Avobenzone issues in Sunscreen method

[brett]

I've perused this site but haven't found anyone with the same problem so I figure it's time for my first post. I won't influence you with my hypotheses yet but I will say this problem has had me lose a lot of sleep over the past few weeks. Also, forgive the typos, as I'm typing this on my phone.

I'm using a Brownlee SPP C18 2.7micron 3x100mm column
The mobile phase is 25% of 1% acetic acid and 75% of a mixture on methanol and acetonitrile

Sunscreens with Avobenzone at 0.1 mg/mL


Avobenzone at 1.1 mg/mL


I'm seeing zero signal for Avobenzone at 0.1 mg/mL when run by itself. I can deal with tailing peaks or coelution but when I don't get any signal, I'm not sure what to do. Any ideas?

[LiVD]

What kind of detector are you using? If it's UV-VIS you might want to try a different wavelength.

[mattmullaney]

Hi Brett,

Guess I should start by saying I Know Nothing About Sunscreens...However...

Sometimes you can find answers to questions you have in the Archives here; I don't know that this is the answer to your troubles as you are using a newer stationary phase:

http://www.lcresources.com/discus/messa ... 20040420pm

As LIVD suggests here, the older thread also mentions to pay attention to wavelength.

Maybe also there are similar methods that may be found on Elseveir's ScienceDirect website, PubMed or Wiley?

Here is a link to Thermo, too:

http://www.thermo.com/eThermo/CMA/PDFs/ ... _50843.pdf

I also do not suggest that your separation approach is incorrect--you know your work a lot better than I do, naturally. I hope that you can find something or get a hint to what you need to do from one of these sources or other places to look.

Best Wishes!

[Consumer Products Guy]

Hmm - we don't have any issues measuring avobenzone, even in a complex SPF mixture. Here's a chromatogram of a finished product (its standard mix looks just like it).
We use RP-18 column and UV or DAD detector, can't disclose more (proprietary). The procedure was cGMP-validated by us.
Also note that the run time is 8 minutes, and isocratic. Avobenzone peak is just before 3.5 minutes.
An important thing to note is that avobenzone is not that stable in some solvents or products; be sure to protect from light.

[paulw]

The chromatogram from Consumer Products Guy is similar to what I have seen (if a bit shorter).

Because you are completely losing it at low concentrations, I might suspect a possible issue with the filter. We found that we tended to lose Avobenzone when using some filter discs as opposed to others. I know for a fact there was significant loss if someone used plastic Uniprep vials.

[Consumer Products Guy]

We filter through 0.45µ PTFE, and use glass autosampler vials. We haven't investigated others for these products.

[brett]

I'm using a PDA right now at 313 and 360 nm (the spectra in the original post is from 313) and have looked at the spectral data as well. The spectral data also shows no signal when Avobenzone is at a normal concentration.

I am also using 0.45micron PTFE filter and glass vials.

The stability of Avobenzone might be worth a deeper investigation. However, I am prepping the samples and running them immediately after. Is it possible that it degrades that quickly?

Additionally, what is still troubling is why it tails so badly (when I actually see it at higher concentrations) while the other sunscreens have decent peak shape.

I have used a different sunscreen method at my previous job with good results, which is why this is so baffling. I even tried using the published method from Perkin Elmer that uses the exact same column specs and I still can't see the Avobenzone. Maybe the next thing to try is a usp standard?

Another interesting note is that I tried using a C8 column using the same method parameters and the avobenzone did show up (although it was coeluting with another peak)! Is it possible I just got a bad column?

[Consumer Products Guy]

I'm seeing zero signal for Avobenzone at 0.1 mg/mL when run by itself.

That's just over half the avobenzone concentration in our solution, so that should be fine.

Sunscreens with Avobenzone at 0.1 mg/mL


Avobenzone at 1.1 mg/mL

The chromatograms confused me also. Is the top chromatogram sunscreen containing 0.1 mg avobenzone/mL?

And the lower one sunscreen containing 1.1 mg avobenzone/mL , so the small peak on the front of peak #2 at 8.5 minutes is the avobenzone?

Can you identify the other three peaks in top chromatogram (disregard that early solvent peak)?

Did this separation previously work for you, and just recently you had issues? Is this a validated test method (in USA, sunscreens are drug products)?

[brett]

The method I'm using is still in the method development stage.

Yes, the first chromatogram I posted is a mix of oxybenzone, octocrylene, avobenzone, octinoxate, and octisalate at similar concentrations. The avobenzone in this mix is 0.1mg/mL.

The second chromatogram was spiked with avobenzone at a relatively high concentration (not actually 1.1 mg/mL). And yes, the small and horribly tailing peak at 8 mins is the avobenzone.

This is the chromatogram I get for avobenzone by itself at 1.1 mg/mL

[Consumer Products Guy]

The method I'm using is still in the method development stage.

Yes, the first chromatogram I posted is a mix of oxybenzone, octocrylene, avobenzone, octinoxate, and octisalate at similar concentrations. The avobenzone in this mix is 0.1mg/mL.

The second chromatogram was spiked with avobenzone at a relatively high concentration (not actually 1.1 mg/mL). And yes, the small and horribly tailing peak at 8 mins is the avobenzone.

This is the chromatogram I get for avobenzone by itself at 1.1 mg/mL

To be blunt: "that sucks!"

I'd change now to different C18 column and/or different mobile phase; what you currently have is not a horse you want to be riding. I's start with something that gives nice peak shape for avobenzone itself, then be concerned about your separation. I understand the goal is - as always - to assay all the APIs in a single assay.

Your product has five SPF APIs, a goodly amount. I've assayed all five of those, but if memory serves I've only looked at combinations of four of them at a time in mixtures, so don't know how your mixture of five would separate using our chromatography conditions as is. You may find that two separations are necessary to quantitate all five.

So big question: I understand the avobenzone situation, but why does your very first chromatogram show only 3 of the other 4 SPF actives? Or is that very early peak an SPF and not a "solvent" or matrix peak?

[brett]

The solvent peak is at 0.5 min and the oxybenzone is the large one at 1.4 minutes. I didn't expect the separation to be easy with 5 analytes but I did expect to actually see 5 peaks!

[paulw]

Like Consumer Products Guy, I'm not sure I ever dealt with those five in that combination, though I did deal with them all in various other combinations. When we had a sunscreen come through with both avobenzone and octocrylene, we typically used a 250mm column to achieve separation. With an 85%MeOH mobile phase isocratic (1mL of acetic per L), we had run times of about 20minutes that achieved separation (including when Homosalate was involved). We had some triethylamine HCl in the mobile phase that, though I never investigated whether it was needed, I just inherited it like that and it wasn't broke. I would expect that the run times could be shortened using columns with smaller particle size and internal diameter like you tried.

Back to your original question. It is possible that the Avobenzone is coeluting with one of your other peaks when present in the small amount. (I read the part where you said you looked at spectra data just now in another post). Do you truly get no uptick at any wavelength when running Avobenzone alone at 0.1 mg/mL?

[Consumer Products Guy]

We use a 50mm column, and that chromatogram I posted was done that way.

[paulw]

I haven't tested a sunscreen in over ten years now. We were using outdated technology, but it worked. I suspect many improvements could be made over that method. Your point about using a 50 mm column and cutting the runtime in half is one huge one.

[Consumer Products Guy]

Years ago we did use THF as mobile phase organic, helped with oxybenzone peak shape. But that was back when Type A columns were in use. We prefer to avoid THF - unless it's needed - because of odor and potential peroxides. But it's a great modifier, that's for sure, can alter selectivities, etc.