Hybridization: sp2 Carbon and the Pi Bond

Ethylene is a flat molecule. That flatness is not decoration — it is the whole chemical personality of the double bond. Every carbon bearing a C=C or C=O owes its planarity to one structural fact. It also owes its restricted rotation and distinctive reactivity to that same fact. The carbon mixes only three of its four valence orbitals into hybrids. It deliberately leaves one $2p$ orbital unhybridized. That orbital stands at right angles to the plane of the other three. That leftover orbital is not a spectator. It pairs sideways with an identical orbital on a neighboring atom. This builds the pi (π) bond. The π bond is the second, weaker, more exposed bond. It turns a single C–C sigma framework into a reactive functional group. These include alkenes, carbonyls, and imines, which dominate organic chemistry. This matters because almost every named reaction you will meet is a story about π electrons. This includes electrophilic addition, nucleophilic addition to carbonyls, and Diels-Alder cycloadditions. It also includes enolate chemistry and aromaticity itself. Understanding why the π bond behaves the way it does requires understanding the orbital picture beneath it. You need to know why it is weaker than a sigma bond per unit overlap. You need to know why it locks geometry. You need to know why it is the first thing an electrophile attacks. Three $sp^2$ hybrids at $120°$ handle the sigma bonding and define the framework. One pure $p$ orbital does something completely different. The big picture connects two separate stories often taught as disconnected facts. These are molecular shape (VSEPR) and reactivity (curved arrows). The $sp^2$ hybridization model is the bridge. It explains geometry mechanistically. Trigonal planar shape is not an assumption bolted onto the atom. It is the consequence of maximizing $s$ - $p$ mixing efficiency for three sigma bonds. It simultaneously hands you the exposed, high-energy, electron-rich π cloud. Every electrophile in this subject is hunting for that cloud. This article builds the $sp^2$ atom from first principles. It explains why three orbitals mix and not two or four. It shows what the resulting hybrids look like quantitatively. It explains how two adjacent $p$ orbitals combine to make a bonding and an antibonding π molecular orbital. It shows why the geometric consequences all fall out of the same nodal picture. These consequences include planarity, $120°$ angles, restricted rotation, and $E/Z$ isomerism. Everything here generalizes directly to Hybridization: sp Carbon and Triple Bonds. It also underlies Pi Bonds and the Double Bond and Conjugation: Overlapping p Orbitals.

Key takeaways

Carbon mixes one $2s$ and two $2p$ orbitals into three $sp^2$ hybrids; one $2p$ orbital is left pure and unhybridized.
The three $sp^2$ hybrids point to the corners of an equilateral triangle, $120.0°$ apart, giving trigonal planar geometry.
Each $sp^2$ hybrid has $33.3\%$ $s$ -character and $66.7\%$ $p$ -character, intermediate between $sp^3$ ( $25\%$ $s$ ) and $sp$ ( $50\%$ $s$ ).
The leftover $p$ orbital on each of two adjacent $sp^2$ carbons overlaps sideways (lobe-on-lobe) to form the π bond, on top of the head-on σ bond.
π overlap is weaker than σ overlap at bonding distance, so C=C ( $\approx 611\text{ kJ/mol}$ total) is less than twice a C–C single bond ( $\approx 347\text{ kJ/mol}$ ), meaning the π component itself is roughly $264\text{ kJ/mol}$ .
Rotation about the double bond destroys π overlap entirely at $90°$ , which is why alkenes are configurationally locked and show $E/Z$ isomerism.
Key idea: $sp^2$ hybridization builds a flat sigma-bonded triangle while deliberately sacrificing one $p$ orbital to sideways overlap — and that sideways overlap, the π bond, is the reactive heart of every double bond in organic chemistry.

Where this lives today

The $sp^2$ /π picture is not a historical curiosity — it is the operating system underneath modern organic and materials chemistry. Every polymer made from vinyl monomers (polyethylene from ethylene, polystyrene from styrene, PVC from vinyl chloride) begins with a radical, cationic, or metal-catalyzed attack on an alkene's π bond, exploiting exactly the lobe-on-lobe electron density mapped out here. Drug design leans on the same logic: the planarity enforced by $sp^2$ centers in amide bonds (the peptide bond is a resonance-stabilized, effectively $sp^2$ -flattened C–N linkage) dictates the rigid backbone geometry that protein secondary structure depends on, and medicinal chemists routinely exploit restricted rotation about $sp^2$ – $sp^2$ linkages (atropisomerism in biaryl drugs) as a design handle for selectivity. In materials science, conjugated polymers and organic semiconductors (used in OLED displays and organic photovoltaics) are built entirely from chains of $sp^2$ carbons whose parallel, unhybridized $p_z$ orbitals overlap continuously down the backbone, delocalizing π electrons over many atoms and creating the low-bandgap, conductive behavior exploited in modern flexible electronics — a direct macroscale descendant of the two-orbital π/π* construction derived above, scaled up to dozens of overlapping $p$ orbitals. Graphene and carbon nanotubes are the ultimate extension: every carbon is $sp^2$ , arranged in an infinite, perfectly flat hexagonal sigma framework, with every leftover $p_z$ orbital merging into one continuous delocalized π cloud responsible for graphene's extraordinary electrical conductivity. In catalysis, transition-metal alkene complexes (the basis of Wacker oxidation, olefin metathesis, and hydroformylation) function by donating π electron density from the alkene HOMO into empty metal $d$ orbitals — a direct application of treating the π MO as a nucleophile, exactly as invoked in the epoxidation worked example above.

Ask the tutor: How does this same $sp^2$ /π framework extend to explain why benzene's six p orbitals give one continuous, unusually stable π system instead of three isolated double bonds?

The physical picture

Start from an isolated, promoted carbon atom with four singly occupied valence orbitals: $2s$ , $2p_x$ , $2p_y$ , $2p_z$ . In Hybridization: The sp3 Carbon all four of these mix evenly to make four equivalent $sp^3$ hybrids pointing to the corners of a tetrahedron, because a carbon with four sigma bonds needs four equivalent, maximally separated lobes. An $sp^2$ carbon is different at the starting gate: it needs to build only three sigma bonds (or a sigma bond plus a lone pair, in nitrogen and oxygen analogues), so nature only mixes three of the four orbitals. Take $2s$ , $2p_x$ , and $2p_y$ ; leave $2p_z$ completely alone. Mixing three orbitals produces three new hybrid orbitals — orbital mixing conserves orbital count, never creates or destroys orbitals. Because each of the three hybrids must contain an equal share of $s$ -character (by symmetry, none of the three sigma bonds in a symmetric trigonal system is special), each is built from $\frac{1}{3}$ part $s$ and $\frac{2}{3}$ parts $p$ -orbitals in combination. This is exactly the $sp^2$ label: one $s$ , two $p$ 's, distributed over three hybrids. The three resulting lobes lie in the $xy$ -plane, separated by exactly $120°$ , because that is the unique arrangement of three unit vectors in a plane that maximizes mutual angular separation and therefore minimizes electron-electron repulsion between the three bonding pairs (a VSEPR argument now derived from orbital algebra rather than assumed). The fourth orbital, pure $2p_z$ , is untouched by this mixing. It has the classic dumbbell shape with a nodal plane exactly coincident with the $sp^2$ triangle's plane — a lobe pointing up out of the page, an equal-and-opposite lobe pointing down. It is orthogonal (zero net overlap when integrated over the sigma framework) to all three $sp^2$ hybrids by symmetry. This is the orbital that becomes the π system. When two $sp^2$ carbons approach each other to form ethylene, one $sp^2$ hybrid from each carbon points directly at the other, head-on, and overlaps to form a cylindrically symmetric sigma bond — the strong, rotationally-invariant bond of the C=C double bond. The two remaining $sp^2$ hybrids on each carbon point outward at $120°$ from that C–C axis and bond to hydrogens, also via head-on σ overlap. That accounts for the sigma framework: five sigma bonds total in ethylene (one C–C, four C–H), all coplanar because all six $sp^2$ hybrids on the two carbons lie in one shared plane once the molecule adopts its equilibrium geometry. Meanwhile the two unhybridized $2p_z$ orbitals, one per carbon, both perpendicular to that same plane, are parallel to each other and close enough to overlap sideways — lobe touching lobe above the plane, and the opposing lobes touching below the plane. This lateral, off-axis overlap is qualitatively different from σ overlap: it has a nodal plane containing both nuclei (the molecular plane itself), and the electron density sits entirely above and below that plane, never on the internuclear axis. That is the π bond.

Figure

Building the sp² carbon: orbital mixing and the resulting geometry

Figure

Ethylene: sigma framework plus the pi bond from sideways p-orbital overlap

The core ideas

The hybrid composition, the bond angle, and the bond-strength decomposition are all quantitative and independently checkable.

\psi(sp^2) = \sqrt{\tfrac{1}{3}}\,\phi(2s) + \sqrt{\tfrac{2}{3}}\,\phi(2p)

where: $\psi(sp^2)$ is one normalized hybrid orbital, $\phi(2s)$ and $\phi(2p)$ are the pure atomic orbitals, and the coefficients-squared ( $\tfrac13$ and $\tfrac23$ ) are the fractional $s$ - and $p$ -character — reading this line: each $sp^2$ hybrid is one-third $s$ -like and two-thirds $p$ -like, which is why it is less directional and slightly larger-lobed than an $sp$ hybrid but more directional than $sp^3$ .

\theta_{sp^2} = 120.0^{\circ}, \qquad \cos\theta = -\tfrac{1}{2}

where: $\theta_{sp^2}$ is the ideal interorbital angle, forced by the requirement that three equivalent hybrids with equal $s$ -character be mutually orthogonal in the overlap sense — reading this line: exactly $120°$ is not empirical, it is the unique planar solution to orbital orthogonality among three equal hybrids.

D_{0}(\text{C=C}) = D_\sigma + D_\pi

where: $D_0(\text{C=C})$ is the total C=C bond dissociation energy ( $\approx 611\ \text{kJ/mol}$ ), $D_\sigma$ is the sigma contribution (approximated by the C–C single-bond value, $\approx 347\ \text{kJ/mol}$ ), and $D_\pi$ is the residual π contribution — reading this line: subtracting the sigma bond energy from the total double-bond energy isolates the π bond's own strength, roughly $264\ \text{kJ/mol}$ , weaker than the σ bond despite both being "one bond" in a Lewis structure.

E_\pi(\phi) \propto -\,S(\phi)^2, \qquad S(\phi) = S_0\cos\phi

where: $E_\pi(\phi)$ is the π-bonding stabilization when one $p_z$ orbital is twisted by angle $\phi$ out of alignment with its neighbor, $S(\phi)$ is the overlap integral at that twist angle, and $S_0$ is the overlap at $\phi = 0$ (perfectly parallel, planar alkene) — reading this line: π overlap falls off as $\cos\phi$ , so at $\phi = 90°$ the overlap and all π stabilization vanish completely, which is the mechanistic origin of the rotational barrier in alkenes.

Derivation AWhy three orbitals mix into exactly three hybrids at 120°, never four or two

Goal. Show that a carbon forming three sigma bonds in one plane is mechanistically required to hybridize as $sp^2$ with $120°$ angles, by orbital-count and overlap-maximization reasoning, not by fiat.

Step 1 — orbital count is conserved. Linear combination of atomic orbitals (LCAO) never changes the number of orbitals in a basis set; it only changes their shape and orientation. Carbon's valence set is $\{2s, 2p_x, 2p_y, 2p_z\}$ , four orbitals. If the atom is to build three equivalent sigma bonds (as in $sp^2$ carbon of ethylene, or the carbonyl carbon of formaldehyde), those three sigma bonds each need one hybrid orbital pointing at a distinct substituent. Mixing exactly three atomic orbitals — $2s$ , $2p_x$ , $2p_y$ — produces exactly three new hybrids; the untouched fourth orbital, $2p_z$ , is left over by necessity, not by choice, because there is no fourth sigma bond to build it for.

Step 2 — equal $s$ -character is forced by symmetry. If the three sigma bonds are chemically equivalent (three C–H bonds in ethylene, ignoring the C=C for a moment as reference), no single hybrid can be "more $s$ -like" than another without breaking that equivalence. The only self-consistent partition of one $2s$ orbital among three equivalent hybrids is $\tfrac{1}{3}$ share each. The remaining $\tfrac{2}{3}$ of each hybrid's character comes from the two $2p$ orbitals in the mixing set, distributed so that the three hybrids are mutually orthogonal.

Step 3 — orthogonality forces exactly $120°$ . Represent each hybrid as a unit vector direction weighted by its $p$ -character. For two hybrids $\psi_i$ and $\psi_j$ built from the same $s$ fraction $c_s^2 = \tfrac13$ and mutually orthogonal ( $\langle\psi_i|\psi_j\rangle = 0$ ), the geometric condition reduces to

c_s^2 + c_p^2\cos\theta_{ij} = 0

with $c_p^2 = 1 - c_s^2 = \tfrac{2}{3}$ . Substituting,

\cos\theta_{ij} = -\frac{c_s^2}{c_p^2} = -\frac{1/3}{2/3} = -\frac{1}{2}

so $\theta_{ij} = 120.0^{\circ}$ exactly. This is the same style of relation that gives $\cos\theta = -\tfrac{1}{3}$ ( $109.5°$ ) for $sp^3$ and $\cos\theta = -1$ ( $180°$ ) for $sp$ — a single formula, $\cos\theta = -c_s^2/(1-c_s^2)$ , spanning all three hybridization states with $c_s^2 = \tfrac14, \tfrac13, \tfrac12$ respectively.

Step 4 — planarity is the only way to satisfy three mutual $120°$ angles. Three vectors that are pairwise separated by $120°$ cannot be arranged non-coplanarly; any attempt to lift one vector out of the plane defined by the other two immediately changes at least one pairwise angle away from $120°$ . Hence the $sp^2$ solution is intrinsically, unavoidably planar — the flatness of the alkene sigma framework is a direct corollary of the orthogonality condition in Step 3, not an independent experimental input.

Step 5 — the leftover $p_z$ is orthogonal to the whole plane. Because $2p_z$ was excluded from the mixing set, it retains its pure nodal structure: a plane of zero amplitude coincident with the $xy$ -plane (the very plane the three hybrids occupy). Any bonding this orbital does must therefore occur off that plane, above and below it symmetrically — which is exactly the geometric definition of a π bond. The derivation of $sp^2$ geometry and the existence of a perpendicular π-capable orbital are thus two faces of the same three-orbital mixing choice.

Go deeper · Derivation B · 3 minConstructing the π molecular orbitals of ethylene and reading off the rotational barrier

Derivation BConstructing the π molecular orbitals of ethylene and reading off the rotational barrier

Goal. Build bonding and antibonding π molecular orbitals from the two $2p_z$ atomic orbitals, and show mechanistically why twisting the C=C bond by $90°$ costs the entire π bond energy.

Step 1 — set up the two-orbital basis. Each $sp^2$ carbon carries one unhybridized $2p_z$ orbital, call them $p_A$ on C1 and $p_B$ on C2, each with a lobe of $+$ phase above the molecular plane and a lobe of $-$ phase below it (an arbitrary but consistent phase convention). When the sigma framework holds the two carbons at their equilibrium C=C distance ( $\approx 1.34\ \text{Å}$ ) with both $p_z$ axes parallel, these two atomic orbitals are the correct starting basis for LCAO treatment of the π system, exactly analogous to how two $1s$ orbitals combine in $\text{H}_2$ .

Step 2 — form the symmetric (bonding) and antisymmetric (antibonding) combinations.

\pi = N_+\big(p_A + p_B\big), \qquad \pi^{*} = N_-\big(p_A - p_B\big)

where $N_+, N_-$ are normalization constants. In $\pi$ , the top lobes of $p_A$ and $p_B$ have the same phase and constructively overlap, building continuous electron density above (and, symmetrically, below) the internuclear axis, with no node between the nuclei in the direction perpendicular to the axis (there is still the permanent nodal plane containing the nuclei, inherent to any $p$ orbital). This constructive, in-phase overlap lowers the energy relative to the isolated atomic orbitals — it is the bonding π orbital, doubly occupied in ground-state ethylene. In $\pi^{*}$ , the phases are opposed, so a new node appears exactly on the perpendicular bisector between C1 and C2, cutting the density into two disconnected antibonding regions. This raises the energy above the atomic $p_z$ level; it is the LUMO of a simple alkene, the orbital electrophiles are shielded from and nucleophiles/reducing agents must populate.

Step 3 — assign electrons and read the bond order contribution. Ethylene contributes two electrons to this two-orbital π system (one from each carbon's leftover $p_z$ electron), and both go into the lower-energy $\pi$ MO. Net π bond order is

\text{(π bond order)} = \frac{(\text{electrons in }\pi) - (\text{electrons in }\pi^{*})}{2} = \frac{2 - 0}{2} = 1

Combined with the sigma bond from the in-plane $sp^2$ – $sp^2$ overlap, the total C=C bond order is 2 — the familiar double bond, now built from two mechanistically distinct orbital overlaps rather than treated as "two lines" of a Lewis structure.

Step 4 — twist the C–C axis and track the overlap integral. Now twist one $\text{CH}_2$ fragment by dihedral angle $\phi$ about the C–C sigma bond, keeping the sigma bond itself intact (sigma overlap, being cylindrically symmetric about the internuclear axis, is unaffected by this rotation). The two $p_z$ axes, originally parallel, now subtend angle $\phi$ to each other. The π overlap integral scales as

S(\phi) = S_0 \cos\phi

because only the component of $p_B$ that remains parallel to $p_A$ contributes to lobe-on-lobe overlap; the perpendicular component overlaps with zero net constructive area by symmetry (it is trying to overlap side-lobe to end-lobe, which integrates to zero over the relevant symmetry).

Step 5 — the barrier maximum at $\phi = 90°$ . At $\phi = 90°$ , $\cos 90^{\circ} = 0$ , so $S(90^{\circ}) = 0$ : the two $p_z$ orbitals are mutually perpendicular, one lying in what was the molecular plane and the other lying perpendicular to it, and they achieve zero net overlap by symmetry (any constructive overlap on one side is exactly canceled by destructive overlap on the other, or the geometry simply removes lobe-lobe contact entirely). All π bonding stabilization is lost at this geometry; the energy cost of getting there, relative to the planar minimum, is the rotational barrier, and it is experimentally close to the isolated π bond strength extracted in the bond-energy derivation ( $\approx 264\ \text{kJ/mol}$ , consistent with measured barriers to alkene isomerization of roughly $250$ – $270\ \text{kJ/mol}$ ).

Step 6 — why this predicts $E/Z$ stereochemistry. Because rotation through the $90°$ twisted, π-broken transition state costs essentially the entire π bond energy — far more than ordinary thermal energies at room temperature can supply — substituents on a C=C double bond cannot interconvert their relative positions without a very high energy input (or without a mechanism that breaks the π bond, such as photochemical excitation to the $\pi^{*}$ state, or acid-catalyzed addition–elimination). This is the mechanistic root of $E/Z$ (cis/trans) isomerism: it is not a rule to memorize, it is a direct consequence of the $\cos\phi$ overlap dependence derived in Step 4.

Worked examples

Example 1 — Foundational: assigning hybridization and counting sigma/pi bonds in propenal (acrolein, $\text{CH}_2=\text{CH–CHO}$ ). Draw the skeleton $\text{C1H}_2=\text{C2H–C3HO}$ . C1 bears two H's and a double bond to C2: three sigma-bonded groups (two H, one C2), zero lone pairs, so C1 is $sp^2$ with a leftover $p_z$ used for the C1=C2 π bond. C2 bears one H, a double bond to C1, and a single bond to C3: again three sigma connections, $sp^2$ , leftover $p_z$ engaged in the same C1=C2 π system. C3 (the aldehyde carbon) bears one H, a single bond to C2, and a double bond to O: three sigma connections, $sp^2$ , leftover $p_z$ forming the C3=O π bond with an oxygen $p_z$ orbital. The oxygen itself is $sp^2$ (two lone pairs occupy two $sp^2$ hybrids, one $sp^2$ hybrid forms the O–C3 sigma bond in the trigonal framework, and oxygen's own leftover $p_z$ supplies the other half of the π bond) — though some treatments simplify carbonyl oxygen as unhybridized for lone-pair counting purposes, the $sp^2$ treatment correctly predicts the observed near- $120°$ C–O behavior in resonance and the coplanarity of the whole enal system. Total count: 3 carbons all $sp^2$ , two independent π bonds (C1=C2 and C3=O), sigma framework fully coplanar because every $sp^2$ center shares the same hybridization plane once the molecule adopts its conjugated, planar minimum-energy geometry — this coplanarity is precisely what allows the two π systems to conjugate, foreshadowing Conjugation: Overlapping p Orbitals.

Example 2 — Applied: predicting the relative rate of epoxidation for two alkenes using π-orbital electron density. Consider m-CPBA epoxidation (electrophilic oxygen transfer) of 2-methylpropene, $(\text{CH}_3)_2\text{C=CH}_2$ , versus ethylene, $\text{CH}_2=\text{CH}_2$ . The reaction mechanism is concerted: the alkene π bond, acting as the nucleophile, attacks the electrophilic oxygen of the peracid, with simultaneous O–O bond cleavage in a butterfly transition state. The rate-determining step's transition state has substantial positive charge character developing on the alkene carbons as the π electrons are drawn toward oxygen. In 2-methylpropene, both methyl groups are attached directly to the $sp^2$ carbon bearing the developing positive character; each methyl group donates electron density into the adjacent, in-plane $\sigma_{C-H}$ orbitals that hyperconjugate with the empty-character $p_z$ orbital developing at the alkene carbon (hyperconjugative stabilization, the same donation invoked for carbocation stability). This raises the ground-state HOMO ( $\pi$ ) energy of 2-methylpropene relative to unsubstituted ethylene — more alkyl substitution pushes π electron density up in energy, making it a better nucleophile toward the electrophilic peracid oxygen. Experimentally, 2-methylpropene reacts with m-CPBA on the order of $10^{2}$ – $10^{3}$ times faster than ethylene under comparable conditions, consistent with the general Markovnikov-type reactivity trend that more substituted, more electron-rich double bonds are more nucleophilic — a direct, mechanistically traceable consequence of how many hyperconjugating C–H sigma bonds are positioned to feed density into the reacting π system.

Example 3 — Exam/competition level: rationalizing why trans-cyclooctene is chiral and isolable, while trans-cyclohexene is not, using the π-overlap/ring-size argument. For a cyclic alkene to adopt the trans (or $E$ ) configuration, the two ring substituents attached to the double bond must lie on opposite faces, which forces the rest of the ring to twist and thread across the double bond plane. From the derivation of $S(\phi) = S_0\cos\phi$ , the π bond is only intact (near-maximal $\cos\phi \approx 1$ ) if the two $p_z$ orbitals remain close to parallel; any ring strain that forces significant pyramidalization at the alkene carbons, or twists the substituents, directly reduces $\phi$ 's deviation from zero and destabilizes the π bond continuously, not catastrophically. In trans-cyclohexene (an 8-atom-counting ring including the two alkene carbons — six total ring atoms), the tether connecting the two alkene carbons is too short to span the required trans geometry without forcing $\phi$ well past the point where $\cos\phi$ collapses the π overlap close to zero; the resulting strain energy is so large (calculated ring strain in excess of $150\text{–}200\ \text{kJ/mol}$ beyond the cis isomer) that trans-cyclohexene is not isolable at ordinary temperatures, existing only as a fleeting, highly reactive intermediate. In trans-cyclooctene (eight ring atoms), the extra two methylenes provide enough conformational slack that the ring can thread across the double bond while keeping $\phi$ close enough to $0°$ that meaningful π overlap, and thus a genuine, isolable double bond, survives; the residual strain (about $40\ \text{kJ/mol}$ above cis-cyclooctene) is enough to make trans-cyclooctene higher in energy, chiral (the twisted ring is not superimposable on its mirror image because the alkene plane and the ring's threading direction break the local symmetry), and kinetically resistant to isomerization back to cis, but not so strained that the π bond fails outright. The entire qualitative trend — impossible at six-membered, difficult but isolable at eight-membered, comfortable by ten-membered rings — falls directly out of tracking how much a given ring size forces $\phi$ away from the $\cos\phi \approx 1$ optimum derived from the π-overlap mechanism.

Limiting cases, scope & common traps

Regime / question	What actually happens	Mechanistic reason
$\phi \to 0°$ (planar alkene)	Maximum π overlap, full bond order 2	$S(\phi) = S_0\cos\phi \to S_0$
$\phi \to 90°$ (twisted alkene, e.g. photoexcited or highly strained)	π bond entirely broken; behaves as a 1,2-biradical/zwitterion	$\cos 90° = 0$ , no lobe-lobe overlap survives
Carbonyl carbon (C=O)	Same $sp^2$ /π picture, but π* is much lower in energy (O is more electronegative)	Polarized π MO: bonding π has more O character, π has more C character — this is why* nucleophiles add to the carbonyl carbon, not oxygen
Allylic/benzylic radical or cation adjacent to a π system	Extra stabilization beyond simple $sp^2$ picture	The p orbital on the adjacent sp³ (or radical/cationic) center is not orthogonal to the π system if rotated into alignment — this is the gateway to Conjugation: Overlapping p Orbitals
Comparing $sp^3$ , $sp^2$ , $sp$	$s$ -character increases $25\% \to 33.3\% \to 50\%$ , angle increases $109.5° \to 120° \to 180°$	Same orthogonality formula $\cos\theta = -c_s^2/(1-c_s^2)$ with $c_s^2 = \tfrac14, \tfrac13, \tfrac12$
Nitrogen/oxygen analogues (imine C=N, carbonyl C=O)	Heteroatom is also $sp^2$ ; lone pairs occupy in-plane $sp^2$ hybrids, not the π system, unless conjugation demands otherwise	Same three-orbital mixing logic, just with lone pairs substituting for one or two of the sigma-bond positions

Common trapMisconception: "the double bond is just two single bonds stacked together"

Treating C=C as mechanistically equivalent to two independent C–C sigma bonds is wrong and misleads reactivity predictions. The two components are orbital-symmetry-distinct: sigma overlap is head-on and cylindrically symmetric about the internuclear axis (free rotation would cost nothing if only sigma existed), while pi overlap is lobe-on-lobe, has an intrinsic nodal plane, is weaker per unit distance, and is destroyed by a $90°$ twist. This asymmetry between the two components is exactly why alkenes show restricted rotation (only the π component, not the σ, opposes twisting) and why the π electrons — not the σ framework — are the electrophile/nucleophile-reactive part of the molecule in addition reactions.

Common trapMisconception: "

sp^2

hybridization requires a double bond to be present"

Boron in $\text{BF}_3$ is $sp^2$ hybridized with a genuinely empty $p_z$ orbital and no double bonds at all — the trigonal planar geometry and the leftover perpendicular $p$ orbital arise from three sigma-bonding requirements, not from the existence of a π bond. The carbanion in an allylic or enolate system is $sp^2$ with a doubly-occupied $p_z$ orbital, again without a formal double bond localized there in every resonance structure. Hybridization state is dictated by the number of sigma-bonded groups plus lone pairs (three groups → trigonal planar → $sp^2$ ), and then one checks separately whether the leftover $p$ orbital happens to be empty, singly occupied, doubly occupied, or engaged in π bonding with a neighbor.

The chemist behind it

The modern hybridization framework is the joint legacy of quantum chemists who translated Lewis's electron-pair bonding into orbital language usable for real molecular geometry predictions.

Origins · 1 minLinus Pauling and how this was discovered

Linus Pauling

1901–1994 · Caltech, 1930s valence bond theory

Pauling formalized hybridization mathematically in his 1931 papers "The Nature of the Chemical Bond," showing how mixing $s$ and $p$ atomic orbitals in different ratios produces directional hybrids whose angles and bond strengths could be calculated, not just postulated from observed geometry. He introduced the language of $sp$ , $sp^2$ , and $sp^3$ hybrid orbitals explicitly, connected hybrid composition to bond strength via overlap maximization, and showed that the resulting orbital pictures reproduce observed molecular shapes — tetrahedral methane, trigonal planar boron trifluoride and ethylene, linear acetylene — as consequences of orbital algebra rather than empirical rules. His 1954 Nobel Prize in Chemistry recognized this and the broader body of work on the nature of the chemical bond and its application to complex substances.

History

- 1916 — G. N. Lewis proposes the shared electron-pair bond, giving chemists the dot-structure/double-bond notation later explained mechanistically by hybridization.

- 1927 — Heitler and London's valence bond treatment of

\text{H}_2

establishes overlap of atomic orbitals as the quantum-mechanical origin of bonding.

- 1931 — Linus Pauling publishes the hybridization formalism, introducing

sp

sp^2

sp^3

notation and directional bond-strength arguments.

- 1933 — Erich Hückel's treatment of ethylene and benzene separates the sigma framework from the pi system explicitly, the direct ancestor of the sigma/pi decomposition used throughout organic mechanism.

- 1951 — Molecular orbital treatments (Mulliken and successors) formalize the bonding/antibonding π and π* picture used to explain UV–vis spectra and pericyclic reactivity of alkenes.

Why does carbon "choose" sp² instead of sp³ when forming a double bond?

Hybridization state follows directly from how many sigma-bonded groups (plus lone pairs) an atom must accommodate, counted first via something like VSEPR logic. A carbon forming a double bond to one neighbor, plus two further single bonds (as in each alkene carbon of ethylene), has exactly three sigma-bond directions to accommodate — the double bond contributes only one sigma bond to that count, since its second bond is the perpendicular π component built from the leftover orbital. Three sigma directions is precisely what $sp^2$ (three hybrids) is built for; using $sp^3$ would leave one hybrid without a sigma-bonding partner and would also fail to produce the pure, unhybridized $p$ orbital needed for π bonding. The double bond does not directly cause $sp^2$ hybridization — rather, both the double bond's existence and $sp^2$ hybridization are simultaneous consequences of the atom needing exactly three sigma connections and one perpendicular p orbital.

Is the pi bond weaker than the sigma bond, and how do we know?

Yes, and the evidence comes from bond-energy bookkeeping. The C–C single bond (pure sigma) has a bond dissociation energy of about $347\ \text{kJ/mol}$ . The C=C double bond's total dissociation energy is about $611\ \text{kJ/mol}$ , not $694\ \text{kJ/mol}$ (twice the single-bond value) as it would be if the two components were equally strong. Subtracting the sigma contribution leaves roughly $264\ \text{kJ/mol}$ for the π component alone — meaningfully weaker than the sigma bond. This matches the orbital picture: head-on sigma overlap achieves better spatial overlap at bonding distance than the necessarily off-axis, lobe-on-lobe geometry of π overlap, so π bonding is intrinsically less efficient per unit distance.

Does hybridization theory have a rigorous quantum mechanical basis, or is it just a bookkeeping convenience?

Hybridization is a mathematically legitimate transformation: any set of orthonormal atomic orbitals can be recombined via a unitary transformation into an equivalent orthonormal set (the hybrids) without changing the total wavefunction or energy of an isolated atom. Its real predictive power appears only once bonds form, because hybrids are chosen specifically to maximize directional overlap with neighboring atoms' orbitals — a choice justified by the variational principle (better overlap lowers the bonding orbital's energy more). Modern ab initio and density functional calculations do not require hybridization as an input, but when you inspect the resulting localized molecular orbitals of a computed alkene, they closely resemble the two-lobed, $120°$ -separated hybrid picture, confirming that hybridization is a useful, physically grounded lens rather than an arbitrary convention.

How does sp² hybridization explain the geometry around a carbocation or a radical, not just double bonds?

A carbocation with three sigma-bonded substituents and no additional lone pair or π bond is still best described as $sp^2$ at the cationic carbon, because it still has exactly three sigma-bonding directions and needs a trigonal planar arrangement to minimize repulsion among those three bonding pairs. The leftover $p_z$ orbital is simply empty rather than engaged in a π bond — this is precisely why carbocations are planar (confirmed by their tendency to react with nucleophiles from either face equally, a hallmark of $sp^2$ , not tetrahedral, geometry) and why adjacent π systems or C–H sigma bonds can donate density into that empty orbital (hyperconjugation), stabilizing the cation. The same logic makes simple alkyl radicals close to planar (or very shallowly pyramidal with a low inversion barrier), since the odd electron occupies that same leftover $p_z$ -like orbital rather than forcing a fourth sigma-bond-style direction.

What happens to the sp² picture when the double bond is part of a larger conjugated or aromatic system?

The basic per-atom picture is unchanged — each participating carbon is still individually $sp^2$ , with three coplanar sigma bonds and one perpendicular $p_z$ orbital — but the π system is no longer confined to a single two-orbital bonding/antibonding pair. When multiple $sp^2$ centers are arranged so that their $p_z$ orbitals are all parallel and contiguous (as in 1,3-butadiene, benzene, or an extended polyene), the individual two-orbital π/π* treatment generalizes to a multi-orbital LCAO problem, producing a whole ladder of π molecular orbitals spanning the entire conjugated framework rather than being localized on one bond. This delocalization, only possible because every relevant carbon shares the same $sp^2$ geometry and coplanar arrangement, is the mechanistic foundation of aromatic stabilization and of extended electronic absorption in conjugated dyes and polymers.

Why can't rotation about a single (sigma-only) C–C bond be restricted the same way rotation about a double bond is?

A pure sigma bond, formed from head-on overlap of two orbitals along the internuclear axis, is cylindrically symmetric about that axis — rotating one end of the molecule about the bond axis does not change the overlap integral at all, because the orbital density distribution around the axis is unchanged by the rotation (think of spinning a football along its long axis; its silhouette from any angle around that axis looks identical). A π bond, by contrast, is built from off-axis lobes with a definite orientation perpendicular to the axis, so rotating one fragment changes the relative orientation of the two $p_z$ lobes and directly reduces the overlap integral, as captured by $S(\phi) = S_0\cos\phi$ . This geometric asymmetry — cylindrical symmetry for sigma, directional dependence for pi — is the entire reason free rotation is essentially unhindered about single bonds but strongly hindered about double bonds.

Is the carbonyl oxygen in a ketone or aldehyde really sp² hybridized like the carbon?

Treating carbonyl oxygen as $sp^2$ is a defensible and commonly used model: it places two lone pairs and one sigma bond (to carbon) into three coplanar $sp^2$ hybrids, leaving a pure $p_z$ orbital perpendicular to the plane to form the π bond with carbon's $p_z$ orbital, consistent with the observed coplanarity of substituents around C=O and with the geometry-dependent reactivity of carbonyls (nucleophilic attack occurs preferentially along the Bürgi–Dunitz trajectory relative to that plane). Some simpler treatments instead describe oxygen's lone pairs as residing in unhybridized $p$ and $s$ orbitals, since oxygen's lone pairs do not need directional bonding character in the same way bonding pairs do; both models are used in practice, and the $sp^2$ version is generally preferred because it correctly reproduces the near- $120°$ geometry oxygen adopts in extended conjugated and resonance-stabilized carbonyl systems.