Restricted Rotation and Molecular Rigidity

Restricted rotation about π-bonded systems is not a museum piece of stereochemistry — it is load-bearing machinery in modern chemistry and biology. Vision depends on it directly: the retinal chromophore in rhodopsin exists in a bent, 11-cis configuration in the dark, and absorption of a single photon supplies enough energy to overcome the C11=C12 π-bond rotational barrier, driving isomerization to the all-trans form; this single restricted-rotation event, reversed by light, is the literal first step of the visual transduction cascade in every rod and cone cell. Protein structure depends on the amide bond's partial-double-bond rigidity (worked out quantitatively above): because the peptide C–N bond is planar and rotation is slow, the protein backbone can be described, to excellent approximation, as a chain of rigid, flat peptide-plane units connected by freely rotating $\phi$/$\psi$ single bonds at the $\alpha$-carbons — the entire Ramachandran-plot framework for predicting protein secondary structure rests on this rigidity assumption. In materials science, molecular photoswitches built on azobenzene ($\text{Ar–N=N–Ar}$) and stilbene ($\text{Ar–CH=CH–Ar}$) exploit exactly this restricted-rotation energy landscape: the thermally stable trans form and the higher-energy, non-interconverting-at-RT cis form can be toggled by light absorption (which populates the twisted, near-orthogonal excited state where π overlap and hence the energy barrier separating the two ground-state geometries collapses), giving reversible, addressable molecular switches used in smart materials, drug-delivery triggers, and molecular machines (work recognized in the 2016 Nobel Prize in Chemistry for molecular machines). In drug design, medicinal chemists deliberately lock double bonds or amide bonds in specific configurations to enforce a bioactive conformation — for example, many kinase inhibitors and macrocyclic drugs use rigid alkene or (E)-configured amide linkers specifically because restricted rotation removes the entropic penalty of searching conformational space, pre-organizing the molecule for tight binding. Even industrial polymer chemistry depends on it: the mechanical properties of natural rubber (all-cis polyisoprene, soft and stretchy) versus gutta-percha (all-trans polyisoprene, rigid and brittle) trace directly back to the fixed, non-interconverting cis/trans geometry frozen in at every double bond along the polymer backbone during synthesis.

Ask the tutor: How does this same p-orbital overlap argument extend to explain why allylic and benzylic systems are stabilized by conjugation, even though those single bonds do rotate freely?

Start with the σ framework. Each alkene carbon is $sp^2$ hybridized: three $sp^2$ orbitals at 120° in a plane, each forming a σ bond (two to substituents, one to the other alkene carbon), plus one unhybridized $2p$ orbital left over, oriented perpendicular to that plane. This leftover $p$ orbital has a node in the $sp^2$ plane and two lobes, one above and one below. When two such carbons are joined, their $p$ orbitals — if kept parallel — overlap sideways above and below the internuclear axis, building a π bonding molecular orbital (in-phase, bonding) and a π* antibonding MO (out-of-phase). Two electrons occupy the bonding π MO, giving a π bond with substantial electron density above and below the molecular plane but with a node in that plane. The crucial geometric fact is that π overlap is a sideways, "face-to-face" overlap, and the magnitude of that overlap depends on how parallel the two $p$ orbital axes remain. If you twist one $\text{CH}_2$ group relative to the other by dihedral angle $\theta$ about the C–C axis, the $p$ orbital on that carbon rotates with it (it is rigidly attached to the $sp^2$ framework at that carbon). The overlap between the two $p$ orbitals scales as the cosine of the angle between their axes:

where $S_\pi(0)$ is the full overlap when the orbitals are perfectly parallel ($\theta = 0$, the planar, undistorted alkene) and $\theta$ is the twist angle about the former double-bond axis. At $\theta = 90°$, $\cos 90° = 0$: the overlap vanishes identically. The two $p$ orbitals are now orthogonal, sitting at right angles to each other, and there is zero net bonding interaction between them — the π bond is completely gone. This is fundamentally different from a σ bond, where rotation about the bond axis does not change the head-on overlap at all (a σ bond is cylindrically symmetric about the internuclear axis, so any dihedral angle gives the same σ overlap). Because breaking the π bond costs real bond energy (not just a steric clash), the potential energy as a function of twist angle $\theta$ rises steeply and smoothly from the planar minimum to a maximum at $\theta = 90°$ — the point of an isolated, non-interacting, twisted ("perpendicular") alkene, sometimes called a pyramidalized or twisted excited-state-like structure. This barrier, roughly equal to the π bond dissociation energy, is what makes cis and trans alkenes configurationally stable, rather than rapidly-averaging conformers.

The consequence for substituents is immediate. Because the $sp^2$ carbons and their attached σ-bonded substituents all lie in a single plane (dictated by the 120° trigonal geometry at each alkene carbon), and because that planar arrangement is locked by the π bond, the four groups attached to a C=C are frozen into one rigid, flat unit. Two arrangements are possible for a disubstituted alkene: the two substituents (say, the two —CH₃ groups of 2-butene) can point to the same side (cis, or Z when priority rules agree) or to opposite sides (trans, or E). Converting one into the other requires passing through the 90°-twisted, π-bond-broken transition state — an enormously costly process at room temperature.

The single governing relationship is the angular dependence of π overlap on twist angle, and the resulting torsional potential energy curve.

where:

• $S_\pi(\theta)$ is the π-orbital overlap integral at twist angle $\theta$,
• $S_\pi(0)$ is the maximum overlap in the planar (untwisted) geometry,
• $\theta$ is the dihedral twist angle about the C=C axis, measured from planarity.

Reading: overlap — and with it π bond strength — falls off as the cosine of the twist angle, reaching exactly zero at 90°. This is a purely geometric consequence of the fact that $p$ orbitals are directional (they have a node through the nucleus) rather than spherically symmetric. Because the π bond energy is (to a first approximation, within simple Hückel-type reasoning) proportional to the overlap, the torsional potential energy takes the form

where:

• $E(\theta)$ is the strain energy of twisting relative to the planar minimum,
• $E_\pi$ is the full π-bond strength (the energy released by forming the π bond from two isolated $p$ electrons),
• the factor $(1-\cos\theta)$ rises from 0 at $\theta=0$ to 1 at $\theta = 90°$ (already all the way "broken" by then) and to 2 at $\theta = 180°$ (equivalent by symmetry back to the planar form, having passed through the barrier).

Reading: the maximum rotational barrier, at $\theta = 90°$, equals essentially the full π bond dissociation energy — twisting a double bond by 90° is energetically equivalent to homolytically breaking the π bond, because that is literally what has happened to the orbital overlap. Finally, the measured thermodynamic barrier connects to observable rate behavior through the Eyring relation, giving the interconversion rate constant:

where:

• $k$ is the rate constant for cis–trans interconversion (bond rotation),
• $k_B$, $h$, $R$ are Boltzmann's constant, Planck's constant, and the gas constant,
• $T$ is absolute temperature,
• $\Delta G^\ddagger$ is the free energy of activation for rotation, essentially $E_\pi$ for a simple alkene.

Reading: with $\Delta G^\ddagger \approx 65$ kcal/mol, the exponential term is so vanishingly small at 298 K that $k$ corresponds to a half-life vastly longer than the age of the universe — cis and trans alkenes simply do not interconvert thermally at room temperature.

Example 1 — Foundational: quantifying the barrier and confirming the cosine law numerically. Ethylene's experimentally measured barrier to rotation about the C=C bond (from thermal and photochemical cis–trans isomerization studies) is $\Delta G^\ddagger \approx 65$ kcal/mol, essentially the strength of the isolated π bond (the total C=C bond energy, about 174 kcal/mol, minus the σ bond contribution of roughly 90 kcal/mol, leaves ≈ 65–84 kcal/mol depending on the partitioning scheme used; we use 65 kcal/mol as the commonly cited π-component figure). Using the overlap law from Mechanism 1, compute the fractional π overlap and the fractional π-bond energy remaining at an off-special twist angle of $\theta = 37°$ (chosen non-round, to test the formula honestly rather than at a convenient value like 30°, 45°, or 60°). Overlap fraction:

If we treat the π-bond energy as scaling with the square of the overlap integral (as is standard in a two-orbital, two-electron treatment where bonding stabilization $\propto S^2/(1-S^2)$, approximated here for simplicity as $\propto S^2$ for a rough energetic estimate):

So at a 37° twist, only about 64% of the π bond strength survives — roughly $0.638 \times 65 \text{ kcal/mol} \approx 41.4$ kcal/mol of π stabilization remains, meaning the molecule has already lost about 23.6 kcal/mol of π stabilization for a fairly modest 37° twist. This illustrates how steeply the energy rises even before reaching the full 90° twist, consistent with the alkene resisting any significant rotation, not just the full 90°. Limiting-case check: as $\theta \to 0$, $\cos^2\theta \to 1$, recovering full π-bond strength (the planar ground state) — consistent, since no twisting should mean no loss of stabilization.

Example 2 — Applied: distinguishing maleic acid and fumaric acid by rigidity, and predicting a real chemical consequence. Maleic acid (cis-butenedioic acid) and fumaric acid (trans-butenedioic acid) are geometric isomers of $\text{HOOC–CH=CH–COOH}$ that do not interconvert at room temperature — direct evidence of restricted rotation. Their properties diverge sharply because of that fixed geometry:

• In maleic acid, the two —COOH groups are cis (same side), forced close together by the rigid, planar C=C framework. This proximity allows maleic acid to lose water intramolecularly on heating, forming maleic anhydride, a five-membered cyclic anhydride, because the two carboxyl oxygens can reach each other across the short "same-side" gap.
• In fumaric acid, the two —COOH groups are trans (opposite sides). Geometrically, they point away from each other across the rigid double bond and simply cannot approach closely enough to cyclize; fumaric acid does not form a cyclic anhydride on simple heating.

Mechanistically: the anhydride-forming step requires one carboxylic acid oxygen's lone pair to attack the carbonyl carbon of the other —COOH group (an intramolecular nucleophilic acyl substitution, forming a 5-membered ring transition state). This attack trajectory is only geometrically accessible when the two carboxyls are cis, because the alkene's planarity and rigidity fix the through-space distance between the two functional groups — a distance that free rotation (were it available) could otherwise have averaged out or optimized on the fly. Because the C=C cannot rotate, the cis geometry is a permanent structural feature, not a transient conformation, and it directly controls which reaction pathway is geometrically feasible. This is a clean illustration of how restricted rotation converts an abstract stereochemical label (cis/trans) into a concrete difference in reactivity, melting point (maleic acid mp 130 °C vs. fumaric acid mp 287 °C, reflecting fumaric acid's more symmetric, better-packing crystal lattice), and even acid strength ($K_{a1}$ of maleic acid is notably larger than fumaric acid's, because the cis carboxyl can intramolecularly hydrogen-bond to stabilize the resulting monoanion — again, only possible because the geometry is locked cis and cannot rotate to relieve or exploit that proximity).

Example 3 — Exam/competition level: predicting and ranking rotational barriers across a conjugated/amide series, with quantitative reasoning. Three bonds are given, and you are asked to rank their resistance to rotation from highest to lowest, justifying each with an orbital argument, and to estimate the barrier for the middle case using bond-order reasoning: (a) the central C–C bond of 1,3-butadiene ($\text{CH}_2=\text{CH–CH}=\text{CH}_2$), (b) the C–N bond of the amide (peptide) group in N-methylacetamide, $\text{CH}_3\text{CONHCH}_3$, (c) the C–C bond of ethane.

Reasoning for (c): pure $sp^3$–$sp^3$ σ bond, no π component at all. Rotational barrier ≈ 2.9 kcal/mol (torsional/hyperconjugative origin only, per Mechanism 2). This is the lowest barrier of the three — essentially free rotation at room temperature, with a rotational rate constant (via the Eyring relation) many orders of magnitude per second.

Reasoning for (a): the central bond of butadiene is formally drawn as a single bond, but each carbon is $sp^2$ hybridized with a leftover, unhybridized $p$ orbital that is part of the conjugated π system spanning all four carbons. Rotating about the central bond partially breaks the conjugative overlap between the two alkene π systems (it does not break a full, localized π bond, since there was never a full π bond there to begin with — the central bond has partial double-bond character from conjugation, roughly bond order 1.1–1.2). The barrier to rotation about this bond is measured at only about 4.9 kcal/mol — larger than ethane's because of residual conjugative stabilization lost on twisting, but far smaller than a true π bond's ~65 kcal/mol, because there is no full, localized π bond centered on that particular bond to begin with. This bond is intermediate.

Reasoning for (b), with quantitative estimate: the amide C–N bond looks, from a naive Lewis structure, like a pure single bond (nitrogen's lone pair drawn as non-bonding). But nitrogen's lone pair sits in a $p$-type orbital that is efficiently positioned, by the planar, $sp^2$-like geometry adopted at nitrogen, to overlap sideways with the carbonyl π* orbital — an $n_N \to \pi^*_{C=O}$ conjugative donation exactly analogous to allylic/conjugation overlap. This donation places significant double-bond character on the C–N bond, describable by the resonance structure

Because that resonance donation requires the nitrogen lone-pair's $p$ orbital to stay parallel to the carbonyl π system (exactly the same $\cos\theta$ overlap law as Mechanism 1, now applied to $n$–π* conjugation rather than π–π overlap), rotating the C–N bond by 90° would destroy this stabilization entirely, just as rotating a true alkene would destroy π bonding. The measured rotational barrier for a typical secondary amide is about 15–20 kcal/mol — much higher than butadiene's 4.9 kcal/mol (because amide resonance donation is a stronger, more complete "partial π bond" than simple diene conjugation) but still well below ethylene's ~65 kcal/mol full π bond (because the amide bond retains substantial single-bond character; the resonance structure with formal C=N is a minor, not dominant, contributor).

Final ranking: amide C–N bond (b, ≈15–20 kcal/mol) > butadiene central bond (a, ≈4.9 kcal/mol) > ethane C–C bond (c, ≈2.9 kcal/mol), all far below a true alkene π bond (≈65 kcal/mol). This ranking is exactly why the amide bond is treated, in protein structure, as effectively planar and rigid on the biological timescale (rotation is slow enough that the cis and trans amide conformers of a peptide bond are treated as distinct, slowly-interconverting states — crucial in Molecular Shape and VSEPR Theory discussions of protein backbone geometry), while the diene's central bond, though somewhat restricted, still allows the s-cis/s-trans conformational equilibrium to be established rapidly at room temperature.

Understanding restricted rotation as an orbital phenomenon is the product of the broader development of molecular orbital theory in the mid-20th century, but the man whose foundational insight into hybridization and orbital overlap made this whole line of reasoning possible was Linus Pauling.