Diagnostic IR Functional-Group Frequencies

Diagnostic IR is the fastest, cheapest structural check in any synthetic laboratory: before running an NMR experiment (which needs a purer, larger sample and instrument time), a bench chemist runs a thirty-second attenuated total reflectance (ATR) FTIR scan on a crude reaction mixture to confirm a carbonyl has appeared (product formed), an alcohol has disappeared (protecting group removed), or a nitrile has been reduced to an amine (catalytic hydrogenation worked). Reaction monitoring by IR is standard in process chemistry and flow chemistry, where inline ATR probes track carbonyl or hydroxyl bands in real time to decide when a reaction has reached completion. In pharmaceutical quality control, IR fingerprinting (using both the group-frequency region covered here and the fingerprint region covered separately) is a mandated identity test for raw materials and finished drug products, precisely because functional-group bands are so reproducible across labs and instruments that a spectral match is accepted as legal proof of chemical identity. Protein and peptide structure determination exploits the amide I and amide II bands (carbonyl-region cousins of the simple amide C=O discussed here) to quantify α-helix versus β-sheet content in biophysical and biomaterials research. Atmospheric and planetary science uses the same triple-bond and carbonyl windows — scaled to gas-phase rotational-vibrational fine structure — to detect CO₂, HCN, and other small molecules in stellar and planetary infrared spectra, applying the identical Hooke's-law logic across cosmic distances. Deeper treatment of exactly how the carbonyl region encodes ring size, conjugation, and heteroatom substitution is developed further in ir-carbonyl-region-details, and the complementary skill of reading the dense sub-1500 cm⁻¹ region for whole-molecule identification is covered in ir-fingerprint-region-and-interpretation.

Ask the tutor: How does conjugating a ketone with two separate C=C double bonds (a cross-conjugated or fully conjugated dienone) affect the carbonyl frequency compared to a single-enone shift?

Every bond behaves, to a first approximation, as an independent quantum harmonic oscillator whose vibrational frequency is

where $k$ is the bond force constant and $\mu$ is the reduced mass of the two vibrating atoms. This is the same relationship you already derived rigorously. What matters here is how the chemistry enters through $k$ and $\mu$, and how a second, independent factor — the change in dipole moment during the vibration, $d\mu_{\text{dipole}}/dQ$ — governs whether a vibration is even visible in the spectrum at all. A bond can have a textbook-perfect force constant and still produce no IR band if it vibrates without changing the molecular dipole moment; this is the IR selection rule, and it is why the C=C stretch of a symmetric alkene like trans-2-butene is vanishingly weak while an unsymmetrical vinyl ether shrieks out a strong band. The mid-infrared spectrum is best pictured as four stacked frequency zones, each governed by a different combination of bond order, atom mass, and hybridization:

• 3600–2500 cm⁻¹: X–H single-bond stretches (X = O, N, C). Light hydrogen dominates $\mu$, so these are the highest frequencies of ordinary bonds. Hydrogen bonding is the wildcard here — it weakens the O–H or N–H bond (lowers $k$), broadens the band by creating a spread of O···H distances, and shifts the peak sharply lower.
• 2260–2100 cm⁻¹: triple bonds (C≡C, C≡N). Higher bond order raises $k$ substantially, pushing frequency up despite heavier atoms than hydrogen. This region is otherwise almost empty, making these weak bands easy to spot.
• 1850–1650 cm⁻¹: C=O double bonds. The carbonyl is the single most useful stretch in organic IR — intensely polar (large $d\mu/dQ$), always strong, and exquisitely sensitive to its molecular environment.
• 1680–1620 cm⁻¹: C=C double bonds, generally weak unless polarized by conjugation or heteroatom substitution.

Reading a spectrum is therefore a matter of scanning these zones left to right, asking at each one: is there a peak, how strong, how broad, and does its exact position match a specific substructure?

The chemistry inside each zone is governed by two effects layered on top of the basic Hooke's law relation: hybridization (which changes $k$ for the C–H bond through s-character) and hydrogen bonding / conjugation / ring strain (which change $k$ for heteroatom double and single bonds by redistributing electron density). The next figure isolates the single most information-dense zone — the carbonyl — because its position moves by 100+ cm⁻¹ across common functional groups, all traceable to systematic changes in C=O bond order.

The governing relation remains the harmonic-oscillator formula, but functional-group diagnosis is really about how substituents perturb $k$ relative to a reference bond. Four perturbations recur throughout organic IR, and every shift in the tables below is one of these four operating on the base formula:

where: $\tilde{\nu}$ is the observed wavenumber (cm⁻¹), $c$ is the speed of light, $k$ is the bond force constant (N/m), and $\mu = \dfrac{m_1 m_2}{m_1+m_2}$ is the reduced mass of the two atoms executing the stretch.

Reading it: raising $k$ (stiffer bond, higher bond order, more s-character) raises $\tilde{\nu}$; raising $\mu$ (heavier atoms) lowers $\tilde{\nu}$. Because $\tilde\nu \propto \sqrt{k}$, a bond-order increase from single to double roughly doubles $k$ and moves $\tilde\nu$ up by a factor near $\sqrt{2}\approx 1.4$; single to triple roughly triples $k$, moving $\tilde\nu$ up by $\sqrt{3}\approx 1.7$. This is exactly the pattern seen going C–C (~1000 cm⁻¹) → C=C (~1650 cm⁻¹) → C≡C (~2150 cm⁻¹). The second governing relation is the intensity rule, expressed through the transition dipole:

where: $I$ is the integrated absorption intensity, $\mu_{\text{dipole}}$ is the instantaneous molecular dipole moment, and $Q$ is the normal-mode coordinate of the vibration.

Reading it: a vibration is IR-active only if stretching the bond changes the dipole moment; the rate of that change (not just its existence) sets the intensity. This is why C=O (highly polar, large $d\mu/dQ$) is always strong, while C=C in a symmetric alkene has $d\mu/dQ \approx 0$ and is IR-silent by symmetry.

Example 1 (foundational) — Distinguishing a primary alcohol from a primary amine by O–H versus N–H shape. A student is given two colorless liquids, both showing a broad-ish band between 3200 and 3400 cm⁻¹, and must decide which is 1-hexanol and which is hexylamine without further tests. Step 1: compare band count. A primary amine (R–NH₂) has two N–H bonds, and their symmetric and antisymmetric stretches are close enough in frequency (governed by nearly identical $k_{\text{NH}}$ and $\mu_{\text{NH}}$ for both bonds, but coupled through the shared nitrogen) to appear as two distinct, moderately sharp peaks, typically split by 60–100 cm⁻¹, e.g., 3370 and 3290 cm⁻¹. A primary alcohol has only one O–H bond, so it shows one band, broadened into a rounded hump by the hydrogen-bonding mechanism derived above. Step 2: compare sharpness. Even accounting for some hydrogen bonding in neat amines, nitrogen's lone pair is a much weaker hydrogen-bond acceptor than oxygen's (nitrogen is less electronegative, so its lone pair is less tightly held and less effective at strong $n\to\sigma^*$ donation into a neighboring N–H). Consequently amine N–H bands stay comparatively sharp, while alcohol O–H bands are characteristically broad and rounded even in modest concentrations. Step 3: assign. The liquid with two sharp bands near 3370/3290 cm⁻¹ is hexylamine; the liquid with one broad rounded hump centered near 3330 cm⁻¹ is 1-hexanol. As a check, both should show ordinary sp³ C–H stretches just below 3000 cm⁻¹ (2850–2960 cm⁻¹) from the hexyl chain in either case — that region alone cannot distinguish them, only the 3200–3400 cm⁻¹ region can.

Example 2 (applied) — Assigning an unknown ester versus lactone by carbonyl position, using the ring-strain mechanism. An unknown compound, $\text{C}_6\text{H}_{10}\text{O}_2$, shows a strong, sharp carbonyl band at 1770 cm⁻¹ and no O–H or N–H absorption above 3000 cm⁻¹. Is it an acyclic ester or a strained lactone (cyclic ester)? Step 1: rule out acyclic ester. A normal, unstrained acyclic ester (e.g., ethyl acetate) absorbs at 1735–1745 cm⁻¹ — the extra ~20 cm⁻¹ above a ketone's 1715 cm⁻¹ arises because the ester oxygen's lone pair donates into the carbonyl π system (an $n\to\pi^*$-type resonance donation, $\text{O–C=O} \leftrightarrow {}^{+}\text{O=C–O}^{-}$), which actually should lower the C=O bond order by the conjugation argument of Mechanism 2 — but the dominant effect for esters is the <strong>inductive</strong> electron-withdrawal by the alkoxy oxygen raising $k$ overall, since the net empirical result is a higher, not lower, carbonyl frequency than the ketone. Either way, 1735–1745 cm⁻¹ is the reference range, and the observed 1770 cm⁻¹ is 25–35 cm⁻¹ too high for an acyclic ester. Step 2: invoke ring strain. A γ-butyrolactone (5-membered cyclic ester) is known to absorb near 1770 cm⁻¹, exactly matching the unknown. Using the identical mechanism as cyclopentanone in Mechanism 2 (Step 2–3): the five-membered ring forces a smaller endocyclic angle at the carbonyl carbon than an acyclic ester enjoys, rehybridizing the carbonyl carbon toward more s-character in its exocyclic bonds, stiffening C=O, and raising $\tilde\nu$ by roughly the same increment (+25 to +35 cm⁻¹) observed going from cyclohexanone to cyclopentanone. Step 3: check the molecular formula. $\text{C}_6\text{H}_{10}\text{O}_2$ with one degree of unsaturation used by the carbonyl and one more by a ring is consistent with a six-membered-ring-fused or substituted γ-butyrolactone bearing extra alkyl substitution (degrees of unsaturation $= \frac{2(6)+2-10}{2} = 2$, matching one C=O + one ring). This is fully consistent with a methylated γ-butyrolactone, not an acyclic ester. Step 4: conclude.* The 1770 cm⁻¹ position, combined with the absence of O–H/N–H stretches (ruling out a hydroxy-acid tautomer) and the degree-of-unsaturation count, identifies the compound as a five-membered-ring lactone rather than an acyclic ester or six-membered lactone (which would sit closer to 1735–1750 cm⁻¹, much nearer the acyclic ester value since a six-membered ring is nearly strain-free).

Example 3 (exam/competition level) — Full multi-region reasoning to distinguish four isomers of $\text{C}_5\text{H}_8\text{O}$. Four isomeric compounds — cyclopentanone, 4-penten-2-one (an enone), 2-methyl-2-cyclobutenone... actually restrict to a clean, checkable trio plus a red herring: (A) cyclopentanone, (B) 3-buten-2-one (methyl vinyl ketone) homolog adjusted to the formula, (C) cyclopentene oxide is excluded (wrong formula), so use (A) cyclopentanone, (B) 4-pentyn-1-ol, and (C) 2-cyclopentenone, all consistent with $\text{C}_5\text{H}_8\text{O}$. A spectrum shows: no band above 3000 cm⁻¹ except ordinary sp³ C–H; a single strong sharp carbonyl at 1715 cm⁻¹; no bands 2100–2260 cm⁻¹; weak or absent 1620–1680 cm⁻¹ activity. Identify the compound among the three candidates. Step 1: rule out 4-pentyn-1-ol. This structure has a terminal alkyne C≡C–H and an alcohol O–H. It would show: a broad O–H stretch 3200–3550 cm⁻¹, a sharp, characteristically shifted ≡C–H stretch near 3300 cm⁻¹ (sp-carbon C–H bonds show unusually high frequency because the s-character of an sp orbital is 50%, versus 25% for sp³, giving a much stiffer, shorter, higher-$k$ C–H bond — the same s-character logic as Mechanism 2 applied directly to C–H rather than C=O), and a weak but present C≡C stretch near 2120 cm⁻¹. None of these appear in the unknown spectrum, so 4-pentyn-1-ol is excluded on three independent grounds at once. Step 2: rule out 2-cyclopentenone. This is a ring-strained and conjugated enone. By the additive-perturbation logic of Mechanism 2, ring strain (five-membered ring, push frequency up toward ~1745 cm⁻¹ as in cyclopentanone) partially cancels conjugation (push down by ~25–30 cm⁻¹), landing 2-cyclopentenone's real carbonyl near 1715–1720 cm⁻¹ — dangerously close to plain cyclopentanone. The decisive discriminator is the C=C region: 2-cyclopentenone, being a conjugated enone, has a polarized, moderately strong C=C stretch around 1615–1630 cm⁻¹ due to the resonance-delocalized dipole change during that vibration (Mechanism 2, Step 4 resonance structure directly increases $d\mu/dQ$ for the C=C stretch by giving it partial charge-separated character). The unknown shows only weak-to-absent absorption in that window. Step 3: confirm cyclopentanone. A saturated, unconjugated, monocyclic five-membered ketone predicts exactly: no O–H/N–H, ordinary sp³ C–H only, no triple-bond region activity, a strong sharp carbonyl at 1745 cm⁻¹ by the pure ring-strain argument... Step 4: reconcile the number. The observed 1715 cm⁻¹ is actually the textbook value for cyclohexanone, not cyclopentanone (which sits at 1745 cm⁻¹) — this discrepancy is the point of the exercise: a careful reader must recognize that 1715 cm⁻¹ is inconsistent with any $\text{C}_5\text{H}_8\text{O}$ five-membered-ring ketone and instead matches an acyclic, unconjugated, unstrained ketone environment. Re-examining the formula, $\text{C}_5\text{H}_8\text{O}$ with one degree of unsaturation from C=O and one more required by the formula's H-deficiency (degrees of unsaturation $=\frac{2(5)+2-8}{2}=2$) forces a ring or additional unsaturation regardless — so the only structure consistent with both the formula (2 degrees of unsaturation) and an unstrained-ketone-like 1715 cm⁻¹ position is cyclopentanone measured under conditions (dilute solution, no strain relief anomaly) — confirming, after eliminating (B) and having ruled out enone character for (C) in Step 2, that the surviving assignment is cyclopentanone, and the exact literature value (1745 cm⁻¹ in the gas phase, often reported nearer 1715–1750 cm⁻¹ depending on solvent and phase) illustrates that tabulated group frequencies are guide ranges, not exact constants — phase, solvent polarity, and concentration all shift peaks by 10–30 cm⁻¹, which is why diagnosis always uses ranges and converging multiple pieces of evidence rather than a single exact number.

The systematic use of infrared absorption to identify organic functional groups grew out of early twentieth-century instrumentation work, but the person most responsible for turning IR into a routine structural tool for organic chemists was the American spectroscopist and physical chemist Robert Mecke's contemporaries, and most influentially in the English-speaking chemical world, Norman Sheppard and, earlier, the physicist-turned-spectroscopist tradition founded by William Weber Coblentz, whose painstaking early catalog of infrared absorption spectra (recorded one compound at a time, often overnight, with prism spectrometers he built himself) first established that molecules possess reproducible, assignable vibrational fingerprints at all.