There is no ONE single strategy for interpreting an NMR spectrum and correlating it to a compound. Spectral interpretation is more like solving a puzzle, and the following approach is merely one method. You may formulate, after many interpretations, another method that you feel more comfortable using.
What does my compound look like?
Draw as accurate a picture of your compound as possible.
What nucleus am I observing?
If you acquire data on more than one nucleus, make sure you know which spectrum belongs to which nucleus.
How many resonances do I expect to see in the absence of any coupling?
Determine which groups are chemically equivalent/inequivalent. Does your compound have a mirror plane that relates nuclei? A mirror plane does not always exist, but for symmetrical molecules (such as a monosubstituted benzene) a mirror plane often is present.
Where do I expect to observe these resonances, i.e. what is their expected chemical shift?
The chemical shift depends on the local environment. Peaks far (9-12 ppm) from TMS are usually unique and identifyable. Peaks close to TMS (0-1 ppm) are usually methyl groups. Typical chemical shifts are given below:
Acids and aldehydes appear between 12-9 ppm. Both aldehydes and acids appear in that region, although for different reasons (right?). To differentiate between acids and aldehydes it is possible to exchange the acid proton with deuterium and make it disappear. Often it is enough to simply add D2O to the sample. The acid signal disappears and a new signal around 5 ppm for DOH appears. Aldehydes do not exchange and so we can tell the difference between the two groups. This technique is very useful for many functional groups where the proton is attached to a heteroatom, especially oxgyen. Alcohols, phenols, and carboxylic acids exchange rapidly and D2O is often sufficient for the exchange process. Amines and thiols need stronger bases, such as NaOD, to force the process to take place.
Arenes and Alkenes occur near 8-5 ppm. Compounds with sp2 C-H fragments fall into this region. Unless the alkene has a strongly electron withdrawing group attached to the double bond, the alkene protons will absorb between 5-6 ppm. The integrated intensity can provide useful hints to discern between alkenes and arenes. Alkenes have at most three protons while arenes can easily have 4-5 protons. To confirm the assignment we need to look carefully at the spin spin pattern.
Aliphatic C-H protons appear between 5-0 ppm. Aliphatic fragments adjacent to heteroatoms or electron withdrawing groups fall into the 2-5 ppm region. Saturated C-H fall into the 2-0 ppm region. If you attach 2 electron withdrawing groups to the carbon, the proton may appear in the 5-6 ppm region where alkenes normally show up. Most organic texts will have tables of common shifts for specific functional groups. Use these tables to make proton assignments. Often, coupling NMR with IR to obtain information about the functional groups that are present is a very good idea.
Possible problems that may appear: (i) some protons can appear over wide ppm ranges, mostly N and O bound protons, making identification difficult; and (ii) the appearance of stray peaks from your solvent (or the solvent used to clean the NMR tube) or an impurity.
What is the integral value or ratio between integral values for each resonance?
Integration values depend on the amount of each kind of nucleus present. Within one given molecule, the ratio between chemically inequivalent nuclei is set, and inspection of the integral values or ratios must follow directly from the molecular formula. Integration can only be compared for resonances arising from the same molecule. For two different molecules, the integral values for similar groups will be quite different if the concentration of the molecules is different!
If you normalize an integral value for a resonance that you can clearly identify, the other integral values can be compared directly. For instance, normalizing a clearly identified CH3 resonance will result in a -CH2- integral value of 2, a C5H5 integral value of 5, etc., for proton fragments on the same molecule only.
If you cannot clearly identify a resonance as belonging to a particular group, a comparison of the ratios of the integral values is necessary. For instance, the ratio of CH3/CH2 integral values should equal 1.5, and the ratio of CH3/C5H5 should equal 0.6, etc., for proton fragments on the same molecule only.
What is the expected multiplicity of each resonance?
Multiplicity can be calculated from M = 2nI + 1 where n = # of equivalent nuclei on adjacent carbons and I = nuclear spin. For the common nuclei that you may encounter, such as 1H, 13C, 19F, 31P, and 195Pt, the nuclear spin is ½ and the above formula collapses to M = n + 1. For a doublet you have 1 adjacent proton, for a triplet you have 2 adjacent protons, and for a quadruplet you have 3 adjacent protons. Some splitting patterns can become very complicated, such as for aromatic protons, and may be difficult to determine. Multiplicity is very important in determining neighboring spin-active nuclei.
Assemble the fragments into a molecule.
Compare your assembled molecule to the molecule you initially drew in Step 1 and rearrange the connectivity as necessary.
Calculate the coupling values, if present.
Coupling is a pairwise relationship. In a simple case of one proton coupling to only one other proton, you MUST observe TWO DOUBLETS that have identical J values. If you observe coupling between a proton and fluorine, JHF will be observed in the 1H NMR (one member of the pair) and JHF will also be observed in the 19F NMR (the other member of the pair). You may observe the pairwise relationship in one NMR spectrum (homonuclear coupling) or between two different spectra (heteronuclear coupling). This the reason for Step 2.
Coupling is a measure of the degree of interaction between two nuclei, and is represented by the letter J (reported in units of Hz). In general, 1J > 2J > 3J, where the superscript refers to the number of bonds between the coupled nuclei. 4J is generally not observed in routine spectra. 2,3J typically ranges from 2-20 Hz.
To calculate J, you will need to know the scan frequency, or SF (PCNMR refers to this as "synthesizer frequency" in the parameter list).
Different nuclei will have different values of SF, and you must check the parameter list to obtain the exact value of SF for your particular experiment before you attempt to calculate J.
Trace through your fragment assembly, comparing it to your the molecule you initially drew.
If you assembled your fragments correctly, the multiplicity, coupling, and chemical shifts should match and in all likelihood you have correctly interpreted your NMR spectrum. If there are inconsistencies, you must (obviously) try again with another assembly of fragments.