Chromatography Forum

Hi everyone,

I have a slight dilemma. I previously tested Avobenzone and Octinoxate in sunscreen samples on my Luna 5u C18 150x4.6 column with no problem. The two peaks are very close to overlapping each other, but causes no problems. I recently purchased a Gemini 5u C18 under advisement from my sales rep, stating that it would be a good replacement and even provide a little more efficiency.

It is true on the efficiency, which is good. But now when i try to analyze Avobenzone and Octinoxate, the Avobenzone is completely overlapped by the octinoxate, to where it can't even be seen. I have tried adjusting my conditions slightly (mobile phase ratio, flow rate), but the Avobenzone remains hidden. Is there anything else I can try that is within reasonable changes to a validated method to keep these two from co-eluting? I would be less than pleased if I had to re-validate the method just because I happened to change the brand of column.

Side note...the Luna is still functioning, so when I need to analyze these two components together, I switch back to the Luna. But it's not going to last forever, so I need to figure out how to adapt this method to the Gemini asap.

My standard conditions: I also analyze other sunscreen actives simultaneously with this method...

MeOH:0.1% TFA (85:15)
325 nm
1.0 mL/min

Thanks!

Did you say that Luna will be discontinued? That seems strange. Also, how do you justify changing the phase in a validated method and still consider it validated?

The selectivity is probably different which would account for your co-elution. You may have to start method development all over again on the new phase.

The Luna C18 stationary phase is still being produced. I think that the sales repesentative was just trying to get you to try a new column.

No, I didn't say the Luna was being discontinued. I purchased the Gemini upon recommendation of my sales rep. In terms of modifying a validated method and still considering it validated:

That is the point of the showing robustness in your validation. By making small but deliberate changes to the method, but still yielding acceptable results. This is also by advisement of the USP, where they have established acceptable parameters on how a method can be modified yet still be compliant and in validation. If a modification of flow rate or a small change in the ratio of your mobile phase causes drastic and unacceptable results, then the method isn't very "robust".

Red - I don't know how you could even try a column from a different supplier, and expect that to be "validated" even if it gave good separation, that's much more than "tweaking" parameters in my book. I think you should send the new column back, say "thank you", and buy another Luna column. My company doesn't use avobenzone but we've briefly looked at its separation. We used a Type"A" RP-18 13 cm long, and mobile phase is 45% (0.4% acetic acid in water) and 55% THF; we got resolution of 6.80 between benzophenone-3 and avobenzone, 2.12 between avobenzone and the octylmethoxycinnamate, and 3.25 between octylmethoxycinnamate and octyl salicylate. We never tried to improve that as there was no business need.

CPG,

Actually the Luna and Gemini are both Phenomenex columns. I was told by my Phenomenex rep that this was their newer better version and that I would see greater efficiency in my assays. So, I went ahead and made the purchase because I thought that since all the characteristics were the same (size, type, etc) that the only difference I would see would be a reduction in my run time (which I have, by about a minute). I had no idea that the selectivity would differ...and I guess the rep should have probably mentioned that to me. I wouldn't have made a different purchase by a different manufacturer, but I guess I wrongly assumed that another Phenomenex column would perform similarly.

Unfortunately this is an issue I've just now had to revisit, so it has been a few months since I made the purchase, thereby nulling my opportunity to return it. Who knew? Learn things every day I guess...=)

I'd still get on the rep's case - he probably had no idea the selectivity was different (or could have been different) or he forgot to inform you. Either case, a decent rep should make it right for you.

Every C18 will give different selectivity (even slightly), depending on your molecules. I don't think changing a column can be considered as measuring robustness. You would need to do a bridge validation at least to show that the results are similar (especially the linearity).

We try different columns all the time and compare them to our validated methods to see if we get better chrom. We've tried the Gemini and got interesting results. I just wish they would tell me what the phase was (I think its linked or something). The Gemini manufacturing is very different than the Luna. If you want a shorter runtime, try the smaller particle size luna. You may get better resolution.