Chromatography Forum

I try to develop a method for amino acids. The separation is fine, but the lifetime of the column is dramatic.

I use a Luna column and I run a gradient from 4%B to 75% B. Mobile phase A is 20 mmol phosphate pH 7.0 and mobile phase B is ACN. Column temperature is 40 degrees.

I inject 1uL sample, where the sample is dissolved in 100mmol sodium borate pH 9.0.

What happens you can see in the figure below. There is the performance shown of a column test mixture before using it for the amino acids and after approximately 100 injections. The retention time remains the same, just the peaks are spitted. If I reverse the column, the performance is much better again. The pressure was in both cases equal.



At this time I have several ideas about the nature of the problem:
 Blocked inlet frit ( but I would expect that the pressure increases).
 Voids due to mechanical shocks (which I didn’t saw in the pressure profile).
 Silica dissolved random inside the column due to the high pH of the mobile phase.
 Silica dissolved at the top of the column due to the high pH of the injection solvent ( I opened the column but wasn’t silica missing at the top).
 Precipitation by phosphate/ACN contact
 Too long equilibration at low %B with phosphate buffer..

I’m very interested in your opinion what would be the most vital explanation (and which above causes would not be likely).

Of course I also look for a solution. I have a few idea’s
 The use of a high pH resistant column (like Xbridge or maybe zorbax eXtend).
 The use of a guard column
 Use an alternative (organic buffer) instead of phosphate.
 ….

Again what would be the most logical next steps.

Thanks in advance

Since the deterioration of columns which you mention would rather split the early peaks, my guess is that you are depositing some crud on the column which is partially removed by reversing the column. Or it is at he beginning of the column and thus the secondary interaction by the crud shows up differently when the column is reversed.
Are you doing a Fmoc-Cl derivatization? Of what?

I concur with HW. A "physical" problem (e.g., a void space at the column head or a partially-plugged frit) would usually affect all the peaks in more or less the same way, with the earlier peaks in an isocratic separation typically showing the effect more obviously.

Speaking if which -- was that test mix run under the same conditions as your amino acids? It certainly looks like an isocratic separation rather than a gradient (peak widths in gradients are usually more or less constant; isocratic peak width increases with retention).

I agree with both HW and Tom. Would like to see your gradient profile. In my opinion it is a problem caused by precipitation of your phosphat buffer mainly in the micro pores of your RP material. Remember that your run at 40°C column temperature. If possible please exchange your phosphat buffer by a volatile buffer.

Sorry for the somewhat incontinence.

The picture I showed is the chromatogram of the test mixture before and after using the column for the amino acid application. I use Uracil, Acetophenone, benzene, toluene and ethyl benzene. This is isocratic with H2O/ACN. Using the isocratic test mix the problem is much earlier detected as with the amino acids. The power of using gradients somewhat masks the column degradation.

I also thought about possible introduced secondary interactions. Therefore I tested the columns with test mix SRM 870 including amytryptyline at pH 7. Beside the spitted peaks, I don’t see severe tailing for the compounds. So no extra free silanol groups are created.

The amino acid samples are just standards up to now. I only use OPA with 3-mercaptopropionic acid as derivatization reagent. So it is not likely that material plugs the frit, beside buffer (which I have filtrated with 0.22um).

As I use fluorescence, phosphate is very friendly buffer. What would be an alternative? I also think that phosphate precipitation in the micro pores or already in the frit is the most vital explanation. This could explain the stable retention time. If silica was dissolved and voids where created the uracil peak would shift to higher retention times.

I mentioned the possibility of "dirt" attaching to the column, changing its surface properties, I was NOT talking about destruction of the stat. phase itself. The explanation for this suggestion has also been given.

I mentioned the possibility of "dirt" attaching to the column, changing its surface properties, I was NOT talking about destruction of the stat. phase itself. The explanation for this suggestion has also been given.

This could be a reason ideed. I will run gradients wihout injecting. After every 10 injections I will check the column performance by the above mentioned test. If the column remains stable, I inject dirt or particles. If the column degradates it is due to precipitation in the mico pores.

I agree with previous posts, and in addition I think some large molecules in your samples may bound to your silica particles or accumulate in the column. What about pore size of your column? A 300 Angstrom or larger pore sized column can fix the problem...