HPLC Method for BAC

[lxBATESxl]

I work for a company that produces a ~0.12% BAC soap. Historically we have followed the exact HPLC method from the USP; however, we have typically gotten results of lower values for the quantification of the active than what has been truly added to the sample. I have postulated that the reason for this is the interference to the signal of the BAC by the other raw materials present in the soap (specifically glycerin and Caltaine C-35).

When calibrating with solutions composed of strictly BAC and water, testing a sample of strictly BAC and water results in an extremely accurate reading by the instrument. Specifically, I was getting 0.1200% and 0.1201% results on an injection of a sample that I formulated to be 0.12%. I followed this up by shooting a sample of our soap (with other raw material components) that I formulated to be 0.12% BAC and the result was between 0.10% and 0.11%. From this I was confident that I could conclude this inaccuracy was due to the matrix of the soap rather than an inconsistency of the HPLC. I confirmed these findings by performing this procedure in triplicate which produced nearly identical results.

Due to this conclusion, I then developed a calibration method that would account for the matrix present in our soap. I made a 0.09%, a 0.12%, and a 0.15% BAC soap solution that included the other raw materials at their regular concentrations. I then injected each of these samples by weighing up 24-25g samples of each and diluting them into a 100mL volumetric with deionized water. I set the calibration level by multiplying the sample weight by the %BAC used in formulation. After calibrating, I made up a separate sample of our soap at 0.12% and ran it as a check standard. Across 15-20 injections, this sample ran in the range of 0.1175-0.1125%.

I concluded that for accurate results it is imperative to account for the interference of the raw materials in the calibration solutions. I then proceeded to perform a validation on this method by testing linearity, repeatability, ruggedness, and an outside verification of a sample. As far as I’m concerned, this method is ready to be implemented into our FDA regulated procedure.
I am wondering if anyone else in the industry does something like this testing scheme by calibrating specific to the matrix of a sample that they intend to test.

One of the other issues with this method is that our BAC has 3 homologues (C12, C14, and C16). One region of concern is that when in the soap matrix the C16 peak is relatively small and almost always requires manual integration. I have spent hours attempting to adjust the integration events to allow for auto-integration of the chromatograms without much success for application across multiple chromatograms. When shooting a sample that does not include the accessory raw materials of the soap, the chromatography is very clean and the C16 peak separates away from the baseline enough for consistent auto-integration with our current integration settings. Our QA personnel is very hesitant to allow us to continue to manual integrate due to possible further questioning from auditors.

Would it be possible to separate glycerin and Caltaine C-35 from the solution to allow for cleaner chromatography?

Caltaine C-35 is Cocamidopropyl Betaine
BAC is also known as BZK (specifically Benzalkonium Chloride)

[Consumer Products Guy]

Yes, benzalkonium chloride typically has 3 or 4 peaks due to homologs. We developed a procedure for liquid soaps that made the BAC peaks to elute as a single peak, and cGMP-validated that, but that is proprietary, sorry.

For your instance:
1. Why don't you just use the areas of the C12 and C14 peaks, those should be consistent with your reference standard, assuming that your reference standard is from the manufacturer of the BAC that you use. If you use a different source of BAC, then all bets are off for that.

2. Coacamidopropyl betaine is an amphoteric surfactant that I suspect is tying up some of the BAC.

This is NOT a straightforward slam-dunk assay.

[Rndirk]

About your first question. Performing "matrix-matched calibration" is very common in chromatography, especially with mass spectrometric detection.

In my opinion, your conclusion that interferences from your matrix (soap) have to be accounted for, is questionable. If I understand correctly, injecting BAC in soap solution (=sample) gives a lower area for the same concentration of BAC in water (=calibration). If the soap would give an interference in your chromatogram, it would be higher. Note that I'm assuming the detection is spectroscopic (UV, DAD).

Could you elaborate about the interference part?

EDIT: I just noticed that the poster above me gave a possible explanation for the lower area!

[lxBATESxl]

Thank you for the responses. Here is some more information that I can provide with a couple of chromatographs. I don't mind sharing this due to the fact that this method is nearly identical to the USP method.

Solvent A: HPLC Grade Acetonitrile,
Mobile Phase B: 0.075M Sodium Acetate, filtered with 0.45um and then adjusted to a pH of 5.0 with glacial acetic acid;
Ratio of 62A:38B at 1mL/min

Column: C18 Column

Detector is a dual wave Deuterium Lamp that is set to:
Wavelength 1: 262 with an attenuation of 0.100
Wavelength 2: 280 with an attenuation of 2.000

The exact raw materials that are included are
1) Di Water,
2) Caltaine C-35,
3) Glycerin,
4) Polyethylene-polypropylene Glycol,
5) Lonza GeoGard221,
6) Nobac BZK NF 50% BAC,
7) Dowanol ePH,
Tetrasodium EDTA,

Here is a picture of a chromatogram of a soap that is composed of ~99.2% water, 0.24% 50%BAC, and the remaining ~0.55% with the remaining raw materials excluding 8.
https://imgur.com/a/I476a

Here is a picture of a soap that is composed of
~86% Water,
~6.5% Caltaine C-35,
~5% Glycerine,
~1% Polyethylene-polypropylene glycol,
-0.24% 50% BAC,
and the remainder with minimal amounts of the other raw materials
https://imgur.com/5hYKhZX

Both of the chromatograph are results from the same day and both are batched up to 0.12% BAC. As you can see, the second chromatograph has noticeably lower peaks, specifically it can be seen in the area.

[James_Ball]

About your first question. Performing "matrix-matched calibration" is very common in chromatography, especially with mass spectrometric detection.

In my opinion, your conclusion that interferences from your matrix (soap) have to be accounted for, is questionable. If I understand correctly, injecting BAC in soap solution (=sample) gives a lower area for the same concentration of BAC in water (=calibration). If the soap would give an interference in your chromatogram, it would be higher. Note that I'm assuming the detection is spectroscopic (UV, DAD).

Could you elaborate about the interference part?

EDIT: I just noticed that the poster above me gave a possible explanation for the lower area!

The lower area for the target analyte can also come from a peak to either side of the target that does not allow the baseline to fall to zero. If that happens then the start and end of the peak will be elevated above true zero and the area will be lower than it should be. If you calibrate using a matrix matched standard it will have the same interference and should help compensate.

If manual integration is documented on how it should be done and why, and you follow that instruction each time then it should pass an audit. Consistency is what matters most, at least with many auditors.

[Rndirk]

The lower area for the target analyte can also come from a peak to either side of the target that does not allow the baseline to fall to zero. If that happens then the start and end of the peak will be elevated above true zero and the area will be lower than it should be. If you calibrate using a matrix matched standard it will have the same interference and should help compensate.

You're right James, my bad.

Seeing the chromatograms, it appears to me that the background is elevated rather than the signal being lower, resulting in a lower area. I think calibration with matrix is the right thing to do, especially since there's not much freedom to change anything?

About the problem with auto-integration for C18. I think the only thing that could help you is getting a sharper and/or higher peak. Different ways to do this but what is allowed?