Sigma and Pi Bonding

The σ/π distinction is not a historical curiosity confined to the classroom — it is load-bearing machinery in modern chemistry and biology. In vision, the retinal chromophore in rhodopsin absorbs a photon that promotes an electron within its extended π system, momentarily weakening a specific π bond enough that thermal rotation about that bond becomes feasible on a biological timescale; this cis-to-trans isomerization of retinal, made possible only because photoexcitation temporarily nullifies π bond order at that position (exactly the $S_\pi \to 0$ physics derived above, achieved electronically rather than mechanically), is the literal first step of animal vision. In synthetic dye and solar-cell chemistry, extended π-conjugated systems (polyenes, porphyrins, phthalocyanines) are engineered so that π molecular orbitals — built from exactly the sideways p-orbital overlaps described here, extended over many centers — have small enough HOMO-LUMO gaps to absorb visible light, the working principle behind dye-sensitized solar cells and organic LEDs. In catalysis, understanding which bonds are σ (rotatable, robust) and which are π (rigid, reactive) is central to designing catalysts: alkene and alkyne π bonds are the reactive handles in hydrogenation, metathesis, and polymerization catalysis (Ziegler-Natta and metallocene catalysts insert monomers by attacking the π system while leaving the σ backbone of the growing polymer chain untouched), while the ubiquity of freely-rotating σ C–C and C–N single bonds is what gives proteins and synthetic polymers their conformational flexibility — the rotatable backbone dihedral angles (φ, ψ) of a protein backbone are, mechanistically, exactly the same "σ bonds rotate freely" physics as ethane's near-zero barrier, while the planar peptide bond (a partial C–N π bond from amide resonance) is rigid for precisely the same $\cos\theta$ reasons that make ethylene planar — this rigidity of the peptide bond is essential to protein secondary structure (α-helices, β-sheets) being well-defined at all. In inorganic and organometallic chemistry, metal–metal multiple bonding (as in [Re₂Cl₈]²⁻ and related quadruply-bonded dimetal complexes) extends the σ/π/δ classification to d-orbital systems and underlies an entire subfield of cluster and multiple-bond chemistry, feeding directly into the electron-counting and bond-order arguments developed further in Bond Order, Length, and Energy and the orbital-energy-level diagrams of Homonuclear Diatomic MOs.

:::quiz[What extra bonding interaction, beyond σ and two π bonds, appears in the Re–Re quadruple bond of [Re₂Cl₈]²⁻?]

• a second independent σ bond from an s orbital
• a δ bond from face-to-face dxy–dxy overlap with two nodal planes ✓
• a hydrogen-bridging bond between the two rhenium centers
• an ionic bond from full electron transfer between the two metals

Ask the tutor: How does the σ/π picture change once we build the full molecular-orbital diagram for a diatomic like O₂, where π* orbitals end up singly occupied and directly explain paramagnetism?

Start with two atoms held a fixed distance apart along what we call the internuclear axis, conventionally the z-axis. Orbitals overlap constructively to form a bond only if they have matching symmetry with respect to that axis, and the way they overlap sorts naturally into two classes. A σ bond forms when the overlapping lobes point directly along the bond axis: s–s, s–pz, pz–pz (head-on), or two sp/sp²/sp³ hybrid lobes pointing at each other. The resulting molecular orbital has no nodal plane containing the internuclear axis — cutting the molecule with any plane that contains the z-axis, the σ orbital looks the same electron density above and below that cut. This is exactly cylindrical symmetry: spin the whole molecule about the bond axis and the σ orbital is unchanged. Every single bond in every Lewis structure is a σ bond, and it is the σ framework that defines a molecule's overall shape (bond angles, VSEPR geometry) because it is built from the hybrid orbitals whose directions are fixed by hybridization. A π bond forms when unhybridized p orbitals (or, in transition-metal chemistry, d orbitals) on adjacent atoms lie parallel to each other, perpendicular to the bond axis, and overlap side-by-side above and below the axis. The resulting orbital does have a nodal plane containing the internuclear axis — the plane in which the two contributing p orbitals themselves have a node. Electron density piles up in two lobes, one above and one below (or one in front, one behind) the axis, with zero density exactly on the axis-containing nodal plane. This lower symmetry — a mirror plane through the axis flips the sign of the wavefunction — is the entire reason π bonds behave differently from σ bonds when the molecule twists. Consider ethylene, H₂C=CH₂. Each carbon is sp² hybridized: three sp² lobes in a plane, forming three σ bonds (two C–H, one C–C), leaving one unhybridized p orbital perpendicular to that plane on each carbon. When the two CH₂ planes are coplanar, the two perpendicular p orbitals are parallel and overlap maximally, forming the π bond — this is the planar, low-energy, "normal" structure of ethylene. Now imagine twisting one CH₂ group about the C–C axis by an angle θ. The C–C σ bond, built from sp² lobes pointing straight along the axis, does not care about θ at all — head-on overlap is unaffected by rotation about the axis connecting the lobe tips. But the two p orbitals, which were parallel at θ = 0°, become progressively less parallel as θ increases, and their sideways overlap integral shrinks. This is the mechanical picture behind the rotational barrier worked out quantitatively below.

The overlap integral between two orbitals $\phi_A$ and $\phi_B$ on adjacent atoms is

where:

• $\phi_A, \phi_B$ are the atomic orbitals contributing to the bond,
• $d\tau$ is a volume element,
• $S_{AB}$ is the (dimensionless) overlap, ranging roughly from 0 (no bonding) to 1 (complete coincidence).

Reading: the bond energy lowering in simple valence-bond/Hückel-type reasoning scales with $S_{AB}$ (more precisely with the resonance integral $\beta \propto S_{AB}$), so anything that drives $S_{AB} \to 0$ kills that bond's contribution to binding. For two p orbitals held a fixed internuclear distance apart but with their axes twisted by an angle $\theta$ relative to each other (one fixed in the yz-plane, the other rotated by $\theta$ about the shared internuclear z-axis), the π-type overlap integral factorizes into a radial part and an angular part:

where:

• $S_\pi(0)$ is the overlap at perfect parallel alignment (planar geometry),
• $\theta$ is the torsional (dihedral) twist angle about the internuclear axis,
• $S_\pi(\theta)$ is the overlap at twist angle $\theta$.

Reading: this cosine law is the mathematical fingerprint of a π bond — it says overlap (and hence π bond strength) falls off exactly as the projection of one p orbital onto the other, vanishing completely at $\theta = 90°$. Contrast this with σ overlap between two orbitals pointing along the bond axis itself, which is independent of $\theta$ to leading order because rotating a lobe about its own axis of pointing does not change how it overlaps head-on with a partner lobe on the same axis. The torsional (π) contribution to the energy as a function of twist angle, in the simplest one-π-electron-pair picture, follows the same cosine dependence (since the interaction energy in a two-orbital system scales with $S$ and hence with $\cos\theta$):

where:

• $E_\pi(0)$ is the π-bond stabilization energy in the planar (aligned) geometry — for ethylene this is essentially the π bond dissociation contribution, roughly 65 kcal/mol (272 kJ/mol) of the total ~174 kcal/mol C=C bond energy,
• $E_\pi(\theta)$ is the π stabilization remaining at twist angle $\theta$,
• the rotational barrier is $\Delta E_{rot} = E_\pi(0) - E_\pi(90°) = E_\pi(0)$, since $\cos 90° = 0$.

Reading: the entire rotational barrier of a double bond, in this picture, equals the π bond energy itself — twisting to 90° does not break the σ bond at all (it survives, holding the two carbons together, hence ethylene does not fall apart at the twisted geometry) but it completely destroys the π bond's stabilization, and re-forming it upon returning to planarity is exactly what the barrier measures.

Example 1 — Foundational: identifying σ and π bonds in N₂. Nitrogen gas, N₂, has the Lewis structure :N≡N: with a triple bond and one lone pair on each nitrogen. Each N atom is best described (in simple VB language) as sp hybridized: two sp lobes at 180°, one holding the lone pair and one pointing at the other N for the σ bond, leaving two unhybridized 2p orbitals ($2p_x$, $2p_y$) perpendicular to the bond axis and to each other. Overlap of the sp lobes head-on along the internuclear axis gives one σ bond; sideways overlap of the two perpendicular pairs of $2p_x$/$2p_x$ and $2p_y$/$2p_y$ orbitals gives two π bonds. Total: 1 σ + 2 π = triple bond, bond order 3, consistent with the very short N–N distance of 109.8 pm and the very large bond dissociation energy of 945 kJ/mol — the largest of any diatomic between two identical main-group atoms, precisely because it is stabilized by two full π bonds on top of the σ bond. As a limiting-case check: removing one π bond (as happens on going to N₂⁺, which loses an electron from the σ-bonding HOMO in the MO picture, but conceptually compare instead to a hypothetical N=N double-bonded species like diazene, HN=NH) leaves bond order 2 and a substantially longer, weaker N–N bond (~125 pm, ~456 kJ/mol) — confirming that each π bond contributes real, quantifiable shortening and strengthening, not merely bookkeeping.

Example 2 — Applied: the rotational barrier of ethylene and why it is not simply "one π bond's worth" numerically. Ethylene's measured barrier to internal rotation about the C=C bond is about 65 kcal/mol (272 kJ/mol) — chemists usually equate this directly with "the π bond energy," and to a good first approximation this is correct, since twisting to 90° should sever π overlap ($\cos 90° = 0$ from the derivation above) while leaving the σ bond and σ framework essentially intact. Let's check the numbers with asymmetric, real thermochemical data rather than round assumed values. The total C=C bond dissociation energy in ethylene (breaking both σ and π completely, to give two ·CH₂ fragments) is about 174 kcal/mol (728 kJ/mol). A typical C–C single bond energy (as in ethane) is about 90 kcal/mol (377 kJ/mol) — but this is not quite the right σ-only comparison, because the σ bond in ethylene forms between sp² carbons (more s-character, shorter, somewhat stronger) rather than sp³ carbons as in ethane. Using the crude subtraction

overestimates the measured rotational barrier of 65 kcal/mol by about 19 kcal/mol. The discrepancy is real chemistry, not an error: twisting ethylene to 90° does not merely turn off the π bond while leaving everything else unchanged — the two ·CH₂ radical-like fragments at the twisted transition state also relax geometrically (pyramidalize slightly, rehybridize) and are stabilized somewhat by the residual σ framework's adjustment, so the measured barrier (65 kcal/mol) is smaller than the naive bond-energy-subtraction estimate (84 kcal/mol). The limiting-case check: as the twist angle $\theta \to 0$, $E_\pi(\theta) = E_\pi(0)\cos\theta \to E_\pi(0)$, recovering full π stabilization and the planar ground state, exactly as observed — ethylene's equilibrium structure is planar with $\theta = 0$, not some intermediate twisted geometry, because $\cos\theta$ is maximized at $\theta=0$.

Example 3 — Competition-level: multiple metal–metal bonding in [Re₂Cl₈]²⁻ and the eclipsed geometry it forces. The rhenium–rhenium quadruple bond in the [Re₂Cl₈]²⁻ ion (Cotton's classic discovery) is the sharpest inorganic test of σ/π/δ bonding logic. Each Re center is square-planar-coordinated by four Cl⁻ ligands, and the two ReCl₄ units are connected along a Re–Re axis. Take the Re–Re axis as z. Each Re contributes a $d_{z^2}$ orbital pointing along the bond axis: head-on overlap of the two $d_{z^2}$ orbitals gives one σ bond. Each Re also contributes $d_{xz}$ and $d_{yz}$ orbitals, perpendicular to the axis; these overlap sideways in pairs ($d_{xz}$–$d_{xz}$ and $d_{yz}$–$d_{yz}$) to give two mutually perpendicular π bonds — exactly analogous to the two π bonds of a C≡C triple bond. That already accounts for a bond order of 3 (σ + 2π), matching a triple bond such as acetylene's. The distinguishing fourth bond comes from the $d_{xy}$ orbitals on each Re, which lie in the plane perpendicular to the Re–Re axis (the same plane as the four Cl ligands) rotated 45° from the Cl–Re–Cl directions. These $d_{xy}$ orbitals overlap face-to-face — all four lobes of one $d_{xy}$ meeting all four lobes of the other — forming a δ (delta) bond, with two nodal planes containing the internuclear axis (rather than the one nodal plane of a π bond). This gives Re≡Re (using ≡ loosely) a total bond order of 4: 1σ + 2π + 1δ. The key geometric consequence, which is the "exam-level" payoff: because the δ overlap requires the four $d_{xy}$ lobes on one Re to line up with the four $d_{xy}$ lobes on the other Re, δ overlap follows a $\cos(2\theta)$ dependence (each lobe pattern has four-fold, not two-fold, symmetry about the axis), so twisting one ReCl₄ unit by 45° relative to the other reduces the δ overlap to zero (since $\cos(2\times 45°) = \cos 90° = 0$) exactly as a 90° twist kills a simple π bond. This is why [Re₂Cl₈]²⁻ is observed crystallographically in the eclipsed (not staggered) conformation — the δ bond, uniquely among metal–metal multiple bonds, imposes a strong electronic preference for a specific torsional angle (θ = 0°, eclipsed) that has no analog in ordinary organic σ-only single bonds and is a stronger and more clearly "extra" rotational constraint than even ethylene's π-based barrier, because it stacks on top of an already-existing σ + 2π framework.

The σ/π language grew directly out of Linus Pauling's valence-bond theory and hybridization framework of the 1930s, formalized further by Erich Hückel's treatment of π electrons in conjugated and aromatic systems (1930-31), which explicitly separated the σ framework (treated as a fixed, non-reactive skeleton) from a delocalized π system (treated quantum mechanically as the seat of reactivity, color, and aromatic stabilization) — the very separation that makes Hückel's method tractable and that underlies simple aromaticity arguments to this day.