The Madelung Constant

The Madelung constant is far from a historical curiosity confined to introductory ionic-bonding problem sets — it is a load-bearing quantity throughout modern solid-state and computational chemistry. Every density-functional or force-field simulation of an ionic or polar crystalline material (battery cathode materials like LiCoO₂ or LiFePO₄, perovskite solar cell absorbers, zeolite catalysts, ceramic oxides) must handle exactly the same conditionally-convergent long-range electrostatic sum, and does so using descendants of Ewald's method — Ewald summation and its modern relative, the particle-mesh Ewald (PME) algorithm, are workhorses of molecular dynamics software used to simulate everything from protein folding in explicit ionic solvent to the stability of battery electrolytes. In materials design, comparing Madelung constants across candidate crystal structures (rock salt vs. zinc blende vs. wurtzite for a given 1:1 compound, or rutile vs. anatase for TiO₂-type stoichiometries) is a first, cheap screening step before expensive full quantum calculations, because it immediately reveals which packing geometry offers the larger purely-electrostatic stabilization — a major driver, though not the only one, of which polymorph a real compound adopts. Mineralogists and geochemists use Madelung-based lattice energy estimates to model the thermodynamic stability of silicate and oxide minerals under the immense pressures found deep in Earth's mantle, where structure type can change (a change in coordination number) and the associated jump in Madelung constant helps explain seismic discontinuities. The concept also underlies the design and interpretation of solid electrolytes for lithium and sodium batteries, where the Madelung potential at a given ionic site (the electrostatic potential generated by the rest of the lattice, closely related to the same sum) helps determine how easily an ion can hop between sites — a high local Madelung stabilization at the resting site versus a shallower one at the transition state is part of what governs ionic conductivity. Finally, the constant is a staple of X-ray and neutron crystallography teaching and research software, where lattice-energy and Madelung-potential calculators are built into standard crystal structure analysis packages, letting a researcher instantly estimate the electrostatic cohesive energy of any newly solved ionic structure.

Ask the tutor: How does adding the Born repulsion term to the Madelung electrostatic energy let us actually predict the equilibrium lattice spacing and lattice energy in the Born–Landé equation?

Start from what you already know: two isolated ions of charges $z_+e$ and $z_-e$ separated by distance $r$ interact with Coulomb energy $u(r) = \dfrac{z_+z_-e^2}{4\pi\varepsilon_0 r}$, negative for oppositely charged ions (stabilizing) and positive for like charges (destabilizing). This is the single ion-pair potential. Now place that ion inside an infinite crystal. It no longer feels just one partner — it feels every other ion in the lattice, at a whole spectrum of distances, alternating in sign as you move outward through shells of like and unlike neighbors. Picture a single Na⁺ ion sitting at the origin of the rock-salt lattice. Its six nearest neighbors, at distance $r$, are all Cl⁻ — attractive. The next shell, twelve Na⁺ ions at distance $r\sqrt2$, is repulsive. The next, eight Cl⁻ at distance $r\sqrt3$, is attractive again. And so on, forever, with the number of ions in each shell growing (the crystal is three-dimensional, so the surface area of a shell of radius $R$ grows as $R^2$) even as the strength of each individual interaction falls as $1/R$. The net energy per reference ion is a sum over the whole crystal:

This sum is what the Madelung constant packages. Because every term scales as $1/r$ and the ionic radii and charges are the same throughout an ideal binary lattice, we can factor out the nearest-neighbor distance $r_0$ and the charge product $|z_+z_-|e^2$, leaving a pure geometric series:

where all the crystallography is hidden inside $A$, the Madelung constant. This is the entire conceptual move: separate "how strong is one bond" (Coulomb's law, known) from "how much does the surrounding lattice help or hurt" (pure geometry, encoded in $A$). The subtlety — and the reason this deserves careful treatment rather than being obvious — is that the sum defining $A$ does not converge absolutely. Summing shell by shell (spheres of increasing radius) gives a different partial-sum trajectory than summing layer by layer in cubes, and the conditionally convergent series only settles to the "correct" physical value when summed in a way that respects the overall charge neutrality of the growing region (each partial sum must itself be electrically near-neutral, or correction terms are needed). This is why sophisticated summation techniques — Evjen's method, Ewald summation — were developed specifically to tame this sum reliably.

Notice the key structural fact embedded in the picture: the coordination shells alternate in sign (attractive, repulsive, attractive, …) and the coefficients (6, 12, 8, 6, 24, …) are the crystallographic coordination numbers of the rock salt lattice — pure geometry, nothing about sodium or chlorine specifically.

The defining equation of the Madelung constant, generalized to any ion $i$ in a lattice of point charges $z_j e$ at distances $r_{ij}$ from it, is

where:

• $A$ is the (dimensionless) Madelung constant for that structure,
• $z_i, z_j$ are the integer ionic charges (in units of $e$) of the reference ion and every other ion $j$,
• $r_{ij}$ is the distance from the reference ion to ion $j$,
• $r_0$ is the reference (nearest-neighbor) distance, used to normalize all other distances.

Reading: each term contributes $+r_0/r_{ij}$ if ion $j$ has the same sign of charge as the reference (repulsive, destabilizing) and $-r_0/r_{ij}$ if opposite (attractive, stabilizing); the sum runs over the entire infinite lattice, and $A$ is what remains after you strip out the charge magnitudes and the absolute length scale, leaving pure geometry. The full lattice electrostatic energy per mole of ion pairs, feeding directly into the Born-Lande Equation, is

where:

• $N_A$ is Avogadro's number (so the energy is per mole of formula units),
• $A$ is the structure-specific Madelung constant derived above,
• $z_+, z_-$ are the cation and anion charges,
• $r_0$ is the equilibrium nearest-neighbor separation,
• the leading minus sign reflects net stabilization (an ionic lattice is lower in energy than separated gas-phase ions).

This is the purely electrostatic (attractive) part of the lattice energy; a full treatment (the Born–Landé equation) adds a short-range Born repulsion term to prevent collapse, but that repulsion is not part of the Madelung sum itself — the Madelung constant concerns only the long-range Coulombic bookkeeping.

Example 1 — Foundational: verifying the CsCl constant from its coordination geometry. The cesium chloride structure is simple cubic, with each ion at the center of a cube of eight oppositely-charged ions (coordination number 8, not 6 as in rock salt). Its accepted Madelung constant is $A_{\text{CsCl}} = 1.762675$. Let's check this is at least qualitatively sensible against the naive first-shell term. The first shell contributes $8/1 = 8$ (fully attractive, 8 nearest unlike neighbors at distance $r_0$) — already far above the converged value of $1.76$, exactly as expected, because the second shell (6 like-charged ions at distance $r_0\sqrt{4/3} \approx 1.1547\,r_0$, since the cube diagonal geometry gives next-neighbor distance $2r_0/\sqrt3$) is strongly repulsive and close by, subtracting heavily:

Already this two-shell partial sum has dropped from 8 to 2.80, illustrating the same rapid, oscillating convergence pattern seen for NaCl, settling eventually (via Evjen or Ewald summation over further shells) to $1.7627$. The key qualitative lesson: CsCl's higher coordination number (8 vs. 6) does not simply mean a proportionally larger Madelung constant — $1.7627$ is barely $0.9\%$ larger than NaCl's $1.7476$, because the higher near-neighbor attraction is substantially offset by closer repulsive shells in the more compact simple-cubic packing.

Example 2 — Applied: using the Madelung constant to rank predicted lattice energies of two real salts. Compare the purely electrostatic (Madelung) part of the lattice energy of MgO (rock salt structure, $z_+ = +2$, $z_- = -2$, $r_0 = 210$ pm) against NaCl (rock salt, $z_+=+1$, $z_-=-1$, $r_0 = 282$ pm), using

Both share the same Madelung constant, $A = 1.747565$, because both crystallize in the rock-salt structure type — this is precisely the point of separating geometry ($A$) from chemistry ($z_+, z_-, r_0$). Compute the prefactor $\dfrac{e^2}{4\pi\varepsilon_0} = 2.307 \times 10^{-28}\ \text{J}\,\cdot\,\text{m}$. For NaCl ($r_0 = 2.82\times10^{-10}$ m, $|z_+z_-| = 1$):

For MgO ($r_0 = 2.10\times10^{-10}$ m, $|z_+z_-| = 4$):

The huge difference — MgO's electrostatic lattice energy is roughly $5.4\times$ more negative than NaCl's — comes almost entirely from the $|z_+z_-| = 4$ versus $1$ charge-product factor and the smaller $r_0$, not from the Madelung constant, which is identical for both. This is exactly why MgO has a far higher melting point (2852 °C) than NaCl (801 °C): the charge and size terms dominate, while $A$ merely sets the common geometric multiplier both salts share by virtue of adopting the same structure. (These are Madelung-only estimates; the full Born–Landé lattice energies, which subtract a short-range repulsion term, are somewhat smaller in magnitude — roughly $-787$ kJ/mol for NaCl and $-3795$ kJ/mol for MgO experimentally — but the ratio and ranking are already correctly predicted by the electrostatic term alone.)

Example 3 — Exam/competition level: why the fluorite constant depends on the convention you choose, and how to convert between them. The fluorite structure (CaF₂-type: Ca²⁺ in an fcc array, F⁻ filling all tetrahedral holes, coordination numbers 8 for Ca²⁺ and 4 for F⁻) has a Madelung constant frequently quoted as $A = 5.03878$ or alternatively $A = 2.51939$, and students are routinely tripped up by this factor-of-2 discrepancy. Resolve it. The ambiguity arises because the Madelung sum can be normalized either per formula unit (treating the reference charge product as $|z_{\text{Ca}} \cdot z_{\text{F}}| = |(+2)(-1)| = 2$, summed over both cation and anion sublattices and referenced to one CaF₂ unit) or per ion pair with a symmetrized charge convention. The two commonly tabulated fluorite constants differ by exactly a factor of 2:

The resolution is to always write the lattice energy formula with an explicit, self-consistent charge-product term and confirm the reference distance used. If we define $r_0$ as the shorter Ca–F distance (the actual nearest-neighbor separation, $\approx 236$ pm) and use $|z_+z_-| = 2$ explicitly in the denominator formula,

then the correct paired constant is $A = 5.03878$, because this convention already has the "extra" charge-multiplicity folded into how the reference sum was constructed (each Ca²⁺ has 8 nearest F⁻, and this sum was built per formula unit including both cation and anion contributions symmetrically). Using $A = 2.51939$ with the same formula and same $|z_+z_-|=2$ would double-count and understate the lattice energy by exactly half.

Check by limiting case: if $z_+$ and $z_-$ were both reduced to a hypothetical $\pm1$ system with the same fluorite geometry (an unphysical but mathematically valid limit), the geometric constant $A$ would not change at all — only the charge-product prefactor would shrink from 2 to 1 — confirming that $A$ is a pure function of geometry, wholly decoupled from the ionic charges, and that all the "which constant convention" confusion lives entirely in bookkeeping of $z_+z_-$, never in $A$ itself. This is the single most important exam trap: always check what charge product and what reference distance a tabulated Madelung constant assumes before plugging it into an energy formula.