Method for pyridine amine derivative

[bunnahabhain]

Dears

I am struggling with a pyridine amine derivative, see picture below (exact structure confidential):

I have the dihydrochloride of this compound. There exists a method using heptanesulphonic acid as ion-pair agent, but UV detection is not sensitive enough. I'd like to go to MS, but was unlucky with all columns I tried:
- Cyanopropyl => broad peaks (2.5 min at low pH, > 10 min at neutral)
- C18 (Gemini-NX) => even broader peaks
- Primesep A (low pH, high NH4 formate, high ACN) => does not elute at all

Any other hints heartly welcome!
Thanks everybody
Jörg

[Andy Alpert]

You didn't say what your mobile phases were, but it's unlikely that they contained any chloride salts. That means that your compounds entered the column in the cloride salt form and probably exchanged chloride for some other anion during the migration through the column. Since different molecules will take part in an exchange at different times during the migration, then the result is a continuum, since the different ion pairs will differ in polarity and hence retention time. That accounts for your broad peaks. This problem is most acute with analytes with a high charge-to-mass ratio, like yours. Solutions:

1) Make sure that the mobile phase contains the same counterion that your analyte has.
2) If that's not possible, then put a high concentration of the anion that's in the mobile phase into the sample solvent prior to injection. Hopefully the anion will compete off the anion originally associated with the analyte (here, chloride).
3) Have a lot of salt in the mobile phase. Example: In HILIC of aminoglycoside antibiotics (which can have 5 amine residues in 4 sugar residues), one doesn't get symmetrical peaks unless the mobile phase contains about 120 mM salt. It can be a volatile salt, like ammonium acetate. For that matter, why not try those conditions with your compounds here? You can use our PolyHYDROXYETHYL A material, but I caution you that best results were obtained with the 1000-A pore version. That's not because aminoglycosides are so large that they posse steric hindrance problems. Rather, it's because the 1000-A material has about 25% of the surface area of the 300-A pore material, making it easier to elute these extraordinarily hydrophilic compounds.

Andy Alpert
PolyLC Inc.

[Kreall]

I would suggest C18 column with no silanols activity. The positivly charged nitrogen is interacting with silanols groups in silica and that is why you have broad peaks.
I would go propably with Ace SuperC18 or Hypersil BDS with 0,1% TFA in mobile phase.

I did once an analysis of Benzalkonium chloride on Hypersil BDS with 0,1% TFA, first time saw almost symetric peaks with quaternary ammonium cation.

[bunnahabhain]

@Kreall

I would suggest C18 column with no silanols activity. The positivly charged nitrogen is interacting with silanols groups in silica and that is why you have broad peaks.

I thought of that, this is why I chose Gemini-NX, it is very well shielded.

I would go propably with Ace SuperC18 or Hypersil BDS with 0,1% TFA in mobile phase.

I'd like to avoid TFA on my mass spec...

I did once an analysis of Benzalkonium chloride on Hypersil BDS with 0,1% TFA, first time saw almost symetric peaks with quaternary ammonium cation.

Funny, I have a method for Benzylkonium chloride on a cyano column with sodium acetate with not too bad peak shape. I will try to replace the sodium for ammonium and see what happens with this compound. Thanks for waking this well lost memory.

You didn't say what your mobile phases were, but it's unlikely that they contained any chloride salts.

Yes sorry. Mainly ammonium acetate and formate in combination with acetonitrile or methanol. Sorry I can't write down all concentrations I tried out... many.

That means that your compounds entered the column in the cloride salt form and probably exchanged chloride for some other anion during the migration through the column. Since different molecules will take part in an exchange at different times during the migration, then the result is a continuum, since the different ion pairs will differ in polarity and hence retention time. That accounts for your broad peaks.

Do you really think this effect is so dramatic for dilute solutions of about 1 - 10 mg/L?

This problem is most acute with analytes with a high charge-to-mass ratio, like yours. Solutions:

1) Make sure that the mobile phase contains the same counterion that your analyte has.
2) If that's not possible, then put a high concentration of the anion that's in the mobile phase into the sample solvent prior to injection. Hopefully the anion will compete off the anion originally associated with the analyte (here, chloride).

This sounds like a very good idea!

3) Have a lot of salt in the mobile phase. Example: In HILIC of aminoglycoside antibiotics (which can have 5 amine residues in 4 sugar residues), one doesn't get symmetrical peaks unless the mobile phase contains about 120 mM salt. It can be a volatile salt, like ammonium acetate. For that matter, why not try those conditions with your compounds here? You can use our PolyHYDROXYETHYL A material, but I caution you that best results were obtained with the 1000-A pore version. That's not because aminoglycosides are so large that they posse steric hindrance problems. Rather, it's because the 1000-A material has about 25% of the surface area of the 300-A pore material, making it easier to elute these extraordinarily hydrophilic compounds.

Andy Alpert
PolyLC Inc.

This column material sounds interesting, but I don't think that our boss will invest in new column materials right now.
Anyway thanks a lot for valuable suggestions that might help with available materials (at least available to me).
Jörg

[Kreall]

Gemini-NX is a column which is shielded to withstand high pH (incorpotaed ethylene groups between silica atoms), not to disacitvate silanols groups (https://www.brechbuehler.ch/Gemini-NX.954.0.html), so I think the gemini wasn't the best choice here.

Why do you want to avoid TFA to mass spec?

[Vlad Orlovsky]

Jorg,

there is no way you can elute double quat from Primesep A, you need to go to a much weaker column (Obelisc R):
http://www.sielc.com/Application-HPLC-S ... isc-R.html

you compound is similar to obidoxime, so you can try short Primesep C column:
http://www.sielc.com/Compound-Obidoxime.html

if X is big enough you can go to Primesep D:
http://www.sielc.com/Application-HPLC-S ... inium.html (you can replace TFA with formic acid in this application. since you are in ion-exclusion mode, you don't need strong ionic mobile phase)

Mixed-mode is always compatible with MS and you can use ammonium formate or acetate, or if interaction is not strong formic or acetic acid

[bunnahabhain]

Gemini-NX is a column which is shielded to withstand high pH (incorpotaed ethylene groups between silica atoms), not to disacitvate silanols groups (https://www.brechbuehler.ch/Gemini-NX.954.0.html), so I think the gemini wasn't the best choice here.

OK I thought this endcapping would also shield the silanol activity. Lesson learned. I tried Zorbax Eclipse XDB and this is much better. Also Andy's suggestion to add acetate to the solution I inject did help a bit.

Why do you want to avoid TFA to mass spec?

I never tried out, but heard so much bad things about killing sensitivity in negative mode unless you flush for days, that I prefer to keep my hands off it.

Jorg,

there is no way you can elute double quat from Primesep A, you need to go to a much weaker column (Obelisc R):
http://www.sielc.com/Application-HPLC-S ... isc-R.html

Obelisc R is the only one of the columns you mentioned, so I will give this a try. However, the molecule is quite more hydrophobic than paraquat, so I guess I will have to increase ACN. Thanks for all suggestions!

Jörg

[Vlad Orlovsky]

if you are doing UV try Obelisc R with low pH (2), and low organic (5-10%) and see if you can get retention. If you get retention your best choice is probably Primesep D or Primesep B2, you will get a perfect peak shape.

Regarding the end-capped columns: You need to remember that column looses end-capping and your peak shape will get worse with time. In Primesep B2 and D no matter what happens the silanols are shielded by basic group on the surface and you quats will never see acidic silanols.

I keep wondering, why people trying to jump through hoops and "fix" approaches they used to do. If you need to dig a hole in the ground you get excavator and don't do it with a fork or spoon.

[bunnahabhain]

Vlad, I can understand you feel a bit upset. I a really don't want to dig a hole with a spoon, but customers and project coordinators are quite bullheaded, sometimes. So if they see: There exists an analytical method that works with a standard column, they ask: Why can't you use it? "Because it is not good enough for our purpose" is not the answer they want to hear. It is also annoying for me (and I guess also for others like me in the same situation) who waste their lifetime on such projects. Finally, I made clear to the responsibles that a new column will be cheaper than more days worth of fiddling around with the existing material.

[Vlad Orlovsky]

I am not upset at all

i sent you email, but it came back with a response that you are out of the office.