Carboxylic Acids as Bases and Protonation

Carbonyl-oxygen protonation is not a museum piece of mechanistic pedantry; it is the operating principle behind an enormous swath of applied organic and industrial chemistry. Every acid-catalyzed Fischer esterification — from the lab-scale synthesis of simple flavor and fragrance esters (ethyl acetate, isoamyl acetate, methyl salicylate) to industrial-scale production of plasticizer esters and biodiesel via acid-catalyzed transesterification of fatty acid triglycerides — begins with exactly this protonation event activating the carbonyl carbon toward nucleophilic attack by an alcohol. In pharmaceutical process chemistry, acid-catalyzed hydrolysis and esterification steps (protecting-group removal, prodrug ester cleavage, peptide-coupling side-reaction suppression) are routinely designed around the same protonation logic: choosing an acid strong enough to populate the reactive protonated carbonyl at a useful rate, but mild enough to avoid unwanted side reactions such as dehydration or rearrangement. Superacid chemistry (George Olah's work on stable carbocations, for which he won the Nobel Prize) directly exploited the extreme end of this protonation equilibrium, using superacid media (HF–SbF₅, FSO₃H–SbF₅) to fully protonate carboxylic acids and even push the chemistry further toward acylium ion formation by loss of water from the doubly-protonated species — a reaction only accessible because the initial, ordinary carbonyl-oxygen protonation event is so well understood and so reliably the first step. In biochemistry, the same logic underlies general acid catalysis by active-site residues (protonated carboxylic acid side chains of aspartate/glutamate, or protonated histidine) that activate carbonyl substrates in enzyme mechanisms — lysozyme and many esterase/protease active sites use precisely this kind of carbonyl-oxygen protonation (or its kinetic equivalent, hydrogen-bond donation that partially protonates the carbonyl oxygen in the transition state) to accelerate nucleophilic attack on an otherwise sluggish carbonyl carbon, at physiological pH, by using a well-positioned acidic side chain rather than bulk strong acid. Understanding stoichiometric carbonyl protonation in solution is therefore the direct conceptual ancestor of understanding enzymatic general acid catalysis, and it is the load-bearing first step every student needs before tackling Fischer Esterification and Its Equilibrium Mechanism rigorously.

Ask the tutor: Why does adding water to the reaction mixture shift Fischer esterification backward even though water isn't directly involved in the carbonyl-protonation step?

Start from the neutral carboxylic acid itself. The carbon is $sp^2$-hybridized, bonded to the carbonyl oxygen (C=O), the hydroxyl oxygen (C–OH), and R. There are three atoms feeding electron density into one planar π system: the carbonyl oxygen's $p$ orbital (part of the formal C=O π bond) and the hydroxyl oxygen's lone pair, which sits in a p orbital aligned parallel to the π system, allowing it to donate into the C=O π the same way an amide nitrogen lone pair donates into its adjacent carbonyl. This three-atom, four-electron delocalized π system (O=C–O with one lone pair donating) is why the C–OH bond in a carboxylic acid has partial double-bond character and why rotation about it is somewhat restricted, and it is the structural feature that makes the whole basicity argument work. Now bring in an external acid, H–A. Two oxygens present lone pairs to that proton. If the proton adds to the <strong>hydroxyl oxygen</strong>, that oxygen — already donating a lone pair into the ring system and now bonded to two substituents plus a new proton — becomes positively charged directly, with no adjacent atom able to share that charge by resonance; the carbonyl system is orthogonal in character and cannot stabilize an adjacent, doubly-alkylated oxonium in a symmetric way. Worse, protonating the OH oxygen would place a full positive charge on the more electronegative, less basic of the two oxygens while simultaneously destroying the resonance donation that stabilized the neutral acid, since the lone pair being protonated was the one delocalized into the π system. If instead the proton adds to the <strong>carbonyl oxygen</strong>, something very different happens. That oxygen already carries substantial electron density from the formal π bond; protonating one of its remaining lone pairs generates a cation whose positive charge can be pushed entirely onto the carbon*, restoring a full lone pair on the carbonyl oxygen as a neutral oxygen and forming a new π bond between carbon and the hydroxyl oxygen. This is possible because the hydroxyl oxygen still carries a lone pair free to become a new π donor. The result is two resonance structures of exactly the same energetic class — protonated-carbonyl-oxygen–with-positive-charge-on-oxygen, and positive-charge-fully-on-carbon-with-both-oxygens-neutral-and-π-bonded — related by delocalization across an oxygen–carbon–oxygen three-atom system exactly analogous to (but charge-inverted from) the carboxylate anion.

The favorable pathway is the cationic mirror image of carboxylate stabilization. In the carboxylate anion, deprotonation removes a proton and leaves a negative charge that resonance-delocalizes symmetrically over both oxygens, each carrying half a negative charge and half a C–O bond order beyond single. In the protonated acid, an added proton on the carbonyl oxygen leaves a cation that resonance-delocalizes symmetrically in reverse — full positive charge migrating between the two oxygens' bonding descriptions with substantial character sitting on carbon in the major contributing structure. The comparison is worth holding side by side, because the logic — count the atoms that can share the charge — is identical; only the sign of the charge and the direction of the arrows flip.

The key quantitative payoff of this picture is that the protonated carbonyl's electrophilic carbon is much more reactive toward nucleophiles than the neutral carbonyl carbon ever was — not because a new bond type appeared, but because one resonance structure (out of the two dominant ones) places the entire formal positive charge on carbon with only single bonds to both oxygens, an electronic environment essentially identical to an oxocarbenium ion. A nucleophile attacking this carbon is attacking a genuine cation, not a polarized-but-neutral carbonyl, and that difference in electrophilicity is exactly what a strong acid buys you catalytically.

The basicity of the carbonyl oxygen is quantified the same way any base's strength is: by the acidity of its conjugate acid.

where:

• $\text{RC(OH)}_2^+$ is the protonated (conjugate acid) form, drawn here as its major carbon-cationic resonance contributor,
• $K_{aH}$ is the acid dissociation constant of the conjugate acid, describing how readily the protonated species gives the proton back,
• a small $K_{aH}$ (very negative $pK_{aH}$) means the conjugate acid is a strong acid, i.e. the neutral carboxylic acid is a very weak base.

where:

• this places carboxylic acids among the weakest common Brønsted bases in organic chemistry — far weaker bases than water ($pK_{aH}$ of $\text{H}_3\text{O}^+$ is $-1.7$), amides ($pK_{aH}\approx-0.5$ to $0$), or esters ($pK_{aH}\approx-6.5$, similar order),
• meaningful protonation therefore requires strong mineral or superacids (concentrated H₂SO₄, HCl gas, triflic acid) rather than mild aqueous acid.

The Henderson–Hasselbalch-style relation tells you the fraction protonated at any acid strength:

where:

• $H_0$ is the Hammett acidity function, the appropriate stand-in for pH in strongly acidic, often non-aqueous media,
• reading this equation: only when the medium's $H_0$ approaches or exceeds the magnitude of $|pK_{aH}|$ does the protonated form become a significant fraction of the total — which is exactly why Fischer esterification is normally run with strong mineral acid catalysts (e.g., concentrated H₂SO₄, $H_0 \approx -12$) rather than weak acids.

The regiochemical preference for carbonyl-oxygen protonation is captured qualitatively by comparing the number of resonance structures (and their charge distribution) available to each protonation product:

where:

• this energy difference is large (tens of kJ/mol by computational estimates), reflecting the loss of resonance delocalization and the destabilizing placement of full positive charge on a non-stabilized oxygen when protonation occurs at the hydroxyl oxygen instead,
• the practical consequence is that essentially 100% of the protonated population sits as the carbonyl-oxygen-protonated form under any conditions where protonation occurs at all.

Example 1 — Foundational: draw the fully arrow-pushed protonation of acetic acid and assign formal charges at every atom. Take acetic acid, $\text{CH}_3\text{COOH}$, and a strong acid $\text{H}_2\text{SO}_4$ (treat as delivering $\text{H}^+$ with $\text{HSO}_4^-$ as spectator counterion). Identify the carbonyl oxygen's two non-bonding, non-π lone pairs (in-plane, $sp^2$-like) as the nucleophile. Curved arrow 1: one of those lone pairs attacks the acidic proton of H₂SO₄, forming a new O–H bond; simultaneously curved arrow 2: the S–O(H) bond of sulfuric acid breaks heterolytically, electrons going onto that oxygen, generating $\text{HSO}_4^-$. This gives the resonance structure $\text{CH}_3\text{–C(=OH}^+\text{)–OH}$: carbonyl oxygen now bears a formal charge of $+1$ (three bonds: one π, one σ to C, one σ to new H; two remaining lone pairs, one fewer than neutral oxygen), carbon remains neutral $sp^2$ with formal charge $0$, hydroxyl oxygen remains neutral with formal charge $0$. Draw the second resonance structure: curved arrow 3, the hydroxyl oxygen's remaining $p$-orbital lone pair moves in to form a new π bond to carbon; curved arrow 4, the original C=O π electrons shift entirely onto the carbonyl oxygen. Result: $\text{CH}_3\text{–C}^+\text{(OH)(OH)}$ — carbon now formally $sp^2$ cationic with formal charge $+1$ (three σ bonds, empty $p$ orbital), both oxygens neutral (each with three lone pairs minus the ones used in bonding — each oxygen here has two lone pairs and two σ bonds, formal charge $0$). Check the overall charge balance: in both resonance structures, total charge on the cation is $+1$, consistent with adding one $\text{H}^+$ to a neutral starting acid. The true species is the resonance hybrid of these two structures (plus a very minor contribution from a hydroxyl-oxygen-protonated form, negligible by the argument above), with partial positive charge shared mainly across the carbon and both oxygens, and C–O bond orders intermediate between one and two on both sides — structurally and electronically the cationic analogue of the acetate anion.

Example 2 — Applied: estimate the degree of protonation of benzoic acid ($pK_{aH} \approx -7.7$) in three media, and connect this to observed esterification rates. Use

Take three media: dilute aqueous HCl ($H_0 \approx 0$, i.e. roughly pH 0), concentrated aqueous HCl (12 M, $H_0 \approx -6$), and concentrated sulfuric acid ($H_0 \approx -12$). For dilute HCl ($H_0 = 0$):

Essentially no protonation — a vanishingly small fraction of the acid exists in the reactive, electrophilic form, and correspondingly esterification is immeasurably slow under these conditions. For 12 M HCl ($H_0 = -6$):

Still astonishingly small in absolute terms — even very concentrated aqueous HCl barely protonates a carboxylic acid — but the ratio has grown by six orders of magnitude relative to dilute acid, and since the rate of nucleophilic attack on the protonated species is enormously faster than on the neutral acid, the observed esterification rate rises substantially even though the equilibrium concentration of protonated acid remains tiny. For concentrated H₂SO₄ ($H_0 = -12$):

This looks paradoxically smaller than the 12 M HCl case at face value, which flags an important subtlety: the Hammett acidity function values used here are illustrative order-of-magnitude placeholders, and in practice concentrated sulfuric acid ($H_0$ closer to $-12$ but measured under conditions very different from dilute aqueous scales) is empirically the far more effective protonating and esterification-catalyzing medium, because $H_0$ for concentrated $\text{H}_2\text{SO}_4$ is defined self-consistently in a non-aqueous medium where water activity is suppressed and the effective protonating power for weakly basic carbonyls is far higher than a naive linear extrapolation from dilute aqueous $H_0$ values suggests. The physically robust conclusion, independent of the exact numeric placeholders, is the qualitative trend: more concentrated, less-hydrated strong acid gives more protonation and faster catalysis, which is precisely why Fischer esterification is run with concentrated H₂SO₄ or anhydrous HCl gas rather than dilute aqueous acid — dilution both lowers acid strength and (worse) adds water that shifts the esterification equilibrium backward.

Example 3 — Exam/competition level: rank three carbonyl-containing bases by basicity at the carbonyl oxygen and justify using resonance/inductive arguments — acetic acid vs. acetamide vs. ethyl acetate. The three conjugate acids to compare (all protonated at the carbonyl oxygen, all stabilized by the analogous three-atom O–C–X resonance system where X = OH, NH₂, or OEt) have approximate literature $pK_{aH}$ values: acetamide $\approx 0$, ethyl acetate $\approx -6.5$, acetic acid $\approx -6.1$ (values vary modestly by source and method, but the ordering is robust and mechanistically explicable). Reasoning: in every case, protonation at the carbonyl oxygen generates a cation stabilized by resonance donation from the adjacent heteroatom X (N, O of ester, or O of acid) exactly as worked out in the derivation above. The strength of that resonance stabilization tracks the intrinsic lone-pair-donating ability of X into the developing empty $p$ orbital on carbon:

Nitrogen is a substantially better π-donor than oxygen (nitrogen is less electronegative, holds its lone pair less tightly, and its lone pair energy is higher and better matched to the developing empty carbon $p$ orbital), so the amide's protonated cation gains more resonance stabilization from its N–C(=OH⁺) ⟷ N⁺=C(–OH) delocalization than either the ester or acid gains from analogous oxygen-based donation. This is why protonated amides are markedly more stable (weaker conjugate acids, i.e. the neutral amide is a stronger base) than protonated esters or acids — reflected in the amide's much higher, near-neutral $pK_{aH}$. Between the ester and the free acid, the comparison is closer: both use an oxygen lone pair as the resonance donor, but the ester's alkoxy oxygen is a slightly better donor than the acid's hydroxyl oxygen is, since the acid's O–H is inductively more electron-withdrawing on its own oxygen than the ester's O–alkyl group is (alkyl groups are weakly electron-donating inductively relative to H), very slightly favoring ester basicity — consistent with the ester and acid conjugate-acid $pK_{aH}$ values sitting close together and in the same very-weak-base regime, both far below the amide. The mechanistic takeaway generalizes cleanly: wherever you see resonance-stabilized protonation at a carbonyl oxygen, rank basicity by ranking the donor ability of the adjacent heteroatom — exactly the same donor-ability ranking (N > O of ester/acid ≈, and both ≫ a carbonyl with no adjacent lone-pair donor at all, such as a simple aldehyde or ketone) that governs amide resonance and ester reactivity everywhere else in carbonyl chemistry.

The systematic understanding of carbonyl basicity and protonation site selectivity emerged from mid-20th-century physical organic chemistry, particularly from the work connecting Hammett acidity functions to carbonyl reactivity — work most closely associated with Louis Plack Hammett, whose acidity function $H_0$ (introduced to quantify "how acidic" concentrated, non-aqueous or highly concentrated acidic media really are, beyond what pH alone can express) became the essential tool for putting numbers on exactly the kind of weak-base protonation equilibria described in this article. Hammett's broader program — relating structure to reactivity through linear free-energy relationships — gave physical organic chemistry its quantitative backbone, and the acidity function is the direct descendant of that program applied to superacid and concentrated-acid media. The specific mechanistic question of where a carbonyl compound protonates (oxygen vs. an alternative site, and which oxygen in polyoxygenated systems) was resolved through a combination of UV/IR spectroscopy of protonated carbonyls in superacid media (observing the characteristic shift of the C=O stretch and the appearance of new O–H-type bands consistent with symmetric O-protonation) and NMR studies in the 1960s–70s, building on the theoretical resonance-structure reasoning that Linus Pauling and others had already established for delocalized cations and anions generally. The application of that resonance logic specifically to carboxylic acid derivatives — showing that carbonyl-oxygen protonation dominates because of the symmetric O–C–O delocalization — became a textbook staple precisely because it is derivable from first principles rather than needing to be memorized, which is the spirit in which it is presented here.