How to Actually Use an IR Spectrum Correlation Table Without Losing Your Mind
The IR spectrum correlation table is one of those reference tools that everyone tells you to memorize early on, but nobody explains why the numbers don't always match what you're looking at. It lists functional group absorptions and their typical wavenumber ranges so you can take an unknown spectrum and start making educated guesses about what structural pieces are present. That is the idea, at least. The reality is messier. I spent years working in an organic chemistry lab where we ran hundreds of IR scans a week on crude reaction products. You learn quickly that the textbook ranges are guidelines, not laws. I still keep a printed correlation table at my bench, but I use it more as a starting point than a definitive answer. Let me walk through how it actually works in practice.
Reading the Ir Spectrum Correlation Table
The table is organized by functional group, with each entry giving you a range in wavenumbers and a description of the expected peak shape. Broad peaks around 3200 to 3600 cm^-1 usually mean O-H stretching. Sharp peaks in the 1650 to 1750 cm^-1 region typically point to C=O stretching. N-H stretches show up around 3300 to 3500 cm^-1 and are usually thinner than O-H peaks. C-H stretches for sp3 carbons sit near 2850 to 3000 cm^-1, while sp2 C-H appears just above 3000 cm^-1. Here is the part that catches people off guard: the same functional group can appear at very different wavenumbers depending on its environment. A carbonyl in a ketone will absorb differently than one in an amide, an ester, or a carboxylic acid. The table gives you ranges for each, but you need to pay attention to the exact compound class. A simple ketone C=O shows up near 1715 cm^-1. An amide C=O drops to around 1650 cm^-1 because of resonance. That shift of 65 wavenumbers matters when you are trying to tell two similar products apart. I ran into this exact problem last year with a Friedel-Crafts acylation that was giving me a messy product mixture. The IR showed a carbonyl peak at 1688 cm^-1, which sat right between where I expected a regular aryl ketone and where an amide would appear. I initially thought the reaction had somehow produced an amide impurity. I had to run NMR to confirm that the lower wavenumber was actually due to conjugation with the aromatic ring, not a completely different functional group. Conjugation can shift a carbonyl stretch down by 20 to 30 cm^-1. The table tells you that, but it does not prepare you for the moment when you see it in your own data and second-guess everything.
Below is the core correlation data I use most often, organized by the region of the spectrum rather than alphabetically, because that is how you actually scan a spectrum when you are trying to work fast. 4000 to 2500 cm^-1 Region O-H stretch: broad, 3200 to 3600 cm^-1 for alcohols. Broader and shifted lower, around 2500 to 3300 cm^-1, for carboxylic acids. N-H stretch: 3300 to 3500 cm^-1, usually sharper than O-H. C-H stretch: sp3 around 2850 to 3000 cm^-1, sp2 above 3000 cm^-1, sp around 3300 cm^-1.
Get the Full Details
2500 to 2000 cm^-1 Region C triple bond stretches: nitriles around 2250 cm^-1, alkynes around 2100 to 2260 cm^-1. These are usually sharp and medium intensity, which makes them relatively easy to spot even in noisy spectra. 1850 to 1500 cm^-1 Region
C=O stretches: acid chlorides near 1800 cm^-1, anhydrides with two peaks around 1820 and 1760 cm^-1, esters near 1735 cm^-1, ketones near 1715 cm^-1, aldehydes near 1725 cm^-1, carboxylic acids near 1710 cm^-1, amides near 1650 to 1690 cm^-1. C=C stretches: alkenes around 1600 to 1680 cm^-1, aromatic C=C in pairs near 1450 to 1600 cm^-1. C=N stretches for imines around 1640 to 1690 cm^-1. 1500 to 400 cm^-1 Region This is the fingerprint region, which means it is unique to each molecule but also much harder to interpret without a reference. C-O stretches for alcohols, ethers, and esters appear between 1000 and 1300 cm^-1. C-N stretches sit around 1000 to 1350 cm^-1. Substitution patterns on aromatic rings give characteristic out-of-plane C-H bends below 900 cm^-1, which can tell you whether a ring is mono-substituted, di-substituted, or tri-substituted.
The fingerprint region is useful when you already know what compound you are dealing with and want to confirm identity by comparing against a known spectrum. It is not helpful when you are trying to figure out what the compound is from scratch. Do not waste time trying to assign every peak in that region on an unknown. I have seen people spend twenty minutes agonizing over the fingerprint peaks of a spectrum when a quick look at the functional group region would have pointed them in the right direction in under two minutes. The correlation table is designed for the functional group region first. That is where the actionable information lives. One thing the correlation table does not warn you about enough is peak overlap. In a real sample, multiple functional groups can absorb in the same narrow range. A C=O stretch and a C=C stretch might both show up near 1650 cm^-1. An overtone of a C=O can appear around 3400 cm^-1 and be mistaken for an O-H or N-H stretch. I once misidentified a weak shoulder near 3400 cm^-1 as residual alcohol on a purified product. It was actually a first overtone of the carbonyl at 1720 cm^-1. The overtone appeared at roughly double the frequency, which is the expected pattern, but I did not catch that until I went back and checked the math.
Another common pitfall is hydrogen bonding. It shifts O-H and N-H stretches to lower wavenumbers and broadens them significantly. A concentrated alcohol sample will show a very broad O-H peak centered around 3300 cm^-1. Dilute it in a non-polar solvent like CCl4 and that peak narrows and shifts up toward 3600 cm^-1. If you are comparing your spectrum to a reference taken under different conditions, the positions will not match exactly. This is why solvent choice matters more than most beginners realize when using an IR spectrum correlation table for identification work. Sample preparation is another area where things go wrong quietly. KBr pellet method is standard for solid samples, but if the KBr absorbs moisture, you will see a broad water O-H peak around 3400 cm^-1 that is not from your sample. I have ruined whole batches of spectra by using improperly dried KBr. The fix is simple: dry the KBr in an oven at 110 degrees Celsius for a few hours before use and store it in a desiccator. Thin film method for liquids has its own issues. If the film is too thick, you get saturation where peaks flatten out completely and lose their shape. You should adjust the path length so that the strongest absorption is no more than 80 percent transmittance. Anything darker and the peak shape is distorted beyond useful interpretation. Let me address something the correlation table alone cannot solve: distinguishing between very similar functional groups. For example, an aldehyde and a ketone both show a strong C=O stretch near 1700 cm^-1. The difference is the aldehyde C-H stretch, which appears as two weak peaks around 2720 and 2820 cm^-1. Those peaks are easy to miss if you are not looking for them, and they are easy to dismiss as noise if you do not know what they should look like. The correlation table lists them, but they require a decent quality spectrum to see clearly. A noisy instrument or a poorly prepared sample will obscure them entirely.
Quantitative analysis is another place where the correlation table falls short. The position of a peak tells you what functional group is present. The intensity of a peak can, in principle, tell you how much is present, but only if you control for path length, concentration, and instrumental variables. For rough screening purposes, IR intensity comparisons work well enough. For actual quantification, you should use a calibrated method like HPLC or NMR instead. IR is fast and cheap for what it does, but it is not a precision tool. If you want a downloadable version of the correlation table for quick reference, most university chemistry department websites host PDF versions that you can pull from their instrumental analysis pages. Agilent and Thermo Fisher also publish updated versions online that include newer functional group data and instrument-specific notes. I keep a copy bookmarked in my browser rather than printing it, since the online versions get updated more frequently than printed handouts. The main limitation you need to accept is that IR spectroscopy simply cannot identify every compound on its own. Isomers with the same functional groups but different arrangements can produce nearly identical IR spectra. Stereoisomers are basically indistinguishable by IR. Compounds with very few polar bonds produce weak spectra with few diagnostic peaks, making correlation table matching unreliable. In those cases, you need NMR or mass spectrometry to fill in the gaps. IR is a screening tool and a structural confirmation tool, not a standalone identification method for complex unknowns.
The correlation table is worth learning because it gives you a framework for reading spectra faster than any other single technique. But it is not a substitute for understanding what is actually happening in the molecule. The wavenumbers move. Peaks overlap. Hydrogen bonding changes everything. Keep the table handy, but trust your eyes on the spectrum more than the textbook ranges when they disagree. That is the practical approach, and it is the one that saves time when you are working against a deadline.
