The Swern Oxidation

What the Swern oxidation does

The Swern oxidation is the go-to way to nudge an alcohol up exactly one oxidation level — a primary alcohol to an aldehyde, a secondary alcohol to a ketone — and then stop dead. That "and then stop" is the whole point. Dunk a primary alcohol in aqueous chromic acid or KMnO₄ and it will not rest at the aldehyde: the aldehyde picks up water to form a gem-diol (a hydrate), and that hydrate is itself an alcohol the metal happily oxidizes onward to the carboxylic acid. The Swern never goes there.

The trick is that the oxidizing atom is sulfur, not oxygen or a transition metal. Dimethyl sulfoxide (DMSO, Me₂S=O) is the oxidant; it net removes two hydrogens from the alcohol through a sulfur-tethered intermediate and departs as the reduced thioether, dimethyl sulfide (Me₂S), while the alcohol's own oxygen stays behind as the new carbonyl. Because DMSO is delivered in a discrete, stoichiometric package that consumes itself, there is no lingering oxidant in the flask to touch the newly minted aldehyde. No metals means no chromium waste and no metal residues in a pharmaceutical intermediate — a genuine practical advantage.

    R-CH₂-OH   ──DMSO, (COCl)₂; then Et₃N, −78 °C──→   R-CHO      (primary → aldehyde)
    R₂CH-OH    ──DMSO, (COCl)₂; then Et₃N, −78 °C──→   R₂C=O      (secondary → ketone)

    net byproducts:  Me₂S  +  CO  +  CO₂  +  Et₃N·HCl

The mechanism, arrow by arrow

There are four movements, and each one is a clean, textbook electron push. The choreography is why the reaction is so reliable.

1. Activate the DMSO. Oxalyl chloride is added to DMSO at −78 °C. The sulfoxide oxygen — a good nucleophile — attacks one carbonyl carbon of (COCl)₂. The collapse expels chloride, then the resulting acyloxysulfonium ion fragments, spitting out CO and CO₂ and leaving the key electrophile: the chlorodimethylsulfonium ion, [Me₂S–Cl]⁺ Cl⁻. This is the "activated DMSO." The gas evolution (CO + CO₂) is visible bubbling and confirms activation.
2. The alcohol attacks sulfur. The alcohol is added. Its oxygen lone pair attacks the electrophilic sulfur of [Me₂S–Cl]⁺, displacing chloride. A proton is lost (chloride or a second equivalent of DMSO takes it) to give the neutral-charged-at-oxygen but cationic-at-sulfur alkoxysulfonium salt: R₂CH–O–S⁺Me₂, Cl⁻. The alcohol's oxygen is now tethered to sulfur, and that C–O is one deprotonation away from becoming a carbonyl.
3. Base makes the ylide. Triethylamine is added. It is too bulky and too weak to rip the α-C–H off the alkoxide carbon directly. Instead it does something subtler: it deprotonates one of the sulfonium methyl groups. That gives a sulfur ylide — a carbanion on a CH₂ next to the positive sulfur, R₂CH–O–S⁺(Me)=CH₂ (equivalently the sulfonium ylide/betaine). This step is why the base has to be added last.
4. Intramolecular deprotonation → carbonyl. The ylide's carbanion is now perfectly placed to reach over and pluck the C–H on the alkoxysulfonium carbon through a five-membered cyclic transition state. As that C–H breaks, the C–O π bond forms, the S–O bond breaks, and dimethyl sulfide leaves. Out drops the aldehyde or ketone, and the sulfur exits as Me₂S carrying the extra methyl-plus-hydrogen as a new C–H bond.

  1. (COCl)₂ + Me₂S=O → [Me₂S-Cl]⁺ Cl⁻  + CO↑ + CO₂↑        (activated DMSO)
  2. R₂CH-OH + [Me₂S-Cl]⁺ → R₂CH-O-S⁺Me₂  Cl⁻ + HCl          (alkoxysulfonium salt)
  3. Et₃N deprotonates an S-CH₃  →  R₂CH-O-S⁺(Me)=CH₂         (sulfur ylide)
  4. cyclic 5-membered TS:  ylide C⁻ grabs the α-C-H  →  R₂C=O + Me-S-CH₃
                                                          (product + Me₂S)

The elegance is in step 4: the base never touches the carbinol carbon. A remote carbanion, generated in situ on the sulfur's own methyl, reaches back through a tidy ring to do the elimination. That intramolecularity is what makes the reaction fast even at −78 °C.

Reagents, order of addition, and conditions

The Swern is famously order-sensitive. Get the sequence wrong and you get a mess of methylthiomethyl ethers instead of your aldehyde.

• DMSO — ~2.0–2.5 equiv, dissolved in dry CH₂Cl₂, is added to the oxalyl chloride solution (not the other way round) at −78 °C. A vigorous evolution of CO and CO₂ signals activation.
• Oxalyl chloride — ~1.1–1.5 equiv. It is the activator of choice because it is fast, generates only gaseous byproducts, and works at cryogenic temperature. (COCl)₂ is moisture-sensitive and lachrymatory; keep everything anhydrous.
• The alcohol — 1.0 equiv, added dropwise after activation is complete (stir ~10–30 min at −78 °C to build the alkoxysulfonium salt).
• Triethylamine — ~4–5 equiv, added last. Then, and only then, warm the reaction to 0 °C or room temperature to complete the elimination.
• Solvent & temperature — anhydrous dichloromethane, dry-ice/acetone bath at −78 °C. Warming before the base is added is the classic beginner error.

Note that oxalyl chloride is stoichiometric, not catalytic — every alcohol molecule consumes its own DMSO and its own activator. The reaction is "reagent-heavy" but forgiving, and everything except the amine salt boils off in workup.

Scope, selectivity, and stereochemistry

The Swern is prized for being mild and chemoselective. Because it runs cold, fast, and without acid or metal, it leaves alone a long list of sensitive functionality:

• Acid-sensitive protecting groups survive: acetals, TBS/TBDPS silyl ethers, Boc carbamates, and trityl groups all come through intact — a stark contrast to acidic Cr(VI) conditions.
• Alkenes, alkynes, and epoxides are untouched — there is no electrophilic oxidant to attack a π bond, unlike ozone or peracids.
• Stereocenters not adjacent to the new carbonyl are preserved. The reaction is not itself stereoselective (it destroys the carbinol stereocenter it oxidizes), but it does not scramble other chirality.

The one stereochemical caveat: substrates prone to enolization can partially racemize at the α-carbon of the product, because the mildly basic Et₃N and the warming step give an enolizable aldehyde a chance to epimerize. For very base-sensitive, α-chiral aldehydes chemists often reach for the Dess-Martin periodinane instead, or run the Swern with a Hünig's-base variant and rigorous low temperature.

Swern vs other alcohol oxidations

	Swern (DMSO/(COCl)₂)	Dess-Martin (DMP)	PCC (Cr(VI))	Jones (CrO₃/H₂SO₄)
1° alcohol gives	Aldehyde	Aldehyde	Aldehyde	Carboxylic acid
Over-oxidation risk	None	None	Low	High (goes to acid)
Metal-free?	Yes	Yes (I(V))	No (Cr)	No (Cr)
Temperature	−78 °C	0–25 °C	25 °C	0–25 °C
Byproducts	Me₂S, CO, CO₂ (smelly, volatile)	Iodinane, AcOH	Cr(III) sludge	Cr(III) sludge
Acid/base-sensitive groups	Tolerated	Tolerated	Mildly acidic	Strongly acidic
Main drawback	Cryogenic; foul odor; α-racemization	Shock-sensitive reagent; cost	Toxic Cr; workup	Only for making acids
Scale	Gram to multi-kg (with care)	Lab scale	Lab scale	Lab scale

Worked example: a Swern in total synthesis

Take a common maneuver in synthesis — oxidizing a primary alcohol to an aldehyde so you can immediately do a Wittig or an aldol. Suppose you have a fragment bearing a free primary alcohol, a TBS-protected secondary alcohol, and an internal alkene, and you need the aldehyde without disturbing anything else.

    (COCl)₂ (1.3 eq), CH₂Cl₂, −78 °C
       add DMSO (2.5 eq) dropwise      → [Me₂S-Cl]⁺  (CO + CO₂ bubble off)
       stir 15 min, −78 °C
    add R-CH₂OH (1.0 eq), −78 °C, 30 min → R-CH₂-O-S⁺Me₂ Cl⁻
    add Et₃N (5 eq), −78 °C → warm to 0 °C → R-CHO  +  Me₂S
    quench: sat. NH₄Cl, extract, wash; use crude aldehyde directly

• Why Swern here: the TBS ether and the alkene both survive; a Cr(VI) reagent risks acid-mediated TBS loss, and permanganate would cleave the alkene.
• Yield: Swern oxidations of unhindered primary alcohols routinely run 85–95%. The aldehyde is usually carried forward crude because it is often too reactive (or too volatile) to purify cleanly.
• Signature detail: chemists watch for gas evolution on DMSO addition (activation) and add the amine only after the low-temperature hold, then warm.

Real-world footprint: the Swern appears in the published routes to countless natural products and drugs — it is one of the handful of oxidations a process chemist reaches for when metal residues are unacceptable and the aldehyde must not over-oxidize, e.g. in prostaglandin, macrolide, and polyketide fragment couplings.

Limitations and side reactions

• The Pummerer trap. If the activated DMSO is allowed to warm before the alcohol reacts, [Me₂S–Cl]⁺ collapses to chloromethyl methyl sulfide, ClCH₂SMe. That electrophile can then convert your alcohol into a methylthiomethyl (MTM) ether, R–O–CH₂–S–Me, instead of the aldehyde. Keeping the reaction at −78 °C until the base is in avoids it.
• The odor. Dimethyl sulfide is detectable at ppb levels and smells of rotting cabbage. It is not especially toxic at these scales but demands a good hood and a bleach quench. This alone limits how happily the reaction scales.
• α-Racemization. Enolizable aldehydes can epimerize under Et₃N during warmup. Use a bulkier, weaker base or keep it cold; or switch to Dess-Martin.
• Reagent load. The Swern consumes multiple equivalents of DMSO, oxalyl chloride, and amine — atom-economical it is not.
• Moisture. Oxalyl chloride and the sulfonium salt are water-sensitive; wet solvent kills the activation and can give sulfide byproducts.

Who discovered it, and when

The idea that DMSO can oxidize an alcohol goes back to Kornblum in the late 1950s (oxidizing alkyl halides and tosylates), and to Pfitzner and Moffatt, who in 1963 first oxidized alcohols directly using DMSO activated by a carbodiimide (DCC) — the Pfitzner-Moffatt oxidation. The carbodiimide route worked but left an annoying, hard-to-remove precipitate of dicyclohexylurea.

In 1978, Kanji Omura and Daniel Swern, working at Temple University, reported that oxalyl chloride is a superior activator: it is faster, works cleanly at −78 °C, and produces only volatile gaseous byproducts. Their paper (Tetrahedron 1978) turned a finicky method into a robust, general-purpose oxidation, and the reaction has carried Swern's name ever since. Daniel Swern was already a distinguished lipid chemist; this became his most-cited work.

Variants and the DMSO family

• Pfitzner-Moffatt (1963). DMSO + DCC (a carbodiimide) + a proton source (H₃PO₄ or TFA). The original; hampered by insoluble urea byproduct.
• Parikh-Doering (1967). DMSO activated by sulfur trioxide-pyridine complex (SO₃·py) with Et₃N. Runs at 0 °C rather than −78 °C — a milder, more scalable cousin, popular when cryogenics are inconvenient.
• Albright-Goldman. DMSO + acetic anhydride. Simple but can give MTM ethers as a stubborn side product.
• Corey-Kim. Uses N-chlorosuccinimide + dimethyl sulfide to generate the same chlorosulfonium electrophile without DMSO or oxalyl chloride — a useful workaround.
• TFAA variant. DMSO activated by trifluoroacetic anhydride; among the most reactive activators, good for hindered substrates.

All of them funnel through the same alkoxysulfonium salt and the same ylide-mediated syn-elimination. The choice of activator is really a choice of temperature, byproduct cleanliness, and how hindered your alcohol is.