Covalent Bonding and Lewis Structures

Covalent bonding explains how atoms share electrons instead of transferring them outright. This differs from Ionic Bonding and Lattice Energy. There, electrons hop from one atom to another. This forms charged ions held together by electrostatic attraction. Covalent bonds form differently. Two atoms — typically nonmetals close together on the Periodic Table and Electron Configurations — each contribute electrons. These electrons form a shared pair that binds the atoms together. Lewis structures are the bookkeeping system chemists use. They track these shared and unshared electrons. A correct Lewis structure tells you which atoms are bonded. It shows how many bonds join them. It also shows where the lone pairs sit. This information becomes the launching point for predicting molecular shape, polarity, and reactivity. This article builds the octet rule from first principles. It gives a step-by-step algorithm for drawing Lewis structures. It works through molecules of increasing complexity. Every rule is checked against a real, countable electron total.

Key takeaways

A covalent bond is a shared pair of electrons; it forms when sharing lowers the total energy of the two atoms compared to their separate atomic states.
The octet rule says main-group atoms tend toward eight valence electrons (a noble-gas configuration); hydrogen instead aims for two.
Lewis structures are drawn by counting total valence electrons, placing single bonds between connected atoms, then distributing remaining electrons as lone pairs, upgrading to double/triple bonds if octets are unmet.
Bond order (single, double, triple) correlates with bond length (longer to shorter) and bond strength (weaker to stronger).
Real molecules sometimes violate the octet rule — odd-electron species, electron-deficient atoms like boron, and expanded octets on period-3-and-beyond atoms are the recognized exceptions.
Key idea: a Lewis structure is a valence-electron ledger — every electron in the starting count must appear somewhere in the final picture, as a bond or a lone pair.

The core insight

Atoms share electron pairs to complete their valence shells; the number of bonds and lone pairs an atom forms is dictated by how many more electrons it needs to reach a stable octet (or duet, for hydrogen).

The physical picture

Picture two hydrogen atoms approaching each other, each with one electron in a 1s orbital. Isolated, neither atom has a filled shell. As they approach, each nucleus begins to attract the other atom's electron, not just its own. If the atoms get close enough, the electron pair concentrates in the region between the two nuclei, and this shared negative charge attracts both positive nuclei simultaneously. That mutual attraction is the covalent bond. It is purely electrostatic in origin, but the pairing and placement of electrons is governed by quantum mechanics, which is why chemists use simplified dot diagrams rather than solving Schrödinger's equation for every bond. The energy of the H–H system as a function of internuclear distance has a minimum: pull the atoms too far apart and the attraction vanishes; push them too close and nuclear-nuclear repulsion (plus electron-electron repulsion) dominates. The bond length, 74 pm for H₂, sits at that energy minimum, and the depth of the well below the separated-atoms energy is the bond dissociation energy, 436 kJ/mol for H₂. Lewis structures compress this physics into dots and lines. A shared pair between two atomic symbols is a bonding pair, drawn as a line; electrons that stay with one atom are lone pairs (or nonbonding pairs), drawn as dot pairs. Only the valence electrons — those in the outermost shell — are shown; core electrons are assumed to stay put and are omitted entirely.

Figure

Energy versus internuclear distance for two hydrogen atoms approaching each other

Now consider a slightly more elaborate case, carbon dioxide, to see how multiple shared pairs and lone pairs distribute across a whole molecule.

Figure

Lewis structure of carbon dioxide, O=C=O, showing all valence electrons

The core relations

The starting point for any Lewis structure is a strict electron count.

N_{\text{total}} = \sum_{\text{atoms}} (\text{valence electrons per atom}) \pm (\text{charge electrons})

where:

$N_{\text{total}}$ is the total number of valence electrons to be distributed as dots and lines,
valence electrons per atom come from the group number in the main-group periodic table (Group 1A → 1, up through Group 8A → 8),
for a cation, subtract electrons equal to the positive charge; for an anion, add electrons equal to the magnitude of the negative charge.

Every bond, single or multiple, removes two electrons from the "unassigned" pool and places them between two atoms:

N_{\text{total}} = 2(\text{number of bonding pairs}) + 2(\text{number of lone pairs})

where:

number of bonding pairs counts every shared pair, so a double bond contributes 2 and a triple bond contributes 3 to this count,
number of lone pairs counts nonbonding pairs, wherever they sit.

The octet-satisfaction check for any atom (other than hydrogen, helium, and a short exception list) is:

(\text{bonds around atom}) + (\text{lone pairs on atom}) = 4

where:

"bonds around atom" counts each connection to a neighboring atom as one, regardless of whether that connection is single, double, or triple — i.e., you count groups of bonds, and a double bond still counts once as one shared region but contributes 2 electrons; for the octet total itself, electrons around the atom (bonding + lone) must sum to 8,
this is normally restated as: electrons around atom = 2×(bonds, counting single=1, double=2, triple=3) + 2×(lone pairs) = 8.

Derivation ABuilding a Lewis structure from the valence-electron count

The goal is to turn a molecular formula into a structure where every atom (except recognized exceptions) has a completed octet and the total electron count matches $N_{\text{total}}$ exactly.

Step 1. Count total valence electrons. Sum the group-number electron counts for every atom in the formula, then adjust for overall charge if the species is an ion.

N_{\text{total}} = \sum_i g_i \pm q

where $g_i$ is the group-number valence count of atom $i$ and $q$ is the ionic charge (subtract for cations, add for anions).

Step 2. Choose a skeletal (connectivity) structure. Identify a central atom — usually the least electronegative nonhydrogen atom, since it can support the most bonds — and attach the remaining atoms to it. Hydrogen and halogens almost always sit on the outside, since they form only one bond each.

Step 3. Place one single bond along every connection drawn in Step 2. Each single bond uses 2 electrons. Subtract these from $N_{\text{total}}$ to get the electrons still available:

N_{\text{remaining}} = N_{\text{total}} - 2(\text{number of bonds drawn})

Step 4. Distribute $N_{\text{remaining}}$ as lone pairs, starting on the outer (terminal) atoms. Give every terminal atom enough lone pairs to complete its octet (or duet for hydrogen) before touching the central atom.

Step 5. Check the central atom's octet. Count electrons already around the central atom (its bonding pairs, doubled, plus any lone pairs already placed). If it has fewer than 8, and no lone pairs remain to give it, convert one lone pair from a terminal atom into a second (or third) bond to the center — this creates a double or triple bond and does not change $N_{\text{total}}$ , since a shifted lone pair simply becomes a shifted bonding pair.

Step 6. Verify the full ledger. Add up every bonding-pair electron (2 per single bond, 4 per double bond, 6 per triple bond) and every lone-pair electron (2 per lone pair) across the whole structure. This sum must equal $N_{\text{total}}$ exactly — if it does not, an electron has been lost or invented somewhere in Steps 3–5, and the structure must be redrawn.

Worked examples

Example 1 — Carbon dioxide, CO₂ (asymmetric electron flow, symmetric molecule) Step 1. Valence electrons: C contributes 4, each O contributes 6.

N_{\text{total}} = 4 + 6 + 6 = 16

Step 2. Carbon is less electronegative and has more available bonding slots than oxygen, so carbon is the central atom, bonded to both oxygens: O–C–O. Step 3. Two single bonds use 4 electrons.

N_{\text{remaining}} = 16 - 2(2) = 12

Step 4. Distribute 12 electrons (6 lone pairs) onto the two terminal oxygens, 3 lone pairs each, giving each oxygen 1 bond + 3 lone pairs = 2 + 6 = 8 electrons around it. All 12 remaining electrons are now used, so none are left for carbon. Step 5. Check carbon: it currently has only 2 bonds × 2 electrons = 4 electrons around it — short of the octet. Convert one lone pair from each oxygen into a second bond to carbon (this does not change the total electron count, since 2 lone-pair electrons become 2 bonding-pair electrons at each oxygen):

\text{O=C=O}

Now carbon has 4 bonds (2 double bonds) → 8 electrons around it, and each oxygen has 1 double bond (4 electrons) + 2 lone pairs (4 electrons) = 8 electrons. Step 6. Ledger check:

\underbrace{2(4)}_{\text{2 double bonds}} + \underbrace{2(4)}_{\text{4 lone pairs, one on each O}} = 8 + 8 = 16

This matches $N_{\text{total}} = 16$ exactly. As a limiting-case check: if CO₂ were mistakenly drawn with single bonds and 3 lone pairs per oxygen (the "before Step 5" picture), the electron count would still total 16, but carbon would have an incomplete octet — showing that a correct electron count is necessary but not sufficient; the octet check in Step 5 is what fixes the actual bonding pattern.

Example 2 — Formaldehyde, CH₂O (asymmetric atoms, verifying every octet individually) Step 1. Valence electrons: C gives 4, each of 2 H gives 1, O gives 6.

N_{\text{total}} = 4 + 1 + 1 + 6 = 12

Step 2. Carbon is central (hydrogen can never be central, since it forms only one bond); connectivity is H–C(–H)–O, i.e., carbon bonded to two hydrogens and one oxygen. Step 3. Three single bonds (C–H, C–H, C–O) use 6 electrons.

N_{\text{remaining}} = 12 - 2(3) = 6

Step 4. Hydrogens need no lone pairs (a single bond already gives each H its duet of 2 electrons). Place all 6 remaining electrons (3 lone pairs) on oxygen: 1 bond + 3 lone pairs = 2 + 6 = 8 electrons around oxygen. Step 5. Check carbon: 3 bonds (2 C–H + 1 C–O) × 2 electrons = 6 electrons around carbon — short of the octet, and no lone pairs remain anywhere else to shift, except on oxygen. Convert one of oxygen's lone pairs into a second C–O bond:

\text{H}_2\text{C=O}

Now carbon has 2 single bonds (to H) + 1 double bond (to O) = 4 bonding regions worth 2+2+4 = 8 electrons around it, and oxygen has 1 double bond (4 electrons) + 2 lone pairs (4 electrons) = 8 electrons. Step 6. Ledger check:

\underbrace{2+2}_{\text{2 C–H bonds}} + \underbrace{4}_{\text{1 C=O double bond}} + \underbrace{4}_{\text{2 lone pairs on O}} = 4 + 4 + 4 = 12

This matches $N_{\text{total}} = 12$ . Limiting case: if oxygen were (incorrectly) given a triple bond instead, carbon would have 5 bonding pairs' worth of electrons (10), blowing past the octet — confirming that the octet ceiling on carbon, not just the raw electron total, forces the double bond rather than a triple bond here.

Limiting cases & common traps

Situation	What actually happens	Why
Hydrogen "needs an octet"	Hydrogen only ever gets 2 electrons (a duet)	Its valence shell is 1s, capacity 2, not 8
Boron in BF₃	Boron ends with only 6 electrons around it	Boron has 3 valence electrons and commonly stays electron-deficient rather than forming a double bond that would strain B–F geometry
Sulfur in SF₆	Sulfur has 12 electrons around it (expanded octet)	Period-3-and-beyond atoms have accessible d-orbital-like flexibility and larger radii, permitting more than 4 bonding regions
Odd-electron species (e.g., NO, ·NO₂)	One atom is left with an unpaired electron	The total valence-electron count is odd, so a full pairing into octets for every atom is mathematically impossible
Multiple bonds "used up" too early	Central atom ends up over-full (more than 8 electrons) if a double bond is added without necessity	Always try single bonds and lone-pair completion first; only promote to multiple bonds if the central atom's octet is otherwise unmet

The chemist behind it

Gilbert N. Lewis introduced the dot-and-line notation for electron pairs in a 1916 paper, proposing that a chemical bond is literally a shared pair of electrons sitting between two atomic cores — a radical departure from the purely ionic bonding pictures dominant at the time. He also framed the octet rule, noting that atoms in stable compounds tend to surround themselves with eight valence electrons, mirroring the electron counts of noble gases. Lewis structures remain the fastest hand-drawn tool for predicting connectivity before moving on to refinements like Formal Charge, Resonance Structures for delocalized electron systems, and ultimately VSEPR and Molecular Shapes for three-dimensional geometry.

Origins · 1 minGilbert N. Lewis and how this was discovered

Gilbert N. Lewis

1875–1946 · Proposed the shared electron-pair bond and the octet rule in 1916, reshaping how chemists picture covalent bonding

Ask the tutor: How do I decide which resonance structure (or structures) is the "best" representation once a molecule has more than one valid Lewis structure?

Why does hydrogen never appear as a central atom in a Lewis structure?

Hydrogen has only a 1s orbital available for bonding, which can hold at most 2 electrons total. It can therefore form exactly one covalent bond and never has room for a second connection, so it can only ever sit on the outside of a structure, attached to one neighbor.

Does a higher bond order always mean a stronger, shorter bond?

Within a comparable pair of atoms, yes: triple bonds are shorter and stronger than double bonds, which are shorter and stronger than single bonds, because more shared electron pairs pull the two nuclei closer and require more energy to separate. Comparisons across different pairs of elements (say, a C–O single bond versus an N=N double bond) need actual bond-energy data rather than bond order alone.

How do you decide whether a molecule needs an expanded octet?

Count the total valence electrons first, then try to build single bonds plus lone pairs (and standard multiple bonds) using only 8 electrons per atom. If that count cannot possibly absorb every electron without overfilling terminal atoms or leaving electrons stranded, and the central atom is in period 3 or beyond, allow it to exceed 8. Period-2 atoms (B, C, N, O, F) never take expanded octets.

Why do chemists bother with Lewis structures if they are such a simplified picture?

They correctly predict connectivity, bond order, and the placement of lone pairs for the overwhelming majority of simple molecules, using nothing more than arithmetic. That connectivity is exactly the input needed for more advanced tools such as formal charge analysis, resonance averaging, and VSEPR-based shape prediction, so getting the dot structure right first pays off at every later stage.