to go to 100 mmol buffer unless you have basic compounds with 3 or 4 strongly basic groups (pK over 10). My guess is that you will be okay with Primesep 100 and 20 mmol ammonium formate buffer pH 3 (if you need LC/MS conditions). At pH 3 all your basic molecules will be protonated and retain by reverse phase cation-exchange mechanism. You can start your run with 0% ACN and 5 mmol ammonium formate and go to 70% ACN with 20 mmol ammonium formate.
This wide gradient will "spread" your peaks apart based on difference in ionic and hydrophobic properties. If for some reason you will have coe-eluting peaks you can change pH a little (from 3 to 3.5) and this will change ionization state for some of your compounds and change property of stationary phase a little.
In case of TFA you can do gradient from 0.05% to 0.2%. At lower pH zwitter-ionic compounds will retain as bases. Acidic compounds will retain as hydrophobic (if pKa is above 2.5) and all compounds will retain by reverse phase mechanism. If you want to use lower buffer concentration (5 mmol of formic acid) you might need to go to a weaker Primesep column. Remember that all these columns have C12 carbon chain which provides hydrophobic interactions and ion-exchange groups on the surface;
Primesep A has pKa below 0
Primesep 100 has pKa of 1
Primesep 200 has pKa of 2
Primesep C has pKa of 3
.....but I would first try whatever you have on hands (Primesep 100). I can guide you through results or help you adjusting conditions after your initial runs.

Regards,

Vlad