Multilevel internal standard calibration

[apuetz]

Hello,

So I am attempting to create a 7 level calibration table within the Agilent Chemstation software. I have ran a calibration curve with my standards being 1% ethanol, 10% ethanol, 20% ethanol, 30% ethanol, 40% ethanol, 50% ethanol, and 60% ethanol. I take 1 mL of each of these standards and add 0.5 mL of acetonitrile to use as my internal standard. When I use excel for my data analysis on my x -axis I have my percentages and the Y-axis is my peak area ratios of Area Standard/Area Internal Standard. My correlation comes out to be in the 0.98-0.99 range and we obtain excellent results. However, when I then create this table in Chemstation and designate acetonitrile as my standard my calibration curve becomes a vertical line and I get a correlation of 0.49. Does anybody have any ideas of why it does this? Any help is appreciated.

[KM-USA]

Sounds like your following or trying similar to USP 611 alcohol determination.

We've validated that for our alcohol-containing finshed products. So what we did for linearity studies was a range of 80-120% of our target alcohol level, keeping the ACN internal standard amount fixed.

But for our finished products, we used a single point Level 1 calibration, and just adjusted the sample amount so the initial solution was about 2% alcohol (like USP 611 states). That would be my recommendation.

We've never failed System Suitability with this either - except occasionally for tailing factor on the alcohol peak, because we validated using the USP 611 column (-624 type) and conditions, which literally are CRAPPY for alcohol. Since supervisor won't let us tweak - within USP 621 Chromatography and FDA-ORA guidelines, we have to deal with that.

Using the USP-611 designated column was a mistake we made, but we were in a hurry, and this way wouldn't need to do robustness studies. We should've used a PEG-type column and done a full validation, that can handle aqueous injections better and has better peak shape.

[apuetz]

Km-USA

You're right I was trying to model after USP 611. The USP method confused me a little with the language though. I was having trouble understanding what values they meant when they referred to concentration of alcohol. I also thought that using a calibration curve would give slightly better results. I do not know that for a fact though. Thank you for your help.

[Bigbear]

Make sure you have the same value for the internal standard for all levels ( when adding a cal level to agilent software it assigns the concentration for that level to everything!).
Also make sure you set the quantitation type to internal standard for the internal standard.

[GasMan]

The reason that you have the "odd" result for your standard is that you do NOT have a multilevel calibration for the internal standard. You only have one point. What result does ChemStation show when you select the calibration curve for the ethanol?

Gasman

[apuetz]

GasMan: When I select the ethanol calibration it shows a linear regression curve with a r value of around 0.98. So if I am reading that right then internal standards are only used for single level calibrations not multi level? So there is no need for the curve? I would run one standard solution and one test solution both with an internal standard present?

[apuetz]

I'm sorry Gasman I may have misread, the software shows results correctly when I select the curve. When I do a multilevel calibration with the internal standard set to the correct concentration at each level the results actually come back close as well as long as the internal standard report is selected.

[GasMan]

When you select the internal standard report the calibration graph should show the two axis as ratios when you make a multilevel calibration.

Gasman

[apuetz]

Yes it does show the two axis as a ratio. So if I were to try and do the USP 611 method exactly I would run the standard solution and test solution together. Then I would create a single point calibration using the standard solution and use the internal standard report to obtain the results of percent ethanol after plugging in my dilution factors?

[WK]

Hi All,
Is there anything in chemstation that allows for different internal standard amounts in the 7 levels of calibration.
I can't see it. I have looked in all help files I can find. If you are weighing out they will be different.
Regards
WK

[apuetz]

I think there is. Under the amounts on the calibration table if you whichever compound you set as an internal standard you can manually alter the amounts per level. You just select amount and type in how much is present I believe. If someone said to determine peak response ratios would that interpret to just the ratio of the areas? Or is there a formula involved in that?