Every structure in this course has so far been asserted. This chapter is about how anyone knows. Four techniques, used together, determine a molecular structure from a sample: IR finds the functional groups, mass spectrometry gives the molecular weight and formula, and the two NMR experiments map the carbon skeleton. IR is the fastest of the four and usually the first one you look at.
What IR measures
Every covalent bond behaves approximately like a spring connecting two masses, stretching and compressing at a specific natural frequency. Infrared light carries just the right energy to excite those vibrations, and a molecule absorbs IR at the frequencies matching its bonds' natural frequencies.
An IR spectrum plots absorption against frequency, expressed as wavenumber in cm⁻¹ — a direct map of which bonds are present. One caveat worth knowing: a vibration is only IR-active if it changes the molecule's dipole moment, which is why the perfectly symmetric C=C of trans-2-butene shows no absorption at all.
Two variables set the frequency
Following the same physics as a real spring, a stronger bond vibrates at a higher frequency — a stiffer spring oscillates faster — and lighter atoms vibrate at a higher frequency, since there is less mass to move.
That explains the whole layout of the spectrum. C≡C (stronger) absorbs near 2150 cm⁻¹, C=C near 1650, C–C near 1000. And any bond to hydrogen appears at very high wavenumber — 2800–3600 cm⁻¹ — because hydrogen is so light.
The absorptions worth recognizing on sight
| Bond | cm⁻¹ | Appearance |
|---|---|---|
| O–H (alcohol) | 3200–3550 | broad, strong |
| O–H (acid) | 2500–3300 | very broad |
| N–H | 3300–3500 | sharper, weaker |
| C–H (sp³) | 2850–2960 | strong |
| C–H (sp², vinyl/aryl) | 3000–3100 | just above 3000 |
| C≡N | 2250 | sharp, medium |
| C≡C | 2100–2260 | weak |
| C=O | 1650–1750 | strong, sharp |
| C=C | 1620–1680 | medium |
| aromatic C=C | 1450–1600 | several medium |
Two shapes are worth internalizing. An alcohol's O–H is broad because hydrogen bonding creates a spread of slightly different environments; a carboxylic acid's is broader still, sprawling across 2500–3300 and often swallowing the C–H peaks. An N–H, which hydrogen-bonds less, is narrower — and a primary amine shows two N–H peaks where a secondary shows one, which distinguishes them at a glance.
The 3000 cm⁻¹ line is a useful marker: C–H absorptions just below it are sp³, and any appearing just above it indicate sp² C–H, meaning an alkene or an aromatic ring.
The carbonyl region in detail
The C=O stretch is the most useful single signal in IR — strong, sharp, and in a region where almost nothing else absorbs. Its exact position then identifies which kind of carbonyl.
| Carbonyl | cm⁻¹ |
|---|---|
| acid chloride | 1800 |
| anhydride | 1760 and 1820 (two peaks) |
| ester | 1735 |
| aldehyde | 1725 |
| ketone | 1715 |
| carboxylic acid | 1710 |
| amide | 1650–1690 |
That ordering is the reactivity ladder of Module 10, read as a frequency. Stronger resonance donation into the carbonyl weakens the C=O, lowering its frequency — and weaker donation leaves a stronger bond that absorbs higher. The acid chloride at 1800 and the amide at 1660 sit at opposite ends of both scales for the same reason. Conjugation lowers any of these by roughly 30 cm⁻¹, since the extended pi system delocalizes the C=O.
Three isomers to distinguish: acetone, propanal, and allyl alcohol.
Strong peak at 1715, no O–H: a ketone. Acetone.
Strong peak at 1725, plus two weak peaks at 2720 and 2820: an aldehyde — that pair of weak C–H stretches is diagnostic for an aldehyde and appears nowhere else. Propanal.
Broad peak at 3350, plus C=C at 1650 and C–H above 3000, and no carbonyl: allyl alcohol.
Three isomers separated in seconds, without any other technique. This is what IR is for.
The fingerprint region
Below about 1500 cm⁻¹ the spectrum becomes a dense tangle of overlapping single-bond stretches and bending vibrations — the fingerprint region. It is too complex to assign peak by peak, but it is unique to each specific compound, so it is definitive for confirming a match against a reference spectrum.
Treat the two halves differently: above 1500, read functional groups; below 1500, compare against a known. Trying to interpret the fingerprint region from first principles is a waste of effort.
How to read a spectrum in thirty seconds
A reliable order. Is there a strong peak at 1650–1800? If so there is a carbonyl, and its exact position says which kind. Is there a broad peak above 3200? An O–H or N–H, with breadth distinguishing them. Is there anything just above 3000, or peaks in the 1450–1600 cluster? Unsaturation, probably aromatic. Anything sharp near 2250? A nitrile.
Four questions, and you have the functional groups. Then hand the skeleton problem to NMR.
What carries forward
IR is the first step in every structure-determination problem in this chapter, and it pairs naturally with the degrees-of-unsaturation calculation from Module 7 — DoU tells you how many rings and pi bonds to look for, and IR tells you what kinds they are. The next three sections supply the skeleton.