Ionic Bonding

Open almost any introduction to this topic and you find a diagram: a sodium atom hands an electron to a chlorine atom, arrow drawn, both look pleased, and the resulting pair is labelled an ionic bond.

Two things about that picture are misleading, and they are the two things worth getting right.

The first is that no such pair exists. There is no NaCl molecule in a grain of salt. There is a crystal in which every sodium ion is surrounded by six chloride ions and every chloride ion by six sodium ions, extending in all directions until the crystal ends. The formula NaCl states a ratio, not a partnership.

The second is that the transfer, taken by itself, does not pay for itself.

The unit the formula describes has a name — a formula unit — and it is a bookkeeping device rather than a physical object. A single visible grain of table salt contains something like 10¹⁸ ions arranged in one continuous lattice, and asking which chloride belongs to which sodium is as meaningless as asking which brick in a wall belongs to which other brick. Everything else on this page follows from taking that seriously.

The transfer runs uphill

Removing the outer electron from a gaseous sodium atom costs 496 kJ per mole. Attaching that electron to a gaseous chlorine atom releases 349 kJ per mole.

496 − 349 = +147 kJ/mol

Making a mole of separated Na⁺ and Cl⁻ ions from separated atoms is endothermic. If the story were really that atoms want full shells and are prepared to swap electrons to get them, sodium chloride would not form at all.

What makes it happen is what comes next. Once the ions exist, they attract one another, and assembling a mole of gaseous Na⁺ and Cl⁻ into a crystal lattice releases 787 kJ. That figure — the lattice energy — dwarfs everything else in the calculation, and it is the actual reason ionic compounds exist.

Born and Haber's accounting, in five steps

The full balance sheet is a Born–Haber cycle, and running one is the clearest demonstration available that lattice energy is doing the work. Every step is an experimentally measured quantity except the lattice energy, which is what the cycle is usually used to find.

Forming solid sodium chloride from sodium metal and chlorine gas:

Adding the column: 107 + 122 + 496 − 349 − 787 = −411 kJ/mol, which is the measured enthalpy of formation of sodium chloride. Three of the five steps cost energy. The single step that returns it returns nearly twice what all three consumed.

This is why "ionic bonding is the transfer of electrons" is only half a definition. The bonding is the electrostatic attraction that follows the transfer, acting between every ion and every other ion in the crystal, and it has no direction and no preferred partner.

Charge and distance set the melting point

The strength of that attraction follows a relationship simple enough to reason with:

lattice energy ∝ (charge on the cation × charge on the anion) ÷ (distance between them)

Double both charges and the energy roughly quadruples. Shrink the ions and it rises further.

Magnesium oxide is the standard demonstration. It is built like sodium chloride, in the same arrangement, but from 2+ and 2− ions that are also smaller. Its lattice energy is about 3,791 kJ/mol against sodium chloride's 787 — nearly five times greater. The consequence shows up in the most easily measured property there is: sodium chloride melts at 801 °C, magnesium oxide at 2,852 °C. Magnesium oxide's refractoriness is why it lines furnaces.

The arrangement itself depends on relative ion sizes. Sodium chloride adopts six-to-six coordination. Cesium chloride, whose cation is much larger, fits eight chloride ions around each cesium and eight cesium around each chloride. Zinc sulfide, with a small cation, manages only four. The same compound type produces three different crystal structures purely from geometry.

Writing the formula: criss-cross, then cancel

An ionic compound is electrically neutral overall, so the formula is whatever ratio of ions makes the charges cancel. The mechanical method: take the size of each charge, ignore the signs, and write each as the other's subscript.

• Al³⁺ with O²⁻ → the 3 becomes oxygen's subscript, the 2 becomes aluminum's: Al₂O₃. Check: 2 × (+3) = +6, and 3 × (−2) = −6.
• Ca²⁺ with PO₄³⁻ → Ca₃(PO₄)₂. Polyatomic ions take brackets before the subscript, because the subscript must multiply the whole group.

The step everyone forgets is cancelling. Mg²⁺ with O²⁻ criss-crosses to Mg₂O₂, which is wrong. The ratio must be reduced to its simplest terms: MgO. The same applies to Ca²⁺ with S²⁻, giving CaS rather than Ca₂S₂. Only formulas of molecular substances keep unreduced subscripts, and ionic compounds have no molecules to keep them in.

The charges themselves, and how to predict them, are covered under ions, cations and anions.

Brittle, insulating and soluble — with caveats on all three

The lattice explains the properties, and the explanations are more interesting than the properties.

Brittleness is the best of them. Metals deform because their structure tolerates layers sliding past one another. An ionic crystal does not: displace one layer by a single ion's width and every ion suddenly faces a neighbour of the same charge. The attraction that held the crystal together becomes repulsion along the whole plane at once, and the crystal splits cleanly. Ionic solids are hard and brittle for the same reason, not despite it.

Electrical behaviour follows from mobility. A solid ionic compound does not conduct, because its charges are locked in place. Melt it, or dissolve it in water, and the ions are free to move — so molten sodium chloride conducts well, while the solid is an insulator.

Solubility is where the textbook line is weakest. Plenty of ionic compounds are barely soluble at all: silver chloride, barium sulfate and calcium carbonate are ionic and stay solid in water. Dissolving requires the energy released by water molecules surrounding the ions to approximately match the lattice energy being dismantled, and for strongly bound lattices it does not come close.

Even sodium chloride, the standard example of a soluble salt, is a surprise on close inspection. Its enthalpy of solution is +3.9 kJ/mol — very slightly endothermic. Salt dissolves not because dissolving releases energy but because it increases disorder; the process is driven by entropy, and a glass of water gets imperceptibly colder as salt goes into it.

Aluminum chloride, which did not read the rule

The rule "metal plus non-metal gives an ionic compound" has a well-defined failure mode, and knowing it is what separates understanding the model from reciting it.

Aluminum chloride is made from a metal and a non-metal. It should be a high-melting crystalline salt. Instead it sublimes at about 180 °C and, in the vapour, exists as discrete Al₂Cl₆ molecules — behaviour that is unmistakably covalent. Beryllium chloride does much the same.

The explanation is polarisation, set out in Fajans' rules. A small cation with a high charge pulls hard on the electron cloud of a neighbouring anion, and if the anion is large and easily distorted, the cation drags enough electron density back toward itself that the bond acquires substantial covalent character. Al³⁺ is small and triply charged, which makes it about as polarising as a common cation gets.

The general lesson is worth carrying beyond this example. Purely ionic bonding is a limiting case, not a category. Every real bond between unlike atoms sits somewhere on a continuum between complete transfer and equal sharing, and where it sits is governed by the electronegativity difference between the two elements. Sodium chloride is close to the ionic end. Aluminum chloride is not, and no amount of consulting the periodic table's colour scheme will reveal that.