HPLC Sunscreen method Validation

[mikhalidis]

Hey everyone,
I've been lurking around here a lot lately and I decided I should finally make an account and post my question. I'm sort of new to this so please forgive me if I ask stupid questions.

First of all, let me introduce my situation. I'm a new hire at the in-house analytical lab of a cosmetic company that manufactures sunscreens. I was asked to develop and validate a method to assay Benzophenone-3, Octocrylene, Octyl methoxycinnamate, Butyl Methoxydibenzoylmethane (avobenzone or Parsol1789), 2-ethylhexyl salicylate, and homosalate in one injection. I am not an Analytical Chemist neither have I had any formal training in HPLC but I have educated myself by reading all relevant literature. The company already had a method in place but that method did not meet the USP and FDA requirements for capacity factors and for the resolution between Octyl methoxycinnamate and Butyl Methoxydibenzoylmethane. I tried several other methods that were published online that successfully fulfilled what I needed but for some reason I still had the same problem of not being able to resolve Octyl methoxycinnamate and Butyl Methoxydibenzoylmethane. I tried 3 different columns and various mobile phases and gradient tables to no avail. So I came here to seek experts help.

I have a Shimadzu VP system with an SIL-10AD autoinjector, two LC-10AT pumps, a degasser, SPD-10A UV-Vis detector that can detect up to two wavelength simultaneously, and a CTO-10A oven. The maintenance that I have performed so far is change the UV lamp and the needle seal.

The old method we used had an 85:15 (Methanol:0.1% Trifluoroacetic Acid) flow rate of 1.0 mL/min at 325nm on a Phenomenex Luna C18(2) 150x4.60mm 5 micron. At 26 degrees celsius with 5 microliters injection. The resolution between Octyl methoxycinnamate and Butyl Methoxydibenzoylmethane was always below 1.8

I have also tried a method using 80:20 (Ethanol: 0.8%Acetic Acid) flow rate of 0.7 mL/min at 313 and 354 nm on an Acclaim 100x4.60 mm 3 micron. At 26 degrees celsius with 5 microliters injection.The resolution between Octyl methoxycinnamate and Butyl Methoxydibenzoylmethane was always below 1.6.

The third method was me just messing around with a 4.6x50mm 2.7 micron Agilent column also to no avail.

Whenever I was able to increase resolution a bit, the peaks become too broad and distorted so I'm thinking my HPLC system has too much dead volume, or I need a smaller size detector cell. Please help in any way you can.

[Consumer Products Guy]

First of all, doesn't USP have a current limit of resolution > 1.5 for the critical pair ? I'm refering to IPC-USP 7th Annual Scientific Meeting, February 6 - 7, 2008, Hyderabad International Convention Center, Hyderabad, India, Revised USP System Suitability Parameters, page 9.

Second, resolution/assay of those six SPF actives in one injection, may not get much better than what you already have (you state "The resolution between Octyl methoxycinnamate and Butyl Methoxydibenzoylmethane was always below 1.8"), and better than 1.5 is your target. Also, homosalate is present as cis- and trans-homosalate, so you really have seven peaks to separate, not six.

Why is "management" so insistent on doing this in one injection? You sample could be assayed/validated for several of the components using Conditions I, then same solutions be assayed/validated using conditions II. Different conditions could include different column (e.g. automated switching valve, different mobile phase/concentrations, maybe just even different temperature.

For example, in our early method development days for mixtures of SPFs, we found some mixtures resolved better at 35C, but other mixtures at 25C, and we had to be ready for whichever mixture product development fianlly decided to go with.

[mikhalidis]

I am aware that homosalate splits into two peaks. I am not having a problem with that. The only problem I am having is the resolution of Avobenzone and Octinoxate while still having decent peak shapes. Even at the 1.8 or 1.6 that I am having, the peaks are soo broad and distorted and there's a lot of peak tailing so RSD is high and I can't validate the method. Also, we have certain products that have all 6 analytes in them that's why they want to do it all in the same injection. Finally, I haven't even found condition that resolve Octinoxate and Avobenzone, if you could lead me to such a thing, I would set up two different method and run each sample twice once in each method to get the results I need.

[Consumer Products Guy]

Our test method conditions are considered proprietary, so I can't share those exactly. Look at the attached separation from Adamson Labs http://www.adamsonlab.com/SUNSCREEN_SPF_30_large.gif

I might suggest trying a modern RP-18 column 50mm x 3mm with mobile phase of 45% THF and 55% H2O with 0.5% acetic acid, at 0.50 ml/min, and starting from there.

[mikhalidis]

Thank you so much! I will start from there. 50 x 3 mm or 3 microns??

[Consumer Products Guy]

Thank you so much! I will start from there. 50 x 3 mm or 3 microns??

50mm x 3mm i.d. 3 micron particle size. Of course, when one resolves one pair, sometimes other components move to muddy up the situation...six is a lot of SPF components, wonder why your product development felt the need to use all six... I'd sure wouldn't want to be the one who needs to document System Suitability for all those each time it's run !!!

[mikhalidis]

Thank god the software I use has a system suitability option built-in. You define the peaks, define the tests, and it automatically kicks out a pretty lengthy report for all the runs and whether they passed the tests or failed. Octocrylene and homosalate are pretty muddy right now lol but I will see what I can do. But so far this looks awesome. I'm getting a 4.2 resolution between Avobenzone and Octinoxate. Thank you sooooooooo much!

[juddc]

A couple of things:

First, I've had good luck with Waters Symmetry C8 3.5uM 100 mm x 4.6mm (or 3.0 or 2.1) and a methanol based MP at low pH (0.1% H3PO4, pH 2.2 works well). I typically achieve Rs of ~2.5 between OMC and P1789. My method separates Benzophenone-3, P1789, OMC, Ethylhexyl salicylate and Octocrylene in that order. I can't tell you about homosalate.

With these analytes, solvent changes can bring about real changes in selectivity, including changes of elution order, and one thing you need to watch for (in my opinion) is the fact the P1789 undergoes a keto-enol tautomerism and you can clearly see both tautomers in the chromatogram (look @ 266 nm for the diketo - you'll see it). I would try to avoid getting any of your analytes to be quantified between these two because the baseline between them isn't quite "at baseline". Nobody EVER mentions this, but I think it's kind of important. Also, lower pH is important to keep the eqiuilibrium toward the enol end of things. Raising the pH - even just to 3 or so with acetate - will give you a somewhat poorer peak shape with P1789.

Shoot me a note at christopher dot judd at basf dot com if I may be of any further assistance!

[mikhalidis]

Thanks Chris I'll be shooting you an email. P.S. I found earlier literature that had a method that was similar to yours. I employed everything in the method which claimed a critical Rs of at least 2.5 for all the analytes I am interested in but the results came nothing close to what the article but maybe because it was a gradient based elution. I'll give your method a try