Bond Order

What bond order actually counts

When two atoms approach, their atomic orbitals overlap and combine into molecular orbitals spread over both nuclei. Constructive overlap produces a bonding orbital, lower in energy than the original atomic orbitals, with electron density piled up between the nuclei where it can glue them together. Destructive overlap produces an antibonding orbital (marked with an asterisk, σ* or π*), higher in energy, with a node between the nuclei that actively pushes them apart. Every pair of atomic orbitals that combine makes exactly one bonding and one antibonding partner.

Electrons in bonding orbitals stabilize the molecule; electrons in antibonding orbitals destabilize it and roughly cancel a bonding pair. The bond order is the net tally of bonding minus antibonding, divided by two so that one shared pair counts as one bond:

Bond order = (bonding electrons − antibonding electrons) ÷ 2

For the hydrogen molecule H₂, both electrons drop into the single σ_1s bonding orbital: (2 − 0) ÷ 2 = 1, a normal single bond. For hypothetical He₂, the two extra electrons are forced into the σ*_1s antibonding orbital: (2 − 2) ÷ 2 = 0. With zero net bonding, He₂ has no reason to exist, which is exactly why helium is a monatomic gas. That single calculation — done before any experiment — tells you whether a diatomic molecule should form at all.

Filling the molecular-orbital diagram

To get a bond order you build the diagram and pour in valence electrons using the same three rules that govern atoms: the Aufbau principle (lowest energy first), the Pauli exclusion principle (two electrons per orbital, opposite spins), and Hund's rule (singly occupy degenerate orbitals before pairing). For second-row diatomics the order of the orbitals is, from the bottom up: σ_2s, σ*_2s, then either the π_2p pair below or above σ_2p, followed by π*_2p and finally σ*_2p.

There is a subtle but famous twist. For the lighter molecules Li₂ through N₂, s–p mixing pushes σ_2p above the π_2p orbitals. For O₂, F₂, and Ne₂ the mixing is weak and σ_2p drops below π_2p. The ordering does not change the bond order, but it changes which orbital the electrons enter — and that is precisely what reveals whether a molecule is magnetic. Take O₂: its final two electrons land one each in the two degenerate π* orbitals, parallel spins, by Hund's rule. Those two unpaired electrons make liquid oxygen paramagnetic — it visibly clings between the poles of a magnet. A naive Lewis double bond predicts a diamagnetic, all-paired molecule and gets this wrong; MO theory nails both the magnetism and the correct bond order of 2. That experiment converted a generation of chemists to the molecular-orbital picture.

Why the number predicts strength and length

Bond order is useful because it is tightly correlated with three measurable quantities. More net bonding pairs deepen the potential-energy well between the atoms, so the equilibrium separation (the bond length) shrinks and the energy needed to pull the atoms apart (the bond dissociation energy or bond strength) climbs. The stiffer well also makes the bond vibrate faster, raising the infrared stretching frequency you would measure by IR spectroscopy. Across the second-row homonuclear diatomics the trend is textbook-clean:

Molecule	Bond order	Bond length (pm)	Dissociation energy (kJ/mol)	Magnetism
H₂	1	74	436	diamagnetic
He₂	0	—	~0 (unbound)	—
Li₂	1	267	105	diamagnetic
B₂	1	159	289	paramagnetic
C₂	2	124	607	diamagnetic
N₂	3	110	945	diamagnetic
O₂	2	121	498	paramagnetic
F₂	1	142	159	diamagnetic
Ne₂	0	—	~0 (unbound)	—

Read down the table and the logic is unmistakable. Nitrogen sits at the peak: three full bonding pairs, the shortest bond, and one of the strongest bonds in all of chemistry at 945 kJ/mol. That ferocious triple bond is why N₂ makes up 78% of the air yet is almost chemically inert, why breaking it in the Haber–Bosch process demands ~450 °C and 200 atm over an iron catalyst, and why nitrogen-rich explosives release so much energy when they finally rearrange into that stable triple bond. Fluorine sits at the other extreme — order 1 but anomalously weak (159 kJ/mol) because lone-pair repulsion between the small, electron-crowded atoms further destabilizes the single bond, making F₂ violently reactive.

Fractional and odd-electron bond orders

Bond order does not have to be a whole number. Any species with an odd electron count, or with electrons delocalized over several atoms, gives a fraction — and those fractions match reality. Adding or removing one electron shifts the order by exactly ½, and you can predict the direction: pulling an electron out of a bonding orbital weakens the bond, while pulling one out of an antibonding orbital strengthens it.

Species	Change from O₂ / N₂	Bond order	Bond length (pm)
O₂⁺ (dioxygenyl)	remove an antibonding e⁻	2.5	112 (shorter, stronger)
O₂ (dioxygen)	reference	2.0	121
O₂⁻ (superoxide)	add an antibonding e⁻	1.5	133 (longer, weaker)
O₂²⁻ (peroxide)	add two antibonding e⁻	1.0	149
N₂ (dinitrogen)	reference	3.0	110
N₂⁺	remove a bonding e⁻	2.5	112 (longer, weaker)
NO (nitric oxide)	11 valence e⁻, one in π*	2.5	115

The oxygen series is a perfect ladder: O₂⁺ at 2.5 is shorter and stronger than O₂, while superoxide and peroxide — both important reactive oxygen species in biology — get progressively longer and weaker as antibonding electrons pile in. Superoxide's bond order of 1.5 is precisely why it is an unstable, damaging radical that cells must destroy with the enzyme superoxide dismutase.

Delocalization gives fractions of a different flavor. In benzene the six π electrons are smeared evenly over the ring, so each carbon–carbon bond is neither single nor double but an average bond order of about 1.5 — which is why all six C–C bonds are identical at 139 pm, intermediate between a typical single (154 pm) and double (134 pm) bond. The carbonate ion CO₃²⁻ spreads one double bond over three equivalent C–O linkages, giving each a bond order near 1.33 and three identical bond lengths, exactly as resonance predicts. Bond order is the bridge between the resonance cartoon and the single, real, delocalized structure.

Two ways to assign bond order

It is worth being clear that "bond order" is computed differently depending on which model you start from, even though the models usually agree on whole molecules:

• Lewis / valence-bond bond order. Simply count the shared electron pairs drawn between two atoms: one line is order 1, two lines order 2, three lines order 3. For molecules with resonance you average over the structures, which is how you get benzene's 1.5. This is fast and works for most organic molecules.
• Molecular-orbital bond order. Use (bonding − antibonding) ÷ 2 from the full MO diagram. This is the only model that correctly handles O₂'s paramagnetism, predicts that He₂ and Ne₂ do not exist, and assigns sensible orders to odd-electron radicals like NO. When the two methods disagree, the MO result is the physically honest one.

More sophisticated computational chemistry refines this further — Mayer, Wiberg, and natural-bond-orbital bond indices extract effective bond orders from a calculated wavefunction and often return non-integer values even for "ordinary" bonds, reflecting partial ionic character and electron correlation. But for teaching, prediction, and quick reasoning, the simple (bonding − antibonding) ÷ 2 rule captures the essential physics.

Where bond order earns its keep

• Predicting whether a molecule exists. A bond order of 0 means no bond — the cleanest possible explanation for why noble gases like helium and neon refuse to form diatomics.
• Reaction energetics. The colossal stability of N₂'s triple bond drives the thermodynamics of nitrogen fixation, combustion of nitrogen-containing fuels, and the energy released by explosives that form N₂.
• Spectroscopy. Higher bond order shifts the IR stretching frequency upward; analysts read bond order straight off a vibrational spectrum (the C≡N stretch near 2200 cm⁻¹ versus a C=N near 1650 cm⁻¹).
• Magnetism and materials. The unpaired antibonding electrons that fix O₂'s bond order at 2 also make it paramagnetic, a property exploited in oxygen sensors and MRI contrast chemistry.
• Biochemistry. The bond orders of superoxide (1.5) and peroxide (1.0) explain their reactivity as reactive oxygen species and the need for antioxidant enzymes.

Common misconceptions

• Bond order must be a whole number. Radicals (NO = 2.5) and delocalized systems (benzene ≈ 1.5) have genuine fractional orders that match measured bond lengths.
• Antibonding electrons "do nothing." They actively destabilize the molecule and cancel a bonding pair; that is the whole reason He₂ has bond order 0.
• More electrons always means a stronger bond. Only bonding electrons strengthen the bond; adding antibonding electrons (O₂ → O₂⁻ → O₂²⁻) lengthens and weakens it.
• Bond order equals bond strength in absolute kJ/mol. It correlates strongly within a family, but lone-pair repulsion makes F₂'s order-1 bond far weaker than H₂'s order-1 bond.
• Lewis and MO always agree. They agree on most molecules, but O₂'s magnetism shows the Lewis picture can be qualitatively wrong even when the order (2) comes out right.