Sciencemadness Discussion Board

Calculating purity

Rich_Insane - 4-2-2014 at 13:53

Hey everyone,

So I'm a little confused as to how purity is calculated for novel compounds. Let's say that I've synthesized a novel compound and I already know that it's the compound I want. I rotovap off all the ether to leave behind a residue that I then dissolve in acetonitrile/water with a little formic acid.

I've already done all of this and I have actually run the sample through the LC-MS. I was able to detect the desired mass by finding it using the EIC tool on the LC-MS software. I noted the retention times for the peaks at which the product was found, then I switched to the DAD tool. What I did was I correlated each peak on the DAD with the peaks on the EIC. Any DAD peak that did not contain a large amount of the desired (expected) mass was considered an impurity. So then I integrated all peaks that exceeded the magnitude of the noise on the DAD. The software I used integrates each peak individually, so I found the peaks that corresponded to my product (in my case, it was mostly two peaks bunched very close together). I took the area under those peaks out of the total area of the entire DAD spectra and then got a number out of it.

From there, I calculated what I think to be the purity (~95%). Is my technique correct? Is there any step that I am missing?

One problem with the compound in question is that it has been synthesized on the scale of 0.1 mmol (not 0.1 mol), so I cannot weigh it. However, I am using a very sensitive time-of-flight mass spectrometer connected to an LC with a C18 column.

forgottenpassword - 4-2-2014 at 14:08

Are you using atlas?

Rich_Insane - 4-2-2014 at 22:08

I am not sure what software I am using. It came with our TOF-LCMS system (Agilent). I use the "Qualitative Analysis" program to do these calculations.

DutchChemistryBox - 4-2-2014 at 23:26

If you can't weigh the sample accurately you should weigh more of it and than dillute it.


As I understand you're using the '100% method'. You better use the extern standard method by making a calibration curve.

DJF90 - 5-2-2014 at 10:59

Quote: Originally posted by DutchChemistryBox  
If you can't weigh the sample accurately you should weigh more of it and than dillute it.


As I understand you're using the '100% method'. You better use the extern standard method by making a calibration curve.


He can't weigh it more accurately because he doesn't have enough of it to do what you suggest.

The second suggestion is much better. You should make up some volumetric solutions of a standard, construct a calibration curve (something like response /mAu against standard /wt%) Then you can prepare your sample (known mass in a known volume) and spike the sample with the standard. Comparing the responses /mAu and using the calibration curve, you can work out the weight of standard present and therefore the weight of sample present, and thus its purity.

The problem with this is your sample may have a different response factor compared to the standard, giving an incorrect assay. You'd have to run a UV-Vis spec of sample and standard, and choose a wavelength where both compounds show a point of inflection. This will minimise error due to response factor.

Personally I prefer assay by 1H-NMR against an internal standard, but it doesn't sound as if you have enough material for that.

Rich_Insane - 6-2-2014 at 11:44

Yea, I'd say comparing to the standard is the best idea. I could use DAD area @ 254 nm or use our UV-VIS at that wavelength. Unfortunately, I cannot obtain a standard because the compound I've synthesized is novel (it's a peptide-nucleic acid chimera).

Typically, we'd be lucky to obtain a few hundred micromoles of product, and I can't scale up as the monomer precursor alone costs over a hundred dollars per gram.

I really wish I worked with the traditional small molecule synthesis stuff -_-

DutchChemistryBox - 6-2-2014 at 13:48

Unfortunately you don't have much choices in such case.

Maybe it is smart to show us your chromatograms. Which mobile fase do you use? It is important to make sure that you have sepperated everything in your sample.