Measuring IR Spectra of Solids

Measuring the IR of a compound is pretty much a standard procedure in an undergraduate organic lab. After all, it's quick, and gives us students a good idea of what functional groups the product actually has; nobody'll let us touch an NMR machine, that's for sure!

Sample preparation for IR spectrum measurement is easy for liquids: just use the stuff neat on a salt plate (well, ok, two salt plates). For solids, however, things are a little bit more complicated.

It's strange, actually. In things like UV-Vis spectroscopy, the Beer-Lambert Law applies. Among other things, this implies that light must pass through the sample. IR spectroscopy, when you get right down to it, isn't all that different from UV-Vis; sure, different motions are excited (the transitions are between vibrational rather than electronic energy levels), but still, light must pass through the sample!

What I find strange is that, alot of the "usual" methods used for preparing solid samples for IR well, don't actually let light pass through the sample very well. But they work!

See, the way my organic lab instructor likes to do IR is to actually dissolve the solid in a solvent and run the solution like a liquid sample. The solvent must be relatively inert, with a simple IR spectrum, like DCM and small alkanes. Since the solution is clear, this makes sense to me. The two problems with this method are: if the solvent dries and the salt plates got stuck*, or, if the sample ended up containing air bubbles. (Of course, don't use too much solvent, either, or the solvent peaks will overwhelm the sample spectrum - usually it's along the lines of a few crystals of solid plus just one drop of solvent.)

However, the "canonical" procedure is to grind the solid up (in a mortar and pestle, say) with some inert mulling agent with a simple, known spectrum (e.g. nujol), and measure the spectrum of the suspension. Presumably, the fact that it's a fine suspension allows the IR light source to interact with the sample while still allowing a (possibly nonlinear) path for the light to go through and reach the detector. (I've never actually done this, by the way.)

I've also heard of the "grind solid sample together with KBr then put it in a pellet press" method. This strikes me as rather odd, as well. Using too much sample will obscure the signal; using not enough sample means that the light will just pass through plain KBr (a substance transparent to IR) and give the spectrum of, say, air. I guess it works much the same way as nujol mulls do: there is "enough" light of each wavelength interacting with and passing through the sample to give results.

What prompted this post was the third way of prepping solid IR samples: thin film on salt plate. (Not in organic lab.) I didn't even realize that "thin film" was the idea until the lab instructor told me. Brief procedure: grind the solid in a solvent that it turns out to be soluble in. Place this suspension/solution on a salt plate as in nujol mulls. Let the solvent evaporate. Sandwich sample with another salt plate, and run the sample dry like that. Sure, now we don't have any solvent interferences, but... Dry sample? With possible airspace and uneven sample thickness (such that light might not pass through)? Really? (But it works! My spectrum was pretty**!)

Postscript: whoever who wrote in the manual to wash my product with water, methanol, and diethyl ether, should be lambasted. The stuff is soluble*** in the latter two solvents! (Admittedly, all solvents were at room temperature, mainly because I need to go up three floors in another building to get ice.) To their credit, CRC handbook's only comment was "slightly soluble in DMSO", but still! I was not pleased when I saw precipitate in the vacuum flask.

*Salt plates are expensive, considering that they are...salt. Meaning NaCl or its relatives (e.g. KBr). The chemistry store at my school deliberately jacked up the price of salt plates during terms when IR-running organic courses are in session, possibly to make sure we students take good care of them. (Or maybe they just want to make a profit.)
**If up-side down from the usual IR spectrum, i.e. peaks point up, not down. Possibly because the machine is an FTIR one?
***Ok, one possibility is that my product is totally different from what it's supposed to be. Considering that the reactants are liquids, and that any potential byproducts (e.g. oligomers) should be less soluble in (various) solvents, however, this is unlikely.