Lewis Acids and Bases

Lewis acid catalysis is a pillar of modern organic and organometallic synthesis, not a historical footnote. Chiral Lewis acid catalysts — BINOL-derived titanium and boron complexes, oxazoline-copper complexes, chiral aluminum salen catalysts — enable catalytic asymmetric Diels-Alder, Mukaiyama aldol, and epoxide-opening reactions that are workhorse steps in pharmaceutical process chemistry, because activating a substrate's LUMO with a chiral Lewis acid transmits stereochemical information directly into the product without stoichiometric chiral auxiliaries. Zeolites and metal-organic frameworks used as industrial cracking and isomerization catalysts function through surface Lewis-acidic aluminum and metal sites that protonate or ionize hydrocarbons at internal pore surfaces, running billions of dollars of petroleum refining on exactly the orbital logic developed here. In biology, Lewis acid-base chemistry underlies metalloenzyme catalysis broadly: zinc in carbonic anhydrase and in many hydrolases acts as a Lewis acid that polarizes a bound water molecule (lowering its effective pKa dramatically so it can act as a nucleophile at physiological pH), and Mg²⁺ ions in polymerase active sites act as hard Lewis acids stabilizing developing negative charge on phosphate oxygens during phosphoryl transfer. Frustrated Lewis pair (FLP) chemistry, developed extensively since the mid-2000s, exploits bulky, sterically mismatched Lewis acid/base pairs that cannot quench each other intramolecularly, and such pairs can heterolytically split H₂ and CO₂ without any transition metal at all — a genuinely new mode of catalysis built entirely on the Lewis framework, now being explored for metal-free hydrogenation catalysis and small-molecule activation relevant to carbon capture.

Ask the tutor: Can you walk through why AlCl₃ is considered a stronger Lewis acid than BF₃ toward most organic carbonyl substrates, using orbital size and polarizability arguments?

Start from orbitals, not from formalism. A Lewis base has some highest occupied molecular orbital (HOMO) that is spatially exposed and energetically high — a lone pair on nitrogen or oxygen, a π bond, or even a filled σ bond in special cases (hyperconjugative donors). A Lewis acid has some lowest unoccupied molecular orbital (LUMO) that is spatially exposed and energetically low — an empty p-orbital on boron or a carbocation, a σ orbital of a polarized C–X bond, or an empty d-orbital / hybrid orbital on a metal center. When these two orbitals approach with the right symmetry and phase match, the HOMO of the base pours electron density into the LUMO of the acid. A new bonding molecular orbital is born, occupied by the two electrons that started life entirely on the base. This is the single event that a curved arrow draws. Compare directly with Brønsted-Lowry chemistry. In HCl + NH₃ → NH₄⁺ + Cl⁻, the nitrogen lone pair (HOMO of ammonia) attacks the σ orbital of the H–Cl bond (LUMO of HCl). The "acid" HCl is really donating its most electrophilic site — the antibonding orbital sitting behind the departing chloride — to accept the nitrogen lone pair. Brønsted acidity is not a different phenomenon from Lewis acidity; it is the specific sub-case where the accepting fragment happens to leave behind a proton. Every Brønsted acid is a Lewis acid delivering H⁺ as the electrophilic center; every Brønsted base is automatically a Lewis base, since anything basic enough to grab a proton has a lone pair to donate. The generalization becomes powerful once you admit acceptors that are not protons. Boron trifluoride has an empty 2p orbital on boron (boron is only 6-electron, sp² trigonal planar) — this is a textbook hard Lewis acid. Anhydrous AlCl₃ similarly has an empty coordination site. A carbocation R₃C⁺ has an empty p-orbital orthogonal to the three σ bonds — it is a carbon-centered Lewis acid, and this is why carbocations are trapped by any nucleophile in solution. Even a neutral carbonyl carbon is a (weak, but real) Lewis acid: the C=O π orbital is low-lying and accessible to nucleophilic attack, which is exactly why Grignard reagents, hydrides, and amines all add to aldehydes and ketones. The carbonyl oxygen's lone pairs, meanwhile, make the carbonyl oxygen itself a Lewis base* toward external Lewis acids like BF₃ or a proton — carbonyl compounds are amphoteric in the Lewis sense, simultaneously an acid at carbon and a base at oxygen, and which face reacts depends on what partner shows up.

The Brønsted-only picture also fails to explain catalysis by Lewis acids, which is ubiquitous in synthesis. Consider TiCl₄ activating an aldehyde toward a Mukaiyama aldol, or ZnCl₂ activating an epoxide toward ring opening, or Sc(OTf)₃ activating an imine toward a Mannich addition. In every one of these, the metal's empty orbital coordinates to the substrate's lone pair (usually oxygen or nitrogen), and this coordination withdraws electron density from the rest of the substrate, lowering its LUMO and making the substrate carbon dramatically more electrophilic toward an external nucleophile. This is activation by orbital-energy tuning, not by protonation, and it is why anhydrous, water-free, strongly Lewis-acidic conditions are used throughout synthetic methodology.

The Lewis definitions, stated precisely:

where: A is any species with an accessible empty orbital (the electron-pair acceptor), :B is any species with an accessible lone pair or π system (the electron-pair donor), and A–B is the new dative (coordinate covalent) bond, formally a single two-electron bond in which both electrons originated on B.

where: $c_A, c_B$ are the orbital coefficients of A and B at the reacting atoms, $\beta$ is the resonance (overlap) integral, and the denominator is the HOMO–LUMO energy gap between donor and acceptor — this is the perturbation-theory expression (from second-order perturbation theory, the same math underlying frontier molecular orbital theory) showing that stabilization grows as the gap shrinks and as the orbital overlap grows. Reading it: a good Lewis acid-base match needs both energetic proximity (small gap) and good spatial/symmetry overlap — the second factor is exactly what HSAB theory tracks qualitatively.

where: $\eta$ is the chemical hardness of a species (Pearson/Parr-Pearson operational definition), a large HOMO–LUMO gap ($\eta$ large) marks a "hard" acid or base (charge-controlled, small polarizable centers, e.g. H⁺, F⁻, Mg²⁺), and a small gap ($\eta$ small) marks a "soft" acid or base (orbital-controlled, large polarizable centers, e.g. I⁻, R–S⁻, Pt(0), carbenium ions with extended conjugation). Reading it: hardness is literally the size of the frontier-orbital gap; soft species have HOMO and LUMO close together and interact best with other soft species (both terms in the perturbation expression favor matched-gap pairs).

Example 1 (foundational) — Classify every species in NH₃ + BF₃ → H₃N–BF₃, and draw the orbital event. Ammonia has a nitrogen lone pair in an sp³ hybrid orbital, energetically high and spatially exposed — this is the Lewis base, donor HOMO. Boron trifluoride is trigonal planar (sp² boron, three B–F σ bonds, 6 valence electrons on B, empty 2p orbital perpendicular to the BF₃ plane) — this is the Lewis acid, acceptor LUMO. The curved arrow runs from the nitrogen lone pair directly into boron's empty p orbital, along the developing N–B axis. As the new N–B σ bond forms, boron must rehybridize from sp² (trigonal planar) to sp³ (tetrahedral) to accommodate the fourth bond — the BF₃ fragment puckers from planar to pyramidal as the adduct forms, exactly as VSEPR (and orbital hybridization) demands for four σ bonds around one center. Formal charges: N goes from 0 to +1 (four bonds now on nitrogen), B goes from 0 to −1 (four bonds, complete octet, formally anionic boron in the adduct H₃N⁺–B⁻F₃). No proton has moved anywhere in this reaction — it is Lewis acid-base chemistry with zero overlap with the Brønsted framework, and yet it is mechanistically identical in electron-pushing logic to the NH₃ + HCl case worked above. As a limiting-case check: if instead of BF₃ we used the far bulkier tri-tert-butylborane, the same electronic argument (empty boron p-orbital, ammonia lone pair) predicts adduct formation, but severe steric clash between the ammonia hydrogens and the tert-butyl groups raises the activation barrier and can prevent adduct formation outright at room temperature — a reminder that the Lewis acid-base framework predicts electronic feasibility, while steric strain (frontier-orbital overlap must still be geometrically achievable) governs whether the adduct is actually observed. This is precisely the basis of "frustrated Lewis pairs," where a bulky Lewis acid and bulky Lewis base cannot approach closely enough to quench each other, yet both retain independent reactivity toward small molecules like H₂.

Example 2 (applied) — Friedel-Crafts acylation: mechanistic role of AlCl₃ as Lewis acid, with numbers. Acetyl chloride, CH₃COCl, reacts with benzene only in the presence of a strong Lewis acid catalyst such as AlCl₃ (a stoichiometric-to-catalytic quantity is typically required because the ketone product itself is Lewis basic and sequesters some AlCl₃, an important practical point). Mechanism: Step 1. AlCl₃'s aluminum center has an empty coordination site (AlCl₃ is a 6-electron, sp² planar Lewis acid, isoelectronic in behavior to BF₃, though a stronger Lewis acid because aluminum is larger, more polarizable, and less electronegative, giving better orbital overlap with soft donors and a genuinely lower-lying acceptor orbital). One chlorine lone pair of acetyl chloride donates into Al's empty orbital:

Step 2. The C–Cl bond heterolyzes fully (assisted by the developing negative charge on the AlCl₄⁻ fragment, which is a good, stabilized leaving group once four Al–Cl-type bonds delocalize the negative charge over a tetrahedral aluminate), generating a resonance-stabilized acylium ion, CH₃–C≡O⁺ ↔ CH�3–C⁺=O, and free AlCl₄⁻. Step 3. The acylium carbon (an sp-hybridized carbocation-like Lewis acid, its LUMO an in-plane p orbital very low in energy because of stabilization by the adjacent oxygen lone pair through the alternative resonance structure) is attacked by the π system (HOMO) of benzene, forming the C–C bond and a benzenonium (arenium) cation — a second Lewis acid-base event, this time with the arene π cloud as donor. Step 4. Loss of H⁺ from the sp³ carbon of the arenium ion (rearomatization, itself a Brønsted-type step, base = AlCl₄⁻ acting as a proton acceptor) regenerates aromaticity and releases HCl + AlCl₃, turning over the catalyst. Quantitatively, this cascade is why the acyl chloride alone (with its own polarized, but only modestly electrophilic C=O carbon, Lewis-acid-strength far below a free acylium) is essentially unreactive toward benzene at room temperature, while the AlCl₃-generated acylium ion (fully charge-separated, LUMO drastically lower) reacts readily: the Lewis acid did not touch the arene at all, it activated the electrophile by removing chloride as AlCl₄⁻, exactly the LUMO-lowering activation logic derived above generalized from "coordination" (BF₃ example) to full ionization (AlCl₃/acyl chloride example) — the two are points on the same continuum of Lewis acid strength.

Example 3 (exam / competition-level) — Predict and rationalize the regiochemistry of a Lewis-acid-catalyzed epoxide opening using HSAB reasoning. Consider 2-methyl-2,3-epoxybutane (an epoxide with one tertiary carbon and one secondary carbon on the ring) treated with (a) NaOMe (a strong, hard, unhindered nucleophile/base, no Lewis acid present) versus (b) catalytic BF₃·OEt₂ with the same methanol nucleophile present in the medium. Case (a), base-mediated (SN2-like) opening. Methoxide (a hard, small, strongly basic nucleophile, orbital-controlled interaction weak, charge-controlled interaction dominant) attacks the epoxide carbon that is least sterically hindered and bears the least positive character in the ground state, because there is no pre-formed cationic character to stabilize — backside attack at the less-substituted (secondary) carbon, classic SN2 geometry, inversion of configuration at that carbon. Case (b), Lewis-acid-activated opening. BF₃ first coordinates the epoxide oxygen lone pair (a Lewis acid-base event exactly analogous to the BF₃-carbonyl case above), which polarizes and weakens both C–O bonds of the strained three-membered ring. Crucially, coordination pulls electron density preferentially through whichever C–O bond can best stabilize developing positive charge — that is the bond to the tertiary carbon, since a tertiary carbocation-like transition state (three alkyl groups hyperconjugating into the empty p orbital) is far more stabilized than a secondary one. The ring opens with substantial SN1-like character: bond-breaking to the tertiary carbon is advanced in the transition state, and the nucleophile (methanol) attacks predominantly at the more substituted (tertiary) carbon.

This is a textbook divergence in regiochemistry driven entirely by whether a Lewis acid is present, and it follows directly from the orbital argument: Lewis acid coordination lowers the LUMO and imparts cationic character preferentially at the carbon best able to stabilize positive charge, switching the mechanism from concerted backside attack (hard nucleophile, no cation stabilization needed) toward attack at the more electrophilic, more substituted carbon (soft, polarizable transition state, orbital-controlled). As a limiting-case check: for a symmetrical epoxide (both ring carbons secondary, e.g., 2,3-epoxybutane), this regiochemical divergence vanishes entirely — both mechanisms converge on the same carbon by symmetry, and only the stereochemical outcome (inversion at both centers under acid-catalyzed conditions with attack occurring preferentially anti to the more hindered face, versus clean single-center inversion under base conditions) distinguishes the two pathways, confirming that the effect described is a genuine substituent-driven electronic effect, not an artifact of the analysis.

Gilbert N. Lewis is one of the most consequential and least publicly famous chemists of the twentieth century — a fact usually attributed to his failure ever to win a Nobel Prize despite (by most historians' reckoning) deserving one for multiple separate contributions. Trained at Harvard, Lewis spent most of his career building the chemistry department at the University of California, Berkeley, into a research powerhouse, mentoring an extraordinary number of future Nobel laureates (Willard Libby, Glenn Seaborg, William Giauque, Harold Urey) while himself introducing the shared electron-pair bond (1916), the octet rule, the "Lewis structure" notation every student learns before they can define a functional group, and — the subject of this article — the generalized acid-base definition (1923), published the same year as the independent and narrower Brønsted-Lowry proton-transfer definition. Lewis deliberately built his acid-base theory to be broader than Brønsted's, arguing that restricting acidity to proton donors artificially excluded chemically identical behavior by species like BF₃ and metal cations — a claim that has been thoroughly vindicated by a century of organometallic and mechanistic organic chemistry, where Lewis acid catalysis is arguably more central to modern synthesis than Brønsted acid catalysis. Lewis also did foundational thermodynamic work (defining fugacity and activity, effectively modernizing chemical thermodynamics for real, non-ideal systems) and was deeply engaged in isotope chemistry, being among the first to isolate pure deuterium oxide (heavy water). He died in his Berkeley laboratory in 1946, apparently of a heart attack suffered while working with hydrogen cyanide, a detail that has fueled decades of historical speculation about whether his death was accidental.