The Reformatsky Reaction

What the Reformatsky reaction does

You want a β-hydroxy ester — a molecule with an ester on one end and a hydroxyl group exactly two carbons away. The obvious retrosynthesis is to add the α-carbon of an ester (as a nucleophile) to the carbonyl of an aldehyde or ketone. The trouble is generating that nucleophile without destroying the ester it lives on. Deprotonate with a strong base and you risk an uncontrolled aldol; make a Grignard from an α-halo ester and the carbanion instantly eats its own ester carbonyl.

The Reformatsky reaction solves this by using zinc as the metal. Zinc inserts into the carbon–halogen bond of an α-halo ester (typically an α-bromo ester like ethyl bromoacetate) to give an organozinc reagent that exists as a stabilized zinc enolate. This enolate is nucleophilic enough to add to an aldehyde or ketone, but so mild that it ignores esters, nitriles, and other reducible groups. After the addition and an aqueous workup, you have a β-hydroxy ester with a brand-new carbon–carbon bond.

    BrCH₂CO₂Et  +  Zn   ──THF──→   [BrZn–CH₂CO₂Et]   (zinc enolate)
                                          │
                                          │  + R–CHO
                                          ▼
                              R–CH(OZnBr)–CH₂CO₂Et
                                          │  H₃O⁺ workup
                                          ▼
                              R–CH(OH)–CH₂CO₂Et       (β-hydroxy ester)

The single most important idea: zinc is deliberately weak. Everything good about this reaction — the chemoselectivity, the tolerance of the ester, the ability to isolate and store the reagent — flows from choosing a metal that makes a soft, non-basic organometallic instead of a hard carbanion.

The step-by-step mechanism

The reaction runs in two distinct phases: forming the zinc enolate, then adding it to the carbonyl.

1. Oxidative insertion. Metallic zinc inserts into the C–Br bond of the α-halo ester. Zn(0) is oxidized to Zn(II) and a carbon–zinc bond forms. The result is an α-zinc ester that is best described not as a carbanion but as a zinc enolate: the negative charge is delocalized onto the carbonyl oxygen, and zinc sits mostly on that oxygen (a C-bound ↔ O-bound equilibrium favoring O). This delocalization is why the reagent is stable and storable — the "carbanion" is spread over three atoms.
2. Coordination of the carbonyl. The Lewis-acidic zinc coordinates the lone pair of the incoming aldehyde or ketone oxygen. This does two jobs at once: it activates the carbonyl toward nucleophilic attack, and it pre-organizes the two partners into a six-membered ring. Both the enolate oxygen and the electrophile oxygen bind the same zinc.
3. C–C bond formation through a chair transition state. The nucleophilic α-carbon of the enolate attacks the electrophilic carbonyl carbon. Because zinc bridges both oxygens, the attack happens through a Zimmerman–Traxler six-membered chair transition state — the same closed transition state that controls aldol stereochemistry. New C–C σ bond forms; the carbonyl π bond breaks; the former carbonyl oxygen becomes a zinc alkoxide.
4. Aqueous workup. The zinc alkoxide (an –O–ZnBr) is protonated by dilute acid (NH₄Cl or dilute H₂SO₄) to liberate the free hydroxyl. Zinc leaves as a soluble salt. You are left with the β-hydroxy ester.

The electron-arrow logic for the key C–C bond-forming step: the α-carbon's electron pair (drawn from the enolate π system) reaches to the carbonyl carbon; the C=O π electrons collapse onto oxygen, which is stabilized by its bond to zinc. Because zinc holds both oxygens, this is an intramolecular-feeling delivery even though two molecules are involved — the reason it is so clean.

        chair (Zimmerman–Traxler) transition state:

              O⋯⋯Zn⋯⋯O
             ╱          ╲
        EtO–C            C–R           Zn bridges BOTH oxygens;
             ╲          ╱              α-C attacks carbonyl C
              C ······ C              through the 6-membered ring
             (α)      (was C=O)

Reagents, activation, and conditions

• The zinc. Zinc dust or granular zinc, almost always activated to strip the passivating ZnO surface layer. Common activators: a trace of I₂, 1,2-dibromoethane, TMSCl, or Zn–Cu couple. Rieke zinc (highly reactive Zn made by reducing ZnCl₂ with potassium) needs no activation and drives sluggish substrates. Ultrasound (sonication) is a popular modern trick: cavitation continuously scrubs the metal surface.
• The α-halo ester. α-Bromo esters are the standard (ethyl bromoacetate, ethyl 2-bromopropanoate). α-Iodo esters are more reactive but pricier; α-chloro esters are usually too unreactive. The halogen must be on the carbon α to the ester carbonyl — that is what stabilizes the resulting enolate.
• The electrophile. Aldehydes react fastest; ketones are slower and more prone to competing side reactions; even esters, imines (giving β-amino esters, the Blaise-adjacent chemistry), and nitriles (the Blaise reaction proper) can serve as electrophiles in variants.
• Solvent. Anhydrous ether, THF, benzene, or toluene. THF is now standard. Rigorously dry — the organozinc is protonated by water back to the plain ester, killing yield.
• Temperature. Often reflux to initiate insertion, then the addition can proceed at or below room temperature. Slow, controlled initiation is the single biggest factor in getting clean product.

Scope, chemoselectivity, and stereochemistry

Chemoselectivity is the headline. Because the zinc enolate is non-basic and mild, functional groups that would be destroyed by a Grignard survive: esters (obviously, since the reagent itself carries one), nitriles, nitro groups, and even some ketones under controlled conditions. This lets chemists carry an ester through a step that installs an alcohol elsewhere in the molecule — a maneuver impossible with RMgX.

Regiochemistry is fixed by the reagent: the new C–C bond always forms at the α-carbon of the ester (the carbon that bore the halogen), and the hydroxyl lands on the former carbonyl carbon. That defines the "β-hydroxy" relationship every time.

Stereochemistry. When both partners bear substituents (e.g. an α-substituted bromo ester adding to an aldehyde), two new stereocenters form and syn/anti diastereomers are possible. The Zimmerman–Traxler chair biases the outcome, but simple substrates give only modest diastereoselectivity. Asymmetric versions use chiral amino-alcohol ligands on zinc or chiral ester auxiliaries; the Honda and Cozzi groups reported catalytic enantioselective Reformatsky reactions reaching >90% ee. There is no absolute stereocontrol without such a control element.

Reformatsky vs related C–C bond-forming methods

	Reformatsky	Aldol (base-mediated)	Grignard addition
Nucleophile	Zinc enolate from α-halo ester	Enolate from α-deprotonation	R–MgX carbanion
How it's generated	Zn inserts into C–Br (neutral)	Strong base (LDA, NaOEt)	Mg inserts into C–X
Tolerates an ester on the nucleophile?	Yes — that's the point	Risky (self-condensation)	No — destroys it
Basicity of reagent	Low (mild, non-basic)	High	Very high
Typical product	β-hydroxy ester	β-hydroxy carbonyl (aldol)	Alcohol (from RCHO/ketone)
Chemoselectivity	Excellent	Moderate	Poor
Reagent stability	Isolable/storable enolate	Must use in situ	Must use in situ
Main failure mode	Slow initiation → self-coupling	Aldehyde self-aldol, dehydration	Attacks own carbonyl

Worked example: ethyl bromoacetate + benzaldehyde

The classic teaching example makes ethyl 3-hydroxy-3-phenylpropanoate.

    PhCHO  +  BrCH₂CO₂Et  ──Zn (activated), THF, reflux then RT──→
                                             │  H₃O⁺ workup
                                             ▼
                          Ph–CH(OH)–CH₂–CO₂Et   (β-hydroxy ester)

• Reagents. Benzaldehyde 1.0 equiv, ethyl bromoacetate 1.2 equiv, zinc dust 1.5 equiv activated with a crystal of I₂ (or use Rieke zinc / sonication).
• Procedure. Reflux the zinc and a small portion of the bromoacetate in dry THF to initiate (a color change and mild exotherm signal insertion). Then add the remaining bromoacetate and the benzaldehyde slowly to keep the α-halo ester concentration low — this suppresses self-condensation and Wurtz coupling.
• Workup. Quench with saturated NH₄Cl (or dilute H₂SO₄), extract, dry, and distill or chromatograph.
• Yield. Typically 60–80% for benzaldehyde; aromatic aldehydes behave well, hindered ketones lower the yield.
• Downstream. Dehydrate the β-hydroxy ester (e.g. with acid or via the mesylate/elimination) to make ethyl cinnamate, an α,β-unsaturated ester — the Reformatsky is a common on-ramp to cinnamates and to β-lactones/β-lactams.

Limitations and side reactions

• Self-condensation. If insertion is slow, unreacted α-halo ester piles up and the zinc enolate attacks its ester carbonyl instead of the aldehyde — giving a β-keto ester dimer. The cure is fast initiation (Rieke/activated zinc, sonication) and slow addition of the halo ester.
• Wurtz-type homocoupling. Two α-halo ester molecules can couple to a succinate diester. Same root cause (excess halo ester), same fix.
• Moisture sensitivity. Any water protonates the organozinc back to the parent ester, cutting yield. Dry glassware, dry solvent, inert atmosphere.
• Sluggish or capricious initiation. The passivating ZnO layer makes the reaction notoriously temperamental to start. Rieke zinc and ultrasound largely tamed this in the modern era.
• Ketone electrophiles. Ketones are less reactive and more hindered, so yields drop and elimination/enolization side paths appear. Aldehydes are the reliable partners.
• Dehydration. β-Hydroxy esters can dehydrate under acidic workup or on heating; sometimes desirable (cinnamates) but a nuisance if you want the alcohol.

Variants and named relatives

• Blaise reaction. Swap the aldehyde/ketone for a nitrile. The zinc enolate adds to the nitrile carbon to give, after workup, a β-keto ester (via a β-enamino ester intermediate). Same zinc-enolate nucleophile, different electrophile.
• Reformatsky with imines. Using an imine (or in situ-generated iminium) as electrophile gives β-amino esters, a direct route to β-lactams (four-membered antibiotic rings) after cyclization.
• Honda/Cozzi asymmetric Reformatsky. Chiral amino-alcohol ligands on zinc deliver enantioenriched β-hydroxy esters, some catalytic in the chiral controller.
• Rieke-zinc Reformatsky. Highly activated zinc metal makes the reaction fast and reliable, extending scope to hindered and less reactive substrates.
• Indium and samarium variants. SmI₂-mediated and indium-mediated Reformatsky-type reactions run under even milder, sometimes aqueous conditions, trading zinc for a different single-electron or organometallic donor.

Historical discovery

The reaction is named for Sergei Nikolaevich Reformatsky, a Russian chemist working at Kazan University, who reported it in 1887. He found that zinc and an α-halo ester react with aldehydes and ketones to give what we now recognize as β-hydroxy esters. The insight was choosing zinc — mild enough to make a usable organometallic that coexists with an ester — decades before the electronic reasons (enolate stabilization, oxophilic Zn(II)) were understood. The reaction predates Victor Grignard's organomagnesium chemistry (1900) and remained, for years, one of the few reliable ways to form a carbon–carbon bond to a carbonyl while carrying an ester group. It is still taught as the archetype of chemoselective organometallic addition and remains in active use, especially in β-lactam and polyketide-fragment synthesis.