by
freemab » Fri May 01, 2009 2:48 am
Such simple questions, and so difficult to answer.
There is no absolute answer. To prove a compound is pure, you need to demonstrate that all reasonable analyses are consistent with the assigned purity. This is usually easier with very pure compounds than with less pure ones, so we strive for high purity even when a lower purity would do as a standard.
There are certain obvious assays to perform: GC for residual solvents. KF for water. Loss on drying. Sulfated ash for metals. The use of these sorts of assays provide assurance that you don't have excessive contamination that might interfere with other testing.
Of course, the basic test for the organic compound itself is an elemental analysis, which should be done. But typically these are not sufficiently precise as we get to higher purities.
Once we are reasonably sure of high purity, a more exact purity can often be determined with spectroscopic and chromatographic methods. A pure compound run on an appropriate chromatographic method should give one peak. (There are problems and conditions that can give rise to two peaks from one compound, which is why the chromatography must be "appropriate".) Better still, use a hyphenated method such as LC-PDA (= LC-DAD) or LC-MS to more precisely demonstrate peak purity or even peak identity. Quantitative NMR also can be used, but my understanding is that it must be used by a competent practitioner because it can give false results if incorrectly applied.
I suspect, however, that you are asking about standards that are NOT so pure. These can be more of a challenge, as I implied. The most direct way of dealing with these is to treat them as secondary standards and to assay them against a highly pure primary standard. This approach allows you to buy or prepare a fairly small amount of the highly pure material and to use it only to quantify the secondary standard - which you then use in your assays.
In this regard, it is well to note that chemical standards are often available from sources like USP for given assays. For the stated assay, they can be trusted. It does not follow that these are universal standards, however. For example, if a USP antimicrobial standard is used in a microbio assay, you can expect its activity to be correct. It does not follow that if you use it as an HPLC standard that it will be correct. This is because a microbio assay determines activity, whereas HPLC determines compounds based upon their separation on a column - which may or may not have anything to do with activity.
Should obtaining a highly pure primary standard not be an option (e.g., it is impossible or excessively expensive to prepare a highly pure primary standard), then each case must be viewed separately and intelligence and creativity must be applied to determine the purity accurately.
For example, suppose your chromatogram shows your main component and a few smaller peaks. The simple-minded approach is to do an area-% calculation to determine purity. In fact, this works well when the peaks have the same chromophore, or at least similar chromophores. Hence, if you can identify the peaks and the chromophores, this approach can be trusted.
If you cannot actually identify the peaks, the spectra will give evidence of the chromophores - radically different spectra cannot be due to the same chromophores. If the chromophores are different, then there is no assurance that the extinction coefficients at any given wavelength will be the same - or even very close at all.
In extremis, it may be worthwhile to do a little preparative chromatography - especially if the mobile phase is completely volatile. Collect the peaks (multiple times perhaps), evaporate off the solvents, and weigh them. (Easier said than done.)
In this regard, it may be worth mentioning detectors like the ELSD. Such detectors are claimed to measure mass, but that's really not true. What they will do for you is to reveal contaminants completely lacking chromophores, and to give you a perspective on your chromatogram that is completely different from what you get from UV. With crude samples, you often see in ELSD peaks that simply don't exist in the UV trace. However, bear in mind that ELSD is highly nonlinear: if peak A is 3 or 4 times the size of peak B, it is probably really only twice the mass. Furthermore, melting point, volatility, and other characteristics profoundly affect the response of compounds.
Similarly, MS is a very powerful tool, but if you cannot ionize a compound, you cannot detect it. Quantification should be done carefully.
As for getting a chemical compound certified - there are probably laboratories that offer such services, but use them with caution. In fact, you might be best to do certifications yourself. Write SOPs describing your procedures, follow your methods, and write your own certificates of analysis. Keep in mind that it's your name to be signed to the bottom, so be sure you know what you're dealing with.
I hope this helps in some small way.
