HELP! Peak areas getting smaller during validation!

[cody84]

Hey I ran some samples during method development and optimized my method. My results were right on target at 0.62% PCMX. Now as I am doing precision and accuracy and range my 100% sample is showing 0.56%. I reran the first sample that was working perfectly and it is also now 0.56%. So I made a dilution of the active by itself in diluent at the exact same concentration as the 100% sample and the peak area is even smaller! I'm not sure what is going on. My standard and sample peak areas are all smaller then before, but the sample peaks are even MORE smaller and so I am getting bad results. Any idea what could be causing this? I use HPLC grade MeOH as diluent. Run is 58% 0.5% Acetic and 42% ACN at 1.0ml/min. I will now run a couple injections of the same samples after having equilibrated all night.

[DR]

Detection method?
Column type?
Did you use MeOH as diluent originally? Do you know that using a diluent with a higher organic concentration than that of your mobile phase can cause a lot of problems, especially at higher injection volumes, if you are running RP?
This reminds me: Injection volume?
Column dimensions?

You have provided us with a gun and a target. We would do better with some light.

[cody84]

Ha sorry, here goes:
Using PDA deriving 280 nm
kinetex C18 100 x 4.6 @ 30 degrees
10 uL injection

And yes MeOH has always been the diluent as the sample will not dissolve in the mobile phase (it's a viscous hand soap with cholorxylenol as active). Do you think I should use the same concentration of MeOH as ACN though?

BTW I have been using this method for a couple weeks now and of course it is just when I start the validation I am getting these problems.

[carls]

Are you saying you're seeing calibration drift, i.e. no recalibration done after initial method development?

Which HPLC system are you using (brand/model)? The injection volume on some HPLC systems is affected by sample viscosity so its important to closely match the viscosity of the sample and standard.

[cody84]

Its a Waters alliance. I was just told by waters do a compression check and try a new column.
The viscosity SHOULD be close as I use only 1g sample into 100 ml MeOH and filter before injecting.

[carls]

Its a Waters alliance. I was just told by waters do a compression check and try a new column.
The viscosity SHOULD be close as I use only 1g sample into 100 ml MeOH and filter before injecting.

Compression check of the pump? Are your retention times the same (+/- 2%) as when you calibrated? What is the %RSD of RT's both between and within a run?

[Consumer Products Guy]

Cody84 - as you might have guessed by my "name", I am also experienced in chloroxylenol assays (at times it seems like I have experience in everything in this field!!!). We've done such assays by both GC after derivatization and by HPLC. For HPLC, our procedure was quite similar to yours, and we did stability assays of products for OTC.

We did not do a formal "modern" validation as our test methodology pre-dated modern cGMP regulations, but we did spike placebo liquid soap and got 100.1% recovery with RSD of 0.5%.

We also used RP-18 column, but narrower bore, and ours was a Type A silica particle, shows the era when we first did this. We also took up our samples in methanol, at about 2.5 grams per 100 ml, then filtered. We also used 280 nm detector wavelength, injected 5 ul, held column temperature at 30C. We used dilute acetic acid in water, and ACN as the mobile phase. We made the chloroxylenol standard up in methanol.

Our ACN percent was relatively low, like yours. However, after the chloroxylenol peak eluted, we instituted a gradient to clean out the column prior to the subsequent injection due to the relatively low % organic in the mobile phase, as there was other stuff from the product matrix held on column (some might be "invisible" to your detector wavelength). You might also want to try 5 ul injection, as "larger" injections of organic solvent can affect chromatography when the organic level in the mobile phase is low. On the other hand, we did have facilities injecting 10 ul with a 4.6mm i.d. column and no issues were reported.

[cody84]

I purged the injector a few times and now everything is ok and I've restarted my run. CPG I'm pretty sure I got the method from you! That was fun when they asked me to include a reference for the method Thanks btw!

I will see how everything goes and then consider including a gradient or smaller injection as it is all 'set in stone' in the protocol.

[cody84]

Ok its not working again, as soon as my sample set got to the samples the areas dropped again.

[cody84]

Yes a calibration drift but the standards are ALSO having a lower peak area, just not as much as the samples. Must be the samples although I can't explain why I worked fine for the last two weeks.

[Consumer Products Guy]

I will see how everything goes and then consider including a gradient or smaller injection as it is all 'set in stone' in the protocol.

Our QA folks would likely say that a protocol should state what test method is used. However, sometimes in a method validation information is obtained which could require a change to a test procedure. This can be detailed in a section for deviations.

Ideally, one would have enough background data to be reasonably sure that the validation would proceed smoothly. But oftentimes real world and ideal do not travel the same path. That previous sentence sounds like I should end it with "grasshopper".

[cody84]

Haha, yea that's true. Here though, they don't understand that. If I don't follow the protocol exactly I will have a lot of explaining to do to people who do not understand what I am even doing. yay.

[Consumer Products Guy]

First question, didn't see this detailed above: are your retention times and peak shapes changing as the peaks become smaller, or have those stayed the same?

Second question: how many hours on that UV lamp?

Cody - it might be also time to check out your HPLC using a known or check standard, to see if everything is stable over time. If you don't have an official type standard, since you're in the consumer products or healthcare business, you might try triclosan, methyl paraben, or propyl paraben with only change being maybe to the ACN percentage. See if such solutions of those change over time.

[cody84]

The retention times and shape are the exact same and the lamp is new, less more than half life left.

[Parham]

Hi.
are you use of a column oven in your method (temperature control).
stability of temp. is so important.
are you change your column?
your mobile phase have a low pH possible and in pH less than 3.5 Repeatability is low.
Change a stable temprature and a new column C18(such as :zorbax,nocleosil,xterra: soutable for low pH).
please time of conditioning with mobile is high 1-2 hours.
test it's i hope be luck.