Torsional Strain and the Origin of the Ethane Barrier

Spin a model of ethane about its central C–C bond and you feel nothing. Real molecules give no mechanical resistance. Yet nature imposes a genuine energetic toll of roughly 12 kJ/mol (2.9 kcal/mol). This toll occurs every time the two methyl groups pass from a staggered arrangement through an eclipsed one and back. This is the smallest, purest example of a conformational energy barrier in all of organic chemistry. For decades chemists argued about where the 12 kJ/mol actually comes from. Is it simple steric bumping between hydrogens six bond-lengths apart? Classical electrostatic repulsion between C–H bond dipoles? Or something more subtle written into the wavefunction itself? The modern, quantitatively defensible answer is orbital, not steric. The staggered conformation is stabilized relative to the eclipsed one by hyperconjugative donation. This donation goes from each filled C–H σ bond into the antibonding σ orbital of a nearby C–H bond on the adjacent carbon. This donation is geometrically optimal — maximal orbital overlap — only when bonds are anti-periplanar-ish, i.e. staggered by 60°. It vanishes by symmetry when bonds eclipse each other. The residual destabilization when eclipsed is often loosely called "torsional strain." This article dissects exactly what that phrase mechanistically means at the orbital level, rather than leaving it as an unexplained buzzword. This matters far beyond ethane. The same σ→σ hyperconjugative logic governs several key preferences. It explains the preference for anti over gauche arrangements in Conformational Analysis of Butane. It explains the anomeric effect in sugars. It explains the staggered geometry preferred at every sp³–sp³ single bond in a protein backbone or a polymer chain. And it ultimately explains the entire conformational machinery behind the chair preference in Cyclohexane Chair Conformation. Torsional strain is the microscopic seed from which the whole discipline of conformational analysis grows. This article builds the orbital picture from first principles. It explains what a σ bond and σ* orbital actually look like in three dimensions. It explains why their overlap depends on dihedral angle exactly as $\cos^2\theta$ . It shows how this produces the sinusoidal rotational potential you see experimentally. It shows how competing explanations (steric repulsion, bond-bending, electrostatics) fail quantitatively where the hyperconjugative model succeeds. Two full mechanistic derivations follow. One derives from orbital overlap and perturbation theory. The other derives from the resulting potential-energy function. Together they pin every number to a defensible physical origin.

Key takeaways

Rotating ethane's C–C bond costs energy sinusoidally; the barrier is about 12 kJ/mol (2.9 kcal/mol), maximal at the eclipsed (0°) geometry and zero at staggered (60°) geometries.
The dominant physical origin is σ(C–H)→σ(C–H) hyperconjugation: each filled C–H bonding orbital donates electron density into an antiperiplanar-ish antibonding C–H orbital on the neighboring carbon.
This donation stabilizes the staggered conformer specifically, because orbital overlap between σ and σ* across the C–C bond scales as $\cos^2\theta$ , peaking near 60°/180° and vanishing at 0°/eclipsed alignments of like-symmetry lobes.
Classical steric repulsion between eclipsing H···H atoms (2.55 Å apart, well beyond the sum of van der Waals radii ~2.4 Å) is real but far too small in ethane to explain the full barrier; it becomes dominant only in bulkier systems like butane.
Natural Bond Orbital (NBO) and photoelectron spectroscopy analyses converge: roughly 75–90% of the ethane barrier is hyperconjugative in character, with the remainder from exchange/steric and electrostatic terms.
Key idea: the ethane barrier exists because staggered conformers gain extra σ→σ* hyperconjugative stabilization that eclipsed conformers cannot, not primarily because eclipsed hydrogens physically collide.

Where this lives today

The σ→σ hyperconjugative logic developed here for the humble ethane barrier is the conceptual backbone of much of modern stereoelectronic chemistry. In carbohydrate chemistry, the anomeric effect — the preference of electronegative substituents at the anomeric carbon of a pyranose to sit axial rather than equatorial — is mechanistically the same $n_{\text{O}} \to \sigma^*_{\text{C–X}}$ donation, just with a lone pair replacing a C–H bond as the donor, and it dictates the reactivity and glycosidic bond stability central to carbohydrate and nucleic acid chemistry (the sugar-phosphate backbone conformation in DNA is governed by related hyperconjugative preferences). In peptide and protein backbones, the preference for staggered arrangements about Cα–Cβ and related single bonds, modulated by exactly this kind of hyperconjugative and steric competition, feeds directly into rotamer libraries used in protein structure prediction and molecular dynamics force fields. In polymer science, the persistence length and helical preferences of vinyl polymers and polysaccharides trace back to the same torsional energetics scaled up over many backbone bonds. Pharmaceutical chemists routinely use torsional-strain reasoning to explain why certain drug conformers are populated in a binding pocket, informing rational design of conformationally locked (bicyclic or fluorinated) analogues that pre-organize a molecule into its bioactive rotamer, exploiting exactly the fluorine-enhanced hyperconjugation seen in Example 2. Computational chemistry software (Gaussian, ORCA, and dedicated NBO packages) routinely reports σ→σ stabilization energies as a standard diagnostic when analyzing any conformational preference, making this once-controversial orbital picture a load-bearing, everyday tool of structural and computational organic chemistry.

Ask the tutor: How does adding a bulky methyl group at one carbon (as in butane) change the balance between hyperconjugative and steric contributions to the rotational barrier?

The physical picture

Picture ethane as two methyl "propellers" joined by a C–C σ bond, each carbon roughly sp³ hybridized. As you twist one methyl group relative to the other, the dihedral angle $\theta$ between a chosen front C–H bond and a chosen back C–H bond sweeps from 0° (eclipsed) to 60° (staggered) and back. The total energy of the molecule, measured as a function of $\theta$ , executes a smooth threefold-periodic oscillation: three equivalent minima (staggered, $\theta = 60°, 180°, 300°$ ) and three equivalent maxima (eclipsed, $\theta = 0°, 120°, 240°$ ). To understand why the staggered geometry is lower in energy, look at a single C–H bond on the front carbon and ask what it "sees" on the back carbon. A C–H bonding orbital, $\sigma_{C-H}$ , is not chemically inert — it is a filled orbital with electron density that can act as a very weak donor, exactly as a lone pair can, just less effectively because bonding electrons are more tightly held. The complementary antibonding orbital, $\sigma^*_{C-H}$ , on a neighboring carbon is empty and can act as an acceptor. When these two orbitals achieve good spatial overlap, second-order perturbation theory tells us the filled orbital is stabilized and the empty orbital is populated infinitesimally — a two-electron, two-orbital interaction exactly analogous to a weak covalent bond forming and breaking as $\theta$ changes. The overlap of $\sigma_{C-H}$ (front) with $\sigma^*_{C-H}$ (back) depends on the dihedral angle between them. Both orbitals have a node perpendicular to their own bond axis, and their extended lobes point roughly along the tetrahedral directions from each carbon. When the dihedral angle is 180° (perfectly anti) or 60° (gauche-like, as in staggered ethane), the back lobe of $\sigma^*_{C-H}$ points substantially toward the front lobe of $\sigma_{C-H}$ , giving good overlap. When the dihedral angle is 0° (eclipsed), the geometry is poor for this particular donor–acceptor pair, though a different pairing (the eclipsing bonds' own back-lobes) becomes momentarily relevant — but symmetry analysis shows the net constructive stabilization is maximized in the staggered geometry summed over all three donor–acceptor pairs per C–C bond.

Figure

Newman projections: staggered vs eclipsed ethane, with the key σ→σ* interaction highlighted

Figure

Orbital overlap of σ(C–H) donor with σ*(C–H) acceptor as a function of dihedral angle

The key qualitative fact those figures encode: overlap between a donor σ and acceptor σ bond is not uniform with rotation — it is a smooth trigonometric function of the dihedral angle, sharply peaked when the bonds are anti-periplanar and rapidly falling off as the dihedral shrinks toward the eclipsed, syn-periplanar geometry relevant to a different* neighboring pair. Summing the three symmetric donor–acceptor interactions around the C–C axis at each dihedral angle reproduces the full sinusoidal rotational potential of ethane.

The core ideas

The total hyperconjugative stabilization energy from a single σ→σ* donor–acceptor pair is given, from second-order Rayleigh–Schrödinger perturbation theory, by

\Delta E^{(2)}_{\sigma \to \sigma^*} = -\frac{\left| \langle \sigma | \hat{F} | \sigma^* \rangle \right|^2}{\varepsilon_{\sigma^*} - \varepsilon_{\sigma}}

where:

$\langle \sigma | \hat{F} | \sigma^* \rangle$ is the off-diagonal Fock (or Kohn–Sham) matrix element coupling the filled $\sigma_{C-H}$ orbital to the empty $\sigma^*_{C-H}$ orbital — proportional to their spatial overlap and to the resonance integral between the constituent atomic orbitals;
$\varepsilon_{\sigma^*} - \varepsilon_{\sigma}$ is the energy gap between the antibonding and bonding orbital, always positive, so $\Delta E^{(2)}$ is always stabilizing (negative);
the numerator scales, to a very good approximation, with $S^2(\theta)$ , the square of the overlap integral between the two orbitals as a function of the dihedral angle $\theta$ between them.

Because $\langle \sigma|\hat F|\sigma^*\rangle \propto \cos\theta$ for two $\sigma$ -type orbitals related by a torsional rotation (this follows from the group-theoretic symmetry of $\sigma$ orbitals under rotation about the internuclear axis — each behaves, in projection, like a vector aligned along its bond, and the overlap of two such vectors separated by angle $\theta$ is a cosine), the stabilization energy from a single interacting pair goes as

\Delta E^{(2)}(\theta) \propto -\cos^2\theta

Summing over the three symmetric donor–acceptor pairs related by the threefold rotation of the methyl groups, and adding the analogous residual steric/exchange (four-electron, two-orbital destabilizing) term, gives the total torsional potential:

V(\theta) = \frac{V_0}{2}\left[1 - \cos(3\theta)\right]

where:

$\theta$ is the C–H/C–H dihedral angle measured from a staggered reference (or equivalently, the deviation of the true dihedral from 60°, depending on convention — here we set $\theta = 0$ at eclipsed);
$V_0$ is the full barrier height, the energy difference between the eclipsed maximum and staggered minimum, experimentally $V_0 \approx 12.0$ kJ/mol (2.9 kcal/mol) for ethane;
the factor $3$ in $\cos(3\theta)$ encodes the threefold symmetry of the methyl rotor: the potential must repeat every $120°$ , since the three hydrogens on each carbon are indistinguishable.

Reading this equation: at $\theta = 0°$ (eclipsed), $\cos(0) = 1$ , so $V = V_0$ , the maximum; at $\theta = 60°$ (staggered), $\cos(180°) = -1$ , so $V = 0$ , the minimum. The function is a pure cosine wave in $3\theta$ , exactly matching the observed threefold barrier measured by microwave and far-infrared torsional spectroscopy of ethane.

Derivation AFrom orbital overlap to the

\cos(3\theta)

torsional potential

Goal. Show, mechanistically, how pairwise σ→σ* hyperconjugative interactions between the three C–H bonds on one carbon and the three C–H bonds on the other combine into the observed threefold-symmetric rotational potential of ethane, and why the staggered geometry is the unique minimum.

Step 1 — Set up the local orbital basis. Take the front carbon's three C–H bonds, each described by a filled bonding orbital $\sigma_i$ ( $i = 1,2,3$ , spaced 120° apart around the C–C axis) and the corresponding empty antibonding orbitals $\sigma_i^*$ . Do the same for the back carbon's three C–H bonds, $\sigma_j$ and $\sigma_j^*$ . Each $\sigma_i$ points outward from the front carbon at the tetrahedral angle, tilted away from the C–C axis; the same is true of each $\sigma_j$ on the back carbon, but the whole back set is rotated by the dihedral angle $\theta$ relative to the front set.

Step 2 — Identify which pairs actually overlap. A front $\sigma_i$ orbital can donate into a back $\sigma_j^*$ orbital, and vice versa (donation is bidirectional and symmetric in ethane by the molecule's overall symmetry). The strength of each individual donor–acceptor interaction depends on the projection of the back C–H bond direction onto the extended axis of the front C–H bond, viewed along the C–C bond — precisely the dihedral angle between that specific pair of bonds. For any chosen front bond $\sigma_i$ , the three back bonds sit at dihedral angles $\theta$ , $\theta + 120°$ , and $\theta + 240°$ relative to it (mod 360°).

Step 3 — Write the overlap-squared as a cosine. Model each $\sigma$ orbital, for the purpose of this angular dependence, as a vector along its bond axis (a good approximation because the interaction is dominated by the p-character lobe extending from each carbon). The relevant matrix element between two such "bond vectors" separated by angle $\phi$ is

\langle \sigma_i | \hat{F} | \sigma_j^* \rangle \propto \cos\phi

so the second-order stabilization from that one pair is

\Delta E^{(2)}_{ij} \propto -\cos^2\phi

This has two maxima per full rotation (at $\phi = 0°$ and $\phi=180°$ , since $\cos^2$ is symmetric) — consistent with hyperconjugation being strongest for both syn- and anti-periplanar alignment, weakest at 90°.

Step 4 — Sum over all three donor–acceptor pairs for a given front bond. For front bond $\sigma_i$ , the total stabilization into the three back $\sigma_j^*$ orbitals is

E_i(\theta) \propto -\left[\cos^2\theta + \cos^2(\theta+120°) + \cos^2(\theta+240°)\right]

Using the trigonometric identity $\cos^2 x = \tfrac{1}{2}(1+\cos 2x)$ , each term becomes $\tfrac12 + \tfrac12\cos(2\theta + 2k\cdot120°)$ for $k = 0,1,2$ . Summing the three cosine terms with arguments $2\theta$ , $2\theta+240°$ , $2\theta+480°$ , note these three phases are spaced 240° apart, i.e. equivalent to three vectors at 120° in the doubled angle — they sum to zero for any $\theta$ (three equally spaced unit vectors sum to zero). Hence

E_i(\theta) \propto -\left[\tfrac32 + 0\right] = -\tfrac32

a constant, independent of $\theta$ ! At the level of a single front bond summed over three back acceptors, the θ-dependence cancels — which correctly signals that the net angular dependence must come from a higher-order (product) term, not from the naive sum of independent pairwise $\cos^2$ terms alone.

Step 5 — Restore the angular dependence via the correct combinatorial counting. The resolution is that the three front bonds and three back bonds are not independent: summing over both indices $i$ and $j$ correctly (nine pairwise interactions total, not just three), and retaining the interference between hyperconjugative stabilization and the competing four-electron Pauli/exchange repulsion (which also depends on $\cos^2\theta$ but with opposite overall sign convention and different phase reference, being large for eclipsed alignment of filled orbitals rather than filled–empty pairs), the leading θ-dependent term surviving after full summation is proportional to $\cos(3\theta)$ — a result confirmed both by explicit NBO deletion calculations and by symmetry: any function periodic in 120° that is expandable in a Fourier cosine series must have its lowest nonzero harmonic at $n=3$ (since $n=1,2$ terms are forced to vanish by the local $C_3$ symmetry of the methyl rotor, exactly as the naive sum in Step 4 demonstrated for $n=2$ ). The physically transparent statement is:

V(\theta) = \frac{V_0}{2}\bigl[1-\cos(3\theta)\bigr]

This is the unique lowest-order threefold-periodic function consistent with $V(0°) = V_0$ (eclipsed maximum) and $V(60°) = 0$ (staggered minimum), and it is what all higher-order hyperconjugative/exchange bookkeeping must ultimately reduce to by symmetry.

Step 6 — Interpret physically. The staggered geometry is a genuine energetic minimum not because any single σ→σ pair is optimized there (Step 4 showed the naive single-pair sum is flat), but because the full multi-orbital interaction — hyperconjugative stabilization competing against four-electron destabilization between filled σ orbitals — is resolved by symmetry into a net threefold wave whose minimum sits exactly at the staggered dihedral. Explicit orbital deletion procedures (removing all σ→σ interactions computationally and re-optimizing) show the staggered preference collapses almost entirely, confirming hyperconjugation, not the trivial pairwise overlap sum, is the operative stabilizing mechanism.

Go deeper · Derivation B · 2 minFrom the torsional potential to the measured barrier height and force constant

Derivation BFrom the torsional potential to the measured barrier height and force constant

Goal. Use the $\cos(3\theta)$ potential to predict (a) the barrier height in the standard units chemists quote, and (b) the torsional force constant governing small oscillations about the staggered minimum, and connect both to spectroscopic and thermodynamic observables.

Step 1 — Confirm the extrema. Differentiate $V(\theta) = \frac{V_0}{2}[1-\cos(3\theta)]$ :

\frac{dV}{d\theta} = \frac{3V_0}{2}\sin(3\theta)

Setting this to zero gives $\sin(3\theta) = 0$ , i.e. $3\theta = 0°, 180°, 360°, \ldots$ , so $\theta = 0°, 60°, 120°, \ldots$ — exactly the eclipsed and staggered geometries, confirmed as the only stationary points, matching chemical intuition that no "intermediate" special conformer exists for ethane.

Step 2 — Classify the extrema via the second derivative.

\frac{d^2V}{d\theta^2} = \frac{9V_0}{2}\cos(3\theta)

At $\theta = 0°$ (eclipsed): $\cos(0)=1$ , so $d^2V/d\theta^2 = +\tfrac{9V_0}{2} > 0$ — wait, a positive second derivative at a claimed maximum would be wrong; check via $V(\theta)$ directly: $V(0°) = \frac{V_0}{2}(1-1) = 0$ . This shows the convention chosen here places $\theta=0$ at the minimum (staggered), not eclipsed — a labeling issue easily fixed by shifting the reference. Redefine $\theta$ so that $\theta=0$ is eclipsed: then $V(\theta) = \frac{V_0}{2}[1+\cos(3\theta)]$ , giving $V(0°)=V_0$ (eclipsed maximum, correct) and $V(60°) = \frac{V_0}{2}[1+\cos180°] = 0$ (staggered minimum, correct). Redo the derivatives with this corrected convention:

\frac{dV}{d\theta} = -\frac{3V_0}{2}\sin(3\theta), \qquad \frac{d^2V}{d\theta^2} = -\frac{9V_0}{2}\cos(3\theta)

At $\theta=0°$ (eclipsed): $d^2V/d\theta^2 = -\tfrac{9V_0}{2} < 0$ , confirming a maximum. At $\theta=60°$ (staggered): $\cos(180°)=-1$ , so $d^2V/d\theta^2 = +\tfrac{9V_0}{2} > 0$ , confirming a minimum. Both classifications now agree with the known chemistry.

Step 3 — Extract the torsional force constant. Near the staggered minimum, expand $\theta = 60° + \delta$ for small displacement $\delta$ (in radians). The potential is harmonic to leading order,

V(\delta) \approx \frac{1}{2}\left(\frac{d^2V}{d\theta^2}\Big|_{60°}\right)\delta^2 = \frac{1}{2}\left(\frac{9V_0}{2}\right)\delta^2

so the effective torsional force constant is

k_{\text{tors}} = \frac{9V_0}{2}

in energy-per-radian² units. Using $V_0 = 12.0$ kJ/mol:

k_{\text{tors}} = \frac{9 \times 12.0}{2}\ \text{kJ/mol/rad}^2 = 54.0\ \text{kJ/mol/rad}^2

This force constant, combined with the reduced moment of inertia of the two methyl rotors about the C–C axis, predicts a torsional vibrational frequency that can be compared directly to the far-infrared/Raman-observed ethane torsion at roughly 275 cm⁻¹ — a nontrivial independent check that the assumed $\cos(3\theta)$ form (and hence the threefold hyperconjugative/steric picture) is quantitatively, not just qualitatively, correct.

Step 4 — Connect to the rotational barrier as measured by NMR/thermodynamics. For a barrier this small, room-temperature thermal energy $RT \approx 2.5$ kJ/mol at 298 K is comparable to $V_0 \approx 12$ kJ/mol, so ethane rotates essentially freely — no NMR line-broadening or conformer populations are observable (unlike hindered rotations with barriers above ~40–50 kJ/mol). The Boltzmann population ratio of a molecule sitting instantaneously at the eclipsed maximum versus the staggered minimum is

\frac{P_{\text{eclipsed}}}{P_{\text{staggered}}} = \exp\left(-\frac{V_0}{RT}\right) = \exp\left(-\frac{12.0}{2.478}\right) = \exp(-4.84) \approx 7.9\times10^{-3}

confirming that at any instant the molecule spends the overwhelming majority of its time near the staggered geometries, consistent with a shallow but real barrier rather than a flat rotor.

Worked examples

Example 1 (Foundational): Verifying the barrier height from bond rotation data. Electron diffraction and microwave spectroscopy give the ethane barrier as $V_0 = 12.0$ kJ/mol. Confirm this is consistent with the measured torsional splitting and use it to predict the energy difference between the eclipsed and a conformation twisted only 20° away from staggered. Using the corrected-convention potential $V(\theta) = \frac{V_0}{2}[1+\cos(3\theta)]$ with $\theta$ measured from eclipsed, a conformation "20° away from staggered" sits at $\theta = 60° - 20° = 40°$ (or equivalently $80°$ , by symmetry — take $40°$ ):

V(40°) = \frac{12.0}{2}\left[1+\cos(120°)\right] = 6.0\left[1 + (-0.5)\right] = 6.0 \times 0.5 = 3.00\ \text{kJ/mol}

So a molecule twisted 20° off the staggered minimum already carries 3.00 kJ/mol of torsional strain — a quarter of the full barrier — illustrating how steeply the potential rises near the minimum (consistent with the harmonic force constant derived above: the harmonic estimate gives $V \approx \tfrac12 k_{\rm tors}\delta^2$ with $\delta = 20° = 0.349$ rad, so $V \approx \tfrac12(54.0)(0.349)^2 = 3.29$ kJ/mol, close to the exact cosine value, confirming the small-displacement harmonic approximation is reasonable but slightly overestimates because $20°$ is not infinitesimally small).

Limiting case check: as $\theta \to 60°$ exactly, $V \to 0$ (bottom of the well, zero strain) and as $\theta \to 0°$ , $V \to V_0 = 12.0$ kJ/mol exactly (full barrier), both recovering the defining endpoints correctly.

Example 2 (Applied): Isotopic and substituent perturbation — why is fluoroethane's barrier different from ethane's? Replacing one hydrogen on ethane with fluorine (giving fluoroethane, CH₃CH₂F) changes the barrier to internal rotation about the C–C bond to roughly 13.9 kJ/mol (3.3 kcal/mol), slightly higher than parent ethane's 12.0 kJ/mol, despite fluorine being much bulkier than hydrogen and naively expected to increase steric clash in the eclipsed form only modestly more than it changes anything else. Explain mechanistically why the barrier rises rather than simply reflecting added sterics. The dominant additional stabilization in the staggered conformer of fluoroethane comes from an enhanced hyperconjugative interaction: fluorine's strong electronegativity lowers the energy of the antiperiplanar C–H σ orbitals on the CH₃ carbon (an inductive effect that pulls electron density from those C–H bonds, lowering $\varepsilon_{\sigma^*}$ ), which increases* the magnitude of the stabilization term

\Delta E^{(2)} = -\frac{|\langle\sigma|\hat F|\sigma^*\rangle|^2}{\varepsilon_{\sigma^*}-\varepsilon_\sigma}

because the denominator shrinks toward zero while the numerator (overlap) is essentially unchanged by the remote substitution. Additionally, the strongest hyperconjugative donor in fluoroethane is the C–F bond's neighboring $\sigma_{C-H}$ orbitals donating into $\sigma^*_{C-F}$ , and this donor–acceptor pair is optimized in the staggered rotamer that places a methyl C–H anti-periplanar to the C–F bond — the classic setup behind the wider "gauche effect" family of interactions. Because this extra staggered-specific stabilization has no counterpart in the eclipsed geometry, the difference between staggered and eclipsed energies (the barrier) grows, even though no dramatic new steric clash has been introduced. This demonstrates that barrier heights are controlled by the differential hyperconjugative stabilization of the staggered well, not simply by the absolute bulk of substituents at the eclipsing positions — a lesson that carries directly into rationalizing the anomeric effect and gauche preferences in more complex systems.

Example 3 (Exam / competition level): Deconstructing the barrier into hyperconjugative and steric contributions using an isodesmic-style orbital-deletion argument. Natural Bond Orbital analyses of ethane's rotational barrier report that deleting all σ(C–H)→σ(C–H) hyperconjugative interactions in the Fock matrix (setting those off-diagonal elements to zero) and re-evaluating the energy difference between eclipsed and staggered geometries reduces the apparent barrier from 12.0 kJ/mol down to approximately $-3$ to $+2$ kJ/mol (values vary modestly by basis set and precise deletion scheme, but the qualitative finding is robust: the staggered preference nearly vanishes or even reverses). Using this data, estimate the fraction of the total barrier attributable to hyperconjugation versus the residual steric/exchange contribution, and discuss what a reversal* of sign upon deletion would mean mechanistically. Take the reported "deletion" barrier as $V_0^{\text{steric-only}} \approx 0$ kJ/mol (an idealized midpoint of the reported range for a clean order-of-magnitude estimate) and the full barrier as $V_0^{\text{full}} = 12.0$ kJ/mol. The hyperconjugative contribution is then

\Delta V_{\text{hyperconj}} = V_0^{\text{full}} - V_0^{\text{steric-only}} = 12.0 - 0 = 12.0\ \text{kJ/mol}

i.e., essentially the entire barrier is attributable to hyperconjugation in this idealized accounting, with the fractional contribution

f_{\text{hyperconj}} = \frac{\Delta V_{\text{hyperconj}}}{V_0^{\text{full}}} = \frac{12.0}{12.0} = 1.00\ (100\%)

More careful decompositions in the literature (accounting for basis-set superposition and orbital relaxation upon deletion, which are not perfectly separable from the "true" steric term) typically quote figures in the 75–90% range for the hyperconjugative share, with the remainder assigned to a small residual four-electron destabilization (Pauli/exchange repulsion between filled σ orbitals, largest when they eclipse) and an even smaller electrostatic term from C–H bond dipole alignment. If the deletion calculation instead gave a negative value for $V_0^{\text{steric-only}}$ — i.e., the eclipsed form becomes more stable than staggered once hyperconjugation is artificially switched off — that would mean the purely steric/electrostatic forces in ethane actually favor the eclipsed geometry (plausible, since eclipsed hydrogens are farther from other atoms' influence in some crude electrostatic models, or because removing σ→σ stabilization removes a term that was masking a small residual preference the other way), and it would sharpen the conclusion that hyperconjugation is not merely the dominant term but is entirely* responsible for the sign of the observed barrier, overriding an intrinsic (if small) steric preference for eclipsing. This is exactly the surprising, robust result that redirected the mechanistic understanding of the ethane barrier away from naive steric pictures during the late 20th century.

Limiting cases, scope & common traps

Situation	What actually happens	Why
Ethane, H···H eclipsing distance ≈ 2.55 Å	Barrier ≈ 12.0 kJ/mol; too large for pure van der Waals repulsion at that distance (sum of radii ≈ 2.4 Å, minimal overlap)	Hyperconjugative σ→σ* stabilization difference dominates, not steric collision
Butane, eclipsed CH₃···CH₃ (methyl over methyl, syn-periplanar)	Barrier rises sharply to ≈ 19–25 kJ/mol locally	Genuine steric/exchange repulsion from bulky methyl groups now adds substantially atop the base hyperconjugative term — foreshadowing Conformational Analysis of Butane
High temperature limit, $T \to \infty$	Rotamer populations become statistically equal (Boltzmann factors → 1 for all θ)	$\exp(-V(\theta)/RT) \to 1$ as $RT \gg V_0$ ; the molecule samples all θ nearly uniformly, though the instantaneous potential shape is unchanged
Deuterated ethane (CD₃CD₃)	Barrier essentially unchanged (~12 kJ/mol), zero-point corrections shift slightly	Electronic/orbital origin of the barrier is isotope-independent to leading order; only vibrational zero-point energy differs slightly
Removing all σ→σ* interactions computationally	Staggered preference collapses to near zero or reverses	Confirms hyperconjugation is necessary, not merely additive, for the observed barrier sign and magnitude
Anomeric center in a pyranose sugar	Analogous but stronger $n_{O}\to\sigma^*_{C-X}$ hyperconjugation drives axial preference	Lone-pair donor replaces σ(C–H) donor, same overlap-vs-dihedral logic, much larger stabilization

Common trapCommon misconception: "the ethane barrier is just steric hindrance between eclipsing hydrogens"

Textbook shorthand often calls the eclipsed destabilization "steric strain," but the H···H distance in eclipsed ethane (~2.55 Å) is not smaller than typical non-bonded contacts elsewhere in organic molecules that carry no comparable strain penalty, and simple van der Waals radii sums (~2.4 Å for two H atoms) suggest only marginal repulsive overlap at that separation. Quantitative orbital-deletion calculations (Example 3) show removing hyperconjugation removes essentially the entire barrier; a purely electrostatic/steric ball-and-spring model of ethane, absent any orbital interaction terms, does not reproduce even the correct sign of the staggered preference reliably. "Torsional strain" is best understood as the energetic penalty for failing to achieve the stabilizing σ→σ* overlap available in the staggered geometry — a lost-stabilization effect, not a repulsive-collision effect, even though loosely calling it "steric" survives colloquially.

The chemist behind it

The modern hyperconjugative reinterpretation of the ethane barrier is closely associated with Frank Weinhold, whose development of Natural Bond Orbital (NBO) theory in the late 1970s and 1980s gave chemists a rigorous, basis-set-independent way to decompose molecular wavefunctions into localized Lewis-like bonding and antibonding orbitals, and to quantify exactly the kind of σ→σ* stabilization energies used throughout this article. Weinhold's collaborations with Kenneth Wiberg and others through the 1990s and 2000s used systematic orbital-deletion procedures — precisely the technique invoked in Example 3 — to argue, against the long-dominant steric-repulsion narrative descended from early 20th-century conformational analysis, that hyperconjugation is the primary driver of the ethane rotational barrier. This was, and to some extent remains, a genuinely contested question in the physical organic chemistry literature, with steric/exchange-repulsion advocates (notably work associated with the Bickelhaupt and Baerends groups using energy-decomposition analysis within density functional theory) offering competing quantitative decompositions that assign a larger share to Pauli repulsion. The pedagogically stable synthesis, reflected in this article, is that both effects are real and both must be included for quantitative accuracy, but the hyperconjugative term is large, systematically direction-dependent (favoring staggered specifically), and essential to explain substituent effects like the fluoroethane example above.

Origins · 1 minFrank Weinhold and how this was discovered

Frank Weinhold

b. 1941 · Developer of Natural Bond Orbital (NBO) theory, University of Wisconsin–Madison, late 20th–early 21st century

History

- 1885 — Baeyer strain theory begins the systematic study of molecular strain, though limited to angle strain in rings, not rotational barriers.

- 1936 — Kemp and Pitzer report the first electron-diffraction/thermodynamic evidence for a genuine rotational barrier in ethane, coining the term "Pitzer strain" for the unexpected energy cost of internal rotation.

- 1950s–1960s — Steric/electrostatic repulsion models dominate textbook explanations of the ethane barrier, largely by default, absent quantitative orbital tools.

- 1980s — Weinhold and coworkers formalize Natural Bond Orbital (NBO) analysis, enabling quantitative decomposition of hyperconjugative contributions to rotational barriers.

- 2001–2003 — Pophristic and Goodman (Nature, 2001) and follow-up work sharpen the debate, using orbital-deletion procedures to argue hyperconjugation, not steric repulsion, is the dominant term; subsequent energy-decomposition analyses (Bickelhaupt, Baerends, and others) push back with alternative partitioning schemes, keeping the precise percentage split an active research question into the 2000s and beyond.

Is the ethane rotational barrier really due to orbital effects, or is this overcomplicating a simple steric picture?

Both effects are physically present, but quantitative decomposition (via NBO orbital-deletion and independent energy-decomposition analyses) shows that removing hyperconjugative σ→σ interactions collapses or reverses the staggered preference, while the pure steric/electrostatic contribution alone is small and even ambiguous in sign. The consensus in modern physical organic chemistry credits hyperconjugation with roughly 75–90% of the barrier's magnitude and, crucially, its correct sign. Calling the eclipsed destabilization "steric strain" is a convenient shorthand embedded in older textbooks, but mechanistically the effect is better described as a lost stabilization* the staggered conformer enjoys and the eclipsed one cannot access, not a repulsive collision between hydrogens that are not, in fact, especially close together.

Why is the ethane barrier so much smaller than the barrier to rotation about a C=C double bond?

The C=C π bond in alkenes involves a genuine, strong covalent π orbital that must be completely broken (both lobes of overlap reduced to zero) to rotate 90°, costing on the order of 250–270 kJ/mol — comparable to a full bond dissociation energy. The ethane barrier, by contrast, involves only a weak, second-order hyperconjugative stabilization (σ→σ*) that is never a full bond in the first place; it is a perturbative correction on the order of a few kJ/mol per interacting pair. The two barriers differ by more than an order of magnitude because one reflects the making/breaking of an actual bond and the other reflects only the gain or loss of a comparatively delicate donor–acceptor stabilization.

Does the ethane barrier depend on temperature or isotopic substitution?

The underlying electronic potential energy surface $V(\theta)$ is essentially independent of temperature and isotopic mass, since both are governed by the Born–Oppenheimer electronic structure at fixed nuclear geometry. However, the observed thermodynamic and spectroscopic manifestations do shift slightly: heavier isotopes (like deuterium) have smaller zero-point vibrational energy in the torsional mode, which can very slightly change the effective barrier felt by the nuclei, and temperature controls the population of different instantaneous torsional angles via the Boltzmann factor, without altering the shape of $V(\theta)$ itself.

How does this connect to the anomeric effect in sugar chemistry?

The anomeric effect is mechanistically the same phenomenon scaled up: instead of a modest C–H σ bond donating into a C–H σ orbital, a much better donor — an oxygen lone pair on the ring oxygen — donates into the σ orbital of the adjacent C–X bond (X = halide, OR, etc.) at the anomeric carbon. Because a lone pair is a far stronger hyperconjugative donor than a bonding pair (it is higher in energy and more available), this donor–acceptor interaction is considerably stronger than in ethane, strongly favoring the geometry (axial at the anomeric center) that achieves optimal antiperiplanar alignment between the lone pair and σ(C–X). The overlap-versus-dihedral-angle logic is identical to the σ(C–H)→σ(C–H) case in ethane; only the identity and strength of the donor orbital changes.

Why doesn't classical steric repulsion alone explain the ethane barrier, given that eclipsing hydrogens are closer together than staggered ones?

The eclipsed H···H distance in ethane is about 2.55 Å, only modestly less than the staggered gauche-like H···H distances and not dramatically inside the sum of van der Waals radii (~2.4 Å for two hydrogens), so any purely repulsive steric term at that separation is small — on the order of a fraction of a kJ/mol in most classical force-field parameterizations, far short of the observed 12 kJ/mol. If steric repulsion alone were responsible, the barrier would be far smaller than observed, and substituent trends (like the fluoroethane increase in Example 2) would be much harder to rationalize without invoking orbital energy shifts.

How does the ethane barrier foreshadow the anti/gauche preference in butane?

In butane, rotation about the central C2–C3 bond experiences the same base σ→σ* hyperconjugative preference for staggered geometries seen in ethane, but now one eclipsing (or gauche) interaction involves bulky methyl groups rather than hydrogens. This adds a genuine, sizeable steric/exchange repulsion term on top of the ethane-like hyperconjugative baseline, which is why the anti conformer of butane is lower in energy than the gauche conformer by roughly 3.8 kJ/mol (despite both being staggered and both benefiting from hyperconjugation) — a purely steric effect layered onto the torsional framework established in ethane.

Can the ethane barrier be measured directly, and how confident are chemists in the 12 kJ/mol value?

Yes: it is measured via far-infrared and Raman torsional spectroscopy (observing the splitting pattern of the hindered internal rotor's energy levels), via microwave spectroscopy of isotopically substituted ethanes, and cross-checked by high-level ab initio and coupled-cluster calculations, all of which converge tightly on $V_0 \approx 12.0$ kJ/mol (2.9 kcal/mol), among the most precisely known small-molecule rotational barriers in existence. This experimental precision is exactly what allows the hyperconjugative-versus-steric debate to be adjudicated quantitatively rather than left as a qualitative argument.