The Ionic Model

The ionic model is not a historical curiosity; it is the default first-pass tool anywhere an inorganic chemist needs a fast, physically grounded number for a salt's stability. In materials science, Born–Landé-type estimates guide the selection of solid electrolytes and battery cathode materials — predicting whether a candidate lattice (say, a doped garnet or NASICON-type ionic conductor) will hold together thermally while still allowing the target ion (Li⁺, Na⁺) to hop between sites, a balance between lattice energy and activation barrier for diffusion. Cement and refractory ceramics design leans directly on the MgO-vs-NaCl style comparison above: engineers choosing furnace linings or high-temperature structural ceramics use charge and radius arguments to anticipate melting points and mechanical hardness before a single sample is synthesized. In geochemistry and mineralogy, lattice-energy systematics (via Kapustinskii-type generalized Madelung sums, which sidestep the need to know a mineral's exact crystal structure) predict which cations substitute for which in silicate and oxide minerals under the pressures found deep in planetary mantles — radius-ratio reasoning underlies models of how Fe²⁺, Mg²⁺, and Ca²⁺ partition among mantle minerals. Pharmaceutical salt selection is a more surprising modern application: when a drug candidate is a weak base or acid, formulators deliberately choose a counter-ion (chloride, mesylate, besylate) whose lattice energy with the drug cation/anion optimizes both crystallinity (for manufacturing and shelf stability) and aqueous solubility (for bioavailability) — a direct, practical descendant of Born–Landé reasoning, now automated in computational salt-screening pipelines. Finally, in fundamental electronic structure theory, the ionic model is the zeroth-order starting point that density functional theory and other first-principles methods are benchmarked against: when a modern DFT calculation reproduces the Born–Haber lattice energy of NaCl to within a kJ/mol, that agreement is itself evidence that the simple electrostatic picture captured the dominant physics correctly a century earlier, with quantum corrections refining rather than overturning it. The same logic scaffolds the more refined Coulombic Attraction and the Ion-Pair Potential treatments used in molecular dynamics force fields for everything from zeolite catalysis to biomolecular electrostatics.

Ask the tutor: How does replacing the Born power-law repulsion with an exponential (Born–Mayer) form change the predicted lattice energy, and why might that form be physically more realistic?

Start from the electronic picture already familiar from atomic properties: an element with low ionization energy (an alkali or alkaline-earth metal) readily gives up one or two electrons, while an element with a strongly negative electron affinity (a halogen, oxygen) readily accepts them. The ionic model asserts that once this electron transfer has happened, the resulting ions can be treated as classical objects — rigid charged spheres with well-defined radii — and that essentially all of the bonding energy comes from the electrostatic field these spheres set up around themselves. There is no directional bonding, no orbital overlap in the covalent sense; the model is isotropic by construction. A given cation attracts every anion in the crystal, not just its nearest neighbors, and repels every other cation, out to infinite range (formally; in practice the sum converges). Two competing requirements set the crystal structure. First, electrostatics favors packing as many oppositely charged neighbors around each ion as geometrically possible, since each additional nearest-neighbor contact adds a large negative (stabilizing) energy term. Second, geometry limits how many anions can be squeezed around a cation without the anions' own hard spheres beginning to touch and repel each other — this is the origin of the radius-ratio rule. Define $\rho = r_+/r_-$ (cation radius over anion radius, using the smaller ion as reference when it is the cation, as is usual for simple binary salts). For a given coordination geometry there is a minimum radius ratio below which the anions surrounding the cation would have to overlap:

• Ρ < 0.155: no stable geometry (linear, coordination number 2, marginal)
• 0.155–0.225: trigonal planar, CN = 3
• 0.225–0.414: tetrahedral, CN = 4 (e.g., ZnS zinc-blende/wurtzite)
• 0.414–0.732: octahedral, CN = 6 (e.g., NaCl rock salt)
• 0.732–1.0: cubic, CN = 8 (e.g., CsCl)

These thresholds come from simple trigonometry: the critical ratio is the one at which the anions are simultaneously touching the central cation and touching each other, so pushing $\rho$ any lower would force anion–anion interpenetration. The rule is only a first approximation — many salts sit close to a boundary and adopt either neighboring structure — but it correctly predicts the crossover from NaCl-type (6:6) to CsCl-type (8:8) structures as radius ratio increases, and from NaCl-type to zinc-blende/wurtzite (4:4) as it decreases. Once the structure type (and hence the specific three-dimensional arrangement of point charges) is fixed, the total lattice energy is obtained by summing the Coulomb interaction of one reference ion with every other ion in the infinite crystal — a geometric series in inverse distances that depends only on the structure type, not on the actual size of the ions. This sum, multiplied by the appropriate charges and by $1/r_0$ (the equilibrium interionic distance), is the Madelung energy. It is opposed by a steep short-range repulsion arising when the electron clouds of adjacent ions begin to overlap (a quantum-mechanical, Pauli-exclusion effect, but modeled here purely empirically as a $1/r^n$ or exponential term). The balance of these two — long-range attraction, short-range repulsion — is exactly the same competition that appears in any interatomic potential, and it is what the ionic model shares structurally with the continuous ion-pair potential treated separately.

The Coulomb potential energy between two point charges $z_+e$ and $z_-e$ separated by distance $r$ is

where: $z_+,z_-$ are the (signed) ionic charges in units of $e$ (so $z_-<0$, making $V<0$, attractive), $e$ is the elementary charge, $\varepsilon_0$ is the vacuum permittivity, and $r$ is the center-to-center distance. Reading this: for oppositely charged ions the energy is negative (bonding) and falls off as $1/r$ — long-ranged compared to covalent bonding, which is why every ion in a crystal "feels" every other ion, not just its neighbors. Summing this interaction over an entire crystal lattice, keeping track of the alternating signs and the multiplicities of ions at each shell of neighbors, produces a numerical constant, the Madelung constant $A$, defined through

where: $N_A$ is Avogadro's number (energy is quoted per mole of ion pairs), $A$ is the dimensionless Madelung constant (≈1.748 for rock salt, ≈1.763 for CsCl, ≈1.638 for zinc blende), and $r_0$ is the equilibrium (experimental) interionic separation. Reading this: $A$ packages all of the geometry of the crystal — every attractive and repulsive neighbor shell, out to infinity — into one number that depends only on structure type, not on which specific ions occupy the lattice. Because point charges alone would collapse to $r=0$, a short-range repulsive term is added, empirically modeled as a Born (power-law) form,

where: $B$ is an empirical constant fixed by requiring force balance at $r_0$, and $n$ is the Born exponent, an integer typically between 5 and 12 that increases with the "hardness"/compressibility of the ions' electron clouds (larger, more polarizable ions have smaller $n$; small, hard ions like F⁻ or Li⁺ have larger $n$). Reading this: this term is essentially a phenomenological stand-in for Pauli repulsion between overlapping closed electron shells — it has no classical electrostatic origin, but it must fall off much faster than $1/r$ (large $n$) so that it only matters at short range. Adding these and minimizing with respect to $r$ (setting the net force to zero at $r=r_0$) yields the Born–Landé equation for the lattice energy:

where: all symbols as above, and the factor $(1-1/n)$ is the reduction of the pure Coulomb energy caused by the repulsive term's contribution at equilibrium. Reading this: the lattice energy is almost the pure electrostatic (Madelung) energy, discounted by a small, $n$-dependent correction (typically 10–20%) that accounts for short-range repulsion; because $n$ enters only as $1/n$, the lattice energy is not very sensitive to the exact value chosen for $n$ — a useful robustness property of the model.

Example 1 — Foundational: lattice energy of NaCl from the Born–Landé equation. Take rock-salt-structure NaCl: Madelung constant $A = 1.7476$, charges $z_+ = +1$, $z_- = -1$, equilibrium interionic distance $r_0 = 2.82\times10^{-10}\ \text{m}$ (the sum of ionic radii, $r_{\text{Na}^+}=1.02$ Å + $r_{\text{Cl}^-}=1.81$ Å ≈ 2.83 Å, close enough to the crystallographic value), and Born exponent $n=8$ (typical for ions of Ne/Ar-like electron configuration, an average of the individual Na⁺ (n≈7) and Cl⁻ (n≈9) values as recommended by Kapustinskii-type averaging). Compute the pure Coulomb (Madelung) term first:

Using $e^2/4\pi\varepsilon_0 = 2.307\times10^{-28}\ \text{J}\,\cdot\,\text{m}$ (a standard combination worth memorizing), $z_+z_- = -1$:

Numerator: $6.022\times10^{23}\times1.7476 = 1.0524\times10^{24}$; times $2.307\times10^{-28}$ gives $2.428\times10^{-4}$.

(the overall sign is negative because $z_+z_-=-1$, restoring the attractive sign properly). Now apply the Born correction $(1-1/n) = (1-1/8) = 0.875$:

This compares well with the Born–Haber experimental value for NaCl of about $-787\ \text{kJ mol}^{-1}$; the ~4% discrepancy is typical of the residual approximations (hard-sphere radii, neglect of van der Waals attraction and zero-point energy, and a crude single averaged $n$). Limiting check: if repulsion were switched off entirely ($n\to\infty$, so $1/n\to0$), $U_L \to U_{\text{Coulomb}} = -861\ \text{kJ mol}^{-1}$ — the raw Madelung energy, more negative (too strongly bound) than reality, exactly as expected since removing repulsion removes the only thing preventing the lattice from being overbound.

Example 2 — Applied: comparing NaCl and MgO to explain relative melting points and hardness. Both salts adopt the rock-salt structure (same Madelung constant, $A=1.7476$), so the entire difference in lattice energy must come from charges and radii. For MgO, $z_+z_- = (+2)(-2) = -4$ — four times the magnitude of NaCl's $z_+z_-=-1$ — and $r_0(\text{MgO}) \approx 2.10\times10^{-10}\ \text{m}$ (smaller than NaCl's 2.82 Å, since Mg²⁺ and O²⁻ are both smaller than Na⁺ and Cl⁻ respectively... in fact O²⁻ is comparable to Cl⁻, but Mg²⁺ is markedly smaller than Na⁺). Using $n=7$ (appropriate for the smaller, harder Mg²⁺/O²⁻ pair):

The numerator is four times the NaCl numerator computed above (since the only change relative to NaCl's numerator is the factor of 4 from the charges): $4\times2.428\times10^{-4} = 9.712\times10^{-4}$. Dividing by the smaller $r_0$:

Apply the Born correction, $(1-1/7)=0.857$:

This is reassuringly close to the Born–Haber value derived above (−3844 kJ mol⁻¹) — the small remaining gap (~3%) is again the fingerprint of MgO's slightly greater covalent/polarization character relative to a perfectly ionic point-charge picture. The physical punchline: MgO's lattice energy is roughly five times more negative than NaCl's, even though $r_0$ only shrank by about 25%. The dominant effect is the quadrupling of $z_+z_-$ (going from singly to doubly charged ions), since lattice energy scales linearly with $|z_+z_-|$ but only as $1/r_0$ with distance. This is exactly why MgO melts at 2852 °C while NaCl melts at just 801 °C, and why MgO is used as a refractory furnace lining while NaCl is not: doubling both ionic charges is a far bigger energetic lever than modest changes in ionic radius. Limiting check: if MgO's ions carried charges of $\pm1$ instead of $\pm2$ (holding $r_0$ fixed), the lattice energy would drop by a factor of exactly 4 (since $U_L \propto z_+z_-$), landing near $-1156\ \text{kJ mol}^{-1}$ — comparable to, even somewhat larger than, NaCl's, entirely consistent with MgO's smaller $r_0$ being the only remaining difference in that hypothetical.

Example 3 — Exam/competition level: how sensitive is the lattice energy to the choice of Born exponent $n$, and what does this reveal about the model's internal consistency? Consider CaF₂ (fluorite structure, Madelung constant $A = 2.519$, $z_+ = +2$, $z_- = -1$ so $z_+z_- = -2$, $r_0 \approx 2.36\times10^{-10}\ \text{m}$, the Ca²⁺–F⁻ contact distance). Suppose a student is unsure whether to use $n=7$ (appropriate for the small, hard F⁻ anion) or $n=10$ (appropriate if one mistakenly averaged in a larger, more polarizable anion). Quantify the resulting spread in predicted lattice energy. First compute the pure Coulomb term (independent of $n$):

Numerator: $6.022\times10^{23}\times2.519 = 1.5169\times10^{24}$; times 2 gives $3.0338\times10^{24}$; times $2.307\times10^{-28}$ gives $6.998\times10^{-4}$.

Now apply the Born factor for each candidate $n$: For $n=7$: $(1-1/7) = 0.8571$, giving

For $n=10$: $(1-1/10) = 0.9000$, giving

The difference between the two choices is

a spread of about 5% of the total lattice energy — noticeable, but small compared to the near-total collapse the model would suffer if $n$ were wrong by an order of magnitude. This is the payoff of the $(1-1/n)$ functional form derived above: because $n$ only ever enters as $1/n$, even a 40% uncertainty in the physically murky Born exponent (7 vs. 10, a real ambiguity, since tabulated $n$ values are only guidelines based on electronic configuration) produces just a ~5% uncertainty in the final energy. The experimental Born–Haber lattice energy for CaF₂ is about $-2611\ \text{kJ mol}^{-1}$, sitting almost exactly between the two estimates — consistent with the real compromise value $n\approx 8$–9 used in practice for a Ca²⁺/F⁻ pair of mixed hardness. Limiting check: as $n\to\infty$ (an infinitely hard, infinitely short-range repulsive wall — the true hard-sphere limit), $(1-1/n)\to1$ and $U_L\to U_{\text{Coulomb}} = -2965\ \text{kJ mol}^{-1}$, the maximal (most negative) lattice energy obtainable from this radius and charge combination; this is the theoretical ceiling the Born-corrected value must always lie above (i.e., be less negative than).

The ionic model as a quantitative energetic theory is inseparable from Max Born, who — working with physicist Alfred Landé — formalized the lattice-energy calculation that bears their joint name. Born had already made his name in the emerging quantum theory (he would later receive the Nobel Prize in Physics in 1954 for the statistical interpretation of the wavefunction), but in 1918, while at the University of Frankfurt/Göttingen orbit of early 20th-century German theoretical physics, he turned classical electrostatics loose on the problem of crystal cohesion. The insight that a $1/r^n$ repulsive term, with $n$ empirically fit from crystal compressibility data, could stabilize an otherwise divergent Coulomb sum was itself a bridge between the old classical electrostatics and the emerging quantum-mechanical understanding of the Pauli exclusion principle, which would only later supply the microscopic justification for why overlapping closed shells repel steeply. Born's collaborator on the compressibility side, Landé, contributed the crucial link between the model's $n$ parameter and measurable bulk compressibility of ionic crystals — turning what could have remained an adjustable-parameter curve fit into a theory with independent experimental checks. The broader context is the German school of crystal-lattice theorists in the 1910s–1920s (including Erwin Madelung, whose 1918 paper gave the lattice summation constant its name), working at the boundary of X-ray crystallography (fresh off Bragg's and von Laue's determinations of real crystal structures) and the nascent quantum theory of matter. It is a nice historical irony that a model built entirely from classical electrostatics turned out to be such an excellent approximation precisely because quantum mechanical exchange and correlation effects — which the model ignores completely — happen to be small for the "ionic" end of the bonding spectrum.