Jacobsen Epoxidation

What the Jacobsen epoxidation does

An ordinary alkene has two identical-looking faces. Add an oxygen atom across the double bond and you get an epoxide — but which of the two mirror-image epoxides you get is a coin flip unless something biases the delivery. The Jacobsen-Katsuki epoxidation is the reaction that rigs the coin. Feed it a plain hydrocarbon cis-alkene, a scoop of a chiral manganese-salen catalyst, and a bottle of bleach, and it hands you back one enantiomer of the epoxide with 90-98% enantiomeric excess.

The killer feature is that the alkene needs no functional handle. Its famous predecessor, the Sharpless epoxidation, only works when the alkene carries an allylic -OH that can grip the titanium catalyst. Millions of interesting alkenes have no such hydroxyl. Jacobsen's manganese oxo doesn't care — it recognizes the alkene by shape alone, squeezing it into a chiral pocket built from tert-butyl groups. That single advance opened enantioselective epoxidation to substrates like cis-β-methylstyrene, indene, dihydronaphthalene, and the chromenes used to make potassium-channel-opener drugs.

                    (S,S)-Mn-salen (cat.)
    cis-alkene  +  [O]  ──────────────────────→   chiral epoxide (one enantiomer)
                    NaOCl, CH₂Cl₂, 0-4 °C, pH ~11

  e.g.   cis-β-methylstyrene  →  (1R,2S)-1-phenylpropylene oxide   (up to 86% ee)

The mechanism, step by step

The catalyst you weigh out — Mn(III)-salen chloride — is not the oxidant. It is a precatalyst that has to be charged up by the terminal oxidant before anything happens.

1. Oxo formation. The terminal oxidant (bleach: OCl⁻; or PhIO, m-CPBA, oxone) transfers a single oxygen atom to the manganese. The metal climbs two oxidation states, from Mn(III) to a high-valent Mn(V)=O — an electron-poor, electrophilic oxomanganese(V) species. This is the true reagent. A base and additives (4-phenylpyridine N-oxide) axially ligate the metal and stabilize this oxo.
2. Face-selective approach. The alkene drifts toward the Mn=O along a skewed, side-on trajectory — not straight down onto the oxygen. The folded, stepped salen ligand walls off three of four approach paths with its 3,3'-tert-butyl groups. Only one prochiral face of a cis-alkene can slide in with its substituents pointing into open space. Everything downstream inherits this choice.
3. Oxygen transfer (stepwise, not fully concerted). The electrophilic oxo attacks one alkene carbon. The best current evidence (from the observed loss of cis/trans stereochemistry) says a short-lived intermediate forms first — either a carbon radical on the other carbon, held next to a Mn(IV)-O, or a manganaoxetane (a four-membered Mn-O-C-C ring). This intermediate lives just long enough to rotate slightly before closing.
4. Ring closure. The C-O bond forms to the second carbon, snapping shut the three-membered epoxide ring and releasing the product. The manganese drops back to Mn(III), ready to be re-oxidized by the next equivalent of bleach.

  step 1:   Mn(III)-salen  +  OCl⁻  →  [Mn(V)=O]-salen  +  Cl⁻     (charge the catalyst)

  step 2:   alkene approaches Mn=O side-on over the chiral pocket → one face selected

  step 3:   Mn(V)=O  +  C=C  →  [ •C–C–O–Mn(IV) ]  (radical / metallaoxetane, brief)

  step 4:   ring closes  →   epoxide  +  Mn(III)-salen   (catalyst regenerated)

The stepwise intermediate in step 3 is why some cis-alkenes give a little of the trans-epoxide — the radical carbon can rotate about the former double bond before the ring shuts. That partial scrambling is the fingerprint that distinguishes the Mn(V)=O radical-rebound pathway from a simple concerted peracid epoxidation, which is always 100% stereospecific.

The catalyst, oxidant, and conditions

The catalyst. "Jacobsen's catalyst" is the manganese(III) complex of a chiral salen ligand — a tetradentate N₂O₂ ligand made by condensing one equivalent of enantiopure trans-1,2-diaminocyclohexane with two equivalents of 3,5-di-tert-butylsalicylaldehyde, then metalating with Mn(II)/air and adding chloride. The cyclohexane backbone sets the absolute configuration; the 3,3'-tert-butyl groups build the walls of the chiral pocket; the 5,5'-tert-butyl groups add solubility and electron density. Both enantiomers — (R,R) and (S,S) — are sold off the shelf, and they give opposite epoxide enantiomers.

The oxidant. The cheapest and most common is household bleach (aqueous NaOCl, buffered to about pH 11 with Na₂HPO₄/NaOH). Alternatives when bleach is inconvenient: m-CPBA + NMO at low temperature, iodosylbenzene (PhIO), oxone, or hydrogen peroxide with the right additive. Bleach makes it a two-phase reaction: catalyst and substrate in dichloromethane, oxidant in the aqueous layer, oxygen shuttling across the interface.

Additives. A donor ligand such as 4-phenylpyridine N-oxide (or pyridine N-oxide, N-methylmorpholine N-oxide) binds the axial site of manganese. It both stabilizes the reactive Mn(V)=O and improves ee and turnover. Typical loadings: 1-8 mol% catalyst, 0.2-0.5 equiv N-oxide additive, run at 0-4 °C to preserve the fragile epoxide.

Conditions in one line: substrate + 2-8 mol% (R,R)-Mn-salen + ~0.2 equiv 4-PhPyNO in CH₂Cl₂, add buffered NaOCl (pH 11.3) at 0-4 °C, stir 2-12 h, watch by TLC. Yields are usually 60-90% with 88-98% ee for good cis substrates.

Scope, selectivity, and stereochemistry

The Jacobsen epoxidation lives and dies by alkene geometry, because the enantioselectivity is entirely a matter of how the alkene threads the chiral pocket.

• cis-disubstituted alkenes — the sweet spot. cis-β-methylstyrene, indene, 2,2-dimethylchromenes, and 1,2-dihydronaphthalene give 88-98% ee. These are the substrates the reaction was built for.
• Trisubstituted alkenes — often good, especially conjugated ones like the chromenes (which have a cis-configured ring double bond).
• trans-alkenes — poor. Both faces clash with the tert-butyl walls almost equally, so ee drops to 20-60%. Use Shi's dioxirane epoxidation instead for trans-alkenes.
• Terminal and 1,1-disubstituted alkenes — poor to moderate ee. There's simply not enough steric differentiation between the two faces.
• Electron-rich alkenes (styrenes, enol ethers, dienes) react fastest, because the electrophilic Mn=O prefers a nucleophilic double bond.

Which enantiomer you get is set by the catalyst you choose. The (S,S)-salen catalyst and the (R,R)-salen catalyst deliver oxygen to opposite prochiral faces. So to switch product handedness you don't change substrate or oxidant — you just reach for the mirror-image ligand off the shelf. That predictability, plus a working mnemonic model for the approach trajectory, is a big part of why the reaction is trusted in synthesis.

Jacobsen vs Sharpless vs Shi epoxidation

	Jacobsen-Katsuki	Sharpless-Katsuki	Shi epoxidation
Catalyst	Chiral Mn(III)-salen	Ti(OiPr)₄ + (+)/(−)-DET tartrate	Fructose-derived chiral ketone → dioxirane
Active oxidant	Mn(V)=O oxo	Ti-peroxide (from TBHP)	Dioxirane (from ketone + oxone)
Terminal oxidant	NaOCl, m-CPBA, PhIO	t-BuOOH (TBHP)	Oxone (KHSO₅)
Alkene requirement	None — unfunctionalized OK	Must be an allylic alcohol	None — unfunctionalized OK
Best geometry	cis-disubstituted	Allylic alcohols of any geometry	trans and trisubstituted
Typical ee	90-98% (cis)	90-99%	90-98% (trans)
Mechanism	Stepwise (radical / metallaoxetane)	Concerted, metal-templated	Concerted spiro/planar TS
Handedness switch	Use (R,R) vs (S,S) ligand	Use (+) vs (−)-tartrate	Use ketone vs ent-ketone
Nobel	—	Sharpless, 2001	—

These three methods are deliberately complementary. Sharpless handles allylic alcohols; Jacobsen handles cis-alkenes with no functionality; Shi covers trans- and trisubstituted alkenes. Between them a synthetic chemist can epoxidize almost any alkene enantioselectively.

Worked example: cis-β-methylstyrene → chiral epoxide

Make (1R,2S)-1-phenylpropylene oxide, a textbook Jacobsen substrate that gives high ee.

    Ph–CH=CH–CH₃  ──(S,S)-Mn-salen (4 mol%), 4-PhPyNO (0.2 eq)──→  epoxide
    (cis)              NaOCl (buffered pH 11.3), CH₂Cl₂, 4 °C, 4 h        ~86% yield, ~86% ee

• Reagents. cis-β-methylstyrene 1.0 equiv; (S,S)-Mn(III)-salen chloride 4 mol%; 4-phenylpyridine N-oxide 0.2 equiv (axial ligand / stabilizer); commercial bleach diluted and buffered with Na₂HPO₄/NaOH to pH ~11.3, 2 equiv active oxygen.
• Setup. Biphasic: alkene + catalyst + additive in CH₂Cl₂ (organic layer); buffered NaOCl added as the aqueous layer; stir vigorously at 0-4 °C.
• Course. The brown Mn(III) precatalyst darkens as the Mn(V)=O forms; the reaction is followed by TLC. 2-6 h is typical.
• Workup. Separate layers, wash organics, filter through silica to remove manganese, concentrate. The chiral epoxide is isolated by chromatography; ee is measured by chiral GC or HPLC.
• Outcome. ~86% isolated yield, ~86% ee. A small amount of trans-epoxide (a few percent) appears — the diagnostic evidence of the stepwise radical intermediate.

Note that the same substrate run with the mirror-image (R,R)-catalyst gives the (1S,2R)-epoxide instead, with essentially the same yield and ee — one reaction, two switchable products.

Real application: the Indinavir / chromene route

The reaction's most cited industrial-scale showcase is the synthesis of chiral chromene epoxides and cis-aminoindanol. The cis-aminoindanol fragment is the key chiral building block of Indinavir (Crixivan), Merck's HIV protease inhibitor. It is made by Jacobsen epoxidation of indene to the enantioenriched indene oxide, which is then opened regio- and stereoselectively (Ritter-type) to install the amino alcohol. Running the epoxidation on plain indene — a completely unfunctionalized cyclic cis-alkene — is exactly what no earlier asymmetric method could do.

The other landmark is the chromene / benzopyran series. 2,2-Dimethylchromenes epoxidized with Jacobsen's catalyst give chiral chromene oxides that are the cores of potassium-channel-opener drugs such as cromakalim and its relatives (BRL-55834, EMD-type antihypertensives). Here the Mn-salen system delivers >95% ee on gram-to-kilogram scale, and the resulting epoxide is opened by an amine or by a pyridone to build the drug.

Limitations and side reactions

• Geometry-limited. The reaction only shines on cis-disubstituted (and some cis-configured trisubstituted) alkenes. trans-, terminal, and 1,1-disubstituted alkenes give mediocre ee. Reach for Shi (dioxirane) epoxidation for trans-alkenes.
• Epoxide stability. Styrene-derived and benzylic epoxides can racemize, ring-open, or rearrange under the basic aqueous conditions if the reaction is over-run. Keep it cold (0-4 °C), quench promptly, don't let it sit.
• Over-oxidation. Electron-rich alkenes and dienes can suffer double epoxidation or oxidative cleavage. Control equivalents of oxidant.
• Catalyst decomposition. The salen ligand slowly oxidatively degrades (the phenol rings and the imine bridge are vulnerable), capping practical turnover numbers. The 3,5-di-tert-butyl groups and axial N-oxide additives are there partly to slow this.
• Manganese removal. The colored manganese has to be scrubbed out of the product (silica filtration, extraction), which matters at pharmaceutical scale for metal-residue limits.
• Biphasic finicking. Bleach epoxidations need pH control and vigorous stirring; too-low pH kills the catalyst, too-high pH slows oxo formation.

Who discovered it, and the salen story

In 1990 two groups published the same breakthrough within months. Eric N. Jacobsen, then a young assistant professor at the University of Illinois at Urbana-Champaign (later Harvard), and Tsutomu Katsuki at Kyushu University in Japan, independently found that manganese complexes of chiral salen ligands epoxidize unfunctionalized alkenes enantioselectively. The reaction is properly the Jacobsen-Katsuki epoxidation, though "Jacobsen epoxidation" and "Jacobsen's catalyst" became the everyday shorthand because Jacobsen's specific ligand — from trans-1,2-diaminocyclohexane and 3,5-di-tert-butylsalicylaldehyde — became the commercial standard.

The word salen is a contraction of salicylaldehyde + en (ethylenediamine), the two pieces of the original ligand family, first made by Pfeiffer in the 1930s. Jacobsen's insight was that a folded, C₂-symmetric salen with the right steric wall could make a manganese-oxo see the difference between the two faces of a plain alkene. The chemistry later grew into the enormously useful Jacobsen hydrolytic kinetic resolution (cobalt-salen splitting racemic terminal epoxides with water) — a related but distinct reaction that solved the terminal-epoxide problem this method could not.

Industrial and safety notes

• Bleach is the green oxidant. A large practical appeal of the method is that the terminal oxidant is cheap, aqueous NaOCl — the byproduct is NaCl and water, not heavy-metal or halogenated waste. That said, NaOCl liberates Cl₂ if acidified, so pH control is a safety issue, not just a yield issue.
• Low temperature is mandatory, both for ee and for epoxide survival. Bench runs sit in an ice bath; scale-up uses jacketed reactors at 0-4 °C.
• Metal residues. Manganese is far less toxic and less regulated than, say, palladium or osmium, which helped the reaction reach pharmaceutical process chemistry. Still, ICH residual-metal limits mean a scrub step is designed in.
• Chiral epoxides are versatile intermediates. The product epoxide is rarely the end goal — it's a springboard, opened by amines, azide, alcohols, or organometallics to make enantiopure amino alcohols, diols, and 1,2-difunctional chains found across drug and agrochemical portfolios.