Friedel-Crafts Acylation

Friedel–Crafts acylation followed by carbonyl reduction (Clemmensen or Wolff–Kishner conditions) remains the standard textbook workaround for the fact that direct alkylation is unreliable — this two-step "Acylation-Then-Reduction to Make Alkylbenzenes" strategy is exactly why acylation is taught immediately alongside, and as the fix for, alkylation's limitations. Industrially, acetophenone and related aryl methyl ketones are manufactured this way, and anthraquinone dyes are built via intramolecular or aromatic diacylation chemistry rooted in the same acylium mechanism. In pharmaceutical synthesis, Friedel–Crafts acylation is a frequent early step for installing an aryl ketone that will later be elaborated by reduction, Grignard addition, reductive amination, or Wittig olefination — a role it plays throughout Multistep Aromatic Synthesis Strategy planning, precisely because its skeletal fidelity (no rearrangement) makes retrosynthetic disconnection through an acyl group trustworthy in a way that disconnection through an alkyl halide is not. The intramolecular version — closing five- and six-membered rings onto an aromatic ring — remains a core method for building the tetralin, indanone, and fused-polycyclic frameworks that appear throughout steroid, terpenoid, and natural-product total synthesis, essentially unchanged in mechanism from Friedel and Crafts's original 19th-century observation.

Ask the tutor: Why does an intramolecular Friedel–Crafts acylation with a four-methylene tether favor a six-membered ring product over a seven-membered one, in terms of transition-state geometry?

Begin with the electrophile-generation step, because everything downstream follows from the structure of what AlCl₃ actually makes. An acyl chloride, R–C(=O)–Cl, has a chlorine bonded to a carbonyl carbon. Aluminium trichloride is a strong Lewis acid: aluminium has an empty p-orbital and only six valence electrons around it, so it eagerly accepts one of the lone pairs on chlorine. This gives an ion pair: R–C(=O)–Cl···AlCl₃ polarizes into $[\text{R–CO}]^{+}\,[\text{AlCl}_4]^{-}$, or in a tighter, more covalent picture, a chlorine-bridged complex that ionizes further as the reaction proceeds. Either way, the electrophilic carbon fragment that attacks the arene is the acylium ion, $\text{R–C}\equiv\text{O}^{+}$. The acylium ion deserves close orbital inspection because its stability is the whole story. Write its two resonance structures:

In the first structure the positive charge sits on carbon with a normal C=O double bond; in the second, one oxygen lone pair has been pushed fully into a second π-bond, giving a formal C≡O triple bond and placing the positive charge on the (more electronegative-tolerant, because it still has a lone pair and complete octet) oxygen. This second resonance form is a major contributor: an sp-hybridized, linear O=C=R⁺ arrangement is stabilized enormously by oxygen's lone pair donation, exactly analogous to the stabilization in a nitrilium ion or in the resonance structure of an amide, except here there is no competing resonance donor (no nitrogen lone pair fighting for the same orbital) — oxygen commits fully. The acylium ion is consequently much more stable, relative to a simple alkyl cation of comparable size, than one might first guess from "it's just a primary-carbon-like cation." This stabilization has an immediate and crucial mechanistic consequence: because the empty p-orbital on the acylium carbon is already filled by π-donation from oxygen, there is essentially no thermodynamic incentive for a 1,2-hydride or alkyl shift. In alkylation, a secondary or primary carbocation rearranges because a more stable (more substituted, or resonance/hyperconjugation-stabilized) cation is reachable one bond away. In acylation, the cation is already about as stabilized as it can be by the adjacent oxygen; any 1,2-shift would have to break the strong C≡O π-interaction to form a plain alkyl cation — a step that is uphill, not downhill. So the carbon skeleton R is delivered to the ring completely unrearranged. The electrophilic aromatic substitution step itself is textbook arenium-ion (Wheland intermediate) chemistry, run with the acylium ion instead of a simple carbocation or halogen electrophile. The arene π-system, viewed as an aromatic HOMO with significant electron density above and below the ring plane, attacks the empty p-orbital of the acylium carbon. This breaks aromaticity temporarily, generating a resonance-delocalized arenium (cyclohexadienyl) cation whose positive charge is shared over three ring positions (ortho, ortho, and para to the point of attack, when we track the σ-complex resonance structures for an unsubstituted ring). Loss of the ring's ipso proton, taken by a base (often the tetrachloroaluminate counterion, or another arene, or eventually water), restores aromaticity and delivers the neutral aryl ketone–AlCl₃ complex: the carbonyl oxygen's lone pair immediately coordinates to the newly regenerated, still-Lewis-acidic AlCl₃ (or to the AlCl₄⁻/ Cl–AlCl₃ system), because a ketone oxygen is itself a competent Lewis base. That final complexation is not a footnote — it dictates two of the reaction's defining practical features. First, since the product ketone sequesters the AlCl₃ as a 1:1 Lewis acid–base adduct, AlCl₃ is consumed stoichiometrically, not catalytically, in acylation (a real practical difference from many Friedel–Crafts alkylations, and one every synthesis student must remember: use ≥1 equivalent of AlCl₃, often 1.1–1.3 equiv, not a catalytic 5–10 mol %). Second, this same complexation withdraws electron density from the ring through the carbonyl, so even before the aqueous workup liberates free ketone, the ring bearing the developing C=O group is aggressively deactivated toward a second electrophilic attack. Both the resonance electron-withdrawal of a carbonyl group and the physical chelation of the remaining Lewis acid onto the newly formed carbonyl oxygen (which competes with, and usually outpaces, complexation to fresh acyl chloride) keep polyacylation from happening. This is why Friedel–Crafts acylation is the well-behaved cousin of alkylation, self-limiting in a way alkylation structurally cannot be.

The overall transformation is summarized by:

where:

• Ar–H is the arene (activated or mildly deactivated arenes work; strongly deactivated arenes like nitrobenzene do not react),
• R–COCl is the acyl chloride (or, equivalently, an acid anhydride (RCO)₂O with AlCl₃, or in intramolecular cases the acid itself with a stronger promoter like PPA or HF),
• the intermediate Ar–CO–R·AlCl₃ is the stoichiometric Lewis acid–base complex that must be hydrolyzed to release free ketone.

The rate law for the electrophilic step, exactly as in other EAS reactions, is second order overall:

where:

• $k$ is the rate constant for arenium-ion formation (the rate-determining step, since the subsequent deprotonation is fast),
• $[\text{R–CO}^{+}]$ is the (typically very low, pre-equilibrium) concentration of free or tightly ion-paired acylium ion generated from RCOCl + AlCl₃.

A key comparative relationship — why acylation resists rearrangement where alkylation does not — can be framed as a stability inequality. Define $\Delta G^{\ddagger}_{\text{shift}}$ as the barrier for a 1,2-hydride/alkyl shift starting from the electrophile-derived cation:

where:

• the acylium shift would have to sacrifice the ~C≡O-level π-stabilization (tens of kcal/mol worth of resonance and orbital overlap) to reach a plain, non-oxygen-stabilized carbocation,
• the alkyl-cation shift instead moves downhill, from a less-substituted to a more-substituted (more hyperconjugatively stabilized) cation.

Reading this inequality: because the barrier on the acylium side is so much higher (indeed the "shifted" product is higher in energy, not just harder to reach), no productive rearrangement pathway competes with direct aromatic attack. This is the single quantitative-flavor statement that explains the qualitative rule "acylium ions don't rearrange."

Example 1 (foundational): Acylation of toluene with acetyl chloride. Toluene (an activated arene, methyl is a weak σ-donor with hyperconjugative π-donation into the ring) is treated with acetyl chloride, CH₃COCl, and one equivalent of AlCl₃, followed by aqueous workup. Step 1. AlCl₃ coordinates the chlorine of CH₃COCl; heterolysis of the C–Cl bond generates the acetylium ion, $\text{CH}_3\text{–C}\equiv\text{O}^{+}$, paired with AlCl₄⁻. This ion is small (methyl group, no branching to worry about) and, per the mechanism above, cannot rearrange — there is no alternative skeleton for a 2-carbon acyl fragment to rearrange into anyway, making this the cleanest possible pedagogical case. Step 2. Regiochemistry is governed by the methyl group's directing effect. The methyl group is an ortho/para director (its filled C–H σ-bonds hyperconjugate into the ring π-system, raising HOMO coefficient density preferentially at ortho and para carbons). Because the incoming acylium electrophile is bulky (an sp-carbon flanked by a full methyl group and an oxygen, plus the developing steric bulk of the transition state resembling the tetrahedral-ish arenium geometry), ortho attack suffers steric clash with the existing methyl substituent. Friedel–Crafts acylations are therefore characteristically highly para-selective, more so than many alkylations, precisely because the acylium electrophile is comparatively large and linear. Step 3. The dominant product is 4′-methylacetophenone (p-methylacetophenone), i.e., 1-(4-methylphenyl)ethan-1-one, with only a minor amount of the ortho isomer and no meta product (meta positions are not stabilized in any resonance structure of the arenium ion for an o,p-director). Step 4. Aqueous acidic workup (dilute HCl or H₂SO₄) hydrolyzes the ketone·AlCl₃ complex: water's oxygen lone pair displaces the ketone oxygen from aluminium (an acid–base ligand exchange, Al being a harder, oxophilic Lewis acid that ultimately prefers hydroxide/water ligands upon aqueous quench), liberating free p-methylacetophenone and forming Al(OH)₃/AlCl₃ hydrolysis products. Check against alkylation: if this had instead been a Friedel–Crafts alkylation with, say, 1-chloropropane/AlCl₃, the initially formed n-propyl cation would rearrange (hydride shift) to the more stable isopropyl cation before ever reaching the ring, giving predominantly isopropylbenzene-type product, not n-propylbenzene. The acylation case shows no such rearrangement risk even in principle, because (as derived above) the acylium ion has no comparably-accessible more-stable rearranged form.

Example 2 (applied): Regiochemistry in acylation of a disubstituted arene — m-xylene. m-Xylene (1,3-dimethylbenzene) is acylated with propanoyl chloride/AlCl₃. Predict the major product and justify it mechanistically. Step 1. Number the ring: methyls at C1 and C3. Each methyl independently directs new substitution to its own ortho/para positions. C1's methyl directs to C2, C4 (ortho) and... note C6 is also ortho to C1, but check occupancy. C3's methyl directs to C2, C4 (ortho) and C6 (para to C3... let's be careful): for a substituent at C1, ortho = C2, C6; para = C4. For a substituent at C3, ortho = C2, C4; para = C6. Step 2. Overlay the two direction sets. Position C2 is ortho to both methyls — doubly activated, but also flanked on both sides by methyl groups, so it is sterically the most hindered site (a peri-like squeeze between two ortho substituents plus the incoming bulky acylium group). Position C4 is ortho to C3's methyl and para to C1's methyl — also doubly activated electronically, and only singly flanked (has one ortho neighbor, C3-methyl, and the other ortho neighbor, C5, is unsubstituted) — much less hindered. Position C6 is ortho to C1's methyl and para to C3's methyl — by the ring's symmetry this position is equivalent to C4 (m-xylene has a mirror plane through C2 and C5). Step 3. Because the acylium ion is a comparatively bulky, linear electrophile (recall Example 1's steric argument), the reaction strongly favors the electronically-activated but sterically-open C4/C6 position over the electronically-doubly-activated-but-sterically-congested C2 position. This is a textbook demonstration that in Friedel–Crafts acylation, steric accessibility frequently overrides a purely additive electronic argument — a subtlety that a naive "count the activating effects" approach would get wrong by predicting C2 as major. Step 4. The major product is therefore 1-(2,4-dimethylphenyl)propan-1-one (acylation at C4, equivalently labeled C6 by symmetry), i.e., 2,4-dimethylpropiophenone, with only a minor amount of the congested C2 (2,6-dimethyl-substituted) isomer. Sanity/limiting check. If the acyl group were replaced by a much smaller electrophile — formyl (as in a formylation, mechanistically distinct via Gattermann–Koch, but useful as a size-limiting thought experiment) — steric discrimination between C2 and C4 would shrink, and the doubly-activated C2 position would become competitively more important. This confirms that the C2/C4 preference genuinely tracks electrophile bulk, not merely some fixed, size-independent electronic rule.

Example 3 (exam/competition level): Intramolecular Friedel–Crafts acylation — a Haworth-type ring closure. 4-Phenylbutanoic acid, $\text{Ph–CH}_2\text{CH}_2\text{CH}_2\text{COOH}$, is treated with a strong promoter (e.g., polyphosphoric acid, PPA, or AlCl₃ after first converting the acid to its acyl chloride with SOCl₂). Predict the cyclization product and explain the ring-size selectivity. Step 1. Convert the carboxylic acid to the acyl chloride (SOCl₂, or directly activate with PPA/HF, which can ionize a carboxylic acid to an acylium-like species under superacidic conditions without isolating the chloride). Either way, the reactive electrophile ultimately generated is an intramolecular acylium ion tethered by a three-carbon chain to the aromatic ring: $\text{Ph–CH}_2\text{CH}_2\text{CH}_2\text{–C}\equiv\text{O}^{+}$. Step 2. Because the electrophile and the nucleophilic arene are now part of the same molecule, the reaction is an intramolecular Friedel–Crafts acylation, and the geometry of ring closure (which ring size forms) is dictated by which cyclization mode reaches a low-strain transition state. Counting atoms from the ipso ring carbon that will bond to the acylium carbon: ring-C(ipso, ortho position) – CH₂ – CH₂ – CH₂ – C(=O)⁺. Closing the bond between the acylium carbon and the ortho ring carbon forms a six-membered ring (the new C–C bond plus the three CH₂'s plus the carbonyl carbon plus the aromatic ipso carbon closes a six-membered ring fused to the benzene ring). Step 3. This six-membered-ring closure is strongly favored by Baldwin's-rules-type reasoning (6-exo-trig-like closure onto the sp²/sp-like acylium carbon) and by the low ring strain of a six-membered ring compared to a strained four- or five-membered alternative that a shorter or longer tether would give. The arenium-ion mechanism proceeds exactly as in the intermolecular case — ring π-attack on the acylium carbon, arenium ion delocalization, loss of the ortho-hydrogen to rearomatize — except now the new C–C bond is also a ring-closing bond. Step 4. The product is 1-tetralone (3,4-dihydronaphthalen-1(2H)-one), the benzo-fused six-membered cyclic ketone, formed in high yield. This exact transformation (with the parent 4-phenylbutanoic acid or its ring-substituted analogues) is a classic step in steroid and terpenoid total synthesis for building the fused six-membered rings of the steroid nucleus, and is mechanistically identical, arrow for arrow, to the intermolecular examples above — it merely uses the tether to guarantee both regiochemistry (only the ortho carbon is geometrically reachable) and molecularity (effectively unimolecular kinetics, hugely favoring cyclization over any competing intermolecular acylation of a second molecule). Limiting case. If the tether were shortened by one carbon (3-phenylpropanoic acid, giving a would-be 5-membered ring fused product, an indanone), the reaction still proceeds — 5-exo cyclizations are also kinetically favorable — but if lengthened to make an 8- or 9-membered ring, the entropic and transannular strain costs rise sharply, and cyclization yields fall off, with competing intermolecular (oligomerization/polymerization) pathways becoming significant. This ring-size dependence is a direct, testable consequence of transition-state geometry, not of any change in the acylium ion's intrinsic electronic behavior.

Charles Friedel (1832–1899), a French mineralogist-turned-organic-chemist, and James Mason Crafts (1839–1917), an American chemist trained partly in Europe, discovered the aluminium-chloride-catalyzed alkylation and acylation of arenes together in Paris in 1877, publishing jointly through the 1880s. Friedel had trained under Adolphe Wurtz and initially worked on mineralogy and silicon chemistry before turning to what became his most famous collaboration; Crafts, who had studied with Bunsen and Wurtz, brought a rigorous quantitative bent to their shared laboratory work at the École des Mines and later the Sorbonne. Their original 1877 report, on the AlCl₃-catalyzed reaction of amyl chloride with benzene, was a serendipitous discovery arising from an attempt to prepare organoaluminium compounds; noticing that HCl gas was evolved and an alkylbenzene formed, they quickly generalized the reaction to acyl halides, recognizing — correctly, well before any orbital-theoretic language existed to describe why — that the acyl version behaved far more cleanly, giving single, well-defined ketone products rather than the mixtures typical of the alkylation. The distinction they observed empirically (single clean product for acyl halides, messy mixtures for alkyl halides) is exactly the rearrangement/polysubstitution dichotomy explained mechanistically above, discovered by these two chemists roughly forty years before Christopher Ingold and others developed the electronic/curved-arrow language used to explain it.