How to interpret your IR spectrum

There seems to be a little confusion about how to interpret your experimental IR spectra and how to use Spartan’s calculated spectrum. I hope the following will help clarify things a little (I recall that I had to interpret about 30-40 IR spectra as a student over a couple of years before this began to feel routine):

1. Experimental spectrum

The signals you see in an IR spectrum are characteristic of your molecule (assuming it is pure). This means that anyone who measures the IR spectrum of this molecule should see the same characteristic signals: same frequencies, same intensities, same signal shapes.

If I were you, I would be strongly tempted to compare my spectrum to an authentic spectrum. The quickest way to do this (but not necessarily the best, see Caveat #2) is to google the IR spectrum. For example, if you believe your sample is ethyl acetate, google ‘ethyl acetate IR spectrum’. You will find a lot of spectra under Images. I would compare my spectrum to one of these and expect pretty close agreement (to within 10 cm-1 or less) everywhere in the spectrum (including fingerprint region).

Caveat #1. The appearance of an IR spectrum (frequencies, intensities, shapes) is influenced by the nature of the sample. Your sample is a ‘neat’ (pure) liquid resting on a so-called ATR accessory. I don’t think Google will turn up many ATR spectra for you, but you shouldn’t have any problem finding ‘neat’ or ‘liquid’ spectra. Look for these before you make your comparison.

Caveat #2. Depending on how you look at it, Google is either brilliant or incredibly stupid. When I google ‘Alan Shusterman’ and click Images, I see some photos of me (brilliant!) and a whole bunch of photos of other people (stupid). Some of them happen to have my name (brilliant!). Others are related to me (or I have published their photo on my web site). There are others that I don’t know at all (stupid).

If you rely on Google to show you an IR spectrum of compound X, go to the original web page and verify that this image is the IR spectrum of compound X.

2. Using Spartan’s calculated IR spectrum

OK, so let’s say you have found a molecule with a boiling point and a published IR spectrum that more or less matches the data you collected in lab. That’s the end of the story, right? What are you supposed to do with the Spartan IR spectrum?

First, only pay attention to Spartan’s calculated spectrum (it may try to show you a gas phase experimental spectrum, but this is not useful – see Caveat #1).

Second, do not expect Spartan’s calculated spectrum to match yours. Discrepancies are partly due to the fact that Spartan is modeling a gas phase molecule, and partly due to the fact that models are only models. Frequencies might be off by as much as 50 cm-1, sometimes more. Intensities rarely agree. (If you want to know why a model and an experiment can look so different from each other, you’ll have to take a higher-level chemistry course that explores the physics and mathematics underlying these models.) So do not look at Spartan’s spectrum as a tool for identifying your unknown. Rely on a published spectrum for identification purposes.

Use Spartan’s calculated spectrum to get ideas about how to assign the signals in your spectrum. When you click on a Spartan peak you will be shown the vibration that corresponds to this peak. Look for stretches of two types: 1) stretches involving a bond between H and another atom, 2) stretches involving double or triple bonds. These vibrations should all appear at higher frequencies than 1500 cm-1, i.e., outside the fingerprint region.

Assignments of lower frequency vibrations – C-C, C-O, and other single bond stretches, bends, wags, etc. – is possible (especially with Spartan’s help), but it takes a lot of experience to make these assignments reliably. You could try and make them – that’s how you get experience! – but don’t take your assignments too seriously. These assignments are certainly not expected for your lab report.

This entry was posted in Labs. Bookmark the permalink.