The Paterno-Buchi Reaction

What the Paterno-Buchi reaction does

Take a carbonyl compound — an aldehyde or ketone — and an alkene, mix them in a photochemically transparent solvent, and shine ultraviolet light. What comes out is an oxetane: a strained four-membered ring built from three carbons and one oxygen. The carbonyl contributes its carbon and oxygen; the alkene contributes its two double-bond carbons. Every atom of both partners ends up in the ring, so the reaction has perfect atom economy.

Formally this is a [2+2] cycloaddition: two π electrons from the C=O and two from the C=C combine to form two new σ bonds. Thermally, a suprafacial-suprafacial [2+2] is forbidden by the Woodward-Hoffmann orbital-symmetry rules — you simply cannot heat a ketone and an alkene together and get an oxetane. Light rewrites the rules. Promoting an electron to an excited state changes the symmetry of the frontier orbitals, and the photochemically excited [2+2] becomes allowed. This is the textbook example of a reaction that is forbidden in the ground state and allowed on the excited-state surface.

        R₂C=O   +   C=C          ──hν──→        oxetane
     (carbonyl)   (alkene)                    O
                                              / \
                                        R₂C     C
                                             \  /
                                              C
     the C and O of the carbonyl + the two alkene carbons = one 4-membered ring

The mechanism, arrow by arrow

The whole reaction hinges on what UV light does to the carbonyl. Follow the electrons:

1. Photoexcitation to the n,π* state. A photon promotes one of oxygen's non-bonding lone-pair electrons (the n orbital) up into the antibonding π* of the C=O. The result is an n,π* excited state. Crucially, oxygen now has only one electron in that lone-pair orbital — it has become electron-deficient and electrophilic, behaving chemically like an alkoxy radical (RO·). The carbonyl carbon simultaneously gains radical character.
2. Intersystem crossing (for most carbonyls). Aryl ketones and aldehydes cross from the initially formed singlet (S₁) to the lower-lying triplet (T₁) n,π* state with near-unit efficiency in a few picoseconds. Benzophenone, the workhorse sensitizer, has an intersystem-crossing quantum yield of essentially 1.0. The reactive species in most Paterno-Buchi reactions is therefore the triplet carbonyl.
3. First C–O bond forms → 1,4-biradical. The electrophilic excited oxygen attacks one carbon of the alkene, forming the first new C–O σ bond. Because the electrons were unpaired, this generates a triplet 1,4-biradical: an oxygen-bearing carbon radical on one end and a free carbon radical on the other, four atoms apart. Regiochemistry is decided here — oxygen adds to give the more stabilized carbon radical (benzylic, allylic, or the more substituted carbon), following the "rule of five" for the favored biradical geometry.
4. Intersystem crossing of the biradical. A triplet biradical cannot close a ring — the two unpaired electrons have parallel spins, and a σ bond needs a singlet pair. The biradical must first flip one spin (intersystem crossing), which takes nanoseconds. During this window the central C–C bond can rotate freely, which is why stereochemistry from the starting alkene is often scrambled.
5. Ring closure. Once the biradical is a singlet, the two radical centers pair up, forming the second σ bond (a C–C bond) and closing the four-membered oxetane ring. The molecule drops back to the ground state as a stable product.

  step 1:  R₂C=O  ──hν──→  [R₂C–O·]*   (n,π* excited, O is electrophilic radical)
  step 2:  S₁  ──ISC──→  T₁            (triplet carbonyl, the reactive species)
  step 3:  O· attacks alkene C  →  ·O–C···C·   1,4-BIRADICAL (triplet)
  step 4:  triplet biradical  ──ISC──→  singlet biradical  (spins now paired)
  step 5:  radical pairing closes the ring  →  OXETANE

The two-step, biradical nature is the single most important feature to remember: unlike the concerted, stereospecific thermal Diels-Alder, the triplet Paterno-Buchi builds its ring through a genuine, rotatable intermediate.

Reagents, light, and conditions

• Carbonyl. Aromatic ketones and aldehydes are the classic partners because they have low triplet energies and clean n,π* absorption in the near-UV. Benzophenone (Ph₂C=O), benzaldehyde, and acetophenone are standards. Aliphatic aldehydes and ketones work too but often react through the singlet and absorb at shorter, harsher wavelengths.
• Alkene. Electron-rich alkenes react fastest: enol ethers, ketene acetals, furans, and alkyl-substituted alkenes. Furan is a signature partner, giving fused bicyclic oxetanes (2,7-dioxabicyclo[3.2.0]hept-3-enes).
• Light source. A medium-pressure mercury lamp with a Pyrex filter (cuts off below ~300 nm) hits the carbonyl n→π* band around 300–366 nm. Modern labs use 365 nm LEDs or 350 nm Rayonet fluorescent tubes for cleaner, cooler irradiation.
• Solvent. Photochemically inert and UV-transparent: benzene, acetonitrile, dichloromethane, or neat alkene. Avoid solvents with abstractable C–H bonds near the carbonyl (like cyclohexane) that let the excited carbonyl abstract hydrogen instead — the competing Norrish-type photoreduction.
• Temperature. Usually run cold (–20 to 25 °C). Lower temperature slows bond rotation in the biradical and can improve both yield and diastereoselectivity by favoring ring closure over fragmentation.
• Atmosphere. Degas with N₂ or Ar — molecular oxygen is a triplet quencher and will short-circuit the triplet carbonyl before it finds the alkene.

Scope, regiochemistry, and stereochemistry

Regiochemistry — the rule of five. For triplet carbonyls, the first bond formed is the C–O bond, and it forms to give the more stable 1,4-biradical. Wagner's "rule of five" rationalizes this: the biradical prefers a geometry in which a five-membered cyclic arrangement of atoms allows efficient spin inversion and closure. The practical upshot: oxygen bonds to whichever alkene carbon leaves the more stabilized radical (benzylic, allylic, or the more substituted carbon) on the other end. With an unsymmetrical enol ether like ethyl vinyl ether, the electron-rich alkoxy terminus dominates the excited-state addition, so oxygen bonds to the alkoxy-substituted carbon, placing the OEt group next to the ring oxygen (a 2,2-diaryl-4-alkoxyoxetane pattern with an acetal-like O–C–O array).

Stereochemistry. Here is the reaction's Achilles' heel. Because the triplet 1,4-biradical lives long enough (nanoseconds) for the central C–C bond to rotate, the geometric information in a cis- or trans-alkene is frequently lost — a cis-alkene and a trans-alkene can converge on the same mixture of oxetane diastereomers. This is diagnostic of the stepwise biradical pathway. Two escape routes exist:

• Singlet chemistry. Some carbonyls (certain aliphatic aldehydes, or excitation directly into a short-lived singlet) react before intersystem crossing. The singlet biradical closes in picoseconds — faster than rotation — so the reaction is suprafacial-suprafacial and stereospecific, retaining the alkene geometry.
• Rigid or constrained systems. When the alkene is locked (a cyclic alkene, or an intramolecular tether), rotation is restricted and diastereoselectivity climbs. Bach's work on chiral-auxiliary and hydrogen-bond-templated Paterno-Buchi reactions has achieved high enantio- and diastereocontrol by rigidifying the biradical.

Paterno-Buchi vs. related cycloadditions

	Paterno-Buchi [2+2]	Thermal Diels-Alder [4+2]	Enone [2+2] photocycloaddition
Ring formed	Oxetane (4-membered, 1 O)	Cyclohexene (6-membered)	Cyclobutane (4-membered, all C)
Partners	Excited C=O + alkene	Diene + dienophile	Excited C=C (enone) + alkene
Driving input	UV light (photon)	Heat (thermal)	UV light (photon)
Symmetry status	Thermally forbidden, photochemically allowed	Thermally allowed	Thermally forbidden, photochemically allowed
Mechanism	Stepwise via triplet 1,4-biradical (usually)	Concerted, one step	Stepwise via 1,4-biradical
Stereospecific?	Often no (triplet scrambles)	Yes (suprafacial-suprafacial)	Often no (biradical)
Key intermediate	1,4-biradical (O + C radicals)	Aromatic-like transition state	1,4-biradical (two C radicals)
Reactive component	The carbonyl is excited	Ground state throughout	The alkene/enone is excited

Worked example: benzophenone + 2-methyl-2-butene

The canonical demonstration Büchi ran in 1954: irradiate benzophenone with a trisubstituted alkene and isolate a single dominant oxetane.

   Ph₂C=O  +  (CH₃)₂C=CH-CH₃   ──hν (350 nm), benzene, N₂, 25 °C──→  2,2-diphenyl-3,3,4-trimethyloxetane

• Reagents. Benzophenone 1.0 equiv, 2-methyl-2-butene in large excess (often used as co-solvent), degassed benzene.
• Light. 350 nm Rayonet lamps or a Pyrex-filtered medium-pressure Hg lamp; irradiate several hours until the yellow benzophenone n→π* color fades.
• Regiochemistry. The excited oxygen adds to the less substituted alkene carbon so that the resulting carbon radical lands on the more substituted, tertiary carbon (more stabilized) — the rule of five in action. Ring closure gives the oxetane with both phenyls on C-2 (next to the ring oxygen).
• Outcome. A 2,2-diphenyl-3,3,4-trimethyloxetane as the major regioisomer, isolated by chromatography. Yields for benzophenone + electron-rich alkenes routinely reach 60–90%.

The furan variant is even more useful synthetically: benzaldehyde + furan under UV gives a bicyclic oxetane in which the fused dihydrofuran ring carries a masked 1,4-dicarbonyl. That single photochemical step, followed by hydrolysis, delivers polyol and carbohydrate fragments in a handful of operations — a strategy Schreiber exploited to make complex sugars.

Real-world applications

• The taxol oxetane. Taxol (paclitaxel), the blockbuster anticancer drug, contains a strained oxetane D-ring that is essential for its microtubule-stabilizing activity. Several total-synthesis campaigns and model studies used Paterno-Buchi chemistry (or Paterno-Buchi-inspired [2+2] logic) to think about installing that oxetane, making the reaction a touchstone in oxetane-containing natural-product synthesis.
• Carbohydrate synthesis (the furan route). Schreiber's "Paterno-Buchi/retro-aldol" strategy converts a furan + aldehyde photocycloadduct into differentiated polyols, giving rapid access to erythro/threo diol arrays and rare sugars in few steps and with high stereocontrol from cyclic substrates.
• DNA and biomolecule photodamage. The same [2+2] photochemistry underlies oxetane-type crosslinks. Although the pyrimidine-pyrimidone (6-4) photoproduct in UV-damaged DNA forms via an oxetane/azetidine intermediate that then rearranges, the Paterno-Buchi framework is exactly how chemists rationalize that biologically critical lesion.
• Oxetanes as medicinal building blocks. Medicinal chemists prize the oxetane ring as a small, polar, metabolically robust replacement for a gem-dimethyl group or a carbonyl. Paterno-Buchi is one of the few direct C–C-and-C–O ring-forming routes to substituted oxetanes, complementing Williamson-type cyclizations.
• Flow photochemistry scale-up. Continuous-flow reactors with high-surface-area LED arrays have made preparative Paterno-Buchi reactions practical on gram-to-kilogram scale, overcoming the classic problem that UV light penetrates only a thin layer of a batch flask.

Limitations and side reactions

• Norrish photoreduction / hydrogen abstraction. The excited carbonyl oxygen is a good hydrogen-atom abstractor. If a solvent or substrate has weak, accessible C–H bonds, the carbonyl abstracts a hydrogen and you get pinacols and reduction products instead of the oxetane. Choose inert solvents and electron-rich alkenes to keep cycloaddition fast.
• Triplet quenching by oxygen. Ground-state O₂ is a triplet and efficiently quenches the triplet carbonyl, killing the reaction and generating singlet oxygen. Rigorous degassing is mandatory.
• Poor stereospecificity. As covered above, the triplet biradical rotates before closing, so E/Z information in the alkene is usually lost — a real limitation when a defined oxetane stereocenter is the goal.
• Regiochemical mixtures. With alkenes that give two similarly stabilized biradicals, the rule of five loses its grip and product mixtures result.
• Oxetane fragility. Oxetanes are strained (ring strain ~106 kJ/mol) and can undergo acid-catalyzed ring-opening or retro-[2+2] under prolonged irradiation, capping conversion and sometimes returning starting materials.
• Electron-poor alkenes fail. Acrylates and fumarates react sluggishly or divert to competing radical polymerization, because their π* is a poor match for the electrophilic excited oxygen.

Discovery: Paternò to Büchi

The reaction is named for two chemists separated by 45 years. In 1909, the Italian chemist Emanuele Paternò (1847–1935), working in Palermo, discovered that irradiating mixtures of carbonyl compounds and alkenes produced new compounds — but the analytical tools of the era could not tell him what the four-membered oxygenated products actually were. His observation sat as a curiosity for decades.

In 1954, George Büchi (1921–1998) at MIT returned to Paternò's reaction with modern spectroscopy, definitively established that the products were oxetanes, mapped the substrate scope, and laid the groundwork for the mechanistic picture. Because Paternò discovered the transformation and Büchi elucidated the products and mechanism, the reaction carries both names. The mechanistic details — the n,π* excited state, the role of the triplet, and the 1,4-biradical — were filled in over the following decades by Turro, Wagner, Kochevar, and others who studied carbonyl photochemistry in depth.

Safety and practical notes

• UV exposure. The 300–366 nm light used here damages skin and especially eyes. Run reactions in a closed photoreactor cabinet or behind UV-blocking shielding; never look at an unshielded medium-pressure Hg lamp.
• Ozone and heat. Medium-pressure Hg lamps generate ozone and significant heat. Ventilate and provide cooling water to the immersion well.
• Degassing. Sparge with inert gas before irradiation, both to remove the triplet-quenching O₂ and to avoid singlet-oxygen chemistry that can oxidize your substrates.
• Solvent choice. Benzene is a good UV-transparent solvent but a carcinogen; acetonitrile or dichloromethane are safer transparent alternatives for most substrates.
• Scale. Because UV light is absorbed in the first millimetre of solution, batch reactions plateau; move to thin-film or continuous-flow photoreactors for preparative scale rather than simply irradiating a larger flask longer.