Lewis Acidity Trends in Main-Group Halides

Boron trihalides and their Lewis acidity trend are not museum pieces: BF₃ (usually as its diethyl etherate, BF₃·OEt₂) is a everyday laboratory Lewis acid catalyst for Friedel–Crafts acylations, Mukaiyama aldol reactions, and epoxide-opening chemistry, chosen specifically because it is a milder, more selective acid than BCl₃ or BBr₃ — its weaker intrinsic acidity (from strong F π-donation) is a feature, giving better functional-group tolerance. BBr₃, by contrast, is prized industrially and in synthesis for demethylating aryl methyl ethers (ROMe → ROH), a reaction that requires a genuinely strong, hard halide-abstracting Lewis acid — directly exploiting its position at the strong end of the boron trihalide series. Aluminum trihalides remain central to petrochemical alkylation and isomerization catalysis and to classical Friedel–Crafts acylation/alkylation on an industrial scale, prized for cost and strength, with the Al₂Cl₆ dimer/monomer equilibrium a textbook illustration of how a highly Lewis acidic center can partially self-quench its own electron deficiency. Silicon tetrahalides and organosilicon halides are central to the entire silicone industry and to semiconductor-grade silicon purification (via SiCl₄ and SiHCl₃ distillation and reduction), where the facile coordination-number expansion of silicon governs both its reactivity toward water (a hazard to be managed) and its use in controlled hydrolysis/condensation to build Si–O–Si networks. Titanium and tin tetrahalides are core Lewis acid catalysts in Mukaiyama-type carbonyl activation and in Ziegler–Natta-adjacent polymerization chemistry. Group-15 pentafluorides underpin modern superacid science: PF₅ and especially SbF₅, combined with HF or HSO₃F, generate the strongest man-made Brønsted acid media, used in mechanistic organic chemistry (isolating and characterizing carbocations directly, work associated with Nobel-recognized superacid chemistry), in industrial isomerization/alkylation processes run under superacidic conditions, and in specialty fluorination chemistry. The same fluoride-ion-affinity logic used to rank PF₅ against SbF₅ here is the direct conceptual bridge into quantitative acidity scales — Hammett acidity functions and beyond — used to define and compare superacid strength, and into the broader hard/soft classification of Lewis acids and bases that organizes which base binds which acid most strongly across the whole periodic table.

Ask the tutor: Why does replacing just one fluorine in BF₃ with a chlorine (BF₂Cl) give an intermediate, non-additive Lewis acidity rather than a simple average of BF₃ and BCl₃?

Consider a trigonal-planar boron trihalide, BX₃. Boron is sp²-hybridized, using three sp² orbitals to form three B–X σ bonds, and it has one unhybridized 2p orbital left over, perpendicular to the molecular plane, that is completely empty. This empty p orbital is the Lewis acid site: it is what accepts the incoming lone pair from a base like pyridine or fluoride. But that same empty p orbital is not inert while BX₃ sits by itself. Each halogen X carries lone pairs in p orbitals of its own, and one of those lone pairs on each halogen is oriented so that it can overlap sideways with boron's empty 2p orbital — a classic π-donor/π-acceptor interaction, exactly analogous to the π system that stabilizes an amide or an allyl cation, except here the "π bond" is delocalized over four atoms (B and three X) and is a net electron donation from X into B. This partial B=X π character does two things: it shortens the B–X bond somewhat relative to a pure single bond, and — most importantly for acidity — it partially fills the orbital that a base would otherwise attack. A boron center with more π-donation into its empty p orbital is a weaker Lewis acid, because there is less empty orbital character left to offer a base. The strength of this π-donation depends on orbital size and energy matching. Fluorine's 2p orbitals are compact and energetically close to boron's 2p orbital (both period 2), so F–B π overlap is excellent — fluorine is a strong π-donor to boron despite being the most electronegative halogen. Chlorine's 3p orbitals are more diffuse and lower-lying relative to boron's small 2p orbital, so the Cl→B π overlap is significantly poorer; bromine and iodine are worse still, both because of the size/energy mismatch and because their higher electronegativity relative to size lowers p-orbital energy less than one might expect. The net result: π-donation ability runs F ≫ Cl > Br > I, exactly the reverse of what naïve electronegativity intuition suggests should matter, and for boron trihalides this π effect dominates the trend, giving the classic experimental Lewis acidity order

(BI₃ is often anomalously placed due to steric strain in adducts, discussed below.) This is measured directly by looking at which BX₃ binds a common reference base (pyridine, or trimethylamine) with the largest enthalpy of adduct formation, or by comparing halide-ion affinities. Now move one row down and one group over: aluminum trihalides, AlX₃. Aluminum is larger, its valence 3p (empty acceptor) orbital is more diffuse and lower in energy relative to the halogen p orbitals, and the geometric mismatch for π-overlap with any halogen — especially small, compact fluorine — is worse than it was for boron. Consequently π-donation is weak across the board for AlX₃, the σ-inductive effect (which favors F pulling density away and leaving Al positively charged and hungry for a base) reasserts itself, and monomeric AlX₃ acidity trends are less inverted, and can even run in the "expected" σ-dominated direction under some comparisons. More strikingly, aluminum trihalides are so electron-poor that they do not remain three-coordinate, planar monomers in the solid or in solution: AlCl₃ dimerizes to Al₂Cl₆, using bridging chlorides that each donate a lone pair to the neighboring aluminum, converting both aluminum centers to four-coordinate, tetrahedral, and largely satisfying the empty-orbital problem intermolecularly instead of intramolecularly by π-donation. This dimerization is itself direct evidence of high intrinsic Lewis acidity — AlCl₃ is such a strong acceptor that even a chloride bridging it to another AlCl₃ is more stable than staying monomeric.

For group 14, the picture changes qualitatively rather than quantitatively. Carbon in CX₄ is tetrahedral and sp³-hybridized: there is no empty low-lying orbital at all (carbon's 3d orbitals are far too high in energy to serve as acceptors), and carbon is already four-coordinate and geometrically saturated. CCl₄ is therefore essentially non-Lewis-acidic — it does not hydrolyze in water at any accessible rate, unlike SiCl₄, which reacts with water almost explosively. The difference is that silicon (and germanium, tin) can access a five-coordinate transition state or intermediate by using low-lying empty 3d-type (more accurately, σ*-hybrid) orbitals or simply by the geometric room afforded by a larger atomic radius, letting a water molecule (or fluoride, or another Lewis base) attack the Si center directly. This is a coordination-number story as much as an orbital-energy story: silicon, germanium, and tin tetrahalides are strong Lewis acids precisely because they can expand past four-coordinate, forming well-characterized five- and six-coordinate anionic adducts such as [SiF₅]⁻ and [SnCl₆]²⁻, whereas carbon simply cannot. Group-15 pentahalides push this logic further. PF₅ is already five-coordinate (trigonal bipyramidal) but remains a good fluoride acceptor because it can expand to six-coordinate octahedral [PF₆]⁻, a strongly stabilized closed-shell anion — this reaction, PF₅ + F⁻ → PF₆⁻, is essentially barrierless and hugely exothermic, which is exactly why PF₅/HF and (even more so) SbF₅/HF and SbF₅/HSO₃F are the superacid-generating combinations: SbF₅ is such a powerful fluoride (and general Lewis base) acceptor that it strips F⁻ (or the conjugate base of a strong Brønsted acid) away, leaving behind an extraordinarily reactive, "naked," highly acidic proton.

The comparative Lewis acidity of a main-group halide MXₙ toward a fixed reference base B can be decomposed conceptually into a σ-withdrawal contribution and a π-donation contribution:

where:

• $\Delta_{\text{acid}}$ is the net thermodynamic driving force for M to accept a lone pair from B (larger = more Lewis acidic),
• $\delta_\sigma(\chi_X)$ grows monotonically with halogen electronegativity $\chi_X$ (F > Cl > Br > I pulls density from M, making M more electrophilic),
• $\delta_\pi(X\to M)$ is the degree to which a filled halogen p lone pair donates into M's empty acceptor orbital, which reduces the accessible electrophilicity at M and is largest when X and M orbitals are close in size and energy.

Reading: whichever term dominates sets the direction of the trend. When $\delta_\pi$ is large and X-dependent (small, hard acceptor orbital on M, as in boron), the trend in $\Delta_{\text{acid}}$ can invert relative to electronegativity. When $\delta_\pi$ is small and roughly constant across X (large, diffuse acceptor orbital, or coordinatively saturated center), $\delta_\sigma$ dominates and the "expected" electronegativity-based order is recovered or coordination-number effects take over entirely. A second relationship quantifies why heavier group-14 and group-15 halides can be acidic at all — the thermodynamics of coordination-number expansion:

where:

• $\Delta G_{\text{exp}}$ is the free energy of forming the higher-coordinate adduct MXₙ·B,
• $\Delta H_{\text{M–B bond formed}}$ is the (favorable, negative) enthalpy of the new dative bond,
• $T\Delta S_{\text{assoc}}$ is the (unfavorable, positive when subtracted) entropy cost of associating two species into one,
• $\Delta H_{\text{rehybridization}}$ is the energetic cost of reorganizing the central atom's orbitals and geometry (e.g. sp² → sp³ at boron, or trigonal bipyramidal → octahedral at phosphorus).

Reading: a halide is a good Lewis acid in practice only if the new dative bond is strong enough to pay for both the entropy loss and any rehybridization penalty. This is why boron (sp²→sp³, modest penalty) forms adducts far more easily than would a hypothetical center needing to break strong existing bonds to rehybridize, and why phosphorus and antimony pentahalides, already close in geometry to their expanded coordination products, expand almost barrierlessly.

Example 1 (foundational): Rank BF₃, BCl�3, and BBr₃ by Lewis acidity toward a fixed base, and justify using bond-length data. Take trimethylamine, NMe₃, as the reference base, and use the experimentally known B–X bond lengths as structural evidence for π-bonding strength: B–F = 130 pm (BF₃ gas phase), B–Cl = 175 pm (BCl₃), B–Br = 187 pm (BBr₃). Compare each to the sum of Pauling single-bond covalent radii: B (82 pm) + F (72 pm) = 154 pm predicted vs. 130 pm observed, a contraction of 24 pm; B (82 pm) + Cl (99 pm) = 181 pm predicted vs. 175 pm observed, a contraction of only 6 pm; B (82 pm) + Br (114 pm) = 196 pm predicted vs. 187 pm observed, a contraction of 9 pm (some contraction persists here too, but proportionally the bond shortening effect, and independently measured adduct enthalpies, place Br lowest in donor strength among the three).

The largest bond-length contraction (relative to a single bond) occurs for B–F, consistent with the strongest π-donation there, tying down the most boron 2p character and leaving the least available for base attack. Measured NMe₃ adduct formation enthalpies confirm the resulting order:

Limiting-case check: if π-donation were negligible for all three halogens (hypothetically, if boron's 2p orbital were much lower in energy than any halogen lone pair, killing the overlap), the trend would collapse to pure σ-inductive control and should instead run BF₃ > BCl₃ > BBr₃, tracking electronegativity directly. The fact that the observed trend is inverted relative to this null hypothesis is itself the diagnostic signature that π-donation, not just electronegativity, is operating.

Example 2 (applied): Explain, quantitatively in reasoning terms, why AlCl₃ is used industrially as a Friedel–Crafts catalyst while BCl₃ — nominally a comparable trihalide — is used far less often for the same purpose despite being a strong Lewis acid. AlCl₃ is famous as the classic Friedel–Crafts catalyst, activating acyl and alkyl halides by abstracting a chloride to generate an acylium or carbocation-like electrophile:

Two structural facts explain why AlCl₃ (not BCl₃) dominates this industrial chemistry. First, aluminum's larger, more diffuse 3p acceptor orbital gives weak π-back-donation from chloride, so AlCl₃'s Lewis acidity is intrinsically very high (poor π-compensation of the σ-withdrawal), and — critically — aluminum's larger radius accommodates the resulting four-coordinate [AlCl₄]⁻ product comfortably, a geometry well suited to stabilizing the halide it has abstracted. Second, and just as important economically, AlCl₃ is far cheaper and easier to handle in bulk (made straightforwardly from aluminum metal or alumina and chlorine/HCl) than boron trichloride, which is more reactive toward adventitious moisture, more expensive, and (as established above) is actually a weaker halide acceptor than BBr₃ due to fluorine-type competition not applying here — but even among the chlorides, once BCl₃ abstracts Cl⁻ to give [BCl₄]⁻, the resulting boron center is four-coordinate tetrahedral, which is a bigger geometric reorganization (from trigonal planar sp² all the way to tetrahedral sp³) than aluminum undergoes if AlCl₃ is already somewhat associated (as Al₂Cl₆) — the net thermodynamics per mole of "active" monomeric acid ends up favoring AlCl₃ for typical Friedel–Crafts substrates.

Limiting-case check: if boron's π-donation from Cl were completely absent (a limiting hypothetical), BCl₃ would be an even stronger acid than observed, more directly comparable to AlCl₃; the actual, smaller-than-naive acidity gap between the two — and BCl₃'s continued practical disadvantage — is consistent with a modest but real residual π-donation even for the larger halogen, plus the practical/cost factors above.

Example 3 (exam/competition level): Predict and rationalize the relative Lewis acidity order for PF₅, PCl₅, SbF₅, and SbCl₅ toward F⁻, and explain why SbF₅ is singled out as the acid component of "magic acid." Set up the governing question as one of fluoride-ion affinity (FIA), the enthalpy of

Three factors matter, layered on top of each other: (1) central-atom size and polarizability (Sb > P favors stronger, more polarizable acceptor orbitals and better stabilization of the incoming anionic charge), (2) the identity of the existing ligands X (F ligands already present make the central atom more electrophilic than Cl ligands, because F is more σ-withdrawing and, unlike in the boron case, there is no competing π-donation pathway of comparable importance once the center is already 5-coordinate and largely σ-bonded), and (3) the stability of the resulting 6-coordinate octahedral anion, which is excellent for both P and Sb (both readily adopt octahedral MF₆⁻/MCl₆⁻-type geometries). Combining (1) and (2): antimony is both larger/more polarizable and — in the specific case of SbF₅ — carries the more σ-withdrawing fluoride ligands already in place, compounding its electrophilicity. The predicted order is

with SbF₅ standing out as the strongest fluoride (and general Lewis base) acceptor of the four, which is exactly the experimentally established order (fluoride-ion affinity measurements and computed values place SbF₅ well above PF₅, and the antimony pentahalides above the corresponding phosphorus ones). This is precisely why SbF₅ mixed with HSO₃F (fluorosulfonic acid) gives "magic acid": SbF₅ abstracts the conjugate base (FSO₃⁻) from HSO₃F so effectively that the remaining proton is left with an extraordinarily weakly coordinating, highly delocalized counter-anion [Sb₂F₁₀(SO₃F)]⁻ or related aggregates, making the medium acidic enough to protonate alkanes.

Limiting-case check: if antimony behaved identically to phosphorus (same size, same orbital energies — the "equal-acceptor" limiting case), the F/Cl ligand effect alone would still predict SbF₅ > SbCl₅ and PF₅ > PCl₅, but the magnitude of the SbF₅ vs PF₅ gap would shrink to reflect only the (smaller) ligand-substitution effect; the large observed gap confirms that the central-atom size/polarizability term is doing real, independent work, not just the ligand identity.

Lewis worked largely in thermodynamics and chemical bonding rather than in the specific halide chemistry discussed here, but the entire conceptual apparatus of this article — asking "where is the empty orbital, and how easily can a base's lone pair reach it" — is a direct, literal application of his 1923 definition to real molecules. The subsequent development of the π-donation/σ-withdrawal analysis for boron trihalides, largely worked out through structural (electron diffraction, later X-ray and computational) and thermodynamic (calorimetric adduct-formation) studies through the mid-20th century, is the natural experimental completion of Lewis's qualitative picture: once chemists could measure bond lengths and adduct enthalpies precisely, the "obvious" electronegativity-based prediction for BX₃ was falsified by data and replaced with the orbital-overlap argument given above.