Heck Reaction

A four-stop loop around one palladium atom

Carbon does not like to bond to carbon on demand. Two unreactive carbon centers — say an aromatic ring carbon and the carbon of a double bond — will sit next to each other indefinitely and do nothing. The Heck reaction solves this by sending a single palladium atom around a loop that grabs one carbon, escorts it to the other, joins them, and then leaves to do it again. One palladium can run this loop tens of thousands of times.

The overall transformation is simple to write. An aryl halide and an alkene go in; a new aryl-substituted alkene and a molecule of HX come out, the HX neutralized by an added base:

Ar–X  +  CH₂=CH–R   ──[Pd(0), base]──>   Ar–CH=CH–R  +  base·HX

   (aryl halide)  (alkene)                (E)-product   (salt)

The double bond is not consumed — it reappears in the product, just shifted and now carrying the aryl group. That is the signature of a Heck reaction and the single fact that distinguishes it from every other palladium coupling: you keep the alkene. The four elementary steps that make this happen are oxidative addition, alkene coordination/insertion, β-hydride elimination, and reductive elimination of HX.

The mechanism, step by step

Walk the loop once. The active species is a coordinatively unsaturated, electron-rich palladium(0) complex, written here as L₂Pd(0) where L is a phosphine or other ligand.

1. OXIDATIVE ADDITION
   L₂Pd(0)  +  Ar–X   ──>   Ar–Pd(II)–X·L₂
   Pd inserts into the C–X bond. Pd goes 0 → +II.
   Aryl–Pd σ-bond forms; halide stays on Pd.

2. ALKENE COORDINATION + MIGRATORY INSERTION (syn)
   Ar–Pd(II)–X  +  CH₂=CHR  ──>  Ar–CH₂–CH(R)–Pd(II)–X
   The alkene π-binds Pd, then Ar and Pd add ACROSS
   the same face (syn). New Ar–C bond is made here.

3. β-HYDRIDE ELIMINATION (syn-coplanar)
   Ar–CH₂–CH(R)–Pd–X  ──>  Ar–CH=CH–R  +  H–Pd(II)–X
   Pd plucks a β-hydrogen syn to itself; the product
   alkene detaches. This step sets E/Z geometry.

4. REDUCTIVE ELIMINATION OF HX (base-assisted)
   H–Pd(II)–X  +  base   ──>  L₂Pd(0)  +  base·HX
   Base strips off HX, Pd returns to 0, ready again.

Two subtleties carry the whole reaction. First, step 2 is syn: the aryl group and palladium land on the same face of the former double bond, which fixes the relative stereochemistry of the alkyl–Pd intermediate. Second, step 3 requires the Pd–C–C–H array to rotate into a syn-coplanar (eclipsed) geometry before the hydride can transfer. Because rotation is free, the molecule selects the rotamer that holds its two bulkiest groups apart — and elimination from that conformer pours out the trans, or (E), alkene almost exclusively.

There is one more consequence of needing a syn β-hydrogen: palladium can only eliminate toward a hydrogen it can reach on the same face. When the carbon that bears the new aryl group has no other syn β-H — as in the Ar–CH₂– carbon after insertion of a monosubstituted alkene — Pd is forced to eliminate back toward the carbon it just left, reforming the alkene in the position that regenerates conjugation with the ring. This regiochemical "ratchet" is why styrene-type products dominate.

Linear vs branched: which carbon gets the aryl group

When the alkene is electron-poor or conjugated (acrylates, acrylonitrile, styrene), insertion is governed by electronics and sterics that funnel the aryl group onto the terminal CH₂ carbon, giving the linear (β-arylation) product — the familiar (E)-cinnamate or (E)-stilbene type. When the alkene is electron-rich (enol ethers, enamides) and the reaction runs under cationic conditions (silver or thallium additives, or triflate leaving groups), the polarization flips and the aryl group migrates to the internal carbon, giving the branched (α-arylation) product.

Neutral pathway (X = Br, I; electron-poor alkene)
   Ar–Pd–X  +  CH₂=CH–CO₂R   →   Ar–CH=CH–CO₂R   (LINEAR, β)

Cationic pathway (X = OTf, or Ag⁺ removes halide; enol ether)
   [Ar–Pd]⁺  +  CH₂=CH–OR'    →   Ar–C(=CH₂)–OR' type (BRANCHED, α)

The control knob is whether the halide stays on palladium during insertion. A neutral, halide-bound Pd inserts through a less-polarized transition state and prefers the linear product. A cationic, halide-free Pd inserts through a more-polarized transition state where the partial positive charge wants to land on the more-substituted carbon, steering branch selectivity. Chemists choose triflates and silver salts precisely to dial this in — the Overman group built an entire body of asymmetric Heck chemistry on the cationic pathway.

Catalysts, ligands, and reaction conditions

A working Heck reaction is a recipe of five ingredients: a Pd source, a ligand, a base, a solvent, and heat. Typical numbers:

• Pd source. Pd(OAc)₂ (reduced in situ to Pd(0)) or Pd₂(dba)₃; loadings of 0.5–5 mol % in classic runs, dropping below 0.01 mol % with the best ligand systems. Turnover numbers above 10⁶ have been reported with palladacycle catalysts.
• Ligand. Triphenylphosphine PPh₃ in early work; bulky, electron-rich phosphines (P(t-Bu)₃, the Buchwald biaryl ligands) for sluggish aryl chlorides; N-heterocyclic carbenes (NHCs) for the most demanding substrates. Ligand-free "Jeffery conditions" use a phase-transfer salt (Bu₄NCl) instead.
• Base. One full equivalent or more — Et₃N, K₂CO₃, NaOAc, Cs₂CO₃, or K₃PO₄. The base is not catalytic; HX is a stoichiometric product.
• Solvent. DMF, NMP, MeCN, dioxane, or water with surfactants. Polar aprotic solvents stabilize the cationic intermediates.
• Temperature. 80–140 °C classically; microwave and ligand-accelerated variants run near room temperature.

Reactivity of the halide tracks the strength of the C–X bond, because oxidative addition must break it. The bond dissociation energies fall in the order C–I < C–Br < C–Cl, and so does the ease of oxidative addition:

Oxidative addition rate:   Ar–I  >  Ar–OTf  ≈  Ar–Br  >>  Ar–Cl

Aryl C–X bond strength (approx.):
   C–I   ≈ 270 kJ/mol      (easiest to add)
   C–Br  ≈ 336 kJ/mol
   C–Cl  ≈ 397 kJ/mol      (needs bulky electron-rich ligand)

Aryl chlorides are cheap and abundant but their strong C–Cl bond makes oxidative addition slow; cracking them open economically — using P(t-Bu)₃ or NHC ligands that pump electron density onto Pd — was a major advance of the late 1990s and is why chloride Hecks are now industrially viable.

Heck vs the other palladium couplings

	Heck	Suzuki	Negishi	Sonogashira
Carbon partner #2	Alkene (C=C)	Boronic acid R–B(OH)₂	Organozinc R–ZnX	Terminal alkyne
Key non-OA step	Migratory insertion	Transmetalation	Transmetalation	Transmetalation (Cu)
Releasing step	β-hydride elimination	Reductive elimination	Reductive elimination	Reductive elimination
Bond formed	C(sp²)–C(sp²)	C–C	C–C	C(sp²)–C(sp)
Product keeps a C=C?	Yes — that's the point	No	No	No (gives an alkyne)
Second metal needed?	None	Boron	Zinc	Copper co-catalyst
Stereochemistry issue	E/Z of new alkene	Retention at carbon	Retention at carbon	—
Byproduct	HX + base salt	Borate salt	Zinc halide	Amine·HX

The structural giveaway is the second column: only the Heck reaction lacks a pre-formed organometallic partner and only the Heck reaction hands you a product with the double bond intact. Suzuki, Negishi, and Sonogashira all run a transmetalation followed by reductive elimination; the Heck runs an insertion followed by β-hydride elimination. Same opening move, completely different finish.

Worked example: ethyl (E)-cinnamate from iodobenzene

The textbook demonstration couples iodobenzene with ethyl acrylate:

C₆H₅–I  +  CH₂=CH–CO₂Et   ──[Pd(OAc)₂ 2 mol%, PPh₃, Et₃N, 100 °C, DMF]──>

   C₆H₅–CH=CH–CO₂Et  (E)   +   Et₃N·HI

   ethyl (E)-cinnamate, typically 85–95% isolated, >98:2 E:Z

Trace it through the cycle. (1) Pd(0) inserts into the C–I bond of iodobenzene to give PhPd(II)I. (2) Ethyl acrylate coordinates and inserts syn, placing the phenyl on the terminal CH₂ and Pd on the carbon α to the ester: Ph–CH₂–CH(CO₂Et)–PdI. (3) Pd has only one set of syn β-hydrogens available — on the benzylic CH₂ — so it eliminates back toward the phenyl carbon, regenerating a double bond now conjugated to both the ring and the ester. The eclipsed rotamer that keeps phenyl and ester anti gives the (E)-isomer. (4) Et₃N strips HI off H–PdI, and Pd(0) is back.

Why is the yield so clean? The product alkene is conjugated and electron-poor; it is a far worse ligand for Pd than the starting acrylate, so it doesn't re-insert. The reaction has a built-in off-ramp, and that thermodynamic trap is a big reason Heck couplings of acrylates are among the most reliable C–C bond formations in the toolbox.

Where the Heck reaction shows up

• Pharmaceuticals. An asymmetric Heck step installs a stereocenter in a noted enantioselective route to the anti-inflammatory naproxen, and Heck couplings appear in syntheses of the antiasthmatic montelukast (Singulair) and in building blocks for several antiviral drugs. The mild conditions tolerate the polar functional groups drug molecules carry.
• Organic electronics. Repeated Heck couplings polymerize aryl dihalides with divinyl monomers into poly(arylene–vinylene)s such as the green-emitting backbone chemistry related to PPV, used in OLED displays and printable electronics.
• Total synthesis. The intramolecular and asymmetric Heck reactions build quaternary stereocenters and fused rings that are hard to reach any other way — Overman's syntheses of complex alkaloids lean on cationic asymmetric Heck cyclizations.
• Agrochemicals and fragrances. Cinnamate and stilbene frameworks made by Heck arylation feed into UV-absorbers, sunscreens (octinoxate is an ethylhexyl cinnamate), and fine-chemical fragrances.

Common misconceptions and pitfalls

• "The Heck is a cross-coupling like Suzuki." It opens the same way (oxidative addition) but it has no transmetalation and no second organometallic. It is a carbopalladation/elimination, not a transmetalation/reductive-elimination. Lumping it with Suzuki obscures why the product keeps a double bond.
• "Insertion sets the E/Z geometry." No — insertion is syn and sets the alkyl–Pd stereochemistry, but the alkene geometry of the product is decided later, during β-hydride elimination, by which syn-coplanar rotamer eliminates. The E-preference comes from minimizing steric strain in that eclipsed transition state.
• "Any halide works." Simple sp³ alkyl halides with β-hydrogens fail because the alkyl–Pd intermediate eliminates before it can reach the alkene. Aryl, vinyl, and benzyl halides — and triflates — are the reliable partners. Aryl fluorides essentially never undergo oxidative addition.
• "Base is just a workup detail." The base is mechanistically essential — it removes HX from H–Pd–X to regenerate Pd(0). Run the reaction without enough base and the catalyst dies as acidic, halide-rich Pd species accumulate. You need at least one full equivalent because HX is a real product.
• "Oxygen doesn't matter." Pd(0) is air-sensitive; trace O₂ oxidizes it and erodes turnover. Most Heck reactions are run under inert atmosphere, although ligand-free aqueous protocols are more forgiving.
• "More ligand is always better." Excess phosphine can over-coordinate Pd and slow oxidative addition or insertion. The Jeffery ligand-free conditions exploit exactly this — sometimes no phosphine and a phase-transfer salt outperform a phosphine recipe.

Variants and refinements

• Intramolecular Heck. Tethering the halide and the alkene in one molecule forms rings and, crucially, quaternary carbon stereocenters. This is the engine of many ring-forming steps in total synthesis.
• Asymmetric Heck. Chiral ligands (BINAP, PHOX) on the cationic pathway create stereocenters with high enantiomeric excess; insertion into a prochiral alkene is the enantio-determining step.
• Oxidative (Fujiwara–Moritani) Heck. Replaces the aryl halide with a plain arene C–H bond, using a stoichiometric or O₂-coupled oxidant to keep Pd cycling. No halide leaving group is needed — a more atom-economical but harder-to-control cousin.
• Reductive Heck. When a reducing agent (formate, silane) is present, the alkyl–Pd intermediate is reduced instead of undergoing β-hydride elimination, so the double bond is not regenerated and a saturated product results — useful when you want the new C–C bond without a residual alkene.
• Ligand-free / heterogeneous. Pd nanoparticles, Pd-on-carbon, and palladacycle precatalysts run Hecks with extremely low leaching, important for removing trace Pd from pharmaceutical products (regulatory limits are often < 10 ppm).