Benzalkonium (C12,C14), poor peak shape, cell path length

[Jimmy Boily]

Hi everybody

I am not an expert in HPLC and I would like to have your ideas about a serious problem with trace analysis of benzalkonium (C12 and C14). The problem is that the peak shape is very poor compared to 2 others customers. I have the same chromatographics conditions. I use a Phenomenex Luna 5 um, 250 x 4.6 mm. mobile phase is optimized to 60/40 ACN: sodium acetate 0.1 M pH 5.0, flow rate 2.0 mL/min, injection volume 100 ul. I have a Thermo finnigan Plus HPLC. that system use a flow cell with path lengh of 50 mm. I begin to think that the problem may be the type of flow cell used. Any ideas ? (I tried to join a chromatogram but I can't..sorry).

[Dirk Hansen]

Hello Jimmy,

what kind of bad peak shape are you observing? Fronting, tailing, split peaks, braod peaks.
Concerning the column I assume it is a Cyano phase and you are following the USP protocol. Is that correct?

Dirk

[HW Mueller]

5cm path length??

[danko]

Hans, I’m not sure but I think there are this 5 cm path length cells for higher sensitivity, which is obviously relevant in this case (i.e. trace analysis).

Anyway the path length per se can not distort the peaks, but the cell volume can be one of various reasons for peak distortion – it depends on the rest of the conditions.

Jimmy, can you look up the cell volume somewhere (e.g. the detector manual or whatever)?
Also, how high go your peaks (in AU or whatever is reported)?

Btw. 2mL/min is quite a flow rate – I must say

Best Regards

[Chromatographer2010]

The best thing to do to determine if your system is your problem is to do a bandspread test.

The 5 sigma band spread should be less than around 10% of your column volume or around 350ul in your case.

I googled the following link describing a bandspread test.

http://www.waters.com/webassets/cms/sup ... 000640.pdf

If the bandspread is okay then its either your column or the fittings on the column assuming the method and chemistry is equivalent to the other customers.

Thermo says that their 50mm light pipe flow cell only has 10ul volume so that shouldn't be the problem.

[danko]

10 µL flow cell is a pretty standard size. So the attention must shift to a couple of other parameters.
Firstly, as Dirk Hansen, mentions above; a clarification of the peak shape abnormality would be helpful in finding the cause.
And then I have some additional questions to Jimmy:

1. When you say: “the same chromatographic conditionsâ€

[Jimmy Boily]

First, I would like to thank you guys ! I apologize the English used is not perfect as a French Canadian...

I resolved 75 % of my issues by changing the D2 lamp (never changed after 1 year since the installation). I also changed a check valve bringing more stability in the pump pressure. After that I try to optimize the method I and found that the use of a column temperature of 60 deg improve a lot the height of the peaks. The lamp was the main cause but I still have some problems.

I have to validate that method for a cleaning validation process. The problem is that we signed an independent consultant to help us with the first method. Unfortunately, the guy is not a specialist in analytical chemistry and cannot help me ! He gives me some acceptance criteria and he compare the chromatogram with other case he had in the past.

The nature of peak shape problem is mainly due to tailing (must be less than 2.0). Tailing seems normal with benzalkonium. I have between 1.50 and 1.80 so its not so bad. The other problem is that we use detergent containing benzalkonium from two source 40 % from benzyl and 40 % from alkyl benzalkonium. I am not sure , but It seems that the co elution time for the C14 and C16 from two different molecules is not exactly the same, I can see a second peak eluting with C14 that increasing the tailing calculation (in fact, maybe this is not tailing). The more is the % of ACN, less is the tailing (the 2 peaks merge together) and the peak height is better. But I decrease the resolution from the C12 and C14 peak. (about 1.20)..maybe the best thing would be to try to convince the consultant that the peaks are resolved even if the criteria is not met (the criteria is not less than 1.5, based on USP.).

I use a flow cell of 50 mm. The retention time of C12 and C14 are about 7,05 et 7,6.

Concerning the column I assume it is a Cyano phase and you are following the USP protocol. Is that correct? Yes.

5cm path length?? Yes. The cell volume is around 10 ul.

Hans, I’m not sure but I think there are this 5 cm path length cells for higher sensitivity, which is obviously relevant in this case (i.e. trace analysis). Yes absolutely, this is a new system introduced by Thermo, the LDM is supposed to be better.

Jimmy, can you look up the cell volume somewhere (e.g. the detector manual or whatever)?
Also, how high go your peaks (in AU or whatever is reported)? The new conditions give me a heigh signal of about 12 mAu for 25 ppm of C14.

When you say: “the same chromatographic conditionsâ€

[Jimmy Boily]

I have a USP standard of benzalkonium (cas 8001-54-5) and I diluted the solution to have comparable intensities with my BAC from the detergent used. I supperposed the chromatogram and what I got is very interesting....

My detergent contains 40 % each of

- n-Alkyl Dimethyl Benzyl Ammonium Chloride (cas 68391-5), M.W = 368.04, C23H42CIN
- n-Alkyl Dimethyl Ethylbenzyl Ammonium Chloride (cas 68956-79-6)

The two different sources of C12, C14 and C16 from different molecule may explain the fact that each species do not coelute in the exact same time resulting in slight differences in Rt (little shoulder is created). in the case of C12, I can't see it because is proportion is less than 5 % in the mixture. In the case of C14, we can see two peaks giving the impression of a tailing problem (little shoulder to the right of the peak at about 80 % height. For the C16, I can really see the two C16 peaks. When I look at the USP standard, I do not have any tailing and the resolution is greater than 2. That STD contains only one BAC molecule. The peak shape is very good. Conclusion: the problem does not come from my method but from the sample type.

Now the question….what can I do with that kind of situation regarding the good HPLC practices? What is the best way to validate that method and to demonstrate the problem…Do I have to consider that I have tailing which is not the case….

[danko]

Hi Jimmy,

Really, I can’t see how a new lamp would improve the peak shape. Are you sure you are not talking about too small peaks rather than distorted ones? Is it sensitivity you’re having trouble with?
Maybe a chromatogram will help the further troubleshooting?

Best Regards

[grzesiek]

" The problem is that we signed an independent consultant to help us with the first method. Unfortunately, the guy is not a specialist in analytical chemistry and cannot help me ! He gives me some acceptance criteria and he compare the chromatogram with other case he had in the past. " - I want his job

Hey, I just read the USP and it seems to me that you are not really following it, why?

[Chromatographer2010]

The fact that your USP standard does not tail or have additional peaks would indicate that you sample is not pure for the give Carbon count. Could be double bonds or some other modification to the BAC.

In terms of good practice you can either modify the method to seperate the contaminants and peak of interest to the requisite height for your tailing calculation (typically 4.4 or 5% of height, I have seen where people use widths at even higher heights), use a tangental skim (for the little shoulder peak) or just accept that the peak is not resolvable and base your tailing criteria on the standard which is pure. The tailing criteria is mainly to indicate that your column and system are operating properly. It does have the secondary usage in that it indicates to some degree that your peak is pure but you already know that is not the case for your sample.

[Jimmy Boily]

[quote="danko"]Hi Jimmy,

Really, I can’t see how a new lamp would improve the peak shape. Are you sure you are not talking about too small peaks rather than distorted ones? Is it sensitivity you’re having trouble with?
Maybe a chromatogram will help the further troubleshooting?

Thank you Danko ! You would be surprised to see the difference, the old D2 lamp was really weak...the peak was distorted and the sensitivity too low, my baseline was also instable. The new lamp had a great impact on the peak shape. If I have time, I will send some chromatograms...Note that for all the other methods I use, there is not real difference in peak shape but for that trace application, tha fact to change the lamp help me a lot.

I am now convinced that this is not a troubleshooting issue or any hardware failure. The sample I use is not pure, as i said, there is 2 sources of BAC. Does someone knows how to join chromatogram in that forum ? that would be helpful..

[kerri]

Jimmy,

Tom has posted instructions on how to add the chromatogram:

viewtopic.php?t=2617



Cheers
Kerri

[Jimmy Boily]

Conclusion...I finally resolved my problems with the benzalkonium chloride (BAC). The detergent used has 2 sources of BAC. The USP STD is pure. When I analysed the USP STD, the peaks were perfect. When I analysed the detergent, there is a kind of co-elution of C14 and C16. I tried everything possible to resolve that problem and there is nothing to do. I had to use another column, the Zorbax Eclipse XDB-CN from Agilent to resolve that problem.
I also changed the D2 lamp. The 50 mm flow cell is not a problem, I have 5 times the signal of 2 others customers with only 100 ul (they use 200 ul). These customers used the luna and did not have any problem..

Thank You everyone

[Jimmy Boily]

Here is the chromatograms
http://fr.tinypic.com/r/2my51jp/4