717 Plus - peak arease decreas only with Normal Phase

[chris taylor]

Good morning,

I am having a problem whereby the active peak in a calcitriol standard has decreased in area to undetectable levels, using a 717 plus autosampler. Strangely, there is a solvent peak present in all injections, indicating that something is being injected. The system is under normal phase conditions. The standard is prepared in 100% miglyol (trigylceride oil) and the mobile phase is made with heptane, THF, n-propanol and ethylene dichloride.

NOW, when I switch over to reverse phase and run our in-house calibration standard (70% MeOH as MP and as diluent for standards), the areas are as expected and RSD is < 0.5%.

I had the analyst move to another LC with a similar configuration to the system under which the original problem occurred and the peak areas are just as expected. And this is using the same solutions, MP and column.

Could it be that the original LC system simply doesn't "like" the conditions? Or is it possibly a viscosity issue between the matrix in which the standard is prepared and the mobile phase? Or could the active be adsorbing to something inside the 717, or possibly there is some plumbing difference between the 2 systems?

Any insight would be much appreciated.

Thanks,

Chris Taylor

[tom jupille]

Could it be that the original LC system simply doesn't "like" the conditions? Or is it possibly a viscosity issue between the matrix in which the standard is prepared and the mobile phase? Or could the active be adsorbing to something inside the 717, or possibly there is some plumbing difference between the 2 systems?

You have pretty much covered the possibilities; I don't know that I can add anything. Off the top of my head, I'd look at viscosity as the most likely culprit. You might try playing with different mixtures of your oil and methanol as the diluent and see if a pattern emerges (maybe a critical concentration of MeOH ???).

[juddc]

If I had to guess, I'd think it might be a viscosity issue with the sample solvent. You can check this by removng your column, weighing your sample vial, running 10 or so injections with a very short run time (0.1 min), and re-weighing the vial. Remember to account for the specific gravity of your solvent. This will tell you whether the sample is actually leaving the vial.

Assuming for a moment that your sample is actually leaving the vial and getting onto the system, another thing occurs to me: If your MP isn't being mixed properly - I'm assuming you're using a gradient pump to do that - you could simply be observing excessive retention of your peak.

I've found the 717 to be gnereally forgiving regarding sample viscosity - I've done GFC on 2000 kD hyaluronic acid solutions at ~0.1 mg/mL with no issues.

Is your void peak in the right place?

Are you sure nothing's leaking anywhere?

[chris taylor]

juddc,

Thanks for your response.
It is an isocratic run. No mixing here.

What is strange is that the active peak area decreased, but the RT did not change. The solvent peak is in the 'right' place.

I am certain there are no leaks. As stated previously, the reverse phase calibration standard areas are just as I would expect them to be and at the right RT so leakage is unlikely, IMO.

Could there be some adsorbance within the 717?

Chris

[remesquaddie]

Have you tried removing the column, use a union instead, and take the system through a passivation procedure, water, 6N Nitric.
Just do not pass the nitric through the membrane degasser if you're using one.
Flush with plenty of water afterwards.
Tony Vella
WWW.HPLCworks.net

Good morning,

I am having a problem whereby the active peak in a calcitriol standard has decreased in area to undetectable levels, using a 717 plus autosampler. Strangely, there is a solvent peak present in all injections, indicating that something is being injected. The system is under normal phase conditions. The standard is prepared in 100% miglyol (trigylceride oil) and the mobile phase is made with heptane, THF, n-propanol and ethylene dichloride.

NOW, when I switch over to reverse phase and run our in-house calibration standard (70% MeOH as MP and as diluent for standards), the areas are as expected and RSD is < 0.5%.

I had the analyst move to another LC with a similar configuration to the system under which the original problem occurred and the peak areas are just as expected. And this is using the same solutions, MP and column.

Could it be that the original LC system simply doesn't "like" the conditions? Or is it possibly a viscosity issue between the matrix in which the standard is prepared and the mobile phase? Or could the active be adsorbing to something inside the 717, or possibly there is some plumbing difference between the 2 systems?

Any insight would be much appreciated.

Thanks,

Chris Taylor

[chris taylor]

juddc,

Finally I tried your suggestion. Today I ran 10 injections at 50ul/injection. My initial weight is 8.45603g and my final weight is 8.00235g. If I calculated correctly, the theoretical value should be 469.25 mg taken from the vial at the end of ten injections at this volume (where total volume, 500ul = 500 mg X the specific gravity of miglyol which is 0.935 - 0.942, I used 0.9385, right in the middle). My actual weight loss is 453.6mg, ~97% of the theoretical value. So in this case it is unlikely the viscosity issue that I had originally thought. I am stumped.

I guess next I will have to try the passivation procedure, in case the active is sticking to the internal plumbing of the instrument...



I hope that gets me somewhere!

Thanks for your help!

[LCbob]

Hi
Have you tried Tom's idea of changing the diluent to MEOH and your oil?

Regs

[chris taylor]

OK we finally figured it out. I had been barking up the wrong tree all along. It was the column! Strange, since the column in question worked perfectly on several other LC systems.

How we finally figured it out: I was running the same routine analysis on a differenct LC (using the column in question) and the peak areas began to decrease, while another peak in the chromatogram began to increase. We ordered another column and the peak was normal. We tried the new column on the original LC system in question and, voila! Peak present.

I am not sure how this peak disappeared on one LC, but was present using other LCs. But we figured it out finally and discarded the old column.

[tom jupille]

Thanks for getting back to us (it's always nice to get the final answer!).