The Woodward-Hoffmann Rules and Orbital Symmetry

In 1965, a puzzle nagged at R. B. Woodward. He was working out the synthesis of vitamin B12. Why did one stereoisomer of a cyclic triene close to a cyclobutene with one relative configuration? A very slightly different substrate closed with the opposite one. There was no obvious steric explanation. Woodward brought the question to Roald Hoffmann. Hoffmann's extended Hückel calculations had been mapping out how molecular orbitals evolve as a ring opens or closes. Within weeks they had the answer. Within a few years that answer had hardened into one of the most quantitatively reliable predictive frameworks in organic chemistry: the Woodward–Hoffmann rules. The rules matter because pericyclic reactions — electrocyclizations, cycloadditions, and sigmatropic shifts — do not go through carbocations, carbanions, or radicals. There are no discrete intermediates. You cannot estimate their stability with pKa tables or radical-stabilization energies. The reactions occur through a single cyclic, concerted transition state. All the bond-making and bond-breaking happens simultaneously, in one orbital-symmetry-controlled step. Whether that concerted pathway is geometrically and energetically accessible — "allowed" — or high-energy and avoided — "forbidden" — is dictated entirely by the symmetry of the orbitals involved. It is not dictated by electronics or sterics in the usual sense. The big picture is this: a concerted pericyclic transition state is itself a set of atomic orbitals arranged in a ring. That ring can only close in a bonding way if the orbital lobes match in phase all the way around. Photochemical excitation reorders which orbital is the frontier orbital. Counting the ring geometry as "same-face" (suprafacial) or "opposite-face" (antarafacial) also changes the phase bookkeeping. So the very same reaction type can flip from forbidden to allowed. This happens just by switching from heat to light. It also happens by switching from a 6-electron count to a 4-electron one. This is not empirical curve-fitting. It follows directly from orbital symmetry conservation along the reaction coordinate. It can be derived independently by three routes. These are frontier orbital phase-matching, the Hückel/Möbius aromatic transition-state analogy, and correlation diagrams. These three routes always agree. This article builds the electron-counting selection rules from first principles. It shows the frontier-orbital and aromatic-transition-state derivations side by side. It works through electrocyclizations, cycloadditions, and a sigmatropic shift. Full curved-arrow and orbital-phase reasoning is included. It sets up, without repeating, the more detailed mechanistic treatments of Electrocyclic Reactions: Ring Opening and Closing, The Diels-Alder Reaction Revisited, and Sigmatropic Rearrangements: Order and Nomenclature.

Key takeaways

Pericyclic reactions proceed through a single cyclic transition state; "allowed" vs "forbidden" is fixed by orbital-phase continuity around that ring, not by steric or inductive effects.
Count the total electrons in the cyclic array of interacting orbitals, then count how many bond components are used antarafacially (twisted, opposite-face) versus suprafacially (untwisted, same-face).
A 4n+2 electron pericyclic array with an even number of antarafacial components is a Hückel (aromatic) transition state — thermally allowed.
A 4n electron pericyclic array with an even number of antarafacial components is a Möbius (antiaromatic) transition state — thermally forbidden, but photochemically allowed.
Photochemical excitation always inverts the rule relative to the ground-state (thermal) case, because promoting an electron flips the symmetry of the relevant frontier orbital.
Correlation diagrams give the same allowed/forbidden answer as electron counting, because they track how bonding MOs of reactant connect, without crossing, to bonding MOs of product only when symmetry is conserved.
Key idea: a pericyclic transition state is allowed exactly when it can be drawn as an aromatic (Hückel, 4n+2, zero or even antarafacial components) or Möbius (4n, odd antarafacial components) array matching the electron count and light/heat conditions — everything else is bookkeeping.

The core insight

Allowedness is an aromaticity problem for the transition state: count π (and σ) electrons around the pericyclic ring, count antarafacial components, and match Hückel/4n+2 or Möbius/4n topology to thermal or photochemical conditions.

Where this lives today

Orbital symmetry control is not a historical curiosity — it is load-bearing in modern synthetic planning, photobiology, and materials chemistry. The vitamin D photochemical/thermal cascade used in Example 3 is a direct physiological instance: skin exposure to UVB light drives a photochemical 6π electrocyclic ring-opening of 7-dehydrocholesterol (a disrotatory-forbidden-turned-photoallowed process, since photochemical control inverts the thermal disrotatory preference to conrotatory) to previtamin D3, which then undergoes the thermally allowed antarafacial [1,7]-H shift analyzed above to give vitamin D3 itself; both steps are textbook Woodward–Hoffmann predictions operating inside human skin. UV-induced cyclobutane pyrimidine dimer formation in DNA, the dominant mutagenic photolesion from sunlight exposure, is the photochemically-allowed [2+2] cycloaddition of Example 2, and its enzymatic reversal by photolyase runs the identical orbital-symmetry logic in reverse using blue light and a radical-anion mechanism. In synthetic methodology, orbital symmetry rules are used prospectively to design cascades: Nicolaou's and Danishefsky's polyene cyclizations, Diels–Alder-based syntheses of steroids and terpenoids, and tandem electrocyclization/cycloaddition sequences in total synthesis (endiandric acid cascades being the canonical example, in which a single molecule undergoes a thermal 8π-conrotatory electrocyclization followed immediately by a 6π-disrotatory electrocyclization, both correctly predicted in direction purely from electron count before the reaction was ever run) rely on knowing in advance which ring-closure stereochemistry is geometrically forced. In materials chemistry, photochromic molecular switches — diarylethenes and fulgides used in optical data storage and smart materials — are engineered explicitly around the photochemical/thermal complementarity: UV light drives a photochemically allowed ring closure that is thermally forbidden (hence thermally stable once formed, a requirement for a useful bistable switch), and visible light or heat drives the reverse ring-opening only through its own allowed channel. Metal-catalyzed formally-forbidden cycloadditions ([2+2] additions catalyzed by titanium, nickel, or photoredox catalysts) exploit exactly the loophole flagged in the scope table: routing a formally forbidden concerted transformation through a stepwise organometallic or radical-ion mechanism sidesteps the orbital symmetry constraint entirely, which is precisely why catalysis, rather than brute thermal forcing, is the practical solution chemists use to access those product classes.

:::quiz[A [2+2] cycloaddition between two simple, unactivated alkenes, both reacting suprafacially, is...]

Thermally allowed and photochemically forbidden
Thermally forbidden and photochemically allowed ✓
Allowed under both thermal and photochemical conditions
Forbidden under both thermal and photochemical conditions

:::quiz[Why does a thermal [1,3]-hydrogen shift not proceed via the concerted antarafacial pericyclic pathway predicted by electron counting?]

The electron count (4) actually requires the suprafacial pathway, which is easy
The single intervening double bond is geometrically too short/rigid to twist into an antarafacial arrangement ✓
Hydrogen atoms cannot participate in any pericyclic reaction
[1,3]-H shifts always have 4n+2 electrons, so they are simply forbidden outright

Ask the tutor: How do I build the actual correlation diagram for the disrotatory closure of hexatriene to 1,3-cyclohexadiene, orbital by orbital?

The physical picture

Picture the transition state of a pericyclic reaction as a closed loop of overlapping p orbitals — literally a ring, whether or not the atoms end up in a ring. In the Diels–Alder reaction the diene's four p orbitals and the dienophile's two p orbitals form a six-membered loop of overlap in the transition state, even though only four of those six atoms are in the diene chain. In the disrotatory or conrotatory closure of a triene to a cyclobutene, the terminal p orbitals rotate to overlap with each other, again closing a ring of orbital overlap. In every case the question "does this loop close with continuous, net-bonding orbital overlap, or does it close with a sign mismatch (a node) built into the ring?" decides whether the transition state is low-energy (allowed) or high-energy (forbidden). This is exactly the logic of aromaticity, transplanted from a stable molecule (benzene) to a transition state. A cyclic array of p orbitals with a 4n+2 electron count and no built-in phase inversion is a Hückel system: continuously overlapping, aromatic, stabilized. The same array with a 4n electron count is destabilized unless a phase inversion (a twist, i.e., one orbital used antarafacially) is introduced somewhere in the loop — that twisted topology is a Möbius system, and Möbius arrays are aromatic (stabilized) at 4n and antiaromatic at 4n+2. Woodward and Hoffmann's insight was that exactly this Hückel/Möbius aromaticity counting, applied to the transition state rather than to a stable ring molecule, predicts which pericyclic pathways are low in energy. The second essential idea is suprafacial versus antarafacial use of a component. A component (a π bond, a σ bond, or a lone pair) reacts suprafacially if both new bonds form on the same face of that component — no twist, no phase inversion introduced. It reacts antarafacially if the new bonds form on opposite faces — necessarily introducing a node, a sign flip, into the cyclic array, because you have connected a top lobe to a bottom lobe on the way around the ring. Suprafacial geometry is stereochemically easy (cis-fused, non-strained) for small rings; antarafacial geometry requires the component to twist through the transition state, which is geometrically difficult for a simple alkene but routine for a long, floppy polyene chain or a σ bond viewed from its backside. Third, promotion from thermal (ground-state) to photochemical (excited-state) control changes the answer because it changes the symmetry of the frontier orbital. In the thermal reaction, the highest occupied molecular orbital (HOMO) — or more precisely, the full set of occupied MOs — determines the phase pattern that must close continuously. Photochemical excitation promotes an electron from the HOMO to the LUMO; the new "frontier" occupied orbital is now the old LUMO, which has one more node and hence the opposite symmetry at the reacting termini. This single change flips the entire allowed/forbidden assignment for every pericyclic reaction type.

Figure

Hückel vs Möbius phase continuity in a six-orbital cyclic array

Figure

Suprafacial vs antarafacial use of a π component

The core ideas

The master selection rule, expressed as a single electron-counting statement, is:

\text{Thermally allowed} \iff \left(\text{total electrons} = 4n+2 \text{ and } N_a \text{ even}\right)\ \text{or}\ \left(\text{total electrons} = 4n \text{ and } N_a \text{ odd}\right)

where: total electrons is the sum of electrons in all components of the cyclic pericyclic array (each π bond contributes 2, each σ bond used contributes 2, each participating lone pair contributes 2); $N_a$ is the number of components used antarafacially; $n = 0, 1, 2, \dots$ Reading: an even number of antarafacial components (0, 2, 4, …) keeps the array in Hückel topology, which wants 4n+2 electrons to be aromatic and stabilized at the transition state. An odd number of antarafacial components (1, 3, …) puts one twist into the loop, converting it to Möbius topology, which wants 4n electrons to be aromatic. Whichever combination gives an aromatic (stabilized) transition state is the thermally allowed pathway; the other combination is forbidden because it would force the transition state to be antiaromatic. Photochemical control inverts this entire statement:

\text{Photochemically allowed} \iff \left(\text{total electrons} = 4n+2 \text{ and } N_a \text{ odd}\right)\ \text{or}\ \left(\text{total electrons} = 4n \text{ and } N_a \text{ even}\right)

where: the electron count and $N_a$ mean the same things as above, but now the process must access the excited-state configuration, so the topology that is "wrong" thermally becomes the one accessed with a lower barrier from the excited state. Reading: light does not change which topology is aromatic; it changes which topology the reacting system actually adopts, because excitation flips the symmetry of the reacting frontier orbital. A process forbidden by heat is (in the simplest orbital picture) allowed by light, and vice versa — this is the celebrated thermal/photochemical complementarity. For the common special case of an all-suprafacial pericyclic process ( $N_a = 0$ , which is even), the rule collapses to the version most students memorize first:

4n+2 \ \text{electrons, all-suprafacial} \implies \text{thermally allowed}

4n \ \text{electrons, all-suprafacial} \implies \text{thermally forbidden (but photochemically allowed)}

where: $n$ is a non-negative integer, so 4n+2 gives 2, 6, 10, 14 … electrons and 4n gives 4, 8, 12 … electrons. These equations are not independent postulates — they are two different bookkeeping methods (aromatic-transition-state counting and frontier-orbital phase matching) for the same underlying fact, which is that the reaction path must conserve orbital symmetry continuously from reactant to product. The two derivations below establish that equivalence explicitly.

Derivation AAromatic transition state counting — Hückel/Möbius analysis

Goal. Show that the electron-counting rule above follows from requiring the cyclic transition-state array to be an aromatic (not antiaromatic) arrangement of p orbitals, by direct analogy with monocyclic aromaticity.

Step 1 — Build the cyclic array of interacting orbitals. In any pericyclic transition state, identify every bond or lone pair that participates directly in the reorganization — this defines the components. For example, in a Diels–Alder cycloaddition the components are the diene π system (4 electrons, one component contributing an array of orbitals across 4 atoms) and the dienophile π bond (2 electrons). Arrange the participating p orbitals as a closed ring, exactly as you would draw benzene's six p orbitals, since in the transition state all these orbitals overlap continuously around a loop connecting the forming/breaking bonds.

Step 2 — Assign a phase to every lobe around the ring. Choose a phase (+ or −) for one lobe arbitrarily, then propagate the phase around the ring using the actual nodal pattern of each component's orbital (for a π bond, adjacent p orbitals in a bonding combination have the same phase; across a node they flip). Walk all the way around the ring back to the starting orbital.

Step 3 — Count the total number of phase inversions (nodes) encountered on the walk. If a component is used suprafacially, the two new bonds it makes are on the same lobe, so no extra sign flip is introduced by that component beyond its own internal nodal structure. If a component is used antarafacially, the new bonds are on opposite lobes (top and bottom), which introduces exactly one additional sign flip relative to the suprafacial case. Let $N_a$ be the number of antarafacially-used components.

Step 4 — Classify the ring topology. If the total number of sign inversions around the ring (from antarafacial usage) is even, the ring closes with the same sign it started with — this is Hückel topology, structurally identical to benzene's continuous p-orbital ring. If the number is odd, the ring closes with the opposite sign — a built-in half-twist — this is Möbius topology, structurally identical to a twisted annulene loop.

Step 5 — Apply the known aromaticity rule for each topology. Hückel systems are aromatic, and therefore stabilized (bonding, low energy) precisely when they hold $4n+2$ electrons — this follows from Hückel molecular orbital theory, where a monocyclic system of $m$ p orbitals has one non-degenerate lowest MO and pairs of degenerate MOs above it, filled completely (closed shell, all bonding/nonbonding MOs occupied) only at electron counts $2, 6, 10, \dots$ . Möbius systems have the reverse pattern — all levels are doubly degenerate (with one non-degenerate level pushed to high energy instead of low), giving a closed, fully bonding shell at $4n$ electrons instead.

Step 6 — Translate into the selection rule. A transition state that is aromatic at the relevant electron count is low in energy — it is allowed. A transition state forced into the antiaromatic count for its topology is destabilized and high in energy — it is forbidden. Substituting the results of Step 5:

(4n+2 \ \text{electrons}) + (\text{Hückel, } N_a \text{ even}) \implies \text{aromatic TS} \implies \text{thermally allowed}

(4n \ \text{electrons}) + (\text{Möbius}, N_a \text{ odd}) \implies \text{aromatic TS} \implies \text{thermally allowed}

which is exactly the master rule stated above; the complementary pairings give antiaromatic (forbidden) transition states.

Limiting-case check. Set $N_a = 0$ (fully suprafacial, the overwhelmingly common experimental case for small rings). Then $N_a$ is even by definition (zero counts as even), forcing Hückel topology, and the rule reduces correctly to "4n+2 thermally allowed, 4n thermally forbidden" — reproducing the simple mnemonic used for Diels–Alder ( $4n+2 = 6$ , allowed) and for disallowed [2+2] cycloadditions ( $4n = 4$ , forbidden, all-suprafacial).

Go deeper · Derivation B · 3 minFrontier orbital phase-matching — direct correlation with FMO theory

Derivation BFrontier orbital phase-matching — direct correlation with FMO theory

Goal. Re-derive the same allowed/forbidden assignment using only the frontier molecular orbitals of the components, showing consistency with Frontier Molecular Orbitals and Reaction Symmetry, and show explicitly how photochemical excitation inverts the answer.

Step 1 — Identify the frontier orbitals of each component. For a thermal, ground-state reaction, the electron flow of interest is HOMO(one component) → LUMO(the other component) — the two frontier orbitals whose overlap forms the new bond(s). For a cycloaddition between two independent π systems, use HOMO of one partner and LUMO of the other; for a unimolecular process like an electrocyclization, use the HOMO of the single conjugated π system (since bond reorganization is intramolecular, both new "bonds" — really, the closing σ bond — come from the same orbital manifold).

Step 2 — Draw the actual nodal pattern of that frontier orbital at the reacting termini. This step requires explicit orbital counting, not memorization: for a linear polyene of $N$ conjugated p orbitals, the $\psi_j$ molecular orbital has $j-1$ nodes, with coefficients on atom $r$ given by the standard Hückel eigenvector

c_{jr} = \sqrt{\frac{2}{N+1}} \sin\!\left(\frac{j r \pi}{N+1}\right)

where: $N$ is the number of conjugated atoms, $j$ is the MO index (1 = lowest energy), $r$ is the atom position (1 to $N$ ). Fill $N$ electrons pairwise from $\psi_1$ upward to find the HOMO index $j_{\max} = N/2$ (even $N$ , neutral polyene). The relative sign of $c_{j,1}$ and $c_{j,N}$ — the two terminal lobes — determines whether the ends of the chain present matching or mismatched phase for ring closure.

Step 3 — Determine geometrically whether matching termini require con- or dis-rotatory closure, or suprafacial/antarafacial cycloaddition. If the terminal HOMO lobes have the same sign on the same face (e.g., both top lobes positive), the termini can bond by rotating in opposite directions (disrotatory), because that motion brings same-phase lobes together while keeping both carbons' rotation moving their top lobes toward each other consistently; this is a suprafacial-suprafacial-type closure with even $N_a$ . If the terminal lobes have opposite sign on the same face, only rotation of both termini in the same direction (conrotatory) brings same-phase lobes into bonding overlap — one terminus must present its bottom lobe to bond, which is the antarafacial-like twist that defines conrotation as the odd- $N_a$ , Möbius-consistent pathway.

Step 4 — Repeat for cycloadditions using both HOMO and LUMO simultaneously. For a [4+2] cycloaddition, overlay the diene HOMO ( $N=4$ , $j=2$ , one internal node, terminal lobes of opposite sign, top face) with the dienophile LUMO (an isolated π orbital, top lobe and bottom lobe of opposite sign by definition, i.e. its own single node). Sliding the dienophile underneath the diene termini, both new σ bonds (C1–C1′ and C4–C4′) form with matching* phase simultaneously when both components react suprafacially — because the diene HOMO's two opposite-sign termini pair naturally with the dienophile LUMO's two opposite-sign lobes on its two faces. No twist is required: $N_a = 0$ .

Step 5 — Apply the photochemical inversion. Under photoexcitation, one component's electron is promoted from HOMO to LUMO. The new reacting (singly occupied, now frontier-relevant) orbital is the old LUMO, which by construction has one more node than the old HOMO. One extra node at the reacting termini flips the relative phase-matching condition worked out in Steps 3–4: whichever rotation sense (or facial combination) was required thermally is now phase-mismatched, and the opposite one (previously forbidden) is now the phase-matched, allowed pathway.

Step 6 — Show numerical agreement with the aromatic transition state count. For the disrotatory butadiene→cyclobutene case: total electrons in the array = 4 (the two π bonds of the diene, viewed as the reacting system, contribute the relevant HOMO), $N_a = 0$ (disrotatory closure keeps both termini using the same face). By the master rule, $4n$ electrons with $N_a$ even is the forbidden Hückel/antiaromatic combination — so thermal electrocyclization of a 4-electron system should be conrotatory ( $N_a = 1$ , odd, Möbius/aromatic, allowed), matching Step 3's finding that a diene HOMO with opposite-sign termini requires conrotatory closure. The frontier orbital route and the aromatic transition-state route agree exactly.

Limiting-case check. Take $N = 6$ (hexatriene → cyclohexadiene): the HOMO is $\psi_3$ , which by the nodal formula has 2 internal nodes and terminal lobes of the same sign (same face). Step 3 then predicts disrotatory closure — $N_a = 0$ even — consistent with 6 electrons being 4n+2, thermally allowed with an even (zero) antarafacial count. This is the opposite stereochemical outcome from butadiene despite both being "electrocyclizations," precisely because the electron count differs by four, flipping the terminal-lobe symmetry of the HOMO from opposite-sign (4 electrons) to same-sign (6 electrons).

Worked examples

Example 1 — Foundational: thermal disrotatory ring-opening of trans-3,4-dimethylcyclobutene. Cyclobutene ring-opening is the microscopic reverse of a 4π electrocyclization, so the same selection rule applies to both directions (microscopic reversibility guarantees the same orbital-symmetry constraint). The breaking σ bond and the two π bonds of the developing diene form the cyclic array: total electrons = 4 (2 from the σ bond being broken, effectively correlating to the diene π system as it opens — conventionally counted as the 4 electrons of the resulting butadiene π system for consistency with the closure direction). By the master rule, 4 electrons with $N_a$ even is forbidden (Möbius/aromatic requires $N_a$ odd at 4 electrons), so the thermally allowed pathway has $N_a = 1$ : conrotatory. Take trans-3,4-dimethylcyclobutene as the substrate — an asymmetric, non-special case (unequal substituents would break symmetry further, but even the trans-dimethyl case is decisive because the two rotation senses are stereochemically distinguishable). Conrotatory opening means both C3 and C4 rotate in the same sense (both clockwise or both counterclockwise, viewed from outside the ring). Starting from the trans diastereomer, with methyls on opposite faces, conrotatory rotation forces one methyl group to rotate "outward" (away from the forming diene plane, ending up in an E-like exterior position) while the other necessarily rotates "inward" is geometrically impossible for same-sense rotation of substituents that started trans — working through the actual geometry (both groups rotate the same way, e.g. both clockwise when viewed along the C3–C4 axis from the same end), a trans substitution pattern conrotates to give the (E,Z) — practically, the singly E,E or E,Z — diene, and experimentally trans-3,4-dimethylcyclobutene opens thermally to give predominantly the (E,E)-2,4-hexadiene, with the cis isomer instead giving the (E,Z) diene. Both results are the classic experimental confirmation (Winstein, and Woodward–Hoffmann's own analysis) that conrotation, not disrotation, operates thermally for this 4-electron system, exactly as the electron count predicts.

Example 2 — Applied: photochemical [2+2] cycloaddition in a synthesis step (thymine dimer / cyclobutane formation). [2+2] cycloadditions between two isolated alkenes are the textbook thermally forbidden pericyclic reaction: total electrons = 4 (2 + 2 from the two π bonds), and the only sterically reasonable geometry is fully suprafacial-suprafacial approach ( $N_a = 0$ , even) — by the master rule this is the forbidden Möbius-mismatched combination (4 electrons wants $N_a$ odd to be allowed). This is precisely why simple alkenes do not thermally dimerize to cyclobutanes under ordinary thermal conditions, despite the reaction being exothermic; the barrier from orbital symmetry mismatch dominates. Under UV irradiation (a concrete, asymmetric example: two thymine bases stacked in DNA, absorbing around 260–270 nm), one alkene is promoted to its π,π excited state. Its reacting frontier orbital is now the (originally) LUMO, π, which has a node between the two carbons rather than the uniform-phase π HOMO. Overlaying this excited-state π (top lobe positive on C1, negative on C2, say) with the ground-state alkene's LUMO in a face-to-face suprafacial-suprafacial approach now gives matching phases at both new bond-forming termini simultaneously — because the extra node in the excited component exactly compensates for the phase requirement that was mismatched thermally. This is the photochemical partner rule: 4 electrons, $N_a = 0$ even, is photochemically allowed* (the complementary rule flips 4n from thermally-forbidden-when-even- $N_a$ to photochemically-allowed-when-even- $N_a$ ). The result is the cyclobutane ring — the notorious cyclobutane pyrimidine dimer that is the principal UV-induced DNA lesion, formed with complete suprafacial, cis-ring-fusion stereochemistry exactly as orbital symmetry dictates, and repaired in cells by photolyase, an enzyme that runs the reaction's radical-mediated reverse.

Example 3 — Exam/competition level: electron count and stereochemical prediction for a [1,7]-sigmatropic hydrogen shift. Consider the thermal [1,7]-H shift interconverting previtamin D3 and vitamin D3 (calciferol) — a genuine synthetic/biological case with an odd, non-round electron count that forces careful bookkeeping rather than pattern-matching. In a [1,7] shift, a σ C–H bond migrates across a conjugated π system spanning the intervening atoms; here the migrating group moves from one end of a triene to the far end, seven atoms away along the chain. Count electrons in the cyclic array: the migrating σ bond contributes 2 electrons, and the conjugated π system it moves across contributes 2 electrons per double bond. A [1,7] shift moves across three intervening double bonds (this is what makes it a "1,7" shift rather than "1,3" or "1,5" — count atoms 1 through 7 along the chain, with new bonds forming at positions 1 and 7): 2 (σ) + 2 + 2 + 2 (three π bonds) = 8 electrons. Since $8 = 4n$ with $n=2$ , the thermally allowed pathway requires $N_a$ odd — at least one component, most practically the migrating hydrogen's σ bond or (far more common structurally) the polyene π system itself, must react antarafacially. For a hydrogen migrating on a long, flexible triene, antarafacial use of the π system is geometrically very feasible: the chain is long and floppy enough to allow the π system to present its top lobe at one terminus and its bottom lobe at the far terminus without excessive strain — unlike a [1,3]-H shift (2 σ + 2 π = 4 electrons, same $4n$ requirement for $N_a$ odd, but there the π system spans only one double bond, far too short and rigid to twist into an antarafacial geometry, which is why thermal [1,3]-H shifts are essentially never observed suprafacially-antarafacially and instead proceed by stepwise radical pathways or not at all). Because the hydrogen atom itself is a spherical 1s orbital (no p-orbital lobes, hence cannot itself present "two faces"), it is essentially always the σ component that stays suprafacial, forcing the π system to supply the required antarafacial twist. For the [1,7] shift in the previtamin D3 / vitamin D3 system, the seven-atom, three-double-bond chain is long enough to accommodate this antarafacial helical twist, and the thermal [1,7]-H shift proceeds readily at physiological and mild thermal temperatures with the predicted suprafacial (H, on σ) / antarafacial (on the triene) geometry — precisely the $N_a=1$ (odd), 8-electron (4n), thermally-allowed combination.

Limiting cases, scope & common traps

Electron count	Topology (N_a)	Thermal	Photochemical
4n+2 (2, 6, 10, …)	Hückel ( $N_a$ even)	Allowed	Forbidden
4n+2 (2, 6, 10, …)	Möbius ( $N_a$ odd)	Forbidden	Allowed
4n (4, 8, 12, …)	Hückel ( $N_a$ even)	Forbidden	Allowed
4n (4, 8, 12, …)	Möbius ( $N_a$ odd)	Allowed	Forbidden

Situation	What actually happens	Why the naive rule can mislead
Small-ring electrocyclizations (4–6π)	Almost always all-suprafacial for the component geometry that matters most (disrotatory/conrotatory refers to the σ-bond termini, not a separate "antarafacial π" twist)	Students sometimes hunt for a literal antarafacial π component in a 6π electrocyclization; the twist here is expressed through rotation sense of the termini, not a physically twisted extended π system
Long-chain sigmatropic shifts ([1,7], [1,9])	Antarafacial π components become geometrically realistic	Below about [1,5], antarafacial geometry is too strained; forbidden pathways are then simply not observed via the concerted route, not "somehow allowed anyway"
Metal-catalyzed formally-forbidden cycloadditions ([2+2] with Ni, Pd, Ti)	Proceed readily at mild temperature	The metal changes the mechanism to a stepwise oxidative-addition/reductive-elimination sequence via metallacycles; orbital symmetry rules apply only to concerted, uncatalyzed pericyclic pathways, not to stepwise organometallic ones
Radical or ionic pericyclic-looking reactions	Can appear to violate the rules	If any discrete intermediate (radical, cation, anion) is on the pathway, the reaction is not pericyclic at all, and Woodward–Hoffmann analysis simply does not apply
Retro-pericyclic reactions	Obey the identical rule as the forward reaction	Microscopic reversibility guarantees the transition state — and hence its orbital symmetry classification — is shared by both directions

The chemist behind it

Robert Burns Woodward (1917–1979) was already the most celebrated synthetic organic chemist of his generation — having completed total syntheses of quinine, cholesterol, cortisone, strychnine, chlorophyll, and, with the vitamin B12 project underway, was pushing the limits of what conscious synthetic planning could achieve — when a specific stereochemical anomaly in a B12 intermediate's electrocyclization forced him to ask a genuinely new theoretical question rather than simply another synthetic one. Roald Hoffmann (b. 1937), a young theoretical chemist trained under William Lipscomb and then working with Lionel Salem and E. J. Corey's group at Harvard, had been developing the extended Hückel method to compute approximate molecular orbitals for organic reaction pathways, including explicit calculations tracking how MO nodal patterns change as a ring opens. Their 1965 collaboration — a famous, fast, and famously productive partnership — produced the original electrocyclization paper within weeks, and a systematic general treatment (culminating in the 1969 monograph The Conservation of Orbital Symmetry) followed within four years. Hoffmann shared the 1981 Nobel Prize in Chemistry with Kenichi Fukui, whose independent frontier-orbital theory provided the alternative derivation route used throughout this article; Woodward, who died in 1979, was not eligible for the shared prize but had already received his own Nobel Prize in Chemistry in 1965, awarded for the totality of his synthetic achievements.

Origins · 1 minRoald Hoffmann and how this was discovered

Roald Hoffmann

b. 1937 · Cornell University, theoretical/computational organic chemistry, 1960s–present

Hoffmann's contribution was to make orbital symmetry conservation computable and diagrammable rather than a qualitative hand-wave: correlation diagrams tracking every reactant MO through to a product MO by symmetry, laid directly on top of Woodward's synthetic intuition, converted a single puzzling stereochemical result into a fully general predictive theory spanning electrocyclizations, cycloadditions, sigmatropic shifts, and group transfer reactions.

History

- 1937 — Roald Hoffmann born in Złoczów, Poland (now Ukraine); survives the Holocaust in hiding before emigrating to the United States.

- 1965 — Woodward and Hoffmann publish the first orbital-symmetry analysis of electrocyclic ring closure, triggered by a stereochemical puzzle in the vitamin B12 synthesis.

- 1965 — Woodward awarded the Nobel Prize in Chemistry for his contributions to the art of organic synthesis.

- 1969 — Woodward and Hoffmann publish The Conservation of Orbital Symmetry, generalizing the rules to cycloadditions and sigmatropic shifts via correlation diagrams.

- 1981 — Hoffmann and Fukui jointly awarded the Nobel Prize in Chemistry for independently developed theories (orbital symmetry conservation and frontier molecular orbital theory) explaining chemical reactivity.

What exactly counts as "one component" when counting electrons in the pericyclic array?

A component is any bond or lone pair whose orbital(s) participate directly in the bond reorganization of the transition state — typically a π bond (2 electrons), a σ bond that breaks or migrates (2 electrons), or a heteroatom lone pair that donates into the array (2 electrons). Substituent groups not directly involved (like a remote methyl) are not counted. In practice, you identify components by drawing the full curved-arrow mechanism first: every arrow's tail sits on a component. Add up 2 electrons per component to get the total, then separately classify each component's facial usage to get $N_a$ .

Does the Woodward–Hoffmann framework apply to reactions with charged intermediates, like E1 eliminations?

No — orbital symmetry conservation, in the pericyclic sense used here, strictly applies only to concerted processes with a single cyclic transition state and no intermediate of any kind (no carbocation, carbanion, or radical sitting in an energy well along the path). E1 eliminations proceed through a discrete carbocation intermediate and are governed by ordinary cation stability and Zaitsev/Hofmann regiochemistry, not by the electron-counting rules here. Distinguishing a genuinely pericyclic mechanism from a stepwise one that merely resembles it (same overall bonds broken and formed) is the essential first step before applying any Woodward–Hoffmann analysis.

How is "antarafacial" different for a σ bond compared to a π bond?

For a π bond, the two faces are simply the top and bottom lobes of the p orbital, and antarafacial use means the new bonds form on opposite faces of that same p orbital — visually a twist of the π system. For a σ bond, the analogous "faces" are the two lobes of the σ (or σ*) orbital pointing away from each atom along the bond axis; antarafacial use of a σ bond means the incoming and outgoing groups attack from the backside of the bond at one terminus and the front side at the other, which is exactly the backside/inversion versus frontside/retention distinction familiar from nucleophilic substitution stereochemistry, now applied within a pericyclic sigmatropic framework.

Can a reaction be "forbidden" thermally but still occur readily at room temperature?

Yes, and this is one of the most common points of confusion. A thermally forbidden concerted pathway has a very high activation energy for that specific mechanism, but the same overall transformation can often occur readily through a completely different, stepwise (diradical or zwitterionic) mechanism that is not pericyclic at all and therefore is not subject to the orbital symmetry constraint. Many apparently "forbidden" [2+2] cycloadditions that nonetheless proceed thermally under forcing conditions are understood to have switched to such a stepwise diradical pathway rather than violating orbital symmetry conservation.

Why do correlation diagrams give the same answer as electron counting?

A correlation diagram tracks each individual reactant molecular orbital, ordered by energy and classified by its symmetry (with respect to whatever symmetry element — a mirror plane or C2 axis — survives throughout the entire reaction path), through to the product molecular orbital of matching symmetry. The reaction is allowed exactly when every occupied reactant MO correlates, without crossing through a high-energy orbital, to an occupied product MO — this "no crossing of occupied levels through the transition state" condition is mathematically the same requirement as demanding that the cyclic array of interacting orbitals close with continuous, low-energy (aromatic) phase matching, which is precisely what the electron-counting rule tests directly and more quickly.

Do the Woodward–Hoffmann rules make quantitative predictions, or only allowed/forbidden qualitative ones?

In their original and most widely used form, the rules are qualitative — they sort pathways into "orbital-symmetry-allowed, low barrier" versus "orbital-symmetry-forbidden, high barrier," and correctly predict stereochemical outcomes (which rotation sense, which face) with essentially perfect reliability across thousands of characterized pericyclic reactions. Quantitative barrier heights and rate constants require full computational treatment (transition-state theory calculations with density functional theory or higher-level methods), but the allowed/forbidden sorting from orbital symmetry alone reliably predicts, before any calculation, which of two competing concerted pathways will actually be observed.