The Wolff Rearrangement

What the Wolff rearrangement does

Start with an α-diazoketone — a ketone that carries a diazo group (=N₂) on the carbon right next to the carbonyl, written R-C(=O)-CH=N₂. Give it a nudge (heat, ultraviolet light, or a silver salt) and two things happen in one step: the diazo group leaves as a molecule of nitrogen gas, and the group R slides — a 1,2-shift — from the carbonyl carbon onto the neighboring carbon the nitrogen just left. The product is a ketene, R-CH=C=O, a molecule with two cumulated double bonds sharing a central carbon.

Ketenes are fast, hungry electrophiles; you almost never isolate one. Instead you generate it in the presence of a nucleophile that pounces on the central carbon:

• Water → a carboxylic acid, R-CH₂-COOH
• An alcohol R′OH → an ester, R-CH₂-CO-OR′
• An amine R′₂NH → an amide, R-CH₂-CO-NR′₂

Count the carbons. You began with an acyl fragment R-CO- (from a carboxylic acid R-COOH) and you end with R-CH₂-COOH — the same molecule with one extra CH₂ spliced in between R and the acid group. That net "grow the chain by one carbon at the acid end" is the whole reason organic chemists reach for the Wolff rearrangement, and it is the payoff step of the Arndt-Eistert synthesis.

The mechanism, arrow by arrow

The diazoketone is already a resonance hybrid. Its two most important forms tell you where the electrons are:

    R-C(=O)-CH=N⁺=N⁻   ⟷   R-C(=O)-CH⁻-N⁺≡N
      (diazo form)            (carbanion/diazonium form)

The C-N bond is the weak point. When the reaction is triggered, that bond breaks and a very stable molecule of N₂ walks away. Two limiting pictures describe what happens next; the truth sits between them.

1. Lose nitrogen. The terminal N₂ departs, taking the C-N bonding electrons with it. If this happens first and alone, it leaves behind a singlet α-ketocarbene (an acylcarbene): R-C(=O)-C̈-H, with an empty and a filled orbital on the same carbon, right beside the carbonyl.
2. Migrate. The R-C bond of the carbonyl pivots. The pair of electrons in the R-C(carbonyl) bond swings over to fill the carbene's empty orbital, carrying R along with it. As R departs the carbonyl carbon, a new carbon-carbon π bond forms between the old carbonyl carbon and the carbene carbon; the oxygen keeps its double bond to the same carbon it was always on (the carbonyl carbon).
3. Ketene. The result is R-CH=C=O — R and the H now sit together on the terminal carbon (the former carbene/diazo carbon), the oxygen stays on the central carbon (the former carbonyl carbon), and the two cumulated double bonds C=C=O share that central carbon.
4. Trap. The nucleophile (H₂O, R′OH, R′₂NH) adds across the electrophilic C=C of the ketene, and after proton shuffling you have the homologated acid, ester, or amide.

   R                       R                        R
    \                       \                         \
     C=O          –N₂         C=O          1,2-shift    CH=C=O   +Nu-H    R-CH₂-C(=O)-Nu
    /            ─────►      /              ─────►                     ─────►
   CH=N₂                    C̈H              (ketene)              (homologated product)
                        (α-ketocarbene:
                         carbene on Cα)

In the concerted version — favored under silver catalysis and for many thermal cases — steps 1 and 2 are one event: R starts migrating as N₂ leaves, so no free carbene ever exists. In the stepwise version — more common under photolysis — the singlet ketocarbene is a real, if fleeting, intermediate. Either way the migrating group keeps hold of its bonding electrons across a tight, suprafacial transition state, which is why stereochemistry is retained (see below).

Triggers, reagents, and conditions

The same diazoketone can be pushed over the barrier three ways, and the choice matters for yield, selectivity, and how much oxirene scrambling you tolerate.

• Silver catalysis (the classic Arndt-Eistert trigger). Silver oxide (Ag₂O), silver benzoate (PhCO₂Ag, often dissolved in Et₃N), or catalytic AgNO₃ with a tertiary-amine base. Run at 0 °C to gentle reflux, directly in the trapping nucleophile — aqueous dioxane for the acid, the alcohol itself for esters. Silver is thought to coordinate the diazo carbon and steer a clean, largely concerted rearrangement.
• Photolysis (mildest, most general). UV light at 254–366 nm, room temperature or below, often in the trapping solvent. Photochemistry tolerates sensitive substrates that would decompose on heating and is the mode used in photoresists. The trade-off is more of the stepwise carbene pathway and hence more oxirene-mediated carbon scrambling.
• Thermolysis. Simply heating the diazoketone, roughly 60–200 °C depending on substrate. Simplest to set up but the most side-reaction-prone (1,3-C-H insertion, dimerization), so it is usually the fallback.

Because the ketene is never isolated, the nucleophile is present from the start. A subtlety: in ester synthesis you want an excess of the alcohol, and in acid synthesis a homogeneous aqueous-organic medium, so the fast ketene is captured before it can dimerize to a diketene or do intramolecular chemistry.

Making the α-diazoketone

The Wolff rearrangement can only start from a diazoketone, so the Arndt-Eistert sequence begins one step earlier. The textbook route:

   R-COOH  ──SOCl₂ (or (COCl)₂)──►  R-COCl  ──2 eq CH₂N₂──►  R-CO-CHN₂  +  CH₃Cl  +  N₂

• Activate the acid. Convert R-COOH to the acid chloride with thionyl chloride or oxalyl chloride.
• Acylate diazomethane. Add the acid chloride to at least two equivalents of diazomethane (CH₂N₂). The first equivalent installs the diazoketone; the second is a sacrificial base that scavenges the liberated HCl. Skimp on CH₂N₂ and the HCl converts your diazoketone into a useless α-chloroketone (R-CO-CH₂Cl).
• Safer modern variants. Diazomethane is toxic, volatile, and explosive against ground glass. Bench chemists increasingly use trimethylsilyldiazomethane (TMSCHN₂) or continuous-flow reactors that make and consume CH₂N₂ in a single small volume, never accumulating it.

Selectivity and stereochemistry

Two facts make the Wolff rearrangement synthetically trustworthy:

• Retention at the migrating carbon. The 1,2-shift is suprafacial and concerted with respect to the migrating group; R never becomes a free, planar radical or cation, so a stereocenter within R survives the migration with its configuration intact. Homologate an enantiopure α-amino acid and you get the β³-amino acid one carbon longer, with the α-stereocenter untouched. This clean chirality transfer is the reason the reaction underpins peptide and β-amino-acid chemistry.
• Migratory aptitude is broad. Alkyl, aryl, vinyl, and even hydrogen migrate. Unlike carbocation rearrangements, migratory aptitude in the Wolff is not strongly controlled by the ability to stabilize positive charge, because the transition state is not a free cation — so groups that would be poor cation-migrators still shift well.

The one selectivity headache is oxirene scrambling. Because the α-ketocarbene can close to a symmetric three-membered oxirene (a C₂O ring) and reopen from the other carbon, the two carbons — the original carbonyl carbon and the original diazo carbon — can lose their identity. Feeding a ¹³C label on the carbonyl carbon shows it partly ends up scrambled between both positions in the product. This is invisible with ordinary substrates but real when you care which carbon is which, and it is worse under photolysis than under silver catalysis.

Wolff vs neighboring rearrangements

	Wolff rearrangement	Curtius rearrangement	Beckmann rearrangement
Starting material	α-diazoketone R-CO-CHN₂	Acyl azide R-CO-N₃	Oxime R₂C=N-OH
Small molecule lost	N₂	N₂	H₂O
Migration origin → terminus	C → C (carbon migrates to carbon)	C → N (carbon migrates to nitrogen)	C → N (anti group migrates to N)
Reactive intermediate	Ketene (via ketocarbene)	Isocyanate (via acylnitrene)	Nitrilium ion
Trapped product	Homologated acid / ester / amide	Amine / carbamate / urea	Amide (or lactam)
Net chain change	+1 carbon (homologation)	Carbon count unchanged; C→N insertion	Ring expansion / C→N insertion
Stereochemistry at migrating C	Retention	Retention	Retention (anti-periplanar group migrates)
Typical trigger	Ag(I), hν, or Δ	Δ (or one-pot DPPA)	Acid catalyst (H₂SO₄, PCl₅)

Worked example: homologate phenylacetic acid

Turn phenylacetic acid (PhCH₂COOH) into 3-phenylpropanoic acid (PhCH₂CH₂COOH) — one carbon longer at the acid end — by a full Arndt-Eistert sequence.

   PhCH₂COOH
      │  (1) SOCl₂
      ▼
   PhCH₂COCl
      │  (2) CH₂N₂ (2.2 eq), Et₂O, 0 °C
      ▼
   PhCH₂-CO-CHN₂            (α-diazoketone)
      │  (3) Ag₂O or PhCO₂Ag, H₂O / dioxane, 60 °C   ← the Wolff step
      ▼
   [ PhCH₂-CH=C=O ]         (ketene, not isolated)
      │  + H₂O
      ▼
   PhCH₂CH₂COOH             (3-phenylpropanoic acid)

• Step 1. Thionyl chloride gives phenylacetyl chloride; strip excess SOCl₂ before the next step (it would destroy diazomethane).
• Step 2. Add the acid chloride to cold ethereal CH₂N₂ (about 2.2 equivalents). The 1-diazo-3-phenylpropan-2-one forms as a pale-yellow solution; watch for N₂ evolution.
• Step 3 (Wolff). Warm the diazoketone with a catalytic silver source in aqueous dioxane. N₂ bubbles off, the benzyl carbon's neighbor migrates, and the transient ketene is hydrated on the spot.
• Yield. Well-behaved substrates give 60–80% of the homologated acid over the sequence. Swap water for methanol in step 3 and you get methyl 3-phenylpropanoate directly.

Real-world applications

• Amino-acid homologation. Arndt-Eistert on an N-protected α-amino acid gives the β-amino acid one carbon longer with the α-stereocenter retained — the standard entry to β-peptides and peptidomimetics that resist proteases.
• Ring contraction. A cyclic α-diazoketone rearranges to a ring-contracted ketene: a cyclohexanone-derived diazoketone becomes a cyclopentane-fused ketene, captured as a cyclopentanecarboxylic acid or ester. This is a workhorse for building strained four- and five-membered rings.
• Diazonaphthoquinone (DNQ) photoresists. The photochemical Wolff is one of the most economically important reactions on Earth. DNQ dissolved in Novolak phenolic resin is the photoactive compound of positive-tone photoresist. UV light triggers a Wolff rearrangement/ring contraction of the naphthoquinone diazide to a ketene, which water traps as an indene-3-carboxylic acid. The exposed regions become soluble in aqueous base (TMAH developer), so the light-struck pattern washes away — the chemistry that has patterned integrated circuits for decades.
• Total synthesis. Chemists use one-carbon Wolff homologation to lengthen side chains in complex targets, and Wolff ring contraction to install cyclobutane-carboxylic acids and other strained rings that are hard to reach any other way.

Limitations and side reactions

• Diazomethane hazard. The upstream diazoketone synthesis usually needs CH₂N₂ — toxic, volatile, and explosive. This is the single biggest reason the reaction is used less than its power warrants; flow chemistry and TMSCHN₂ mitigate but do not eliminate the concern.
• Competing carbene chemistry. The α-ketocarbene can do a 1,3-C-H insertion, cyclopropanate a nearby alkene, or be trapped by the solvent instead of rearranging. Silver catalysis and photolysis at low temperature bias toward the desired Wolff pathway.
• Ketene side reactions. If the trapping nucleophile is slow or absent, the ketene dimerizes (to a β-lactone/diketene) or reacts intramolecularly. Keep the nucleophile in excess and the ketene concentration low.
• Oxirene scrambling. When carbon identity matters (isotope labeling, regiochemically defined products), the reversible oxirene intermediate scrambles the two candidate carbons, more so under photolysis.
• α-substitution required. You need a diazo group α to the carbonyl; substrates that resist forming a clean diazoketone (or that have acidic α-protons prone to side reactions) are poor candidates.

Discovery and history

Ludwig Wolff, a German chemist at Jena, reported in 1902 that heating diazoketones (or treating them with silver oxide) produced carboxylic acids and their derivatives one carbon longer — the rearrangement that carries his name. The photochemical variant came decades later. The mechanistic picture of a ketene intermediate came later, as the chemistry of ketenes and carbenes matured. In the 1930s Fritz Arndt and Bernd Eistert packaged the rearrangement into a reliable, general homologation protocol — acid → acid chloride → diazoketone → silver-catalyzed Wolff → homologated acid — the sequence now universally called the Arndt-Eistert synthesis. The oxirene controversy was settled by ¹³C- and ¹⁸O-labeling studies in the mid-20th century, which caught the symmetric intermediate red-handed by the carbon scrambling it produces.

Safety and practical notes

• Diazomethane. Generate and use it behind a blast shield, in flame-polished (never ground-glass) apparatus, at low temperature, and only in the quantity you need. Prefer in-line generation or TMSCHN₂ where possible.
• Nitrogen evolution. The rearrangement liberates N₂ gas; run in vented apparatus and expect gas evolution as the diagnostic that it is working.
• Silver waste. Silver salts are the cleanest trigger but are costly and toxic to aquatic life; collect and recover silver residues rather than discarding them.
• Ketene reactivity. Ketenes are lachrymators and potent acylating agents. They are consumed in situ here, but design the workup so no free ketene survives.