Sciencemadness Discussion Board

TLC to identify unknowns

cnidocyte - 11-9-2010 at 08:59

I know the theory behind TLC but have very little experience with it. Is there some kinda database of TLC results that were obtained using a specific type of TLC plate and solvent which can be used to identify unknown compounds? Lets say I do TLC on what I know to be pure NaCl and it travels 5mm up the TLC plate in 10 minutes. If I then wanted to find out what an unknown mixture contained and I replicated the conditions of my first TLC experiment, would all NaCl in the unknown mixture travel 5mm up the plate in 10 minutes? In other words could I identify NaCl in the unknown by comparing the TLC plate to that of the previous experiment?

crazyboy - 11-9-2010 at 09:35

TLC is commonly used to determine if a reaction has progressed. A TLC plate is often spotted with a mixture of the reaction mixture at T=0 then compared with spots of the reaction mixture at T=X. TLC is useless when trying to determine the identity of an unknown compound with no basis for comparison.

Paddywhacker - 11-9-2010 at 15:48

Quote: Originally posted by cnidocyte  
... In other words could I identify NaCl in the unknown by comparing the TLC plate to that of the previous experiment?


Not with any great certainty.

A better scheme is to add some third compound to both the standard sample and the unknown sample. The positions of the observed spots would be measured relative to the position of the spot from the added compound. This way you would compensate for variations of conditions between separate runs.

DDTea - 11-9-2010 at 16:16

In any analysis you do, you can have two of the following, but not all three: it can be fast, cheap, or accurate :P. TLC is fast and cheap.

Truthfully, I have not looked into whether or not there are standard tables collected of retention times/factors in TLC. This is more common with HPLC and GC. With TLC, as crazyboy mentioned, you're usually just looking for qualitative data: "Has the reaction occurred?" or "How many compounds are in this mixture?" or "What is the nature of this compound (polar/non-polar)?" When an unknown sample is run next to a standard, it can help in identifying the unknown compound (if you already have a hunch about what it is); otherwise it simply says that the unknown and the standard are of similar size and solubility in the mobile phase.

But as with anything, it's going to be limited by your creativity and determination. Could you find a way to standardize TLC data? Sure, but I'd say it's barely worth the effort even compared to simple column chromatography. Definitely consider 2D TLC if you want to do any rigorous experimentation.
Quote:
A better scheme is to add some third compound to both the standard sample and the unknown sample. The positions of the observed spots would be measured relative to the position of the spot from the added compound. This way you would compensate for variations of conditions between separate runs.


Actually, this is a good idea. Maybe then you could do 3D or 4D TLC and come up with some multivariable regression model to aid in identification of compounds via TLC.

In any case, between two different runs, there's going to be a lot of variability: temperature, the amount of material you spotted... You'll definitely have to find a way to control those factors before you can even start taking standard data.

[Edited on 9-12-10 by DDTea]

aonomus - 11-9-2010 at 20:51

Not to nitpick, but if you were to run TLC, chances are you are looking for organics, not inorganics. You have to visualize it somehow, either by UV (a UV fluorescent compound is mixed in with the silica and binder, any compounds will mask the fluorescent compound casting a shadow), or by a stain (permanganate oxidizes, ninhydrin binds to amines, bromocresol green for pKa <5.0, etc). NaCl as far as I know is a non-UV active compound, and probably doesn't react with many stains (if any?), its not common to look for inorganic salts on a TLC, typically you want to see the contents of a reaction mixture, to see if your starting material has reacted or not.


Anyway, in terms of comparisons to a 'library', its not terribly common to look for a magic book that contains retention factors in TLC plates. You have a single solvent system, and you can observe the retention factors between experiments and conditions, however its not terribly quantitative since you can overload your plate causing streaking, etc.

Even comparing between the same TLC runs is tenuous at best. Comparing an experiment where you use the exact same plate, the exact same solvent system, and the same method for spotting (loading, etc), you may *still* observe fluctuating retention times. Depending on the adsorbed water on the stationary phase, you could get fluctuations (I have in the past, so now I bake all of mine just prior to use, but I don't use TLC much anymore).

DDTea is right, you would run your reaction mixture against another spot of each starting material and/or reagent (that you know will show up on a TLC plate with your visualization method of choice), you aren't trying to run a random sample (say a solvent extracted sample of dirt) and expect a database to show hits for pesticides, etc. You have a general idea of what you know is there, so you can look for anything that isn't.

In HPLC, you can't always compare one chromatogram to another either, especially because multiple solvents can be used in gradients to selectively resolve and elute compounds off as efficiently as possible, not everyone will use the same method. What HPLC does give you is repeatability (accurately repeatable retention times given a properly designed method), quantification (the area under the peaks on the chromatogram can be integrated and used to give compositional information), and high resolution power (theoretical plates ++++ compared to TLC).

tetrahedron - 24-10-2012 at 05:45

it's conceivable to separate aqueous cations by chromatography and visualize the spots too, e.g. taking advantage of the complexes they form (especially valid for transition metals) and their other characteristic reactions.

an example with Fe3+, Ni2+, Cu2+: http://infohost.nmt.edu/~jaltig/ModelChromReport.pdf