The Barton-McCombie Deoxygenation

What it does, and why it is hard

Removing a hydroxyl group and putting a hydrogen in its place — turning R-OH into R-H — sounds trivial, but it is one of the harder transformations in classical organic chemistry. You cannot simply "reduce" an alcohol the way you reduce a ketone; the oxygen has no π bond to grab, and the C-O single bond is strong and inert. Ionic routes struggle because a hydroxide (or even a tosylate) is a poor enough leaving group that Sₙ2 with hydride (LiAlH₄) is sluggish on anything but the most activated substrates, and Sₙ1 ionization risks rearrangements, eliminations, and scrambling of neighboring stereocenters.

The Barton-McCombie deoxygenation sidesteps all of that by leaving ionic chemistry behind and going radical. The insight, published by Derek Barton and Stuart McCombie in 1975, is that you should not try to break the C-O bond directly. Instead, you decorate the oxygen with a thiocarbonyl group (C=S) that acts as a radical antenna: a tin radical grabs the sulfur, and that single event sets up a fragmentation that snaps the C-O bond from the far side, spitting out a carbon radical which is then quenched with a hydrogen atom. The whole thing runs at 80 °C in refluxing toluene or benzene under neutral, essentially anhydrous conditions — utterly gentle to acid- and base-sensitive functionality elsewhere in the molecule.

The reaction is two operations, not one

It is important to see that Barton-McCombie is always a two-step sequence:

1. Derivatize the alcohol into a thiocarbonyl. Convert R-OH into R-O-C(=S)-X, where X is most classically SMe (a xanthate), but can also be an imidazolyl group (from thiocarbonyldiimidazole) or an aryloxy group (from an aryl chlorothionoformate). This step installs the all-important C=S.
2. Run the radical reduction. Heat the thiocarbonyl derivative with a tin hydride (tributyltin hydride, Bu₃SnH) and a radical initiator (AIBN). The radical chain does the deoxygenation, leaving R-H.

  Step 1 (xanthate formation):
     R-OH  --NaH-->  R-O⁻  --CS₂-->  R-O-C(=S)-S⁻  --MeI-->  R-O-C(=S)-SMe

  Step 2 (radical deoxygenation):
     R-O-C(=S)-SMe  --Bu₃SnH, AIBN, PhMe, 80 °C-->  R-H  +  [MeS-C(=O)-S-SnBu₃ etc.]

Only the second step is the "Barton-McCombie" mechanism proper, but the two are inseparable in practice — the whole appeal is that the ordinary alcohol is the starting material.

The radical-chain mechanism, arrow by arrow

The reduction is a classic radical chain: an initiation step makes the first radical, then a two-part propagation cycle turns over many times per initiator molecule.

1. Initiation. AIBN (azobisisobutyronitrile) is warmed above ~65 °C and homolyzes, expelling N₂ and giving two 2-cyanoprop-2-yl radicals. Each of those abstracts the hydrogen atom from Bu₃Sn-H, generating a tributyltin radical, Bu₃Sn·. The weak Sn-H bond (bond dissociation energy ≈ 78 kcal/mol) makes this abstraction easy.
2. Tin adds to sulfur. The soft, thiophilic Bu₃Sn· radical adds to the sulfur of the C=S double bond (not the oxygen). This produces a carbon-centered radical on the thiocarbonyl carbon — a stabilized alkoxythiocarbonyl radical, R-O-C·(-SMe)(-S-SnBu₃). Tin's affinity for sulfur is the reason it targets S, not O.
3. β-Scission fragments the C-O bond. This is the key deoxygenation event. The radical on the thiocarbonyl carbon rearranges: the C-O bond to the substrate breaks homolytically, and the electron pair reorganizes so that the carbon now forms a strong C=O double bond. Out comes a carbon radical R· on the former alcohol carbon, and a new, very stable byproduct — an S-stannyl thiocarbonate/dithiocarbonate — where the C=O has formed. Forming a C=O (~180 kcal/mol) at the expense of a C-O single bond (~85 kcal/mol) is the thermodynamic engine that makes the fragmentation irreversible and drives the whole reaction forward.
4. Hydrogen delivery closes the deoxygenation. The naked carbon radical R· abstracts a hydrogen atom from a fresh molecule of Bu₃Sn-H. This is the step that actually installs the new C-H bond, giving the deoxygenated product R-H — and it simultaneously regenerates a new Bu₃Sn· radical, which re-enters step 2 to carry the chain.

  Initiation:   AIBN --Δ--> 2 (CH₃)₂C·(CN) + N₂
                (CH₃)₂C·(CN) + Bu₃Sn-H --> (CH₃)₂CH(CN) + Bu₃Sn·

  Propagation:  (a)  Bu₃Sn·  +  R-O-C(=S)-SMe  -->  R-O-C·(-SMe)(-S-SnBu₃)
                (b)  R-O-C·(-SMe)(-S-SnBu₃)   -->  R·  +  O=C(-SMe)(-S-SnBu₃)
                (c)  R·  +  Bu₃Sn-H           -->  R-H  +  Bu₃Sn·   (chain carries on)

Every part of the design is deliberate: the C=S antenna gives the tin radical a soft target; the fragmentation is powered by forming a C=O bond; and tin's weak Sn-H bond makes it a good hydrogen donor while its strong Sn-S and Sn-O bonds sop up the sulfur and oxygen fragments. Take away the thiocarbonyl and none of it happens — a plain alcohol is inert to Bu₃Sn·.

Reagents, initiators, and real conditions

• The thiocarbonyl "handle." Xanthate: treat the alcohol with NaH (or KH) to form the alkoxide, add CS₂, then quench with MeI (or, for hindered cases, use imidazole-based reagents). Thiocarbonylimidazolide: stir the alcohol with 1,1′-thiocarbonyldiimidazole (TCDI) — neutral, no strong base needed. O-Aryl thionocarbonate: react the alcohol with phenyl (or pentafluorophenyl) chlorothionoformate and DMAP/pyridine — the most reactive derivative, best for primary and hindered alcohols.
• Hydrogen donor / chain carrier. Tributyltin hydride (Bu₃SnH) is the classic. Typically 1.1-2.0 equiv, added slowly (or kept dilute) to favor productive H-transfer over side reactions.
• Initiator. AIBN, 5-20 mol%, decomposes cleanly at ~70-80 °C (half-life ~1 h at 80 °C). For lower temperatures use triethylborane/O₂ (Et₃B/air), which generates ethyl radicals at 0 °C or even −78 °C; for higher temperatures use di-tert-butyl peroxide or ACCN.
• Solvent and temperature. Refluxing toluene (111 °C) or benzene (80 °C) are standard; the reaction is run under inert atmosphere and typically anhydrous. Dilute conditions suppress premature H-abstraction of the tin radical.
• Concentration matters. Because step (b) — fragmentation — must outrun step (c) — H-abstraction — happening prematurely on the alkoxythiocarbonyl radical, keeping Bu₃SnH concentration low (slow addition) is a practical lever for sluggish (e.g. primary) substrates.

Scope, selectivity, and stereochemistry

The method's reach and its quirks both come from the fact that the pivotal intermediate is a free carbon radical.

• Secondary alcohols are the sweet spot. They fragment to secondary radicals, which are stable enough for clean, fast β-scission. This is the reason Barton-McCombie is the default deoxygenation of a secondary OH in complex-molecule synthesis.
• Primary alcohols are harder. Primary radicals are higher energy, so fragmentation is slower; these substrates usually need the more reactive thionocarbonate or imidazolide derivatives, higher temperatures, and careful control of tin-hydride concentration.
• Tertiary alcohols are usually the wrong tool. They form the most stable radical (good) but are hard to convert into a xanthate in the first place because the crowded tertiary carbon resists acylation on oxygen; other deoxygenation routes are typically used.
• Stereochemistry is scrambled at the deoxygenated carbon. A planar radical intermediate erases the stereocenter that was at the alcohol carbon. If that carbon ends up as CH₂, no problem. If it still bears three different substituents, the hydrogen is delivered to the less hindered face and the product diastereomer is set by the substrate's shape — often a feature, since it lets you install a ring-fusion H with predictable facial selectivity.
• Functional-group tolerance is excellent. Because the conditions are neutral and radical, esters, ethers, amides, ketals, and many protecting groups survive untouched. This is a major reason the reaction thrives in total synthesis.
• Watch for radical side reactions. An alkene positioned five or six atoms away from the carbon radical can be trapped by 5-exo/6-exo cyclization before the H-transfer — sometimes a nuisance, sometimes deliberately exploited to build a ring in the same pot (radical cascade deoxygenation-cyclization).

Barton-McCombie vs other ways to remove or reduce a C-O bond

	Barton-McCombie	Ionic (tosylate + LiAlH₄)	Clemmensen / Wolff-Kishner
Starting material	Alcohol (via xanthate)	Alcohol (via sulfonate)	Ketone / aldehyde (C=O)
Net transformation	R-OH → R-H	R-OH → R-H	R-C(=O) → R-CH₂
Mechanism	Radical chain (addition-fragmentation)	Sₙ2 hydride displacement	Ionic (acid/base, metal)
Conditions	Neutral, 80 °C, anhydrous	Basic, needs a good leaving group	Strong acid (Clemmensen) or strong base + heat (Wolff-Kishner)
Best substrate	Secondary alcohols	Primary / unhindered secondary	Carbonyls only
Stereocenter fate	Destroyed (planar radical)	Inverted (Sₙ2) or scrambled (Sₙ1)	N/A (was sp² carbonyl)
Rearrangement risk	Low (no cation)	High for 2°/3° (carbocation)	Low
Tolerates acid-sensitive groups?	Yes	Depends on base sensitivity	No (Clemmensen is strongly acidic)
Main drawback	Toxic tin, two steps	Elimination / rearrangement on hindered C	Only works on carbonyls

Worked example: deoxygenating a secondary alcohol on a sugar

A textbook application is removing a secondary hydroxyl from a carbohydrate to make a "deoxy sugar" — exactly the sort of substrate where acidic or basic conditions would wreck the acetals and glycosidic bonds. Take a partially protected methyl glycoside bearing one free secondary OH.

  Step 1 — make the xanthate:
     sugar-OH  +  NaH (1.1 eq), THF, 0 °C, 30 min
               +  CS₂ (excess), 30 min
               +  MeI (1.5 eq), 1 h  -->  sugar-O-C(=S)-SMe   (~85-95%)

  Step 2 — radical deoxygenation:
     sugar-O-C(=S)-SMe  +  Bu₃SnH (1.5 eq), AIBN (10 mol%),
                           toluene, 110 °C (reflux), 2-3 h  -->  deoxy-sugar   (~80-90%)

• Why it works here. The acetal and glycoside protecting groups are inert to the neutral radical conditions; only the derivatized carbon is touched.
• Practical notes. Degas the solvent (radicals are quenched by O₂), add AIBN in portions if the reaction stalls, and monitor for the disappearance of the xanthate (it has a diagnostic UV/TLC signature).
• Cleanup. The tin byproducts are the pain point: a KF or DBU workup, or filtration through a plug pre-treated to bind tin, is standard before chromatography.

Where it earns its keep: total synthesis

• Complex natural products. Barton-McCombie is a staple of the late-stage toolkit for taxol, erythromycin-type macrolides, steroids, prostaglandins, and countless terpenoids — anywhere a hydroxyl was useful for setting stereochemistry earlier in the route but is not wanted in the final skeleton. Deoxygenate it away once its job is done.
• Deoxy sugars and nucleosides. The synthesis of 2′-deoxy and 2′,3′-dideoxynucleosides (relevant to antiviral drugs) uses radical deoxygenation to remove ribose hydroxyls without disturbing the base or the glycosidic bond.
• Radical cascade construction. Because the fragmentation produces a carbon radical, chemists deliberately place an alkene nearby so the radical cyclizes onto it before H-transfer — building a new ring and removing an oxygen in a single operation.
• Correcting an oxidation state. When a synthesis installs an OH as a byproduct of a needed C-C bond-forming step (e.g., after an aldol or a Grignard addition that had to happen at a carbonyl), Barton-McCombie is the clean way to erase the leftover oxygen.

Limitations, side reactions, and modern fixes

• Organotin toxicity and purification. Bu₃SnH and its residues are toxic and cling to products. Fixes: tin-free hydrides such as tris(trimethylsilyl)silane (TTMSS), catalytic-tin protocols (Bu₃SnCl/NaBH₄ or PMHS regenerating the hydride in situ), and phosphorus-based donors like hypophosphorous acid (H₃PO₂) or its amine salt (EPHP) that carry the same chain with no tin at all.
• Premature reduction of the xanthate. If Bu₃SnH is too concentrated, the alkoxythiocarbonyl radical can grab a hydrogen before it fragments, returning a formate/thioformate instead of the deoxygenated product. Keeping the tin hydride dilute (slow addition) favors fragmentation.
• Unwanted radical cyclization or elimination. Nearby π systems can trap the carbon radical; strained systems can undergo ring-opening. These are predictable from the substrate and either avoided or exploited.
• Reluctant primary substrates. Primary radicals form slowly; switch to the more reactive O-aryl thionocarbonate or thiocarbonylimidazolide and raise the temperature.
• Oxygen sensitivity. Dissolved O₂ intercepts radicals and kills the chain — thorough degassing is mandatory.

Who discovered it, and when

The reaction was reported in 1975 by Sir Derek H. R. Barton and his postdoc Stuart W. McCombie in the Journal of the Chemical Society, Perkin Transactions 1 ("A New Method for the Deoxygenation of Secondary Alcohols"). Barton — already a 1969 Nobel laureate for his work on conformational analysis — was a master of radical chemistry, and this was one of a family of "Barton reactions" (alongside the Barton nitrite ester reaction and the Barton decarboxylation) that put controlled radical chemistry into the synthetic mainstream at a time when radicals were still viewed as messy. The choice of a thiocarbonyl antenna plus a thiophilic tin radical was the crucial design idea; it converted an inert C-O bond into a tractable radical fragmentation. In the decades since, the same addition-fragmentation logic seeded a whole family of xanthate-based radical reactions (including Zard's xanthate-transfer chemistry), and the tin-free variants have made the deoxygenation practical on scale.