In this example we'll try to deduce the structure of a compound with formula C7H14O6 from its 1H-NMR spectrum and its 2D H/H COSY.
Let's see how far we can go...
The 1H-NMR spectrum was recorded on a 9.5 Tesla instrument where 1H resonance happens at 400 MHz. Solvent is D2O.
Make a list of once-over observations and preliminary conclusions:
1) The chemical shifts of all 1H's are within 3.2-4.4 ppm suggesting that all 1H's are under the influence of the electronegativity of the O atoms.
2) The sum of all integrals amounts 10 1H's. The 4 remaining from the molecular formula could be OH's that end up giving a common signal together with the residual H2O of the solvent.
3) The individual integrals indicate that almost every signal belong to single 1H's, exception made of those at 3.6 and 3.53 ppm.
4) The relatively high chemical shif of the signal at 3.53 ppm, its singlet shape and its integral of 3, suggest that we are dealing with a OCH3 group.
6) The shapes of all signals (except that at 3.53 ppm) is quite complex suggesting that the 1H's form a complex network of couplings that cannot be figured out from the 1H-NMR spectrum.
Let's see how far the 2D H/H COSY lets us go:
1) As usual, start by tracing a diagonal (blue line).
Draw the different found correlations by means of the corresponding paralelograms and take note of the related chemical shifts:
Summing up, we have obtained the following fragments of coupled 1H's:
Let's do the exercise of measuring coupling constants for the first time. We'll do it from the 1H-NMR spectrum in which the multiplets have been expanded and their peaks pick-peaked in Hz.
Coupling constants deduced from peak separations are indicated in green numbers. Try to measure them by yourself...
We may now complete our correlation map with the measured coupling constants:
We've got the following:
1) A OCH3 group (singlet at 3.3 ppm) isolated from the rest of 1H's.
2) One group of 7 coupled 1H's in the way the 2D H/H COSY allowed us to deduce.
Is this enough info to figure out the structure?
Unfortunately it isn't!!! But we're close!!!
We do need additional data, namely 13C-NMR and 2D H/C correlated spectra.
I just beg you a bit of patience...
