VSEPR Theory: Electrons Repel, Shapes Emerge

One principle: electron domains avoid each other

Every prediction in VSEPR theory flows from a single physical idea: regions of electron density around a central atom repel each other, so they spread out as far as geometry allows. Like charges push apart — that is the entire engine.

Chemists call each region an electron domain: a bonding pair (whether single, double, or triple bond — each bond is one domain) or a lone pair. The domains arrange themselves to maximize their separation, and the resulting arrangement is the electron geometry.

Lone pairs complicate the picture because they repel more strongly than bonding pairs. A bonding pair is stretched between two nuclei, but a lone pair belongs to one nucleus alone, so it spreads wider and hugs the central atom more closely. This extra bulk squeezes the neighboring bonds together, shrinking bond angles below the ideal values — a theme that runs through every example ahead.

Counting domains: the AXE method

To apply VSEPR you first inventory the domains with AXE notation. A stands for the central atom, X counts the bonding domains, and E counts the lone pairs. The steric number — the total domains, X + E — selects the geometry.

Take ammonia, NH3. Its Lewis structure shows three N−H bonds and one lone pair on nitrogen: AX3E, steric number 4. Water, H2O, has two bonds and two lone pairs: AX2E2, also steric number 4. Carbon dioxide, O=C=O, has two double bonds and no lone pairs on carbon: AX2, steric number 2 — remember, each multiple bond is a single domain.

The notation is compact but complete: from AX3E you know everything VSEPR needs. A frequent error is counting a double bond as two domains or forgetting lone pairs entirely — both corrupt the steric number and wreck the prediction. Count regions, not electrons.

The five electron geometries

Five steric numbers cover nearly every main-group molecule:

• 2 domains — linear, 180°. The domains point opposite each other. Examples: BeCl2, CO2.
• 3 domains — trigonal planar, 120°. Three domains at the corners of a triangle. Examples: BF3, SO3.
• 4 domains — tetrahedral, 109.5°. Four domains toward tetrahedron corners. Examples: CH4, NH4+.
• 5 domains — trigonal bipyramidal, 90° and 120°. Three around the equator, two at the poles. Examples: PCl5.
• 6 domains — octahedral, 90°. Six domains toward octahedron corners. Examples: SF6.

These are the electron geometries — the arrangements of all domains. They are ideals: perfectly symmetric, perfectly spaced. Real molecules deviate when lone pairs enter, which is exactly what the next section quantifies.

Lone pairs squeeze harder

Here is VSEPR's most testable prediction: as lone pairs replace bonding pairs, bond angles shrink. The sequence CH4 → NH3 → H2O demonstrates it cleanly.

Methane has four bonding domains and no lone pairs: the H−C−H angle is the ideal 109.5°. Ammonia swaps one bonding pair for a lone pair (AX3E): the lone pair's extra repulsion squeezes the H−N−H angle to 107°. Water swaps another (AX2E2): two lone pairs compress the H−O−H angle to 104.5°.

The repulsion ranking behind this is lone pair–lone pair > lone pair–bonding pair > bonding pair–bonding pair. Lone pairs are fatter because a single nucleus holds them loosely, letting them sprawl.

This is a worked-logic prediction, not memorization: given any AXE formula, you can estimate whether its angles sit at, slightly below, or well below the ideal. Examiners love this chain precisely because it tests understanding rather than recall.

Electron geometry is not molecular geometry

A subtle but crucial distinction: electron geometry counts every domain, while molecular geometry names only where the atoms sit. Lone pairs shape the molecule but are invisible in its name.

Ammonia illustrates the split. Its four domains (AX3E) adopt the tetrahedral electron geometry — but the molecular geometry, describing the three hydrogens around nitrogen, is called trigonal pyramidal. Water's domains are tetrahedral too, yet its molecular geometry is bent.

When the central atom has no lone pairs, the two names coincide: CH4 is tetrahedral in both systems, CO2 linear in both. The divergence appears exactly when lone pairs occupy domains.

Keeping the two vocabularies separate prevents the most common VSEPR error: calling ammonia 'tetrahedral.' The electrons are tetrahedral; the molecule is pyramidal. The full naming system for molecular shapes gets its own article next.

Where VSEPR stumbles

VSEPR is a model, and models have borders. It assumes electron domains behave as simple repelling regions, which works beautifully for main-group elements but falters where d orbitals participate — many transition-metal complexes adopt geometries VSEPR never predicts.

Very heavy atoms can also surprise: their large, diffuse orbitals and relativistic effects bend the simple repulsion picture. And VSEPR says nothing about bond lengths or energies, only about angles and arrangements.

None of this diminishes its value. For the molecules of introductory chemistry — the hydrides, halides, and oxides of the main groups — VSEPR predicts shapes correctly the overwhelming majority of the time, in seconds, from a Lewis structure alone. It is the fastest reliable shape predictor in chemistry, which is why every chemist learns it first and reaches for heavier theory only when it fails.

Common mistakes

Real-world applications

• Predicting molecular shapes before spectroscopy

• Explaining bond-angle trends in teaching labs

• Estimating polarity from geometry

• Designing ligands for catalysts