Does internal standard have tohave purity known?

[rodrigaoman]

When using internal standard to make very accurate quantitative measurements, does internal standard need to have its purity checked or I can assume it's 100 % pure because it's a relative method of measurement?

[chromatographer1]

While its purity is not required, it should be determined that whatever impurities are present in the internal standard do not affect the analyte being measured. IE, that no peak from the internal standard interferes with the peak being measured. Likewise any internal standard should be STABLE, and UNREACTIVE in the context of the analysis.

It is also advisable that the internal standard be as pure as possible to reduce any possible regulatory displeasure.

It should respond to the means of detection in a similar manner as the measurement analyte. Finally, the internal standard should be taken from a single lot of material that has been demonstrated to be homogeneous in content, so old results can fairly be compared to new results.

For example, using benzene as an internal standard to measure impurities in pentane, it is imperative that any impurities in benzene do not match any impurities in pentane (which is not always true, by the way). However, if no impurities in benzene match the retention time of pentane and ONLY pentane is being measured, then it should be suitable as an internal standard for the measurement of pentane.

One should choose an internal std carefully, being aware of any possible challenges to the choice selected.

Think twice and choose once.

best wishes,

Rod

[Peter Apps]

In addition to what Rod said.

This depends on how you are using the internal standard to calculate concentration of analyte.

If you are comparing peak area of analyte with peak area(s) for known quantities of internal standard, then you have to know the purity of the internal standard.

If you are calibrating as the amount of analyte vs the ratio of analyte peak area to peak area for a constant quantity of IS added to each calibration level, then you do not need to know the IS purity (and you do not need to assume that it is 100% pure).

If you are using internal standards for very precise and accurate techniques such as double isotope dilution then you will probably be drawing up full uncertainty budgets for the whole analysis, and purity of IS (both in chemical and isotope terms), with its uncertainty, would be a term in the uncertainty budget.

Peter

[HW Mueller]

Rod, Peter, are you talking about calibrating (or whatever word you use these days) with an external standard, then using an internal standard only to check runs in the cases where you say that purity of IS is unimportant?

[Peter Apps]

Hello Hans

Both and neither

My first example would be if the analysis was e.g. for decane, with undecane added to the samples (i.e. "internal") in different amounts to generate the calibration of peak area vs amount of alkane. This is limited to those cases where the detector (and the sample prep) is not selective, and is most useful when there are strong matrix effects, and no sample without the analyte is available. The alternative would be known additions of the analyte, and extrapolation of the calibration back to the Y intercept.

The second example includes your case of adding an IS only to correct for sample preparation losses, injection volume, inlet discrimination etc, and also includes doing the calibration by adding a constant amount of IS, and a varying amount of target analyte, and plotting the ratio of analyte area : IS area against analyte quantity.

I am sure that the terminology is not as clearly understood, or as consistently used, as it might be. And doubtless a guideline will be drawn up for everyone to debate and ignore

Peter

[rodrigaoman]

Thank you, everybody.
I should have said, but I'll use external calibration and I'll use internal standar for my samples and standards just to check the runs.
So, by what you guys have said, I don't need to know the purity of the IS.
I know it's higher than 98%, but I'll lose so much time determining it. If there's no need, I won't lose this time.

I know the amount of IS added has to be constant. I'll add weighted amounts of internal standard. So it's impractible to add the exact mass of IS for every sample. But of course the added mass will be as close as possible (ex. 0.00505 mg; 0.00512 mg; 0.00498 mg). I'll apply a correction for the very small differences.

I have checked the IS impurities and none of them have the same retention time of my analyte.

Is my approach right?

Thanks,

Rodrigo

[chromatographer1]

Why not add your IS to the dissolution solvent and then it will be exact for all the samples, assuming you dissolve them at the same or a known concentration.

Otherwise, it would be as I would do it.

Rod

[rodrigaoman]

Because I'll have to calculate measurement uncertainty. If I add volume, the uncertainty will be higher than if I add mass. I need very small uncertainties. Uncertainties of solutions gravimetrically prepared are smaller than the ones volumetrically prepared. That's why we weight everything. We even weight solvents and work with ug/g or mg/g concentrations, instead of ug/mL or mg/mL.

Rodrigo

[chromatographer1]

That sounds fine then. Full speed ahead.

Rod

[Consumer Products Guy]

When using internal standard to make very accurate quantitative measurements, does internal standard need to have its purity checked or I can assume it's 100 % pure because it's a relative method of measurement?

If you use the exact same internal standard solution and take the same volume when you make your Rf standard mixture that you are adding to each sample, then the amount of internal standard (and, thus, its purity) doesn't even enter the equation for % analyte. Or am I missing something in the above thread?

Really: I just wanted to use the word "thus" for the first time !!!

[Peter Apps]

Hi Rodrigo

Your approach is right, but weighing IS separately into every sample at the masses that you give is likely to be a significant source of uncertainty, even if you correct for actual mass you have only three significant figures of resolution.

Details will depend on exactly how you are preparing your samples, and at what stage of the process you add the IS. Since you are weighing solvents anyway (excellent practise in my opinion) you can reduce repeatability uncertainty for IS addition by making up the IS in sufficient solvent to prepare all the samples and standards that you need to run (as CPG suggested). Measuring the mass of solvent and the mass if IS that you added you have mass fraction of IS in solvent. Measuring mass of solvent added to sample (or standard) you have mass of IS added. There will be uncertainty terms for all the masses, but since your mass of IS added is larger the relative uncertainty should be smaller, there will be a new term for the total mass of solvent in which you dissolved the IS, but that should also be small since the mass will be quite large, and the uncertainty in mass of solvent added to the samples is already in the equation with your current scheme.

As I am sure you are aware, solvent evaporation is probably the single biggest contributor to uncertainty.

Peter