Amino column separation reproducability

[benzech]

I have been using an Amino column (which I am un familiar with for the most part) to purify a component of a soil matrix which may be an amino acid or a carboxylic acid. The retention time and peak shape was stable until I upped the inj. vol. from 10 to 100 ul. Then I started getting all sorts of weird things, peak broadening, retention time varying by a minute or so, and half the material eluting in a broad ugly region four or five minutes before the actual peak elutes. The Mobile phases are A: 20mM ammonium formate, pH 2.8 and B: 98:2 ACN/H20 The condtions start with 100% B and ramp up the A. We expect to use MS hence the volatile buffer. The reproducibility is bad no matter how much we equilibrate before the run.

Any ideas?

[Bryan Evans]

Some possibilities:

- column was fouled with crud at large inj. volume
- injection volume overload
- mass overload

What column dimensions are you using? Sample solvent?
Sample cleaning procedure?

[benzech]

Thank you for your ideas, an alltech 4.6 x 150, 5 micron nuleosil amino column is used. the sample is in water (not ideal, but we had anticipated being able to get away with RP conditions). The sample was cleaned up with EtOAc partition and it is transparent and light-brown colored. The sample was syringe filtered through PTFE, 45 u before injection.

Still, column became clogged and had to be backflushed just after I posted my first comments.

I had considered column overload, but the peak looks fine some times and not at others, there is no reproducibility.

[JGK]

The sample was syringe filtered through PTFE, 45 u before injection.

Why use a hydrophobic filter medium? A nylon or PVDF membrane of similar pore size would be better with an aqueous based sample.

[Bryan Evans]

Maybe the salt in MP + salt in sample crashes out at high % organic?
Maybe you can change B to 90:10 acn:water

[Uwe Neue]

The ugly early peak of the analyte could be due to lack of pH adjustment of the sample, or due to the high water content of the sample. It sounds similar to DMSO injections in RP. See if you can get away with a reasonably high acetonitrile concentration in the sample without precipitating it.

[lmh]

Possibility (1): This is unfortunately exactly what I'd expect when you use what amounts to a HILIC method and inject in the strongly eluting solvent. You'll have to get the water content of the samples down, or it's always going to be edgy. Fortunately you already have an estimate of what water content you can use; a 10uL injection works OK, so you know you can put 10uL of water on your column without immediately eluting things.

Personally I'd take a sample at the concentrated end of what you are working with, and do a series of dilutions in acetonitrile to 10%, 20%, 40% etc., and see what you can get away with. As Uwe correctly said, sooner or later there's a danger of precipitating the sample. Leave the dilutions sitting in the autosampler for an hour or two before injecting. Sometimes things take time to precipitate.

Possibility (2): yes, you've got a very dirty sample, which may make diagnosing the problem a bit more complicated. You might be able to sort out any water-overload problems better by working with a clean standard until you've got that bit of the method sorted out.