Concerted Metalation-Deprotonation: How a Carboxylate Cleaves a C–H Bond

What CMD is: a base that grabs the proton while the metal grabs the carbon

Concerted metalation-deprotonation (CMD) is the pathway by which a late transition-metal complex bearing a carboxylate ligand — most often acetate (⁻O₂CCH₃) or pivalate (⁻O₂CtBu) — cleaves an otherwise inert aromatic or vinylic C–H bond. The defining feature is that two events happen in one elementary step: the metal forms a new M–C σ-bond to the carbon while a pendant carboxylate oxygen abstracts the hydrogen as it leaves. There is no discrete carbanion, no Wheland arenium cation, and no free proton — the C–H bond is broken heterolytically inside a single, cyclic transition state.

The name is due to Keith Fagnou and Sergei Gorelsky, whose DFT studies (2006–2008) showed that Pd(OAc)₂-mediated arene metalation proceeds not by electrophilic attack but by this concerted proton shuttle. In parallel, David Davies, Stuart Macgregor, and coworkers described essentially the same process and named it ambiphilic metal–ligand activation (AMLA), emphasizing that the transition state is ambiphilic: the metal is electrophilic toward carbon while the carboxylate oxygen is basic toward hydrogen. AMLA-6 (a six-membered version) and CMD refer to the same geometry; the terms are used interchangeably in the literature, with CMD dominant in synthetic papers.

Crucially, CMD is not oxidative addition of the C–H bond, and it is not σ-bond metathesis in the classic d⁰ sense. In oxidative addition the metal is formally oxidized by two units and both H and C end up bound to the metal; in CMD the metal oxidation state is unchanged (a Pd(II) stays Pd(II)), the hydrogen leaves on the carboxylate as a carboxylic acid, and only carbon binds the metal. This makes CMD the workhorse for d⁶–d⁸ late metals — Pd(II), Ru(II), Rh(III), Ir(III) (and, less commonly, Pt(II), which often sits toward the IES limit) — that resist the +2 oxidation-state jump of a classical oxidative addition.

The six-membered transition state and its electron bookkeeping

Picture a square-planar Pd(II) center carrying a κ²- or κ¹-bound acetate. As the arene approaches, one carboxylate oxygen remains bonded to Pd while the second (or the same, in a κ¹→proton-shuttle picture) swings toward the ortho C–H. The transition state closes a six-membered ring: Pd–O–C(=O)···H···C(aryl), with Pd completing the loop back to the aryl carbon. The forming Pd–C bond, the breaking C–H bond, and the O···H interaction are all partially formed at the saddle point — the geometric hallmark that DFT captures and that distinguishes CMD from a stepwise mechanism.

The electron count is conserved throughout. Consider Pd(OAc)₂ activating benzene: the aryl C–H σ-electrons are redistributed so that the carbon donates a two-electron σ bond to Pd (a new anionic aryl ligand, X-type, contributing 2 e⁻ in the ionic count) while the proton departs as part of acetic acid, which leaves the coordination sphere or re-coordinates as a neutral L-type acid. The metal goes from a 16-electron mononuclear Pd(II)(OAc)₂ fragment (Pd(OAc)₂ itself is a trimer, Pd₃(OAc)₆, in the solid state) to a 16-electron Pd(II)(Ar)(OAc)·HOAc-type species — Pd remains d⁸, +2, and the count is bookkept without any redox change. This is why CMD sits comfortably in a Pd(0)/Pd(II) or Pd(II)/Pd(IV) catalytic cycle without introducing an unwanted oxidation event of its own.

The energetics favor an early-to-central transition state whose barrier is exquisitely sensitive to the C–H bond's acidity and to the geometric strain of bending the arene into the cyclic array. Gorelsky's dissection separates the CMD barrier into a distortion (deformation) term — the cost of pyramidalizing/bending the arene and the Pd–carboxylate fragment — and an interaction term. For a series of arenes the deprotonation energy (proportional to C–H pKₐ) tracks the barrier closely: the more acidic the C–H, the lower the CMD barrier. This is exactly opposite to electrophilic aromatic substitution, and it is the single most useful predictive statement about CMD selectivity.

Why it matters: it reverses the rules of arene selectivity

Classical Friedel–Crafts and electrophilic palladation preferentially functionalize electron-rich arenes at the most nucleophilic position. CMD does the reverse. Because the barrier scales with C–H acidity, electron-poor arenes — pentafluorobenzene, polyfluoroarenes, azoles, electron-deficient heterocycles — are among the most reactive substrates. Pentafluorobenzene, whose lone C–H has a pKₐ near 25–29 (much more acidic than benzene's ~43), undergoes direct arylation with remarkable ease under Pd/carboxylate catalysis. A synthetic chemist who internalizes "most acidic C–H reacts fastest" can predict CMD regiochemistry across a huge range of heteroarenes.

This inversion is what made direct arylation a practical alternative to cross-coupling. Instead of pre-installing an organometallic partner (a boronic acid for Suzuki, a stannane for Stille), one couples an aryl halide directly with an arene C–H bond. The atom economy improves and a stoichiometric organometallic waste stream disappears. Much of the pharmaceutical and materials literature on synthesizing biaryls, fused heterocycles, and conjugated polymers (e.g., direct-arylation polymerization of thiophenes for organic electronics) rests on CMD as the C–H cleavage step.

CMD also underpins the broad family of carboxylate-assisted, chelation-directed C–H functionalizations developed by Lutz Ackermann (Ru), Frank Glorius, Melanie Sanford, Jin-Quan Yu, and others. In Ru(II)-catalyzed arylations and annulations, added carboxylate (KOPiv, KO₂CMes / MesCO₂H) is essential; Ackermann has shown that carboxylates such as 1-adamantanecarboxylate or mesitylenecarboxylate dramatically accelerate the metalation, precisely because they lower the CMD barrier. The carboxylate is not a spectator — it is a co-catalyst that performs the deprotonation.

A worked example: pivalic acid, KIE, and the Fagnou arylation

The canonical demonstration is Fagnou's 2006–2008 direct arylation of arenes with aryl bromides, Pd(OAc)₂ or Pd(0)/phosphine, a base such as K₂CO₃ or Cs₂CO₃, and — the key discovery — catalytic pivalic acid (PivOH, ~30 mol%). Without PivOH many couplings are sluggish; with it, yields jump and previously unreactive arenes engage. Mechanistically, the carbonate deprotonates PivOH to pivalate, which binds Pd and becomes the internal base for CMD. Pivalate outperforms acetate because its bulky tert-butyl group and higher basicity organize a lower, less crowded six-membered transition state.

Two experiments nail the mechanism:

• Kinetic isotope effect. Replacing the reacting C–H with C–D gives a primary KIE of roughly k_H/k_D ≈ 3–6 (values around 3–5 are typical for Pd/pentafluorobenzene-type systems, and larger for some intramolecular cases). A primary KIE of this magnitude means the C–H bond is broken in or before the rate-determining step — direct evidence that deprotonation is concerted with metalation rather than a fast post-rate-limiting step.
• Electronic trend. Competition experiments with substituted arenes show electron-withdrawing groups accelerate metalation. A Hammett-type analysis gives a positive ρ (buildup of negative character / anionic-like carbon at the transition state), consistent with the carboxylate removing a proton and the aryl becoming carbanion-like as it binds Pd.

Put numbers to the free-energy surface: DFT for Pd(OAc)₂ + benzene places the CMD barrier around 25–30 kcal/mol, and for pentafluorobenzene several kcal/mol lower — enough to explain why the perfluoroarene reacts at 80–120 °C while benzene needs harsher conditions or does not react at all. Swapping acetate for pivalate typically shaves an additional ~2–4 kcal/mol off the computed barrier, matching the experimental rate acceleration Fagnou observed.

Limits, subtleties, and where CMD blurs into other mechanisms

CMD is a family of related transition states, not a single rigid geometry, and several subtleties matter in practice:

• κ² vs κ¹ carboxylate and the proton shuttle. Whether both carboxylate oxygens engage (a true bidentate-derived shuttle) or one oxygen deprotonates while the other stays on the metal is system-dependent. Macgregor's AMLA-6 label emphasizes the six-membered ring; some carbonate- or amide-assisted variants form other ring sizes.
• Not the only internal base. Carbonate, phosphate, and even a coordinated amido or acetamidate ligand can perform the same concerted deprotonation. Jin-Quan Yu's mono-N-protected amino acid (MPAA) ligands are thought to deliver the proton to the amide carbonyl in a CMD-like step, which is how they impart enantioselectivity to C–H activation.
• KIE is diagnostic but not proof. A large KIE shows C–H cleavage is rate-limiting; it does not by itself distinguish CMD from oxidative addition or σ-bond metathesis. The full case for CMD rests on the carboxylate dependence, the electron-poor-arene preference, and the computed six-membered saddle point together.

There is genuine, ongoing discussion about the exact character of the transition state. The CMD vs AMLA distinction is largely semantic (both describe the six-membered, ambiphilic pathway), but authors do debate whether specific systems are better described as internal electrophilic substitution (IES) — a related model championed by Robert Bergman, Karen Goldberg, and Tom Cundari for some Ir and Pt amido/hydroxo systems — in which the arene is attacked by an electrophilic metal and the proton is delivered to a cis heteroatom ligand. IES and CMD sit on a mechanistic continuum: as the metal becomes more electrophilic and the C–H less acidic, the description shifts from carboxylate-base-driven (CMD) toward metal-electrophile-driven (IES). The honest position is that these are limiting descriptions of one ambiphilic process, and a given catalyst may lie anywhere between them.

Finally, CMD does not rescue every C–H bond. Very electron-rich, sterically hindered, or high-pKₐ aliphatic C–H bonds remain difficult, and directing groups are usually still needed to bring the metal into range and to override the intrinsic acidity-based selectivity. The distortion term also penalizes activations that require severe arene bending, so geometry — not just pKₐ — sets real-world scope.

History and reach: from cyclometalation curiosity to a synthetic staple

The seeds predate the name. Cyclometalation and orthometalation of aryl imines and azobenzenes with Pd(OAc)₂, studied since the 1960s–70s (Cope, Trofimenko, and others), already showed that acetate-bearing palladium cleaves ortho C–H bonds cleanly. For decades this was rationalized loosely as "electrophilic" metalation. The reframing came from computation: Sanford, Fagnou, Gorelsky, Davies, and Macgregor in the mid-2000s showed the transition states were the concerted, carboxylate-assisted six-membered structures now called CMD/AMLA, overturning the electrophilic picture for these systems.

Fagnou's pivalic-acid discovery (Org. Lett. and JACS, 2006–2008) was the moment CMD became a design principle rather than a mechanistic footnote: chemists began adding carboxylates deliberately to accelerate C–H cleavage. The idea propagated across metals and reaction classes — Ackermann's carboxylate-assisted Ru(II) catalysis, Glorius's and Satoh–Miura's Rh(III)/Cp* annulations, Yu's ligand-accelerated and enantioselective Pd C–H activation, and direct-arylation polymerization in materials science. In essentially all of these, a carboxylate performing a concerted deprotonation is the shared mechanistic heart.

The intellectual payoff is a predictive, additive-tunable model for C–H bond cleavage: choose an electron-poor or acidic C–H, add pivalate, and expect metalation; expect a primary KIE; expect selectivity that inverts Friedel–Crafts. No Nobel Prize has been awarded specifically for CMD, but it is inseparable from the broader C–H functionalization revolution that has reshaped how complex molecules — pharmaceuticals, agrochemicals, and conjugated materials — are assembled from simple C–H bonds rather than pre-functionalized coupling partners.

CMD contrasted with electrophilic aromatic metalation (SEAr / Wheland-type) for arene C–H functionalization

Feature	CMD (carboxylate-assisted)	Electrophilic metalation (SEAr)
Proton fate	Removed by carboxylate O in the same step as M–C bond forms	Lost to external base after Wheland intermediate forms
Arene electronics	Electron-poor / acidic C–H react fastest (e.g., polyfluoroarenes)	Electron-rich rings react fastest
Kinetic isotope effect	Primary, ≈ 3–6 (C–H cleavage rate-limiting)	Small or none (deprotonation not rate-limiting)
Key transition state	Six-membered M–O···H–C cyclic, ambiphilic	Arenium (Wheland) σ-complex, cationic
Regioselectivity driver	C–H pKₐ + geometric distortion energy	Directing-group resonance / charge stabilization
Rate signature	First order in [carboxylate]; PivOH accelerates	Independent of carboxylate; needs π-nucleophilic arene