Industrial Catalytic Cycles: Hydroformylation, Wacker, Metathesis, and Cross-Coupling

Hydroformylation remains one of the largest homogeneous catalytic processes on Earth by volume, producing butanal (from propene) that is hydrogenated to n-butanol and condensed (via aldol) to 2-ethylhexanol, the alcohol used to make the plasticizer bis(2-ethylhexyl) phthalate found in flexible PVC. Rh/triphenylphosphine and Rh/biphosphite (bulky, wide-bite-angle ligand) catalyst systems under the Union Carbide/Davy-Johnson-Matthey LP Oxo process route roughly 75% of world propene toward linear butanal with selectivities exceeding 95:5 linear:branched. The Wacker process still supplies a meaningful fraction of global acetaldehyde, although routes via ethanol oxidation compete regionally; more importantly, the mechanistic logic of Wacker-type nucleopalladation — alkene activation by Pd(II), external nucleophilic attack, β-hydride elimination — is the working template for a huge swath of modern Pd(II)-catalyzed C–O, C–N, and C–C bond-forming reactions used in fine-chemical and pharmaceutical synthesis, including Wacker-type cyclizations that build oxygen and nitrogen heterocycles found in natural product total synthesis. Olefin metathesis, after the mechanistic clarification by Chauvin and the practical, air-stable catalysts developed by Robert Grubbs (Ru-based) and Richard Schrock (Mo- and W-based alkylidenes), underpins ring-closing metathesis routes to macrocyclic pharmaceuticals (the hepatitis C drug boceprevir's precursor macrocycle and related HCV protease inhibitors are built by RCM), the industrial Shell higher olefins process and related routes to specialty α-olefins, and self-healing polymers and elastomers made by ring-opening metathesis polymerization of strained cyclic alkenes like norbornene and dicyclopentadiene. Palladium cross-coupling is arguably the single most-used carbon-carbon bond-forming method in modern pharmaceutical and agrochemical manufacturing: Suzuki couplings alone appear in the synthesis of blockbuster drugs including losartan (antihypertensive) and numerous kinase inhibitors, Negishi coupling is prized for functional-group tolerance with sensitive organozinc partners, and the Heck reaction remains central to industrial syntheses of naproxen precursors and various UV-filter and fragrance molecules. The same oxidative-addition/transmetalation/reductive-elimination skeleton, tuned with different metals and ligands, also underlies Buchwald-Hartwig C–N coupling for amine arylation, a reaction so central to medicinal chemistry that entire ligand families (SPhos, XPhos, RuPhos) were engineered specifically to accelerate its rate-limiting steps.

:::quiz[Why does the industrial Wacker process require a high concentration of chloride ion even though the rate law shows an inverse dependence on [Cl⁻]?]

• Chloride is a spectator ion with no mechanistic role
• Chloride is needed to keep palladium in solution as a soluble chloro complex and suppress Pd(0) black formation ✓
• Chloride directly attacks the coordinated alkene as the nucleophile
• Chloride oxidizes Pd(0) back to Pd(II), replacing the role of CuCl₂

:::quiz[In the metallacyclobutane mechanism of olefin metathesis, what distinguishes a "productive" retro-[2+2] ring-opening from a "non-productive" one?]

• Productive opening occurs only with Schrock-type catalysts, non-productive only with Fischer-type
• Productive opening breaks the same two bonds that were formed in the original [2+2], while non-productive opening breaks the other diagonal ✓
• Productive opening breaks the diagonal bonds different from those formed in cycloaddition, releasing a new alkene and new alkylidene, while non-productive opening reverses the original cycloaddition exactly
• Non-productive opening only happens above 100 °C

Ask the tutor: How does the choice between a Fischer-type and a Schrock-type metal carbene change which alkene substitution patterns a metathesis catalyst can tolerate?

All four processes share a structural logic: a coordinatively unsaturated metal center picks up a substrate, some bond reorganizes on the metal (insertion, oxidative addition, cycloaddition), and eventually a stable organic product leaves while the metal is regenerated in its original oxidation state and coordination environment, ready for another substrate. The catalytic cycle is drawn as a ring precisely because the resting state (or one of several resting states) reappears identically at top and bottom of the diagram.

Hydroformylation converts a terminal alkene, CO, and H₂ into an aldehyde with one more carbon than the starting alkene:

The Rh-based Wilkinson-type catalyst, HRh(CO)(PPh₃)₃, is a d⁸, 18-electron, trigonal-bipyramidal species that must first dissociate a phosphine to open a coordination site. The alkene binds, inserts into the Rh–H bond (this is Migratory Insertion into an M–H bond, the reverse of β-hydride elimination), CO binds and inserts into the new Rh–alkyl bond to give an acyl, oxidative addition of H₂ occurs at the resulting Rh(I) 16-electron acyl species to give an octahedral Rh(III) dihydride-acyl, and reductive elimination releases the aldehyde and regenerates the 16-electron Rh(I) hydride-CO fragment, which recombines with phosphine to close the loop.

The Wacker process oxidizes ethylene to acetaldehyde using PdCl₄²⁻ as the stoichiometric oxidant for the organic substrate and Pd(0) reoxidation by CuCl₂ (itself reoxidized by O₂) as the industrial trick that makes the whole thing catalytic in palladium:

Ethylene binds to Pd(II), an external water molecule attacks the coordinated alkene from the face opposite the metal (anti attack — this stereochemical fact, established by deuterium-labeling and stereochemical studies, rules out a syn-insertion "hydroxypalladation" mechanism in the classical picture, though modern mechanistic consensus for the industrial chloride-rich process leans toward anti attack while some ligand systems show syn pathways), giving a β-hydroxyethyl-Pd(II) intermediate. β-Hydride elimination generates a Pd–H bound to vinyl alcohol, which reinserts with the opposite regiochemistry (a formal isomerization) to place the oxygen on the carbon still bound to palladium, giving an oxygen-stabilized Pd-enolate. Reductive elimination (formally a β-hydride elimination from the alkoxide) releases acetaldehyde and Pd(0), which is reoxidized by two equivalents of Cu(II).

Olefin metathesis exchanges the alkylidene fragments of two alkenes:

A metal alkylidene (carbene) Ln M=CHR (see Metal Carbene and Carbyne Complexes) coordinates an alkene edge-on (Metal-Alkene and Metal-Alkyne Complexes) and undergoes a formal [2+2] cycloaddition to give a four-membered metallacyclobutane. This ring can open in either of two ways: back to starting materials (unproductive) or across the other diagonal to release a new alkene and a new metal alkylidene (productive). Iterating this over many alkene molecules statistically scrambles alkylidene fragments among all alkenes present, driving the reaction to whatever product distribution is thermodynamically (for cross metathesis) or entropically (for ring-closing metathesis, which expels a small volatile alkene like ethylene) most favorable.

Palladium cross-coupling reactions all share the skeleton oxidative addition → transmetalation (or insertion) → reductive elimination:

Pd(0)L₂ oxidatively adds the R–X bond (X = halide or pseudohalide) to give R–Pd(II)–X, a square-planar d⁸ 16-electron species. This complex then exchanges its halide for the organic group R′ carried by a main-group organometallic (boronic acid/ester in Suzuki, organozinc in Negishi, organostannane in Stille, organosilane in Hiyama), a step called transmetalation. The resulting R–Pd(II)–R′ complex undergoes reductive elimination to forge the R–R′ bond and regenerate Pd(0)L₂. The Heck reaction substitutes migratory insertion of an alkene into the Pd–R bond, followed by β-hydride elimination, for the transmetalation step, so it needs no organometallic coupling partner at all.

The regiochemistry of hydroformylation is controlled by which carbon of the alkene bonds to rhodium during migratory insertion, which sets up either a linear (n) primary alkyl or a branched (iso) secondary alkyl:

where:

• $k_n, k_{iso}$ are the effective rate constants for insertion giving the linear vs branched Rh-alkyl,
• $\theta$ is the Tolman cone angle of the phosphine (steric bulk favors the less hindered, linear-alkyl pathway),
• $\beta$ is the natural bite angle of a chelating diphosphine (wide bite angles like in Xantphos strongly favor linear product by forcing equatorial-equatorial coordination that sterically discriminates against secondary-carbon binding),
• $[\text{PPh}_3]$ is the excess free phosphine concentration, which shifts the pre-equilibrium toward the bulkier, more selective 18-electron resting state.

Reading: this is not a single-step rate law but a statement that ligand sterics and geometry — not electronics alone — dictate whether the alkene inserts to place Rh on the terminal or internal carbon, which is the single choice that determines linear versus branched product. For the Wacker process, the classical rate law (Henry, 1959-63 kinetic studies) is:

where:

• $[\text{PdCl}_4^{2-}]$ and $[\text{C}_2\text{H}_4]$ appear with simple first-order dependence,
• $[\text{H}^+]^{-1}$ reflects that the active nucleophile is hydroxide/water in a pre-equilibrium deprotonation step, so more acid slows the reaction,
• $[\text{Cl}^-]^{-2}$ reflects two chloride pre-equilibria: dissociation of Cl⁻ from PdCl₄²⁻ to open a coordination site for ethylene, and a second Cl⁻ dissociation to open a site for the incoming water/hydroxide nucleophile.

Reading: high chloride (needed industrially to keep Pd in solution and suppress Pd black) inhibits the reaction kinetically even though it is thermodynamically required — a genuine catalyst design trade-off. For metathesis, the position of equilibrium in a cross metathesis between two symmetric internal alkenes reflects a simple statistical/thermodynamic combination:

where:

• the numerator counts the unsymmetrical cross product (formed in two indistinguishable ways, hence squared concentration in the statistical limit $K_{eq}=4$),
• deviations from $K_{eq}=4$ reflect real thermodynamic (steric/electronic) preferences among the three alkenes, not just combinatorics.

Reading: cross metathesis alone gives a statistical mixture unless one product is removed from the system (as in ring-closing metathesis expelling ethylene, or ethenolysis using excess ethylene to push the equilibrium toward terminal-alkene products). For Pd cross-coupling, the overall rate is normally governed by whichever step has the highest barrier for a given substrate/ligand combination:

where:

• $k_{OA}$, $k_{TM}$, $k_{RE}$ are the oxidative addition, transmetalation, and reductive elimination rate constants,
• electron-poor, sterically hindered aryl halides typically make oxidative addition fast but reductive elimination slow (electron-rich Pd(II) resists reductive elimination), so bulky electron-rich phosphines (P(t-Bu)₃, SPhos, XPhos) are used specifically to accelerate the slow step.

Reading: ligand design in cross-coupling is diagnostic chemistry — you first identify which step limits turnover for your substrate class, then choose a ligand engineered to accelerate exactly that step.

Example 1 (foundational) — electron counting through the hydroformylation cycle. Track the d-electron count and total electron count (ionic method) at each Rh intermediate for 1-hexene hydroformylation with HRh(CO)(PPh₃)₃. Start: HRh(CO)(PPh₃)₃. Assign H⁻ (hydride, X-type, 2e via ionic method contributing as anionic ligand), CO (L-type, 2e), three PPh₃ (L-type, 2e each). Metal oxidation state: Rh with one X-type hydride ligand and overall neutral complex means Rh is formally +1 (Rh(I), d⁸).

After losing one PPh₃: HRh(CO)(PPh₃)₂, still Rh(I) d⁸:

1-Hexene (C₄H₉CH=CH₂) binds as an L-type 2-electron alkene donor: 16 + 2 = 18e alkene adduct (a transient intermediate). Migratory insertion converts the alkene (L-type) and hydride (X-type) into a single alkyl (X-type) ligand — net loss of one ligand slot (alkene + hydride, 2 donor "slots," become one alkyl X-type ligand), dropping back to 16e: hexyl-Rh(CO)(PPh₃)₂, Rh(I) d⁸,

CO binds (18e alkene-like adduct transiently, 18e), then migratory insertion of CO into the Rh–alkyl bond gives the acyl, again a single X-type ligand, back to 16e: heptanoyl-Rh(CO)(PPh�3)₂. Oxidative addition of H₂ adds two new X-type ligands (2 hydrides) and formally increases Rh oxidation state by two units, Rh(I) → Rh(III), d⁸ → d⁶:

Wait — check ligand count carefully: this Rh(III) intermediate carries acyl, 2×H, CO, 2×PPh₃ = 6 ligands around an octahedral center, giving

(the arithmetic above double-counts one phosphine in the first line; correctly there are two phosphines contributing 2e each = 4e total, plus acyl 2e, two hydrides 4e, CO 2e, giving $6+2+4+2+4=18e$), an 18-electron, coordinatively saturated octahedral Rh(III) species, as required for a stable oxidative-addition product. Reductive elimination of the acyl and one hydride releases heptanal (C₆H₁₃CHO) and regenerates 16e Rh(I) HRh(CO)(PPh₃)₂, which recombines with free PPh₃ to regenerate the 18e resting state HRh(CO)(PPh₃)₃, closing the loop exactly where it started.

Example 2 (applied) — predicting cross metathesis outcome for a terminal alkene mixture. Two terminal alkenes, 1-octene (C₆H₁₃CH=CH₂) and methyl oleate-type methyl acrylate-terminated alkene (approximate as R–CH=CH₂ with R = CH₂CH₂CO₂Me for simplicity), are combined 1:1 with a Grubbs II catalyst. Estimate qualitatively what cross product forms and why ethylene is a byproduct. Each terminal alkene reacts with the Ru=CHPh benzylidene (or, after the first turnover, Ru=CH₂ methylidene) via [2+2] cycloaddition to a metallacyclobutane, then retro-[2+2] can expel either the original alkylidene (unproductive) or a new alkylidene bearing the alkyl chain (productive), releasing styrene (first turnover) or, in subsequent turnovers, ethylene (H₂C=CH₂) as the alkylidene fragment exchanges with a second terminal alkene. Concretely: Ru=CH(C₆H₁₃) (from octene) reacts with the acrylate-alkene's terminal CH₂=CH–R′ to form a metallacyclobutane with substituents C₆H₁₃ and R′ on the 3- and 4-positions; retro-[2+2] across the productive diagonal releases the cross product C₆H�13CH=CHR′ (a type II cross metathesis product) and regenerates Ru=CH₂. This Ru=CH₂ methylidene is the catalytically active species that continues turning over both alkene partners, ejecting ethylene as the low-molecular-weight byproduct each time a terminal alkene reacts productively — precisely because both starting materials are monosubstituted (terminal), so every productive metathesis event that consumes two terminal-alkene equivalents nets one ethylene molecule as a side product; this ethylene continuously escapes the reaction vessel (bp −104 °C), which is exactly the entropic/volatility driving force that pulls the statistical cross-metathesis equilibrium toward the desired 1,2-disubstituted cross product rather than sitting at the symmetric mixture. Using the statistical relation from the "core ideas" section with the two homodimers as the reference alkenes,

predicts roughly a 50% cross product / 25% / 25% homodimer split at full statistical equilibrium in the absence of any thermodynamic or ethylene-loss bias; because both homodimers are also 1,2-disubstituted internal alkenes (not more volatile than the cross product), true selectivity enhancement in this particular pairing relies on subtle thermodynamic differences (e.g., steric strain differences between homodimer and heterodimer), illustrating why cross metathesis selectivity is harder to control than ring-closing metathesis, where the ethylene-loss driving force is unambiguous and large.

Example 3 (exam/competition level) — diagnosing the rate-limiting step in a Suzuki coupling from a kinetic isotope/Hammett-type dataset. A Suzuki coupling of para-substituted aryl bromides (4-X-C₆H₄Br) with phenylboronic acid, catalyzed by Pd(PPh₃)₄/base, is run for a series of X = OMe, H, CF₃, and the pseudo-first-order rate constants (relative to X = H ≡ 1.00) are found to be

A Hammett plot of $\log(k_{\text{rel}})$ against $\sigma_p$ gives a straight line with slope $\rho \approx +2.4$. Identify which step of the catalytic cycle is turnover-limiting and justify it mechanistically. A large positive $\rho$ means electron-withdrawing substituents (large positive $\sigma_p$, e.g., CF₃, $\sigma_p = 0.54$) accelerate the reaction sharply, while electron-donating substituents (OMe, $\sigma_p = -0.27$) slow it down. Oxidative addition of Ar–Br to Pd(0) is an electron-rich metal attacking an electrophilic C–Br σ orbital; this step is accelerated by electron-withdrawing aryl substituents because they lower the energy of the C–Br σ orbital and stabilize developing negative charge on the ring in the oxidative-addition transition state — exactly the sign and rough magnitude ($\rho$ in the +1 to +3 range is typical for concerted three-centered oxidative additions to electron-rich d¹⁰ Pd(0)) expected if oxidative addition, not transmetalation or reductive elimination, is rate-limiting. This is consistent with the general observation that with a highly active, well-ligated Pd(0)/monophosphine catalyst and excess boronic acid/base (transmetalation kept fast and non-limiting by using excess nucleophile), electron-poor aryl halides couple fastest and oxidative addition is turnover-limiting — the opposite substituent trend would instead implicate reductive elimination (which is retarded, not accelerated, by electron-withdrawing R groups on the departing aryl, since electron-poor Pd(II) centers are typically slower to reductively eliminate) as rate-limiting.