Grignard and Organolithium Addition to Aldehydes and Ketones

Grignard and organolithium additions to aldehydes and ketones remain a workhorse of pharmaceutical and fine-chemical synthesis precisely because they build C–C bonds with predictable regiochemistry and (with chiral catalysts or chiral ketone substrates) tunable stereochemistry. Industrially, Grignard chemistry is used at multi-ton scale to manufacture fragrance alcohols (e.g., additions to substituted cyclohexanones and aromatic aldehydes for the perfume industry), and in the synthesis of active pharmaceutical ingredients where a tertiary alcohol stereocenter must be installed with control — statin side-chain intermediates and various antihistamine syntheses have historically relied on Grignard or organolithium addition steps. Tamoxifen-related and steroid syntheses have used Grignard addition to introduce quaternary-adjacent tertiary alcohol centers that are difficult to access any other way. In modern asymmetric catalysis, chiral additives (chiral amino alcohols such as (−)-DAIB or chiral Lewis-base catalysts) are used to convert the diethylzinc addition to aldehydes — a close cousin of the Grignard/organolithium addition mechanistically, using organozinc rather than organomagnesium reagents, and famously developed by Noyori — into a highly enantioselective process, illustrating how the fundamental Bürgi–Dunitz addition mechanism taught here underlies an entire modern subfield of catalytic asymmetric C–C bond formation. Organolithium reagents specifically remain indispensable in total synthesis for late-stage, highly reactive additions where maximum nucleophilicity is required, and in metalation chemistry (directed ortho-metalation, halogen-metal exchange) that generates the very organolithium nucleophiles used in carbonyl addition steps downstream. In an academic and pedagogical sense, this reaction is also the conceptual bridge to Alcohol Products and Retrosynthesis of C–C Bonds: every time a synthetic chemist looks at a target alcohol and mentally "disconnects" a C–C bond adjacent to the C–OH carbon, they are invoking — in reverse — exactly the mechanism worked through above, imagining which carbonyl compound and which organometallic nucleophile could be combined to rebuild that bond forward.

Ask the tutor: Why does axial attack dominate for small nucleophiles on cyclohexanones, and how does that change for bulky nucleophiles like tert-butylmagnesium bromide?

Start from what is already known about the reagent: in a Grignard reagent R–MgX, the carbon–magnesium bond is polarized C^(δ–)–Mg^(δ+) because carbon (electronegativity ≈ 2.5) is more electronegative than magnesium (≈ 1.3). The same logic, even more extreme, holds for R–Li (Li electronegativity ≈ 1.0), which is why organolithiums are, bond for bond, even more reactive and more strongly basic than their magnesium cousins. In both cases the working approximation — never perfectly literal, but mechanistically decisive — is that the carbon carries substantial carbanion character: a lone pair in an sp³ orbital, held in electrostatic and covalent partnership with a cationic metal center. Meet this carbanion with a carbonyl compound and you have paired the best available nucleophile with one of the best available electrophiles. The carbonyl carbon is sp² hybridized, part of a C=O π system in which oxygen's higher electronegativity polarizes the π bond so that carbon bears a partial positive charge and a low-lying, accessible π antibonding orbital. This π orbital is the target: filled orbital (carbanion lone pair, HOMO) attacks empty orbital (carbonyl π, LUMO) in classic frontier-molecular-orbital fashion. Geometrically, the nucleophile does not approach oxygen or attack straight down the C=O axis. It approaches from an angle of roughly 107° relative to the C=O bond, tilted away from the oxygen lone pairs and toward best overlap with the extended lobe of the π orbital — the Bürgi–Dunitz angle, deduced originally from crystal-structure surveys of nucleophile...carbonyl contacts and rationalized by the requirement of maximal HOMO–LUMO overlap while minimizing electron-electron repulsion with oxygen's lone pairs. As the nucleophile's orbital begins to overlap with π*, the C=O π bond begins to break, and electron density flows onto oxygen, converting it from a neutral carbonyl oxygen into an alkoxide oxygen bearing a full negative charge (paired, in the immediate product, with the Mg²⁺ or Li⁺ counter-ion in a four-membered or aggregated transition state). A second essential feature of the mechanism, often underplayed at the introductory level, is that the metal cation is not a spectator. In real ethereal solvents, Grignard reagents exist substantially as the Schlenk equilibrium mixture of RMgX and R₂Mg/MgX₂, often further aggregated as dimers, and the magnesium is coordinated by two ether oxygens (from THF or Et₂O). During the addition step, the electrophilic Mg²⁺ center simultaneously coordinates the incipient alkoxide oxygen, effectively acting as a Lewis acid catalyst that pre-organizes and stabilizes the transition state — lowering the activation barrier by neutralizing developing negative charge on oxygen at the very moment it is forming. This cyclic, six-membered-ring-like transition state (carbon nucleophile, carbonyl carbon, oxygen, magnesium, and often a bridging halide or a coordinated ether oxygen) is a good mental picture for why the addition is fast, essentially irreversible, and only weakly reversible in the reverse (retro-addition) direction under the reaction conditions.

Once the new C–C bond and the alkoxide are formed, the carbon that was the carbonyl carbon is now sp³ and tetrahedral, bonded to the newly delivered R group, whatever substituents were already present (H and/or alkyl), and the oxygen (now an alkoxide, tightly ion-paired to the metal). This metal alkoxide is completely stable under the anhydrous, basic reaction conditions — magnesium and lithium alkoxides do not spontaneously revert to starting materials, and they are themselves poor electrophiles (no more empty π* orbital on carbon), so the reaction stops cleanly at the mono-addition stage. Protonation only happens deliberately, on aqueous acidic workup (dilute HCl or saturated NH₄Cl solution), which delivers a proton to the alkoxide oxygen to give the neutral alcohol and precipitate the magnesium salts.

The governing relationships here are less "equations" in the physics sense and more structural bookkeeping rules that fall directly out of the mechanism, together with the frontier-orbital energetics that set relative rates.

where:

• $\text{Nu}^-$ is the carbanion-like alkyl or aryl group delivered from the organometallic reagent,
• $R'$ (and, for a ketone, a second $R''$) are the substituents already present on the carbonyl carbon,
• the product is drawn as the tightly ion-paired metal alkoxide, the species that actually exists in the reaction flask before workup.

The product-class rule follows directly from counting substituents on the new sp³ stereocenter:

where:

• an aldehyde carbon carries one carbon substituent (R′) and one H, so the product carbinol carbon carries two carbon substituents (R′, R) and one H — a secondary alcohol,
• a ketone carbon carries two carbon substituents (R′, R″), so the product carbinol carbon carries three carbon substituents (R′, R″, R) — a tertiary alcohol,
• formaldehyde carries zero carbon substituents (two H's), so the product carbinol carbon carries one carbon substituent (R) and two H — a primary alcohol.

A rate relationship, qualitative but essential for predicting chemoselectivity, follows from steric and electronic accessibility of the carbonyl LUMO:

where:

• $k$ is the second-order rate constant for nucleophilic addition of a given organometallic reagent to the carbonyl compound,
• aldehydes are faster because they have only one alkyl (or aryl) substituent, giving less steric hindrance to the Bürgi–Dunitz approach and less electron donation into the carbonyl (making the carbon more electrophilic than in a ketone),
• esters react far more slowly per addition step because the ester oxygen's lone pair donates into the carbonyl π system, reducing electrophilicity, and (mechanistically distinct) esters undergo a subsequent elimination of alkoxide to regenerate a ketone carbonyl, allowing a second equivalent of RMgX/RLi to add — a different reaction entirely, foreshadowing ester-to-tertiary-alcohol chemistry, not treated further here.

Example 1 — Foundational: methylmagnesium iodide plus propanal. Propanal, CH₃CH₂CHO, is treated with one equivalent of methylmagnesium iodide (CH₃MgI) in dry diethyl ether, followed by aqueous acid workup. Identify the mechanism, the immediate organometallic intermediate, and the final product, and state whether the product is chiral. Mechanistic walk-through. The carbonyl carbon of propanal is sp², bonded to an ethyl group, a hydrogen, and the oxygen. The methyl carbanion (delivered from CH₃MgI, whose C–Mg bond is polarized C^(δ–)–Mg^(δ+)) attacks along the Bürgi–Dunitz trajectory. One curved arrow: C(of CH₃)–Mg bonding pair → carbonyl carbon, forming the new C–C bond. Second curved arrow: C=O π electrons → oxygen, generating the alkoxide. The magnesium iodide cation coordinates the developing alkoxide oxygen throughout, stabilizing the transition state.

Aqueous acid workup protonates the alkoxide oxygen:

Result. The product is butan-2-ol, CH₃CH₂CH(OH)CH₃, a secondary alcohol (as expected from aldehyde + one R group), formed with the carbinol carbon bonded to four different groups: OH, H, ethyl, and methyl. This is a genuine stereocenter, so the product is chiral. Because the achiral starting materials (propanal, achiral methyl Grignard) and achiral reagents provide no chiral influence anywhere in the mechanism, the Bürgi–Dunitz attack occurs with equal probability on the two faces of the planar carbonyl (Re and Si faces are equally accessible), giving a racemic mixture — (±)-butan-2-ol, 50:50 R and S. This is the essential limiting-case check: whenever no chiral catalyst, chiral auxiliary, or pre-existing stereocenter is present anywhere in the substrate or reagent, addition to a planar sp² carbonyl is necessarily non-stereoselective, and a racemate results even though the product is chiral.

Example 2 — Applied: a two-step synthesis of a named tertiary alcohol, triphenylmethanol, from an ester intermediate check. A student wants to make triphenylmethanol, Ph₃C–OH, and proposes: react phenylmagnesium bromide (2.2 equivalents) with methyl benzoate, PhCO₂CH₃, in ether, then work up with aqueous acid. Verify this route mechanistically and explain why two equivalents of Grignard reagent are required (a case that also foreshadows why esters behave differently from aldehydes/ketones in this reaction family). Mechanistic walk-through, first addition. Methyl benzoate's carbonyl carbon is flanked by phenyl and by –OCH₃ (the ester oxygen donates a lone pair into the carbonyl π system by resonance, which is why esters are less electrophilic than simple ketones — a lower initial rate, consistent with $k_{\text{ketone}} \gg k_{\text{ester}}$ stated earlier). Even so, one equivalent of PhMgBr attacks: curved arrow from the Ph–Mg bonding pair to the carbonyl carbon, curved arrow from C=O π electrons to oxygen, giving a tetrahedral alkoxide bearing Ph (original), Ph (newly delivered), OCH₃, and O⁻MgBr — four groups on one carbon, one of which is a leaving group (OCH₃⁻, stabilized as methoxide). The collapse step (unique to esters, absent for simple ketones). Because this tetrahedral intermediate carries a genuine leaving group (–OCH₃, unlike a simple ketone alkoxide, which has no leaving group and is a dead-end product), it collapses: a curved arrow from the alkoxide oxygen's lone pair reforms the C=O π bond, and simultaneously a curved arrow from the C–OCH₃ σ bond kicks out methoxide, Mg-coordinated, as a leaving group. This regenerates a new carbonyl compound — specifically, benzophenone, Ph₂C=O — in situ. Second addition. Benzophenone is itself a ketone, fully electrophilic, and reacts with the second equivalent of PhMgBr exactly as in the general mechanism: curved arrow from Ph–Mg bond to carbonyl carbon, curved arrow from C=O π electrons to oxygen, giving the alkoxide Ph₃C–OMgBr. Workup.

Result. Triphenylmethanol, Ph₃C–OH, is obtained, but the stoichiometry genuinely requires two equivalents of PhMgBr per equivalent of ester, because the first equivalent is consumed converting the ester into a ketone (releasing methoxide as a byproduct, not incorporated into the product), and only the second equivalent performs the "final" addition that builds the alcohol. This is mechanistically distinct from simple ketone or aldehyde addition (only one equivalent needed per new C–C bond) precisely because the ester tetrahedral intermediate has an expellable leaving group that the ketone/aldehyde tetrahedral intermediate lacks — the central reason esters are treated as a separate reaction class even though the initial attack step looks identical.

Example 3 — Exam/competition level: diastereoselective addition to a chiral cyclohexanone (Felkin–Anh-flavored reasoning). Methylmagnesium bromide is added to 4-tert-butylcyclohexanone. Predict which face of the carbonyl is attacked preferentially, name the major diastereomer product, and justify using a conformational/steric argument grounded in the Bürgi–Dunitz trajectory. Setting up the conformational picture. 4-tert-Butylcyclohexanone adopts, overwhelmingly, the chair conformation with the bulky tert-butyl group equatorial (its A-value, ≈ 4.9–5.0 kcal/mol, is so large that the axial conformer is negligibly populated — essentially a conformational lock). With t-Bu equatorial, the ring carbon bearing it is fixed, and the carbonyl carbon (C1) sits across the ring with two distinguishable faces of attack: nucleophilic addition can occur from the axial trajectory (nucleophile approaches perpendicular-ish to the mean ring plane, ending up axial in the product, pushing the new alkoxide equatorial) or from the equatorial trajectory (nucleophile approaches along a flatter angle, ending up equatorial, pushing the alkoxide axial). Torsional and steric analysis (why axial attack is generally preferred for small nucleophiles on unhindered cyclohexanones). Equatorial attack, though seemingly less hindered by 1,3-diaxial interactions, actually suffers greater torsional strain in the transition state: the incoming nucleophile's trajectory along the equatorial-like path eclipses the C2–C3 (and C6–C5) ring bonds' substituents more severely as the carbon rehybridizes from sp² toward sp³. Axial attack, by contrast, follows a trajectory that is roughly staggered with respect to the neighboring ring C–H bonds as rehybridization proceeds — a lower-torsional-strain pathway. For a small, unhindered nucleophile like methylmagnesium bromide attacking an unhindered cyclohexanone (no adjacent axial substituents blocking the axial trajectory at C2/C6), this torsional argument dominates, and axial attack is preferred. Curved-arrow mechanism, axial trajectory. The methyl carbanion (from CH₃MgBr) approaches C1 from the axial direction, roughly along the trajectory that becomes the new axial C–CH₃ bond; one curved arrow from the Mg–C bonding pair to C1, one curved arrow from the ring C=O π electrons to the oxygen, generating the alkoxide, which ends up in the equatorial position (since the incoming group took the axial position, geometry dictates the oxygen substituent is pushed equatorial). Result. The major product, after aqueous workup, is 1-methyl-4-tert<strong>-butylcyclohexan-1-ol with the methyl group axial and the OH group equatorial — i.e., the trans relationship between the (larger, oxygen-bearing) OH and the t-Bu group is not the descriptor typically used here; more precisely, chemists report this as the isomer with OH equatorial (the thermodynamically expected orientation for the new, larger substituent group after axial-trajectory attack). This is the classically cited outcome in Grignard additions to 4-tert-butylcyclohexanone with small nucleophiles: axial attack dominates, delivering the alkyl group axial and the hydroxyl equatorial, typically in a ratio around 9:1 to 10:1 (methyl case) favoring this diastereomer over the alternative (OH axial, methyl equatorial) — illustrating that even without any classical "stereocenter created from scratch" ambiguity, the ring already fixes three-dimensional space, and the new C–C bond's face-selectivity is governed by transition-state torsional and steric effects layered directly on top of the Bürgi–Dunitz addition trajectory.

Victor Grignard (1871–1935) developed the magnesium-mediated carbon-carbon bond formation that bears his name while working under Philippe Barbier at the University of Lyon, building on Barbier's earlier, less reliable magnesium-mediated coupling of allyl halides with ketones. Grignard's key insight, published in 1900, was to prepare the organomagnesium reagent cleanly and separately, in dry ether, before introducing the carbonyl electrophile — turning an erratic one-pot procedure into a controllable two-step protocol that any competent chemist could reproduce. This procedural discipline, more than any single new bond type, is what made the reaction explode in usefulness: within a decade "Grignard reagents" were a standard tool across organic laboratories worldwide, used for natural product synthesis, structure elucidation via degradation, and eventually industrial-scale alcohol and pharmaceutical manufacture. Grignard shared the 1912 Nobel Prize in Chemistry with Paul Sabatier, cementing organometallic chemistry as a permanent, central pillar of synthetic method.