Oxidative Addition & Reductive Elimination

Two steps, one metal, opposite directions

Think of a transition metal as a tiny vise with electrons to spare. Oxidative addition (OA) is the metal grabbing a molecule and tearing one of its bonds in half — both fragments end up clamped to the metal as new ligands. In doing so the metal hands two of its own electrons into the new metal–ligand bonds, so its formal oxidation state goes up by two and its d-electron count goes down by two. Reductive elimination (RE) runs the movie backwards: two ligands already on the metal couple to each other, leave together as a single new molecule, and dump two electrons back onto the metal. Oxidation state drops by two; d-count climbs back by two.

The canonical example is the addition of methyl iodide to a low-valent metal:

Oxidative addition:    LₙM(0)  +  CH₃–I   →   LₙM(II)(CH₃)(I)
                       d¹⁰              two new σ-bonds, d⁸

Reductive elimination: LₙM(II)(CH₃)(Ph) →   LₙM(0)  +  CH₃–Ph
                       d⁸               one new C–C bond, d¹⁰

Both events happen at a single atom. That is the whole trick of organometallic catalysis: instead of two organic molecules having to find each other and react directly (slow, hard to control), they take turns binding to the same metal, which lines them up and lowers the barrier to coupling.

           Pd(0)
          ╱      ╲   reductive
oxidative╱        ╲  elimination
addition╱          ╲  (forms R–R')
  (Ar–X) ╲        ╱
          ╲      ╱
       Pd(II)(Ar)(X) → transmetalation → Pd(II)(Ar)(R')
                       (swaps X for R')

How oxidative addition actually happens

"Oxidative addition" names a result, not a single mechanism. There are at least three distinct pathways, and the right one depends on the substrate:

• Concerted three-centered. Non-polar substrates like H₂ and Si–H add in one step through a triangular transition state — the metal slides into the σ-bond, both new bonds forming as the old one breaks. This is stereospecific and gives cis products. It is how Vaska's complex adds H₂.
• SN2-like. The electron-rich metal attacks the carbon of an alkyl halide from the back side, displacing halide with inversion of configuration at carbon. Halide then re-binds. This dominates for primary alkyl halides and tracks nucleophilic-substitution reactivity (I > Br > Cl).
• Radical / single-electron. The metal donates one electron to the C–X bond, generating a radical pair. This pathway scrambles stereochemistry at carbon and is common with hindered or activated substrates and with some nickel systems.

For aryl halides — the bread and butter of cross-coupling — the operative step is a concerted insertion of Pd(0) into the C–X bond, often after a ligand has dissociated to give a reactive 12- or 14-electron Pd(0) fragment. The barrier scales with the C–X bond dissociation energy, which is why the halide identity matters so much (see the table below).

The electron bookkeeping

The "oxidative" and "reductive" labels come straight from formal oxidation-state counting. In the ionic (donor-pair) convention, each bond to the metal is assigned entirely to the more electronegative atom. Adding H₂ to an Ir(I) center illustrates the bookkeeping:

trans-IrCl(CO)(PPh₃)₂   +   H₂   →   IrH₂Cl(CO)(PPh₃)₂
   Ir(+1), d⁸, 16e⁻                    Ir(+3), d⁶, 18e⁻

  ΔOS = +2     Δd = −2     Δelectron count = +2 (gained two H⁻ ligands)

Each hydride (H⁻) is a two-electron X-type donor, so the metal's electron count rises by 4 from two ligands but it loses 2 from being oxidized — net +2 in the 18-electron count, exactly enough to take a 16-electron Ir(I) complex to a coordinatively saturated 18-electron Ir(III). A useful rule of thumb: oxidative addition needs at least two open coordination sites (or the ability to make them) and at least two spare d-electrons. That is why d¹⁰ Pd(0) and d⁸ Ir(I)/Rh(I) are such good OA partners, and why d⁰ Ti(IV) cannot do classical OA at all.

Substrate reactivity: it all comes down to the bond you break

C–X bond	Approx. BDE (kJ/mol)	Relative OA rate	Practical note
Ar–I	~272	fastest	Adds readily even with simple PPh₃ ligands
Ar–OTf (triflate)	—	fast	Pseudo-halide; great for phenol-derived substrates
Ar–Br	~339	fast	The everyday workhorse electrophile
Ar–Cl	~397	slow	Needs bulky electron-rich phosphines or NHCs
Ar–F	~526	negligible	Essentially inert to classical OA
Ar–OMe	~420	very slow	Possible with Ni and strong-donor ligands

The pattern is monotonic: weaker C–X bonds add faster, so reactivity falls I > OTf ≈ Br > Cl ≫ F. The practical revolution of the last 25 years — bulky, electron-rich ligands like tri-tert-butylphosphine, SPhos, XPhos, and N-heterocyclic carbenes — was driven by the need to make cheap, abundant aryl chlorides couple at room temperature. Electron-rich ligands make the metal a better electron donor (lowering the OA barrier); bulk forces low coordination numbers that keep the metal reactive and then accelerate the final reductive elimination.

What controls reductive elimination

Reductive elimination is the bond-forming, product-releasing step, and it has its own rules. The two groups that couple must be cis on the metal, so the first requirement is geometric. Beyond that:

• Sterics help. Bulky ligands push the two coupling groups together and relieve strain when they leave, so crowding accelerates RE — the opposite of OA, where crowding slows things.
• Bite angle matters. Wide-bite-angle bisphosphines compress the angle between the two reacting ligands. Xantphos (bite angle ≈ 108°) and DPEphos speed C–N reductive elimination dramatically; this is the design principle behind efficient Buchwald–Hartwig amination.
• Electronics. Electron-poor metals (and π-acceptor ligands like CO) make RE faster because the metal "wants" its electrons back. Adding an electron-withdrawing ligand can switch on an otherwise stuck elimination.
• Coordination number. Dropping from a four-coordinate Pd(II) to a three-coordinate T-shaped intermediate is often the true trigger; a ligand dissociates first, then C–C bond formation is fast.

Ease of RE also depends on what you are coupling: C(sp²)–C(sp²) (biaryl) is fast, C(sp²)–C(sp³) is moderate, and C(sp³)–C(sp³) is hard — slow RE plus a competing β-hydride elimination is exactly why alkyl–alkyl coupling stayed difficult for decades.

Energetics and real numbers

For palladium oxidative addition into aryl bromides, computed and measured free-energy barriers (ΔG‡) typically fall in the 60–90 kJ/mol range with common phosphine ligands, dropping toward the low end for electron-poor aryl bromides and electron-rich Pd. Aryl chlorides push ΔG‡ up by roughly 20–40 kJ/mol, which — through the Eyring relation — costs you several orders of magnitude in rate at room temperature unless you compensate with a better ligand.

A 35 kJ/mol higher barrier at 298 K:
  rate ratio = exp(−ΔΔG‡ / RT)
             = exp(−35,000 / (8.314 × 298))
             = exp(−14.1)
             ≈ 7 × 10⁻⁷

So an unactivated aryl chloride can be a million-fold slower to add than the corresponding bromide with the same ligand — which is precisely the gap that SPhos-class ligands close. On the thermodynamic side, oxidative addition of H₂ to Vaska's complex is exothermic by about ΔH ≈ −60 kJ/mol and is readily reversible near room temperature, which is why the complex can be used to reversibly bind small molecules. The full Pd cross-coupling cycle is downhill overall because reductive elimination forms a strong new C–C bond (~350 kJ/mol) and regenerates the catalyst.

Where this drives real chemistry

Oxidative addition and reductive elimination are not textbook curiosities — they manufacture drugs, agrochemicals, and materials at industrial scale. The 2010 Nobel Prize in Chemistry went to Richard Heck, Ei-ichi Negishi, and Akira Suzuki for the palladium-catalyzed cross-couplings built on exactly these two steps.

• Suzuki–Miyaura coupling. Pd(0) does OA into an aryl halide, transmetalation hands an aryl group over from a boronic acid (base-assisted), then RE spits out the biaryl. It builds the biaryl cores of countless pharmaceuticals; losartan (a blood-pressure drug) and the fungicide boscalid are made this way on multi-ton scale.
• Buchwald–Hartwig amination. Same OA/RE skeleton but forming a C–N bond, turning aryl halides plus amines into arylamines — the workhorse for making the nitrogen-rich rings in medicinal chemistry.
• Monsanto / Cativa acetic-acid process. Rhodium (Monsanto) or iridium (Cativa) repeatedly adds CH₃–I oxidatively, migrates CO, and reductively eliminates acetyl iodide. Billions of kilograms of acetic acid per year ride on this OA/RE loop.
• Hydroformylation. Rhodium and cobalt catalysts toggle oxidation states to turn alkenes + CO + H₂ into aldehydes — one of the largest-volume homogeneous catalytic processes in the world.

Common misconceptions and pitfalls

• "Nothing is really oxidized." The metal genuinely is — formally. Its oxidation state and d-electron count change by two even though no electrons leave the molecule. The bookkeeping is real and predicts reactivity.
• Assuming OA is always concerted. Aryl and H₂ substrates add concertedly with retention of geometry; primary alkyl halides go SN2 with inversion at carbon; hindered substrates go radical with racemization. The mechanism sets the stereochemistry — don't assume.
• Forgetting the cis requirement for RE. Two trans ligands cannot reductively eliminate directly; the complex must first isomerize to cis. A complex that looks ready to couple may just be stuck in the wrong geometry.
• Treating sterics the same for both steps. Bulk slows oxidative addition (harder to approach the bond) but speeds reductive elimination (groups pushed together). A ligand optimized for one step can sabotage the other; good ligands balance both.
• Confusing reductive elimination with β-hydride elimination. RE couples two ligands into one product; β-hydride elimination ejects an alkene and leaves a metal hydride. For alkyl–metal intermediates, β-hydride elimination is the chief unwanted side reaction that limits sp³–sp³ coupling.
• Thinking only Pd works. Ni, Rh, Ir, and Pt all do OA/RE. Nickel is cheaper and inserts into tough C–O and C–Cl bonds; it is the basis of a whole modern wave of earth-abundant-metal coupling.