Sigma Bonds from Orbital Overlap

Every rotatable bond descriptor used in drug-design software — the count of "rotatable bonds" that predicts oral bioavailability via Lipinski/Veber-style rules — is a direct, practical application of the σ-bond cylindrical-symmetry principle: software identifies acyclic σ single bonds (excluding those with restricted rotation from conjugation, amide character, or ring membership) as the flexible hinges of a drug molecule, because those are exactly the bonds whose overlap-derived energy does not resist twisting. Polymer science relies on the same idea at a larger scale: the flexibility of a polyethylene chain, its ability to adopt random coil conformations that determine rubber elasticity and glass transition temperature, comes entirely from free rotation about the backbone C–C σ bonds, modulated only by the same gauche/anti steric and hyperconjugative effects worked out for butane. Engineering a polymer's stiffness (e.g., introducing aromatic rings or amide linkages, as in Kevlar) means deliberately breaking this free rotation by inserting bonds with partial π character or steric locks. In molecular spectroscopy and NMR, the distinction between σ-only bonds (fast rotation, averaged coupling constants and chemical shifts on the NMR timescale) and partial-double-bond character (slow rotation, observable rotamer populations, e.g., the two distinct N-methyl signals in DMF at room temperature from restricted amide rotation) is used daily to diagnose conjugation and electronic delocalization directly from a spectrum. In catalysis and organometallic chemistry, metal–ligand σ-bonds (e.g., M–alkyl, M–H) formed by head-on overlap between a metal d/s/p hybrid and a ligand orbital are the reactive sites for σ-bond metathesis and oxidative addition — reactions whose stereochemical outcomes are rationalized by tracking exactly which orbital lobes point where, the same overlap-geometry reasoning developed here for main-group σ bonds.

Ask the tutor: How does the overlap integral S change quantitatively with bond length, and why does that explain why shorter bonds are usually stronger?

Picture two atoms, A and B, separated by a distance $r$, each bearing one orbital that can point toward the other atom — a 1s orbital on hydrogen, or an sp³ lobe on carbon. As the atoms approach, these orbitals begin to overlap: the amplitude of orbital A's wavefunction at points near B becomes nonzero, and vice versa. Quantum mechanically, when this overlap is significant, the system's true eigenstates are no longer "electron on A" and "electron on B" separately but linear combinations of the two atomic orbitals — the molecular orbitals of the bond. The defining geometric feature of a σ interaction is that both orbitals have their major lobe directed along the line connecting the two nuclei, and neither orbital has a node containing that axis. An s orbital is spherically symmetric, so it always satisfies this trivially. A p orbital satisfies it only when it is the $p_z$ orbital oriented along the internuclear axis (call that axis $z$) — the $p_x$ and $p_y$ orbitals, whose lobes point perpendicular to the axis, cannot form σ overlap with anything sitting on the $z$-axis; they are exactly the orbitals available for π bonding instead. An sp³ hybrid, being a mixture of one s and three p orbitals arranged so that one lobe points straight at the neighboring atom, is automatically σ-symmetric in the bond-forming lobe. Once you fix the axis, look at what happens as you rotate one fragment about that axis, holding the axis fixed. An s orbital, being spherical, is completely unaffected — there is no azimuthal angle to speak of. A $p_z$ orbital pointed along $z$ is also unaffected by rotation about $z$: its shape (two lobes, one positive, one negative, stacked along $z$) has cylindrical symmetry around its own axis. So both orbitals that participate in σ bonding are themselves cylindrically symmetric about the bond axis, and any overlap integral built from them inherits that same cylindrical symmetry. This is the geometric seed from which free rotation grows.

Contrast that immediately with π overlap, built from two $p_x$ (or $p_y$) orbitals whose lobes sit off-axis, above and below (or in front of and behind) the internuclear line. That overlap region is not centered on the axis; it is a pair of lobes hovering to one side. If you rotate one $p_x$ orbital by 90° about the bond axis so that it becomes $p_y$, its lobes swing out of alignment with the other atom's $p_x$ lobes, and the overlap collapses toward zero. That is the entire mechanistic reason π bonds resist rotation and σ bonds do not: σ overlap is built from axially symmetric lobes, so overlap survives any rotation about the axis; π overlap is built from off-axis lobes, so overlap is destroyed by such a rotation. The second essential picture is the molecular-orbital diagram generated by any head-on overlap: two atomic orbitals combine to give one bonding MO (lower in energy, no node between the nuclei, density piled up in the internuclear region) and one antibonding MO, denoted σ (higher in energy, one node perpendicular to the bond axis, density pushed away from the internuclear region). Filling the bonding MO with the shared electron pair — one from each atom, spins paired — is what "forming a σ bond" means* at the orbital level.

The overlap between two orbitals $\phi_A$ and $\phi_B$ centered on atoms A and B is quantified by the overlap integral,

where: $\phi_A,\phi_B$ are the (real, normalized) atomic or hybrid orbitals, and the integral runs over all space. $S$ ranges from $-1$ to $+1$; $S=0$ means no bonding interaction is possible by symmetry (e.g., an s orbital with a perpendicular p orbital), and $|S|$ close to its maximum for the orbital sizes involved means strong, effective overlap. For two orbitals combined by the variational LCAO method, the bonding and antibonding molecular orbitals and their energies are

where: $\alpha = \int \phi_A^* \hat H \phi_A\, d^3r$ is the atomic (Coulomb) energy of an electron in $\phi_A$ alone, $\beta = \int \phi_A^* \hat H \phi_B \,d^3r$ is the resonance (exchange) integral (negative in sign for a bonding interaction), $S$ is the overlap integral above, and $N_\pm$ are normalization constants. Reading this: the $+$ combination, $\psi_+$, is stabilized (bonding, σ) and the $-$ combination, $\psi_-$, is destabilized (antibonding, σ) relative to the isolated atomic orbitals, and the splitting* between them grows directly with $|\beta|$, which itself grows with $|S|$ — bigger head-on overlap, bigger bonding stabilization. The cylindrical-symmetry statement that governs rotation is captured by writing the σ-orbital's electron density in cylindrical coordinates $(\rho,\theta,z)$ about the bond axis:

where: $\rho$ is the radial distance from the bond axis, $z$ is position along the axis, and $\theta$ is the azimuthal angle around the axis. Reading this: because the density has no $\theta$-dependence, the energy of the σ bond (which depends on the electron density) cannot change as one fragment is rotated by any angle $\theta \to \theta + \Delta\theta$ about that axis — this is the formal statement of "free rotation." Finally, the rotational-barrier-relevant quantity is the torsional potential energy $V(\phi)$ as a function of dihedral (twist) angle $\phi$ about the σ bond, decomposed for a threefold-symmetric single bond (like ethane's C–C) as

where: $\phi$ is the C–C–C–H (or H–C–C–H) dihedral angle, and $V_3$ is the full barrier height (energy difference between eclipsed and staggered conformations). Reading this: this entire $V(\phi)$ comes from secondary effects (steric repulsion, hyperconjugative donor–acceptor interactions between neighboring bonds) layered on top of a σ framework whose own overlap integral is exactly $\phi$-independent — the barrier exists despite the σ bond, not because of it.

Example 1 (foundational): Building σ(H–F) and predicting its polarity from overlap symmetry. Consider HF. Fluorine's valence configuration is $2s^2\,2p_x^2\,2p_y^2\,2p_z^1$ (taking $z$ as the H–F axis), so the singly occupied $2p_z$ orbital is available for bonding, while $2p_x$ and $2p_y$ are filled lone pairs pointing perpendicular to the bond axis. Hydrogen contributes its $1s$ orbital. Both $1s_H$ and $2p_{z,F}$ have their major lobes on the internuclear axis and no angular ($\theta$) dependence about that axis — a valid σ pair. Applying the same secular-equation logic as Step 2–4 of the first mechanism (now with $\alpha_H \ne \alpha_F$, since the atoms differ), the bonding combination is

because fluorine's orbital sits at lower energy (fluorine is more electronegative, $\alpha_F < \alpha_H$), and the bonding MO always leans toward the lower-energy atomic orbital. This asymmetric mixing coefficient is precisely the orbital-level origin of the bond dipole: the electron pair in $\psi_\sigma$ spends more time near fluorine, giving H–F a dipole moment of $1.82\ \text D$ pointing H$^{\delta+}$→F$^{\delta-}$. Despite this unequal sharing, the bond is still classified as σ because the overlap geometry — head-on, axial, θ-independent — is unchanged; polarity is a property of the coefficients, not of the overlap symmetry. Limiting check: if the two atoms were identical ($\alpha_H = \alpha_F$, i.e., the H₂ or F₂ case), the mixing coefficients become equal, $|c_H|=|c_F|$, exactly recovering the nonpolar bonding MO of the previous derivation — confirming the formula degrades correctly to the homonuclear limit.

Example 2 (applied): Estimating the eclipsed-conformer energy penalty in butane's central C–C bond from a mix of torsional terms. Butane, CH₃–CH₂–CH₂–CH₃, rotating about the central C2–C3 bond, is governed by a torsional potential with both threefold (ethane-like, from the two methyls/methylenes) and steric (methyl–methyl clash) contributions. Using the standard Ryckaert–Bellemans-type expansion is overkill here; instead reason directly from the σ-bond-is-rotationally-free principle plus superposed sterics. At $\phi = 0°$ (fully eclipsed, methyl groups eclipsing each other) the σ(C2–C3) bond itself contributes zero extra energy (per the mechanism above), but the two terminal methyl groups now sit only about $2.5$–$3.0$ Å apart, well inside their combined van der Waals radii sum ($\sim 4.0$ Å for two methyls), producing a steric clash worth roughly $4$–$5$ kcal/mol above the anti ($\phi = 180°$) minimum — this is the experimentally known totally eclipsed butane conformer penalty (measured barrier $\approx 4.5$–$5.0$ kcal/mol above anti, i.e., roughly $1.6$–$2.1$ kcal/mol above the gauche minimum, itself $\sim 0.9$ kcal/mol above anti). Concretely, take the anti conformer as the zero of energy: gauche sits at $+0.9$ kcal/mol (from residual pentane-type steric strain even in a staggered arrangement) and totally eclipsed at $+4.5$–$5.0$ kcal/mol, so the increment purely from moving methyl past methyl at $\phi=0°$ relative to at $\phi=180°$ is

This value is far larger than ethane's $V_3 \approx 2.9$ kcal/mol precisely because the substituents are bulkier methyls rather than hydrogens — confirming that the barrier magnitude tracks steric bulk and hyperconjugative alignment, never the σ-bond overlap integral itself, which by the second mechanism is provably flat in $\phi$ for both molecules. Limiting check: replacing both terminal methyls with hydrogens collapses this system back to ethane, and the formula/reasoning correctly predicts the barrier drops to the much smaller, purely steric+hyperconjugative $V_3 \approx 2.9$ kcal/mol — consistent with the actual measured ethane barrier.

Example 3 (exam/competition level): Ranking three C–C σ bonds by overlap-derived bond strength using hybridization and orbital size. Rank the C–C σ bonds in (a) ethane (Csp³–Csp³), (b) propyne, H–C≡C–CH₃ (the Csp–Csp³ bond), and (c) but-1-en-3-yne-type fragment where a bond is Csp²–Csp², by bond dissociation energy, and justify using overlap/orbital-size reasoning alone (no lookup table logic). Step 1 — s-character controls orbital size and directionality. An sp³ hybrid is $25\%$ s character; sp² is $33\%$ s; sp is $50\%$ s. Because s orbitals are smaller and more contracted (lower principal quantum number contribution to spatial extent, more tightly bound, closer to the nucleus) than p orbitals of the same shell, a hybrid with more s-character is more compact and more sharply directional (a "fatter, shorter" lobe pointed more precisely along the bond axis). Step 2 — relate compactness/directionality to overlap integral $S$. The overlap integral $S = \int \phi_A \phi_B\,d^3r$ is maximized when both lobes are well-directed (little "leakage" of amplitude off-axis) and appropriately size-matched at the actual internuclear separation. Higher s-character hybrids, being smaller and pulled in closer to their own nucleus, allow the two bonded atoms to approach to a shorter equilibrium distance $r_{eq}$ while still achieving strong on-axis overlap — because the hybrid's electron density is concentrated closer to the nucleus and more sharply peaked along the bond direction, less overlap is "wasted" off-axis at short range. Step 3 — apply the ranking. Bond (b), Csp–Csp³ (propyne's C–C single bond connecting the sp carbon of the triple bond to the sp³ methyl carbon) and especially any Csp–Csp bond would have the highest effective s-character-weighted overlap on the sp side; bond (c), Csp²–Csp² has intermediate s-character (33%); bond (a), Csp³–Csp³ (ethane) has the least (25%). Since bond strength tracks with s-character-driven overlap quality and the resulting bond contraction, the strength ranking is

This is borne out numerically: ethane's C–C bond length is $1.54$ Å with BDE $\approx 90$ kcal/mol; propyne's C(sp)–C(sp³) bond is $1.46$ Å with BDE $\approx 97$–$102$ kcal/mol; a typical Csp²–Csp² single bond (as in the C2–C3 bond of buta-1,3-diene or the analogous enyne fragment) sits around $1.48$–$1.50$ Å with BDE intermediate between the two, consistent with the s-character ordering. Limiting check: if all three bonds involved the same hybridization on both carbons (the actual sp³–sp³ case, ethane), the ranking argument correctly reduces to "no difference expected from s-character," matching the fact that ethane's own bond is the internal reference point (weakest of the three) rather than an outlier — the trend is self-consistent at the sp³=sp³ limit.

The mechanistic language of σ and π bonds — orbitals overlapping head-on versus sideways — emerged from the fusion of Lewis's electron-pair bond with the new quantum mechanics of the late 1920s, chiefly through the competing but complementary valence-bond and molecular-orbital programs.