How VSEPR Predicts Molecular Shape: A Practical Guide

Molecules are not flat drawings. Each one has a three-dimensional shape, and that shape controls how it fits into other molecules, how it looks in a spectrum and whether it is polar. Valence shell electron pair repulsion theory, usually called VSEPR, gives a simple way to predict those shapes from a count of electron pairs.

This guide explains the core idea, how to count electron domains, how lone pairs change the answer and how to check predictions against measured bond angles. The angles quoted are standard values that you can confirm in any general chemistry reference.

The core idea

VSEPR starts from one assumption. Electron pairs around a central atom repel each other, so they arrange themselves as far apart as possible. The molecule's shape follows from that arrangement.

Electron pairs include both bonding pairs, which are shared between two atoms, and lone pairs, which belong to one atom only. Each bond counts as one region of electron density, whether it is a single, double or triple bond. A double bond is still one region, which is why CO₂ is linear even though it has four bonds to oxygen in total.

Step 1: count the electron domains

Find the central atom and count its valence electrons. Then count how many electron domains surround it. Each bond is one domain, and each lone pair is one domain.

For example, in methane, CH₄, carbon has four bonds and no lone pairs, so it has four electron domains. In ammonia, NH₃, nitrogen has three bonds and one lone pair, giving four domains. In water, oxygen has two bonds and two lone pairs, also giving four domains.

Step 2: pick the base geometry

The electron domain count sets the base arrangement. These are the ideal geometries that maximise distance between domains.

Electron domains	Base arrangement	Ideal angle(s)
2	Linear	180°
3	Trigonal planar	120°
4	Tetrahedral	109.5°
5	Trigonal bipyramidal	90° and 120°
6	Octahedral	90°

The base arrangement describes the positions of all electron domains, bonding or not. The molecular shape is named by looking only at the atoms. Lone pairs occupy positions but are not counted in the shape name. As a rough check, a measured bond angle should sit near the ideal value for its base arrangement, and a value far outside that range signals that the domain count or the lone pair placement needs another look.

Step 3: apply lone pair corrections

Lone pairs sit closer to the central atom than bonding pairs do, because they are attached to only one nucleus. They therefore take up more room and push the bonding pairs closer together.

The usual repulsion order is lone pair to lone pair strongest, then lone pair to bonding pair, then bonding pair to bonding pair weakest. Each lone pair reduces the bond angles a little.

Molecule	Electron domains	Lone pairs	Molecular shape	Approximate bond angle
CO₂	2	0	Linear	180°
BF₃	3	0	Trigonal planar	120°
CH₄	4	0	Tetrahedral	109.5°
NH₃	4	1	Trigonal pyramidal	about 107°
H₂O	4	2	Bent	about 104.5°
PCl₅	5	0	Trigonal bipyramidal	90° and 120°
SF₆	6	0	Octahedral	90°

The drop from 109.5° in methane to 107° in ammonia to 104.5° in water is the cleanest demonstration of the lone pair effect. Water has two lone pairs, so its angle shrinks the most.

Worked example: sulfur tetrafluoride

Sulfur tetrafluoride, SF₄, is a useful test because it is not symmetric. Sulfur has four bonds and one lone pair, giving five electron domains. The base arrangement is trigonal bipyramidal.

Lone pairs prefer equatorial positions in a trigonal bipyramid, because there they have fewer close neighbours at 90° angles. Placing the single lone pair in an equatorial position leaves the molecule with two axial fluorines and two equatorial fluorines. The name for this shape is seesaw.

This example shows why the geometry matters. The seesaw shape makes SF₄ polar, while a symmetric shape such as SF₆ is nonpolar.

Worked example: xenon tetrafluoride

Xenon tetrafluoride, XeF₄, has four bonds and two lone pairs, giving six electron domains. The base arrangement is octahedral. The two lone pairs sit opposite each other, which leaves the four fluorine atoms in a flat square around the xenon. The shape is square planar.

Predicting the square planar shape requires placing the lone pairs as far apart as possible. A naive guess that ignores lone pair positions would miss this result.

Polarity follows from shape

Molecular shape decides whether bond dipoles cancel. Each polar bond has a dipole pointing toward the more electronegative atom. In a symmetric shape, the dipoles can cancel to zero. In an asymmetric shape, they add up to a net dipole.

Carbon dioxide has two polar C=O bonds that point in opposite directions in a linear arrangement, so they cancel and the molecule is nonpolar. Water has two polar O–H bonds at 104.5°, and they do not cancel, so water has a large net dipole. That dipole is why water dissolves salts and forms hydrogen bonds so readily.

Shape therefore matters even when bond types are the same. Two molecules can contain identical bonds and still have different polarity depending on their geometry.

Limits of the model

VSEPR is a model, not a complete theory of bonding. It predicts shapes well for main-group molecules, especially those with central atoms in rows two and three of the periodic table. It works less cleanly for some transition metal complexes, where d electrons and ligand fields also influence geometry.

The model also treats electron domains as equal in size, which is an approximation. Real bond angles depend on electronegativity differences, multiple bonds and steric crowding of large substituents. A molecule with bulky groups may deviate from the ideal angle even when the domain count is simple.

When a predicted angle does not match a measured value, the mismatch is often informative. It can point to factors such as partial double-bond character or to a shape that was not considered in the first pass.

Common mistakes

The most frequent error is counting lone pairs in the shape name. Ammonia is trigonal pyramidal, not trigonal planar, even though its four domains are arranged tetrahedrally. Another error is forgetting that a double bond is one domain. A third is placing lone pairs in axial positions in a trigonal bipyramid without checking which arrangement minimises repulsion.

A quick check is to draw the Lewis structure first, count domains from it and then match the count to the base arrangement table. If the shape name and the base arrangement differ, the lone pairs are the reason.

Key takeaways

VSEPR predicts molecular shape by treating electron domains as charged regions that repel and spread apart as far as possible. Count bonds and lone pairs around the central atom, choose the base arrangement, then adjust the shape name by ignoring lone pairs.

Lone pairs take up more space than bonding pairs and compress bond angles. That effect explains why water is bent at 104.5° while methane is near 109.5°.

Shape controls polarity. Symmetric arrangements can cancel bond dipoles, while bent, pyramidal and seesaw shapes usually leave a net dipole. Use VSEPR as a fast, reliable first pass for main-group molecules, and treat exceptions as a signal to look more closely.