Group 13 Trihalides and Adduct Formation
Melt aluminium chloride and something strange happens. The vapor above the liquid is twice as heavy as AlCl₃ "should" be. The molecule has quietly dimerized into Al₂Cl₆. This structure consists of two AlCl₄ tetrahedra fused at an edge. It is held together by chlorine atoms. Each donates a lone pair into an empty orbital on the neighboring aluminum. Meanwhile BCl₃, built from the element one row up in the same group, shows no such behavior at all. It stays resolutely monomeric, planar, and only reluctantly forms adducts with strong bases. Same group, same +3 oxidation state, same nominal Lewis acidity. Yet the structural chemistry is wildly different. Explaining that difference is the whole point of this article. Group 13 trihalides (BX₃, AlX₃, GaX₃, InX₃, TlX₃) are the textbook Lewis acids of the periodic table. Their reactivity is the gateway to an enormous swath of applied chemistry. Friedel–Crafts catalysis depends on it. So do Ziegler–Natta and metallocene polymerization co-catalysts. Chlorosilane and semiconductor precursor chemistry rely on it too. So does the industrial production of aluminum itself. All of this hinges on an MX₃ center's appetite for an extra pair of electrons. Yet that appetite is satisfied differently depending on what's available. If no external base is present, some trihalides feed on themselves. They form halide-bridged dimers or higher oligomers — a kind of self-adduct. If an external Lewis base (ether, amine, halide ion) is present, that self-association is broken open. It is replaced by a discrete 1:1 or 1:2 adduct. Understanding which trihalides dimerize, why, and how easily they surrender that dimer to an incoming base is the central skill this article builds. The reasoning draws on three ideas that should already feel familiar. First, the geometry of an sp²-hybridized, six-electron boron or aluminum center (Boron Halides and Lewis Acidity). Second, the general behavior of aluminum in its +3 state (Aluminium Chemistry and the +3 State). Third, simple periodic trends in orbital size and π-overlap efficiency. What's new here is putting them together to explain a structural fork in the road. Small, first-row boron forms strong B–X π-bonds that satisfy its electron deficiency internally. So it stays monomeric and is a comparatively mild, sterically fussy Lewis acid. Larger group 13 atoms — aluminum, gallium, indium — cannot form efficient p_π–p_π bonds with halogens (poor 2p/3p, 2p/4p, 2p/5p overlap). So they crave the extra electron density enough to dimerize via three-center-four-electron bridging bonds. They also attack incoming Lewis bases far more readily. By the end you will be able to predict, from first principles, whether a given MX₃ is monomeric or dimeric in the gas and solid phases. You will be able to rank the Lewis acidity of the trihalides across a period and down a group. You will be able to explain why BF₃ is the outlier among boron halides. You will also be able to describe the mechanism by which an external base cracks open the Al₂Cl₆ dimer. This sets up the broader theme of Amphoterism in the p-Block, where the same acid–base logic reappears in aqueous and oxide chemistry.
- MX₃ (M = B, Al, Ga, In, Tl; X = halogen) are trigonal-planar, sp²-hybridized, six-valence-electron molecules — inherently electron-deficient at M.
- BX₃ stays monomeric because small 2p orbitals on B and X overlap well, forming a genuine B–X π-bond that partially fills the empty p orbital.
- AlCl₃, AlBr₃, GaCl₃, and related heavier trihalides dimerize (M₂X₆) via halogen bridges — 3-center-4-electron bonds — because 3p/4p/5p π-overlap with halogens is too poor to quench the Lewis acidity internally.
- Lewis acid strength toward a common base runs BBr₃ > BCl₃ > BF₃ for boron (reduced π-donation from F despite higher electronegativity) but AlCl₃ > BCl�3 down the group at fixed halogen (larger M, weaker M–X π-bonding, more available orbital space).
- Adduct formation converts trigonal-planar (or dimeric, edge-bridged tetrahedral) MX₃ into a tetrahedral, sp³ MX₃L unit; the dimer must first be cracked apart, costing bridge-bond energy that is repaid by the new dative bond.
- Fluoride affinity and gas-phase enthalpies of adduct formation with a reference base (Cl⁻, NMe₃, pyridine) are the quantitative currency for ranking these acids.
- Key idea: whether MX₃ dimerizes or stays monomeric, and how eagerly it grabs an external base, are both governed by the same competition — internal π-donation from halogen lone pairs into the empty p orbital on M versus that orbital's availability to an outside donor.
Where this lives today
Anhydrous AlCl₃ is still the workhorse Lewis acid catalyst for industrial Friedel–Crafts acylation and alkylation — used, for example, in the manufacture of anthraquinone dyes, detergent alkylbenzenes, and various pharmaceutical intermediates — precisely because its ready dimer-to-monomer-adduct interconversion lets it activate carbonyl and alkyl halide substrates on demand. BF₃, milder and more selective, is supplied industrially as its ether or amine adducts (BF₃·OEt₂, BF₃·NH₃) specifically because those adducts are bench-stable, easy to meter and handle, yet still release active BF₃ under reaction conditions — a direct, practical exploitation of the adduct-formation thermodynamics developed above. In polymer chemistry, alkylaluminum halides and related group 13 Lewis acids serve as co-catalysts/activators for Ziegler–Natta and metallocene olefin polymerization, where their job is to abstract a halide or alkyl group from a transition-metal precatalyst by exactly the same Lewis-acid logic — an empty p orbital on Al reaching for a lone pair or bonding electron pair it can pull away. Boron-based analogues, especially the strong, sterically encumbered Lewis acid tris(pentafluorophenyl)borane, B(C₆F₅)₃, are prized industrial and academic activators for exactly the same reason, engineered to be even more electron-poor at boron than BF₃ or BCl₃ by replacing halogens with strongly electron-withdrawing, poorly π-donating perfluoroaryl groups. Gallium and indium trihalides find use as precursors in metal-organic vapor-phase epitaxy (MOVPE) and related routes to III–V semiconductors (GaAs, InP, GaN-related materials), where their gas-phase monomer/dimer equilibria and their adduct chemistry with amine or phosphine ligands directly control precursor volatility, decomposition pathway, and the purity of the deposited semiconductor film. In each of these settings — dye synthesis, polymer catalysis, or crystal growth — the operative question is the same one this article has been building toward: how easily does this particular group 13 center's empty orbital get filled, and by what?
Ask the tutor: Why does GaCl₃ dimerize in the gas phase yet many Ga(III) compounds in aqueous solution behave as simple hydrated ionic Ga³⁺ centers instead — what changes between the two environments?
The physical picture
Every group 13 trihalide MX₃ starts from the same electron count: the central atom contributes three valence electrons, each halogen contributes one to a σ bond, giving three M–X σ bonds and leaving one empty p orbital perpendicular to the MX₃ plane. VSEPR treats the central atom as having three σ-bonding domains and no lone pair, so the idealized geometry is trigonal planar, D₃ₕ, with X–M–X angles of 120°. This is a six-valence-electron center — an isolated MX₃ molecule is not an octet species, and that empty p_z orbital is the reactive site for everything that follows. The first way a trihalide can address this deficiency is intramolecular: a filled p orbital (or lone pair) on each halogen can donate sideways into the empty p_z orbital on M, forming a partial π bond delocalized over the M(p_z)–3×X(p_π) system. This works best when the p orbitals involved are similar in size and diffuseness, which is exactly the condition satisfied by B (2p) and F, Cl, Br, or even I to a lesser degree, all of whose relevant lone-pair orbitals are reasonably compact for period 2–4 elements bonding to a period-2 center. The result: BX₃ molecules carry genuine partial π character in the B–X bonds, the B–X bond is shorter than a pure single bond would predict, and the molecule's Lewis acidity, while real, is muted because that empty orbital is already partly occupied by π donation. Moving down the group to Al, Ga, In, the central atom's valence p orbital is 3p, 4p, or 5p — much more diffuse and higher in energy than the 2p orbitals on the halogens. The size and energy mismatch kills efficient π overlap: Al–Cl π bonding is weak, so the aluminum p_z orbital stays largely empty and hungry. Two such hungry AlCl₃ units satisfy each other instead: a chlorine on one AlCl₃ swings a lone pair into the empty orbital of the neighboring AlCl₃, and symmetry (each Al needs the favor returned) produces a chlorine-bridged dimer, Al₂Cl₆, with two 3-center-4-electron Al–Cl–Al bridges. Each aluminum ends up four-coordinate and (pseudo-)tetrahedral, and the whole molecule adopts a D₂ₕ, edge-shared bitetrahedral structure — two AlCl₄ tetrahedra sharing an edge (two bridging Cl atoms), with four terminal chlorines completing the outer positions.
Fluorine is the interesting exception on the boron side: F is the smallest, most electronegative halogen, and its 2p lone pairs are held tightly, yet BF₃ has the weakest Lewis acidity of the boron trihalides toward almost every common base, not the strongest as naive electronegativity arguments would suggest. The resolution is that fluorine's compact, energetically low 2p orbitals overlap outstandingly well with boron's 2p_z orbital, so BF₃ enjoys the strongest B–X π-donation of the series, leaving the least residual electrophilicity at boron. Going BF₃ → BCl₃ → BBr₃ → BI₃, the halogen p orbitals become larger and more diffuse, the size mismatch with boron's compact 2p grows, π-donation weakens, and the empty boron orbital becomes progressively more available to outside bases. Lewis acidity therefore increases in the order BF₃ < BCl₃ < BBr₃ (BI₃ often continues this trend, though steric bulk of iodine complicates direct comparison). When an external Lewis base L (ether, amine, phosphine, halide ion) is introduced to a dimer like Al₂Cl₆, it doesn't need to compete with the bridging chlorines forever — it attacks a terminal aluminum, and the strain of adding a fifth coordinate group tips the balance toward breaking a bridge bond, converting Al₂Cl₆ + 2L into two molecules of the tetrahedral 1:1 adduct AlCl₃L. The reaction is thermodynamically favorable whenever L is a better donor than a bridging chloride ion is (nearly always true for amines, ethers, and phosphines), because a genuine dative M←L bond from a good σ-donor is stronger than the compromise 3-center-4-electron bridge bond it replaces.
The core ideas
The energetics of dimerization and adduct formation are best organized around two linked equilibria and the thermodynamic quantities that control them.
where: ΔH_dim is (negative of) the strength of two 3-center-4-electron bridge bonds formed, and ΔS_dim is necessarily negative (two molecules become one), so dimerization is favored at low T and disfavored as T rises — exactly why AlCl₃ vapor shifts from mostly Al₂Cl₆ toward monomeric AlCl₃ as temperature increases toward ~400 °C and above.
where: D(M←L) is the dative bond dissociation energy of the newly formed adduct bond; larger (more negative ΔH_ads) means a stronger Lewis acid–base pairing. Because forming an adduct from a dimeric trihalide first requires breaking a bridge, the overall reaction enthalpy for that case is a sum:
where: ΔH_bridge-break is the (positive, endothermic) energy to convert half of M₂X₆ into monomeric MX₃, and the factor 1/2 reflects that one dimer yields two monomers, each of which then binds one L. Adduct formation is observed whenever ΔH_overall is sufficiently negative — i.e., whenever the new dative bond more than compensates for the bridge energy lost. Quantitatively, Lewis acidity is compared using a fixed reference base so that only the acid varies. Two standard scales:
where: a larger FIA means MX₃ binds F⁻ more strongly and is a stronger Lewis acid on this particular scale; BF₃ has an unusually high FIA (it is an excellent fluoride acceptor, forming very stable BF₄⁻) even though it is the weakest acid of the boron trihalides toward amine/ether donors — a reminder that Lewis acidity rankings are base-dependent, not absolute.
where: comparing K_eq for a series of MX₃ against one fixed base L (commonly NMe₃ or pyridine, measured calorimetrically or by NMR/IR shift methods) gives the practical, base-specific Lewis acidity order used in most tables (e.g., the Gutmann–Beckett or Childs methods).
Goal. Show, from orbital size and symmetry, why aluminum's empty p orbital remains chemically "hungry" (favoring dimerization) while boron's analogous orbital is substantially quenched by π-donation from chlorine (favoring the monomer).
Step 1 — Identify the acceptor orbital. In both BCl₃ and AlCl₃, the central atom is sp²-hybridized, using three sp² hybrids for the three σ M–Cl bonds and leaving one pure p orbital (p_z, perpendicular to the MCl₃ plane) empty. This p_z orbital is the Lewis-acidic site: it is unoccupied, σ-non-bonding, and available for any filled orbital of correct symmetry (π or lone-pair donor) to overlap with it.
Step 2 — Compare orbital sizes for the π-donation pathway. Each chlorine has filled 3p lone pairs in its own valence shell; one lone pair per Cl (perpendicular to the Cl–M axis, in the molecular plane... more precisely, out-of-plane π symmetry) can, in principle, donate into M's empty p_z. Boron's acceptor orbital is 2p — compact, with a radial extent comparable in scale to a chlorine 3p orbital, giving reasonable (though not perfect) overlap. Aluminum's acceptor orbital is 3p — considerably larger and more diffuse than boron's 2p, and critically, poorly size-matched to chlorine's 3p donor orbital because overlap integrals fall off sharply when orbital radial nodes and extents differ substantially between the two centers even though the principal quantum numbers happen to coincide (3p–3p): symmetry match is necessary but not sufficient — energetic and radial matching matter as much, and the highly polarized, ionic-leaning Al–Cl bond leaves the chlorine lone pairs poorly disposed to donate efficiently into the more diffuse, higher-energy Al 3p orbital.
Step 3 — Quantify with structural evidence. Experimentally, the B–Cl bond length in BCl₃ (1.75 Å) is measurably shorter than the sum of covalent radii would predict for a pure single bond, consistent with partial π character reinforcing the σ framework. The Al–Cl terminal bond length in Al₂Cl₆ (≈2.06 Å) is close to a normal single-bond value, and no comparable shortening evidence exists for hypothetical monomeric AlCl₃'s Al–Cl bonds — meaning the aluminum p_z orbital in the monomer would remain largely empty and highly electrophilic if the molecule stayed monomeric.
Step 4 — Let a second molecule solve the problem. Because internal π-donation cannot quench Al's electron deficiency, a second AlCl₃ unit — itself equally electron-deficient but possessing chlorine lone pairs to spare — steps in as an external Lewis base. One chlorine lone pair on molecule 2 donates into the empty p_z orbital of molecule 1's aluminum, and symmetry demands the reverse also happens (each Al is equally hungry), producing two bridging Al–Cl–Al 3-center-4-electron units. Each Al becomes 4-coordinate, each pair of bridging Cl atoms contributes one lone pair to each of the two adjacent Al centers, and both aluminum centers achieve a fully occupied valence shell (formally 8 electrons around each Al, counting the shared bridge pairs once for each center it stabilizes).
Step 5 — Contrast with boron. Because BCl₃'s p_z orbital is already substantially engaged by intramolecular π-donation from its own three chlorines, there is much less driving force for a second BCl₃ molecule to act as an external base; the residual electrophilicity at boron is real but modest, and BCl₃ remains monomeric in the gas phase and liquid, only dimerizing under extreme conditions or not at all under ordinary ones. This is the structural fork: efficient self-π-donation (small, period-2 M) → monomer; inefficient self-π-donation (larger, period ≥3 M) → dimer via halogen bridging.
Go deeper · Derivation B · 2 minEnergetics of adduct formation: why L "wins" against a bridging chloride
Goal. Derive, in enthalpy terms, the condition under which an external Lewis base L successfully cracks open the Al₂Cl₆ dimer to give two AlCl₃L adducts, and show the numbers are consistent with observation for L = a typical ether or amine.
Step 1 — Write the overall reaction as a thermochemical cycle. The transformation of interest is:
Break this into two conceptual steps: (i) dissociate the dimer into two monomeric AlCl₃ fragments, costing the dimerization enthalpy in reverse, and (ii) let each monomeric AlCl₃ bind one L, releasing the dative bond energy.
Step 2 — Assign approximate numbers. The gas-phase dimerization enthalpy of Al₂Cl₆ (i.e., 2 AlCl₃(g) → Al₂Cl₆(g)) is experimentally around −126 kJ/mol (values in different compilations range roughly −100 to −130 kJ/mol depending on method; we use −126 kJ/mol here as a representative literature figure), so breaking the dimer back into two monomers costs +126 kJ/mol.
Step 3 — Assign the adduct bond strength. Calorimetric and computational estimates for a reasonably strong O- or N-donor binding to AlCl₃ (e.g., diethyl ether or trimethylamine) give a dative bond enthalpy on the order of −80 to −110 kJ/mol per AlCl₃···L bond; take a representative value of −95 kJ/mol per bond, and remember there are two such bonds formed (one per monomer):
Step 4 — Sum the cycle.
Step 5 — Interpret. A net exothermicity of roughly −64 kJ/mol per mole of Al₂Cl₆ consumed (i.e., about −32 kJ/mol per AlCl₃L formed, if you prefer per-monomer bookkeeping — though the −64 kJ/mol figure already refers to the whole dimer-to-two-adducts reaction) confirms that a moderately good donor easily outcompetes a bridging chloride, explaining why donor solvents (THF, Et₂O) and amines reliably de-aggregate Al₂Cl₆ into monomeric 1:1 adducts. The qualitative lesson generalizes: any base whose Al←L bond is stronger than half the dimer's bridging bond energy (here, half of 126, i.e. > 63 kJ/mol) will spontaneously crack the dimer. Weak donors (e.g., very bulky or poorly basic species) may fail this test and leave Al₂Cl₆ substantially intact or only partially adducted, which is exactly what is seen with sterically hindered or very weakly basic ligands.
Limiting case check. If L were an extremely weak base with a dative bond enthalpy of only, say, −40 kJ/mol per AlCl₃···L bond, step (ii) would contribute only 2 × (−40) = −80 kJ/mol, giving ΔH_overall = +126 − 80 = +46 kJ/mol — an endothermic, unfavorable process, correctly predicting that Al₂Cl₆ would not be cracked open by such a weak donor and would remain dimeric in its presence, consistent with the observation that Al₂Cl₆ persists in noncoordinating or weakly coordinating solvents (e.g., aromatic hydrocarbons) while it dissolves as monomeric adducts in ethers and amines.
Worked examples
Example 1 (foundational): Electron counting and geometry prediction for GaCl₃. Gallium contributes 3 valence electrons, each of three chlorines contributes 1 electron to a σ bond, giving 3 bonding pairs and 0 lone pairs on Ga in the free monomer — VSEPR class AX₃E₀, predicting trigonal planar, 120° angles, D₃ₕ symmetry, exactly analogous to BCl₃. Because gallium's valence acceptor orbital is 4p (even more diffuse than aluminum's 3p), self π-donation from chlorine is negligible, and by the same logic developed above, GaCl₃ dimerizes to Ga₂Cl₆ in the vapor and solid, isostructural with Al₂Cl₆ (two GaCl₄ tetrahedra sharing an edge). Checking the limiting case: if gallium's acceptor orbital were as compact as boron's (it is not, but hypothetically), we would predict a monomer instead — confirming that orbital diffuseness, not merely "group 13-ness," is what drives the dimerization criterion.
Example 2 (applied): Ranking Lewis acidity for a Friedel–Crafts catalyst choice. A chemist needs to choose between AlCl₃ and BCl₃ as a catalyst for a Friedel–Crafts acylation that requires strong activation of an acyl chloride's carbonyl oxygen (a hard, moderately basic donor site). Toward this kind of hard O-donor substrate, AlCl₃ is empirically and mechanistically the stronger, more effective acid: aluminum's larger, more diffuse, more available 3p acceptor orbital (never fully quenched by weak Al–Cl π-donation, and further activated once the Al₂Cl₆ dimer is cracked by the substrate acting as the incoming base) binds the carbonyl oxygen more strongly than boron's smaller, partially π-satisfied 2p orbital does. This matches real industrial practice: AlCl₃ is the classic stoichiometric (often superstoichiometric) Friedel–Crafts catalyst, while BCl₃ is reserved for more specialized syntheses (e.g., some directed borylations or where milder, more selective activation and easier product separation are wanted). As a limiting-case sanity check: were the substrate instead F⁻ itself (the base used in the fluoride-ion-affinity scale), the ranking would partly invert for the boron series specifically (BF₃ becomes a very strong F⁻ acceptor because Al–F versus B–F bond-strength and lattice/solvation terms shift the comparison) — illustrating again that "Lewis acidity" is not one universal number but a base-dependent quantity.
Example 3 (exam/competition-level): Predicting and rationalizing an unusual coordination number. A gas-phase electron diffraction study finds that InCl₃ vapor, unlike AlCl₃ or GaCl₃, exists to a significant extent as a monomer at moderate temperature, with only partial dimerization even below 300 °C, and its solid-state structure is a layered (YCl₃-type) lattice with six-coordinate In rather than a molecular M₂X₆ dimer. Explain, from periodicity, why indium's behavior diverges from aluminum's and gallium's, and predict what happens to TlCl₃. Reasoning: Indium's 5p valence orbital is even larger and more diffuse than gallium's 4p, so in principle the "gas-phase monomer is very electron-deficient, wants to dimerize" argument should apply even more strongly — but a competing effect now dominates: indium's greater atomic radius allows it to achieve a higher coordination number without excessive steric strain, favoring extended ionic/covalent lattices with octahedral In in the solid rather than discrete molecular dimers, while in the gas phase (where lattice stabilization is unavailable) some monomer survives because the driving force to dimerize specifically (rather than form larger aggregates or exist as isolated ionic-character monomers) is weaker for the larger, more electropositive In center — indium chemistry begins to shade toward the ionic, higher-coordinate behavior typical of a metal halide MX₃ salt rather than a discrete molecular Lewis acid. Thallium, further down and further destabilized in the +3 state by the inert-pair effect, continues this trend even more strongly: TlCl₃ is thermodynamically much less stable relative to TlCl (the inert 6s² pair resists oxidation/bonding participation), TlCl₃ is a comparatively poor, less commonly encountered Lewis acid framework, and Tl(III) halide chemistry is dominated by complex ions and instability toward reduction to Tl(I) rather than by clean, BX₃/AlX₃-style molecular or dimeric Lewis-acidic trihalide behavior. The limiting-case check: as one imagines M becoming arbitrarily large and electropositive (extrapolating past Tl toward a fully "ionic" limit), the entire covalent-dimer-versus-monomer framework breaks down and the compound is better described as an ionic solid MX₃ with a coordination number set by radius-ratio rules, not by molecular Lewis acid–base bonding at all — exactly the boundary this comparison is designed to expose.
Limiting cases, scope & common traps
| Situation | Expected behavior | Why |
|---|---|---|
| BX₃, X = F, Cl, Br (gas/solution) | Monomeric, trigonal planar | Efficient B(2p)–X(np) π-donation quenches Lewis acidity internally |
| AlX₃, GaX₃, X = Cl, Br, I (gas/solid, moderate T) | Dimeric M₂X₆, edge-shared tetrahedra | Poor M(np, n≥3)–X π-overlap leaves empty p orbital "hungry"; self-adduct forms |
| AlF₃ | Extended ionic lattice, not molecular dimer | F⁻ is small/hard/ionic-bonding-favoring; high lattice energy wins over discrete molecular structure |
| InCl₃, TlCl₃ | Partial/no dimerization; higher coordination numbers or lattice structures; Tl(III) thermodynamically fragile | Larger, more electropositive/inert-pair-affected centers favor ionic or higher-CN solid-state structures over discrete Lewis-acidic molecules |
| MX₃ + strong donor L (ether, amine, phosphine) | Cracks dimer (if present); forms tetrahedral 1:1 MX₃L adduct | New dative bond stronger than bridge-bond energy it replaces |
| MX₃ + very weak/bulky donor | Dimer may persist, adduct forms only partially or not at all | Dative bond too weak to outcompete existing bridge/self-association energy |
| High temperature, gas phase | Dimer shifts toward monomer | ΔS_dim unfavorable for association; entropy wins as T rises |
The chemist behind it
The structural puzzle of aluminum chloride's vapor density anomaly was worked out over the first decades of the twentieth century as physical chemists began applying vapor-density and, later, electron-diffraction methods systematically to inorganic halides — but the person whose name is most tightly bound to the systematic study of these trihalides' Lewis acid–base chemistry, and to the language ("Lewis acid," "Lewis base," "coordinate covalent bond") used throughout this whole area, is Gilbert N. Lewis.
Origins · 1 minGilbert Newton Lewis and how this was discovered
Lewis reframed acid–base chemistry away from proton transfer and toward electron-pair donation and acceptance, publishing the generalized definition in the 1920s (fully articulated in his 1923 book Valence and the Structure of Atoms and Molecules). The framework gave chemists a language for exactly the phenomenon central to this article: an "acid" like BCl₃ or AlCl₃ is simply any species with an accessible empty orbital capable of accepting an electron pair, and a "base" is any species with a lone pair to donate — the underlying idea that unifies the dimerization of Al₂Cl₆ (a molecule acting as its own base) with its adduct chemistry toward external donors (ammonia, ethers, halide ions). Lewis never personally resolved the Al₂Cl₆ structure — that came from later electron-diffraction and X-ray work — but the conceptual vocabulary in which every sentence of this article is written is his.
Why does Al₂Cl₆ exist as a dimer but not as a trimer or higher oligomer?
A dimer is the smallest aggregate that lets every aluminum reach four-coordination and every bridging chlorine donate one lone pair to each of its two neighboring aluminums — the electron-deficiency problem is exactly and completely solved with two bridges. Going to a trimer or chain would require either leaving some Al under-coordinated or introducing additional bridging chlorines that provide no further stabilization once each Al is already tetrahedral and each Cl already has its lone pairs appropriately distributed. Geometrically, the edge-shared bitetrahedral D₂ₕ dimer is also a comfortable, low-strain arrangement, whereas larger rings or chains would introduce unfavorable angle strain without any compensating bonding gain. This is analogous to why many p-block halide-bridged species (e.g., some boron hydride or aluminum alkyl dimers) stop at the dimer rather than oligomerizing further.
Does BCl₃ ever dimerize under any conditions?
Essentially no — BCl₃ is monomeric as a gas, liquid, and even in the solid state, because boron's compact 2p acceptor orbital is already substantially satisfied by intramolecular π-donation from its own three chlorines, leaving very little residual driving force for a second BCl₃ to act as an external base. This is a genuine structural contrast with AlCl₃, GaCl₃, and InCl₃(partially), and it is one of the cleanest diagnostic differences between period-2 and heavier group 13 trihalide chemistry.
How does adduct formation change the coordination geometry at the group 13 center?
The central atom goes from three-coordinate, sp²-hybridized, trigonal planar (in the free monomer, or effectively four-coordinate pseudo-tetrahedral if starting from a dimer) to four-coordinate, sp³-hybridized, (distorted) tetrahedral once the new dative bond to the incoming Lewis base is formed. Bond angles around M relax from the ideal 120° (planar) toward angles closer to 109.5° (tetrahedral), and the M–X bonds typically lengthen slightly because M's electron density is now shared with an additional ligand, weakening each existing M–X σ bond marginally.
Why is AlCl₃ such an effective Friedel–Crafts catalyst compared to, say, AlF₃?
AlCl₃ is molecular, readily sublimes, dissolves (as adducts) in organic reaction media, and its dimer is easily cracked open by the incoming acyl chloride or alkyl halide substrate to generate a reactive, strongly electrophilic monomeric AlCl₃-substrate adduct in situ. AlF₃, by contrast, is an extremely stable, high-lattice-energy ionic solid (fluoride being small, hard, and forming very strong Al–F ionic/covalent bonds throughout an extended lattice) that is essentially insoluble and unreactive under normal conditions — its Lewis acidity is real in principle but practically inaccessible because breaking apart the lattice costs far more energy than any organic substrate's coordination could repay.
Is Ga₂Cl₆ isostructural with Al₂Cl₆, and does gallium show any distinctive chemistry here?
Yes — Ga₂Cl₆ adopts essentially the same edge-shared, chlorine-bridged bitetrahedral structure as Al₂Cl₆, for the same orbital-size reasoning (gallium's 4p acceptor orbital is even more diffuse than aluminum's 3p, so π-donation from chlorine is likewise inefficient and dimerization via bridging is favored). One distinctive wrinkle in gallium (and, more so, indium) chemistry is the increasing viability of the +1 oxidation state alongside +3, driven by a weak but real inert-pair-type effect building up before it becomes dominant at thallium; this occasionally produces mixed-valence or Ga(I)/Ga(III) species that have no boron or aluminum analogue.
How do chemists experimentally distinguish a monomer from a dimer for these trihalides?
Gas-phase molecular weight (historically by vapor density, now more precisely by mass spectrometry) directly reveals whether the effective molar mass matches MX₃ or M₂X₆. Electron diffraction and X-ray crystallography give the actual bond lengths and angles, distinguishing trigonal-planar monomers (all M–X bonds equivalent, 120° angles) from bridged dimers (two distinct M–X bond lengths — shorter terminal, longer bridging — and characteristic bridging M–X–M angles well below 120°). Vibrational spectroscopy (IR/Raman) also distinguishes the two because the symmetry-allowed vibrational modes differ sharply between D₃ₕ monomers and D₂ₕ dimers.
What happens to Lewis acidity trends when the halogen is fixed but the base is varied instead of the metal?
Base identity can reorder acid rankings because different bases probe different aspects of the acid's electronic structure (hard vs. soft character, steric demand, σ- vs. π-donor ability) — this is the essence of the Hard–Soft Acid–Base framework layered on top of the simple π-donation picture developed here. A given trihalide might rank as a strong acid toward a hard, small O-donor like F⁻ or Et₂O but rank differently toward a soft, bulky P-donor like PPh₃, because steric clash with the trigonal MX₃ plane and orbital hardness/softness matching both shift depending on which base is chosen — which is precisely why Lewis acidity tables always specify the reference base used.