Not enough information about your samples and proposed methods to answer some of your questions, but here is some info which may help you.

• Choice of standards should not be made at random or based on convenience. They need to relate to the exact type of sample being analyzed. A related peptide for one and a related sugar for the other.
• While an internal standard (IS) is designed to allow you to take into account variations, it is only one point on a calibration curve. To cover a range of stds, you should still create a multi-point calibration curve for each compound that requires quantification.
• We have scientific reasons for using either an internal (ISTD) or external standard (ESTD). It has everything to do with which one is most appropriate. *I have attached a link to a basic article ("External vs. Internal Standard Calibration in HPLC") on this topic which may help you better understand this topic and perhaps answer many of your questions.

http://hplctips.blogspot.com/2014/03/external-vs-internal-standard.html

The choice of an internal standards very much depends on the actual problem you are trying to solve (which analytes in which matrix ...).

Generally speaking isotopically labelled standards are the best choice if you want to account for matrix effects hampering your quantification. Often such labelled substances are not availavble, so one would have to think about a substance which has a chemical structure very close to the actual analyte, but at the same time will not likely be present in the sample matrix.

In case the matrix of your samples does not show a high variability, it might be good enough to use an external (but matrix matched) calibration with a recovery test now and then in your sample series. 

What the gentlemen said above is completely true and straight to the point. Especially the talk about Matrix Effect mentioned by J. Janda. If you are targeting only 1 analyte in each analysis, let's say the peptide, that doesn't mean there are not any other compounds existing, which may cause ionization enhancement or suppression. If there was no Matrix Effect, using external calibration should be enough, whereby you construct a calibration curve with multiple concentrations (points). This is true in other types of detection whereby there is no ionization taking place, i.e. fluorescent detector. But since you are using the MS detector, then you have to be sure that a Matrix Effect is taking place.

To correct the Matrix Effect, you need to use an Internal Standard and still use the external calibration simultaneously. The selection of an Internal Standard is a tedious task that requires months of lab testing if there is no previous published work on the same topic. Isotopically labelled standards should serve the task of an internal standard, but not every single analyte has a ready Isotope in the market to use as Internal Standard. Thus, you may choose another compound as Internal Standard, and run multiple testing experiments to test the suitability of such compound as Internal Standard for your analyte.

I hope that you have a better understanding now about the importance of Internal Standards in MS analyses. Good Luck.

I assume, we should never assume, that you know the structure of your compounds you want to quantitate.  If this is the case, depending on the size of the peptides, small peptides up to about 20 amino acid residues can be synthesized with a C13 labeled in one or several of the amino acids.  This gives you the standard with the same retention time but the molecular weight that is at least 1MW greater.  Your MS/MS is more than sensitive to differentiate these.  You can kill two birds with one stone. You can add this to the sample at the beginning of sample prep and determine recovery. Because it should behave exactly like your molecule. And you can quantitate by external or internal methods.  The downside to this is that it can cost from about a thousand to a few thousand to produce.  If you have a synthesis group at your institution, work with them and they might able to do it and then you can publish a paper together.

The comments above are very good and external standard with recovery is always a good option.  More time but less expensive for the materials.

If you go the route to use some compound as a general standard be ready to defend that when you publish. It will be a big point of contention.

Good Luck.

Dear Eloi

Sorry for not having time to write you earlier.

I read about your problem with the internal standard (IS) for quantitative determinations.

I'd recommend to use midazolam (CAS # 59467-70-8; GM 325.78, C18H13ClFN3).

Mass spectrum has a very stable base with a peak at 326 (APCI or ESI interfaces) that it can use the area calculation standard.

You can take the standard 5 ml vials used in medicine, which have a concentration of 1 g / l.

I diluted by adding 5 ml vial of 15 ml of analytical  water.

I recommend this dilution to add to your sample, 500 micro liters.

The substance is very stable and it provide the correct calculation of the analyte amount.

For quantitative calculations I recommend to use only SI method..

External standard method could fail if more than 30%.

Good luck with that

Mihai Ionică